1. Introduction
Antitumor B (ATB) or Zeng Sheng Ping is a botanical drug, consisting of water extract or powder of six plants:
Sophora tonkinensis Gagnep.,
Polygonum bistorta L.,
Prunella vulgaris L.,
Sonchus brachyotus DC.,
Dictamnus dasycarpus Turcz., and
Dioscorea bulbifera L (
Table 1). The anticancer effect of ATB has been demonstrated in mouse models of bladder cancer [
1], lung cancer [
2,
3], and oral cancer [
4]. Our group successfully identified several key active compounds (KACs) of ATB in the A/J mouse oral tissues that are capable of inhibiting oral cancer cell proliferation and likely contribute to the chemoprevention effects of ATB [
3]. The four major KACs responsible for the anticancer activities of ATB were matrine (Matr), dictamine (Dict), fraxinellone (Frax), and maackiain (Maac) [
3]. ATB has also shown significant chemopreventive efficacy against human esophageal and oral cancers in several clinical studies in China [
5,
6,
7,
8]. ATB is currently undergoing clinical trials (
ClinicalTrials.gov Identifier: NCT04278989) in cancer patients for chemopreventive efficacy against oral cancer in the United States.
In chemoprevention clinical trials, where trial duration is usually longer (1–5 years or more), ensuring patient compliance is essential for a successful study design and execution. Therefore, patient monitoring for adherence to the medication regimen without using invasive procedures would significantly improve trial integrity and data quality of chemopreventive clinical studies. Our animal studies have shown that matrine is secreted via saliva in mice after i.p. administration of ATB [
9]. Therefore, we hypothesized that matrine secreted in human saliva can be used for compliance monitoring and act as a possible indicator of drug exposure at the target site. There has been no report published so far that examines the secretion of matrine in human saliva and very limited information about the human pharmacokinetics (PK) of ATB herbal mixture is available. Matrine is of particular importance as it is also reported to have anticancer activities [
10,
11] and is used as a quality control marker of ATB as per the Chinese Pharmacopeia.
Therefore, the purpose of the present study is to advance the understanding of ATB pharmacokinetic behaviors in preparation for oral cancer chemoprevention trials in humans. In the present study, we aim to: (1) Characterize the pharmacokinetics of ATB-KACs in healthy humans; (2) Conduct interspecies scaling of the PK parameters from mice to estimate the PK profiles of ATB-KACs in rats and humans; and (3) Model the correlation between drug concentrations in plasma and saliva such that saliva concentrations of active compounds can be successfully used to represent their plasma and oral cavity exposure, and monitor patient compliance in the upcoming and future clinical trials.
4. Discussion
The human PKs of ATB revealed that we can use the salivary secretion of Matr, one of the key active compounds of ATB at the site of action (i.e., oral cavity) as a measurement of matrine exposure and marker of patient compliance in long-term chemoprevention clinical studies of ATB. The PBPK modeling studies indicated that PBPK modeling can be developed to describe the PK behaviors of Matr across three different species.
This is the first human study using saliva samples for drug therapeutic monitoring and patient compliance for a complex botanical extract (e.g., ATB). This significantly improves the study robustness and quality of long-term chemoprevention clinical trials by monitoring patient adherence to a study protocol, which should positively impact the quality and reliability of the collected data [
24,
25]. Additional benefit of using Matr concentration in saliva for oral cancer trials comes from its anti-inflammatory and anti-proliferative activity in the oral cavity [
26]. Most importantly this study has provided the needed exposure information at the site of action that allows for the complete monitoring of drug PK profile using saliva samples for the time between patient visits during the window of opportunity pharmacokinetic (NCT03459729) and efficacy (NCT04278989) clinical trials of ATBs. Lastly, saliva samples represent a non-invasive, patient-friendly self-sampling method, which may reduce cost and improve patient retention in the study.
The saliva and plasma concentration profiles of Matr were successfully co-modeled using Model C (
Figure 5 and
Figure 6) which included the active secretion of Matr into saliva, presumably via an OCT transporter expressed in salivary glands [
27,
28]. The model structure includes plasma, saliva, and tissue compartments. Matr was rapidly absorbed (K
a = 1.66 ± 0.27 h
−1) from the gastrointestinal tract, whereas comparatively slower absorption was estimated from the oral cavity (K
as = 0.39 ± 0.14 h
−1). However, later is expected to have minor contribution to overall drug absorption due to tablet dosage form in the current PK study.
After absorption, Matr mostly stayed in plasma as evident from a significantly large volume of distribution of the central compartment (V
p = 1.42 ± 0.08 L/kg) as compared to saliva (V
s = 0.02 ± 0.01 L/kg) and tissue (V
t = 0.62 ± 0.04 L/kg), indicating significant protein binding of Matr. The estimated vs. of Matr in saliva was reasonable based on the normal daily production of saliva (0.5–1.5 L) in humans [
29]. This also explains the higher drug concentrations of Matr in saliva samples.
The model estimated a rapid secretion of Matr from tissue to saliva (Kts = 23.42 ± 2.33 h−1), whereas uptake of Matr from plasma into the tissues (Kpt = 0.11 ± 0.02 h−1) was the rate limiting step. In a pilot study performed in our lab in F344 rats, intravenous infusion of Matr resulted in the salivary excretion of Matr within 15 min. Model-estimated faster rate of reabsorption of salivary-secreted Matr from the oral cavity into plasma (Ksp = 1.25 ± 0.28 h−1) was comparable to its absorption from the gut.
The estimated high K
m values (279.52 ± 5.73 mg/L equal to 1125.5 µM) of Matr transport from plasma to saliva via an OCT transporter were comparable to K
m values of metformin (saliva/plasma concentration ratio = 0.29–0.39) uptake by OCT1, OCT2 (majorly expressed in kidney) and OCT3 (majorly expressed in saliva) in the range of 285–3170 µM [
30]. However, K
m values of Matr need to be verified further in OCT-expressed cells.
The involvement of active transporter was evident from a significantly high ratio of saliva to plasma concentration at each time point for all study participants (
Supplementary Material Figure S1). The saliva-plasma level correlation allowed for the estimation of population PK parameters and provide useful information that can be integrated with a sparse blood sampling approach to provide more detailed and complete population pharmacokinetic profiles. The saliva concentration profile can be used as a surrogate for oral cavity exposure in the PK/PD modeling during the clinical trials and can help inform dose adjustment decisions, which is often done in chemotherapeutic settings [
31,
32,
33,
34].
The PBPK model of Matr successfully explains the distribution and clearance of Matr after oral and i.v. administration of ATB in different species with a good simulation of observed data (
Figure 8,
Table 7), and closely predicting the PK parameters in humans based on the rodent’s data. The rodent PK and metabolic profiling of other KACs suggest high pre-systemic clearance in humans, which is consistent with low KACs exposure in blood and saliva after oral administration of ATB tablets. The salivary drug concentration profile was not included in the PBPK model because most of the excreted drug will be reabsorbed in the GI tract and the involvement of transporter that is needed more investigation. Overall, the PBPK model of Matr can be successfully applied to build a predictive human PK profile from rodent PK data, which can provide insights into calculating the required dose of ATB to achieve the desired exposure in the oral cavity. Through oral administration, matrine exhibits dual PK profiles which were further analyzed through a population model and a PBPK model, and the compound tracer (e.g., Matr) could serve as a competitive biomarker that enables further development of the salivary secretion-based PK/PD analysis.
Several limitations with the current study exist that should be acknowledged and kept in mind while interpreting its results. First, we missed a few data points at the initial time points due to technical difficulty in blood sampling. However, no significant impact on the quality of the model due to missing data points suggests that we can successfully use sparse sampling for PK parameter estimation in the clinical trial design. Second, the total sampling duration was not ideal (ideal last time point ~ 5 × half-life) for estimating the half-life of Matr, because Matr apparent half-life was longer than expected, probably due to the extended elimination phase resulting from probable entero-salivary recycling (
Figure 5, see
Supplement Material Section S5 for more discussion). Therefore, in future clinical trials, saliva samples will be collected daily for 14 days which will allow us to use co-modeling approach to determine the PK parameters and more accurately estimate the half-life of KACs. Third, urinary excretion is the main disposition pathway reported for Matr in humans [
11] with possible involvement of renal clearance but the study design did not include the collection of urine samples, which would have provided the actual value of renal excretion rates of Matr in PBPK modeling. Additionally, the OCT transporter is proposed to involve in the renal clearance and the salivary excretion of Matr [
28,
35], which is needed to improve the physiological relevance of the models and will be investigated in future studies. Lastly, though the sample size (
n = 8) for a single-dose PK study in healthy humans was optimum, the population pharmacokinetic estimation can further improve with a more representative US demographic.