Introduction

Nausea and vomiting are common and distressing side effects associated with chemotherapy for malignant disease and are the principal cause of discontinuation of cancer chemotherapy. Prevention of these side effects is important for ensuring continuation of chemotherapy and for improving patient quality of life [1–6]. One of the main mechanisms of nausea and vomiting induced by antineoplastic agents is activation of 5-HT3 receptors, which exist on vagus nerve afferent fibers in the small intestine mucous membrane [7–10]. Antineoplastic drugs have been shown to elicit direct effects on the intestine and to induce the release of serotonin from enterochromaffin cells in the small intestinal mucosa, where more than 90% of the total body serotonin is contained. It is thought that released serotonin stimulates vagal afferent fibers through 5-HT3 receptors located in the vagal afferent terminals in the gastrointestinal tract and initiates sensory signals to the area postrema and the emetic center, thereby initiating nausea and vomiting. 5-HT3 antagonists competitively inhibit serotonin at its specific binding sites, 5-HT3 receptors, and thereby elicit an antiemetic effect. Therefore, receptor occupancy may be a more appropriate indicator of the antiemetic activity of a 5-HT3 antagonist than its dose or blood concentration.

In this study, we analyzed receptor occupancy (ΦB) of 5-HT3 antagonists by integrating pharmacokinetic and receptor-binding kinetic parameters published in the literature based on receptor occupancy theory [11–13, 31–34] in order to quantitatively determine the relationship between ΦB and antiemetic efficacy.

Materials and methods

Pharmacokinetic and pharmacodynamic parameters of 5-HT3 antagonists

The area under the plasma concentration-time curve (AUC0–∞) [14–22] and value of the plasma unbound fraction (fu) [23–27] of both parenteral preparations (granisetron hydrochloride, ondansetron hydrochloride, azasetron hydrochloride, and ramosetron hydrochloride) and oral preparations (granisetron hydrochloride, ondansetron hydrochloride, azasetron hydrochloride, ramosetron hydrochloride, and tropisetron hydrochloride) of 5-HT3 antagonists were obtained from the reported data of a phase I clinical trial. The value of the pharmacodynamic parameter receptor dissociation constant (KI) of each drug was obtained from reported results of in vitro binding inhibition experiments using radioactive ligands in the cerebral cortex of rats [28–30]. The dose stated in the product package insert of each drug for use in Japan was taken as the standard dose of each drug.

Calculation of receptor occupancy of 5-HT3 antagonists

5-HT3 antagonists bind to receptors by competing with serotonin, which is an agonist for 5-HT3 receptors, and the efficacy of the antagonists is elicited by displacement of serotonin from 5-HT3 receptors [13]. These molecular interactions can be represented as follows:

$$A + R \rightleftarrows AR,\quad K_{{\text{A}}} = A \times R/AR$$
(1)
$$B + R \rightleftarrows BR,\quad K_{{\text{B}}} = B \times R/BR$$
(2)

where R is the concentration of unoccupied receptors, A and B are the effective (unbound) concentrations of an agonist (serotonin) and an antagonist near 5-HT3 receptors, respectively, and KA and KB represent the dissociation constants of the agonist and antagonist, respectively. The total receptor concentration (R0) is the sum of the occupied and unoccupied receptors:

$$R_{0} = R + AR + BR$$
(3)

The occupancy of 5-HT3 receptors by the antagonists (ΦB) can be expressed by the following equation:

$$ \Phi _{{\text{B}}} = \frac{{BR}} {{R_{0} }} \times 100 $$
(4)

Substitution of Eqs. 1–3 into Eq. 4 gives the following equation:

$$\Phi _{{\text{B}}} = \frac{B}{{K_{{\text{B}}} {\left( {1 + A/K_{{\text{A}}} } \right)} + B}} \times 100$$
(5)

Therefore, ΦB can be predicted from the drug concentration near receptors and dissociation constants. The concentration of a 5-HT3 antagonist near receptors (B) was approximated by the unbound drug concentration in plasma (Cpf, nanomoles), since it can be assumed that drug transfer across capillary vessels is due to simple diffusion and there is no active transport process. Plasma unbound drug concentrations at steady-state (Cssf) were calculated from AUC0–∞ and fu after intravenous or oral administration of each drug at the standard dose. Metabolites of each 5-HT3 antagonist were not taken into consideration in this study, since their contribution to antiemetic efficacy is considered to be small. On the other hand, ΦB should vary with change in concentration of serotonin (A) as an agonist. It was assumed that the value of A was smaller than that of KA (150 nM) (A«KA) [28] and that its variation after administration of a 5-HT3 antagonist would not affect the value of ΦB. Accordingly, ΦB could be predicted by the following equation:

$$ \Phi _{{\text{B}}} = \frac{B} {{K_{{\text{B}}} + B}} \times 100 = \frac{{{\text{Cp}}^{{\text{f}}} }} {{K_{{\text{B}}} + {\text{Cp}}^{{\text{f}}} }} \times 100 $$
(6)

Relationship between receptor occupancy of 5-HT3 antagonists and antiemetic action after administration of cisplatin

We analyzed the relationship between the average receptor occupancy of 5-HT3 antagonists (ΦB) predicted as described above and antiemetic effects. We obtained complete vomiting inhibition rates from the antiemetic effects observed with preventive and curative intravenous administration and with preventive oral administration of 5-HT3 antagonists after a single intravenous administration of cisplatin (not less than 50 mg/m2) found in clinical studies conducted in Japan [35–45]. It was also assumed that each drug examined acted as a complete antagonist, and the antiemetic effect (EB) of a 5-HT3 antagonist after administration of an antineoplastic agent could therefore be expressed as a function of ΦB.

$$E_{{\text{B}}} = f{\left( {\Phi _{{\text{B}}} } \right)}$$
(7)

Results

Pharmacokinetic and pharmacodynamic parameters of 5-HT3 antagonists

AUC0–∞, fu, and KI values of each 5-HT3 antagonist at standard (curative) doses obtained from the literature are shown in Tables 1 and 2. Table 1 summarizes the parameters for parenteral drugs, and Table 2 summarizes the parameters for oral drugs. As shown in Tables 1 and 2, the KI values of the drugs examined differed by up to almost 40-fold, although the KI value of each drug is lower than the 150 nM of serotonin. The AUC0–∞ values of the drugs examined differed by up to almost 80-fold.

Table 1 Pharmacokinetic and pharmacodynamic parameters of four parenteral 5-HT3 antagonists
Full size table
Table 2 Pharmacokinetic and pharmacodynamic parameters of five oral 5-HT3 antagonists
Full size table

Receptor occupancies of 5-HT3 antagonists

Predicted values of average receptor occupancy (ΦB) after intravenous administration of each 5-HT3 antagonist at the standard dose are listed in Table 3, and those after oral administration are shown in Table 4. The value of ΦB after intravenous administration was predicted to be relatively high (not less than 65%), although there were large differences in the standard dose KI value and Cssf among the drugs examined. The value of ΦB after oral administration was predicted to be more than 50%.

Table 3 Average 5-HT3 receptor occupancies of four 5-HT3 antagonists after intravenous administration of standard doses (Cssf unbound concentration of 5-HT3 antagonists at steady-state, Φ B average 5-HT3 receptor occupancy of 5-HT3 antagonist)
Full size table
Table 4 Average 5-HT3 receptor occupancies of five 5-HT3 antagonists after oral administration of standard doses (Cssf unbound concentration of 5-HT3 antagonists at steady-state, Φ B average 5-HT3 receptor occupancy of 5-HT3 antagonist)
Full size table

Relationship between receptor occupancy of 5-HT3 antagonists and antiemetic effects after administration of cisplatin

Regarding efficacy of 5-HT3 antagonists in cisplatin therapy, the relationships between standard dose and plasma unbound drug concentrations at steady-state (Cssf) and complete vomiting inhibition rates for prevention or cure [35–45] were analyzed. No significant positive correlations were found in these relationships.

The relationships between average receptor occupancy (ΦB) and complete vomiting inhibition rate for prevention and cure are shown in Fig. 1. Significant (P<0.05) positive relationships were found between complete vomiting inhibition rate of the 5-HT3 antagonists after intravenous and oral administration for prevention.

Fig. 1
figure 1

Relationships between antiemetic effects and average receptor occupancy of 5-HT3 antagonists: A intravenous administration for prevention; B intravenous administration for cure after development of vomiting; C oral administration for prevention

Full size image

Discussion

When drug action is elicited by specific receptors, it is important to clarify kinetically the relationship between drug concentration in the region of the site of action of the drug and binding of the drug to receptors. Therefore, receptor occupancy of a drug is considered to be a rational index of curative effectiveness. In this study, we predicted the average receptor occupancy by integrating pharmacokinetic and pharmacodynamic parameters after administering 5-HT3 antagonists, aiming to establish the relationship between ΦB and efficacy.

In this study, we assumed that 5-HT3 receptor antagonists are transported across capillary walls by simple diffusion, not mediated by a special transport mechanism, and that drug permeability across vessels is not the rate-determining step for exhibiting their efficacy. Thus, the drug concentration in the intercellular space of the small intestine was assumed to be equivalent to the plasma unbound drug concentration. Confirmation of the validity of this assumption would require measurements of unbound drug concentrations in the gut using a microdialysis technique. No such data are currently available.

The data used in this study were obtained from many reports. Pharmacokinetic data were obtained from reported results of clinical studies using standard doses, and data on receptor binding affinity were obtained from reported results of studies using the same animal species and the same tissue region.

Organ-derived and species-derived differences in the KI values (as receptor dissociation constants) among the many 5-HT3 antagonists are shown in Fig. 2. The relationship between KI values in the rat cerebral cortex and intestine is shown in Fig. 2A, and the relationship between KI values in the rat intestine and human jejunum is shown in Fig. 2B. Significant correlations were found in both relationships, with correlation coefficients of 0.93 (P<0.001) and 0.90 (P<0.01), respectively, suggesting that organ-specific and species-specific differences in the KI values of 5-HT3 receptors are small [46]. Moreover, it has been reported that there is a good correlation between the KI value in the cerebral cortex and the pA2 value of the depolarization reaction in the vagus nerve of the rat [29]. Therefore, the use of the KI value in the cerebral cortex in this study seems to have been appropriate.

Fig. 2
figure 2

Difference in 5-HT3 receptor dissociation constants between rat and human. A Relationship between KI values in rat cerebral cortex and rat gut; B relationship between KI values in rat gut and human jejunum

Full size image

It is thought that the concentration of released serotonin around 5-HT3 receptors differs among antineoplastic drugs. Since data on amounts of serotonin released by antineoplastic drugs at the gastrointestinal mucous membrane were not available, it was not possible to take this possible difference into consideration in this study. It has been reported that the concentrations of serotonin and its metabolite, 5-HIAA, in plasma do not change but that urinary 5-HIAA excretion increases after antineoplastic treatment in cancer patients experiencing nausea and vomiting [47, 48]. Therefore, determination of time courses of the amounts of serotonin released by antineoplastic agents would be required for detailed analysis.

The predicted average receptor occupancy of all of the drugs examined in this study were relatively high (more than 65%) after intravenous administration and more than 50% after oral administration. These results can be explained by the spare receptor theory. Linear relationships were found between average receptor occupancy and complete vomiting inhibition rates of all of the 5-HT3 antagonists used in cisplatin therapy, with correlation coefficients of 0.99 and 0.91 (P<0.05) after intravenous and oral administrations for prevention, respectively. These results suggest that 5-HT3 receptor occupancy is an appropriate index of clinical efficacy of 5-HT3 receptor antagonists, with higher receptor occupancy indicating more extensive antiemetic action. Concerning ondansetron, dosage differs between Japan (4 mg for oral and intravenous administration) and the US (24 mg for oral administration and 32 mg for intravenous administration). The average receptor occupancy after these dosages in the US have been found to be 85.4% for oral administration and 94.1% for intravenous administration, respectively. This suggests that the dosage of ondansetron in Japan is low.

The analysis based on the receptor occupancy theory of the antiemetic effect of a 5-HT3 receptor antagonist used in this study should be very useful for not only estimating the rational dosage regimen but also determining the standard dose of a new drug using experimental data obtained from a preclinical study.