Abstract
Morphine blood-brain barrier (BBB) transport is governed by passive diffusion, active efflux and saturable active influx. These processes may be associated with nonlinear concentration-dependencies which impact plasma and brain extracellular fluid (brainECF) pharmacokinetics of morphine. In this study, we aim to evaluate the impact of nonlinear BBB transport on brainECF pharmacokinetics of morphine and its metabolites for different dosing strategies using a physiologically based pharmacokinetic simulation study. We extended the human physiologically based pharmacokinetic, LeiCNS-PK3.0, model with equations for nonlinear BBB transport of morphine. Simulations for brainECF pharmacokinetics were performed for various dosing strategies: intravenous (IV), oral immediate (IR) and extended release (ER) with dose range of 0.25-150mg and dosing frequencies of 1-6 times daily. The impact of nonlinear BBB transport on morphine CNS pharmacokinetics was evaluated by quantifying (i) the relative brainECF to plasma exposure (AUCu,brainECF/AUCu,Plasma) and (ii) the impact on the peak-to-trough ratio (PTR) of concentration-time profiles in brainECF and plasma. We found that the relative morphine exposure and PTRs are dose dependent for the evaluated dose range. The highest relative morphine exposure value of 1.4 was found for once daily 0.25mg ER and lowest of 0.1 for 6-daily 150mg IV dosing. At lower doses the PTRs were smaller and increased with increasing dose and stabilized at higher doses independent of dosing frequency. Relative peak concentrations of morphine in relation to its metabolites changed with increasing dose. We conclude that nonlinearity of morphine BBB transport affect the relative brainECF exposure and the fluctuation of morphine and its metabolites.
Highlights
Nonlinear transport affects relative morphine exposure in brainECF.
Nonlinear transport affects PK fluctuations of morphine in brainECF.
Nonlinear transport affects brainECF PK relationship of morphine and its metabolites.
1. Introduction
Morphine is an opioid with an important place for the treatment of acute and chronic pain. The main metabolites of morphine in humans are morphine-3-glucuronide (M3G) and morphine-6-glucuronide (M6G) (Christrup, 1997; De Gregori et al., 2012; Frölich et al., 2011). M3G displays a relatively low affinity for opioid receptors and has no analgesic activity. In fact, an opposite effect, hyperalgesia, has been reported (Frölich et al., 2011; Gabel et al., 2022). M6G, however, is capable of eliciting profound analgesic activity, and has even been propose to as the main drive of the analgesic effects of morphine treatment (Klimas and Mikus, 2014; Murthy et al., 2002).
Many pharmacological studies on morphine and its metabolites effect have been performed, but these studies have typically only considered its plasma pharmacokinetics and not the target site pharmacokinetics. However, morphine and its metabolites first need to cross the blood–brain barrier (BBB) to reach the brain extracellular fluid (brainECF) where they can bind with opioid receptors in the brain. Thus, brainECF concentrations, and not plasma concentrations, should therefore be considered the target site concentration driving the effect. The rate and extent of BBB transport of morphine, M3G, and M6G are different, as has been shown by microdialysis studies in rats. Beside passive transports, para- and transcellular, morphine, M3G and M6G are actively transporter. For morphine, both the P-glycoprotein (P-gp) (Chaves et al., 2017; Letrent et al., 1999; Xie et al., 1999) and probenecid-sensitive transporters (Tunblad et al., 2003) act as BBB efflux transporters, while morphine has a saturable active influx by a yet unidentified BBB influx transporter (Groenendaal et al., 2007; Xie et al., 1999). In rats, it has been shown that blocking P-gp increases the plasma and spinal cord M6G concentrations (Lötsch et al., 2002) while in humans no p-gp related changes in plasma pharmacokinetics were observed (Skarke et al., 2004). The same study in humans showed that probenecid treatment decreases M6G plasma clearance, suggesting that M6G is a substrate for the probenecid-sensitive efflux transporter in the human body (Skarke et al., 2004). It has been reported that GLUT-1 and a digoxin-sensitive transporter can actively efflux M6G but with a weak capacity (Bourasset et al., 2003). For M3G, no P-gp interaction at the level of the BBB has been reported (Xie et al., 1999), while there is a possible involvement of a probenecid-sensitive efflux transporter (Xie et al., 2000). Earlier studies shows the different transport mechanism involved in BBB transport of morphine and its metabolites that can influence their CNS exposure.
When considering BBB transport, constant concentrations at equilibrium (steady-state conditions) and linear pharmacokinetic relationships are often assumed. The ratio from a particular (unbound) concentration in brain and in plasma is used (i.e., a fixed Kp,uu,BBB value), without considering (plasma) concentration-dependency (Wright et al., 2011). The concentration-dependency should be considered for drugs with potentially nonlinear active BBB transport processes, and drugs with metabolites that compete in binding to the same receptor(s). Since morphine and its metabolites are known to be affected by nonlinear BBB transport processes, dosing schedules and/or formulations may impact the ultimately observed exposure at the target site.
The aim of this study is to evaluate the impact of nonlinear BBB transport on relative CNS exposure of morphine and its active metabolites for broad range of dosing regimens and formulations. To that end, we will apply a physiologically based pharmacokinetic (PBPK) CNS modeling approach. This PBPK CNS model, the LeiCNS-PK3.0, predicts within two-fold error the unbound drug concentrations at different CNS compartments (Saleh et al., 2021). We expand the LeiCNS-PK3.0 PBPK model with concentration-dependent BBB transport processes of morphine. The area under the curves (AUC) and the peak-to-through ratio (PTR) for unbound plasma and unbound brainECF pharmacokinetic profiles of morphine and its metabolites were compared to assess the effect of nonlinear BBB transport.
2. Methods
2.1 Nonlinear transport blood brain barrier
The nonlinear BBB transport of unbound morphine in the LeiCNS-PK3.0 model is described by a concentration-dependent Kp,uu,BBB function. To derive this function, a previously published pharmacokinetic model was used that described nonlinear BBB transport of morphine in rats, which included passive diffusion, active efflux and saturable influx transport (Groenendaal et al., 2007). To obtain an equation for Kp,uu,BBB, this nonlinear model was simulated for rat for a wide range of doses between 0.1 – 500 mg/kg as a continuous infusion for 24 hours to obtain steady-state profiles. We then fitted a power function to relate the plasma unbound steady state concentrations to the Kp,uu,BBB, resulting in the following power function Kp,uu,BBB = 5.4902*Css,u,plasma-0.552.
2.2 LeiCNS-PK3.0 PBPK model
For this study the previously published CNS PBPK model, LeiCNS-PK3.0 (figure 1B), was used as base model (Saleh et al., 2021). Briefly, this comprehensive model consists of a plasma and multiple CNS and cerebrospinal fluid (CSF) compartments. Between the brain microvasculature and brainECF and CSF compartments the BBB and the blood-CSF-barrier (BSCFB) are incorporated. The multiple physiological compartments are connected through cerebral blood, brainECF and CSF flows. Furthermore, this model takes into account pH values in each compartment, as well as brain non-specific tissue binding. As input into the LeiCNS-PK3.0 model-, on one hand, previously published human population plasma pharmacokinetic model for morphine and metabolites following IV and oral dosing was used (figure 1A; table 1) (Oosten et al., 2017). The physicochemical properties of morphine and its metabolites were provided to the model (table 1).
To describe physiological processes such as active BBB transport, the model includes asymmetry factors (AF). This value can be seen as the “pure” extent of drug distribution at the barrier, without influences of other elimination routes such as brainECF bulk flow, which in our model are explicitly separated. The AF are calculated as influx and efflux ratios at steady state and includes Kp,uu,BBB values. If a Kp,uu,BBB is 1, mainly passive transport is dominating, and the AFinflux and AFefflux will be equal. If Kp,uu,BBB is lower than 1, AFefflux will be calculated and AFinflux set to 1, and the other way around when Kp,uu,BBB is higher than 1.
In order to do simulations for humans, a human Kp,uu,BBB value is needed as input. The Kp,uu,BBB describing morphine nonlinear transport at BBB for human has not been determined. Therefore, the calculated rat Kp,uu,BBB power function was used with a translational factor based on a transporter protein expression ratio at human versus rat BBB (abbreviated as fAFBBB) to correct the AF. This factor is only applied when a drug or metabolite is a substrate of a transporter. For morphine, two efflux (P-gp and probenecid-sensitive) and one influx transporter was taken into account. The mean protein expression level of P-gp in humans is 4.21 fmol/μg total protein (Al-Majdoub et al., 2019; Shawahna et al., 2011; Uchida et al., 2011) and in rats this is 19.28 (Al Feteisi et al., 2018; Hoshi et al., 2013), resulting in a ratio of 0.22. For other transporters, such as the probenecid sensitive transporter and the saturable influx transporter, no expression information is available. In this case it was assumed that expression in humans and rats is equal. Since only P-gp is identified an fAFBBB of 0.22 was used for the translation of rat value of AFBBB to that of human. For M3G and M6G no information on the exact transporters is available and for this an fAFBBB of 1 is used.
2.3 Simulation scenarios
Morphine and metabolite brainECF distribution simulations for human were performed for a period of seven days in order to reach a steady state exposure. A dose range of 0.25-150 mg for intravenous (IV), oral immediate release (IR) and extended release (ER) formulations. All the doses are administered once, twice, four and six times a day.
2.4 Evaluation of simulation scenarios
To compare the relative morphine exposure, the AUC ratio of brainECF over plasma AUC at steady-state was used (equation I). The results are also compared for the advised clinical doses for the different formulations.
The pharmacokinetic profile fluctuations were evaluated at day seven by comparing the peak-to-trough (PTR) calculated as in equation (II) (Tozer and Rowland, 2016). PTR is calculated by the highest concentration Cmax minus lowest concentration Cmin divided by the average concentration Cav.
2.5 Sensitivity analysis
A sensitivity analysis was performed for the fAFBBB parameter to evaluate the effect of variations of this parameter on the brainECF exposure (AUCECF). Perturbations of 0.25-2 fold changes in steps of one-quarter of fAFBBB parameter was simulated.
2.6 Software
Simulations for nonlinear BBB transport and LeiCNS-PK3.0 models were performed using the package RxODE version 1.1.5 and for sensitivity analysis the additional PKNCA package version 0.9.5 using R version 4.1.3.
3. Results
3.1 Relative morphine exposure
To compare the effect of nonlinear BBB transport processes, the relative morphine exposure in the brainECF to plasma was compared for the different formulations at steady state (day seven after treatment start). For all the administration routes, low morphine doses administrations at low frequency resulted in a relative higher exposure of unbound morphine in the brainECF than in plasma, while increasing dose and frequency led to an increased exposure in plasma compared to brainECF (figure 2). For almost all administrations, the relative morphine exposure was 1 or lower expect for ER and IR administration of 0.25mg once-a-day. For the metabolites, no differences in relative metabolite exposure have been observed (results not shown). These results show that at low doses (<0.5mg) and low frequency (<twice a day) administrations relative morphine exposure is higher in brainECF than plasma.
3.2 Morphine peak-to-trough ratios
To investigate the effect of nonlinear transport on the fluctuation in pharmacokinetic profiles, the PTR concentration ratios for the three dosing regimens were compared. The PTR versus dose in plasma was stable over the simulated dose range, while increasing the frequency, the PTR decreased as expected (figure 3). For brainECF there was no stable PTR versus dose range observed. The PTR increased with increasing dose and stabilized at higher doses for IV (figure 3), IR and ER (supplementary figure S1 and S2 respectively). The brainECF PTR over the dosages shows the impact of the saturable influx transporter at lower doses resulting in a nonstable PTR over the dosages. This indicates that steady-state plasma PK profile is not representative for the brainECF PK profile.
3.3 Nonlinearity effect on metabolite distribution
To study the effect on nonlinear BBB transport of morphine and its metabolites, the profiles are compared for different IV administrations at steady-state. Comparing the peak concentrations (Cmax) of morphine to that of metabolites showed changes with increasing dose. At low dose of 1 mg morphine Cmax was higher compared to M3G Cmax and with increasing dose, M3G Cmax became higher and the difference in peak concentrations of morphine and M3G increased further (figure 4). For morphine versus M6G, this difference in relative Cmax was other way around, the difference in Cmax decreased with increasing dose (figure 4). When the metabolite to morphine exposure ratio at steady state was compared (AUCbrainECF,metabolite / AUCbrainECF,morphine) an increase in this exposure ratio with increasing dose was observed. For M3G/morphine exposure, the ratio at lower doses were above 1 and with increasing dose this ratio increased. Same effect was also observed for M6G/morphine exposure, but at lower dose this ratio was lower than 1 and at higher doses it increased above 1. From these results we can conclude that due to nonlinear BBB transport of morphine, the relation between morphine and metabolites brainECF peak concentrations and brainECF exposure ratios changed in relation to dose changes.
3.4 Sensitivity analysis
To simulate human morphine brainECF distribution, rat to human AFBBB factors were translated using the fAFBBB parameter. Morphine is transported by P-gp and one unidentified efflux and one unidentified influx transporter. To calculate fAFBBB to translate rat to human Kp,uu,BBB values, the unidentified transporters are assumed to be equally expressed at rat and human BBB. Sensitivity analysis was performed by varying the fAFBBB value to evaluate the possible changes in the brainECF AUC in case the human BBB transporter expression would deviate from rat values. We find that the highest impact of a change in fAFBBB would be at lower morphine doses, where the contribution of influx transport is the largest (figure 5). For IV administrations of once-a-day, up to 20 mg, an increase in fAFBBB would result in an increase brainECF AUC. The opposite effect was observed for doses higher than 20mg once a day, where an increase in fAFBBB leading to a decrease in brainECF AUC. For IR, the possible effect of fAFBBB changes on brainECF AUC was similar, only the shift in effects occurred at a higher dose of 70mg once a day. For ER, this shift was even at higher dose of 110 mg once a day. The sensitivity analysis showed that the possible largest effect of fAFBBB change would be for a dose of 1 mg within the simulated range of 1 to 150 mg. For ER administrations, a decrease of 75% of fAFBBB would lead to a decrease of 65% brainECF AUC and an increase of 200% would result in an increase of 74% brainECF AUC.
4. Discussion
In this study we evaluated the impact of nonlinear BBB transport on distribution of morphine and its active metabolites in the brainECF, by expanding the LeiCNS-PK3.0 PBPK model with nonlinear BBB transport processes. We showed that nonlinear BBB transport of morphine affects the relative target site exposure and PTR, as well as the relation of morphine to its metabolite brainECF exposure.
Our model predicts the importance of including nonlinear BBB transport to evaluate human brainECF concentration of morphine and its metabolites. For ethical reasons no such direct information can be obtained from human. In this study, nonlinear BBB transport was implemented for predicting morphine brainECF pharmacokinetics, based on previous in vivo mice and rat studies that provided quantitative information on plasma concentration dependent BBB influx transport and non-saturable BBB efflux transport processes (Groenendaal et al., 2007; Xie et al., 1999). The extended LeiCNS-PK3.0 model (Saleh et al., 2021; Yamamoto et al., 2017a, 2017b) needs as input a plasma concentration-dependent human Kp,uu,BBB (nonlinear BBB transport) of morphine, but such data are not available. So, rat values for concentration-dependent Kp,uu,BBB values were derived (Groenendaal et al., 2007), and used in combination with rat to human transporter expression translational factor, fAFBBB, to correct the AFBBB. Here the fAFBBB is the relative expression factor of transporters on the BBB in rat and human (Al-Majdoub et al., 2019; Shawahna et al., 2011; Uchida et al., 2011; Yamamoto et al., 2018). Transporters at BBB play a crucial role in drug exposure at brainECF. For this reason, using the relative expression factor as translational factor from rat to human is useful (Yamamoto et al., 2018).
We assumed both unidentified transporters for morphine BBB influx and efflux are equally expressed in rat and human, while the expression of P-gp was scaled from rat to human based on available relative expression values (Al-Majdoub et al., 2019; Shawahna et al., 2011; Uchida et al., 2011; Yamamoto et al., 2018). The sensitivity analysis has shown the possible impact of changes in the fAFBBB on the brainECF exposure. The results indicate the importance of identification of these transporters, mainly at lower doses. The morphine brainECF exposure might be mainly at lower doses higher or lower with higher or lower fAFBBB, respectively. For morphine transport across the BBB, P-gp is the only identified active transporter. Another efflux transporter is probenecid dependent as best current knowledge, and furthermore, there is an unidentified saturable influx transporter. Probenecid is known to be an inhibitor for many transporters including multidrug resistance associated proteins (MRPs). The organic anion transporter 1 and 3 (oat1, oat3) and organic anion transporting polypeptide 1 and 2 (oatp1, oatp2) are also inhibited by probenecid (Sugiyama et al., 2001). The possible influx transporter could be the organic cationic transporter 1 (OCT1). Previous study have shown that OCT1 plays a role in hepatocellular saturable and concentration-dependent uptake of morphine (Tzvetkov et al., 2013) and for some cationic compounds in rat and human transfected hepatocytes (Umehara et al., 2007). Whether these suggested transporters are involved in morphine transport their presence at human BBB and transport of morphine should be confirmed.
This study has shown that nonlinear BBB transport mainly affects morphine brainECF pharmacokinetics at lower dose and lower dosing frequencies for IV, oral IR and ER formulations. With increasing dose, the influx BBB transport of morphine becomes saturated, and its BBB transport becomes mostly linear with plasma concentrations. This nonlinear BBB transport effect is outside the clinical dosing regimens for adult (FDA, 2012, 1984), suggesting no direct impact of nonlinear BBB transport on morphine treatment to adults. For pediatrics, nonlinear BBB transport might have more impact on the treatment regimens. Clinical dosing regimens of morphine in pediatrics depends on their weights resulting for example in oral regimens for pediatrics younger than 12 years a maximum of 200-500 mcg/kg every 4 hours with a maximum of 5mg per day (Unknown author, 2012). With this, the total dose administered of morphine in pediatrics might be within the nonlinear BBB transport dosing range whereby relative more morphine exposure is at the brainECF. Therefore, possible effects due to higher morphine exposure at the brainECF could be taking into account when administered to pediatrics.
To our best knowledge, nonlinear BBB transport is applicable for morphine, but not for M3G and M6G. The effect of nonlinear BBB transport on the relation of morphine with its metabolites at brainECF has not been studied before. We found that increasing plasma concentrations result in different brainECF concentrations ratios of M3G/morphine and M6G/morphine. This may have an impact on their relative receptor binding. The target receptors of morphine and its metabolites are the mu1, mu2, delta and kappa opioid receptors (Imming et al., 2007; Kristensen, 1995). These receptors are predominantly present in the CNS (Peng et al., 2012). For morphine and M6G to exerts their analgesic effect, they should bind to the mu-opioid receptors (Rainville, 2002; Vanderah, 2010; Yamada et al., 2006) and therefore compete with each other. M3G, on the other hand, has a low potency for the mu-opioid receptor (Frölich et al., 2011). In humans, it has been debated that M3G can cause hyperalgesia by binding to Toll-like receptor 4 and may lead to cross-talk of the Toll-like receptor 4 and mu-opioid receptor (Gabel et al., 2022). By binding to the mu-opioid and Toll-like receptors, morphine and M6G actives the Gi-protein and β-arrestin while M3G has a lower potency and activates scaffold proteins (Frölich et al., 2011; Gabel et al., 2022). Altogether, this indicates the need for understanding brainECF exposure as step one, followed by further exploration on the consequences on receptor binding.
One aspect not yet taken into account in this study is the possible metabolism of morphine within the CNS. This study assumed only metabolism of morphine in the liver and used plasma pharmacokinetics with fixed metabolite fractions (Oosten et al., 2017). However, a previous study has shown possible M6G formation from morphine in human brain homogenates (Yamada et al., 2003), and also in rat microglia (Togna et al., 2013). This is to be further investigated in future research.
5. Conclusion
In conclusion, our simulations indicate that nonlinear BBB transport of morphine and its metabolites may affect exposure in brainECF target site concentrations, in particular at lower doses, enabled by an in silico PBPK modeling approach using the LeiCNS-PK3.0 model.
6. Conflict of interest
NA
7. Funding
This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 848068. This manuscript reflects only the authors’ view and the European Commission is not responsible for any use that may be made of the information it contains.
8. Authors contribution
Berfin Gülave: Conceptualization, Data collection, preparation and analysis, Model simulations, Writing original draft, review and editing. Divakar Budda: Conceptualization, Data collection, preparation and analysis, Model simulations, Writing original draft, review and editing. Mohammed AA Saleh: Data preparation and analysis. JG Coen van Hasselt: Conceptualization, Writing original draft, review and editing. Elizabeth CM de Lange: Conceptualization, Writing original draft, review and editing.