Department of Pharmacology and Toxicology, Medical College of
Virginia/Virginia Commonwealth University, Richmond, Virginia
Fentanyl and morphine are administered to human neonates and infants to
provide analgesia and sedation during painful and stressful procedures.
These opioids have often been shown to produce tolerance and dependence
during continuous intravenous infusion. In neonatal animals, morphine
produces tolerance and dependence, yet little is known about fentanyl.
This report describes the first model for studying opioid tolerance and
dependence in neonatal animals with use of osmotic minipumps. Postnatal
day 6 rat pups were anesthetized and then remained naive or were
surgically implanted subcutaneously with Alzet osmotic minipumps
containing either saline or fentanyl (100 µg/kg/hr). Tolerance and
dependence were assessed 72 hr after implantation. The ED50
values for fentanyl antinociception in the tail-flick test were not
different between naive and saline pump-implanted animals. However, the
fentanyl pump-implanted animals were tolerant to fentanyl. The
tolerance observed was not the result of gender, developmental changes, fentanyl distribution or changes in fentanyl metabolism. These results
indicate that continuous administration of fentanyl via osmotic minipump can render normal neonatal rats tolerant and physically dependent on fentanyl in 72 hr. Withdrawal precipitated by
naloxone (5 mg/kg s.c.) in the fentanyl pump-implanted animals was
characterized by increased spontaneous activity,
micturition/defecation, wall climbing, abdominal stretching, tremors,
scream on touch and spontaneous vocalization. This new model may
provide a tool for studying the long-term consequences of neonatal
opioid exposure in juvenile and adult animals.
 |
Introduction |
Reports indicate that physicians
are now more likely to use opioids to manage pain in neonates and
infants than in the past (Purcell-Jones et al., 1987
; Yaster
1987; Sukhani, 1989; McLaughlin et al., 1993). Concern has
developed over the possible overzealous management of pain and its
consequences with the increased use of opioids. Opioids are routinely
administered i.v. to provide continuous analgesia and sedation during
extracorporeal membrane oxygenation and mechanical ventilation for the
treatment of life-threatening pulmonary diseases in neonates and
infants (Arnold et al., 1990, 1991; Roth et al.,
1991; Leuschen et al., 1993). Considerable evidence
indicates that iatrogenic tolerance and dependence can develop in this
population receiving fentanyl or morphine by continuous intravenous
administration (Maguire and Maloney, 1988; Norton, 1988; Arnold
et al., 1990, 1991; Noerr, 1990; French and Nocera, 1994;
Katz et al., 1994; Franck and Vilardi, 1995). Iatrogenic tolerance was indicated when a given dose of fentanyl or morphine became ineffective and the patient required increasingly larger doses
to provide the same level of analgesia observed with smaller doses.
Physical dependence on morphine and fentanyl was characterized by the
presence of withdrawal signs when the drug was discontinued. Fifty to
84% of neonates removed from fentanyl exhibited abstinence within a
24-hr period and 48% exhibited signs with morphine withdrawal (Norton,
1988; Arnold et al., 1990; French and Nocera, 1994). To our
knowledge, nothing is known about the long-term consequences of
iatrogenic tolerance and dependence in neonates and infants as they
grow into juveniles and adults.
Although studies have investigated opioid antinociception in neonatal
animals, surprisingly little is known about neonatal tolerance and
physical dependence. Furthermore, these studies have investigated
morphine, but no studies have examined fentanyl. Morphine tolerance and
dependence have been observed in neonatal rats, but reports disagree
about the age at which this can occur. Some researchers report
tolerance in 9-day-old rats (Van Praag and Frenk, 1991; Barr and Wang,
1992), whereas others did not observe tolerance until the rats were
15-days-old (Fanselow and Cramer, 1988; Windh et al.,
1995). In each of these studies, morphine was repeatedly administered
by bolus injection with a variety of dosing and injection schedules.
Such differences could account for the divergent results among these
studies. In addition, it is likely that repeated administration leads
to wide fluctuations in central nervous system opioid concentrations
that could affect the development of tolerance. Furthermore, bolus
injections require repeated handling and stress of neonatal rats and
dams that could affect the development of tolerance. These concerns led
us to begin using subcutaneously implanted Alzet osmotic minipumps to render neonatal rats tolerant and physically dependent on opioids. Minipumps have the advantage of delivering drug at a constant rate to
provide stable plasma and tissue opioid concentrations for long periods
of time. The constant infusion of opioid by minipumps may also reduce
the toxicity often associated with bolus drug administration and
minimize the stress and handling of the neonates and dam. In addition,
minipump drug delivery more closely mimics the intravenous route by
which opioids are administered to human neonates and infants.
Therefore, we developed a model of neonatal rat tolerance and
dependence to test the hypothesis that continuous fentanyl
administration would render neonatal rats tolerant and physically
dependent on fentanyl. Our characterization of the model addresses
shifts in fentanyl potency as a result of gender, age, postnatal
development and metabolism.
| |
Methods |
Source of neonatal rats.
Breeding of adult animals occurred
in the animal care facilities at the Medical College of Virginia.
Nulliparous female Sprague-Dawley rats (Harlan Sprague Dawley,
Indianapolis, IN) were paired with males of the same strain and from
the same supplier. Approximately 2 weeks before parturition, the dams
were housed individually in plastic cages with bedding and checked for
pups on day 21, with the day of birth designated as PND0. The pups were
culled to groups of 10 consisting of five males and five females. Extra pups were fostered to dams with less than 10 live births.
Surgical implantation of Alzet osmotic minipumps.
Fentanyl
hydrochloride was dissolved in sterile isotonic saline and cold
sterilized by filtration through a Millex-HV, 25 mm, 0.45 µm syringe
filter (Millipore Corp., Bedford, MA). Alzet 1003D osmotic minipumps
were loaded with fentanyl or saline in a laminar flow hood by sterile
procedures as described in "Alzet Osmotic Minipumps: Technical
Information Manual" from Alza Corp., Palo Alto, CA. The Alzet 1003D
pump is recommended for animals weighing at least 10 g and infuses
solution at 1 µl/hr for 72 hr (Alza Corp., Palo Alto, CA). The loaded
pumps were then primed by placing them in sterile isotonic saline at
37°C for 3 hr before implanting them in the rats. Pump delivery
beginning at 4 hr (Alza Corp.) allowed the neonatal rats 1 hr to
recover from anesthesia. Therefore, time zero of the study began 1 hr
after implantation of the pump.
Neonatal rats were briefly anesthetized with methoxyflurane. After
induction of anesthesia (as noted by the absence of the righting reflex
and foot pinch response), the pups were placed on a 37°C heating pad.
The skin on the back was swabbed with 70% ethanol approximately 1.5 cm
from the tail. Sterile scissors were used to make a 1-cm incision
approximately 1.5 cm from the base of the tail (fig. 1).
Alza Corp. recommends that the incision be made at the base of the neck
in adult animals. However, in early implantation trials of neonates we
observed that the rostral position of the incision and the pump
interfered with feeding posture. By making the incision near the base
of the tail the incision and caudal position of the pump did not
interfere with movement and the feeding posture of the animal. After
the incision, a sterile preloaded Alzet 1003D osmotic minipump was
inserted under the skin and the incision was closed with wound
autoclips. We are currently using Vetbond Tissue Adhesive (3 M Animal
Care Products, St. Paul, MN), which decreases postsurgical tissue
damage induced through normal maternal care. The pump was inserted so that the delivery port was 180° from the incision to prevent drug leakage from the incision. The area was swabbed with betadine, and the
animal was allowed to recover. The animals were injected i.p. with
20,000 U of potassium penicillin G to prevent infection and 0.5 ml of
isotonic saline s.c. to prevent hypovolemia according to IACUC
(Institute on Animal Care and Use Committee) guidelines at the Medical
College of Virginia. Within each litter of five females and five males,
two rats were anesthetized but remained naive, whereas eight were
randomly assigned to receive saline- or fentanyl-filled pumps. The pups
were returned to the dam and were challenged 72 hr later with fentanyl
s.c. for generation of dose-response curves.
View larger version (101K):
[in this window]
[in a new window]
|
Fig. 1.
Photograph of PND6 rats before and after
implantation of osmotic minipumps. Rats were anesthetized and implanted
with saline- or fentanyl-filled Alzet 1003D osmotic minipumps as
detailed in Methods. After a 72 hr infusion period, the animals were
injected with fentanyl s.c. and tested for antinociception in the
tail-flick test.
|
|
Tail-flick test.
For tests of antinociception, neonatal rats
removed from the dam were kept warm (37°C) on a heating pad, because
they are unable to thermoregulate on their own and hypothermia can
contribute to a diminished pain sensitivity (Phifer and Terry, 1986).
The tail-flick test used to assess antinociception was developed by D'Amour and Smith (1941) and modified by Dewey et al.
(1970). Before injection of drug, the base-line (control) tail-flick
latencies were measured for each animal. As in previous studies of
neonatal rats in this laboratory, the intensity of the heat stimulus
was adjusted to yield a base-line latency of 3 to 4 sec, and a 10-sec cut-off was used to prevent tissue damage (Enters et al.,
1991; McLaughlin and Dewey, 1994). At the appropriate time after drug administration the test latency was measured and the data were transformed to the %MPE according to the method of Harris and Pierson
(1964). This was calculated as: %MPE = [(test latency 
control latency)/(10 control latency)] × 100. Time-course data were analyzed by a two-factor (treatment and time) ANOVA followed
by the Tukey's test for post hoc analyses. The
ED50 value and 95% confidence limits for dose-response
curves and potency ratios were calculated by the method of Tallarida
and Murray (1987).
Fentanyl equivalent levels.
In other experiments, naive PND6
and PND9 animals received fentanyl for determination of antinociceptive
ED50 values and brain fentanyl equivalent levels. For each
dose of unlabeled fentanyl used for the dose-response curves,
radiolabeled fentanyl (0.045 µCi) was included for determination of
plasma and brain fentanyl equivalent levels. After the assessment of
antinociception, the pups were sacrificed via decapitation,
and brain and plasma levels were collected for determination of
fentanyl equivalent levels. The brains were homogenized in 1 ml of
distilled water and then solubilized overnight with tissue solubilizer
TS-2 (200 µl/mg tissue). After overnight solubilization, the solution
was neutralized with concentrated hydrochloric acid. Budget-Solve
scintillation cocktail was added and samples were counted in a Beckman
scintillation counter. Blood samples were collected in
ethylenediaminetetraacetate-coated microcentrifuge tubes after
decapitation of the pups. The blood samples were centrifuged at 14,000 rpm for 10 min and then 50 µl of plasma was removed, 10 ml of Budget
Solve was added and samples were counted in a Beckman scintillation
counter. EC50 values (i.e., nanograms of
fentanyl equivalents per gram of brain tissue eliciting 50% MPE) in
the brain were calculated by the method of Tallarida and Murray (1987).
Cytochrome P450 3A ELISA assay.
Analysis of liver and brain
P450 3A levels was conducted according to the procedures described in
the kit from Amersham Life Sciences (Arlington Heights, IL).
Seventy-two hours after beginning the experiment, the livers and brains
were removed from PND9 rats that remained naive or were implanted with
saline- or fentanyl-filled pumps. The tissue was homogenized in
ice-cold 50 mM Tris with 0.3 mM phenylmethyl sulfoxide with a
glass/Teflon homogenizer before centrifugation. The tissue was
centrifuged at 1100 × gav for 3 min. The
supernatant was saved and the pellet was resuspended in Tris buffer and
centrifuged again for 3 min. Both supernatants were combined and
centrifuged at 15,800 × gav for 10 min.
The supernatant was then centrifuged at 100,000 × gav for 30 min to isolate the microsomal
fraction. The microsomal pellet was resuspended in Tris buffer and
centrifuged again at 100,000 × gav for 30 min. The pellet was resuspended and the protein levels were adjusted to
a concentration of 500 µg/ml in preparation for the ELISA assay, which is based on a one-site immunoenzymometric format. Standards and
samples were solubilized and coated onto a 96-well microtiter plate.
Binding to the wells was detected and quantitated with use of a primary
rabbit antibody specific to rat P450 3A, a secondary antirabbit
immunoglobulin-horseradish peroxidase antibody conjugate and a
tetramethylbenzidine substrate. Cross-reactivity of the primary
antibody between the cytochrome P450s 1A1, 2B1 and 4A1 is less than 1%
(Amersham Life Sciences). The color developed and absorbance detected
with a 450-nm filter was proportional to the cytochrome P450 3A content
of the standard and samples. Sample concentrations of cytochrome P450
were determined by interpolation from a standard curve by use of the
Softmax (version 2.32) program (Molecular Devices Corporation, Menlo
Park, CA). The standard curve was constructed with microsomes prepared
from the livers of rats which were treated with the P450 3A inducer,
dexamethasone. The microsomes were calibrated against pure cytochrome
P450 3A. ANOVA was used to compare treatment group differences in the
liver and brain.
Withdrawal testing.
Naloxone (5 mg/kg s.c.) was administered
to naive neonates, or neonates chronically infused with saline or
fentanyl (100 µg/kg/hr) for 72 hr. After naloxone administration, the
animal was moved to a cage for a 25-min observation period. A cage
measuring 50 × 31 cm was marked with a grid of 30 squares (8 × 7 cm) for measurement of spontaneous activity. The average number of
lines crossed in 25 min was expressed as lines per animal. Other
behaviors were quantified as the number of animals exhibiting the sign
to the total number of animals observed. These behaviors included
micturition/defecation, face washing, wall climbing, abdominal
stretches, tremors, scream on touch and spontaneous vocalization.
Drugs and solutions.
Crystalline fentanyl hydrochloride
(National Institute on Drug Abuse, Bethesda, MD) was dissolved in
sterile pyrogen-free isotonic saline (Baxter Healthcare Corp.,
Deerfield, IL). Alzet 1003D osmotic minipumps (Alza Corp., Palo Alto,
CA) were filled with either fentanyl or isotonic saline. Animals were
anesthetized with methoxyflurane (Metofane®, Pitman-Moore, Mundelein,
IL). Wound autoclips (Clay Adams Co., Parsippany, NJ) were used to close the surgical incision. Potassium penicillin G (various suppliers) was diluted and injected i.p. to prevent infection.
| |
Results |
Fentanyl tolerance in neonatal rats.
Experiments were
conducted to test the hypothesis that continuous fentanyl
administration renders neonatal rats tolerant to fentanyl. Before
beginning the tolerance study, experiments were conducted to establish
the time course of fentanyl antinociception. Antinociception was tested
in naive PND9 rats with a dose derived from McLaughlin and Dewey
(1994). PND9 rats were used because this was the age at which fentanyl
tolerance was measured. In figure 2, the peak time
effect of fentanyl (40 µg/kg s.c.) was 10 min, with antinociception
lasting 40 min in PND9 rats.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 2.
Time course of fentanyl antinociception.
Opioid-naive PND9 rats were injected with saline s.c ( ) or fentanyl
( , 40 µg/kg) before testing for antinociception. Significant
differences between corresponding saline and fentanyl groups was
calculated by ANOVA followed by the Tukey's test (* P < .05).
|
|
For experiments on fentanyl tolerance, PND6 rats remained naive or were
surgically implanted with Alzet 1003D osmotic minipumps as detailed
under "Methods." After a series of implantation trials we
determined that PND6 rats (15-18 g) could receive the minipump (fig.
1). This pump has been successfully used to infuse insulin-like growth
factors I and II in PND10 rats (Glasscock et al., 1992). Rats in this study were infused with saline (1 µl/hr) or fentanyl at
100 µg/kg/hr for 72 hr. The infusion dose was based on an earlier study in this laboratory which indicated that 100 µg/kg of fentanyl corresponded to the ED84 value in the tail-flick test
(McLaughlin and Dewey, 1994). Seventy-two hours later, base-line
tail-flick latencies were measured before fentanyl dose-response curves
were generated in naive, saline- and fentanyl-infused rats. Base-line tail-flick latencies at 72 hr were not different among the groups, which indicated that fentanyl infused from the pump did not elicit antinociception (fig. 3A). However, the potency of
acutely administered fentanyl was significantly reduced in
fentanyl-infused rats compared with saline-infused rats (fig. 3B). Not
only were the ED50 values increased (table
1), but potency ratio calculations revealed a 4.1-fold
decrease in the potency of fentanyl. It is noteworthy that the potency
of fentanyl in saline pump-implanted rats was the same as naive
animals, which indicated the absence of a pump effect on
antinociception.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 3.
(A) Base-line tail-flick latencies in naive,
saline- and fentanyl-infused rats. PND6 rats remained naive or were
surgically implanted with 1003D Alzet osmotic minipumps. Seventy-two
hours after infusion of saline (1 µl/hr) or fentanyl (100 µg/kg/hr), base-line tail-flick latencies were obtained from PND9
rats before administration of fentanyl, as described in panel B. (B)
Tolerance to the antinociceptive effects of fentanyl. After obtaining
base-line tail-flick latencies, fentanyl was administered to naive
( , solid line), saline (, dashed line, 1 µl/hr)- or fentanyl
( , dotted line, 100 µg/kg/hr)-infused PND9 rats. Ten minutes
later, test latencies were obtained for calculation of %MPE. Each
dose-response curve represents 20 to 25 rats.
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 1
Tolerance to fentanyl in neonatal rats after a 72-hr infusion of
fentanyl (100 µg/kg/hr) from osmotic minipumps
PND6 rats remained naive or were implanted with an osmotic minipump
containing saline or fentanyl. Three days later, the rats were tested
for antinociception in the tail-flick test 10 min after s.c.
administration of fentanyl.
|
|
Role of gender in fentanyl tolerance.
The role of gender in
the degree of fentanyl tolerance was also examined. As seen in table
2, naive male and female rats were equally sensitive to
the acute antinociceptive effects of fentanyl. In addition, both male
and female fentanyl-infused rats were tolerant to the acute
antinociceptive effects of fentanyl 72 hr later. The ED50
values between tolerant male and female rats were not different, which
indicated the absence of an effect of gender in tolerance development.
View this table:
[in this window]
[in a new window]
|
TABLE 2
Role of gender in fentanyl tolerance in neonatal rats
PND6 rats remained naive or were implanted with an osmotic minipump
containing saline or fentanyl. Three days later, the rats were tested
for antinociception in the tail-flick test 10 min after s.c.
administration of fentanyl.
|
|
Role of postnatal development in fentanyl antinociception.
It
could be argued that postnatal development, and not tolerance,
contributed to the apparent reduction in the potency of fentanyl.
Therefore, experiments were conducted to compare the potency of
fentanyl in PND6 and PND9 rats. Dose-response curves and
ED50 values between PND6 and PND9 rats revealed no effect of postnatal development on the potency of fentanyl (fig.
4A, table 3). Although it was not
significant, the PND9 pups appeared to have an increased sensitivity to
fentanyl.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 4.
(A) Role of postnatal development in fentanyl
antinociception. Opioid-naive PND6 () and PND9 () rats were
injected s.c. with fentanyl 10 min before the tail-flick test. These
animals also received 0.045 µCi of 3H-fentanyl for
determination of brain fentanyl equivalent levels. Each dose-response
curve represents 30 to 35 rats. (B) Correlation between brain fentanyl
equivalent levels and fentanyl dose. Brain fentanyl equivalent levels
calculated for the rats in panel A were graphed as a function of
fentanyl dose.
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 3
Effect of rat age on fentanyl ED50 values and corresponding
brain EC50 values in neonatal rats injected s.c. with fentanyl
Antinociception in the tail-flick test was measured 10 min after s.c.
injection of fentanyl. Corresponding fentanyl equivalent levels were
determined by coadministration of 0.045 µCi of 3H-fentanyl
with unlabeled fentanyl.
|
|
In addition, it could be argued that postnatal development affected the
distribution of fentanyl to the brain, thereby reducing the potency of
fentanyl. Animals that were tested for antinociception to unlabeled
fentanyl in figure 4A, also received 3H-fentanyl (0.045 µCi). Our results demonstrate that the EC50 values
(i.e., nanograms of fentanyl equivalents/gram brain tissue eliciting 50% MPE) were similar in PND6 and PND9 rats (table 3). Furthermore, brain fentanyl equivalent levels increased in proportion to dose and did not differ between the groups (fig. 4B). These results
indicate that postnatal development did not affect the distribution of
fentanyl to the brain over the 72-hr infusion period.
Role of metabolism in fentanyl tolerance.
We also tested the
hypothesis that fentanyl tolerance reflected an increase in liver
and/or brain cytochrome P450 3A levels that more rapidly metabolized
acutely administered fentanyl. The livers and brains were removed from
naive, saline- and fentanyl-infused PND9 animals 72 hr after beginning
the experiment. However, the data reveal that cytochrome P450 3A levels
were not significantly elevated in the livers (F(2,13) = 0.552, P = .5908) or brains (F(2,13) = 1.442, P = .2824) of any treatment group (table 4). These results
indicate that chronic fentanyl administration did not induce an
increase in P450 3A levels, thereby ruling out the contribution of P450
3A metabolism to the expression of fentanyl tolerance.
View this table:
[in this window]
[in a new window]
|
TABLE 4
Comparison of tissue cytochrome P450 3A concentrations in naive, saline
and fentanyl pump-implanted PND9 animals
Microsomal membranes were prepared, and the concentration of P450 3A
was measured by an ELISA assay as detailed under "Methods." The
data are expressed as nanograms of P450 3A per mg protein ± S.E.
|
|
Physical dependence on fentanyl.
Other experiments were
conducted to test the hypothesis that continuous fentanyl
administration renders neonatal rats physically dependent on fentanyl.
Fentanyl-infused PND9 rats administered naloxone (5 mg/kg s.c.)
displayed a precipitated withdrawal syndrome (table 5).
Fentanyl-infused rats displayed the highest level of spontaneous
activity, which often resulted in wall climbing behavior. A percentage
of animals in each treatment group spontaneously micturated, whereas
defecation without diarrhea occurred in only fentanyl-infused rats.
Half of the fentanyl-infused rats exhibited abdominal stretching
similar to the visceral nociception elicited by
p-phenylquinone. Fentanyl-infused animals had severe tremors characterized by head turning and shaking. Finally, fentanylinfused neonates exhibited spontaneous vocalization characterized by periodic high-pitched screaming. During tactile stimulation while handling the
animals after the 25-min period, the animals emitted high pitched
screams.
View this table:
[in this window]
[in a new window]
|
TABLE 5
Behavioral profile of opiate withdrawal precipitated by 5 mg/kg
naloxone in PND9 rats
Rats remained naive or were infused 72 hr with saline (1 µl/hr) or
fentanyl (100 µg/kg/hr) from osmotic minipumps. PND9 rats injected
with naloxone (5 mg/kg s.c.) were observed for 25 min for signs of
physical dependence. These behaviors are indicated as the number of
animals exhibiting the sign to total number of animals observed.
|
|
| |
Discussion |
Fentanyl tolerance in neonatal rats.
Our results support the
hypothesis that continuous fentanyl administration renders neonatal
rats tolerant to fentanyl. Seventy-two hours after implantation of the
pumps, no antinociception was exhibited as the result of the fentanyl
infusion (fig. 3A). However, fentanyl-infused rats exhibited tolerance
to the acute antinociceptive effects of fentanyl (fig. 3B). To our
knowledge, this is the first reported attempt at rendering neonatal
rats tolerant to fentanyl or any phenylpiperidine opioid. Others have
demonstrated that neonatal and infant rats can become tolerant to
repeated injections of morphine (Fanselow and Cramer, 1988; Van Praag
and Frenk, 1991; Barr and Wang, 1992; Windh et al., 1995).
Repeated morphine injections may have been successful because
morphine's long duration of action resulted in the maintenance of
sufficient morphine tissue levels between each injection. Unlike
morphine, fentanyl's duration of action is less than 1 hr (fig. 2).
Thus, the development of fentanyl tolerance would require many
injections around the clock to provide sustained tissue levels for
several days. In addition, repeated injections would stress both the
neonates and the dam, a potential confound in the expression of
tolerance. The Alzet osmotic minipump was used to provide sustained
fentanyl brain levels, while limiting the handling stress of the
neonates and dams to a single time.
Although osmotic minipumps may provide sustained tissue opioid levels,
the investigator must determine the dose of drug to be infused from the
minipump. In nearly every report of opioid infusion in adult rats, the
minipump dose was derived empirically from ED50 values of
acute antinociception (Adams and Holtzman, 1990; Dierssen et
al., 1990; Paronis and Holtzman, 1992), or the antinociception
elicited by the minipump during the first day of infusion (Stevens
and Yaksh, 1989a, b). The infusion dose was based on an earlier study
in this laboratory which indicated that 100 µg/kg of fentanyl
corresponded to the ED84 value in the tail-flick test
(McLaughlin and Dewey, 1994). Yet, the empirical approach provides no
information on the brain levels achieved throughout the infusion
period. Measurement of brain and plasma opioid levels may provide one
approach to deciding minipump infusion dose. For example, in an early
study of morphine pellets, Patrick et al. (1975) initially
measured brain morphine levels associated with antinociceptive doses of
morphine. They demonstrated that a brain morphine concentration of 200 ng/g would elicit >80% antinociception. Brain morphine levels
achieved by the morphine pellet were shown to exceed 200 ng/g
throughout the 72-hr development of tolerance. In like manner, minipump
doses could be adjusted to provide brain opioid concentrations
equivalent to any degree of antinociception. Future studies in this
laboratory will include measurement of brain and plasma opioid levels,
so that minipump doses can be adjusted to provide sustained
antinociceptive concentrations.
It is noteworthy that the doses used in this model of neonatal rat
tolerance/dependence differ from those used in humans. Numerous studies
have shown that rodents require significantly higher opioid doses to
elicit antinociception. This is likely the result of pharmacokinetic
and pharmacodynamic differences between species. Furthermore, human
neonates and infants have been shown to require larger doses of
fentanyl than adults. For example, during extracorporeal membrane
oxygenation fentanyl is usually administered as a bolus dose of 5 to 20 µg/kg, followed by continuous fentanyl infusion ranging from 1 to 53 µg/kg/hr (Maguire and Maloney, 1988; Arnold et al., 1990;
Leuschen et al., 1993). The fentanyl doses used in human
neonates and infants are 4.5- to 32-fold higher than those used in
adults because of differences in pharmacokinetics and pharmacodynamics.
Role of gender in fentanyl tolerance.
It was also important to
assess whether gender contributed to the expression of neonatal rat
fentanyl tolerance. Several reports indicated that sexually mature male
and female rats are differentially sensitive to opioid antinociception
(Islam et al., 1993; Kepler et al., 1991), as
well as nonopioid antinociception (Kiefel and Bodnar, 1992).
Specifically, centrally administered morphine elicited a greater
magnitude of antinociception in male than in female rats (Kepler
et al., 1991). However, our results clearly demonstrated two
findings: 1) Naive male and female rats were equally sensitive to the
acute antinociceptive effects of fentanyl, and 2) fentanyl infusion
produced an equal degree of tolerance between male and female rats.
Thus, in the early postnatal period gender did not play a role in
fentanyl antinociception and tolerance. To our knowledge, the earliest
ages at which gender begins to affect opioid activity remains to be
determined.
Role of postnatal development in fentanyl antinociception.
Studies were also conducted to test the hypothesis that postnatal
development, and not tolerance, contributed to the apparent reduction
in the potency of fentanyl. However, our results indicate that
postnatal development did not affect the potency of fentanyl. Fentanyl
antinociception had both an early onset and short duration of activity
nearly identical with the effect of fentanyl in adult rats (Hug and
Murphy, 1981; Van den Hoogen and Colpaert, 1987; Millan, 1989). In like
manner, ED50 (95% C.L.) values for PND6 and PND9 pups were
no different than those of adult rats (Van den Hoogen et
al., 1988; Jang and Yoburn, 1991; Paronis and Holtzman, 1992).
Interestingly, the ED50 values in this study were nearly identical with those of PND3 rats (McLaughlin and Dewey, 1994). The
absence of differences in the distribution of fentanyl equivalents into
the brain also supports the absence of an effect of postnatal development on antinociception (fig. 4B). However, the apparent lack of
a role of postnatal development on fentanyl antinociception must be
addressed. In each of the cited studies, the radiant heat stimulus was
adjusted to elicit a 3- to 4-sec base-line latency in both neonatal and
adult rats. Obviously the radiant heat intensity must be increased as
animals age because of the enlargement of the tail and the degree of
keratinization. Yet by maintaining base-line latencies at 3 to 4 sec,
the potency of fentanyl appears to remain the same regardless of animal
age.
Our results also indicate that mu opioid systems are
sufficiently developed by the early postnatal period for neonatal rats to respond to fentanyl. In a study examining the opioid receptor ontogeny, rat brainstem mu opioid receptor density increased
progressively from PND1 to PND21. At PND21, opioid receptor density was
not significantly different from adult rats (Xia and Haddad, 1991). Thus, although opioid receptor density increases with age, PND6 and
PND9 responses to fentanyl did not change.
Role of metabolism in fentanyl tolerance.
It could be argued
that the fentanyl tolerance was caused by an increase in fentanyl
metabolism by liver and/or brain cytochrome P450 3A.
N-Dealkylation of fentanyl into nor-fentanyl by hepatic P450
3A is the major route of metabolism of fentanyl and related phenylpiperidines (Hug and Murphy, 1981; Lavrijsen et al.,
1990; Silverstein et al., 1993; Mautz et al.,
1994). Another pathway involves amide hydrolysis of fentanyl into
despropionyl fentanyl, which is subsequently converted by
N-dealkylation into despropionyl nor-fentanyl (Hug and
Murphy, 1981). P450 3A can also metabolize steroids, phenobarbital,
ethanol and some antibiotics. In addition, some of these agents have
been shown to induce P450 3A by increasing its enzyme activity and
tissue concentrations (Gonzalez, 1990; Ryan and Levin, 1990). To our
knowledge, no studies have determined whether fentanyl or related
phenylpiperidines can also induce P450 3A and alter the metabolism of
this class of opioids. There is one study demonstrating that P450
induction by chronic ethanol consumption in dogs increased the
metabolism of fentanyl (Gvozdenovic et al., 1993). Yet,
whether or not fentanyl could induce its own metabolism was unknown.
Our results provide two observations about P450 3A in neonatal rats.
First, PND9 rats possess measurable levels of P450 3A in the liver and
brain. Others have reported the presence of P450 3A in the liver of
PND4 rats and in the brains of adult rats (Jayyosi et al.,
1992; Borlakoglu et al., 1993). Second, our results indicate
that the chronic fentanyl infusion did not increase the levels of P450
3A. The ELISA method cannot address the possibility that fentanyl
induced an increase in P450 3A enzyme activity, although evidence
exists that fentanyl does not effect cytochrome P450 3A activity (Loch
et al., 1995). Finally, our results indicate that P450 3A is
not involved in the expression of fentanyl tolerance in neonatal rats.
Fentanyl-induced physical dependence.
Our results support the
hypothesis that continuous fentanyl administration renders neonatal
rats physically dependent on fentanyl. After naloxone administration,
the fentanyl-infused rats exhibited a robust withdrawal syndrome. The
primary signs were increased spontaneous activity, wall climbing
behavior, defecation, abdominal stretches, severe tremors, screaming to
touch and spontaneous vocalization. These signs were very similar to
those observed in chronic morphine-treated PND9 rats (Windh et
al., 1995). It can be argued that fentanyl-infused neonatal
behaviors constitute a true withdrawal syndrome. First, the similarity
of signs with Windh et al. (1995) suggests that physical
dependence was the result of chronic mu opioid receptor
activity. Second, Windh et al. (1995) reported that these
signs occurred in neonates undergoing either passive or precipitated
withdrawal. And third, these signs were triggered by naloxone only in
fentanyl-infused rats. The onset of signs began almost immediately,
peaked by 10 min and were reduced in severity by 25 min. Other signs
such as weight loss and ultrasonic vocalization have been reported in
neonatal rats chronically treated with morphine (Fanselow and Cramer,
1988; Barr and Wang, 1992). These behaviors did not resemble the
classic withdrawal syndrome observed in adults, which includes wet dog shakes, jumping, teeth chattering, piloerection and ptosis (Blasig et al., 1973; Wei et al., 1973). Therefore, even
though neonates are developmentally incapable of expressing adult signs
of dependence, neonates are capable of expressing signs appropriate to
their age.
Summary.
In conclusion, chronic fentanyl administration
via osmotic minipumps appears to render neonatal rats
tolerant and dependent on fentanyl. Tolerance was neither the result of
postnatal development nor of an increase in the metabolism of fentanyl
by P450 3A. Future studies will be conducted to examine the long-term
consequences of chronic postnatal opioid exposure, so that the
potential effects of iatrogenic tolerance and dependence in human
infants can be examined.
PND, postnatal day;
s.c., subcutaneous;
%MPE, percent maximal possible effect;
ANOVA, analysis of variance;
ED50, 50% effective dose;
EC50, 50% effective
concentration;
ELISA, enzyme-linked immunosorbent assay;
C.L., confidence limits.
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Top
Abstract
Introduction
