Effect of tyrosine kinase inhibitors on renal handling of creatinine by MATE1

Creatinine is actively secreted across tubular epithelial cells via organic cation transporter 2 (OCT2) and multidrug and toxin extrusion 1 (MATE1). We previously showed that the tyrosine kinase inhibitor (TKI) crizotinib inhibits OCT2-mediated transport of creatinine. In the present work, we examined the inhibitory potency of TKIs, including crizotinib, on MATE1-mediated transport of creatinine. Then, we used the kinetic parameters estimated in this and the previous work to predict the potential impact of TKIs on serum creatinine level (SCr) via reversible inhibition of creatinine transport. Crizotinib inhibited [14C]creatinine uptake by MATE1-overexpressing cells, and the inhibitory effect increased with incubation time, being greater in the case of pre-incubation or combined pre-incubation/co-incubation (pre/co-incubation) than in the case of co-incubation alone. The inhibition was non-competitive, with Ki values of 2.34 μM, 0.455 μM and 0.342 μM under co-, pre- or pre/co-incubation conditions, respectively. Similar values were obtained for inhibition of [3H]MPP+ uptake by MATE1-overexpressing cells. Gefitinib, imatinib, pazopanib, sorafenib, and sunitinib also inhibited MATE1-mediated creatinine uptake. Further, all these TKIs except pazopanib inhibited [14C]creatinine uptake by OCT2-overexpressing cells. In rat kidney slices, the ratio of unbound tissue accumulation of TKIs to extracellular concentration ranged from 2.05 to 3.93. Prediction of the influence of TKIs on SCr based on the renal creatinine clearance and plasma maximum unbound concentrations of TKIs suggested that crizotinib and imatinib might increase SCr by more than 10% in the clinical context. Accordingly, it is necessary to be cautious in diagnosing TKI-induced renal failure only on the basis of an increase of SCr.

We recently reported that crizotinib is an inhibitor of OCT2-mediated uptake of creatinine, and we found that its inhibitory effect after pre-incubation or combined pre-incubation/co-incubation (pre/co-incubation) was greater than that in the case of co-incubation alone 20 . Furthermore, the inhibitory effect of crizotinib on OCT2 was substrate-dependent. It is important to know whether inhibition of MATE1, which mediates renal secretion of creatinine at the apical membrane, by TKIs shows similar characteristics. Therefore, in the present study we examined on the inhibitory effects of TKIs, including crizotinib, on MATE1-mediated uptake of creatinine. Furthermore, we used the kinetic parameters obtained in the present and previous work to model the effect of TKI-induced inhibition of creatinine transporters on SCr in order to assess whether it might be significant in relation to the clinical diagnosis of TKI-induced renal failure.

Results
Inhibitory Effect of Crizotinib on Creatinine and MPP + Uptake by MATE1-Overexpressing HEK293 Cells. Uptake of [ 14 C]creatinine by HEK293/MATE1 cells was significantly higher than that by mock cells, and since the uptake increased almost linearly up to 3 min (Supplemental Fig. 1A), initial uptakes were evaluated at 2 min in subsequent experiments. On the other hand, although uptake of [ 14 C]creatinine by HEK293/MATE2K cells was comparable to that by mock cells, clear uptake of MPP + was observed (mock: 5.76 μL/mg/1 min; HEK293/MATE2K: 19.1 μL/mg/1 min). Then, we examined the time dependence of inhibition of MATE1-mediated [ 14 C]creatinine uptake by crizotinib. We found that the extent of the inhibition increased time-dependently, reaching a plateau after 60 min pre-incubation (Fig. 1A). Therefore, the pre-incubation time was set at 60 min in subsequent studies. Next, the inhibitory effect was compared among co-incubation, pre-incubation, and pre/co-incubation protocols at a fixed substrate concentration. Co-, pre-, and pre/co-incubation with crizotinib concentration-dependently reduced the uptake of [ 14 C]creatinine by HEK293/MATE1 cells with IC 50 values of 2.16 ± 0.31 μM, 0.804 ± 0.132 μM, and 0.573 ± 0.102 μM, respectively ( Fig. 2A). Similarly, co-, pre-, and pre/co-incubation with crizotinib reduced the uptake of [ 3 H]MPP + by HEK293/MATE1 cells with IC 50 values of 2.66 ± 0.46 μM, 1.45 ± 0.22 μM, and 0.868 ± 0.097 μM, respectively (Fig. 2B). On the other hand, MATE2K-mediated [ 3 H]MPP + uptake was inhibited by crizotinib with IC 50 values in the range of 8.8 ± 3.4 μM (Supplemental Fig. 1B), showing that the inhibitory potency of crizotinib on MATE2K was less than that on MATE1. Accordingly, in the present study, we considered that TKIs would predominantly influence MATE1mediated creatinine secretion, and that the effect on MATE2K could be neglected.
Kinetic Analysis of the Inhibitory Effect of Crizotinib on Creatinine Uptake by MATE1-Overexpressing HEK293 Cells. To study the inhibitory mechanism of crizotinib on MATE1-mediated [ 14 C]creatinine uptake, the concentration dependence of creatinine uptake by HEK293/MATE1 cells in the presence and absence of crizotinib was examined ( Fig. 3A-C). The results are summarized in Table 1. Under all treatment conditions (co-, pre-incubation and pre/co-incubation), crizotinib reduced the apparent V max value, while the K m value was essentially unaffected, demonstrating that crizotinib is a non-competitive inhibitor of creatinine uptake by MATE1 under all these incubation conditions. The K i values were 2.34 μM, 0.455 μM, and 0.342 μM under co-, pre-or pre/co-incubation conditions, respectively. Intracellular Accumulation of Crizotinib during Pre-incubation. We hypothesized that the time dependence of the inhibition (pre-incubation effect) is due to intracellular accumulation of crizotinib. Accordingly, to test this idea, we examined the intracellular accumulation of crizotinib using monolayers of MDCKII/MATE1 cells in a Transwell system. Pre-incubation of the cells with crizotinib (2 μM) added on the basolateral side alone reduced [ 14 C]creatinine uptake from the apical side (Fig. 4). In this experiment, the observed crizotinib concentration in the apical side medium after the pre-incubation was 0.105 ± 0.015 μM, which is below the Ki value under the co-incubation condition. Accordingly, since the observed reduction of creatinine uptake from the apical side cannot be accounted for by the crizotinib concentration in the apical   side medium, these findings confirm that the pre-incubation effect is due to the intracellular accumulation of crizotinib.

Inhibitory Effects of Other TKIs on Creatinine Uptake by OCT2-and MATE1-Overexpressing HEK293 Cells and TKI accumulation in Rat Kidney Slices. To study the inhibitory potential of other
TKIs on OCT2-and MATE1-mediated uptake of [ 14 C]creatinine, the concentration dependence of creatinine uptake by HEK293/MATE1 and HEK293/OCT2 cells was examined in the presence of several TKIs. Firstly, in order to confirm that the pre-incubation effect is not specific to crizotinib, the effects of co-and pre/co-incubation with imatinib on [ 14 C]creatinine uptake by HEK293/MATE1 cells were examined. Imatinib reduced [ 14 C]creatinine uptake with IC 50 values of 0.466 ± 0.053 μM and 0.0651 ± 0.0087 μM under co-and pre/co-incubation conditions, respectively (Supplemental Fig. 2), demonstrating that the pre-incubation effect is not specific to crizotinib. Since pre-incubation is considered to be clinically relevant and results in greater inhibitory potency, the pre/ co-incubation condition was used to examine the effects of other TKIs. As shown in Fig. 5, [ 14 C]creatinine uptake by HEK293/MATE1 cells was decreased in the presence of each of imatinib, gefitinib, pazopanib, sunitinib, and sorafenib in a concentration-dependent manner. The order of inhibitory potency for MATE-mediated uptake of [ 14 C]creatinine was imatinib > crizotinib > pazopanib > sunitinib > sorafenib > gefitinib. In addition, [ 14 C]creatinine uptake by HEK293/OCT2 cells was inhibited by imatinib, gefitinib, sunitinib, and sorafenib, though not by pazopanib (Fig. 6), and the order of inhibitory potency of TKIs for OCT2-mediated uptake of [ 14 C]creatinine was sunitinib > crizotinib > imatinib > gefitinib > sorafenib > pazopanib. The IC 50 values of TKIs for OCT2-and MATE1-mediated uptake of creatinine are summarized in Table 2.
We also examined the tissue accumulation of crizotinib in rat kidney slices. The uptake of crizotinib at 37 °C was markedly higher than that at 4 °C (Fig. 7), suggesting carrier-mediated uptake of crizotinib into rat kidney slices. Because the uptake attained a steady-state at 60 min, the K p,uu values of TKIs were calculated at 120 min. The K p,uu value of crizotinib was 3.28 ± 0.28, which suggests that crizotinib is concentratively accumulated in proximal tubular cells. The K p,uu values of gefitinib, imatinib, pazopanib, sorafenib, and sunitinib were also determined, and ranged from 2.05 ± 0.28 (pazopanib) to 3.93 ± 0.15 (sunitinib), as shown in Table 2.
The determined IC 50 and K p,uu values were used, in conjunction with parameters taken from a previous clinical study of 11 subjects 10 , to predict the increment of SCr due to inhibition of transporter-mediated creatinine uptake by TKIs. As shown in Table 3, the results indicate that crizotinib and imatinib could cause an increase of SCr of more than 10% in the clinical context.

Discussion
Most TKIs have the potential to cause renal failure, and since SCr is a major indicator of acute and chronic renal failure, it is clinically important to understand the potential of these drugs to cause reversible inhibition of renal creatinine transporters in order to avoid incorrect diagnosis of drug-induced renal failure. We previously showed that crizotinib in the clinically relevant concentration range has the potential to inhibit OCT2-mediated creatinine transport 20 . In the present study, therefore, we focused on the potential of TKIs, including crizotinib, to inhibit uptake of creatinine by MATE1, the other major renal creatinine transporter. Moreover, we aimed to use the estimated kinetic parameters to predict the extent of SCr increase that might be caused by TKIs through inhibition of transporter-mediated uptake of creatinine. Firstly, we considered the effect of pre-incubation on the inhibitory potential of crizotinib for MATE1-mediated transport of creatinine, because kidney tissues are clinically exposed to administered drugs for a long period of time and many previous studies have been done without pre-incubation. As shown in Fig. 1, crizotinib inhibited MATE1-mediated uptake of [ 14 C]creatinine in a time-dependent manner, as we had previously found in the case of crizotinib inhibition of OCT2 20 . Accordingly, pre-incubation for an appropriate period is essential for adequate evaluation of the inhibitory activity of drugs towards creatinine transport by OCT2 and MATE1. Secondly, many drugs may be transported by OCT2 and/or MATE1 to various extents 21-25 , so we examined whether the inhibitory effects of TKIs on MATE1 are substrate-dependent. In this context, we previously reported that the IC 50 value of crizotinib for MPP + uptake by OCT2 was 10-20 times higher than that for creatinine uptake 20 . Here, however, in contrast to the case of OCT2, we found that the IC 50 values of crizotinib for MATE1 were comparable with creatinine and MPP + as substrates (Fig. 2). This apparent discrepancy may be due to different inhibition mechanisms; crizotinib appears to be a competitive inhibitor of OCT2, but a non-competitive inhibitor of MATE1 (Fig. 3). The mechanism of the the crizotinib-mediated reduction of V max value of creatinine uptake by MATE1 is uncertain. One possibility would be a decrease of MATE1 protein in the cellular membrane due to the altered trafficking, but this seems unlikely, since our preliminary study showed no change of intracellular signals of green fluorescent protein (GFP)-fused MATE1 protein in response to crizotinib. Accordingly, other mechanisms such as structural changes of MATE1 due to an allosteric effect of bound crizotinib, or dephosphorylation or degradation of MATE1, which could be associated with the pharmacological effect of crizotinib, should be considered. It has been reported that another TKI, dasatinib, reduced OCT2 activity by inhibiting the Src family kinase YES1, resulting in dephosphorylation of OCT2 19 . Therefore, the mechanism by which crizotinib inhibits MATE1 might involve dephosphorylation of MATE1 via inhibition of an unknown kinase(s). In addition, we found that creatinine was not transported by MATE2K in the present study, contrary to a previous report 6 . This apparent discrepancy might be explained by differences in the amounts of MATE2K expressed in the two models. In any event, crizotinib was a markedly less potent inhibitor of MATE2K than MATE1. Therefore, MATE2K was not incorporated in the equation for prediction of creatinine clearance, as its contribution to the effect of TKIs is presumed to be small.
Next, to aid in predicting the inhibitory potential of TKIs on MATE1 in the clinical context, we estimated the intracellular concentrations of TKIs, because MATE1 excretes intracellular creatinine to the tubular lumen. Tissue accumulation, K p,uu , of crizotinib in kidney was estimated using rat kidney slices. For concentration setting, it is preferable to use the clinically relevant unbound concentration. However, the clinically unbound concentration of some TKIs is as low as about 10 nM. It is technically infeasible to use such low concentrations experimentally, because TKIs are extensively adsorbed on the experimental apparatus; therefore, the concentration of TKIs was unified to 1 μM in this study. In vitro uptake of crizotinib by the kidney slices showed a significant temperature dependence, suggesting that it is transporter-mediated, at least in part. Although the exact mechanism of crizotinib uptake into proximal tubular epithelial cells is unclear, it has been reported that several TKIs are substrates of solute carrier transporters, including OCT2 26 . Thus, transporters expressed in tubular epithelial cells, such as OCT2 and organic anion transporter 3 (OAT3), may play a role in accumulation of TKIs in kidney. Nevertheless,  it will be important in the future to establish the mechanism of the pre-incubation effect in order to ensure that the prediction of SCr change is soundly based. Finally, we used the estimated kinetic parameters in a modeling analysis to investigate whether inhibition of transporters by TKIs might cause a clinically significant increase of SCr. As shown in Table 3, based on the renal creatinine clearance equation 10 and the reported C max,u , crizotinib could increase SCr by 14.5%, and the increment is increased to 25.3% when the K p,uu value is taken into consideration. We did not incorporate MATE2K in the equation for prediction of SCr, because creatinine was negligibly transported by HEK293/MATE2K cells, and also because crizotinib was a much less potent inhibitor of MATE2K, compared with OCT2 and MATE1. Moreover, the intrinsic clearances for OCT2 and MATE1 are similar, since the protein amounts of OCT2 and MATE1 are comparable in human kidney 27 , and a previous study well predicted the clinical increase of SCr based on this assumption 10 . The clinically reported eGFR reduction in patients treated with crizotinib was 23.9% (n = 38, 95% confidence interval (CI): 21.3-26.6%) 28 . Bearing in mind that crizotinib has a long half-life (t 1/2 ) of 42 hr, the predicted increase of SCr estimated from C trough,u might also be consistent with the clinical observation. Our modeling analysis also showed that imatinib could increase SCr by 42.0% and 29.4% based on C max,u and C trough,u , respectively, and these values increased to 52.6% and 42.6%, respectively, after incorporation of K p,uu . Again, these values are broadly consistent with the clinically reported decrease of eGFR of 27.1% (mean value calculated from 7 subjects 29 ). Considering a report that increasing effect of imatinib on SCr was reversible 13 , our conclusion here regarding imatinib appears to be clinically relevant. These results support the idea that specific interaction of some TKIs with renal creatinine transporters may lead to a reversible increase of SCr in the absence of renal failure. However, since it is also possible that crizotinib and imatinib cause true adverse events in kidney, we emphasize that it is desirable to use not only SCr, but also additional renal function markers such as blood urea nitrogen, to distinguish between reversible and true drug-induced renal failure. Moreover, it should be noted that incorporation of K p,uu values in the prediction may or may not be necessary depending on the mechanism of the pre-incubation effect; it is necessary if the pre-incubation effect is caused by kinase inhibition, but not if the pre-incubation effect is determined by the intracellular concentration of drugs. Further study will be needed to clarify the mechanism in order to improve the prediction of the changes in creatinine disposition.   In conclusion, the present study demonstrated that TKIs such as crizotinib and imatinib have the potential to inhibit creatinine transport via MATE1, as well as OCT2, at clinically relevant concentrations. The resulting reversible increase of SCr could lead to incorrect diagnosis of drug-induced renal failure.
Rats were housed three per cage with free access to commercial chow and tap water, and were maintained on a 12 h dark/light cycle in an air-controlled room (temperature, 24.0 ± 1 °C; humidity, 55 ± 5%). All animal studies were approved by the Kanazawa University Institutional Animal Care and Use Committee (Permit number, AP-163750), and were performed in accordance with the university guidelines.
Uptake by Transporter-expressing Cells. The uptake experiment was conducted as described previously 20 . HEK293 cells were plated at a density of 1.7 × 10 5 cells/cm 2 on 24-well plates and cultured for two days before uptake assay. The cells were pre-incubated with 0.5 ml of transport medium (TM) (125 mM NaCl, 4.8 mM KCl, 1.2 mM KH 2 PO 4 , 1.2 mM CaCl 2 , 1.2 mM MgSO 4 , 5.6 mM D-glucose, and 25 mM HEPES, pH 7.4) for an appropriate time at 37 °C. Cells for MATE1 assays were pre-incubated in TM containing 30 mM NH 4 Cl for 20 min to establish an outwardly directed H + gradient; this was done concomitantly with the pre-incubation with TKIs. To measure co-, pre-or both (pre/co-) incubation effects of TKIs, the cultured cells were pre-incubated with TM in the absence or presence of TKIs for predetermined periods, and uptake was initiated by replacing 0.25 ml of TM with TM containing [ 14 C]creatinine or [ 3 H]MPP + in the absence or the presence of TKIs. The cells were incubated at 37 °C. Uptake was terminated by washing the cells three times with 0.5 ml of ice-cold TM, and the cells were solubilized in 0.25 ml of 0.01% (v/v) Triton-X. Radioactivity was measured with a liquid scintillation counter (Hitachi Aloka Medical, Ltd., Tokyo, Japan). Part of the lysate was used for the determination of total protein amount with a protein assay kit (Bio-Rad Laboratories, Hercules, CA). For uptake assay using the Transwell system, MDCKII cells were plated at a density of 2.2 × 10 5 cells/cm 2 on a 12-well Transwell filter insert for 5 days before uptake study. After adding TM containing 2 μM crizotinib for 60 min, the medium on both sides was removed and the cells were washed three times with ice-cold TM. Uptake study was initiated by adding TM containing [ 14 C]creatinine and TM alone on the apical and basolateral sides, respectively. Uptake was terminated by washing the cells three times with ice-cold TM, and intracellular [ 14 C]creatinine concentration was determined by the method described above. In the experiments with crizotinib, non-labeled creatinine was used and the concentration was determined by LC-MS/MS. After initiation of uptake, the medium on the apical side was collected at 2 min, and the crizotinib concentration was determined.
Uptake by Rat Kidney Slices. Uptake study by rat kidney slices was carried out as described previously, with slight modifications 31 . Whole kidneys of male rats were sliced (0.3 mm thick) with a microslicer (Zero 1; Dosaka EM, Kyoto, Japan). The slices were immediately put in ice-cold oxygenated transport buffer (130 mM NaCl, 4.8 mM KCl, 1.2 mM CaCl 2 , 1.2 mM MgSO 4 , 1.2 mM KH 2 PO 4 and 25 mM HEPES, adjusted to pH 7.4). Three or four slices per rat, each weighing 5 to 20 mg, were randomly selected and pre-incubated in a 12-well plate with 2.0 ml/well of oxygenated transport buffer at 37 °C for 5 min. After pre-incubation, the kidney slices were placed in transport buffer containing TKIs to initiate the uptake reaction. After uptake at 37 °C for an appropriate time, each slice was rapidly removed from the transport buffer, washed twice in ice-cold transport buffer, blotted on filter paper and weighed. Non-specific adsorption of each substrate was estimated as the absorption on a slice incubated at 4 °C for 1 min. The tissue accumulation of TKIs was measured by liquid chromatography-tandem mass spectrometry (LC-MS/MS) as described below. Data analysis. Experiments were conducted at least twice, and the results are shown as representative values in individual experiments. Uptake by transporter-expressing cells was expressed as the cell-to-medium (C/M) ratio (μL/mg protein), obtained by dividing the uptake amount by the concentration of substrate in the TM. MATE1-or OCT2-mediated uptake was obtained after subtraction of the uptake by mock cells.

Measurement of
Kinetic parameters were calculated by means of a nonlinear least-squares method using KaleidaGraph (Synergy Software, Reading, PA), and are shown as mean +/− statistical errors from a single experiment. The inhibitory effect of TKIs was expressed as a percentage of the control, and the IC 50 value was obtained by means of the following Eq. (1); where [I] is inhibitor concentration. Kinetic parameters for uptake were obtained using the Michaelis-Menten Eq. (2); where v, V max , K m , and [C] are initial uptake rate, maximal uptake rate, Michaelis constant, and substrate concentration, respectively. K i of crizotinib was obtained from the following Eq. (3); where V max,app is the apparent maximal uptake rate in the presence of inhibitor. The unbound renal proximal tubule epithelial cell-to-medium concentration ratio (K p,uu ) based on the steady-state uptake at 37 °C and 4 °C was defined according to Eq. (4) 32 ;