Introduction

Top Abstract Introduction Methods Results Discussion References

Dynorphin A(1-17) is an endogenous opioid peptide that is distributed throughout the central nervous system (Khachaturian et al., 1982; Maysinger et al., 1982), suggesting that it might serve multiple regulatory functions (Smith and Lee, 1988). An abbreviated peptide, dynorphin A(1-13), was synthesized and determined to be equally potent as the natural substance in vitro (Goldstein et al., 1979). Although the effects of shorter residues have been examined (Chavkin and Goldstein, 1981; Corbett et al., 1982; Takemori et al., 1993; Yoshimura et al., 1982), the (1-13) amino acid fragment has been most frequently used by researchers to study the dynorphinergic pharmacological actions and biobehavioral effects, and it is the fragment used in the present study to characterize the effects of dynorphin in human subjects.

Previous research using in vitro assays indicates that dynorphin A(1-13) is a kappa-selective ligand. Its effects in guinea pig ileum are reduced after pretreatment with the kappa-preferring agonist ethylketocyclazocine (Huidobro-Toro et al., 1981; Oka et al., 1982) and demonstrate selective cross-tolerance to kappa agonists (Huidobro-Toro et al., 1981; Wüster et al., 1980, 1981; Yoshimura et al., 1982). In both rhesus monkey and human brain tissue, dynorphin exhibits >60-fold kappa receptor binding selectivity (Emmerson et al., 1994; Pfeiffer et al., 1981). The in vivo effects of dynorphin are also generally kappa-like, but not as reliably so, and some reports indicate this peptide may act through nonopioid (e.g., NMDA) mechanisms. Early studies found that centrally administered dynorphin produced both analgesia and paralysis (Herman, 1982; Herman and Goldstein, 1985; Przewlocki et al., 1983; Spampinato and Candeletti, 1985); these effects were not antagonized by naloxone (Long et al., 1988; Massardier and Hunt, 1989; Stevens and Yaksh, 1986; Walker et al., 1982a, 1982b). At subneurotoxic doses, dynorphin analgesia in thermal tests is often weak (Hayes et al., 1983; Piercey et al., 1982) or absent (Stevens and Yaksh, 1986; Tung and Yaksh, 1982; Walker et al., 1982a, 1982b) but evident in chemical and pressure tests (Hayes et al., 1983; Hooke et al., 1995; Kaneko et al., 1983; Nakazawa et al., 1985). These data suggest a kappa-opioid or nonopioid profile because kappa drugs are weak against heat stimuli but effective against other noxia (Tyers, 1980; Upton et al., 1982). Dynorphin produced effects in rhesus monkeys consistent with a partial kappa agonist profile, including temperature-dependent analgesia that was blocked by the kappa antagonist nor-BNI but not the mu antagonist clocinnamox, and attenuation of analgesia produced by full kappa agonists (Butelman et al., 1995). Dynorphin also engenders unique behavioral signs in rhesus monkeys (i.e., facial flushing, vocalizations, salivation, sedation and muscle relaxation) (Aceto and Bowman, 1992; Butelman et al., 1995). In summary, the pharmacodynamic effects of dynorphin vary depending on the test system (in vitro vs. in vivo) and assay type and species (rodent vs. primate). It is clear that its effects are not mu-like, but the kappa-opioid profile originally shown in vitro should not necessarily be expected in the intact animal, in which unique effects may be mediated by nonopioid mechanisms. The present study extended previous observations by obtaining assessments of both mu- and kappa-like opioid effects in human heroin abusers familiar with the effects of mu-opioid drugs.

The pharmacokinetics of dynorphin are also complex and may affect its pharmacodynamic profile. Dynorphin is rapidly metabolized by peptidases (Herman et al., 1980; Leslie and Goldstein, 1982; Young et al., 1985), and degradation products are detectable in plasma within minutes (Chou et al., 1994). A major metabolite, [des-Tyr1]dynorphin A(2-17), has nonopioid properties (Hooke et al., 1995), whereas the effects of other main metabolites, such as (1-12) and (2-13) fragments and tyrosine, are still unknown. Thus, the duration of direct effects from dynorphin A(1-13) might be brief due to its rapid metabolism. Furthermore, the selectivity of the effects of dynorphin seen after its administration could be low because the peptide and its metabolites are in flux (Leslie and Goldstein, 1982). Presumably, it would be advantageous to study the effects of dynorphin during a continuous infusion (i.e., when a steady state level is achieved) to maximize the stability and selectivity of its actions. This approachwith assessments both during and after a continuous drug infusionwas used in the present study.

Preclinical data suggest that dynorphin could have important clinical applications. In opioid-tolerant animals, dynorphin enhances morphine analgesia (Friedman et al., 1981; Schmauss and Herz, 1987; Song and Takemori, 1992; Tulunay et al., 1981) and antagonizes morphine respiratory depression (Woo et al., 1983). The ability of dynorphin to increase pain relief and reduce respiratory toxicity from opioids could certainly benefit opioid abusers and chronically treated pain patients. Furthermore, dynorphin can attenuate the signs and/or symptoms of spontaneous withdrawal in morphine-dependent animals (Aceto et al., 1982; Green and Lee, 1988) and heroin-dependent humans (Wen and Ho, 1982; Wen et al., 1984). Khazan et al. (1983) reported that dynorphin A(1-13) (0.25 mg/kg i.v.) could substitute for morphine (10 mg/kg i.v.) in physically dependent rats and that abstinence signs were "minimal" during saline substitution for dynorphin. Takemori et al. (1992, 1993) used the naloxone-precipitated withdrawal paradigm to study the ability of dynorphin to attenuate signs of physical dependence in mice that had been implanted with morphine pellets for 3 days. This model can be used to study potential treatment effects when a candidate medication is given before or after antagonist challenge. Takemori et al. (1992) showed that the administration of dynorphin 5 min before or 2 min after naloxone in morphine-dependent animals attenuated precipitated withdrawal effects, but dynorphin administered before naloxone was relatively more effective. Whether this treatment effect is opioid mediated is unclear because Takemori et al. (1993) found that not only dynorphin A(1-13) and (1-17) but also the nonopioid (2-17) fragment could suppress naloxone-precipitated withdrawal. Although dynorphin has been safely administered to human chronic pain patients for analgesic testing (Wen et al., 1985, 1987), there are no published data on whether dynorphin can modulate naloxone-precipitated opioid withdrawal in humans.

There were two objectives in this study, pursued as two phases during a single protocol. In phase 1, we evaluated the effects of dynorphin in nondependent, opioid-experienced subjects. The characterization of the human pharmacology of dynorphin may provide clues about its receptor mechanism of action and yield practical information on its dose effects, duration of action and safety for future clinical studies. In phase 2, we modeled preclinical studies that have shown that dynorphin can modulate the expression of naloxone-precipitated withdrawal after morphine exposure (Takemori et al., 1992, 1993). To specifically look for withdrawal modulation effects in humans, we extended our previous work with the acute morphine physical dependence paradigm (Azorlosa et al., 1994; Bickel et al., 1988; Heishman et al., 1989a, 1989b, 1990; Kirby et al., 1990) to determine whether the preclinical findings could be replicated in human volunteers.

	Methods

Top Abstract Introduction Methods Results Discussion References

Participants. The U.S. Food and Drug Administration and the local institutional review board approved this study. Volunteers were recruited from the Baltimore community by newspaper advertisements and word-of-mouth to participate in inpatient drug research. All acceptable volunteers (ages 18 to 50 years) showed objective signs of current intravenous opioid use (needle tracking marks and positive urine for opioids at screening) and reported sporadic opioid (i.e., heroin) use but were not physically dependent on entering the study. This was confirmed by the absence of a response to naloxone (10 mg/70 kg i.v.) challenge before the first experimental session. Volunteers provided a medical history and received a physical examination; those selected had no chronic health problems and were not taking prescribed medications. Subjects were excluded if they reported a history of treatment for psychiatric or alcohol problems. Volunteers were required to provide a urine sample that tested negative for opioids and a Breathalyzer sample that was negative for alcohol at the time of admission. After admission to the residential research unit, each participant provided written informed consent, was oriented to the study procedure and was paid per diem for successful study completion. Volunteers lived on the research unit for ~5.5 weeks, during which pre- and post-experimental naloxone challenges and all test sessions for phase 1 and phase 2 were conducted.

Study Design

Subjects participated sequentially in a two-phase study testing the direct effects (phase 1: first 2 weeks) and the withdrawal modulatory effects (phase 2: second 2 weeks) of dynorphin.

Direct effects. The phase 1 design was within-subject with four test sessions, scheduled twice per week, to ensure independence of test conditions. To increase safety, dynorphin doses were administered in an ascending series, beginning with placebo and progressing from 0.1 to 0.32 to 1.0 mg/kg. Dynorphin dose selection was in part based on preclinical studies; in addition, the 1 mg/kg i.v. dose matches that being used by another group studying the effects of dynorphin in opioid-dependent subjects (Specker et al., 1996) and is half of the maximum dose that has recently been studied in opioid-dependent chronic pain patients (Balla et al., 1993).

Modulation of naloxone-precipitated effects. The phase 2 design was within-subject with four test conditions, scheduled twice per week, to ensure independence of test conditions. Subjects received morphine either 15 hr (subjects 1-3) or 18 hr (subjects 4-6) before each test session. There were four test conditions: dynorphin with low- vs. high-dose naloxone and placebo with low- vs. high-dose naloxone. The two naloxone challenge doses were intended to produce different levels of withdrawal symptoms. During test sessions, a 15-min dynorphin or placebo pretreatment infusion was followed 5 min later by naloxone challenge. The lower naloxone challenge dose was always administered in the first two test sessions, and the higher dose of naloxone was always administered in the final two test sessions. This constrained randomized crossover design was adopted for safety reasons because higher-dose challenges could be cancelled if any adverse effects were detected with lower-dose challenge. The order of placebo and active dynorphin administration at each naloxone dose level was randomized.

Drugs

Dynorphin A(1-13) and mannitol placebo in 10-mg vials⁴ were stored at 20° F. Doses were reconstituted under aseptic conditions within 1 hr before injection. The dynorphin solution passed through a millipore filter attached to a 30-ml syringe, which was loaded in a programmable infusion pump. Dynorphin was always given as a 15-min continuous intravenous infusion (volume, 20 ml). Morphine sulfate (15 mg/70 kg; DuPont, Wilmington, DE) was administered intramuscularly in a volume of 2.5 ml. Naloxone hydrochloride (1 or 3 mg/70 kg and 10 mg/70 kg; DuPont) was administered in a volume of 5 ml as a 5-min manual, continuous intravenous infusion. Sterile 0.9% NaCl was the vehicle for morphine and naloxone. All drugs were administered under double-blind conditions.

Procedure

Before each session, the subject ate breakfast (completed by 7:00 a.m.) and was not permitted to smoke cigarettes from midnight before the session until the end of the session. Caffeine intake was prohibited throughout the residential stay. The volunteer was escorted from the residential unit to a quiet, light-controlled test area for all test sessions. Each test session consisted of a predrug base-line and a postdrug testing phase. Physiological signs, subjective symptoms and observer-rated withdrawal signs were measured at base line and periodically throughout the session (see below).

Direct effects. Phase 1 sessions began at 9:30 a.m. Intensive data recording and safety monitoring occurred throughout the session. Base-line measures were taken at session start (time, 10 min). At 9:40 a.m. (time, 0 min), the 15-min dynorphin or placebo intravenous infusion (injection 1) was initiated, and at 9:55 a.m. (time, 15 min), the infusion ceased. The first postdrug assessment took place at 9:50 a.m. (time, 10 min). At 10:00 a.m. (time, 20 min), a 5-min saline placebo infusion (injection 2) began, and it concluded at 10:05 a.m. This procedure served to maintain double-blind drug administration for phase 2, which involved naloxone challenge at this time. Starting at 10:05 a.m. (time, 25 min) and every 10 min thereafter until 115 min postdrug, the volunteer completed the assessment battery.

Modulation of naloxone-precipitated effects. In phase 2, volunteers received morphine on the residential unit the day before each test session. Subjects were administered three intramuscular injections of 15 mg/70 kg morphine spaced at 5-hr intervals (i.e., 9:00 a.m., 2:00 p.m. and 7:00 p.m.) for subjects 1 to 3 or at 4-hr intervals (i.e., 8:00 a.m., 12:00 noon and 4:00 p.m.) for subjects 4 to 6; in all cases, the cumulative dose was 45 mg/70 kg. The test session took place the following morning. During this session, the subject received two intravenous injections, with timing as described for phase 1. The first injection was a 15-min continuous intravenous infusion of placebo or 1 mg/kg dynorphin (counterbalanced order within each challenge dose assessment). The second was a 5-min continuous intravenous infusion of a low or high dose of naloxone (ascending order). High dose was always 10 mg/70 kg. Low dose was 3 mg/70 kg for subjects 1 to 3 and 1 mg/70 kg for subjects 4 to 6. Phase 2 assessment time points were identical to phase 1. However, session time was defined as relative to start of naloxone (rather than dynorphin) infusion (i.e., naloxone infusion was time 0-5 min).

Test Session Measures

During test sessions, four subjective report measures were administered sequentially at each assessment time point in the order presented below:

Subjective effects. The Drug Effect Questionnaire had six VAS questions that were analyzed individually. Subjects rated from 0 (not at all) to 100 (extremely) the items high, any drug effect, good drug effect, bad drug effect, withdrawal sickness and drug liking.

Physiological signs. Immediately after completion of the subjective test battery at each session time point, pupil diameter was assessed by taking Polaroid photographs of the right eye (in dim lighting at 3× magnification). Pupil diameter was measured directly from the photographs (to an accuracy of 0.01 ml) using precision calipers. The scorer was blind to experimental condition. Respiration rate was measured continuously by chest bellows connected to an Apple IIgs microcomputer; analog pressure changes were converted to breaths/min for each 1 min of the session. Heart rate and blood pressure (systolic, diastolic and mean arterial) were recorded every 1 min by using a Sentron automatically inflatable cuff. Skin temperature was recorded every 1 min using a thermistor taped to the distal end of the nondominant index finger. Respiration rate, skin temperature, heart rate and blood pressure data were averaged across 10 min of base line and for 10-min time intervals during the remainder of the session.

Observer-rated signs. A trained research assistant was present during test sessions to observe the subject for objective opioid withdrawal signs. The research assistant, who was blind to test condition, rated six withdrawal signs using a modified Himmelsbach (1941) withdrawal severity scale (Eissenberg et al., 1996). The observer rated each of the signs on a 3-point scale, where 0 = none, and 1 and 2 were individualized for each sign. Signs and their ratings were: yawning, 1 = one yawn during the observation period (since last measurement), 2 = two or more yawns during observation period; lacrimation, 1 = manually induced tearing of the eye, 2 = spontaneous tearing of the eye; rhinorrhea, 1 = watery sound when sniffing in either nostril, 2 = spontaneous runny nose; perspiration, 1 = can palpate sweat on skin, 2 = can observe sweat without touching subject; gooseflesh, 1 = can feel bumps after manually stimulating skin, 2 = spontaneous piloerection (without touching) and restlessness, 1 = one behavioral sign, 2 = two or more behavioral signs. Behavioral signs of restlessness were scored when the subject shifted position at least once during the observation period, moved legs rhythmically for >5 sec, fidgeted with hands for >5 sec or made verbal comments expressing a desire for the procedure to end. The six scores for the individual items were summed to form a composite observer rating score (maximum composite score, 12).

Morphine Pretreatment Measures

On days when the volunteer remained on the residential unit and received three intramuscular injections (saline during phase 1, morphine during phase 2), data were collected 30 min before and 1 hr after each injection, resulting in six data points for each divided dose pretreatment. Morphine effects are described in "Results" to show that these subjects did respond in an appropriate fashion to the opiate agonist drug.

Subjective effects from pretreatments were reported using the Opiate Adjective Checklist and VAS items similar to those used in test sessions (high, any drug effect, good drug effect, bad drug effect, like drug effect, dislike drug effect and desire more drug). Physiological measures collected by nursing staff at these same time points included pupil diameter (Polaroid photo), oral temperature, auscultatory systolic and diastolic blood pressure, manually determined pulse rate and respiration rate (observed by counting chest movements).

Data collection time-line. Subjective report, observer rating and pupil measures were obtained periodically throughout each test session. Order of measurement was Drug Effect Questionnaire, Opiate Adjective Checklist, Drug Identification Questionnaire, Side Effect rating scales, pupil photograph and observer ratings. For phase 1 (sessions 1-4), data were collected starting at 15 min (15 min base line) before the start of dynorphin administration (time, 0) and at 10, 25, 35, 45, 55, 65, 75, 85, 95, 105 and 115 min after the start of 15-min dynorphin infusion. For phase 2 (sessions 5- 8), the identical drug injection and assessment schedule was used to maintain the double-blind, but session time 0 was defined as the start of naloxone infusion. Base-line data were collected 10 min before naloxone administration. Postchallenge data were collected at 5, 15, 25, 35, 45, 55, 65, 75, 85 and 95 min after the start of the naloxone infusion.

Data Analysis

Measures were pupil diameter, respiration rate, skin temperature, heart rate, systolic and diastolic blood pressure, Drug Effect Questionnaire (six VAS scores), individual item and subscale scores for the Opiate Adjective Checklist, Drug Identification Questionnaire and individual item and composite scores of observer-rated withdrawal signs.

Dynorphin direct effects. Sessions 1 to 4 were designed to evaluate the dose- and time-related effects of dynorphin. Data were analyzed using two-way dynorphin dose × session time ANOVAs. Tukey's post hoc tests were used to evaluate the significance of active dynorphin (0.1, 0.32 and 1 mg/kg) vs. placebo score differences at each time point; critical difference scores were based on the two-way interaction error term.

Morphine pretreatment effects. On days before sessions 5 to 8, morphine pretreatment effects were evaluated. Because initial analysis indicated that morphine pretreatment effects did not vary across sessions, results from within-session time (6 points: 30 min before and 1 hr after each of the three intramuscular injections) ANOVAs are reported. Tukey's post hoc tests were used to evaluate changes in postdrug response relative to predrug (i.e., 30 min before injection 1) base line.

Modulation of naloxone-precipitated effects. Sessions 5 to 8 were designed to evaluate the ability of dynorphin to alter naloxone-precipitated withdrawal effects. Initial data analysis revealed no difference in precipitated effects from the two low naloxone doses (1 and 3 mg/70 kg); therefore, these low doses were collapsed into a single level of this naloxone dose factor. Data were thus analyzed using three-way dynorphin dose (0 vs. 1 mg/kg) × naloxone dose (low [1 or 3 mg/70 kg] vs. high [10 mg/70 kg]) × session time ANOVAs. Tukey's tests were used to identify significant dynorphin (1 mg/kg) vs. placebo score differences for high vs. low naloxone dose conditions at each time point; critical difference scores were based on the three-way interaction error term.

	Results

Top Abstract Introduction Methods Results Discussion References

Phase 1: Direct Effects

Subjective effects. Dynorphin produced transient increases in some subjective ratings that peaked during or at the first measurement point after infusion, depending on the measure, and then quickly dissipated. Figure 1 shows that dynorphin produced dose- and time-dependent changes in subjects' reports of feeling "any drug effect." Scores were significantly elevated for both the 0.32 and 1 mg/kg doses at the during-infusion (10 min) and first postinfusion (25 min) measurement points, dose × time F(33,165) = 4.97. 

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Dose- and time-dependent effects of dynorphin for the VAS "any drug effect." Shaded vertical bar, critical difference score (Tukey's test) based on the dose × time interaction. , , Mean drug response was significantly different than placebo response at comparable time points.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Dynorphin dose effects on peak (± S.E.M.) ratings (top) of bad drug effect and good drug effect. Peak ratings for individual subjects always occurred during infusion (10 min) or at the first time point after infusion (25 min). , , Mean responses were significantly greater than for the associated placebo condition.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Dynorphin dose effects on the number of drug identifications as placebo (left), opiate agonist (middle) and opiate antagonist (right). Data indicate the summed frequency of identifications made by six subjects at two measurement time points (10 and 25 min) in each dose condition. , , , Number of identifications in the active dynorphin dose condition is significantly different from the 0 mg/kg condition.

Physiological signs. Dynorphin did not significantly change any of the physiological signs measured here. Mean pupil diameter values ranged from 4.8 to 5.7 mm across the four test (i.e., dose) sessions and showed no consistent within-session (i.e., time related) trends. Most other physiological signs showed progressive changes from preinjection base line to end of the session, which resulted in significant time effects but no dose × time interaction. Mean finger temperature (i.e., averaged across doses) increased from 91.5° F at base line to 94.5° F at end of the session. Mean respiration rate decreased from 19.5 breaths/min at base line to 16 breaths/min. Mean heart rate decreased from 75 beats/min at base line to 70 beats/min. Mean systolic blood pressure decreased from 120 mm Hg at base line to 112 mm Hg. Diastolic blood pressure did not show systematic changes. However, one volunteer experienced a transient (<5 min) decrease in diastolic pressure to 40 mm Hg with the high dose immediately after infusion.

Phase 2: Morphine Pretreatment Effects

Subjective effects. Relative to predrug base line, a significant increase in good drug effect ratings (from 0 to 30 on the 0-100 scale; mean across 4 pretreatment days) was observed after the first morphine dose (15 mg/70 kg i.m.), with no further score increases observed after additional 15-mg doses administered at 4- to 5-hr intervals, session time F(5,25) = 12.02. Similar elevations were seen on drug "high," F(5,25) = 6.68, and drug liking, F(5,25) = 9.27. Scores on the agonist scale of the Opiate Adjective Checklist were also significantly elevated after the first daily exposure to 15 mg of morphine, F(5,25) = 8.12, with no further increases after cumulative administrations. Mean predrug score on this measure was 7.7; mean score at 1 hr postdrug was 12.2.

Physiological signs. In contrast to the lack of physiological activity of dynorphin, 15 mg/70 kg morphine significantly constricted pupils (from 5.5 mm predrug to 4.1 mm), with further constriction to 3.5 and 3.1 mm after additional spaced doses within the same day, F(5,25) = 17.01. Relative to base line, morphine 15 mg/70 kg also produced mean increases in oral temperature (from 98.0 to 98.5° F), pulse rate (from 78.7 to 83.5 beats/min) and systolic blood pressure (from 121.8 to 133.8 mm Hg), but these changes were not statistically significant; higher cumulative doses of 30 and 45 mg/70 kg did not produce an additional incremental effect.

Phase 2: Modulation of Naloxone-Precipitated Effects

Subjective effects. Naloxone challenge given on the day after morphine treatment produced characteristic opioid withdrawal symptoms and signs during the test session, including significant increases in reported withdrawal sickness, bad drug effect and opiate antagonist symptoms, time F(11,55) values = 8.01, 5.33 and 4.48, respectively. As figure 4 shows, withdrawal sickness scores rose rapidly, remained significantly elevated for 35 min after naloxone injection, and then returned gradually to base line during the remainder of the session. As shown in figure 4, both the intensity and duration of the opioid withdrawal response were similar for the two naloxone doses administered. However, the opioid antagonist scale (adjective checklist) items backache, naloxone dose F(1,5) = 10.0, and flushing, naloxone dose × time F(11,55) = 2.50, showed significantly greater effects with high vs. low naloxone doses, whereas other antagonist items (muscle cramps, gooseflesh, sick to stomach and skin damp/clammy) and the antagonist subscale total score showed trends in the same direction (values for P < .10).

View larger version (28K): [in this window] [in a new window]   Fig. 4.   Dose- and time-related effects of dynorphin and naloxone on ratings of withdrawal sickness (top) and skin temperature (bottom). Left, placebo () vs. 1 mg/kg i.v. (, ) dynorphin treatment effect at the low naloxone dose (either 1 or 3 mg/70 kg i.v.; see text). Right, dynorphin effect at the high naloxone dose (10 mg/70 kg i.v.). Session time 0, start of the 5-min naloxone infusion. Shaded vertical bar, critical difference score (Tukey's test) from the dynorphin × naloxone × time interaction; data points above (top) or below (bottom) the shaded bar indicate that postnaloxone response significantly differed from prenaloxone baseline. , Dynorphin and placebo conditions differed significantly.

Physiological signs. Similar to the subjective effect measures, naloxone produced time-, but not dose-, related changes in physiological signs of withdrawal. Pupil diameter, heart rate and systolic and diastolic blood pressure each showed significant biphasic (increasing then decreasing) patterns after naloxone, time F(13,65) values = 2.88, 3.67, 2.92 and 6.48, respectively. As figure 4 (bottom) illustrates, skin temperature showed a biphasic (decreasing then increasing) trend across session time, F(13,65) = 3.82, P < .07. For most physiological signs, mean changes were significant from 5 to 25 min after naloxone (both doses). After naloxone, mean pupil size increased ~1 mm, heart rate increased ~3 beats/min, systolic blood pressure increased ~8 mm Hg, diastolic blood pressure increased ~6 mm Hg and skin temperature decreased ~4° F (fig. 4). No significant changes were found for respiration rate. Dynorphin did not significantly modulate naloxone-precipitated changes in physiological signs of withdrawal, although dynorphin tended to exaggerate and prolong naloxone-precipitated decreases in skin temperature, dynorphin × naloxone × time F(13,65) = 2.48, P < .07, as shown in fig. 4.

Observer-rated withdrawal signs. Objective opioid withdrawal signs increased after antagonist challenge. Overall increases in lacrimation, rhinorrhea and yawning items and the observer-rated total score were significant, time F(11,55) = 10.73, 3.91, 7.62 and 9.97, respectively. The 10 mg/70 kg naloxone dose produced significantly longer-lasting lacrimation than the lower dose, naloxone dose × time F(11,55) = 2.52; this trend was also observed in the total observer score, F(11,55) = 1.19, P < .07. Consistent with other measures, dynorphin did not attenuate observer-rated signs of opioid withdrawal.

	Discussion

Top Abstract Introduction Methods Results Discussion References

This study showed that dynorphin A(1-13), at doses within the range of those that have been given systemically to nondependent rhesus monkeys (Aceto et al., 1982; Aceto and Bowman, 1992; Butelman et al., 1995) and human volunteers maintained on opioids (Balla et al., 1993; Wen et al., 1984; Wen and Ho, 1982), had pharmacological activity in nondependent humans. This itself is notable because as an opioid neuropeptide, the ability of dynorphin to produce drug effects of any type might be limited by low bioavailability from systemic administration. The compound produced changes in subjective effects that were time- and dose-related. Subjective effects from the two higher doses (0.32 and 1 mg/70 kg) generally peaked during the 15-min intravenous infusion, although some volunteers showed a peak response at 15 min after the end of the infusion (fig. 1). In a comparable study of the duration of analgesic action of dynorphin in rhesus monkeys not tolerant to opioids, effects from 0.1 mg/kg were negligible; effects after acute bolus doses of 0.32 and 1 mg/kg i.v. were apparent at 15 min, peaked at 30 min and returned to base line at 60 min postdrug; whereas effects from 3.2 mg/kg lasted up to 75 min (Butelman et al., 1995). The present data in human subjects are consistent with these primate analgesic test data in showing similar dose-effect separation but provide a duration of action estimate that is relatively shorter for subjective effects. An important procedural difference between studies is that for safety reasons, a 15-min intravenous drug infusion was used in this human study as opposed to a 15-sec intravenous infusion in the rhesus monkey study.⁵ This could be expected to reduce both the intensity and duration of effects. In fact, peak effects occurred during infusion and dissipated quickly. The transient effects of dynorphin in this study are consistent with its rapid metabolism; however, more thorough analysis of pharmacokinetic/pharmacodynamic relationships will be needed to determine whether the metabolites of dynorphin A(1-13) could mediate some of the effects observed in this study.

Opioid-abusing volunteers who were sensitive to the typical subjective effects of the mu agonist morphine in this study provided mixed subjective reports of the effects of dynorphin. As dynorphin dose increased, some reported good drug effects and some reported bad drug effects; all ratings increased in magnitude with increasing dose. As the dose increased, volunteers also provided drug identification responses consistent with their VAS reports (table 1). The subjects who reported good drug effects were more likely to identify dynorphin as an opiate agonist, although opiate antagonist identifications were less reliably made by those reporting bad drug effects. Taken alone, these data suggest that the effects of dynorphin are dissimilar to those of mu-opioid receptor agonists because the latter routinely produce only good drug effect and opioid agonist identifications (Greenwald et al., 1996; Preston and Bigelow, 1991).

There were interesting individual differences in the pattern of subjective ratings, particularly at the high dose of dynorphin. Three of the six subjects predominantly liked dynorphin, whereas the other three did not (table 1). Two likers (subjects 2 and 3) experienced only good drug effects and two nonlikers (subjects 5 and 6) experienced only bad drug effects. Although this study used a small sample, we explored the data set to identify factors that might explain this group difference. The only apparent difference was that likers had longer opiate abuse histories (and were older) than the nonlikers. Because all volunteers had extensive histories of opiate use, additional work would be needed to confirm whether this unexpected but interesting finding merits further attention. The mixed pattern of drug effects within and between subjects in this study could be due to pharmacological factors, such as mixed agonist activity at both mu and kappa receptors, but also individual differences in neurobiological function.

Data from the present study suggest that the effects of dynorphin in humans are not primarily (if at all) mediated by mu-opioid receptors. This is consistent with preclinical studies of this peptide, which suggest that its pharmacological effects may be more similar to kappa-opioid or nonopioid compounds than mu-opioids. The present findings, however, are not sufficient to specify further the mechanism(s) of the effects of dynorphin. Mixed or indirect actions at known opioid receptors could be involved, as could nonopioid mechanisms. Drug identifications do suggest that the effects overlap those of the opioid class, but these data should be cautiously interpreted. Volunteers were informed that they might receive opiate drugs, naloxone and dynorphin ("which resembles a natural substance in the brain") and thus may have expected dynorphin to produce opiate-like effects because it was administered in the context of other opiates. This would explain the high frequency with which subjects endorsed opiate-like qualities. However, this instructional set does not explain why subjects rated dynorphin as having both opiate-like (agonist) and withdrawal-like (antagonist) characteristics as drug dose increased under double-blind conditions (fig. 3). In phase 1, opioid agonists were not administered before dynorphin and withdrawal signs/symptoms were not detected; thus, a parsimonious interpretation of these data is that antagonist-like drug identifications of dynorphin by these opiate-experienced subjects may simply reflect a familiar or generic label for "bad" drug effects (i.e., consistent with their VAS ratings, see fig. 2).

It is similarly difficult to interpret the profile of effects observed as kappa like. Kappa-opioids can produce diuresis, perceptual distortions, dysphoria, sedation, psychotomimesis and sleep disruptions in human subjects (Kumor et al., 1984, 1986; Pfeiffer et al., 1986; Pickworth et al., 1986), with some weak miotic and respiratory depressant effects (Jaffe and Martin, 1990). We have noted that dynorphin did not produce appreciable physiological (e.g., miotic, respiratory) changes. Although the full array of potential kappa-opioid signs described above was not systematically assessed during the session time-line, our informal observations indicated that they did not occur at the acute doses tested, with the possible exception of dysphoria in some subjects (i.e., nonlikers). Therefore, the data tentatively suggest that the effects of dynorphin are also not particularly kappa like.

In summary, the effects of dynorphin are not consistent with those of mu agonists and, on the basis of limited human data, also appear to differ from those of kappa agonists. Alternatively, dynorphin could produce its unique pattern of effects through other pathways, such as indirect actions at mu receptors or nonopioid mechanisms. Because the major goals of this study were to establish the safety and duration of action of dynorphin, measurements were intensive and focused on opioid characteristics. Future studies using different methods (e.g., drug discrimination), non-mu drug controls (e.g., kappa-opioids and nonopioids), pharmacokinetic analyses (e.g., to establish any influence of psychoactive metabolites) and measures that assess effects across multiple drug classes (e.g., Addiction Research Center Inventory) may provide novel, useful information concerning the mechanism of action of dynorphin.

One important goal of this study was to determine whether dynorphin would modulate opioid withdrawal signs or symptoms, as has been demonstrated in previous studies (Takemori et al., 1992, 1993; Wen et al., 1984; Wen and Ho, 1982). Dynorphin (1 mg/kg) did not prevent or attenuate the signs or symptoms of naloxone-precipitated withdrawal after acute morphine (45 mg/70 kg) pretreatment. Instead, where trends were observed, these were opposite to the predicted effects, suggesting that dynorphin might potentiate rather than attenuate the effects of naloxone. The failure to observe opioid withdrawal attenuation occurred despite the manipulation of naloxone doses (1-10 mg/70 kg), a feature designed to increase chances of observing the withdrawal-modulating effects of dynorphin.

One explanation for the failure to observe predicted effects is that the ability of dynorphin to modulate opioid physical dependence may only occur with relatively greater duration and/or intensity of morphine exposure. Previous studies in mice have assessed the effects of dynorphin on naloxone-precipitated withdrawal after more extensive exposure to high-dose morphine (i.e., 75-mg pellet implanted for 3 days and left intact during withdrawal assessment) (Takemori et al., 1992, 1993). In another study (Aceto et al., 1982), rhesus monkeys were administered 3 mg/kg s.c. every 6 hr for 90 days. In the only study of the effects of dynorphin on acute opioid physical dependence, a high morphine pretreatment dose (100 mg/kg s.c.) was given 6 hr before naloxone challenge (Hooke et al., 1995). Pretreatments (whether acute or chronic) in each of the above studies would no doubt have produced a level of physical dependence markedly greater than that produced in this study. The higher level of opioid exposure in previous studies may have engendered quantitatively and/or qualitatively different perturbations of opioid receptors, second-messenger systems or other processes that may be more susceptible to dynorphin regulation.

A second reason for failure to observe withdrawal modulation might be low sensitivity of the precipitated withdrawal model that was used. Dynorphin might be more likely to exhibit dependence-modulating effects in a human spontaneous withdrawal model, which would also have greater clinical relevance. However, a preliminary report by Specker et al. (1996) of a completed study that used a spontaneous withdrawal model did not find a convincing dynorphin (0.15-1.0 mg/kg) antiwithdrawal effect. These results are consistent with findings from earlier human studies. For example, Wen and Ho (1982) and Wen et al. (1984) treated different groups of chronic heroin-dependent volunteers with test compounds including dynorphin, other opioid peptides or saline only after their withdrawal symptom scores were observed to rise 100% above initial levels measured at inpatient treatment admission. Under those conditions, dynorphin A(1-13) (0.06 mg/kg i.v.; a dose that was notably lower than in this study) reduced withdrawal symptoms 20% to 30% (from pretreatment levels) for ~45 min. Other test compounds (-endorphin and [D-Ala²,D-Leu⁵]enkephalin) produced similar treatment effects. However, saline treatment also significantly reduced withdrawal symptoms 10% to 25% (relative to pretreatment levels) for ~20 min. The roughly equivalent effects of placebo and dynorphin reported from these earlier studies fail to support a withdrawal-modulating effect of dynorphin in humans. To assess sensitivity of the precipitated withdrawal methodology, it would be useful in future medication-screening studies to incorporate comparison treatments that are able to modulate the precipitated withdrawal response. In one previous study using this approach, oral clonidine attenuated some effects of naloxone-precipitated withdrawal (Sullivan et al., 1994). Morphine treatment might also be used to reduce the expression of physical dependence but only when given after naloxone because morphine given before naloxone in the already-dependent subject would be expected to enhance withdrawal effects (Azorlosa et al., 1994; Brase et al., 1976; Sofuoglu et al., 1990).

A final possible explanation for the absence of a dynorphin effect on opioid withdrawal could be low statistical power in the present study. This seems unlikely, however, because modulation of withdrawal intensity has been detected in previous human studies of naloxone-precipitated withdrawal with similar sample sizes of five to eight subjects when dose or time parameters were manipulated (e.g., Eissenberg et al., 1996; Greenwald et al., 1996; Heishman et al., 1989a, 1989b; Kirby et al., 1990). Furthermore, as previously noted, no trends in the predicted direction were observed. Therefore, it is likely that a larger sample size would either confirm a withdrawal-exaggerating effect of dynorphin or confirm a no-effect result.

In conclusion, dynorphin (0.1-1 mg/kg i.v.) produced significant dose- and time-dependent subjective effects, but not physiological changes, in opioid-experienced volunteers who were not physically dependent at the time of testing. The subjective effects of dynorphin were mixed (good and bad drug effects), whereas morphine produced only good drug effects and miosis. Thus, dynorphin was shown to be pharmacologically active in humans, producing effects that were neither mu nor kappa like. Dynorphin at 1 mg/kg appeared safe; it produced no serious side effects when administered intravenously over 15 min. Contrary to predictions from preclinical research, dynorphin did not significantly attenuate the expression of naloxone-precipitated withdrawal after acute morphine exposure. Rather, trends were observed on some measures suggesting potentiation of the effects of naloxone. Thus, with the use of an acute opioid-dependence model with relatively mild-intensity naloxone-precipitated effects, the ability of dynorphin to attenuate opioid withdrawal in human volunteers was not evident. Whether dynorphin might be an effective antiwithdrawal agent in human subjects with more severe opioid dependence (analogous to animal subjects in preclinical studies) and/or in spontaneous withdrawal models should be investigated before concluding that the withdrawal-attenuating properties of this neuropeptide are absent in human volunteers.