Cadmium exposure enhances organic cation transporter 2 trafficking to the kidney membrane and exacerbates cisplatin nephrotoxicity

Abstract

Renal accumulation and exposure of cadmium originating from pollution in agricultural land and the prevalence of cigarette smoking remains an unneglectable human health concern. Whereas cadmium exposure has been correlated with increased incidence of a variety of kidney diseases, little is known pertaining to its effect on renal drug disposition and response in patients. Here, we report that cadmium exposure significantly increased the activity of organic cation transporter 2 (OCT2), a critical renal drug transporter recommended in United States Federal Drug Administration guidance for assessment during drug development. Cadmium enhanced OCT2 trafficking to the cell membrane both in vitro and in vivo. Mechanistically cadmium-mediated OCT2 translocation was found to involve protein-protein interaction between serine/threonine-protein kinase AKT2, calcium/calmodulin and the AKT substrate AS160 in in vitro cellular studies. The formed protein complex could selectively facilitate phosphorylation of AKT2 at T309, which induced translocation of OCT2 to the plasma membrane. Moreover, cadmium exposure markedly exacerbated nephrotoxicity induced by cisplatin, an OCT2 substrate, by increasing its accumulation in the mouse kidney. Consistently, there was a significant correlation between plasma cadmium level and alteration of renal function in cervical cancer patients who underwent chemotherapy with cisplatin. Thus, our studies suggest that membrane transporter distribution induced by cadmium exposure is a previously unrecognized factor for the broad variation in renal drug disposition and response.

The heavy metal cadmium (Cd) is an environmentally prevalent toxicant posing increasing risk to human health worldwide, due to human activities,¹ the increasing number of tobacco smokers worldwide,² and lack of efficient intervention.³ Although intoxication by high levels of Cd exposure is rarely seen nowadays, low levels of daily Cd exposure remain a major renal health concern, as Cd is a cumulative toxicant with an extremely long biological half-life of 10–30 years in the human body.⁴ Even non-occupationally exposed individuals can accumulate very high levels of Cd in tissues, particularly the kidney (e.g., 16 μg/g of wet tissue on average, or ~140 μM).⁵ Cd exposure may alter multiple cellular signaling pathways, induce mitochondrial damage, and culminate in cell death by apoptosis or necrosis.⁶ Mounting evidence has linked chronic low levels of Cd exposure to the cause of various diseases, including cancer, cardiovascular diseases, diabetes, renal tubular disease, bone damage, and obstructive pulmonary disease.⁷ By contrast, little is known about whether Cd exposure affects renal drug disposition and response in patients, as for instance, by interfering with the drug transporter proteins expressed in the kidney, where drugs are eliminated and the highest Cd accumulation is seen.^5,8

Membrane transporters play a critical role in disposition of and response to drugs.⁹ The International Transporter Consortium, in collaboration with the US Food and Drug Administration, has therefore issued recommendations on transporter function assessment during drug development.^9–13 Among the transporters recommended for assessment are organic cation transporters 2 (OCT2) and multidrug and toxin extrusion transporters (MATEs) in the kidney. The basal uptake via OCT2 and the apical efflux via MATEs in human renal proximal tubular cells is believed to be essential for urinary elimination of many cationic drugs,^14,15 such as cisplatin, a potent and economic anti-cancer drug, of which the clinical application is dose-limited by its nephrotoxicity.^16,17 Both animal and human studies have demonstrated that cisplatin-induced renal damage is causatively related to the function of OCT2/MATEs.^17–19

In this study, we report, for the first time, that Cd exposure could significantly increase the activity of OCT2 by enhancing its trafficking to cell membrane both in vitro and in vivo. We mechanistically revealed that Cd exposure could trigger the formation of a protein complex consisting of serine/threonine-protein kinase AKT2, Ca/calmodulin (Ca/CaM), and the AKT substrate AS160 (160 kDa), which selectively facilitate the phosphorylation of AKT2 at T309 to stimulate the translocation of OCT2 to plasma membrane. Moreover, Cd exposure was found to significantly increase the accumulation of cisplatin in mouse kidney, and dramatically exacerbate cisplatin-induced nephrotoxicity. Importantly, we observed a significant correlation between plasma Cd level and alteration of renal function in cervical cancer patients who were administered chemotherapeutic cisplatin. Our results demonstrate that Cd exposure, through drug transporter regulation, affects renal drug disposition and response.

RESULTS

Cd exposure increased OCT2 activity and its membrane expression

Cd has been reported to be a moderate substrate of OCTs^20,21 and MATEs.²² Our initial intention was to examine the potential of Cd to inhibit OCT/MATE-mediated drug disposition. With stable human embryonic kidney (HEK)293 cells expressing human OCT2 (HEK-human(h)OCT2), we unexpectedly found that Cd pretreatment could increase the activities of human (h) OCT2 in a time- and concentration-dependent manner (Figure 1a; Supplementary Figure S1A). In contrast, the activity of hMATE1 was not influenced (Supplementary Figure S1B). Interestingly, we observed steeper concentration-dependent effects of Cd on hOCT2 activity with shorter incubation times. Whereas the exact mechanism was unknown, we conceived of 2 possible explanations. First, Cd exposure might induce the expression of its binding proteins, such as metallothionein, over time. The intracellular free Cd ion concentrations might be relatively higher given our shorter times of exposure, such as 20 minutes and 4 hours, as compared to 12 hours and 96 hours. Second, the longer exposure might lead to a certain type of cytotoxicity that compromised transporter activities. For convenience, we conducted most of our cellular studies in vitro at relatively high concentrations (albeit ≤100 μM) with a short exposure time (20 minutes), to quickly achieve an effective intracellular free Cd level. We showed that the increased activity of hOCT2 by Cd exposure was accompanied by a significant increase of the maximal transport rate (V_max), and only a minor change in the kinetic affinity constant (K_m) of the probe substrate metformin (Figure 1b; Table 1), suggesting a regulatory effect by Cd exposure on the abundance of functional hOCT2 protein. With immunoblotting analysis, the abundance of hOCT2 protein in cell membrane was found to increase by Cd exposure, in a dose-and time-dependent manner, with no change in total hOCT2 proteins (Figure 1c; Supplementary Figure S1C). The immunocytochemistry assay confirmed this finding by showing a more intense and clear staining of hOCT2 in cell membrane after Cd exposure (Figure 1d).

Figure 1 |. Cadmium (Cd) exposure increases organic cation transporter 2 (OCT2) activity by enhancing its membrane expression.

(a) Cellular uptake of metformin in human embryonic kidney (HEK)–human (h) OCT2 cells with and without exposure to various concentrations of Cd at various times. (b) Dose-dependent cellular uptake of metformin in HEK-hOCT2 cells with and without exposure to Cd (50 μM) for 20 minutes. (c) Immunoblotting analysis of total and membrane fraction of hOCT2 for HEK-hOCT2 cells with and without exposure to Cd for 20 minutes. (d) Confocal images of immunostaining of hOCT2 (red) and 4′,6-diamidino-2-phenylindole (DAPI; blue) staining of nuclei for HEK-hOCT2 with and without exposure to Cd for 20 minutes. The white arrows indicate membrane expression of hOCT2. Bar = 20 μm. (e) Ex vivo accumulation of metformin (n = 3 per group) and (f) immunoblotting analysis of total and membrane fraction of mouse (m) Oct2/1 with the slices of mouse kidney being exposed with and without Cd (50 μM) for 20 minutes. (g) Renal accumulation of metformin (n = 4 per group) and (h) immunoblotting analysis of total and membrane fraction of mOct2/1 in the kidney tissues isolated from the mice treated with and without Cd-metallothionein (CdMT) complex (100 μg Cd/Kg, i.p.) for 6 hours. (a–d,f,h) Figures are representative of 3 independent experiments. (c,f,h) Pan-cadherin and Na⁺/K⁺ ATPase are given as the positive control (Cont), and β-actin as the negative Cont, for membrane proteins. An abundance of membrane protein was normalized by total protein. (a,b,e) Data are presented as mean ± SD; (g) data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 (a): analysis of variance followed by post hoc Tukey analysis; (e,g): 2-tailed Student’s t test. ATPase, adenosine triphosphatase. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.

Table 1 |.

Effect of cadmium (Cd) exposure on the kinetics of organic cation transporter 2–mediated metformin uptake in human embryonic kidney 293 cells^a

	Vmax (μM/g protein/min)	Km (mM)
Control	0.54 ± 0.08	2.9 ± 1.3
Cd2+ (50 μM)	1.2 ± 0.07	1.9 ± 0.3

K_m, kinetic affinity constant; V_max, maximal transport rate.

^aThe apparent V_max and K_m values were calculated by fitting the data from Figure 1b to the Michaelis-Menten equation. Data are given as mean ± SD.

The increased activity of OCT2 by Cd exposure in the cells was verified in mouse kidney tissues. In the ex vivo studies, the accumulation of metformin was increased significantly after slices of mouse kidney were pre-incubated with Cd for 20 minutes (Figure 1e), accompanied by an increased abundance of membrane fraction of mouse (m) Oct2 and Oct1 (Figure 1f). Next, we examined the in vivo effect of Cd exposure on renal Oct activity by injecting mice with Cd-metallothionein (CdMT), which has been widely employed to achieve quick Cd exposure in the kidney²³ (Supplementary Figure S2). As with the in vitro and ex vivo results, CdMT injection was found to significantly enhance the accumulation of metformin in mouse kidney (Figure 1g) and increase the abundance of mOct2/1 protein in the cell membrane (Figure 1h). Altogether, our observations from the in vitro, ex vivo, and in vivo studies have demonstrated an inductive effect of Cd exposure on OCT2 activity and membrane translocation.

Selective phosphorylation of AKT2 at T309 was involved in increased activity of OCT by Cd exposure

Consistent with previous reports,²⁴ we observed that Cd could dramatically increase the phosphorylation of AKT at T308/309, in both cells and mouse kidney (Figure 2a; Supplementary Figure S3). The extent of AKT phosphorylation was well aligned with the increase of OCT2 activity by Cd (Figures 1a and 2a). Thus, we hypothesized that the phosphorylation of AKT at T308/309 might be involved in Cd-mediated membrane translocation and increased activity of hOCT2. For evidence to support this possibility, we blocked Cd-mediated AKT phosphorylation at T308/309 (Figure 2b) by a selective inhibitor, MK-2206, which significantly reduced Cd-mediated increase of hOCT2 activity (Figure 2c) and its membrane expression (Figure 2d). Interestingly, our data showed that Cd-mediated AKT phosphorylation was isoform-dependent. Knockdown of AKT2, but not AKT1, dramatically reduced the level of Cd-mediated AKT phosphorylation (Figure 2e) and abolished the increased activity of hOCT2 by Cd exposure (Figure 2f). This finding was further validated by overexpressing the AKT2 T309A mutant (Supplementary Figure S4), which eliminated the inductive effect of Cd exposure on hOCT2 activity, as compared to wild-type AKT2 (Figure 2g). However, we noticed that Cd could subtly increase the expression of AKT1 in the presence of AKT1 small, interfering (si)RNA treatment, and that of AKT1 and total AKT in the presence of AKT2 siRNA. Whereas the exact reason warrants further investigation, it is likely that Cd exposure affects cellular AKT protein turnover under certain conditions. Altogether, our results indicate that AKT2 phosphorylation at T309 in HEK293 cells could be a critical molecular regulator involved in Cd-mediated hOCT2 membrane translocation.

Figure 2 |. Selective phosphorylation of serine/threonine-protein kinase AKT2 at T309 is involved in cadmium (Cd)-induced organic cation transporter 2 (OCT2) activation.

(a) Immunoblotting analysis of phosphorylated (p) and total AKT in human embryonic kidney (HEK)–human (h) OCT2 cells with and without exposure to various concentrations of Cd at various times. An abundance of pAKT at T308/309 was normalized by total AKT. (b) Immunoblotting analysis of phosphorylated and total AKT in HEK-hOCT2 cells with and without Cd exposure in the presence and absence of the AKT phosphorylation inhibitor MK-2206. (c) Cellular uptake of metformin in HEK-hOCT2 cells after being treated with or without MK-2206 (4 μM) for 10 minutes, followed by exposure with and without Cd for 20 minutes. (d) Immunoblotting analysis of total and membrane fraction of hOCT2. The HEK-hOCT2 cells were treated with or without MK-2206, followed by exposure with or without Cd for 20 minutes. An abundance of membrane protein was normalized by the total protein. (e) Immunoblotting analysis of phosphorylated and total AKT, AKT1, and AKT2 in HEK-hOCT2 cells, and (f) cellular uptake of metformin in HEK-hOCT2 cells. (e,f) Cells were transfected with scrambled small, interfering RNA (siRNA), AKT1 siRNA, or AKT2 siRNA, followed by exposure with or without Cd for 20 minutes. (g) Cellular uptake of metformin in HEK-hOCT2 cells. Cells were overexpressed with either wild-type (WT) or mutant AKT2 T309A, followed by exposure with or without Cd for 20 minutes. All figures are representative of 3 independent experiments. (b,e), β-actin was used as a loading control (Contr). (d) Pan-cadherin served as a positive Contr and β-actin as a negative Contr for membrane proteins. (c,f,g) Data are shown as mean ± SD. *P < 0.05, **P < 0.01, ^## P < 0.01, ***P < 0.001; ^###P < 0.001; (f) analysis of variance followed by post hoc Tukey analysis; (c,g): 2-tailed Student’s t test. DMSO, dimethylsulfoxide; n.s., not significant. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.

AS160 and Ca/CaM were involved in the increased activity of OCT2 by Cd exposure

The GTPase-activating protein (GAP) protein AS160 has been recognized as the downstream effector of AKT2 in the signal cascade of insulin-induced glucose transporter type 4 (GLUT4) translocation.²⁵ Similarly, we found that knockdown of AS160 (Figure 3a) significantly decreased the effect of Cd exposure on hOCT2 activity (Figure 3b), and abolished the increased membrane expression of hOCT2 at 75 KD due to Cd exposure (Figure 3c). However, an increase of the basal level of hOCT2 was also observed by AS160 knockdown, which might be attributable to disturbed hOCT2 trafficking, although the exact mechanism is unknown. Whereas the phosphorylation of AS160 is essential to translocation of GLUT4,^25,26 we observed that the phosphorylation of AS160 at T642 was not enhanced by Cd exposure (Figure 3d), nor was the overexpression of AS160 4P, a mutant that loses 4 major phosphorylation sites (Supplementary Figure S5), affected by the Cd-mediated increase of hOCT2 activity (Figure 3e). Thus, our data suggested that AS160 was involved in Cd-mediated hOCT2 membrane translocation in a manner different from that in insulin-induced GLUT4 translocation. Cd exposure has an impact on the Ca/CaM signaling pathway,²⁷ and there is a known CaM-binding site in AS160,^28,29 leading us to investigate whether Ca/CaM was involved in Cd-mediated AKT-AS160 signaling and increased hOCT2 activity. We found that the CaM inhibitor calmidazolium (CMZ) and knockdown of CaM (Figure 3h) caused a significant decrease in the level of increase of hOCT2 activity by Cd exposure (Figure 3f and i). The increased membrane expression of hOCT2 by Cd exposure was eliminated by CMZ treatment as well (Figure 3g). The involvement of Ca was also demonstrated. We showed that Cd led to a dose-dependent increase of the intracellular Ca signal (Supplementary Figure S6A), and that depletion of Ca eliminated Cd-mediated increase of hOCT2 activity (Supplementary Figure S6B). In addition, the effect of CaM knockdown was able to be rescued by overexpression of wild-type CaM protein, but not by the CaM 1234 mutant, which loses Ca binding capacity³⁰ (Figure 3j). Furthermore, we mutated the CaM-binding sites²⁸ (L842G/W843G, LW/GG) in both wild-type AS160 and the AS160 4P mutant (Supplementary Figure S7) and found that both AS160 mutants LW/GG and 4P+LW/GG in HEK-hOCT2 cells resulted in significant, similar decreases in the inductive effect of Cd on hOCT2 activity (Figure 3k). Overall, our results demonstrated that an interaction between AS160 and Ca/CaM played an essential role in a Cd-mediated increase of hOCT2 activity.

Figure 3 |. Calmodulin (CaM) and AKT substrate AS160 are key regulators involved in cadmium (Cd)-induced human organic cation transporter 2 (hOCT2).

(a) Immunoblotting (IB) analysis of AS160; (b) cellular uptake of metformin; (c) IB analysis of total and membrane fraction of hOCT2, after human embryonic kidney (HEK)–hOCT2 cells being transfected with scrambled and AS160 small, interfering RNA (siRNA). (b,c) Cells were exposed with and without Cd for 20 minutes after treatment with AS160 siRNA. (d) IB analysis of total and phosphorylated (p) AS160 after HEK-hOCT2 cells were exposed to Cd for 20 minutes. (e) Cellular uptake of metformin after HEK-hOCT2 cells were overexpressed with wild-type (WT) and 4P mutant of AS160, followed by Cd exposure for 20 minutes. (f) Cellular uptake of metformin; (g) IB analysis of total and membrane fraction of hOCT2, in HEK-HOCT2 cells after treatment with or without CaM inhibitor calmidazolium (CMZ, 5 μM) for 10 minutes, followed by Cd exposure for 20 minutes. (h) IB analysis of CaM; (i) cellular uptake of metformin, after HEK-hOCT2 cells were transfected with scrambled and CaM siRNA, followed by Cd exposure for 20 minutes. (j) Cellular uptake of metformin in HEK-hOCT2 cells after transfection with CaM siRNA, then rescue by overexpression of vector, wild-type (WT), and 1234 mutant of CaM, followed by Cd exposure for 20 minutes. (k) Cellular uptake of metformin in HEK-hOCT2 cells after overexpression with WT and 3 mutants (LW/GG, 4P, and 4P+LW/GG) of AS160, followed by Cd exposure for 20 minutes. (a,h) β-actin was used as a loading control (Cont). (c,g) Pan-cadherin served as a positive Cont and β-actin as a negative Cont for membrane proteins. Abundance of membrane protein was normalized by total protein. All panels in this figure are representative of 3 independent experiments. (b,e,f,i–k) Data are shown as mean ± SD. *P < 0.05, ^#P < 0.05, **P < 0.01, ^##P < 0.01, ***P < 0.001, ^###P < 0.001 (2-tailed Student’s t test). n.s., not significant. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.

Cd-mediated AKT phosphorylation at T308/T309 was both Ca/CaM and AS160 dependent

We further investigated whether AKT phosphorylation is affected by AS160 and Ca/CaM. Intriguingly, we found that Cd-mediated AKT phosphorylation could be dose-dependently decreased by CMZ (Figure 4a) and by the treatment of BAPTA-AM, an intracellular Ca²⁺ chelator (Supplementary Figure S8A) in cells, indicating a regulatory effect by Ca/CaM on AKT phosphorylation. The role of Ca/CaM was further evidenced by the knockdown of CaM or overexpression of CaM 1234, in which Cd-mediated AKT phosphorylation at T308/T309 was significantly lower than that in the scrambled control or overexpression of wild-type CaM, respectively (Figure 4b; Supplementary Figure S8B). Likewise, we found that AS160 was also required for the full Cd-mediated AKT phosphorylation at T308/T309 (Figure 4c). In particular, we observed that overexpression of the AS160 mutant (LW/GG; Supplementary Figure S7) that was deficient of the CaM-binding site could significantly reduce AKT phosphorylation as well (Figure 4d), suggesting an important role of CaM-AS160 interaction in Cd-mediated AKT phosphorylation. In addition, a co-immunoprecipitation assay showed that Cd exposure could substantially enhance the binding of both AKT and CaM to AS160, which was abolished by overexpression of the AS160 LW/GG mutant (Figure 4e) and by CMZ (Figure 4f). Thus, Cd-mediated binding of AKT with AS160 was likely dependent on the binding of CaM with AS160. Taken together, our results supported a mechanistic model in which Cd exposure could trigger the formation of a complex of AKT, CaM, and AS160. The formed complex may facilitate the phosphorylation of AKT2 at T308/309, which serves as a switch for the release of vesicles containing membrane proteins such as OCT2 from AS160 (Figure 4g; see Discussion).

Figure 4 |. Cadmium (Cd)-induced serine/threonine-protein kinase (AKT) phosphorylation at T309 is both calmodulin (CaM)-and AKT substrate AS160–dependent.

(a) Immunoblotting analysis of total and phosphorylated (p) AKT in human embryonic kidney–human organic cation transporter 2 (HEK-hOCT2) cells after being treated with and without calmidazolium (CMZ) for 10 minutes, followed by exposure with and without Cd for 20 minutes. (b–d) Immunoblotting analysis (left) and quantification (right) of total and pAKT in HEK-hOCT2 cells, after being (b) transfected with scrambled and AS160 small, interfering RNA (siRNA), (c) transfected with scrambled and CaM siRNA, or (d) transfected with wild-type (WT) and the LW/GG mutant of AS160, followed by exposure with and without Cd for 20 minutes. β-actin was used as a loading control. (e,f) Co-immunoprecipitation (co-IP) analysis (left) and quantification (right) of the interaction between AKT, CaM, and AS160 in HEK-hOCT2 cells (e) overexpressed with WT and the LW/GG mutant of AS160, followed by treatment with and without Cd for 20 minutes, or (f) overexpressed with WT of AS160, then treated with and without CMZ, followed by exposure with and without Cd for 20 minutes. (g) The model of Cd-induced translocation of hOCT2 into the membrane for expression. All panels in this figure are representative of 3 independent experiments. (b–f) Data are shown as mean ± SD. *P < 0.05, **P < 0.01 (2-tailed Student’s t test). Cont, control; PIP3, phosphatidylinositol (3,4,5)-trisphosphate. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.

Cd exposure exacerbated cisplatin nephrotoxicity by increasing cisplatin accumulation in mouse kidney

As a well characterized uptake transporter for cisplatin, OCT2 is a critical determinant of the intracellular concentration of cisplatin and its nephrotoxicity in both mice and humans.^17,18 Thus, we examined whether Cd exposure would exacerbate cisplatin nephrotoxicity by enhancing cisplatin accumulation in mouse kidney. As shown above, renal Cd exposure via CdMT injection significantly increased metformin accumulation (Figure 1g), and expression of mOct2/1 protein in mouse kidney (Figure 1h), in contrast, had no effect on the expression of copper transporter 1 (Supplementary Figure S9), another important renal uptake transporter for cisplatin.³¹ CdMT also enhanced the phosphorylation of Akt at T308/309 in a dose-dependent manner (Figure 5a) and significantly increased the accumulation of cisplatin in mouse kidney (Figure 5b). The inductive effect of Cd exposure on hOCT2-mediated cisplatin uptake was also confirmed in a cellular uptake assay (Figure 5c). Both CdMTand cisplatin alone caused no obvious renal damage histologically (Figure 5d). In contrast, necrosis containing eosinophilic amorphous material and pyknotic debris was found in the group that received the combined treatment of CdMT plus cisplatin (Figure 5d). Both CdMTand cisplatin alone resulted in only minor renal injury, indicated by elevated expression of Kim-1 and Lcn2, two sensitive molecular biomarkers of renal injury.^32,33 In contrast, dramatically increased expression of Kim-1 and Lcn2 were seen in the mice that received the combined treatment, compared with the other 3 groups (Figure 5e).

Figure 5 |. Cadmium (Cd) exposure exacerbates cisplatin nephrotoxicity by increasing its accumulation in mice kidney.

(a) Immunoblotting analysis of total and phosphorylated Akt in the kidney after mice being administrated with Cd-metallothionein (CdMT; i.p.) for 6 hours. (b) Quantification of cisplatin in the kidney after mice were treated with CdMT (100 μg Cd/Kg, i.p.) for 6 hours, then given a single dose of cisplatin (4 mg/kg, i.p.) for 3 hours. n = 5 per group. (c) Cellular accumulation assays for cisplatin in human embryonic kidney–human organic cation transporter 2 (HEK-hOCT2) cells after exposure with and without Cd for 20 minutes. Data are representative of 3 independent experiments. (d) Representative images of hematoxylin and eosin–stained kidney sections from mice in the following groups: Control (cont), CdMT (100 μg Cd/Kg, i.p.), cisplatin (10 mg/kg, i.p.), and CdMT (100 mg Cd/Kg, i.p.) + cisplatin (10 mg/kg, i.p.). Mice were treated first with CdMT, then with cisplatin 6 hours later, and killed 72 hours after the administration of cisplatin. (e) Quantitative polymerase chain reaction analysis for the expression of Kim-1 and Lcn2 in mice kidney from the indicated groups shown in (d). (f) Representative images of hematoxylin and eosin–stained kidney sections from mice in the following groups: cont, Cd (10 ppm for 10 weeks in drinking water), cisplatin (10 mg/kg, i.p.), and Cd (10 ppm for 10 weeks in drinking water) + cisplatin (10 mg/kg, i.p.). Mice were treated with Cd (10 ppm) in drinking water for 10 weeks, then with a single i.p. dose of cisplatin, and killed 72 hours after the administration of cisplatin. n = 4 or 5 per group. (g) Quantitative polymerase chain reaction analysis for the expression of Kim-1 and Lcn2 in mouse kidney from the indicated groups shown in (f). (h) Quantification of cisplatin concentrations in the kidney of mice from the indicated groups shown in (f). (d,f) A tubulointerstitial injury score was examined and compared. Black arrows indicate necrosis containing eosinophilic amorphous material and pyknotic debris. Enlarged views are shown in every image. Bar = 200 μm. (e,g) Gene expression was normalized to Gapdh mRNA levels. (c) Data are presented as mean ± SD. (b,e,g,h) Data are shown as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 (e,g: analysis of variance with Tukey post hoc analysis; b,c,h: 2-tailed Student’s t test). ns, not significant. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.

To further mimic Cd exposure in the general human population, mice were also exposed chronically with and without Cd (10 ppm) in drinking water for 10 weeks. These mice were then given a single dose of cisplatin (10 mg/kg, i.p.) or vehicle control treatment. Similar to the above study with CdMT treatment, both Cd (10 ppm) exposure and cisplatin treatment alone caused no renal damage histologically, whereas apparent necrosis was found in the kidney of Cd-exposed mice that received cisplatin treatment (Figure 5f). Likewise, the level of Cd (10 ppm) exposure alone had no effect on the expression of Kim-1 and Lcn2; in contrast, it substantially increased their expression when mice were co-treated with cisplatin (Figure 5g). Similar to the CdMT study results, a significant increase of cisplatin accumulation in mouse kidney was observed as well (Figure 5h). Altogether, our results indicate that Cd exposure can exacerbate the nephrotoxicity of cisplatin by enhancing its accumulation in mouse kidney.

The nephrotoxicity of cisplatin was correlated with the plasma level of Cd in cervical cancer patients

To extend the cellular and mouse studies further, we conducted a clinical study to examine the correlation between Cd exposure and cisplatin nephrotoxicity in cervical cancer patients who received chemotherapeutic cisplatin. We found that the Cd level in the plasma was significantly correlated with the change of creatinine after cisplatin chemotherapy (R = 0.2978; P = 0.0496; Figure 6a). A median plasma Cd level of 0.099 ppb with the interquartile range of 0.094–0.124 ppb has been reported for a general population.³⁴ We thus divided our patients into 3 groups according to their plasma Cd levels (ppb): low (≤0.080), medium (0.080–0.12), and high (≥0.12). We performed multivariate linear analysis of different factors to determine their correlation to the serum Cd level, and we found that the “change of creatinine” was the only significantly dependent variable (Table 2). Of note, no differences in demographic or other clinical characteristics, including the dose of cisplatin, were found among the groups. In addition, when the estimated glomerular filtration rate was compared, the trend of renal function loss became clearer as the plasma level of Cd increased (Figure 6b).

Figure 6 |. Cadmium (Cd) exposure is correlated with cisplatin nephrotoxicity in cervical cancer patients.

(a) Linear regression analysis of the correlation between the Cd level and the change of creatinine in cervical patients’ plasma pre- and postcisplatin chemotherapy. (b) Estimated glomerular filtration rate (eGFR) pre- and postcisplatin chemotherapy in low– (≤0.08 μg/L), medium– (0.08 ~ 0.12 μg/L) and high– (≥0.12 μg/L) Cd exposure group. (a,b) Actual P values are given to each panel in this figure. Each circle, square, and triangle represents an individual patient. (b) Data are shown as mean ± SEM. *P < 0.05. (b) Two-tailed Student’s t test.

Table 2 |.

Multivariate linear analysis of dependent variables in relation to the serum cadmium (Cd) level in cervical cancer patients

Dependent variables	P value	95% CI			Multiple comparisons between groups (P value)
		Serum Cd level (μg/L)
		Low (≤0.08, n = 7)	Medium (0.08–0.12, n = 21)	High (≥0.12, n = 16)	Low vs. medium	Low vs. high	Medium vs. high
Change of creatinine (μmol/L)	0.038a	(−3.41, 3.40)	(0.187, 4.72)	(2.86, 6.83)	0.119	0.015a	0.085
Age (yr)	0.903	(46.9, 59.1)	(51.0, 59.1)	(47.3, 58.4)	0.656	0.772	0.848
Body weight (kg)	0.893	(50.2, 66.1)	(52.5, 60.6)	(54.2, 61.0)	0.678	0.885	0.727
Height (cm)	0.722	(155.6, 160.7)	(154.1, 158.7)	(155.0, 159.2)	0.423	0.545	0.819
Dose of cisplatin (mg)	0.397	(39.5, 65.2)	(54.0, 62.7)	(46.1, 59.0)	0.395	0.904	0.202

CI, confidence interval; n, sample size.

^aP < 0.05.

DISCUSSION

Environmental exposure to Cd predominantly accumulates in kidney and has been linked to various kidney diseases. However, little is known about the effect of this toxic metal on renal drug disposition and response. In this study, we conducted a comprehensive examination of the effect of Cd exposure on the renal drug transporter OCT2, and thus the disposition and response of OCT2 substrates. We established Cd as an activator of OCT2 translocation to the membrane both in vitro in cells and in vivo in mice. We mechanistically unraveled that the Cd-mediated hOCT2 translocation was dependent on the formation of a protein complex consisting of AKT2, Ca/CaM, and AS160. For the first time, the clinical implication of Cd exposure in terms of drug disposition and response was exemplified by findings that Cd exposure led to increased accumulation and aggravated nephrotoxicity of the OCT2 substrate cisplatin in mice. Furthermore, a significant correlation between Cd exposure and cisplatin nephrotoxicity was found in cervical cancer patients. Collectively, our studies demonstrate that Cd exposure may increase the activity of OCT2 by enhancing its trafficking to cell membrane, leading to altered renal drug disposition and response mediated by this transporter protein.

Although substantially different, the mechanism underlying the Cd-mediated increase of OCT2 activity is reminiscent of that for insulin-stimulated GLUT4 translocation. We demonstrated that AKT2 phosphorylation at T309 was an essential step involved in Cd-mediated OCT2 translocation, which is critical to insulin-stimulated GLUT4 translocation in adipose and muscle cells as well.^35,36 However, in contrast to the selective phosphorylation of AKT2 at T309 by Cd exposure, insulin-mediated increase of the secondary messenger PIP3 in cell membrane non-selectively activates both of the AKT isoforms, AKT1 and AKT2, at both S473/474 and T308/309.^37,38 The difference between Cd and insulin was further evidenced when they were compared in our in vitro and in vivo studies (Figure 1a, e, and g; Supplementary Figure S10A and B), in which insulin treatment had little efffect on metformin transport mediated by the activities of OCTs.

We found that AS160, a critical downstream effector of the insulin-AKT2-GLUT4 pathway, was also involved in the Cd-mediated increase of OCT2 activity. In insulin signaling transduction, the phosphorylated AKT2 could further phosphorylate AS160 at its phospho-Akt substrate (PAS) motifs, leading to disassociation of AS160 from Rab proteins residing in vesicles containing GLUT4.^25,39 The released Rab protein would be activated to bring the vesicles of GLUT4 to the cell membrane. Nevertheless, the specific role of AS160 in the Cd-mediated increase of OCT2 activity seemed to be different. Cd exposure did not enhance the phosphorylation of AS160 at the PAS motif, and Cd-mediated hOCT2 translocation was not able to be eliminated by overexpression of the PAS mutant (AS160 4P). Subsequently, we found that the binding of CaM to AS160 was crucial in the Cd-mediated increase of OCT2 activity. Notably, the involvement of CaM has been implicated in the contraction-stimulated GLUT4 translocation in skeletal muscle cells.^28,29 However, the PAS motifs remain necessary for the full activation of contraction-stimulated GLUT4 translocation, but they are not required in Cd-mediated OCT2 translocation. More than 60 Rab protein members have been discovered in humans,⁴⁰ many of which have been confirmed by an interaction with AS160.⁴¹ Thus, it is possible that different Rab proteins are present in the vesicles of GLUT4 and OCT2, and that the PAS phosphorylation and/or Ca/CaM binding would be required for activation of Rab proteins associated with GLUT4 vesicles.²⁹ In contrast, binding of Ca/CaM to AS160 might be crucial for the activation of Rab proteins associated with OCT2 vesicles. On the other hand, our data indicate that AS160 might not be the only TBC (Tre-2/Bub2/Cdc16) protein involved, since both knockdown of AS160 and overexpression of its mutants, the LW/GG and 4P+LW/GG, could not fully eliminate a Cd-mediated increase of OCT2 activity. Actually, more than 40 different TBC proteins are present in humans and mice,⁴¹ many of which could function as a specific Rab GTPase-activating protein. Further characterization of the specific TBC/Rab proteins involved in Cd-mediated OCT2 protein trafficking is thus warranted.

We found that Cd exposure could enhance the binding of AKT2 and CaM to AS160, and that Cd-mediated phosphorylation of AKT2 at T309 was both Ca/CaM and AS160 dependent. Our data indicate that AKT2 is the major AKT isoform involved, although we could not fully rule out AKT1. The role of CaM in facilitating AKT phosphorylation is consistent with the results of previous studies,^42,43 in which CaM was proposed to bind to AKT proteins in the conserved CaM-binding motif and to direct the complex to plasma membrane for AKT phosphorylation. Based on our findings, we proposed that AS160, Ca/CaM, and AKT2 form a protein complex after Cd exposure. The formed complex might move to an area adjacent to PIP3-enriched cell membrane, where AKT2 was phosphorylated. Thereafter, the phosphorylation of AKT2 probably resulted in critical conformation change in the complex, leading to disassociation between AS160 and a Rab protein, which eventually initiated the translocation of OCT2 vesicles to cell membrane (Figure 4g). Further work is needed to delineate the details of our model. Given that only HEK cells were used in our study, it is important and necessary to verify our model in additional cell models and in vivo. In addition, it is of great interest to determine how the complex formation is initiated by Cd exposure. Although Ca was required in Cd-mediated increase of OCT2 activity, artificially increased levels of Ca in the cytoplasm by the ionophore A23187 did not have any effect on OCT2 activity (Supplementary Figure S11). It has been reported that Cd ions could displace Ca ions and bind to the Ca-binding pockets of CaM.⁴⁴ Hence, it is possible that a Ca-CaM-Cd complex was formed, which had a relatively high affinity to AKT2 and initiated the formation of the protein complex consisting of CaM, AKT2, and AS160, leading to the selective phosphorylation of AKT2. Future studies are needed to determine the mechanism in detail.

Cd exposure has been reported to be in the range of 10–60 μg/day (in “uncontaminated” European areas) through 150–200 μg/day (in contaminated areas in Japan) to 1850 μg/day (in China, mainly from rice with high Cd levels).^5,45 Even though 90%–95% of ingested Cd remains unabsorbed,⁴⁵ Cd exposure is unfortunately cumulative, with an extremely long biological half-life of 10–30 years in humans.^4,46 More than a dozen studies with autopsy tissues have reported that non-occupationally exposed individuals may accumulate a significant amount of Cd in the kidney (e.g., 0.62–61.3 μg/g tissue, or ~5.5–545 μM).⁵ In our animal studies, both acute CdMT treatment and 10 weeks of 10 ppm Cd exposure via drinking water led to mouse renal Cd accumulation (from ~5.0 to ~15.6 μg/g tissue; Supplementary Figure S12) in the concentration range of those in human renal tissues. The Cd concentrations causing OCT function changes in our cellular studies and animal studies were clinically relevant to those reported for human kidney, an organ highly involved in drug disposition. Although the human plasma Cd level is much lower than those in many tissues such as the kidney, it can reflect exposure.³⁴ In our clinical study, the plasma Cd level was significantly correlated with the change of creatinine level after cisplatin chemotherapy. These findings at different levels strongly suggest that Cd exposure may lead to clinically significant alteration in renal drug disposition.

Our studies in mice have clearly demonstrated that the nephrotoxicity of cisplatin could be dramatically exacerbated by Cd exposure. Although it is a known caveat that Cd itself can cause injury to renal tubular cells, we only observed limited damage under the current dose of CdMT alone, which is consistent with findings from other studies.^47,48 The profound deterioration in histology and the dramatic increase in expression of the molecular biomarkers of renal injury observed in the mice that received both CdMT and cisplatin treatment might be attributed largely to the significantly increased uptake of cisplatin mediated by Oct transporters in the kidney. Our animal findings also support the observed correlation between cisplatin-induced renal function loss and a higher plasma Cd level in cervical cancer patients. The multivariate analysis of the clinical data demonstrated that renal function loss, rather than other factors, was correlated with serum Cd level in these patients. Of note, the small sample size of this retrospective clinical study might limit the power to detect the difference, as shown by a marginal significance (P = 0.0496; Figure 6a) for the correlation analysis and by non-significance when the estimated glomerular filtration rate was compared between pre- and postcisplatin treatment; however, the P value drops from 0.95 in the low plasma Cd group to 0.09 in the high plasma Cd group (Figure 6b).

In human kidney, 2 primary transporters are now known to be involved in transporting cisplatin into the tubular cells—OCT2 and Ctr1.^17,31 Of note, supporting data at the clinical level are available only for OCT2.^17,18 In rodent kidney, in addition to Oct2, Oct1 is also involved in the basal uptake of cisplatin into the tubular cells.⁴⁹ On the apical side, consistent findings have implicated a role by MATE transporters in the secretion of cisplatin into the urine.¹⁹ Whereas the present study focuses on the primary renal transporter system OCT/MATE for cisplatin transport, and thus the interaction between Cd and cisplatin in causing nephrotoxicity, it is likely that Cd exposure affects the activities of additional membrane transporters. Cd exposure seemed not to affect the overall expression of Ctr1 (Supplementary Figure S9); however, its effects on the function and protein trafficking of this particular transporter remain unknown. We are currently assessing, with proteomic approaches and individual transport assays, the impact of Cd exposure on the expression of membrane proteins and the broad function of multiple transporters in the kidney. It should also be noted that the transport mechanisms for Cd itself are very interesting but very complicated. Our recent findings and those from others have indicated that OCT2 and MATEs are Cd ion transporters.^20–22 However, given their moderate affinity toward Cd ions, they are probably not the primary transport mechanisms for Cd. Other mechanisms or transporters, such as the receptor-mediated endocytosis of the Cd²⁺-protein complex, calcium channels, zinc transporters, transient receptor potential channels, diva-lent metal-ion transporters, amino acid/cystine transporters, and the adenosine triphosphatase multidrug resistance protein 1 (MDR1), have been implicated in transporting Cd across the membrane.^50,51 Future research is needed to examine whether Cd exposure affects these transport mechanisms and the homeostasis of Cd itself.

In conclusion, the present study demonstrates that Cd exposure may increase the activity of OCT2 by enhancing its trafficking to the membrane, and exacerbate cisplatin nephrotoxicity by increasing its renal accumulation mediated by the transporter proteins. Our findings provide a foundation to further uncover Cd exposure as a potential determinant of the broad variation in drug disposition and response.

METHODS

Cellular accumulation assay

Metformin and cisplatin were used in the cellular uptake studies in HEK-hOCT2 and HEK-hMATE1 cell lines. The uptake times of metformin and cisplatin in HEK-hOCT2 were set to 20 seconds and 30 minutes, respectively, according to previous studies.^52,53 The same protocol of metformin uptake in HEK-hMATE1 was used as described in our previous publication.²²

Animal studies

The wild-type mice (C57BL/6J background) were used at 10–16 weeks of age. For ex vivo accumulation studies, slices of mice kidneys were incubated with metformin for 5 minutes after being exposed with and without Cd for 30 minutes. The radioactivity was measured by the liquid scintillation analyzer and normalized by the weight of each slice. For in vivo pharmacokinetic studies, metformin or cisplatin was injected into mice 6 hours after CdMT administration. Tissue accumulation of metformin and cisplatin was determined by a liquid scintillation analyzer and an inductively coupled plasma mass spectrometer (Agilent 7700, Santa Clara, CA), respectively. For in vivo pharmacologic and toxicologic studies, mice were given either a single dose of CdMT or vehicle, or treated with and without Cd in drinking water for 10 weeks. Cisplatin or vehicle was administered 6 hours after CdMT injection, or right after 10 weeks of Cd exposure. Kidney samples were harvested 72 hours later for histology, protein, and mRNA analysis.

Clinical studies

Cervical cancer patients were confirmed with cytologic and histologic diagnoses and treated with chemotherapeutic cisplatin for 5 to 8 cycles. The patient charts were reviewed retrospectively to collect creatinine data pre- and postcisplatin chemotherapy. One venous blood sample was collected for each patient for quantification of Cd by inductively coupled plasma mass spectrometry.

Details on the methods and materials used in the present study are given in the Supplementary Materials and Methods.

Translational Statement.

Cadmium (Cd) is an environmentally prevalent toxic metal posing increasing risk to human health worldwide. Our study took a first step to understand the impact of Cd exposure on drug disposition and response. Specifically, the present study has demonstrated that Cd exposure increases the activity of important renal organic cation transporters and exacerbates the nephrotoxicity of cisplatin by increasing its renal accumulation mediated by these transporters. Clinicians should be aware that Cd exposure might serve as a determinant for the variation in the disposition of and response to clinically important drugs such as cisplatin, and try to avoid Cd–drug interaction.