Unexpected relationship between plasma protein binding and the pharmacodynamics of 2-NAP, a CCK1-receptor antagonist

Abstract

What is already known about this subject?

• Two chemically diverse CCK1 receptor antagonists have been shown clinically to inhibit CCK-evoked contraction of human gallbladder [2, 3]. These studies have not examined the relationship between plasma concentration and effect, the latter usually considered to be predictive from the free drug concentration [8].

• We wanted to examine our novel CCK1 receptor antagonist in this validated model and also to explore its PK-PD relationship.

What this study adds

• 2-NAP inhibited CCK-evoked human gallbladder contraction in vivo but at a plasma free concentration that was, in theory, too low to have achieved adequate CCK1 receptor occupancy.

• The study serves as a caveat to the assumption that free plasma concentration can be used to predict pharmacological effect.

Aims

To study the pharmacokinetics and pharmacodynamics of 2-NAP (2-naphthalenesulfonyl-L-aspartyl-(2-phenethyl)amide), a selective CCK1 receptor antagonist in healthy volunteers.

Methods

2-NAP was given to 12 healthy male volunteers in an ascending dose, safety and PK phase 1a study by 1 h i.v. infusion (0.6–9.6 mg kg⁻¹ h⁻¹). A further 12 healthy male volunteers received i.v. CCK-8S (6.25 pmol kg⁻¹ h⁻¹) to produce gallbladder contraction, measured by ultrasound recordings of gallbladder volume, and the effect of concurrent i.v. 2-NAP administration was studied. Plasma protein binding in vitro and ex vivo was measured by ultrafiltration and by equilibrium dialysis.

Results

2-NAP was generally well tolerated, displayed linear pharmacokinetics and a very high degree of plasma protein binding (99.9%). A 105 min i.v. CCK-8S infusion induced a reduction in gallbladder volume of 14.9 (±7.0) ml during placebo co-infusion and this was reduced to 2.4 (±5.9) ml when 2-NAP was co-infused with CCK-8S (P = 0.00024, paired t-test, mean change 12.5 ml; 95% CI For mean 7.4, 18.3 ml). This extent of inhibition was consistent with a 2-NAP total plasma concentration of 36 µm, but when protein binding corrections were made, the ‘free concentration’ of 2-NAP was only 0.04 µm, a value much less than the average equilibrium dissociation constant of 2-NAP for human CCK1 receptors (∼0.7 µm).

Conclusions

The pharmacological effect of a drug is usually considered to be determined by its free concentration. However, the complete inhibition of CCK-8S-evoked gallbladder contraction by a free plasma concentration of 0.04 µm 2-NAP was much greater than would have been predicted from simple drug-receptor occupancy theory and cautions against the general use of free concentration of drug for predicting pharmacological effect.

Introduction

Peripheral physiological actions of cholecystokinin mediated by CCK1 receptors include the contraction of gallbladder smooth muscle, stimulation of pancreatic enzyme synthesis and secretion, as well as inhibition of gastric acid secretion and emptying. CCK1 receptor antagonists have been explored in man previously (e.g. loxiglumide [1, 2], devazepide [3]), and represent useful pharmacological tools and potentially valuable therapeutic agents. Here we have examined 2-NAP, a hydrophilic (log P_{chloroform:pH7.4buffer} = −0.9), CCK1 receptor antagonist with a pK_B across species (including man) of between 6 and 7 (dissociation constant range 0.1–1 µm) [4, 5]. 2-NAP has also been shown to be highly selective (∼300-fold) for CCK1 over the hormone-family related CCK2/gastrin receptors in isolated tissues and radioligand binding assays, in addition to 49 other pharmacological loci. The high selectivity and good safety profile determined from preclinical studies and the likelihood of minimal penetration to the CNS [6] indicated that 2-NAP was a potentially useful clinical tool for studying peripheral CCK physiology and pathophysiology.

The results of the first administration of 2-NAP into man are reported here. The objectives were to assess the safety, tolerability and pharmacokinetics of 2-NAP following intravenous infusion and then to conduct a pharmacodynamic study of CCK-8S-evoked gallbladder contraction. The studies reported here are part of a larger series of investigations in healthy volunteers conducted at Drug Development Scotland Ltd, Dundee and at King's College School of Medicine and Dentistry, London which illustrated an unusual relationship between plasma concentration and pharmacodynamic effect.

In the pharmacodynamic study, the aim was to achieve a 2-NAP free plasma concentration of ∼6.5 µg ml⁻¹ (equivalent to 14.5 µm), which is 20 times the equilibrium dissociation constant (K_B) at CCK1 receptors in human tissues [4, 5, see Discussion). If the underlying CCK-8S dose–response curve is assumed to be a rectangular hyperbola and 2-NAP behaves in a reversible competitive manner then this concentration of 2-NAP would be expected to produce a readily-detectable reduction of 94% in the response to an ED₅₀ dose of agonist. However, in practice due to the high degree of plasma protein binding even the highest dose which could be studied did not achieve more than 1% of the target free plasma concentration. Notwithstanding this, the compound was highly effective. A possible explanation for this apparent anomaly is discussed.

Methods

Volunteer studies

2-NAP was given in an ascending dose, single-blind, placebo-controlled, phase 1a design, in two overlapping groups of six healthy male volunteers (aged 18–29 years), at 14 day intervals. Each subject received three treatments which comprised of one placebo (saline), plus two different doses of 2-NAP. The treatments were given as 60 min i.v. infusions of 500 ml.

In a phase 1b study, an i.v. infusion of 2-NAP was administered prior to a CCK-8S i.v. infusion and the effect on gallbladder volume measured. The study was a placebo-controlled, crossover study in 12 healthy male volunteers, studied on two occasions with 7 days between treatment sessions. On the morning of each study day, a baseline i.v. infusion of saline was given for 60 min. The saline infusion was either continued or substituted with 2-NAP (9.6 mg kg⁻¹ h⁻¹ i.v.) for a further 60 min. CCK-8S infusion (6.25 pmol kg⁻¹ h⁻¹ i.v.), a dose shown from a previous study to produce an approximate 50% reduction in basal, gallbladder volume [1], was started 15 min after commencement of the 2-NAP infusion and continued for a total of 105 min. Gallbladder volume was measured by ultrasound at 15 min intervals for 4 h.

Both studies were conducted following informed and written consent at Drug Development Scotland Ltd, Ninewells Hospital & Medical School, Dundee after approval by the Tayside Committee on Medical Ethics.

Chemical assay of 2-NAP

Concentrations of 2-NAP in human plasma and urine were determined using solid phase extraction followed by reverse phase HPLC (Bond ElutTMc8 column; mobile phase 40% acetonitrile containing 0.1% acetic acid at 2 ml min⁻¹ flow rate) with fluorescence detection (excitation wavelength 240 nm, emission wavelength 340 nm). The average CV and the limits of quantification were 4.0% and 10 ng ml⁻¹, respectively.

Gallbladder volume measurement

Gallbladder volumes were measured by real time ultrasonography using an Aloka SSD-500 convex sector/linear scanner and UST-93 N-3.5 MHz convex sector probe similar to that previously described [7].

Plasma protein binding

A range of drug concentrations (corresponding to the plasma concentrations measured in the pharmacokinetic studies) was used for in vitro studies in which plasma protein binding was determined by ultrafiltration (anisotropic hydrophilic ultrafiltration membrane, ambient temperature, 2000 × g) and by equilibrium dialysis (Dianorm apparatus with a regenerated cellulose membrane: Type Spectrapor 2, molecular weight cut off ∼13 000 Da). The proportion of 2-NAP bound to plasma proteins was also determined ex vivo, using ultrafiltration following a 1 h i.v. infusion of 2-NAP (0.6–9.6 mg kg⁻¹ h⁻¹) obtained during the phase 1a study. 2-NAP was assayed by HPLC.

Analysis of results

The area under the 2-NAP plasma-concentration–time curve was determined using Simpson's trapezoidal rule. Data are presented as mean (±SD) except where stated otherwise. The change in gallbladder volume, measured as the difference between mean values before, (a) at the end of the 105 min CCK-8S infusion and (b) at the end of the 1 h 2-NAP infusion (and corresponding to 45 min infusion of CCK-8S), was determined for each subject. The differences between the co-infused placebo and 2-NAP treatment arms were compared using Student's t-test for paired data.

Compounds

CCK-8S, was imported from E. R. Squibb & Sons, Montréal, Canada. 2-NAP was manufactured by Palmer Research, Holywell, UK and prepared as 98% pure, lyophilized powder by the Department of Pharmaceutical Development, University of Strathclyde, Scotland. Both drugs were dissolved in 0.9% w/v saline.

Results

Safety

Over the phase 1a and 1b study, treatment was generally well tolerated and there were no withdrawals from either study. There were no significant changes in blood pressure, pulse, ECGs, blood chemistry, haematology, and stool frequency or consistency. The exceptions to this were one subject reporting central epigastric discomfort for 6 h starting 1.5 h after receiving 7.2 mg kg⁻¹ h⁻¹ 2-NAP unaccompanied by any physical signs. A second subject reported lumbar and abdominal pain following 9.6 mg kg^–1 h⁻¹ 2-NAP plus CCK-8S followed by nausea and vomiting accompanied by elevations in blood urea and creatinine; these returned to normal upon follow up.

Pharmacokinetics of 2-NAP in the phase 1a study

Peak 2-NAP plasma concentrations and areas under the plasma concentration vs. time curves increased linearly with dose (0.6–9.6 mg kg⁻¹ h⁻¹, Figure 1). At the dose used for the phase 1b study (9.6 mg kg⁻¹ h⁻¹), clearance was 90.6 (±15.0) ml h⁻¹ kg⁻¹ and apparent volume of distribution (V_ss) was 129 (±15.3) ml kg⁻¹; these were not significantly different at the various infusion rates. Analysis of individual urine samples showed that the proportion of the dose excreted unchanged in the urine over 24 h was 21% (±5.9%) and was independent of dose; 95% (±6%) of this fraction was excreted during the first 6 h after dosage. Similar PK parameters were seen in the phase 1b study (data not shown).

Figure 1.

Mean plasma concentrations of 2-NAP over 24 h in 12 healthy volunteers. 2-NAP was infused intravenously for 1 h at the following rates (•) 0.6, (○) 1.2, (▴) 2.4, (▵) 4.8, (▪) 7.2, (□) 9.6 mg kg⁻¹ h ⁻¹. Each line is the mean data from four subjects. The hatched line shows the equilibrium dissociation constant (K_B) of 2-NAP for human CCK1-receptors [4, 5 and Discussion]. Filled bar = duration of i.v. infusion. Inset: Relationship between 2-NAP dose and AUC(0,t) (area under the total plasma concentration-time curve from time = 0 to last quantifiable concentration)

Inhibition of CCK-8S-induced gallbladder contraction – phase 1b study

The reduction from baseline in gallbladder volume following 45 min CCK-8S infusion was 10.8 (±8.0) ml during 1 h placebo co-infusion. This was significantly more than the change observed following 1 h co-infusion with 2-NAP, i.e. −2.7 (±5.3) ml (P = 0.00035, paired t-test, mean difference between treatments 13.5 ml; 95% CI 7.9, 19.2 ml). The total plasma concentration of 2-NAP at this time point was 74.3 (±9.8) ⁻¹ ml⁻¹ (166 µm).

Inspection of Figure 2 shows that over the 45 min period following termination of the 2-NAP infusion, but while the CCK-8S infusion was maintained, there was still a substantial difference between the two treatment arms and thus this was analyzed further. The reduction from baseline in gallbladder volume, 14.9 (±7.0) ml, following 105 min infusion of CCK-8S in the placebo co-infusion arm of the study was also significantly greater than that observed in the 2-NAP co-infusion arm, i.e. 2.4 (±5.9) ml (P = 0.00024, paired t-test, mean difference between treatments 12.5 ml; 95% CI 7.4, 18.3 ml; Figure 2). The total plasma concentration of 2-NAP at this time point was 16.3 (±6.3) µg ml⁻¹ (36 µm).

Figure 2.

Time course of the effect of i.v. infusion for 105 min of CCK-8S (6.25 pmol kg⁻¹ h⁻¹) on gallbladder volume (mean ± SEM) following the initiation of a 60 min i.v. infusion of 2-NAP (9.6 mg kg⁻¹ h⁻¹) or saline to 11 healthy male volunteers. Horizontal bars show the period of i.v. infusions

The gallbladder could not be accurately visualized in one subject during one of the phases of the study and the data from this subject were excluded from the analysis.

Protein binding

From the phase 1a study an ultrafiltration assay revealed that the free plasma concentration of 2-NAP at the end of a 60 min i.v. infusion (9.6 mg kg⁻¹ h⁻¹) was 0.096 (±0.006) µg ml⁻¹ corresponding to 99.9% plasma protein binding. Furthermore, from in vitro studies, the percentage of plasma protein bound 2-NAP was found to be 99.5%, 99.9%, 99.8% and 99.7% when determined by ultrafiltration at total concentrations of 4, 13, 40 and 130 µg ml⁻¹, respectively, corresponding to the total plasma concentration range observed in man (∼3–100 µg ml⁻¹). Similarly, when determined by equilibrium dialysis in vitro, 2-NAP was >99% bound to plasma proteins over this range.

Discussion

Following i.v. infusion of 2-NAP there was complete inhibition of the CCK-8S-evoked gallbladder contraction, whether gallbladder volume was measured at the end of 2-NAP infusion or at the end of the CCK-8S infusion (i.e. 1 h after stopping the 2-NAP co-infusion). The corresponding total plasma concentrations of 2-NAP were 166 µm and 36 µm, respectively. A previous functional study [4] using human gallbladder tissue indicated a 2-NAP equilibrium dissociation constant (K_B) for CCK1 receptors of 1.5 µm (pK_B 5.8). This value is similar to that obtained in radioligand binding assays on human gallbladder (K_i 0.35 µm (pK_i, 6.46) [5]) and human CCK1 receptors expressed in PC-3 cells: K_i 0.67 µm (pK_i 6.18, E. Harper personal communication). Assuming an average pK_i of 6.15 (i.e. equilibrium dissociation constant of 0.72 µm) then the total plasma 2-NAP concentrations were 230 or 50 times greater, depending upon the time point of comparison. However plasma protein binding from in vitro and ex vivo approaches, using various techniques, yielded the consistent finding that only 0.1% of 2-NAP was expected to be free. As it is generally assumed that the unbound drug concentration determines the pharmacological effect [8, 9], it seemed to us surprising that such high plasma protein binding would permit any discernible inhibition, given free 2-NAP plasma concentrations (i.e. 0.17 µm or 0.04 µm, respectively) that were substantially less than the equilibrium dissociation constant for CCK1 receptors. For example, simple drug receptor competition theory would have predicted only 10% and 3% reductions, respectively, by these free concentrations of 2-NAP against an ED₅₀ dose of CCK-8S. Evidently, in this phase 1b pharmacodynamic study, the free concentration of 2-NAP in the domain of the receptors behaved as though it was closer to the total concentration measured in the plasma. This interpretation indicates that the compound distributed significantly from the vascular to the extravascular receptor compartment.

Compounds which are highly plasma protein bound (e.g. methylene blue) and which do not leave the vascular compartment, have a volume of distribution of about 40 ml kg⁻¹, which corresponds to the plasma volume per kg body weight. However, 2-NAP had a V_ss of 129 and 109 ml kg⁻¹ after 1 h infusion from the phase 1a and phase 1b studies, respectively, supporting the assertion of flux through the capillary endothelium. The driving force for this is likely to be due more to free diffusion and interactions with tissue proteins, than to adsorption to extravascular lipophilic sites, owing to low lipophilicity of 2-NAP (log Pchloroform:pH 7.4 buffer = −0.9 [4]). Thus the distribution of the molecule, despite high plasma protein binding, permits access to the receptor domain. For the compound to behave as though the total plasma concentration is available to the receptor, all the processes involved in the distribution kinetics of 2-NAP must be occurring substantially faster than the relative off rate of the molecule from CCK1 receptors. In this way, although the steady-state free concentration in the receptor domain may be very low, the receptors effectively have access to the total amount of 2-NAP present.

Van Der Graaf et al. [10] noted that pK_i values of adenosine A1 ligands from radioligand binding assays in brain homogenates correlated almost identically with pKA estimates from rat heart rate assays based on using whole blood concentrations but not from using the free plasma concentrations. Interestingly, others have noted that plasma protein binding data can underestimate the amount of drug than can penetrate another extravascular compartment, namely the blood brain barrier [11, 12]. Nevertheless, the mechanism for these phenomena is unknown.

The current findings emphasize the need to exercise caution in extrapolating the observation of high plasma protein binding in the selection of target plasma concentrations required to produce a given pharmacological effect. In some cases, it is evident that the target dose can be predicted simply from the free plasma concentration (e.g. phenytoin [13]). However, with 2-NAP, this was not the case; whether this is due to the relatively unusual combination of an anionic drug of hydrophilic character, very high plasma protein binding and low V_ss, remains to be seen with further compounds.

The pharmacokinetic-pharmacodynamic and safety findings from these two studies were sufficiently encouraging to warrant further exploration. However, at the outset of a subsequent phase 1b study, an incident of abdominal pain, following the administration of the drug (from which an uneventful recovery ensued), led to a suspension of its development.