![[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}
|
|
|
|
|
||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Division of Pharmaceutics and Biopharmaceutics, Faculty of Pharmaceutical Sciences (S.C.S.A.-Y., K.W.R.), and Child and Family Research Institute, Department of Obstetrics and Gynecology, Faculty of Medicine (D.W.R., N.G.), University of British Columbia, Vancouver, British Columbia, Canada
(Received October 18, 2005; accepted February 24, 2006)
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
; Farrell and Risau, 1994; Davson and Segal, 1996) by controlling the exchange of compounds between two barrier structures—one located between the blood and brain extracellular fluid (ECF), termed the blood/brain barrier, and the second between the blood and cerebrospinal fluid (CSF), known as blood/CSF barrier. Diphenhydramine [2-(diphenylmethoxy)-N,N-dimethylethylamine (DPHM)] is a potent histamine H1-receptor antagonist (Douglas, 1980) widely used for its antiallergic properties, as well as for its antiemetic, sedative, local anesthetic, and hypnotic effects (Runge et al., 1992; Ernst et al., 1993; Pontasch et al., 1993). Like other "first-generation" antihistamines, DPHM occupies central H1-receptors to result in drowsiness, sedation, incoordination and with higher doses, convulsions, and death (Douglas, 1980; Nicholson, 1983; Koppel et al., 1987; Gengo et al., 1989). However, there are limited data on the CNS levels of the drug or on the mechanisms of transfer involved. Results from previous studies in rats, guinea pigs, and rabbits suggest that DPHM enters the brain tissue and CSF extremely rapidly to achieve CNS concentrations exceeding those in plasma (Glazko and Dill, 1949a,b; Takasato et al., 1984; Goldberg et al., 1987). Because only 
2.5% of DPHM (pKa 9.0) is un-ionized at the physiological pH, the above results cannot be explained by the passive diffusion of this un-ionized form through the blood-brain and blood-CSF barriers. Moreover, there is evidence for saturable BBB transporter mechanisms for lipophilic, amine drugs (Pardridge et al., 1973, 1984; Spector, 1988; Yamazaki et al., 1994a,b,c). This includes mepyramine, a histamine H1-antagonist. The available data suggest that these compounds cross the blood-brain and blood-CSF barriers by both simple diffusion of the un-ionized lipid-soluble form and by carrier-mediated transport of the ionized form (Pardridge et al., 1984; Goldberg et al., 1987; Yamazaki et al., 1994a). In addition, there is evidence that various substances can inhibit the actions of this transport process. In vivo (rat carotid injection technique), brain uptake of mepyramine is inhibited by DPHM (Yamazaki et al., 1994a). Furthermore, both in vivo and in vitro (bovine brain capillary endothelial cells) studies demonstrate that propranolol (PRN) inhibits mepyramine uptake (Yamazaki et al., 1994; Yamazaki et al., 1994). Together, these data led us to hypothesize that PRN could inhibit brain (and perhaps CSF) DPHM uptake. The purpose of our studies then was to use in vivo microdialysis (MD) in chronically instrumented adult ewes to investigate the transport processes of DPHM across the adult ovine blood-brain-barrier using two different experiments. The first involved stepped infusions of DPHM at five different dosing rates to assess blood-brain CSF and blood-brain ECF drug concentration relationships in relation to variations in drug dose and hence plasma drug levels. The second involved coadministration of DPHM and PRN to examine whether PRN alters blood-brain CSF and blood-brain ECF DPHM relationships. |
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Experimental Protocols. This was a single-site, nonrandomized, open-label, two-period, single sequence study with the stepped infusion experiment followed by the DPHM-PRN coadministration study. There was a washout period of at least 2 days between the study sessions. The details of the two study periods are described below.
DPHM step-infusions. The protocol involved bolus i.v. loading doses of DPHM (to hasten the achievement of steady state), followed by i.v. infusion of the drug using an infusion pump (model 600-000; Harvard Apparatus Inc., Dover, MA) at five different rates, with each infusion rate lasting 7 h. The DPHM loading dose was 0.15 mg/kg, and the infusion rates were 1.5, 5.5, 9.5, 13.5, and 17.5 µg/kg/min. During the infusions, arterial blood samples (3 ml) were collected hourly. MD sampling began at the onset of the infusion. The microdialysis pump (model PHD2000; Harvard Apparatus Inc., Holliston, MA) infusion rate was 2 µl/min, and 60-min cumulative samples of CSF and ECF were collected throughout the duration of the experiment. Both blood and MD samples were collected up to 18 h after the end of the last infusion step.
Coadministration of DPHM and PRN. DPHM was infused at rate 4 (i.e., 13.5 µg/kg/min) as described in DPHM step-infusions above for 8 h. After 4 h, PRN was coinfused (1.5 mg/kg loading dose, 20.0 µg/kg/min) (Jones and Ritchie, 1978; Mihaly et al., 1982; Czuba et al., 1988) for the remaining 4 h of the DPHM infusion. Blood and MD samples were collected as described above during the infusion and for up to 12 h after infusion.
Retrodialysis. MD probe recovery was determined using the retrodialysis technique (De Lange et al., 1998). The MD dialysate (degassed, sterile lactated ringer solution) contained a calibrator ([2H10]-DPHM) at a concentration of 400.0 ng/ml. The probe recovery rate was determined by comparing the input and output concentrations of the calibrator as follows:
 | (1) |
Free-fraction drug concentration (CCSF or CESF) at the MD sampling site was equal to the diphenhydramine concentration in the output dialysate ([DPHM]dialysate)/recovery rate.
Physiological Recording. Arterial pressure was measured using strain-gauge manometers (Ohmeda Inc., Madison, WI) and heart rates from a cardiotachometer (Astro-Med, West Warwick, RI). All variables were recorded on a Grass K2G polygraph (Astro-Med) and on a computerized data acquisition system (chart v4.2; ADInstruments, Grand Junction, CO).
Plasma Protein Binding of DPHM. Determination of plasma protein binding/unbound fraction (CPu) of DPHM was achieved using the equilibrium dialysis procedure described by Yoo et al. (1990) in the 7-h steady-state plasma sample from each infusion step of the five-step infusion studies. In the case of the DPHM-PRN coadministration study, a sample collected at 8 h of the infusion was used for measurement.
Drug Analysis. The concentrations of DPHM (CPt, CPu, CCSF, and CECF) in all samples were measured using a gas chromatographic-mass spectrometric assay capable of simultaneously measuring DPHM and [2H10]-DPHM with a limit of quantitation of 2.0 ng/ml (Tonn et al., 1993).
Statistical and Data Analysis. All pharmacokinetic modeling was performed using WinNonlin, version 1.1 (Scientific Consulting Inc., Apex, NC).
Five-step infusion study. Volume of distribution (Vd) and total body clearance (ClT) were calculated using the following respective equations (Gibaldi and Perrier, 1982):
 | (2) |
 | (3) |
Data were plotted using Microsoft Excel 2000 (Microsoft Corporation, Mountain View, CA). All data are reported as mean ± S.D. Statistical analyses were performed using JMP IN, version 3.2.1 (SAS, Cary, NC). The significance level was p < 0.05 in all cases.
Calculation of fCSF and fECF. The extent of DPHM transfer into the brain in this study was calculated by relating the CSF and ECF total area under the plasma concentration versus time curve (AUC0
) values to the plasma AUC0 value to yield the fCSF and fECF ratios. Specifically, using fCSF as an example:
 | (4) |
). This method of characterizing drug transfer across the BBB has been used in numerous other MD studies for many different drugs, including acetaminophen, atenolol, gabapentin, zidovudine, morphine-6-glucuronide, lamotrigine, phenobarbital, and felbamate (Wang et al., 1993; Wong et al., 1993; de Lange et al., 1994; Luer et al., 1999; Bouw et al., 2001; Potschka et al., 2002).
DPHM-PRN coadministration. Vd, ClT, fCSF, and fECF values in these experiments were estimated using the equations listed above. However, for comparisons of the DPHM alone and DPHM-PRN coadministration periods, DPHM concentrations and AUC values for the periods 0 to 4 h and 4 to 8 h, respectively, were used. Thus, to assess the effect of PRN on the fCSF value, for example, area under the CSF concentration versus time curve from 0 to 4 h (AUCCSF 04)/area under the plasma concentration versus time curve from 0 to 4 h (AUC04) was compared with area under the CSF concentration versus time curve from 4 to 8 h (AUCCSF 48)/area under the plasma concentration versus time curve from 4 to 8 h (AUC48).
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
|
DPHM-PRN Coadministration Study. Figure 2 depicts the concentration-time relationships for the plasma, CSF, and ECF compartments from the DPHM-PRN coadministration study. Consistent with results from the five-step infusion study, elimination in all three fluids followed two-compartment pharmacokinetics (Table 6).
Tables 4 and 5 compare the DPHM concentrations in plasma, CSF, and ECF and CCSF and CECF to CPu ratios before and after propranolol coadministration. After PRN administration, CPt tended to be lower, although this did not achieve statistical significance. However, the percentage decrease (12.8 ± 6.3%) was significantly different from 0. In contrast, CECF and CCSF concentrations increased during PRN administration, and for CCSF, the increase was statistically significant (Table 4). For both CECF and CCSF, the percentage increases (88.1 ± 45.4% and 91.6 ± 34.3%, respectively) were statistically significant. Similar to the five-step infusion study, the CNS concentrations were higher than the CPu. As shown in Table 5, the CCSF/CPu and CECF/CPu ratios tended to increase during PRN infusion, but because of interanimal variability, the changes were not significant. However, the percentage increases in the ratios were significantly different. The protein binding value with propranolol coadministration was 84.4 ± 10.5%, which was not significantly different from the value obtained in the five-step study (86.1 ± 2.3%).
|
|
Table 6 provides a summary comparing the pharmacokinetic parameters before and after PRN administration. ClT tended to be higher with PRN coadministration, whereas Vd was lower, but these changes were not statistically significant. The elimination half-life in plasma tended to be lower during PRN and the half-life values in CSF and ECF higher, but none of these changes were statistically significant. The f ratios for both CSF and ECF were significantly increased.
Physiological Responses. In the five-step infusion study, during the administration of infusion steps 4 to 5, three animals showed symptoms of agitation including restlessness, tremor, excessive bleating, and heavy breathing. These symptoms disappeared 1 to 1.5 h after the end of the infusion. During steps 4 and 5, which involved the highest infusion rates, the mean arterial pressure (110 ± 2.41 mm Hg for step 4 and 110 ± 0.98 mm Hg for step 5) was not significantly different from the mean baseline value of 111 ± 0.41 mm Hg. In terms of mean heart rate, no significant changes from baseline were observed during all five infusion steps.
In the DPHM-PRN study, no significant difference was observed in mean arterial pressure after PRN coinfusion (101 ± 1.47 mm Hg) compared with DPHM administration alone (102 ± 0.86 mm Hg). However, a significant drop in mean heart rate was observed with PRN coadministration (77.7 ± 3.47 beats per minute) compared with DPHM alone (95.5 ± 2.37 beats per minute), and this fall in heart rate persisted for the full 20-h duration of the experiment (Fig. 3).
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
). The tight junctions primarily restrict the entry of proteins and other large hydrophilic molecules into the CNS (Davson and Segal, 1996; Habgood et al., 2000). For lipophilic compounds not significantly bound to plasma proteins, there is a good correlation between the BBB permeability coefficient and the octanol/water partition coefficient, provided the molecular mass is <400 to 600 Da (Levin, 1980). However, numerous transporters are present in brain endothelial cells, which transfer substances into the CNS at rates higher than could occur via simple diffusion. There is also evidence for saturable transporter mechanisms in the BBB for a number of lipophilic amine drugs, including PRN and the histamine H1-antagonist mepyramine (Pardridge et al., 1984; Yamazaki et al., 1994b). There are two possible paths for a substance from the CNS to return to the systemic circulation. One is efflux (transporter-mediated or not) via brain or choroidal blood. The second route involves efflux via bulk flow of CSF draining into either the lymphatic system or venous blood (Bradbury et al., 1972). The latter phenomenon is termed the sink effect, whereby the brain concentrations of different compounds under steady-state conditions are different from each other and lower than the unbound concentration in blood (Davson and Segal, 1996), as a consequence of the continuous removal of the substances via the CSF.
In this study, we applied the microdialysis technique to investigate the blood-brain CSF and blood-brain ECF DPHM relationships in two different experiments. In the first experiment, the results from the five-step infusion showed that brain concentrations increased correspondingly to increases in dose, suggesting that the transfer of DPHM into the CNS was a concentration-dependent process. However, considering the high degree of plasma protein binding (86.1 ± 2.3%), the CNS DPHM concentrations were actually higher than the free DPHM concentration in plasma (Tables 2, 3, 4, 5). This suggested that the entry of DPHM into the CNS was most likely due to an active transport process, because if passive diffusion was the only driving force, then free plasma DPHM concentrations should be comparable with the CNS levels. In fact, our findings indicated that DPHM concentrations were at least 2 times higher than free plasma concentrations (Table 3). As mentioned earlier, there are data suggesting that lipophilic, amine compounds such as DPHM can cross the blood-brain and blood-CSF barriers by both simple diffusion of the un-ionized lipid-soluble form and by carrier-mediated transport of the ionized form (Pardridge et al., 1984; Goldberg et al., 1987; Yamazaki et al., 1994a).
The postulation of an active transport process is supported by the CCSF/CPu and CECF/CPu ratios, both ranging from 2 to 3 (Table 3). Transfer of DPHM into the CNS was rapid after administration; this was not a surprising observation considering the highly lipophilic nature of this compound [octanol/water partition coefficient 1862 (Douglas, 1980)]. Other reports also suggest rapid distribution of DPHM into tissues, with the maximum tissue uptake occurring at 1 to 3 min after i.v. injection (Drach et al., 1970). The close similarities between the two CNS drug concentrations (Table 2 and Fig. 1) suggest that drug clearances in the choroid plexus and the cerebral cortex are comparable with each other. This observation can further be assessed by comparing the fCSF and fECF values (0.4 ± 0.2 and 0.4 ± 0.2, respectively), which were not significantly different from each other. Two-compartment pharmacokinetics were observed in the DPHM elimination profiles of all three fluids in the five-step infusion studies, with the CNS compartments declining at the highest rates (Table 6). This was reflected in the half-life (t1/2ßplasma, 10.8 ± 5.4 versus t1/2ßCSF, 3.6 ± 1.0; t1/2ßECF, 5.3 ± 4.2 h) values. The more rapid elimination of DPHM from the brain compared plasma is associated with higher concentrations of the drug in the CNS compared with the unbound plasma concentration (Table 2). This is probably due to active transport of the ionized DPHM into the brain, as discussed previously, so that more of the total circulating concentration of the drug is available to the brain compared with other organs and tissues.
The rapid efflux of DPHM from CNS could be due to the bulk flow of CSF (sink effect) and also involvement of a transporter-mediated efflux mechanism for the drug. The validity of the above assumption can be assessed by examining the results from the DPHM-PRN experiment. The CSF DPHM concentration increased significantly during PRN coadministration and for both CSF and ECF the percentage increase in drug levels was statistically significant (Table 4). In addition, both the fCSF and fECF increased after the coadministration of propranolol (from 0.40 ± 0.20 and 0.40 ± 0.20 to 0.69 ± 0.03 and 0.95 ± 0.05, respectively), with fECF being significantly increased. The trend for an increase in the CSF and ECF DPHM elimination half-life after PRN is also consistent with reduced clearance of the drug from the brain. However, these changes were not statistically significant. Although the PRN infusion was not continued into the elimination phase, the persistence of the PRN-elicited bradycardia for the entire duration of the experiment (Fig. 3) indicates that there was sufficient PRN still present during the elimination period for this action and thus perhaps also for an effect on DPHM transfer across the BBB. Overall, the findings suggest lower DPHM clearances from the CNS after PRN coadministration.
As mentioned earlier, besides the bulk flow of CSF, efflux of substances (transporter-mediated or not) can occur via brain or choroidal blood back to the systemic circulation. To date, there is no information showing that PRN lowers CSF formation or its secretion rates and that it has any interference with the passive diffusion process of substances back to the cerebral circulation. In addition, besides a slight decrease in heart rate, PRN causes no systemic or cerebral physiologic changes in sheep (O'Brien et al., 1999). In another study, PRN infusion did not significantly change choroid plexus blood flow in sheep (Townsend et al., 1984). Results from these studies are consistent with our current findings in that only heart rate, but not arterial pressure, was affected by PRN. Therefore, the lowered brain clearances are probably due to lowered rates of efflux of the drug. Propranolol can be involved in this process in one or both of the following manners— by directly inhibiting the efflux mechanism and/or by competing with DPHM for the efflux process. Both of these actions could be responsible for the observed lowered rates of CNS clearances; however, the exact mechanism cannot be elucidated by our current experiments.
In contrast to the situation with DPHM concentrations in CSF and ECF, the plasma total DPHM level fell by 12.8% during PRN coadministration. This may have been due to an increase in systemic clearance of the drug, given the trend for an increase in this variable during PRN and for a decrease in t1/2ß (Table 6). DPHM has a high hepatic extraction in sheep, and thus its hepatic clearance is largely dependent on hepatic blood flow (Kumar et al., 1999). However, PRN has been reported to decrease hepatic blood flow in humans (Zoller et al., 1993; Orszulak-Michalak, 1995); thus, the mechanisms involved in the decreased plasma DPHM concentration are unclear.
In summary, using in vivo microdialysis in chronically instrumented adult ewes, we have demonstrated that DPHM enters the brain rapidly after administration by passive diffusion and an active transfer process. Drug concentrations are markedly higher in the brain relative to unbound plasma levels and this may, in part, explain the significant CNS effects of the drug. DPHM clearances from brain ECF and CSF were similar and faster than plasma clearance, as indicated by the relatively short half-lives in the brain. The rapid efflux of DPHM from the CNS could be due to the bulk flow of CSF (sink effect) and also a transport-mediated efflux mechanism for the drug. The DPHM-PRN coadministration study suggests that PRN inhibits an efflux rather than influx mechanism. This latter finding was rather unexpected and warrants further investigation.
|
Footnotes |
|---|
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
ABBREVIATIONS: BBB, blood-brain barrier; ECF, extracellular fluid; CSF, cerebrospinal fluid; DPHM, diphenhydramine; CNS, central nervous system; PRN, propranolol; MD, microdialysis; AUC, area under the curve; CPt, total plasma DPHM concentration; CPu, free plasma DPHM concentration.
1 Current affiliation: PK/PD, Clinical Pharmacology, Quintiles, Inc., Kansas City, MO. 
Address correspondence to: Dr. K. Wayne Riggs, Faculty of Pharmaceutical Sciences, 2146 East Mall, University of British Columbia, Vancouver, BC, V6T 1Z3, Canada. E-mail: riggskw{at}interchange.ubc.ca
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. C.S. Au-Yeung, K. W. Riggs, N. Gruber, and D. W. Rurak The Use of Microdialysis for the Study of Drug Kinetics: Central Nervous System Pharmacokinetics of Diphenhydramine in Fetal, Newborn, and Adult Sheep Drug Metab. Dispos., August 1, 2007; 35(8): 1285 - 1291. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
|
 |
 |
 |
