Association of Plasma Creatinine with Systemic Exposure to S-1 and Oxaliplatin in Two Types of Chronic Kidney Disease Animal Models

Abstract

Chronic kidney disease (CKD) is a serious chemotherapy-associated clinical condition. CKD complicates the pharmacokinetics of anticancer drugs, requiring personalized dosing strategies to minimize toxicity. S-1 and oxaliplatin (the SOX regimen) are widely used for gastrointestinal cancer treatment. However, the specific association between systemic drug exposure and renal biomarker levels in CKD remains unclear. This study evaluated the pharmacokinetics of S-1 and oxaliplatin in 2 CKD model rats (5/6 nephrectomy and adenine-induced) and examined their relationships with renal biomarkers. S-1 (2 mg/kg as tegafur) and oxaliplatin (5 mg/kg) were administered separately, and plasma levels of tegafur, 5-fluorouracil (5-FU), 5-chloro-2,4-dihydropyridine, oxaliplatin, and platinum were measured by LC-tandem MS. Systemic exposure to S-1 and oxaliplatin was higher in the CKD model rats than in the normal group, with 5-FU levels being particularly higher in the adenine-induced model than in the 5/6 nephrectomy model. The area under the curve values of 5-FU and platinum were strongly correlated with plasma creatinine (P_Cr) levels (r = 0.79 and 0.88, respectively). Population pharmacokinetic analysis identified P_Cr as a significant covariate of 5-FU clearance. A nomogram constructed using P_Cr-based simulations demonstrated the feasibility of individualized S-1 dosing. Overall, our findings suggest that P_Cr is a practical biomarker to guide S-1 dose optimization and highlight the importance of pharmacokinetics-based strategies for patients with cancer and CKD.

INTRODUCTION

Patients with cancer and chronic kidney disease (CKD) exhibit elevated all-cause and cancer-specific mortality rates.¹⁾ Appropriate dose adjustments based on renal function are crucial to effectively treat these patients while minimizing the risk of toxicity. However, the pharmacokinetics of patients with cancer and CKD, particularly drug metabolism and clearance, are complicated owing to multiple factors such as reduced liver enzyme activity and altered expression of renal drug transporters.^2,3) These pathophysiological changes contribute to unpredictable drug disposition and make dose optimization challenging. The SOX regimen, combining S-1 and oxaliplatin, is widely used as a 1st-line treatment for colorectal cancer in East Asia^4,5) due to the increasing number of outpatients receiving chemotherapy and the aging cancer population. Therefore, precise SOX dose adjustments are necessary for patients with cancer and CKD.

Recently, clinical studies have explored dose adjustments of the SOX regimen for patients with cancer and CKD.^5,6) Adjusting S-1 dosing based on renal function reduces 5-fluorouracil (5-FU) toxicity, underscoring the need for individualized dosing in patients with CKD.⁵⁾ The currently used approach, which determines S-1 dosage based solely on creatinine clearance (C_Cr) and the patient’s body surface area, is possibly insufficient as it does not fully account for the inter-individual variability in pharmacokinetics, resulting in inadequate drug exposure and compromised therapeutic outcomes in patients with cancer and CKD. Watanabe et al. examined the relationship between C_Cr and oxaliplatin toxicity in patients with metastatic colorectal cancer and reported no significant correlation with adverse events, suggesting that a reduction in the oxaliplatin dose based on C_Cr is not recommended for these patients.⁶⁾ However, a previous case report showed a patient with colorectal cancer and mild preexisting CKD developing severe acute tubular necrosis requiring dialysis following oxaliplatin administration.⁷⁾ The necessity for oxaliplatin dose modification in patients with CKD remains controversial.

Determination of the relationship between anticancer drug exposure and renal biomarker levels will facilitate the development of predictive, biomarker-guided treatment strategies for patients with cancer and CKD. However, direct clinical investigation remains challenging due to the inter-patient variability in CKD severity and confounding factors, such as comorbidities and concomitant medications.⁸⁾ Additionally, large sample sizes are often needed to account for heterogeneity in clinical backgrounds, complicating statistical analysis.⁹⁾ Animal studies facilitate controlled and practical evaluation of pharmacokinetics and biomarker relationships.

Various preclinical CKD models reflecting different renal pathophysiological aspects have been developed.¹⁰⁾ Among them, the 5/6 nephrectomy and adenine-induced CKD models are commonly used as foundational models.¹⁰⁾ However, the most appropriate rat model to study dose adjustments in patients with cancer and CKD remains unclear. Investigation of both S-1 and oxaliplatin in these distinct CKD models will facilitate comparative analysis of the mechanisms by which CKD alters the SOX pharmacokinetics, providing valuable insights into the mechanisms by which renal dysfunction affects drug exposure and its correlations with renal biomarker levels. Therefore, this study aimed to elucidate the relationships between S-1 and oxaliplatin pharmacokinetics and renal biomarker levels in different CKD animal models.

MATERIALS AND METHODS

Materials

Tegafur and potassium oxonate were obtained from the Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). Additionally, 5-chloro-2,4-dihydropyridine (CDHP) was obtained from Combi-Blocks Co., Ltd. (San Diego, CA, U.S.A.). Oxaliplatin was obtained from Wako Pure Chemical Corporation (Osaka, Japan). Elplat (5 mg/mL oxaliplatin) was supplied by Yakult Honshu Co., Ltd. (Tokyo, Japan). All reagents were of analytical grade.

Preparation of CKD Model Rats

Ten-week-old male Wistar rats (n = 43) purchased from Nippon SLC Co., Ltd. (Hamamatsu, Japan) were housed in a temperature-controlled room under a 12/12-h light/dark cycle and provided free access to standard rat chow and water. Two CKD models were established: the 5/6 nephrectomy CKD model, in which renal impairment was rapidly induced by removing two-thirds of the renal mass, and the adenine-induced CKD model, in which progressive renal damage was caused by crystalline deposits in the renal tubules; the adenine-induced CKD model exhibited a lower mortality rate than the 5/6 nephrectomy CKD model.¹¹⁾ To compare the pharmacokinetics of different CKD models, these 2 models were selected to account for variations in disease progression, underlying pathophysiologies, and potential impacts on pharmacokinetics.^12,13) To prepare the 5/6 nephrectomy CKD model rats,¹¹⁾ an incision was made in the skin and muscles behind the final ribs to expose the kidneys. One-third of the upper and lower parts of each kidney was removed. Hemostasis was achieved, and the kidneys were subsequently repositioned into the abdominal cavity. The abdominal muscles and the skin were closed with sutures. One week after left partial nephrectomy, a total right nephrectomy was performed following the same procedure, with an additional step of fully ligating the hilum of the right kidney.

Adenine-induced model rats were established by administering an adenine solution as previously described,^12–14) with minor modifications. Briefly, adenine was dissolved in a 0.5% methylcellulose solution and orally administered at 400 mg/kg for five days. The normal group did not undergo 5/6 nephrectomy or adenine treatment. All experimental procedures were approved by the Institutional Review Board and were performed according to the Kyoto Pharmaceutical University (Kyoto, Japan) Guidelines for Animal Experimentation (Approval Number: A22-53).

Biochemical Parameter Analysis

Under intraperitoneal anesthesia using 0.375 mg/kg medetomidine, 2.0 mg/kg midazolam, and 2.5 mg/kg butorphanol, blood samples were collected from the left jugular vein, and urine samples were collected from the bladder through a catheter before the pharmacokinetic experiment to measure the biochemical parameters. To assess biochemical markers such as blood urea nitrogen (BUN), aspartate aminotransferase (AST), and alanine aminotransferase (ALT) levels, blood samples were collected from a separate cohort of animals (n = 4 per group), independent of those used in the pharmacokinetic study. This was necessary because the volume of blood required for these analyses, combined with serial sampling in the pharmacokinetic study, could have affected the accuracy of both the biochemical and pharmacokinetic data. In contrast, P_Cr and C_Cr were measured using plasma and urine samples collected from the same animals used in the pharmacokinetic analysis. Urine samples were collected using a bladder catheter to evaluate the urine creatinine (U_Cr) levels. Blood and U_Cr levels were determined using LabAssay Creatinine (Wako Pure Chemical Corporation). Then, renal function of the rats was confirmed using the blood and urine samples to determine the C_Cr values as follows:

  

$$C_{Cr}=(UF\times U_{Cr})/(P_{Cr}\times 1000)$$(1)

where C_Cr is creatinine clearance (mL/min/kg), UF is the urine flow rate (µL/min/kg), U_Cr is the urine concentration of creatinine (mg/dL), and P_Cr is the plasma concentration of creatinine (mg/dL). BUN, AST, and ALT levels were measured by Kyoto Biken Laboratories Inc. (Kyoto, Japan). BUN levels were determined using the urease–glutamate dehydrogenase method. AST and ALT levels were measured using the transferable method described by the Japan Society of Clinical Chemistry.

Pharmacokinetic Study of CKD Model Rats

Rats were divided into 3 groups: normal (n = 9), 5/6 nephrectomy CKD (n = 12), and adenine-induced CKD groups (n = 10). To avoid the influence of drug–drug interactions and to allow for independent pharmacokinetic evaluation, S-1 and oxaliplatin were administered to different individual animals within each experimental group. S-1 was prepared by mixing 4.54 mg at a molecular ratio (tegafur : CDHP : potassium oxonate = 1 : 0.4 : 1) and dissolving it in 2 mL of carboxymethyl cellulose. S-1 consists of tegafur, a prodrug of 5-FU; CDHP, a renally excreted inhibitor of dihydropyrimidine dehydrogenase (DPD), the rate-limiting enzyme in 5-FU catabolism; and potassium oxonate, which reduces gastrointestinal toxicity. S-1 (2 mg/kg tegafur) was orally administered. Oxaliplatin (5 mg/mL; Elplat) was intravenously administered via the jugular vein at 5 mg/kg. Oxaliplatin is a platinum-based chemotherapeutic agent that is eliminated by renal excretion. The dosages of both drugs were determined based on the clinical dosing guidelines and previous animal studies.¹⁵⁾ Blood samples (0.3 mL) were collected from the jugular vein 48 h after S-1 administration and 2 h after oxaliplatin administration. The sampling times were determined according to our previous study.¹⁵⁾ Plasma concentrations of tegafur, 5-FU, and CDHP from S-1 and intact oxaliplatin and platinum from oxaliplatin were determined via LC-tandem MS using the API 3200 triple-quadrupole mass spectrometer (Sciex, Framingham, MA, U.S.A.), following our previously reported method.¹⁵⁾ Oxaliplatin exerts anti-tumor effects by forming cross-links between platinum and DNA.¹⁶⁾ Plasma platinum and intact oxaliplatin concentrations were measured using our previously described method.^15,17) The precision and accuracy of the analysis were within acceptable limits (<15%), with a lower limit of quantification <0.01 µg/mL.

Pharmacokinetic Analysis

Non-compartmental Analysis (NCA)

An NCA was performed using the plasma S-1, intact oxaliplatin, and platinum concentration data with the NCA program of Phoenix WinNonlin software (version 8.2, Certara U.S.A. Inc., Princeton, NJ, U.S.A.). The area under the plasma concentration–time curve from the time of dosing to the last time (AUC_last) and to infinity (AUC_0–∞) were determined using the linear trapezoidal rule. Owing to the difficulties in accurately assessing the terminal elimination phase in some CKD models, AUC_last was used to evaluate the S-1 components. Half-life (t_1/2) was calculated using the terminal slope (k_e), which was determined by the linear regression of the terminal phase of the plasma concentration–time curve. Total clearance (CL_tot) was calculated by dividing the dose by AUC_0–∞, and the volume of distribution (V_d) was determined by dividing CL_tot by k_e. Additionally, correlations between the obtained pharmacokinetic parameters, specifically AUC_last, and renal biomarker levels were investigated. These parameters were selected to accurately reflect the total drug exposure, which is essential for examining the relationships with biomarker levels. Power regression analysis was performed to further investigate the quantitative relationship between drug exposure and renal biomarker levels.

Population Pharmacokinetic Analysis

The influence of P_Cr and C_Cr levels on the exposure to 5-FU, CDHP, intact oxaliplatin, and platinum was evaluated. These routinely used clinical biomarkers were incorporated as covariates in the population pharmacokinetic analysis to assess their predictive value for drug exposure. In population pharmacokinetic analysis, plasma concentrations of S-1, intact oxaliplatin, and platinum in the normal and CKD groups were analyzed using the nonlinear mixed-effects modeling software, Phoenix NLME version 8.2 (Certara U.S.A. Inc.). The schematic of the pharmacokinetic model is shown in Supplementary Fig. S1. A population pharmacokinetic model was established to fit the data by estimating the typical structural model parameters for S-1, intact oxaliplatin, and platinum, as well as the inter-individual and residual variabilities. The inter-individual and residual variability models were determined according to the 2× log-likelihood (−2LL) values, goodness-of-fit (GOF) plots, and parameter estimate coefficients of variation (CVs). Error models for individual and residual variability were selected based on Akaike’s Information Criteria, GOF plots, and CV% estimates for the parameters. Initially, a base model was constructed for each drug, followed by the exploration of covariates. The first-order conditional estimation-extended least-squares method was used to examine the population parameters and their variabilities. The pharmacokinetic model for S-1 was adapted from a previously reported model.¹⁸⁾ Two conversion rate constants for the metabolism of tegafur to 5-FU were incorporated into the base model of tegafur. The k_i value, the inhibitory constant of CDHP for DPD, was obtained from a previously published report.¹⁸⁾ Importantly, since 5-FU is a metabolite of tegafur and its volume of distribution (V_FU) is not directly identifiable, V_FU was fixed at 0.7, as established in a prior population pharmacokinetic analysis.¹⁹⁾ Other parameters were fixed when they could not be reliably estimated from the current dataset, particularly when doing so improved model convergence and avoided overparameterization.¹⁸⁾ For oxaliplatin, a 2-compartment model with linear elimination was selected for intact oxaliplatin and platinum as the base model based on our previous study.²⁰⁾

Covariate Analysis

After constructing the base model, covariates were tested for statistical significance, and the possible relationships between individual estimates of the pharmacokinetic parameters and covariates were explored. A stepwise forward inclusion (p < 0.05) and backward elimination (p < 0.001) process was used for covariate searching. The covariates in the analysis included the continuous variables of P_Cr and C_Cr. For covariates, the following equation was used for modeling:

  

$${\theta }_i={\theta }_{TV}\cdot CO{V_i}^x$$(2)

where 𝜃_TV represents the typical population value for the pharmacokinetic parameter, θ_i denotes the individual parameter value, COV_i indicates the covariate value for the ith individual, and x is the coefficient that describes the covariate influence on the parameter. Graphical and statistical analyses were conducted to validate the final model, including GOF plots comparing the observed and population-predicted concentrations and conditional weighted residuals. Additionally, prediction-corrected visual predictive checks (pcVPC; n = 1000) and nonparametric bootstrap (n = 1000) were performed to evaluate the model predictability.

Model-Based Simulation of Plasma 5-FU Concentration

As P_Cr was a covariate of 5-FU clearance, alterations in the AUC values of 5-FU at different P_Cr levels were simulated based on the final model. The simulations were performed using P_Cr levels of 0.3–4.0 mg/dL. This P_Cr range was selected based on the distribution observed across the normal, 5/6 nephrectomy, and adenine-induced CKD groups in this study and was slightly extended to represent a broader spectrum from normal renal function to mild and severe CKD. Plasma 5-FU concentrations were simulated across this spectrum of P_Cr levels using the final pharmacokinetic model. The nomogram depicted the relationship between P_Cr levels and the corresponding AUC values of 5-FU. The target AUC was set at 2.5–5.0 µg · h/mL based on the previously reported values and AUC values of the normal group.

Statistical Analysis

Data are represented as the mean ± standard deviation. Multiple groups were compared via one-way ANOVA, followed by the Bonferroni post hoc test. Correlation analyses between pharmacokinetic parameters and renal biomarker levels were also performed to evaluate the drug exposure relationships. Statistical significance was set at p < 0.05.

RESULTS

Biochemical Parameter Comparison in CKD Model Rats

Table 1 presents the biochemical parameters in the normal, 5/6 nephrectomy CKD, and adenine-induced CKD model rats. AST and ALT levels were not significantly different in the CKD model rats. BUN levels were higher in the 5/6 nephrectomy and adenine-induced CKD groups than in the normal group. P_Cr levels were increased in the CKD groups compared to those in the normal group, with a 2.6-fold increase in the 5/6 nephrectomy CKD group and a 5.5-fold increase in the adenine-induced CKD group. Consistently, C_Cr values were significantly reduced in both CKD groups. C_Cr values were decreased by approximately 93 and 95% in the 5/6 nephrectomy and adenine-induced CKD groups, respectively, compared to those in the normal group.

Table 1. Biochemical Parameters in the Normal and 5/6 Nephrectomy and Adenine-Induced Chronic Kidney Disease (CKD) Model Rats

Biochemical parameters	Unit	Normal	CKD
			5/6 nephrectomy	Adenine
AST	U/L	93.8 ± 25.6	66.3 ± 9.2	98.0 ± 31.2
ALT	U/L	39.0 ± 10.7	39.0 ± 15.3	66.8 ± 30.7
BUN	mg/dL	15.7 ± 2.8	54.0 ± 23.5a)	135.2 ± 63.4a,b)
PCr	mg/dL	0.53 ± 0.19	1.4 ± 0.5	2.9 ± 0.9a,b)
CCr	mL/min/kg	13.2 ± 8.8	0.93 ± 0.44a)	0.61 ± 0.31a)

a) p < 0.05, 5/6 nephrectomy and adenine-induced CKD model rats vs. normal rats. b) p < 0.05, adenine-induced CKD model rats vs. 5/6 nephrectomy CKD model rats. Statistical significance was evaluated via one-way ANOVA, followed by the Bonferroni post hoc test. Each value represents the mean ± standard deviation of P_Cr and C_Cr: n = 9 (normal), 12 (5/6 nephrectomy), and 10 (adenine-induced); AST, ALT, and BUN: n = 4 (each group). ALT: alanine aminotransferase; AST: aspartate aminotransferase; BUN, blood urea nitrogen; C_Cr: creatinine clearance; P_Cr: plasma creatinine.

NCA of S-1 and Intact Oxaliplatin in CKD Model Rats

S-1

Plasma concentration–time curves of tegafur, 5-FU, and CDHP after oral administration of S-1 in the normal, 5/6 nephrectomy CKD, and adenine-induced CKD model rats are shown in Figs. 1A–1C. Pharmacokinetic parameters of tegafur, 5-FU, and CDHP after S-1 oral administration are listed in Supplementary Table S1. Maximum plasma concentration (C_max) values of tegafur were significantly lower in the adenine-induced and 5/6 nephrectomy CKD groups than in the normal group. Moreover, a 3.7-fold increase in AUC_last of 5-FU was observed in the adenine-induced CKD group compared to that in the normal group, whereas the AUC_last of 5-FU was unaltered in the 5/6 nephrectomy group. Additionally, CL of CDHP in the adenine-induced group was 0.26 times lower than that in the normal group.

Fig. 1. Mean Plasma Concentration–Time Profiles of S-1 and Oxaliplatin Components in Normal and CKD Model Rats

Mean plasma concentration–time profiles of (A) tegafur, (B) 5-FU, (C) CDHP, (D) intact oxaliplatin, and (E) platinum after the oral administration of S-1 and bolus injection of oxaliplatin (5 mg/kg). Different symbols represent the normal (○), 5/6 nephrectomy CKD (◍), and adenine-induced CKD (●) groups. Results are represented as the mean ± S.D. S-1 dosing group: normal (n = 5), 5/6 nephrectomy CKD (n = 5), and adenine-induced CKD (n = 5). Oxaliplatin dosing group: normal (n = 4), 5/6 nephrectomy CKD (n = 7), and adenine-induced CKD (n = 5). 5-FU: 5-fluorouracil; CDHP: 5-chloro-2, 4-dihydropyridine; CKD: chronic kidney disease.

Oxaliplatin

Plasma concentration–time curves of intact oxaliplatin and platinum after bolus injections of oxaliplatin in the normal, 5/6 nephrectomy CKD, and adenine-induced CKD model rats are shown in Figs. 1D, 1E. Pharmacokinetic parameters of intact oxaliplatin and platinum after a bolus injection of 5 mg/kg oxaliplatin are listed in Supplementary Table S2. The half-life of intact oxaliplatin in the adenine-induced group was extended compared to those in the normal and 5/6 nephrectomy CKD groups. The AUC_0–∞ of intact oxaliplatin was significantly higher in the adenine-induced CKD group (4.2 ± 0.4 µg · h/mL) than in the normal group (2.1 ± 0.7 µg · h/mL). Accordingly, total clearance of intact oxaliplatin was approximately 46% lower than that in the normal group. The half-life of platinum in the adenine-induced CKD group was 2.2 times higher than that in the 5/6 nephrectomy CKD and normal groups. Furthermore, the AUC_0–∞ of platinum was 3.1 times higher than that of intact oxaliplatin in the adenine-induced group.

Correlation between AUC Values and Renal Biomarker Levels

As shown in Fig. 2, the power regression curve indicated the significant correlations between the AUC values of 5-FU, CDHP, intact oxaliplatin, and platinum and P_Cr levels and C_Cr values. The regression model revealed a nonlinear relationship between drug exposure and renal biomarker parameters. AUC values of 5-FU, CDHP, and platinum exponentially increased following the elevation of P_Cr levels, indicating reduced drug clearance with the onset of CKD. A similar trend was observed with decreasing C_Cr values, indicating an inverse relationship between renal function and systemic drug exposure.

Fig. 2. Correlations between AUC Values and Renal Function Markers (P_Cr and C_Cr) in Normal and CKD Model Rats

Correlations between the AUC values and plasma creatinine levels and creatinine clearance for 5-FU, CDHP, oxaliplatin, and platinum after the oral administration of S-1 (2 mg/kg tegafur) and bolus injection of oxaliplatin (5 mg/kg). Different symbols represent the normal (○), 5/6 nephrectomy CKD (◍), and adenine-induced CKD (●) groups. Results are represented as the mean ± S.D. S-1 dosing group: normal (n = 5), 5/6 nephrectomy CKD (n = 5), and adenine-induced CKD (n = 5). Oxaliplatin dosing group: normal (n = 4), 5/6 nephrectomy CKD (n = 7), and adenine-induced CKD (n = 5). AUC: area under the concentration–time curve.

Population Pharmacokinetic Analysis of CKD Model Rats

S-1

Estimated population pharmacokinetic parameters and bootstrap validation results of S-1 are presented in Table 2. CV% estimates for all fixed-effects parameters were in the range of 8.8–17.5%. CV% for the inter-individual variability of each parameter was 7.7–38.3%, indicating good accuracy and reliability.

Table 2. Population Pharmacokinetic Parameters of S-1 in the Normal and 5/6 Nephrectomy and Adenine-Induced Chronic Kidney Disease (CKD) Model Rats Orally Treated with S-1 (2 mg/kg Tegafur)

Parameters of S-1	Unit	Final model		Bootstrap (n = 1000)
		Estimate	CV%	Median	2.5th – 97.5th
Fixed-effects parameters (θ)
Vtegafur	L/kg	0.8	14.9	0.8	0.6–0.95
ka tegafur	1/h	0.7	Fixa)	0.7
ka2	1/h	0.05	10.8	0.05	0.03–0.07
VFU	L/kg	0.7	Fixa)	0.7
km	1/h	0.03	24.5	0.03	0.024–0.034
CLFU = 0.46 × PCr x	L/h/kg	0.46	12.7	0.47	0.34–0.72
x		−1.1	9.3	−1.1	−0.61–−1.7
VCDHP	L/kg	0.3	8.8	0.4	0.2–0.6
ka CDHP	1/h	1.04	Fixa)	1.04
ke CDHP	1/h	0.023	Fixa)	0.023
k12 CDHP	1/h	8.64	17.5	6.52	1.65–12.9
k21 CDHP	1/h	1.35	12.0	1.34	0.78–2.31
ki	μmol/L	1.4	Fixa)	1.4
Inter-individual variability (ω)
Vtegafur	%	28.7	7.7	27.7	27.5–27.8
CLFU	%	41.7	14.9	38.1	37.6–38.5
VCDHP	%	6.7	8.2	6.8	6.84–6.86
k12 CDHP	%	52.4	23.6	60.2	58.1–62.4
k21 CDHP	%	82.5	38.3	96.3	90.5–102.1
Residual variability (σ)
Ctegafur	%	62.7	6.2	62.6	47.8–78.0
CFU	%	20.3	6.6	19.5	13.0–25.3
CCDHP	%	11.7	5.9	11.4	8.7–14.1

a) Fix data were based on previous studies. V_tegafur: distribution volume of tegafur; k_{a tegafur}: absorption rate constant of tegafur; k_a2: conversion rate constant of the first-pass metabolism of tegafur; V_FU: distribution volume in the 5-fluorouracil compartment; k_m: rate constant of the conversion of tegafur to 5-fluorouracil in the systemic circulation; CL_FU: clearance from the 5-fluorouracil compartment; V_CDHP: distribution volume of the gimeracil compartment; k_a CDHP: absorption rate constant of gimeracil; k_{e CDHP}: elimination rate constant of gimeracil; K_{12 CDHP}: transfer rate constant from the central to the peripheral compartment; K_{21 CDHP}: transfer rate constant from the peripheral to the central compartment; k_i: inhibitory constant; C: plasma drug concentration.

Based on the −2LL value and GOF plots, a proportional error model was selected to indicate the plasma concentrations of S-1 components, although additive and multiplicative error models were also considered. After covariate analysis, P_Cr was identified as a significant covariate of CL_FU. The final model of CL_FU was indicated as follows:

  

$$C{L}_{FU}=0.46\times {P_{Cr}}^{-1.1}\left(\text{L}/\text{h}/\text{kg}\right)$$(3)

Each parameter estimate obtained from bootstrap validation was similar to that in the original dataset, indicating that the pharmacokinetic model parameters were adequately estimated. Based on the GOF plots, the final model was deemed appropriate for the data. The pcVPC plots in Figs. 3A–3C show that most of the observed S-1 concentrations, including tegafur, 5-FU, and CDHP concentrations, were included within the 90% confidence intervals from the simulation data, and 5, 50, and 95% quantiles fell within the 95% confidence intervals, indicating good prediction by the final model.

Fig. 3. pcVPC Plots of the Final Pharmacokinetic Models in Normal and CKD Model Rats

pcVPC plots for the final pharmacokinetic models of (A) tegafur, (B) 5-FU, (C) CDHP, (D) intact oxaliplatin, and (E) platinum in the normal (○), 5/6 nephrectomy CKD (◍), and adenine-induced CKD (●) groups. Open circles indicate the individual observations. Solid and dashed lines indicate the 5, 50, and 95th percentiles of the predictions and observations, respectively. The shaded dark gray area indicates the 95% confidence interval for the median. Light gray areas indicate the 95% confidence intervals for the corresponding model-predicted percentiles. pcVPC: prediction-corrected visual predictive check.

Oxaliplatin

A 2-compartment model with linear elimination was used to describe the pharmacokinetics of intact oxaliplatin and platinum in the normal, 5/6 nephrectomy CKD, and adenine-induced CKD model rats. Population pharmacokinetic parameter estimates and bootstrap validation results are shown in Table 3. CV% estimates for all pharmacokinetic parameters were in the range of 7.6–28.4% for oxaliplatin and 14.9–21.5% for platinum, indicating that the model precision was acceptable, except for the inter-individual variability in CL_2oxaliplatin. The proportional error model provided the best description of plasma concentrations of oxaliplatin and platinum. None of the remaining covariates tested were identified as statistically significant sources of inter-individual variability in the pharmacokinetics of intact oxaliplatin and platinum. The median value of each model parameter, estimated using the bootstrap procedure, was similar to that of the original dataset. The pcVPC plots of oxaliplatin and platinum (Figs. 3D, 3E) show that the observed values largely overlapped with the median, 5th, and 95th percentiles of the simulation data. These validation results confirmed that the population pharmacokinetic model was reliable and sufficient for simulating the plasma oxaliplatin and platinum concentrations.

Table 3. Population Pharmacokinetic Parameters of Intact Oxaliplatin and Platinum in the Normal and 5/6 Nephrectomy and Adenine-Induced Chronic Kidney Disease (CKD) Model Rats Treated Following Bolus Injections of Oxaliplatin (5 mg/kg)

Parameters	Unit	Final model		Bootstrap (n = 1000)
		Estimate	CV%	Median	2.5th–97.5th
Intact oxaliplatin
Fixed-effects parameters (θ)
V1 oxaliplatin	L/kg	0.33	8.6	0.33	0.27–0.40
V2 oxaliplatin	L/kg	0.99	7.6	0.98	0.24–2.2
CL1 oxaliplatin	L/h/kg	2.9	10.6	2.9	2.2–3.75
CL2 oxaliplatin	L/h/kg	0.52	28.7	0.54	0.30–0.92
Inter-individual variability (ω)
V1 oxaliplatin	%	27.8	10.3	26.3	25.9–26.6
V2 oxaliplatin	%	21.6	8.1	18.9	18.7–19.0
CL1 oxaliplatin	%	13.9	28.2	13.8	13.6–13.9
CL2 oxaliplatin	%	>100	52.3	94.7	91.4–98.1
Residual variability (σ)
Coxaliplatin	%	19.9	10.4	19.8	15.9–23.4
Platinum
Fixed-effects parameters (θ)
V1 platinum	L/kg	0.17	14.9	0.17	0.13–0.23
V2 platinum	L/kg	0.31	21.5	0.35	0.21–0.65
CL1 platinum	L/h/kg	0.63	19.5	0.63	2.2–3.75
CL2 platinum	L/h/kg	0.79	17.5	0.79	0.30–0.92
Inter-individual variability (ω)
V1 platinum	%	56.2	23.1	52.7	51.8–53.5
V2 platinum	%	32.6	16.2	32.7	31.9–33.4
CL1 platinum	%	61.8	20.7	62.1	61.1–63.1
Residual variability (σ)
Cplatinum	%	29.3	7.1	28.8	25.2–32.6

V_{1 oxaliplatin}: distribution volume of oxaliplatin in the central compartment; V_{2 oxaliplatin}: distribution volume of oxaliplatin in the peripheral compartment; CL_{1 oxaliplatin}: clearance of oxaliplatin from the central compartment; CL_{2 oxaliplatin}: oxaliplatin intercompartmental clearance; C_oxaliplatin: plasma concentration of oxaliplatin; V_{1 platinum}: distribution volume of platinum in the central compartment; V_{2 platinum}: distribution volume of platinum in the peripheral compartment; CL_{1 platinum}: clearance of platinum from the central compartment; CL_{2 platinum}: platinum intercompartmental clearance; C_platinum: plasma platinum concentration.

Model-Based Simulation of Plasma 5-FU Concentration

Figure 4A shows the observed AUC values of 5-FU in 15 rats. The AUC value of the adenine-induced CKD group exceeded the target AUC range. Figure 4B shows the nomogram for AUC_last of 5-FU based on the P_Cr levels and the pharmacokinetic model of 5-FU after S-1 treatment in rats with CKD. A nonlinear increase in 5-FU exposure was simulated based on the P_Cr levels in CKD. For instance, 1 adenine-induced CKD model rat exhibited a P_Cr level of 2.98 mg/dL and a 5-FU AUC of 28.4 µg · h/mL, significantly exceeding the target range. This indicated the need to reduce the initial dose from 2 to 1 mg/kg to achieve the target AUC. These results highlight the importance of renal function-based dose adjustments of S-1.

Fig. 4. Observed AUC Values of 5-FU and a Nomogram for S-1 Dosing Based on Plasma Creatinine Levels

(A) Observed AUC values of 5-FU in 15 rats. (B) A nomogram showing the recommended total daily doses of S-1, according to the AUC values of 5-FU and P_Cr levels in CKD model rats. P_Cr: plasma creatinine.

DISCUSSION

While oxaliplatin-induced acute kidney injury has been reported, its incidence is relatively low and usually transient.²¹⁾ In contrast, CKD is more prevalent and persistent in cancer patients.²²⁾ Therefore, in this study, we focused on CKD to evaluate its impact on the pharmacokinetics of S-1 and oxaliplatin, and to explore the potential for dose individualization in such patients. We compared 2 CKD models, 5/6 nephrectomy and adenine-induced CKD, to evaluate the drug exposure relationship between renal biomarker levels and S-1 and oxaliplatin pharmacokinetics. Both models showed elevated renal biomarker levels, with more severe renal impairment observed in the adenine-induced CKD group. Moreover, the pharmacokinetic study revealed a significant increase in the AUC of 5-FU following S-1 administration in the adenine-induced CKD group. However, no substantial changes in the AUC were observed in the 5/6 nephrectomy CKD group. Population pharmacokinetic analysis revealed P_Cr as a covariate of 5-FU clearance, and model-based simulations highlighted the potential for establishing personalized dosing strategies based on renal biomarker levels in patients with cancer and CKD.

To confirm CKD in both the 5/6 nephrectomy and adenine-induced CKD groups, P_Cr levels, BUN levels, and C_Cr values were measured. Both CKD models exhibited reduced C_Cr values and high P_Cr and BUN levels compared to those in the normal groups, consistent with previous reports.^23,24) Moreover, P_Cr and BUN levels were higher in the adenine-induced CKD group than in the 5/6 nephrectomy CKD group. The nonsurgical approach to induce CKD using adenine causes renal dysfunction by forming 2,8-dihydroxyadenine, which crystallizes in the renal tubules, causing injury, inflammation, atrophy, and fibrosis.²⁵⁾ Upon comparing the renal function parameters and applying clinical CKD staging criteria used in humans,²⁶⁾ previous studies have attempted to estimate human CKD stages in rodent models based on P_Cr thresholds.²⁷⁾ In our study, the 5/6 nephrectomy model exhibited a mean P_Cr level of 1.4 ± 0.5 mg/dL, suggesting moderate CKD (stage 3), whereas the adenine-induced model showed a higher P_Cr level of 2.9 ± 0.9 mg/dL, indicating more advanced renal dysfunction with stage 4 or higher in humans.^27,28) Although species differences limit direct extrapolation, these P_Cr-based thresholds provide a reasonable basis for estimating CKD severity in rodent models.

Our pharmacokinetic study indicated that the AUC value of 5-FU and the t_1/2 and AUC values of CDHP in the adenine-induced CKD group were higher than those in the normal and 5/6 nephrectomy CKD groups. In clinical practice, 5-FU-derived adverse events are elevated in patients with cancer and renal impairment receiving S-1 due to delayed CDHP elimination.⁵⁾ Notably, this was not observed in the 5/6 nephrectomy CKD group; however, the adenine-induced CKD group showed behavior consistent with this clinical observation. In the 5/6 nephrectomy group, unexpectedly elevated apparent clearance of CDHP was observed, despite renal impairment. This anomaly is likely attributable to reduced bioavailability, as suggested by the lower plasma concentrations of both tegafur and CDHP in this group. Gastrointestinal absorption may have been reduced by ischemia–reperfusion injury associated with the surgical model.²⁹⁾ Since our 1-compartment model in the population pharmacokinetic analysis of S-1 did not account for such changes in absorption, these factors likely limited the usefulness of P_Cr as a covariate for CDHP clearance. These findings suggest that the adenine-induced CKD model mimics the pharmacokinetic alterations in patients with CKD receiving S-1 more closely than the 5/6 nephrectomy model.

In the adenine-induced CKD group, the increased systemic exposure of 5-FU may not only be attributed to the elevated CDHP levels but also to enhanced absorption of tegafur or accelerated conversion to 5-FU. Previous studies have shown that adenine-induced CKD can alter gastrointestinal physiology, such as increasing intestinal permeability and modifying luminal pH,³⁰⁾ which may facilitate greater absorption of orally administered drugs. Additionally, the flip-flop phenomenon known to occur after oral tegafur administration complicates the interpretation of 5-FU pharmacokinetics, as the terminal phase reflects absorption or conversion rather than elimination.³¹⁾ Although the half-life estimation of 5-FU in this study should be interpreted with caution due to the limited sampling window, the shorter t_1/2 observed in the adenine model may suggest faster conversion of tegafur to 5-FU. However, the activity of hepatic CYP3A and CYP2A, which mediate tegafur bioactivation in rats, was not evaluated in this study, as our primary aim was to explore the relationship between renal biomarkers and systemic exposure to anticancer drugs. Nevertheless, previous studies have demonstrated that CKD pathophysiology can affect the expression and activity of drug-metabolizing enzymes.³²⁾ Notably, adenine-induced CKD has been reported to cause significant reductions in renal and hepatic CYP450 activity, particularly CYP1A1, whereas 5/6 nephrectomy models show less pronounced or isozyme-specific alterations, such as preserved CYP3A expression.³²⁾ In addition, these differences between the 2 CKD models may contribute to the distinct pharmacokinetic profiles of 5-FU observed in this study. A more detailed mechanistic investigation would require intravenous administration of tegafur and comparison with oral dosing to distinguish between changes in absorption and metabolism.

For intact oxaliplatin and platinum, a prolonged t_1/2 and increased AUC_0–∞ in the CKD group were observed compared to the normal group. These results could potentially increase the risk of toxicity in patients with CKD undergoing oxaliplatin-based chemotherapy. Although systemic oxaliplatin concentrations increased in the CKD models, this may not be solely attributable to reduced renal excretion. Oxaliplatin is rapidly converted into various biotransformation products after administration, which then irreversibly bind to macromolecules in vivo, leading to elimination through nonrenal pathways.^33–36) As these processes are not directly dependent on renal function, P_Cr alone may not sufficiently explain the variability in oxaliplatin clearance. A decrease in the volume of distribution suggests impaired tissue uptake, potentially caused by uremia-related physiological changes.³⁷⁾ While tissue platinum concentrations were not assessed in this study, future investigations should address tissue distribution to clarify its impact on both efficacy and toxicity. Oxaliplatin exhibits a lower risk of nephrotoxicity than other platinum-based anticancer agents, and dose modification of oxaliplatin-based chemotherapy is uncommon in patients with CKD.⁶⁾ However, altered pharmacokinetics possibly contribute to oxaliplatin-induced toxicity, as demonstrated by the relationship between drug exposure and dose-limiting toxicity in our animal study.²⁰⁾ Collectively, these results suggest that the pharmacokinetic evaluation of intact oxaliplatin aids in optimizing oxaliplatin dosing while minimizing the risk of toxicity in patients with CKD.

Population pharmacokinetic analysis of S-1, intact oxaliplatin, and platinum revealed that P_Cr is a covariate of the clearance of 5-FU. P_Cr significantly improved the model fit for 5-FU clearance, likely reflecting the indirect impact of CKD on systemic 5-FU exposure through prolonged DPD inhibition by CDHP.³⁸⁾ These findings support the utility of P_Cr as an indirect but effective predictor of 5-FU pharmacokinetics in CKD settings. Our P_Cr-based modeling and simulation approach might provide a novel and clinically applicable tool for individualized S-1 dose adjustment in patients with cancer and CKD. The developed nomogram enables clinicians to estimate the optimal dosing based on P_Cr levels. Moreover, simulations confirmed its effectiveness in controlling systemic 5-FU exposure, thereby supporting safer chemotherapeutic management.

This study has several limitations. First, long-term monitoring of CKD and corresponding biomarker levels was not performed in this study. Only P_Cr levels and C_Cr values were selected as renal function markers in this study owing to their widespread use in pharmacokinetic studies and clinical practice. However, other biomarkers, such as liver fatty acid-binding protein and neutrophil gelatinase-associated lipocalin, should also be evaluated further to understand CKD progression and its impact on drug clearance. Second, this study used a single-dose regimen instead of a multiple-dose regimen. This approach was used to assess the pharmacokinetic effects of CKD without the confounding influence of drug accumulation or adaptive physiological changes due to repeated dosing.³⁹⁾ Owing to the variability in CKD animal models and the lack of a universal gold standard for chemotherapy pharmacokinetic evaluation, this single-dose study used a controlled environment to assess the renal function-related pharmacokinetic changes. Finally, as with all preclinical studies, these results cannot be directly extrapolated to humans. Although the population pharmacokinetic model was built using a limited number of rats, it highlights the strength of this approach in identifying covariates such as P_Cr. While CKD is multifactorial and relying on a single biomarker may be simplistic, the model suggests that P_Cr-based strategies could support biomarker-guided dose individualization. Further clinical validation is needed to assess its applicability in patients with cancer and CKD.

CONCLUSION

In conclusion, our findings reveal the exposure–renal biomarker relationship of 5-FU following S-1 administration by using 2 different CKD model rats. Population pharmacokinetic analysis identified P_Cr as a covariate of 5-FU clearance, and P_Cr-based modeling and simulations highlighted its potential to inform individualized dose optimization for patients with cancer and CKD. Although renal biomarker levels were not predictive of oxaliplatin exposure, the observed elevation in systemic levels underscores the importance of monitoring. These findings support the potential utility of pharmacokinetic model-informed SOX dosing strategies in clinical settings, pending further validation.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

This article contains supplementary materials.

REFERENCES

• 1)  Ishii T, Fujimaru T, Nakano E, Takahashi O, Nakayama M, Yamauchi T, Komatsu Y. Association between chronic kidney disease and mortality in stage IV cancer. Int. J. Clin. Oncol., 25, 1587–1595 (2020).
• 2)  Velenosi TJ, Urquhart BL. Pharmacokinetic considerations in chronic kidney disease and patients requiring dialysis. Expert Opin. Drug Metab. Toxicol., 10, 1131–1143 (2014).
• 3)  Tangri N, Rastogi A, Nekeman-Nan C, Hong LS, Ozaki A, Franzén S, Sofue T. Dapagliflozin utilization in chronic kidney disease and its real-world effectiveness among patients with lower levels of albuminuria in the U.S.A. and Japan. Adv. Ther., 41, 1151–1167 (2024).
• 4)  Sunami E, Kusumoto T, Ota M, et al. S-1 and oxaliplatin versus tegafur-uracil and leucovorin as postoperative adjuvant chemotherapy in patients with high-risk stage III colon cancer (ACTS-CC 02): a randomized, open-label, multicenter, phase III superiority trial. Clin. Colorectal Cancer, 19, 22–31.e6 (2020).
• 5)  Booka E, Imamura CK, Takeuchi H, Hamamoto Y, Gomi D, Mizukami T, Ichiyama T, Tateishi K, Takahashi T, Kawakubo H, Soejima K, Boku N, Tanigawara Y, Kitagawa Y. Development of an S-1 dosage formula based on renal function by a prospective pharmacokinetic study. Gastric Cancer, 19, 876–886 (2016).
• 6)  Watanabe D, Fujii H, Matsuhashi N, Iihara H, Yamada Y, Ishihara T, Takahashi T, Yoshida K, Suzuki A. Dose adjustment of oxaliplatin based on renal function in patients with metastatic colorectal cancer. Anticancer Res., 40, 2379–2386 (2020).
• 7)  Filewod N, Lipman ML. Severe acute tubular necrosis observed subsequent to oxaliplatin administration. Clin. Kidney J., 7, 68–70 (2014).
• 8)  Janus N, Launay-Vacher V, Byloos E, Machiels JP, Duck L, Kerger J, Wynendaele W, Canon JL, Lybaert W, Nortier J, Deray G, Wildiers H. Cancer and renal insufficiency results of the BIRMA study. Br. J. Cancer, 103, 1815–1821 (2010).
• 9)  Launay-Vacher V, Janus N, Deray G. Renal insufficiency and cancer treatments. ESMO Open, 1, e000091 (2016).
• 10)  Kim K, Anderson EM, Thome T, Lu G, Salyers ZR, Cort TA, O’Malley KA, Scali ST, Ryan TE. Skeletal myopathy in CKD: a comparison of adenine-induced nephropathy and 5/6 nephrectomy models in mice. Am. J. Physiol. Renal Physiol., 321, F106–F119 (2021).
• 11)  Takayama J, Takaoka M, Matsumura Y. Acute and chronic renal failure model in rats and mice. Nippon Yakurigaku Zasshi, 131, 37–42 (2008).
• 12)  Diwan V, Brown L, Gobe GC. Adenine-induced chronic kidney disease in rats. Nephrology, 23, 5–11 (2018).
• 13)  Rahman A, Yamazaki D, Sufiun A, Kitada K, Hitomi H, Nakano D, Nishiyama A. A novel approach to adenine-induced chronic kidney disease associated anemia in rodents. PLOS ONE, 13, e0192531 (2018).
• 14)  Tokunaga A, Miyamoto H, Fumoto S, Nishida K. Effect of chronic kidney disease on hepatic clearance of drugs in rats. Biol. Pharm. Bull., 43, 1324–1330 (2020).
• 15)  Tanaka T, Kobuchi S, Ito Y, Sakaeda T. Pharmacokinetic evaluation of oxaliplatin combined with S-1 (SOX) chemotherapy in a rat model of colorectal cancer with acute kidney injury: predictive renal biomarkers for dose optimization. Xenobiotica, 53, 613–620 (2023).
• 16)  Seetharam R, Sood A, Goel S. Oxaliplatin: pre-clinical perspectives on the mechanisms of action, response, and resistance. Ecancermedicalscience, 3, 153 (2009).
• 17)  Kobuchi S, Kai M, Ito Y. Population pharmacokinetic model-based evaluation of intact oxaliplatin in rats with acute kidney injury. Cancers, 13, 6382 (2021).
• 18)  Inoue S, Ohtani H, Tsujimoto M, Hori S, Sawada Y. Development of a pharmacokinetic model to optimize the dosage regimen of TS-1, a combination preparation of tegafur, gimeracil, and oteracil potassium. Drug Metab. Pharmacokinet., 22, 162–168 (2007).
• 19)  Kobuchi S, Ito Y, Nakano Y, Sakaeda T. Population pharmacokinetic modeling and simulation of 5-fluorouracil incorporating a circadian rhythm in rats. Xenobiotica, 46, 597–604 (2016).
• 20)  Kobuchi S, Katsuyama Y, Ito Y. Mechanism-based pharmacokinetic-pharmacodynamic (PK-PD) modeling and simulation of oxaliplatin for hematological toxicity in rats. Xenobiotica, 50, 223–230 (2020).
• 21)  Ulusakarya A, Misra S, Haydar M, Habert H, Castagne V, Gumus Y, Delmas-Marsalet B, Machover D. Acute renal failure related to oxaliplatin-induced intravascular hemolysis. Med. Oncol., 27, 1425–1426 (2010).
• 22)  Ciorcan M, Chisavu L, Mihaescu A, Gadalean F, Bob FR, Negru S, Schiller OM, Grosu ID, Marc L, Chisavu F, Dragota Pascota R, Apostol A, Ivan V, Schiller A. Chronic kidney disease in cancer patients, the analysis of a large oncology database from Eastern Europe. PLOS ONE, 17, e0265930 (2022).
• 23)  Huang MC, Chen ME, Hung HC, Chen HC, Chang WT, Lee CH, Wu YY, Chiang HC, Hwang SJ. Inadequate energy and excess protein intakes may be associated with worsening renal function in chronic kidney disease. J. Ren. Nutr., 18, 187–194 (2008).
• 24)  Yokozawa T, Zheng PD, Oura H, Koizumi F. Animal model of adenine-induced chronic renal failure in rats. Nephron J., 44, 230–234 (1986).
• 25)  Klinkhammer BM, Djudjaj S, Kunter U, Palsson R, Edvardsson VO, Wiech T, Thorsteinsdottir M, Hardarson S, Foresto-Neto O, Mulay SR, Moeller MJ, Jahnen-Dechent W, Floege J, Anders HJ, Boor P. Cellular and molecular mechanisms of kidney injury in 2,8-dihydroxyadenine nephropathy. J. Am. Soc. Nephrol., 31, 799–816 (2020).
• 26)  Nelson AW, Mackinnon B, Traynor J, Geddes CC. The relationship between serum creatinine and estimated glomerular filtration rate: implications for clinical practice. Scott. Med. J., 51, 5–9 (2006).
• 27)  Saito H, Miyakoshi N, Kasukawa Y, Nozaka K, Tsuchie H, Sato C, Abe K, Shoji R, Shimada Y. Analysis of bone in adenine-induced chronic kidney disease model rats. Osteoporos. Sarcopenia, 7, 121–126 (2021).
• 28)  Chevalier RL. The proximal tubule is the primary target of injury and progression of kidney disease: role of the glomerulotubular junction. Am. J. Physiol. Renal Physiol., 311, F145–F161 (2016).
• 29)  Takeda F, Oda M, Terasaki M, Kubota A, Asada K, Ichimura Y, Kojima H, Saitoh H. Downregulated expression of organic anion transporting polypeptide (Oatp) 2b1 in the small intestine of rats with acute kidney injury. Drug Metab. Pharmacokinet., 40, 100411 (2021).
• 30)  Bartochowski P, Gayrard N, Bornes S, Druart C, Argilés A, Cordaillat-Simmons M, Duranton F. Gut-kidney axis investigations in animal models of chronic kidney disease. Toxins, 14, 626 (2022).
• 31)  Yoshisue K, Kanie S, Nishimura T, Chikamoto J, Nagayama S. Effect of dimethylnitrosamine-induced liver dysfunction on the pharmacokinetics of 5-fluorouracil after administration of S-1, an antitumour drug, to rats. J. Pharm. Pharmacol., 61, 1643–1651 (2009).
• 32)  Guévin C, Michaud J, Naud J, Leblond FA, Pichette V. Down-regulation of hepatic cytochrome P450 in chronic renal failure: role of uremic mediators. Br. J. Pharmacol., 137, 1039–1046 (2002).
• 33)  Lévi F, Metzger G, Massari C, Milano G. Oxaliplatin: pharmacokinetics and chronopharmacological aspects. Clin. Pharmacokinet., 38, 1–21 (2000).
• 34)  Arnould S, Hennebelle I, Canal P, Bugat R, Guichard S. Cellular determinants of oxaliplatin sensitivity in colon cancer cell lines. Eur. J. Cancer, 39, 112–119 (2003).
• 35)  Pinato O, Musetti C, Farrell NP, Sissi C. Platinum-based drugs and proteins: reactivity and relevance to DNA adduct formation. J. Inorg. Biochem., 122, 27–37 (2013).
• 36)  Koellensperger G, Hann S. Ultra-fast HPLC-ICP-MS analysis of oxaliplatin in patient urine. Anal. Bioanal. Chem., 397, 401–406 (2010).
• 37)  Cunha RS, Santos AF, Barreto FC, Stinghen AEM. How do uremic toxins affect the endothelium? Toxins, 12, 412 (2020).
• 38)  Hironaka S, Sadachi R, Machida N, Iwasa S, Yamada Y, Sasako M, Yoshikawa T, Boku N, Terashima M. Association of renal function with the safety and efficacy of cisplatin plus S-1 therapy and docetaxel plus cisplatin plus S-1 therapy in patients with advanced gastric cancer: an exploratory analysis of JCOG1013. Jpn. J. Clin. Oncol., 52, 14–23 (2022).
• 39)  Wason S, Mount D, Faulkner R. Single-dose, open-label study of the differences in pharmacokinetics of colchicine in subjects with renal impairment, including end-stage renal disease. Clin. Drug Investig., 34, 845–855 (2014).