Morphine and fentanyl are widely used opioid analgesics for pain control. Despite the usefulness of these drugs, each has some serious drawbacks. For example, morphine-induced constipation is a serious systemic side effect that affects 80% of patients1 and can have a negative impact on both the patient's quality of life2 and adherence to their drug regimen.3 A delayed time of onset can also be an issue in outpatient use of opioids in conditions such as breakthrough cancer pain. Oral morphine and buccal fentanyl are frequently prescribed in treatment of breakthrough cancer pain4 and typically require 22 to 30 minutes or 5 to 10 minutes,5,6 respectively, to initiate analgesic effect. Directly targeting these drugs to the central nervous system (CNS) could significantly reduce side effects and improve time to analgesic onset with these drugs.
A fraction of intranasally applied drugs may rapidly distribute directly from the nasal cavity to the CNS, effectively bypassing the blood-brain barrier (BBB) resulting in faster onset times and lower systemic concentrations.7 Intranasal morphine instillation in rats has resulted in approximately 10% direct transport from the nasal cavity to the CNS resulting in significantly larger brain concentrations of morphine in the brain and less systemic exposure compared with IV administration.8 Using this route of delivery with opioid pain medications could improve the time of onset and side-effect profile of these drugs.
Despite its potential, direct “nose-to-CNS” distribution is underutilized in clinical settings. This is attributable to a lack of drug delivery devices to reach the olfactory region. The standard nasal pumps used in clinical trials for intranasal morphine and fentanyl typically deposit <5% of the dose in the olfactory region.9 The olfactory epithelium, which is located between the nose and the brain, is only 3% to 10% of the surface area of the nasal cavity in humans.10,11 Its position in the upper nasal cavity with turbinate restriction makes consistent delivery of drug to this region challenging. The olfactory region in the upper back quantile of the nasal cavity encompasses a narrow, approximately 1- to 2-mm-wide slit. Typical nasal sprays tend to have a wide aerosol plume angle that mainly deposits on the respiratory epithelium9 located at the lower nasal cavity. Drugs deposited in the lower nasal cavity and respiratory epithelium could potentially penetrate through the trigeminal pathway,12 but evidence for this pathway is more limited.
Therefore, we have developed a pressurized olfactory delivery (POD) device that is able to deposit a majority of drug at the olfactory region of the nasal cavity in rats. The objective of this study was to determine whether increasing the fraction of the opioids morphine and fentanyl deposited at the olfactory region of the nasal cavity would lead to a larger fraction of drug being directly transported from the nasal cavity to the CNS. If so, intranasal morphine and fentanyl could be developed with faster onset times and potentially fewer systemic side effects. We determined that preferential nasal olfactory region deposition of morphine and fentanyl produced more rapid analgesic onset and significant direct nose-to-CNS distribution compared with either nasal respiratory deposition or systemic administration in rats.
Morphine sulfate and fentanyl citrate were purchased from Sigma-Aldrich (St. Louis, MO). Beuthanasia-D (Schering-Plough Animal Health Corp., North Chicago, IL) was used for euthanasia at the end of the distribution studies. The drug doses were dissolved in 0.9% saline solution (Hospira, Inc., Lake Forest, IL). Morphine and fentanyl liquid chromatography / mass spectrometry (LC/MS) standards were purchased from Cerilliant (Round Rock, TX). All other materials used were reagent grade.
Adult male Sprague-Dawley rats (weight, 200–300 g; Harlan, Indianapolis, IN) were housed under a 12-hour light/dark cycle with Harlan Teklad rat chow and water provided ad libitum. Animals were cared for in accordance with institutional guidelines, and all experiments were approved by the University of Washington Animal Care and Use Committee under IACUC #2372-12.
Preparation of Drug Formulations
Morphine and fentanyl were stored at −20°C as a lyophilized powder. On the day of each experiment, the necessary doses of morphine and fentanyl were solubilized in 0.9% saline solution (Sigma-Aldrich, St. Louis, MO). Each formulation had a pH of 7.0 for each of the doses tested. The nasally administered doses were given in volumes of 7 to 11 μL and the intraperitoneal (IP) doses were given in volumes of 0.45 to 0.55 mL depending on the weight of the animal.
POD Device Construction and Characterization
The overall construction of the POD nasal aerosol device is diagrammed in Figure 1. A pressurized nitrogen gas supply was connected to a standard 2-valve pressure regulator. Plastic tubing with a 200-psi pressure rating connected the pressure regulator to the inflow connection of a pneumatic solenoid (Cramer Decker Industries, Santa Ana, CA). The solenoid was regulated with a foot pedal–actuated GraLab 555 digital timer (GraLab Corp., Centerville, OH). On the outflow connection of the solenoid was a 3-mL cylinder, which made up the device body and fit securely to the solenoid. On the end of the device body was a custom-fit aerosol nozzle with a 0.8-mm outside diameter. The nozzle was fitted with an aerosol insert composed of a small length (2.0 mm) of metal cylinder, which had 2 spiral fluid passages in which the fluid/gas mixture traveled. This served to mix the nitrogen and liquid drug to create an aerosol output to enhance penetration into the nasal cavity toward the cribriform plate area while minimizing pressure on the nasal epithelia. Finally, a fluorinated ethylene-propylene liner was placed over the outside of the metal tip to protect the nasal epithelia from being damaged by the nozzle during use. A fluorescence assay was used to determine what percentage of the total desired volume was dispensed from the device with each actuation. A solution of 50 μg/mL fluorescein was dispensed from the device into a well in a 24-well plate that was prefilled with 1 mL deionized H2O. After collection of the aerosol, the solution was mixed by pipetting up and down several times. The fluorescence of the solution was measured on a PerkinElmer 1420 multivariable fluorescence plate reader (PerkinElmer, Waltham, MA). The fluorescence signal of the collected volume was compared with the expected signal to calculate the percent of drug expelled from the device with each actuation.
The basic operation of the POD device in rats was as follows: the animal was anesthetized with isoflurane, the dose was loaded into the device, and the nozzle was carefully placed 8.0 mm into the rat's nasal cavity and pointed in the direction of the cribriform plate. The foot pedal switch was then pressed to actuate the solenoid for 0.1 second to discharge the aerosol dose.
Nose Drop Administration Targeting the Nasal Respiratory Region
To determine any differences in opioid distribution after enhanced olfactory deposition, a simple nose drop method was developed as a control route of administration specifically targeting the nasal respiratory region. The rats were anesthetized with isoflurane and retained in a natural position resting on their stomachs. Drops of 3 μL were placed near the rat's naris and naturally inhaled into the nasal cavity. This was repeated in the opposing naris every 20 seconds until the required dose volume was administered.
Nasal Deposition Experiments
Deposition was tested using both the POD device and nose drops. Under pentobarbital anesthesia, the rats (n = 6) received 10 μL dye solution of 0.1% Coomassie blue with the POD device or nose drop administration. Shortly after administration was complete (<5 minutes), the animals were overdosed with 250 mg/kg pentobarbital. The nasal cavity was then bisected at the septum, the septum was removed, and the tissues were examined grossly for dye localization. The nasal tissues were then decalcified, sectioned, and stained with hematoxylin and eosin stain according to the method of Young.13 These sections were then examined under a Zeiss Axiovert 200 fluorescent microscope (Carl Zeiss, Inc., Jena, Germany) using the methods of Chandler et al.14 to analyze the epithelial integrity.
Tail-Flick Analgesia Experiments
Each group of animals (n = 9) initially underwent 3 days of placebo testing to obtain a baseline reading for the tail-flick test and to acclimate the animals to handling. Each animal was exposed to 5% isoflurane in an induction box for 2.5 minutes. The animal was then removed and given a 10-μL dose of 0.9% saline solution, pH 7.4, via nose drops, POD, or IP injection. After the placebo dose was administered, the animals were allowed to fully wake from the isoflurane in a padded tray. Then, at 5, 10, 30, 45, 60, 90, and 120 minutes, each animal was wrapped gently in a towel, their tail was placed in room temperature water (18°C ± 0.5°C) for 5 seconds, the tail was quickly dried, and then the distal 3 cm of the tail was placed in 55°C ± 0.5°C water. The time until tail removal was measured with a digital stopwatch.
After the initial placebo trials, the same procedure was repeated 3 times, every other day over 5 days, with each rat receiving a single dose of either morphine or fentanyl by nasal spray, nose drops, and IP injection in a randomly chosen order. The cutoff time, at which the tail would be removed from the water to prevent tissue damage, was set at 10 seconds for all tail-flick trials.
Three days after a group of animals completed the tail-flick studies, they were cannulated in preparation for and immediately before a plasma pharmacokinetic experiment. Animals were anesthetized with 2% isoflurane (Novaplus; Hospira). Body temperature was maintained at 37°C by a heating pad (Fine Science Tools, Inc., Foster City, CA). For pharmacokinetic experiments, the femoral vein was cannulated for blood draw with PE-10 polyethylene tubing (Becton, Dickinson and Company, Franklin Lakes, NJ) connected to a blunt-tipped 23-gauge needle (Becton, Dickinson and Company). The PE-10 tubing was inserted 3 cm into the femoral vein to ensure that blood sampling was from the vena cava. Animals were maintained on anesthesia for 30 minutes before drug administration and pharmacokinetic analysis. After the final blood draw, the catheter was removed from the femoral vein, the proximal end of the femoral vein was tied off with suture string to assure hemostasis, and the incision stitched with suture string (Harvard Apparatus, Holliston, MA).
After a 30-minute resting period under anesthesia after surgery, the animals (n = 3) were administered a single dose of either morphine (1.0, 2.5, or 5.0 mg/kg) or fentanyl (7.5, 15.0, or 25.0 μg/kg). Then, at 5, 10, 30, 45, 60, 90, and 120 minutes after drug administration, 300 μL of blood was drawn from the femoral vein catheter. The blood was collected in a 1-mL syringe (Becton, Dickinson and Company) and transferred to a microcentrifuge tube for blood/plasma separation. The tubes were immediately centrifuged at 8000g for 5 minutes, and then the plasma was removed and frozen on dry ice. At the end of the experiment, 2.0 mL of sterile 0.9% saline solution was injected via the femoral vein catheter to replace the removed blood volume.
CNS Tissue Distribution Experiments
One week after the pharmacokinetic experiments, the animals used in the pharmacokinetic studies were used in CNS tissue distribution experiments. The animals (n = 6) were briefly anesthetized with a 2.5-minute exposure to 5% isoflurane. A dose of morphine (2.5 mg/kg) or fentanyl (15 μg/kg) was administered via POD, nose drops, or IP injection. At 5 minutes, the animals were killed with an overdose of Beuthanasia-D. Immediately after death, the animal was decapitated. A syringe was used to collect 1.0 mL of blood from the trunk. This plasma was processed in the same way as the plasma from the pharmacokinetic experiments, and was used in the brain/plasma concentration ratio data. The base of the skull and the parietal bones were quickly removed. The brain was removed within 2 minutes of death. The brain was dissected into the cortex, diencephalon, brainstem, cerebellum, and olfactory bulbs. The cervical spinal cord was also collected from the body. The olfactory bulbs were the final brain tissues collected.
Analysis of Morphine and Fentanyl by LC/MS
A fixed volume (20 μL) of either morphine-D6 or fentanyl-D6 (Cerilliant, Palo Alto, CA) was added into each tissue and plasma sample to act as an internal standard. Tissue samples were homogenized in 5- to 10-times volume of 0.1 M borate buffer, pH 8.9, and centrifuged for 10 minutes at 1000g. The tissue supernatant and plasma samples were passed over Certify solid-phase extraction cartridges (Varian, Palo Alto, CA) and eluted with methylene chloride/isopropanol/ammonium hydroxide (80:20:2). After elution, the samples were evaporated under N2 gas until dry. The samples were resuspended in 75 μL of mobile phase that consisted of 92% of 0.05% acetic acid and 8% acetonitrile for morphine samples or 40% 10 mM ammonium acetate and 60% acetonitrile for fentanyl samples. An Agilent HPLC/MS series 1100 series B with autosampler (Agilent Technologies, Inc., Santa Clara, CA) was used for quantification. The injection volume was 5 μL. The morphine samples were passed over a Zorbax SB-C8 column (Agilent Technologies) with a flow rate of 0.25 mL/min. The ionization setting was API-ES in the positive ion mode with a capillary voltage of 1400 V.
For both morphine and fentanyl, a standard curve was created on the day of analysis according to the same process described for the samples. Each standard curve was linear with a coefficient of linear regression R2 > 0.99. In addition, 2 quality-control samples with a known amount of drug were processed on the day of analysis to ensure day-to-day consistency of the analytical assay.
All tail-flick test values are presented as a percentage of maximal possible effect (%MPE), which is defined as:
The cutoff time was set at 10 seconds in the tail-flick studies to prevent tissue damage from occurring. Area under the curve (AUC) values from all experiments were calculated using the trapezoidal rule without extrapolation to infinity. Pharmacokinetic values were obtained from 1-compartmental modeling in WinNonlin (Pharsight Corp., Mountain View, CA). Tail-flick data were compared using repeated-measures analysis of variance (ANOVA). Plasma and tissue concentrations were compared using a 1-way ANOVA with a Tukey posttest. AUCeffect/AUCplasma ratios were calculated from individual animals so they could be statistically compared with a 1-way ANOVA. All statistical analyses were performed using SigmaPlot software version 11.0 (Systat Software, Inc., San Jose, CA).
A direct transport percentage (DTP%) to the brain was calculated to determine the amount of drug in the brain that was distributed directly from the nasal cavity to the CNS. Because analgesic effect has been shown to correlate well with morphine concentrations in the extracellular brain fluid,15 we used analgesic effect in place of brain concentrations. The DTP% is used to estimate the amount of drug in the brain that cannot be accounted for by systemic distribution. The DTP as defined was calculated as follows16:
We also performed hysteresis analysis to determine the time delay in plasma drug concentration with corresponding analgesic effects observed typically for systemic administration. This analysis would reveal whether the POD device-mediated-enhanced opioid delivery to the CNS affects the time lag of opioid analgesic effects compared with after oral or IV administration. To do so, plasma opioid concentration for any time point within the time course of plasma concentration was plotted against the effects observed at that specific time point for each dose of morphine or fentanyl. The time sequences were included as arrows in these plots for rats treated with morphine or fentanyl. These plots were used as a mean to evaluate consequence of direct delivery of opioid by POD device to CNS delivery as drug traveling from the plasma to the CNS would give rise to a faster onset in plasma drug concentration before or at the same time as rise in analgesic effect.
Evaluation of Drug Administration Methods
We first evaluated the dose residuals and nasal deposition of the POD device. Using a fluorescence assay to measure residuals, the POD device dispensed a 10-μL dose with 99.3% ± 2.6% accuracy. Using Coomassie blue dye, the deposition within the rat's nasal cavity was determined after delivery with either the POD device or our nose drop method. When using 10 μL of dye, the POD administration resulted in deposition on the olfactory epithelium area of the nasal cavity with none or very little observed on the respiratory region, trachea, or esophagus. Administering the dye by our nose drop method resulted in deposition primarily on the respiratory epithelium. No noticeable dye staining was apparent in the olfactory region, or in the trachea or esophagus. In the nasal epithelial histopathology assessment of membrane integrity, there was no noticeable nasal damage observed in the POD-administered naris compared with the controlled naris.
Time Course of Drug Concentrations and Analgesic Effects
As shown in Figure 2A, IP administration of 5.0 mg/kg morphine in rats resulted in significantly higher plasma concentrations compared with those treated with an equivalent dose using nose drops or the POD device. The plasma morphine concentrations after IP administration were significantly higher (P < 0.05) during the first 30 minutes with a Tmax(plasma) at 20.0 minutes with a Cmax of 597.6 ng/mL (Table 1). The plasma morphine concentrations at 10 minutes were approximately 2.6-fold higher in IP-treated rats than in those treated with either type of nasal administration. The only point with a significant difference between the POD device and nose drop–treated rats (P < 0.05) was at 5 minutes, when the plasma concentration in the POD device group was nearly 2-fold higher than that in the nose drop group. Both the POD spray and the nose drops resulted in a Tmax(plasma) of approximately 30 minutes and Cmax values of 314.5 and 297.4 ng/mL, respectively (Fig. 2). Although there was a significant trend of higher plasma concentrations after IP administration in the first 45 to 60 minutes, there were no statistical differences among the 3 test groups in overall plasma morphine exposure (AUC0–120) for animals treated with the 2 low doses of 1 or 2.5 mg/kg (Table 2). At 5.0 mg/kg morphine, AUC0–120 for rats treated with IP administration was significantly higher compared with those treated with either type of intranasal administration (Table 2).
The tail-flick analgesic test revealed that at 5.0 mg/kg morphine, rats treated with the POD device compared with the nose drops exhibited a significantly higher analgesic effect (P < 0.05) during the first 60 minutes (Fig. 2C). In addition, POD device treatment resulted in a significantly higher analgesic effect (P < 0.05) than IP injection at the 5-minute time point. The animals treated with morphine in the POD device exhibited Tmax(effect) at 10 minutes and remained at approximately the same level during the next 50 minutes. The overall effects, AUC0–120 for animals treated with morphine in either the POD device or by IP injection, were significantly greater (P < 0.05) than those for animals treated with nose drops (Table 2). There was no significant difference in analgesic AUC between rats treated with the POD device or IP administration after 5.0 mg/kg morphine.
We next tested the effects of the POD device in delivery of a more hydrophobic opioid, fentanyl. The plasma concentrations of 15 μg/kg fentanyl were significantly higher (P < 0.05) after both POD and nose drop administration than after IP administration over the course of 120 minutes (Fig. 2B). Although the 7.5 and 25 μg/kg doses led to similar relative plasma profiles as the 15 μg/kg dose, they did not lead to significant plasma concentration differences between routes of administration. In rats treated with 15 μg/kg fentanyl with either the POD device or by IP injection, the Tmax(plasma) was very rapid, 1.70 and 1.46 minutes, respectively (Table 3). The plasma concentrations for rats treated with fentanyl nose drops resulted in a Tmax(plasma) at 6.50 minutes with high variation, but all animals had the same experimental Tmax(plasma) and overall shape of the plasma fentanyl curve. This rapid onset and decrease was also observed for the 7.5 and 25 μg/kg doses after POD administration. At the 5-minute time point, the plasma concentrations were 5 and 6.7 times higher for the nose drop and POD device groups compared with the IP injection group.
POD administration of 15 μg/kg fentanyl to the olfactory region of the nasal cavity resulted in a much greater analgesic effect (Fig. 2D) at 5 minutes compared with nose drop or IP administration (P < 0.05 vs nose drops and P < 0.01 vs IP administration). It should be noted that at the 5-minute time point, more than half of the POD-treated rats did not pull their tails from the water bath before the 10-second maximal cutoff time (to prevent tissue damage). Therefore, the analgesic effect from the POD spray at 5 minutes was underestimated. After 5 minutes, however, the analgesic effect from the POD spray decreased quickly. After this initial time point, there was no significant difference in the analgesic effects among the 3 routes of administration. At the 5- and 10-minute time points, animals treated with fentanyl nose drops also exhibited a significantly higher analgesic effect (P < 0.05) than those treated by IP injection. The overall fentanyl effects, AUC0–120 for POD device treatment, were significantly less (P < 0.05) than those with either nose drop or IP treatment (Table 4).
Comparison of Plasma Drug Exposure and Overall Analgesic Effects
POD administration of morphine resulted in a significantly lower plasma AUC after a 5.0 mg/kg dose and a nonsignificantly lower plasma AUC at 2.5 mg/kg (Table 2). POD administration resulted in nonsignificantly higher plasma AUC at the 1.0 mg/kg dose. In addition, POD administration resulted in a significantly higher effect AUC compared with nose drops at 1.0 and 5.0 mg/kg doses and a significantly higher effect AUC compared with IP administration after a 1.0 mg/kg dose. POD administration also resulted in 1.64-, 1.61-, and 2.24-fold increases in the AUCeffect/AUCplasma ratio compared with IP administration after 1.0, 2.5, and 5.0 mg/kg doses at each dose tested (Table 2). There was no significant difference in this AUC ratio between IP and nose drop administration. The DTP% after POD administration was estimated to be 38.5%, 38.1%, and 55.0% after 1.0, 2.5, and 5.0 mg/kg doses, respectively. The DTP% after nose drop administration was estimated to be 0%, 10.9%, and 14.8% after 1.0, 2.5, and 5.0 mg/kg doses, respectively.
For fentanyl, because neither nose drop nor POD administration led to significantly higher AUCeffect/AUCplasma ratios compared with systemic IP administration, all of the opioid drug causing analgesic effect could be accounted for from blood distribution, and thus, a DTP% was not calculated. In addition, there were no significant differences between the plasma-normalized brain concentrations after nose drops, POD spray, or IP administration (Fig. 3). In both the forebrain and midbrain-cerebellum-brainstem-spinal cord (MCS), the POD spray resulted in a nonsignificant trend of higher blood-to-brain ratios at 5 minutes after administration of 15 μg/kg fentanyl.
Hysteresis Analysis to Evaluate First Passage of Drug from Nose to CNS
To evaluate plasma drug concentration–effect response in relationship to time-course plasma drug concentrations, we performed hysteresis analysis by plotting the plasma drug concentration versus analgesic effects according to the time sequence for rats treated with morphine or fentanyl. These plots are useful in determining any direct to CNS delivery because drug traveling from the plasma to the CNS should result in an increase in plasma concentration before or at the same time as an increase in analgesic effect. Any direct nose-to-CNS distribution should result in an increase in analgesic effect preceding an increase in plasma concentrations. As shown in Figure 4, several differences among rats treated with nose drops, POD device, or IP injection become apparent. The point on each graph at (0,0) represents the baseline tail-flick test before dosing. It is semiartificial and was primarily added for easier visualization of the data sequence. After a 2.5 mg/kg morphine dose (Fig. 4, A–C), both nose drop and IP administration resulted in a counterclockwise hysteresis, whereas the POD administration resulted in a clockwise hysteresis. However, rats treated with 15 μg/kg fentanyl nose drops exhibited no discernible hysteresis (except the baseline point). Those treated with the POD device exhibited a clockwise hysteresis whereas the IP group showed a counterclockwise hysteresis (Fig. 4, D–F). All 3 fentanyl effect-concentration curves were plotted on the same axis scale, so the plot for IP administration appears much smaller because of low plasma fentanyl concentration and analgesic effect.
Drug Concentration in the CNS Tissues
CNS tissue concentrations were determined as a secondary measure of CNS drug distribution in addition to analgesic effect. The upper cervical spinal cord was processed with other parts of the brain. The brain concentrations of morphine at 5 minutes after POD administration of 2.5 mg/kg to the olfactory region of the nasal cavity were higher than either nose drop administration to the respiratory region of the nasal cavity or IP injection (Fig. 5). Although after POD device treatment the olfactory bulb morphine concentration was much higher, 7.60 or 39.58 times that of nose drop or IP administration, the results did not reach significance. However, POD device–treated animals exhibited higher morphine brain (excluding the olfactory bulbs and including the upper cervical spinal cord) concentrations at 5 minutes (P < 0.05). Administration with the POD device to the nasal olfactory region resulted in 1.68 times the brain concentration after nose drops and 3.38 times the brain concentration after IP administration. There was no significant difference between brain concentrations at 5 minutes after nose drop or IP administration.
Similar differences were observed when the brain tissue morphine concentration was normalized by plasma morphine concentration (Fig. 6). Again, the POD device provided a trend toward larger blood-normalized olfactory bulb morphine concentrations (but the data were not statistically significant). In the remainder of the brain, the POD device provided significantly higher (P < 0.05 vs nose drops and P < 0.01 vs IP injection) blood-normalized brain morphine concentrations. The blood-normalized brain morphine concentrations at 5 minutes for POD device– treated animals were 2.27 and 6.48 times the concentrations of those treated with nose drops and IP injection.
Nasal delivery to the olfactory region 5 minutes after a dose of 15 μg/kg fentanyl with the POD device resulted in significantly higher (P < 0.05) concentrations in the brain than after either nose drop administration to the respiratory region or IP injection (Fig. 7). In this study, the forebrain included the olfactory bulbs. At 5 minutes, the POD nasal device with fentanyl resulted in 1.63-fold increased concentration compared with nose drop application and a 5.61-fold increase compared with IP administration. The results in the MCS were similar. POD spray application resulted in 1.56 and 5.32 times the fentanyl concentration in the MCS when compared with nose drop administration or IP injection, respectively. In addition, animals treated with the POD device exhibited significantly larger (P < 0.05) plasma concentrations at 5 minutes after 15 μg/kg fentanyl administration. The POD device resulted in 1.28 times higher plasma fentanyl concentrations compared with nose drops and a 4.5 times higher concentration than IP administration.
We have developed drug administration techniques in rats to specifically deposit drug at the nasal olfactory region (POD) or the nasal respiratory region (nose drops). Using morphine and fentanyl, which exhibit significantly different hydrophobicity and ability to penetrate the BBB, we found that there are significant differences in distribution between drug delivery to the nasal respiratory epithelium compared with the nasal olfactory epithelium for both fentanyl and morphine. In this study, we used the analgesic effect, as determined by the tail-flick latency test, as a measure of the opioid drug distribution to the CNS. Pharmacokinetic studies have determined the use of effect to be a suitable surrogate for extracellular brain concentrations.17,18 The correlations between plasma pharmacokinetics and effect allow a mechanism to determine any direct nose-to-CNS distribution after olfactory nasal administration. Delivering a more hydrophilic morphine to the olfactory region of the nasal cavity with the POD device resulted in increased direct transport from the nasal cavity to the CNS, increasing the analgesic effect while maintaining low plasma values. However, POD-administered fentanyl deposition on the olfactory region resulted in a very strong and rapid analgesic effect that could be related to the first pass of drug to the brain that readily cleared into the blood. As a result, the overall time course of plasma drug concentration did not appear to differ significantly from that of nose drop nasal administration (Fig. 2).
The morphine data demonstrate that in addition to morphine uptake into the blood from the nasal olfactory region, which was similar to uptake at the respiratory region, there was further increased uptake into the CNS after deposition at the olfactory region of the nasal cavity. This additional uptake into the CNS observed after olfactory delivery was rapid. This direct transfer is likely attributable to passive diffusion along the fluid-filled channels that connect the olfactory region of the nasal cavity with the subarachnoid space surrounding the olfactory bulb.19 At the initial time point measured, POD administration did not result in higher plasma concentration, but did result in a significantly higher analgesic effect compared with the other 2 routes of administration. The data were plotted as plasma concentration versus effect hysteresis curves to more easily visualize the relationship among the changes in plasma levels and brain levels to distinguish differences in distribution among the 3 routes of delivery.
This rapid distribution into the CNS after POD administration is seen in the plasma concentration versus analgesic effect curves, where POD administration of morphine resulted in a clockwise hysteresis curve compared with the other 2 routes, which displayed a clockwise hysteresis. This counterclockwise hysteresis could be explained by either acute tolerance to morphine or to the pain test, or it could be attributable to direct and rapid distribution to the brain from the nasal cavity.20 The clockwise hysteresis observed after POD administration of morphine does not seem to be attributable to sensitization,21 which would be observed as a decreased effect at the later time points, and is instead likely to be primarily attributable to a rapid increase in effect before the increase in plasma concentration. Thus, this clockwise hysteresis after administration with the POD device is more likely attributable to direct transport from the nasal cavity to the CNS, such as that observed by Westin et al.8,22
This conclusion is supported by the brain-to-plasma ratios observed 5 minutes after administration with the POD device, which were significantly higher than those observed after nose drop or IP administration. An increased brain-to-plasma concentration ratio at this early time point indicates that the drug distributed into the brain through an alternate nonplasma route such as direct nose-to-CNS distribution. This increase in the brain/plasma ratios, especially at the earliest time points, was also noted in a study by Westin et al.8 in which brain-to-plasma AUC ratios of 0.5 were observed in the first 10 minutes after nasal administration. Morphine was visually observed with autoradiography to distribute from the nasal cavity to the lamina propria surrounding the olfactory bulb. In our study, a large difference was also observed in the olfactory bulb morphine concentrations; however, these differences were not significant, likely because of the method of collection of the olfactory bulb and the small sample size (n = 3).
Westin et al. calculated the percentage of drug delivered directly to the brain, or DTP%, over the course of 240 minutes. They reported that from 0 to 60 minutes and 0 to 240 minutes, the DTP% morphine was 48% and 10%, respectively. In this study, we found that the DTP% after nose drops was between 0% and 10% depending on the dose; however, this may not be significant because the AUCeffect/AUCplasma ratios were not significant, whereas after POD administration, the DTP% was between 38% and 55%. It is difficult to directly compare our DTP% values with those of Westin et al. because they used brain concentrations whereas we used analgesic effect. They also did not calculate a 0- to 120-minute DTP%. However, it is clear from our study that POD delivery to the olfactory region resulted in a higher DTP% compared with nose drop administration to the respiratory epithelium.
In contrast to the enhanced CNS distribution observed after POD administration of morphine to the nasal olfactory region, nose drop distribution to the nasal respiratory region displayed little to no evidence of direct transport from the nasal cavity to the brain. The AUCeffect/AUCplasma ratios after respiratory epithelium and IP administration were not significantly different across the doses tested and resulted in a counterclockwise hysteresis typical of systemic administration. IP administration was chosen as a systemic method of administration because the plasma time-course profile of morphine is similar after nasal administration and IP administration15; thus, the AUCs can be directly compared. These data indicate that morphine applied to the respiratory epithelium distributed primarily into the capillaries of the nasal cavity where it was distributed to the CNS across the BBB.
Administration of lipophilic fentanyl (log P = 3.9) to the olfactory region resulted in a different pattern of distribution compared with hydrophilic morphine (log P = 0.8). Immediately after POD administration of fentanyl, the plasma and analgesic levels were very high and then both decreased dramatically. In fact, POD-administered doses of 7.5 and 15.0 μg/kg led to significantly lower AUCeffect compared with nose drop administration. The hysteresis curve after POD administration of a 15.0 μg/kg dose indicated slight hysteresis but only at the initial 5-minute point, and this seems to have been caused by the dramatic decrease in analgesic effect with the slight decrease in plasma effect. Although this could indicate some direct transport of fentanyl from the olfactory region to the CNS, this conclusion is not supported by the measured brain tissue concentrations. The forebrain and the MCS contained higher fentanyl concentrations 5 minutes after POD administration to the olfactory region. However, when normalized by plasma, there was no statistical difference among any of the routes of administration. In addition, both forms of nasal administration resulted in lower AUCeffect/ AUCplasma ratios compared with IP delivery, indicating that all 3 routes of administration resulted in CNS distribution from the bloodstream.
Our data suggest that POD administration of fentanyl to the olfactory region of the nasal cavity primarily leads to uptake into the capillaries of the olfactory epithelium followed by distribution to the CNS across the BBB. Fentanyl is a very lipophilic compound, with a nasal bioavailability of 71%.23 Thus, fentanyl applied to either the respiratory or olfactory epithelium would likely be rapidly absorbed into the bloodstream leaving little drug to distribute directly to the CNS. The rapid redistribution of fentanyl would likely make any direct nose-to-CNS delivery difficult to detect. Investigators conducting other studies of nasally administered lipophilic drugs such as estradiol24 and lidocaine25 have also been unable to observe any direct distribution from the nasal cavity to the CNS. Interestingly, we observed an increased level of analgesia and brain concentrations 5 minutes after POD administration to the olfactory epithelium compared with nose drop delivery to the respiratory region of the nasal cavity. One possible explanation is that the higher density of blood vessels in the lamina propria of the olfactory region26 could lead to a greater rate of initial fentanyl uptake into the blood.
There were significant differences in distribution after deposition on the nasal olfactory or nasal respiratory epithelium with both drugs. Interestingly, the plasma drug concentrations and effect profiles of both fentanyl and morphine after nose drop administration to the respiratory epithelium were the most similar to values obtained from human nasal administration of these opioid drugs,27,28 with morphine having Tmax plasma values between 10 and 30 minutes, and fentanyl having Tmax values between 5 and 10 minutes. This is not surprising considering that most of these trials use standard nasal pumps, which primarily deposit drug on the respiratory epithelium.29 Thus, low-volume (<15-μL) nose drop administration using our methods in rats may be an appropriate model for nasal administration in humans with a standard nasal pump. In contrast, POD distribution to the olfactory region in rats could be representative of localized olfactory delivery within the nasal cavity in humans.
POD administration of morphine resulted in an increased analgesic effect compared with respiratory epithelial deposition, and lower plasma values compared with systemic administration. Thus, administering morphine to the olfactory region of the nasal cavity could be an effective, noninvasive method of opioid administration, which could decrease systemic side effects such as opioid-induced constipation. Olfactory delivery of fentanyl resulted in a very high immediate analgesic effect because of rapid distribution into the bloodstream. This could be of clinical value to reduce the onset time and dose required for outpatient fentanyl use.
Name: John D. Hoekman, PhD.
Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.
Conflicts: John D. Hoekman has equity interest in Impel NeuroPharma, Inc. (Seattle, WA). The results in this work, along with others, prompted the founding of a startup company, Impel NeuroPharma, to create nasal drug delivery devices to better penetrate the olfactory region of the nasal cavity. In full disclosure of conflict of interest, both authors are inventors on a patent for the POD nasal spray device and both authors also own stock in a startup company, Impel NeuroPharma, founded to develop olfactory-targeted nasal drug delivery devices.
Attestation: John D. Hoekman has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Rodney J. Y. Ho, PhD.
Contribution: This author helped design the study, analyze the data, and write the manuscript.
Conflicts: Rodney J. Y. Ho has equity interest in Impel NeuroPharma, Inc. (Seattle, WA). The results in this work, along with others, prompted the founding of a startup company, Impel NeuroPharma, to create nasal drug delivery devices to better penetrate the olfactory region of the nasal cavity. In full disclosure of conflict of interest, both authors are inventors on a patent for the POD nasal spray device and both authors also own stock in a startup company, Impel NeuroPharma, founded to develop olfactory-targeted nasal drug delivery devices.
Attestation: Rodney J. Y. Ho has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.
The authors thank Linda Risler for her help with LC/MS analysis.