Methods for improving the accuracy of predictions of human plasma concentrations using pharmacokinetic prediction software

Abstract

To provide a method for improving prediction accuracy of concentration in human plasma of an orally ingested compound by combining a parameter obtained by an in vitro testing and in vivo kinetics prediction software.SOLUTION: There is provided a method for predicting a concentration in human plasma of an orally ingested compound by using in vivo kinetics prediction software, in which a predicted value of the software is corrected by a parameter obtained by a cell testing. It is preferable that the cell testing be a membrane permeability testing using an established human intestinal tract epithelial cell. Further, it is preferable that the established human intestinal tract epithelial cell be Caco-2 cell derived from human colon adenocarcinoma.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a method for improving the accuracy of predicting the human plasma concentration of an orally ingested compound by combining pharmacokinetic prediction software with in vitro testing. [Background technology]

[0002] Generally, the biological effects of a compound show a concentration-dependent reaction, so when considering the physiological effects and toxicity of various compounds, including compounds contained in food, the in vivo concentration after oral ingestion, especially the plasma concentration, is important (Non-Patent Document 1). After being absorbed through the digestive tract, a compound is distributed throughout the body via the circulatory bloodstream, metabolized in the liver, etc., and then excreted from the body. In order to analyze the plasma concentration, it is important to understand the pharmacokinetics.

[0003] Animal testing is excellent for examining the biological effects of compounds based on their pharmacokinetics, but there are problems with species differences with humans and the cost of testing (Non-Patent Document 2). In recent years, the idea of ​​adhering to the 3R principles for conducting animal testing (Reduction: reducing the number of experimental animals, Refinement: reducing the pain of experimental animals, and Replacement: replacing with alternative methods that do not use animals) has spread internationally, and it is now necessary to take ethical issues into consideration (Non-Patent Document 3).

[0004] Ultimately, the biological effects of compounds that are actually ingested cannot be clarified without conducting clinical trials and epidemiological surveys in humans, but it is not easy to verify efficacy and safety, including plasma concentrations, through human clinical trials.

[0005] Under these circumstances, the development of prediction software for obtaining pharmacokinetic information of compounds, mainly focusing on pharmaceuticals, is underway (Non-Patent Document 4). Although such methods can quickly and easily predict the plasma concentration of a compound when it is orally ingested based on the chemical structure of the compound, there are a wide variety of factors that affect the change in plasma concentration in the living body (for example, in the case of gastrointestinal absorption, not only transcellular pathways dependent on lipid solubility and pKa, but also permeation of compounds through paracellular pathways, compound-specific uptake and excretion by transporters, intracellular metabolism, etc.), and pharmacokinetic prediction software alone cannot provide high reliability, and further research is required (Non-Patent Document 4).

[0006] As a method for improving the prediction accuracy of human plasma concentration by pharmacokinetic prediction software, a method of combining with measured values ​​obtained from pharmacokinetic studies using model animals has been reported (Patent Document 1). Although this method has been shown to be capable of accurately predicting the transition of human plasma concentration, it requires pharmacokinetic studies using model animals, and issues remain in terms of test period, cost, and ethics.

[0007] In order to resolve issues related to animal testing and human clinical trials, such as ethics, costs, and test duration, there is a strong demand for technology that can predict with high accuracy the pharmacokinetics of orally ingested compounds, in particular the human plasma concentration, without relying on model animals or human subjects. [Prior art documents] [Patent documents]

[0008] [Patent Document 1] Patent application 2017-93262 [Non-patent literature]

[0009] [Non-Patent Document 1] Bull. Natl. Inst. Health Sci., 115, 1-14(1997) [Non-Patent Document 2] Journal of the Japanese Society of Nutrition and Food Science Vol. 59 No. 5 277-282 (2006) [Non-Patent Document 3] Japanese Pharmacological Journal 151, 48-51 (2018) [Non-Patent Document 4] Survey of the international situation regarding the development and practical application of new risk assessment methods for food, including in silico assessment methods (FY2015 Comprehensive Survey on Food Safety) Summary of the Invention [Problem to be solved by the invention]

[0010] The main objective of the present invention is to provide a method for improving the accuracy of predicting the human plasma concentration of an orally ingested compound by combining parameters obtained from in vitro tests with pharmacokinetic prediction software. [Means for solving the problem]

[0011] As a result of extensive research conducted in light of the above-mentioned object, the present inventors discovered that by combining the results of membrane permeability tests using Caco-2 cells with pharmacokinetic prediction software, the prediction accuracy of human plasma concentrations is significantly improved compared to predictions using the pharmacokinetic prediction software alone, and thus completed the present invention.

[0012] That is, the present invention includes the following inventions: Item 1: A method for predicting the human plasma concentration of an orally ingested compound using pharmacokinetic prediction software, in which the software's predicted value is corrected using parameters obtained from a cell test. Item 2. The method according to item 1, wherein the cell test is a membrane permeability test using a human intestinal epithelial cell line. Item 3. The method according to item 1 or 2, wherein the parameter obtained in the cell test is an absorption rate constant ka. Item 4. The method according to any one of Items 1 to 3, wherein the established human intestinal epithelial cell line is Caco-2 cell derived from human colon adenocarcinoma. Item 5. The method according to any one of Items 1 to 4, wherein the pharmacokinetic prediction software is ACD / Percepta. Item 6. The method according to any one of Items 1 to 5, characterized in that the human plasma concentration of a "group of compounds having structural features of a flavonoid skeleton and its glycosides" is predicted with high accuracy using the pharmacokinetic prediction software. Effect of the Invention

[0013] According to the present invention, the accuracy of prediction of human plasma concentration by pharmacokinetic prediction software can be improved by using parameters obtained in a membrane permeability test using Caco-2 cells.

[0014] Furthermore, the present invention makes it possible to rapidly and simply predict human plasma concentrations without relying on model animals or human subjects. [Brief description of the drawings]

[0015] [Figure 1] 1 is a graph showing the ratio of the actual measured human plasma concentration of a compound to the predicted values ​​by each alternative method test. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0016] The present invention will be described in detail below.

[0017] The method for predicting the human plasma concentration of a compound using an alternative to animal testing of the present invention combines software for predicting the physicochemical properties and pharmacokinetics of a compound from its chemical structure with parameters obtained from alternative testing methods.

[0018] The above-mentioned compound pharmacokinetic prediction software is software that predicts the physicochemical parameters of a compound, such as lipid solubility (log P) and acid dissociation constant (pKa), based on a quantitative structure-activity relationship model constructed based on information on known compounds, predicts pharmacokinetic parameters related to absorption, distribution, metabolism, and excretion, such as absorption rate constant (ka), elimination rate constant (ke), solubility in the gastrointestinal tract, distribution volume (Vd), and hepatic first-pass effect, from these predicted physicochemical parameters, and finally calculates the human plasma concentration from these predicted pharmacokinetic parameters and the dose using a mathematical model, such as a multi-compartment absorption model. Although not particularly limited, ACD / Percepta can be used most preferably.

[0019] Examples of parameters obtained in alternative method tests include an absorption rate constant (ka) calculated based on a membrane permeability coefficient (Papp) obtained in a membrane permeability test, an elimination rate constant (ke) calculated from a metabolic stability test, and a distribution volume (Vd) calculated based on a plasma protein binding rate obtained in a plasma protein binding test. Although there is no particular limitation, in the present invention, the absorption rate constant (ka) calculated from a membrane permeability test using an established human intestinal epithelial cell line can be most preferably used.

[0020] Examples of the above-mentioned human intestinal epithelial cell line include human colon cancer-derived cell lines Caco-2, HT-29, T84, etc., and are not particularly limited, with Caco-2 being the most suitable.

[0021] There are no particular limitations on the compounds whose human plasma concentrations can be predicted with high accuracy, but the present invention is most suitable for use with the human plasma concentrations of "compounds having structural characteristics of a flavonoid skeleton and its glycosides." EXAMPLES

[0022] Next, examples of the present invention will be described in detail, but the present invention is not limited thereto.

[0023] ACD / Percepta (Advanced Chemistry Development) was used as software for predicting the pharmacokinetics of a compound from its molecular structure.

[0024] The predicted human plasma concentration of a compound calculated by combining the above software with an alternative test to animal testing was compared with the actual measured human plasma concentration of the compound reported in the literature.

[0025] [Example 1] Combined membrane permeability test using Caco-2 cells <Method> Caco-2 cells were seeded in a 24-well transwell and maintained for 18 days on a CacoReady cell plate (ReadyCell). After 4 hours of incubation in a CO2 incubator at 37°C, the medium was replaced with D-MEM medium containing 10% fetal bovine serum (FBS) and 1% non-essential amino acids, and the cells were cultured for 60 to 72 hours. The volume of the liquid in the culture vessel was 0.3 mL on the apical side and 0.9 mL on the basal side.

[0026] After culturing for 60 to 72 hours, the resistance was confirmed to be 1000 Ω·cm2 using a sterilized Millicell ERS-2 (Millipore). If the resistance was low, the medium was replaced again and the cells were cultured for 48 hours.

[0027] After checking the resistance, the medium was replaced with 0.25 mL / well of HBSS(+), 4.5 g / L Glucose, 10 mM MES, pH 6.5 on the apical side and 0.75 mL / well of HBSS(+), 4.5 g / L Glucose, 10 mM HEPES, pH 7.4 on the basal side, and the apical inserts were placed on the basal compartment and left to stand for 1 minute. This operation was performed three times in total, and the third time was left to stand for 30 minutes in a CO2 incubator at 37°C.

[0028] After leaving the plate for 30 minutes, the medium was removed, and the test substance dissolved in DMSO was diluted with the apical medium to a final concentration of 10 μmol / L and a DMSO concentration of 0.05-0.1%, and added to the apical inserts at 0.3 mL / well. Immediately after addition, 0.05 mL / well of the solution was collected from the apical inserts and used as the 0 hr. apical sample. 0.75 mL / well of basal medium was added to the basal side. The plate was cultured in a CO2 incubator at 37°C for 2 hours. Triplicate experiments were performed for each test substance.

[0029] Eight compounds were used as test substances: Daidzin, Daidzein, Genistin, Genistein, Glycitin, Hesperidin, Naringin, and Rutin.

[0030] After 2 hours of incubation, the culture medium in the apical inserts and basal compartments was collected and used as a 2-hr apical sample and a 2-hr basal sample. The 0-hr apical sample, the 2-hr apical sample, and the 2-hr basal sample were appropriately diluted with a 50% acetonitrile, 0.1% formic acid solution and subjected to quantification by liquid chromatography-tandem mass spectrometry (LC-MS / MS).

[0031] For the LC-MS / MS quantification, an ACQUITY UPLC I-Class (Waters) was used as the liquid chromatography device, and a QTRAP6500 (SCIEX) was used as the mass spectrometer. The measurement conditions are shown in Tables 1 and 2. Of the eight compounds listed in Table 2, the compounds marked with *1, *2, and *3 were measured by the same method.

[0032] [Table 1]

[0033] [Table 2]

[0034] After collecting the culture medium, Lucifer yellow dissolved in DMSO was diluted with the apical medium to a final concentration of 100 μmol / L and DMSO concentration of 0.3% to confirm the maintenance of tight junction formation. 0.25 mL / well of the diluted solution was added to the apical inserts. 0.75 mL / well of the basal medium was added to the basal inserts. The cells were cultured in a CO2 incubator at 37°C for 1 hour.

[0035] After 1 hour of incubation, 0.2 mL of each culture medium was collected from the apical inserts and basal compartments, and measured with serially diluted Lucifer yellow standard solutions using a fluorescence plate reader with an excitation wavelength of 485 nm and an emission wavelength of 535 nm filter. The fluorescence intensity in the basal compartment was confirmed to be 0.7% or less of that in the apical inserts.

[0036] From the quantitative values ​​of each test substance, the apparent membrane permeability coefficient Papp was calculated, and the absorption rate constant ka was calculated according to Non-Patent Documents 5, 6, and 7.

[0037] [Comparative Example 1] Combined with a membrane permeability test using an artificial lipid bilayer membrane <Method> A PAMPA plate (Corning) was left at room temperature for 30 minutes or more to thaw, and the test substance dissolved in DMSO was diluted with PBS (pH 6.5) to a final concentration of 50 μmol / L and a DMSO concentration of 0.25-0.5%, and 0.3 mL was added to the receiver plate (donor side). 0.2 mL of PBS (pH 7.4) was added to the filter plate (acceptor side). The plate was left to stand at room temperature for 5 hours. Triplicate experiments were performed for each test substance.

[0038] After standing for 5 hours, the solutions in the receiver plate and the filter plate were each collected, appropriately diluted with a solution of 50% acetonitrile, 0.1% formic acid, and subjected to quantification by LC-MS / MS.

[0039] The types of test substances and their quantification by LC-MS / MS were determined in the same manner as in Example 1.

[0040] From the quantitative values ​​of each test substance, the apparent membrane permeability coefficient Papp was calculated, and the absorption rate constant ka was calculated according to Non-Patent Documents 5, 6, and 7.

[0041] (Non-Patent Document 5) Construction of a prediction model for drug absorption and metabolism in the intestinal tract (Nagoya City University dissertation, FY2016) (Non-patent document 6) Pharm Res., 12(3), 413-420, (1995) (Non-patent document 7) Int J Pharm., 186(2), 119-125(1999)

[0042] [Comparative Example 2] When combined with a metabolic stability test using HepaRG cells <Method> HepaRG cells (Biopredic) were suspended in HepaRG Thawing / Plating / General Purpose Medium at 7.2 × 10 5 Prepare 0.1 mL / well (7.2×10 4 The cells were seeded at 1000 cells / well onto a Collagen I-coated MW96 plate (Biopredic) and cultured in a CO2 incubator at 37°C for 24 hours.

[0043] After 24 hours of culture, the medium was replaced with HepaRG Maintenance and Metabolism Medium and cultured for an additional 60 to 72 hours.

[0044] After culturing for 60 to 72 hours, the medium was replaced with serum-free induction medium and cultured for an additional 24 hours.

[0045] After 24 hours of culture, the medium was removed, and the test substance dissolved in DMSO was diluted with 50% acetonitrile solution, and then diluted with Serum-free Induction Medium to a final concentration of 1 μmol / L and added. The amount of medium in the culture vessel was 0.1 mL, the DMSO concentration was 0.001%, and the acetonitrile concentration was 0.5%. In order to measure the test substance concentration in the culture solution 0, 0.5, 1, 1.5, 3, 18, and 24 hours after addition, the test substance was added at 1 time point / well in parallel at two points.

[0046] After each incubation period, 50 μL of the culture supernatant was collected, 150 μL of cold acetonitrile + 0.8% formic acid solution was added, and the mixture was centrifuged at 5,000 g for 20 minutes at 4°C. The supernatant was diluted with an equal amount of water and subjected to quantification by LC-MS / MS.

[0047] The types of test substances and their quantification by LC-MS / MS were determined in the same manner as in Example 1.

[0048] The elimination rate constant ke was calculated from the quantitative values ​​of each test substance at each incubation time according to Non-Patent Documents 8 and 9.

[0049] (Non-patent document 8) Drug Metab Dispos., 44(4), 527-533(2016) (Non-patent document 9) Pharm Res., 27(10), 2150-2161(2010)

[0050] [Comparative Example 3] Combined with plasma protein binding test <Method> The test substance dissolved in DMSO was diluted with plasma to a final concentration of 10 μmol / L and added to the plasma chamber of a Rapid Equilibrium Dialysis (RED) Device Single-Use Plate with Inserts (Cut-Off or Exclusion Limit: 8.0 kDa) (Thermo Fisher Scientific). The total volume of the liquid in the chamber was 0.3 mL, and the DMSO concentration was 0.1%.

[0051] 0.55 mL of dialysis buffer (100 mM PBS, pH 7.2, 150 mM NaCl) was added to the buffer chamber and shaken at 250 rpm at 37° C. for 4 hours. Triplicate experiments were performed for each test substance.

[0052] After shaking for 4 hours, 0.05 mL was collected from each chamber, 0.05 mL of dialysis buffer was added to the solution collected from the plasma chamber, and 0.05 mL of plasma was added to the solution collected from the buffer chamber. 0.3 mL of acetonitrile + 0.1% formic acid solution was added and mixed, and the mixture was left to stand on ice for 30 minutes.

[0053] After standing for 30 minutes, the mixture was centrifuged at 15,000 g for 10 minutes at 4° C. The supernatant was diluted with an equal volume of 0.1% formic acid solution and subjected to quantification by LC-MS / MS.

[0054] The types of test substances and their quantification by LC-MS / MS were determined in the same manner as in Example 1.

[0055] The plasma protein binding rate was calculated from the quantitative value of each test substance, and the distribution volume Vd was calculated according to Non-Patent Document 10. [Non-Patent Document 10] Drug Metab Dispos., 38(7), 1159-1165(2010)

[0056] <Result> Figure 1 shows the ratio of the actual values ​​of human plasma concentrations to the predicted values ​​by each alternative method test (hereinafter, actual value / prediction value). When ACD / Percepta was combined with the membrane permeability test results using Caco-2 cells, the actual value / prediction value improved from 0.95-249.2 when predicting using only ACD / Percepta to 0.93-9.1. On the other hand, when the membrane permeability test results using an artificial lipid bilayer membrane were combined, the actual value / prediction value was 1.4-220.0. In addition, when the metabolic stability test results using HepaRG cells and the plasma protein binding test results were combined, the actual value / prediction value was 1.1-61.1 and 2.1-101.8, respectively. From these results, it can be confirmed that the improvement in the accuracy of prediction of human plasma concentrations when ACD / Percepta is combined with the membrane permeability test results using Caco-2 cells is remarkable. In Figure 1, 1 indicates ACD / Percepta only, and 2 to 5 indicate the actual / predicted values ​​when ACD / Percepta is combined with the results of a membrane permeability test using Caco-2 cells (2), a membrane permeability test using an artificial lipid bilayer membrane (3), a metabolic stability test using HepaRG cells (4), and a plasma protein binding test (5).

Claims

1. A method for predicting the human plasma concentration of an orally ingested compound using software that predicts physicochemical parameters of a compound, such as lipid solubility (log P) and acid dissociation constant (pKa), based on a quantitative structure-activity relationship model constructed based on information on a known compound, and predicts at least the absorption rate constant (ka) and, if necessary, other pharmacokinetic parameters related to absorption, distribution, metabolism, and excretion, such as elimination rate constant (ke), solubility in the gastrointestinal tract, volume of distribution (Vd), and hepatic first-pass effect, from the predicted values ​​of these physicochemical parameters, and finally predicts the human plasma concentration using a mathematical model, such as a multi-compartment absorption model, from the predicted values ​​of the pharmacokinetic parameters and the dose, A method for correcting the absorption rate constant of the software using parameters obtained in a membrane permeability test using Caco-2 cells, which are an established human intestinal epithelial cell line derived from human colon adenocarcinoma.

2. A method for accurately predicting the human plasma concentration of a "group of compounds having structural characteristics of a flavonoid skeleton and its glycosides" using the method for correcting the absorption rate constant of the software described in claim 1 using the software.

Images