Systemic Concentrations Can Limit the Oral Absorption of Poorly

Apr 23, 2013 - with inlets that were modified in-house as described.8 The spray ...... Hammock, B. D.; Busse, R.; Fleming, I. Soluble epoxide hydrolas...
0 downloads 0 Views 999KB Size
Download PDF
Article pubs.acs.org/molecularpharmaceutics

Systemic Concentrations Can Limit the Oral Absorption of Poorly Soluble Drugs: An Investigation of Non-Sink Permeation Using Physiologically Based Pharmacokinetic Modeling Po-Chang Chiang,*,† Hank La,§ Haiming Zhang,‡ and Harvey Wong*,§ †

Department of Pharmaceutics, ‡Process Chemistry, and §Drug Metabolism and Pharmacokinetics, Genentech, South San Francisco, California 94080, United States S Supporting Information *

ABSTRACT: In the early drug discovery environment, poorly soluble compounds with suboptimal potency are often used in efficacy studies to demonstrate in vivo preclinical proof-of-concept for new drug discovery targets and in preclinical toxicity studies to assess chemical scaffold safety. These compounds present a challenge to formulation scientists who are tasked with improving their oral bioavailability because high systemic concentrations are required. Despite the use of enabling formulations, increases in systemic exposure following oral delivery are often not achieved. We hypothesize that in some cases non-sink intestinal permeation can occur for poorly soluble compounds where their high systemic concentrations can act to inhibit their own oral absorption. Rats were given a 30 mg/kg oral dose of 1,3-dicyclohexyl urea (DCU) alone or concurrently with deuterated DCU (D8-DCU) intravenous infusions at rates of 13, 17, and 22 mg/kg/h. D8-DCU infusions dose dependently inhibited DCU oral absorption up to a maximum of 92%. Physiologically based pharmacokinetic modeling was utilized to understand the complex interaction between high DCU systemic concentrations and its effect on its own oral absorption. We show that high systemic concentrations of DCU act to suppress its own absorption by creating a condition where intestinal permeation occurs under non-sink conditions. More importantly, we identify relevant DCU concentrations that create the concentration gradient driving the intestinal permeation process. A new parameter, the maximum permeation extraction ratio, is proposed and provides a simple means to assess the extent of non-sink permeation. KEYWORDS: non-sink permeation, oral absorption, physiologically based pharmacokinetic modeling, PBPK, nanoparticle

■

INTRODUCTION The number of poorly soluble compounds that are being evaluated in today’s pharmaceutical industry setting continually increases.1 One of the direct consequences of low solubility is poor oral bioavailability. The oral absorption of drugs is complex and is dependent on the sequential processes of dissolution and intestinal membrane permeation (Figure 1A). The rate and extent of oral absorption are a result of the interaction between these two processes. The dissolution rate, described commonly by the Noyes−Whitney equation (eq 1), is influenced by compound characteristics such as solubility in intestinal fluid and more controllable properties such as particle size. The gradient between the solubility of a compound in intestinal fluid and the compound concentration in the intestinal section of interest serves as the driving force for dissolution. Following dissolution in the intestinal lumen, drug in solution exists as free molecules or is solubilized by bile micelles.2 Solubilization in micelles can act to enhance intestinal membrane permeation by assisting the movement across the unstirred water layer adjacent to the intestinal membrane.3 A gradient between the relevant concentrations in the intestinal lumen and in the systemic circulation serves to drive the intestinal permeation process. Ultimately, the amount of drug © 2013 American Chemical Society

absorbed is determined by the rate limiting step of the entire oral absorption process. On the basis of the described oral absorption process, the poor oral absorption of compounds can be commonly categorized as dissolution, solubility, or permeability ratelimited.2 Dissolution rate-limited absorption occurs when the intestinal permeation rate far exceeds the dissolution rate such that dissolved drug is rapidly removed from the intestinal fluid, resulting in dissolution being the rate limiting step to oral absorption. Solubility rate-limited absorption occurs when compound concentrations in intestinal fluid reach the solubility limit. As the solubility limit is reached, additional compound cannot be dissolved from the solid form. Therefore, reductions in particle size are expected to improve oral absorption for dissolution rate-limited but not solubility rate-limited absorption. Finally, permeability rate-limited absorption occurs when Special Issue: Impact of Physical Chemical Drug-Drug Interactions from Drug Discovery to Clinic Received: Revised: Accepted: Published: 3980

February April 15, April 23, April 23,

15, 2013 2013 2013 2013

dx.doi.org/10.1021/mp400088q | Mol. Pharmaceutics 2013, 10, 3980−3988

Molecular Pharmaceutics

Article

Figure 1. The oral absorption process with intestinal permeation under classical sink (A) and non-sink (B) conditions.

preclinical in vivo efficacy models to achieve a better understanding of the biological target of interest. In the early stages of a drug discovery program, compounds often have less than optimal potency, thus necessitating high systemic concentrations. For toxicology studies, the goal is usually to achieve systemic concentrations that are multiples of anticipated therapeutically active concentrations which can be challenging for compounds with very poor solubility. Providing they are tolerated, enabling solution formulations often containing organic solvents are used in attempts to increase oral exposure in both preclinical efficacy and toxicology studies. Despite the use of enabling formulations, increases in systemic exposure following oral delivery are often not achieved. We hypothesize that in some cases non-sink conditions can occur for poorly soluble compounds in these studies whereby the systemic concentrations of compound are high enough that they begin to impact the intestinal permeation process by causing a reduction in the concentration gradient between relevant concentrations in intestinal lumen and systemic circulation (Figure 1B). This condition is enabled by a lower achievable concentration of drug in solution in the intestinal lumen due to poor solubility. Although theoretically possible, to our knowledge, non-sink conditions have never been demonstrated in vivo and are often not considered as a possible reason for poor oral absorption. In vivo preclinical proof-of-concept efficacy studies and toxicology studies mentioned in the preceding paragraph are two situations where such conditions can be encountered in the early drug discovery phase. Therefore, the objective of the current study is to experimentally demonstrate non-sink permeation by investigating the effect of systemic concentrations of a low solubility tool compound, 1,3-dicyclohexyl urea (DCU), on its own oral absorption.7 Intravenous infusions of a DCU (in the form of deuterated DCU; D8-DCU) nanosuspension formulation are utilized in this study in order to achieve high systemic DCU concentrations that approach the solubility of the compound in plasma. Finally, physiologically based pharmacokinetic modeling is utilized to understand the complex interaction between systemic concentrations of DCU

the dissolution rate far exceeds the permeation rate. In this case, concentrations of orally administered compound in intestinal fluid remain lower than the solubility limit. For most compounds, the intestinal fluid concentrations in the lumen are much higher than systemic concentrations such that intestinal membrane permeation occurs under “sink” conditions (Figure 1A). Therefore, in the case of permeability rate-limited absorption, intestinal permeation is often rate-limited by the ability of the compound to move across the unstirred water layer or the intrinsic membrane permeability of the compound across membranes.2 As with the case described for permeability rate-limited absorption, in cases of dissolution and solubility rate-limited absorption, the assumption is that absorption occurs under sink conditions. In recent years, there has been an increase in the utilization of complex physiologically based pharmacokinetic (PBPK) models to dynamically integrate the interplay between drug dissolution and intestinal permeation to provide a prediction of oral absorption.4,5 Such exercises enable visualization of which process is rate limiting and also provide a means to identify which changes in controllable properties of molecules can help to improve oral absorption. Despite improvements in our understanding of oral absorption through the use of PBPK models, there remains much to learn. This was illustrated in a recent exercise to predict the oral pharmacokinetics of a large set of compounds assembled by pharmaceutical industry using PBPK models.6 At best, only 23% of the predictions of oral pharmacokinetics made were considered to be performed with a medium to high degree of accuracy.6 With the increase in the numbers of poor solubility compounds being developed, activities that are necessary for progression of either a specific compound or a chemical scaffold in the drug discovery setting are often hampered by the inability to achieve adequate exposure following oral administration. In particular, activities such as in vivo preclinical proof-of-concept studies in animal models, and preclinical toxicology assessments require the high systemic exposures following oral administration. In the case of in vivo preclinical proof-of-concept studies, compounds are often dosed in 3981

dx.doi.org/10.1021/mp400088q | Mol. Pharmaceutics 2013, 10, 3980−3988

Molecular Pharmaceutics

Article

Figure 2. (A) One compartment model with nonlinear elimination fit to D8-DCU plasma concentration. (B) D8-DCU observed and predicted (dotted line) plasma concentration−time profile when infused intravenously at three infusion rates. Associated pharmacokinetic parameters estimated following model fitting is presented on the right.

250B-10R) established an electrical field between the spray head and the grounded particle collection substrate. The distance between the spray head and substrate was approximately 4.0 cm, and the applied high voltage ranged between 11 and 13 kV. The infusion speed of the D8-DCU stock was fixed at 1 uL/min. The spray modes of the system were monitored by viewing the liquid meniscus at the exit of the spray head. The viewing was accomplished through illuminating the liquid meniscus using a diffusing light source and observing its shape via a microscope (Infini-Van Video Microscope), digital camera (model KR 222, Panasonic), and high resolution monitor (Sony Trinitron, 1028 × 256 pixels). The compound cone-jet mode operation of the ESI system ensures production of monodisperse droplets; compressed nitrogen was also applied to facilitate the solvent evaporation when drug particles were collected on aluminum foil. Collected particles were then dried for 24 h using an in-house vacuum. The solid-state properties of the nanoparticles were evaluated by PXRD at room temperature with a Rigaku (Texas, USA) MiniFlex II desktop X-ray powder diffractometer. The formulation was prepared by reconstituting the nanoparticles in the IV vehicle (1% Cremophor EL and 5% PVP in pH 7.4 phosphate buffered saline). Particle size was analyzed using a Nanotrac (PA, USA) instrument. Formulation concentration was confirmed by LC/ MS/MS. Formulation for DCU Oral Suspension. DCU suspension was prepared by wet milling DCU in the oral vehicle consisting of 0.5% methylcellulose and 0.2% Tween 80 in water. Particles were characterized by PXRD. Particle size was determined using a Microtrac (PA, USA) instrument. Formulation concentration was confirmed by LC/MS/MS. Determination of Solubility in Buffer, Simulated Gastic Fluid (SGF), and Fasted State Simulated Intestinal Fluid (FASSIF). Solubility of DCU in phosphate buffer (pH 7.2), SGF (pH 1.2), and FASSIF (pH 6.5) was determined by adding an excess of DCU into either buffer, SGF, or FASSIF followed by equilibration at 37 °C for >12 h. Following equilibration, excess drug was removed by centrifugation and

and DCU oral absorption under non-sink conditions. We show that high systemic concentrations of a poorly soluble compound can suppress its own oral absorption. More importantly, we identify the relevant DCU concentrations creating the gradient driving the intestinal permeation process. A new parameter, the maximum permeation extraction ratio (MPER), is proposed and provides a means to assess the extent of non-sink permeation.

■

EXPERIMENTAL SECTION Experimental Material. HPLC grade acetonitrile was obtained from Burdick & Jackson (Muskegon, MI), reagent grade formic acid was obtained from EM Science (Gibbstown. NJ), DCU was purchased from Sigma-Aldrich (St. Louis, MO), and deuterated DCU (D8-DCU) was made in-house (Genentech, Inc., South San Francisco, CA). Formulation for D8-DCU Intravenous (IV) Nanosuspension. The D8-DCU nanosuspension formulation was prepared using an electrospray device and a stock solution of D8-DCU (2 mg/mL in ethanol). The electrospray device used in our study was purchased from the University of Minnesota with inlets that were modified in-house as described.8 The spray head consists of three coaxially arranged stainless steel capillaries. The inner capillary tube has an inner diameter (ID) of 300 μm and an outer diameter (OD) of 785 μm, the middle capillary tube has an ID of 1150 μm and an OD of 1575 μm, and the outer capillary tube has an ID of 1956 μm. The capillary arrangement in the spray head creates three separate liquid flow channels, allowing separate introduction of liquid from the inner, middle, and outer capillary tubes. The spray head is connected to a high-voltage power supply via a high voltage cable. For this experiment, only the inner capillary was used. The inner, middle, and outer liquids are connected to separate flow channels via Teflon tubes connected to three Harvard syringe pumps (model PHD 2000). Metal substrate was used in the studied system as the electrical ground reference. A DC high-voltage power supply (Bertan, model 3982

dx.doi.org/10.1021/mp400088q | Mol. Pharmaceutics 2013, 10, 3980−3988

Molecular Pharmaceutics

Article

Figure 3. Rat oral physiologically based pharmacokinetic model. Abbreviations are defined in the Experimental Section.

Menten constant in μM), and Vd (the volume of distribution in L/kg) are presented as the parameter estimate followed by the coefficient of variation (CV) for the estimate in parentheses. All other pharmacokinetic parameters were calculated by noncompartmental methods as described in Gibaldi and Perrier9 using WinNonlin version 3.2 (Pharsight Corporation; Mountain View, CA, USA). Parameters are presented as an arithmetic mean ± standard deviation Physiologically Based Pharmacokinetic (PBPK) Modeling. A nine intestinal segment oral absorption PBPK model was constructed in ModelMaker version 4.0 (Oxford, UK) based upon a modified version of the advanced compartmental absorption and transit model (ACAT).4,10 The configuration of the described model is shown in Figure 3. The first segment represents the stomach followed by seven small intestine (SI) segments and the colon segment. Two compartments were assigned for each of the nine segments representing the gastrointestinal (GI) tract and were designated to contain either solid (S) or dissolved (Dis) drug. Rate constants describing gastric emptying (KS), small intestine compartmental transit (KT), and colon emptying (KC) were set 4 h−1, 4.76 h−1, and 0.0172 h−1, respectively, in order to reflect the GI transit times presented for the Rat Model in GastroPlus (Simulations Plus, Inc., Lancaster, CA). Specific pH for the GI compartments were set at pH values presented for the Rat Model in GastroPlus as follows: stomach, pH 4.15; SI1, pH 6.5; SI2, pH 6.6; SI3, pH 6.7; SI4, pH 6.75; SI5, pH 6.8; SI6, pH 6.95; SI7, pH 7.1; colon, pH 6.7. Solubility for the stomach compartment was set at 2.0 μg/mL (solubility of DCU in simulated gastric fluid at stomach pH). The volume in the stomach compartment was set at 3 mL. Aside from the stomach, the volume of the lumen for each GI compartment was calculated using following equation: Vlumen = πLR2, where L is the length of the intestinal section, R is the radius. A radius of 0.2 cm was used for all SI compartments, and 0.8 cm was used for the colon. The following lengths (in cm) were used for each GI compartment as per the Rat Model in GastroPlus: SI1, 17.01; SI2, 12.53; SI3, 9.23; SI4, 6.8; SI5, 5.01; SI6, 3.69; SI7, 2.72; colon, 0.7 Dissolution Process. Dissolution rate of solid drug in the oral absorption PBPK model was governed by the following equation based upon the Noyes−Whitney equation:

the concentration of dissolved drug in either buffer, SGF, or FASSIF was determined by LC/MS/MS. Each determination was performed in triplicate. Solid sample harvested centrifugation was checked by PXRD. As DCU contains no ionizable groups, it was assumed that the solubility of DCU was the same from pH 6 to 7.4 in FASSIF. The measured values from these studies were used as inputs for the PBPK model described below. DCU and D8-DCU Pharmacokinetic Study in Rats. The purpose of this study was to examine the effect of intravenous infusions of D8-DCU on the oral administration of DCU in male Sprague−Dawley rats (Sino-British SIPPR/BK Laboratory Animal Co., Ltd., Shanghai, China). At study initiation, rats used weighed from 200 to 350 g and were 7−9 weeks of age. All animals were fasted overnight before dosing. Four groups of rats (n = 4 per group) receive a single 30 mg/kg oral dose of DCU as suspension prepared in MCT (0.5% methylcellulose and 0.2% Tween 80 in water). In addition to the single 30 mg/ kg oral dose of DCU, groups 2, 3, and 4 received an intravenous infusion of D8-DCU at 13, 17, and 22 mg/kg/h for 3 h (formulated as a 4.8 mg/mL D8-DCU nanosuspension), starting at 2 h before the administration of the oral dose and continuing 1 h post oral dose. For group 1, blood samples (approximately 200 μL per sample) were collected from the carotid artery at predose, 0.167, 0.333, 0.5, 0.67, 0.83, 1, and 2 h post oral dose. For groups 2−4, blood samples were collected at the following time points relative to the start of the D8-DCU infusion: predose, 1, 1.5, 2 (just before the 30 mg/kg oral dose is given), 2.167, 2.333, 2.5, 2.67, 2.83, 3, and 4 h post start of the D8-DCU infusion. All blood samples were collected into tubes containing K2 EDTA and then chilled on ice until centrifugation. Samples were centrifuged within 1 h of collection. Plasma was collected and kept frozen on dry ice before storage at approximately −70 °C. Concentrations of DCU and D8-DCU in plasma were quantitated by LC/MS/ MS. Pharmacokinetic Modeling of D8-DCU Infusion Data. Parameters describing the pharmacokinetics of D8-DCU were estimated by fitting a one compartment model with nonlinear elimination as described in Figure 2A to D8-DCU mean concentration−time data using SAAM II (Saam Institute, University of Washington, Seattle, WA). Estimates of Vmax (the maximum elimination rate in μM/h/kg), Km (the Michaelis− 3983

dx.doi.org/10.1021/mp400088q | Mol. Pharmaceutics 2013, 10, 3980−3988

Molecular Pharmaceutics dXsolid 3D =− (Csolubility_IF − CGItract) × Xsolid ρrh dt

Article

intestinal lumen into the systemic circulation under non-sink conditions is defined as follows:

(1)

where Xsolid is the amount of undissolved DCU in the S compartments, D is the DCU diffusion coefficient (default = 5.4 × 10−4 cm2/min), ρ is the drug particle density (DCU = 1.35 g/cm3), r is the particle radius (0.4 μm), h is the diffusion layer thickness (0.4 μm based on if r < 30 μm, h = r; see ref 2), Csolubility_IF is the measured solubility in simulated intestinal fluid (FASSIF; DCU = 3.59 at pH from 6.0 to 7.4 (measured value 3.59 ± 0.67 μg/mL)), and CGItract is the concentration of dissolved DCU in the GI tract segment. Upon the basis of the equation above, dissolution of DCU was assumed to be driven by the difference between the FASSIF solubility and the concentration of dissolved DCU in the GI segment. Intestinal Membrane Permeation Process. The rate of membrane permeation of dissolved DCU from the GI tract into the system circulation in the PBPK model was governed by the following equation which assumes that absorption occurs via passive permeability:10

MPER =

CsolFREE

(4)

Upon the basis of the equation described above, under sink conditions, MPER would range from 0.9 to 1. Non-sink conditions for permeation would occur at MPER of CsolFREE (solubility in buffer), then the rate of permeation would be governed by the following equation:

■

RESULTS Intravenous Nanosuspension and Oral Formulation Characterization. The D8-DCU nanosuspension was characterized in order to ensure its proper preparation. D50 and D90 of D8-DCU particles in the IV formulation were determined to be 0.23 and 0.37 μm, respectively. PXRD characterization of the solid form of the nanomaterial indicated no discernible change in crystal form due to the electrospray process. Using a previously published theoretical calculation8 and the D50 listed above, the calculated dissolution time of an average particle was less than 3 s. Upon in vivo intravenous delivery, the actual dissolution time should be much more rapid because turbulent blood flow in the vein should serve to both reduce the diffusion boundary thickness and rapidly disperse the injection formulation minimizing local concentration effects.8 The D50 and D90 of the DCU particles in the oral suspension prepared by wet milling method7 was also characterized and determined to be 2.4 and 6.7 μm, respectively. DCU and D8-DCU Pharmacokinetic Study in Rats. The concentration−time profile following varying infusion rates of D8-DCU is presented in Figure 2B. Concentrations of D8DCU increased progressively with increased infusion rate. D8DCU concentrations dropped rapidly following the end of the infusion at 3 h post infusion start. Within 1 h of ending the infusion, D8-DCU concentrations were