The TRPM6/EGF Pathway Is Downregulated in a Rat Model of Cisplatin Nephrotoxicity

Cisplatin-induced hypomagnesemia is described in humans and rats, but the underlying mechanisms are still unclear. Recent studies have shown that epidermal growth factor (EGF) stimulates Mg2+ re-absorption in the distal convoluted tubule via the Mg2+ channel TRPM6. This study investigates the role of TRPM Mg2+ channels, claudines, and EGF in the Mg2+ homeostasis in a rat model of cisplatin-induced nephrotoxicity. Wistar rats were given 2.5 mg/kg cisplatin per week for 3 weeks and were euthanized 4 or 9 weeks after the first administration. The cisplatin treatment significantly increased the fractional excretion of Mg2+. Real-time RT-PCR and/or Western blots were performed to assess the renal expression TRPM6, TRPM7, claudin-16, claudin-19, EGF, EGF receptor (EGFR) and EGFR-pathway components. The renal mRNA expression of TRPM6 and EGF showed a significant decrease after cisplatin treatment, while the TRPM7, claudin-16 and EGFR expressions remained stable. The claudin-19 mRNA expression was significantly upregulated after cisplatin treatment. Western blotting confirmed the mRNA expression data for the claudins, but an showed upregulation of EGFR only at week 9. The role of the EGFR pathway, involving Pi3-AKT-Rac1, in cisplatin-induced nephropathy, could not be substantiated in further detail. This study shows that cisplatin treatment results in EGF and TRPM6 downregulation in the rat kidney, causing renal Mg2+ loss. Our results are in line with the hypothesis that EGF influences the renal expression or activation of TRPM6 and plays a significant role in Mg2+ loss in medication-induced nephropathy.


Introduction
About 43% of the patients treated with cisplatin develop hypomagnesemia due to renal magnesium (Mg 2+ ) loss [1]. Moreover, it was shown that acute MgCl 2 infusions lead to a significantly higher urinary Mg 2+ excretion in cisplatin treated rats and humans [2,3]. In rats, the Mg 2+ depletion enhances cisplatininduced nephrotoxicity, significantly increasing plasma creatinine and plasma urea [4]. However, the mechanisms underlying these observations have not yet been described.
The major site of passive Mg 2+ re-absorption is the thick ascending limb (TAL) where 70% of the Mg 2+ is reabsorbed. The tight junction proteins claudin-16 (also known as paracellin-1) and claudin-19 are the key players in this paracellular transport [5,6].
Recently, two other ion channels with an important role in the Mg 2+ homeostasis were identified, TRPM6 and TRPM7. They belong to the transient receptor potential subfamily Melastatin (TRPM). TRPM6 is predominantly expressed in absorbing epithelia. In the kidney, TRPM6 is present in the distal convoluted tubule (DCT), known as the main site of active transcellular Mg 2+ re-absorption along the nephron. TRPM7 is ubiquitously expressed and implicated in cellular Mg 2+ homeostasis, cell death, and cell cycle regulation [7,8].
Our research group and others reported that hypomagnesemia in calcineurin inhibitor (CNI)-induced nephropathy is related to the downregulation of epidermal growth factor (EGF) and TRPM6 [9,10,11]. Recently, it was shown that EGF stimulates Mg 2+ re-absorption in the DCT via its receptor on the basolateral membrane and via activation of TRPM6 in the apical membrane. The EGF-mediated stimulation of TRPM6 occurs through signaling via SRC kinases and Rac1 in vitro, thereby redistributing endomembrane TRPM6 to the plasma membrane [12,13]. This pathway has not been studied thoroughly in cisplatin-induced nephropathy although it was shown that cisplatin administration results in a decreased renal expression of EGF and a decreased urinary EGF excretion in the rat [14].
This study aimed to unravel the molecular mechanisms of cisplatin-induced hypomagnesemia. The expression profiles of several proteins involved in the Mg 2+ re-absorption, i.e. TRPM6, TRPM7, claudin-16, claudin-19 and EGF, were analyzed as well as the interaction of TRPM6 with EGF. Moreover, the role of the EGF receptor (EGFR) and its signaling pathway was studied through the expression analysis of several signaling proteins.

Animal model
Twenty-four male Wistar rats (Charles River, Brussels, Belgium), weighing between 180 and 200 g, were housed two to four per cage, at 22 6 2uC with a 12h dark-light cycle. They had free access to water and were fed a normal diet (0,30% Mg, Carfil Labofood, Oud-Turnhout, Belgium).
The ethical approval for the animal studies was obtained from the Medical Ethical Committee on Animal Experimentation of the University of Antwerp, Belgium (file 2009/17).

Experimental set-up
Rats were divided into two groups (each containing twelve animals) and treated as follows.
Group 1 (control rats): the rats received a weekly intraperitoneal injection of saline in a volume equal to the cisplatin-treated group during 3 weeks.
Group 2 (cisplatin-treated rats): the rats received a weekly intraperitoneal injection of cisplatin (2.5 mg/kg/w, Cisplatin Hospira 100 mg/100 ml ONCO-TAIN H , Hospira, Warwickshire, UK) during three weeks. This dose of cisplatin and the three-week time interval between the first administration and the euthanasia result in minimal tubular damage and allows a focus on the functional effects in the rat model. A previous study showed that this cisplatin treatment scheme causes hypomagnesemia in the rat [2].
This experiment was performed twice to obtain data 4 or 9 weeks after the first cisplatin administration.

Sample collection
One week before the first cisplatin administration, the rats were caged individually in a metabolic cage to obtain a 5-hour urine collection. At the same time, 0.6 ml blood was collected from the tail vein. Twenty-eight days after the first cisplatin/saline administration, the rats were caged individually to obtain a second 5-hour urine collection. Afterwards, six rats of each group were euthanized by an overdose of pentobarbital (100-150 mg/kg) intraperitoneally. Blood samples were taken from the vena cava inferior at the time of sacrifice. The kidneys were quickly prelevated and decapsulated. The right kidney was snap-frozen in liquid nitrogen and stored at -80uC until the RNA extraction was performed. The left kidney was cut into 2 mm thick transverse slices and processed for further histological analysis using different fixation procedures (described below). To evaluate the long term effect of cisplatin, the six remaining rats of each group were sacrificed 63 days after the first dose of cisplatin.
Serum and urine creatinine, Mg 2+ , Na + and K + levels Serum and urine creatinine, Mg 2+ , Na + and K + levels were measured using an indirect potentiometric method with a dimension Vista 1500 System (Siemens Healthcare Diagnostics, Deerfield, USA).
The FE of Na + was calculated as: FE Na = 1006 (U Na 6P Cr )/ (P Na 6U Cr ). With U Na urinary excretion of Na + (mmol/l), P Cr plasma creatinine (mg/dl), P Na plasma Na + (mmol/l) and U Cr urinary excretion of creatinine.

Morphological assessment of cisplatin-induced lesions
For light microscopy, the rat kidney slices were fixed in formol during 24h, embedded in paraffin, cut in 5 :m-sections and stained with periodic acid-Schiff reagent or Sirius red collagen stain. Sirius red-stained sections were scanned with an Olympus BX61 Motorized Research Microscope (Olympus Corporation, Tokio, Japan) equipped with AnalySIS pro 5.0 software (Olympus). After the processing of the image, the total amount of collagen was measured and expressed as the percentage of the total cortex.

Quantification of inflammatory cells
The formol-fixed, paraffin-embedded renal tissue was stained with the ED1 monoclonal antibody (Acris Antibodies, Hiddenhausen, Germany), directed to an antigen of tissue macrophages and peripheral blood granulocytes. Infiltration was quantified by counting immunoreactive cells in 28 randomly chosen areas (magnification6200) of the cortex [15]. An observer, blinded for the experimental groups, examined 30 glomeruli per renal section. Inflammatory cells contained within large blood vessels or surrounded by erythrocytes were excluded from all counts. The results were expressed as absolute numbers of immunoreactive cells per mm 2 or per glomerular cross section.

mRNA isolation and cDNA synthesis
Total RNA was extracted from the kidneys using a columnbased technique (RNeasy Minikit, Qiagen, KJ Venlo, Netherlands). Purified total RNAs were treated with DNase to obtain DNA-free RNA (Turbo DNase free, Ambion, Applied Biosystems, Lennik, Belgium). cDNA was synthesized using the Transcriptor First Strand cDNA Synthesis Kit (AMV) (Roche Applied Science, Indianapolis, IN, USA).

Real-time PCR
To examine the mRNA expression of TRPM6, EGF, EGFR, Pi3, Akt and Rac1, real-time PCRs were performed using the LightCycler Taqman Master (Roche, Vilvoorde, Belgium). TGFb, PAI-1, TRPM7, claudin-16 and claudin-19 mRNA expressions were examined using the LightCycler FastStart DNA Master plus SYBR green I (Roche, Vilvoorde, Belgium). The PCRs were performed as previously described [9]. All genes were normalized against the housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH), which has been described as a stable reference gene for renal tissue in this experimental set-up [16] and expressed as the normalized ratio (NR). The renal sodium chloride channel (NCC) mRNA expression was determined to evaluate the distal tubule damage.

Statistics
The results are presented as a mean 6 SD. The differences between the cisplatin-treated groups and the controls at a specific time point (4 or 9 weeks) were analyzed using generalized linear models, which included 'group' and 'time after treatment' as predictors for each different dependent variable. The correlation between TRPM6 mRNA and EGF mRNA was measured using a Pearson correlation coefficient. Statistical analysis was performed using SPSS (version 20.0) for Windows. P-values less than 0.05 were considered statistically significant.

Functional parameters (table 1)
No rats died during the experiment. All experimental groups showed similar body weights and laboratory values (serum creatinine, creatinine clearance, serum Mg 2+ , FE Mg 2+ , serum Na + , serum K + and FE Na + ) at the start of the study (data not shown). At the time of sacrifice, the mean body weight in the control groups was significantly higher than in the cisplatin-treated groups (p,0.05). The serum creatinine values were significantly higher in the cisplatin-treated group after 4 weeks of treatment (p = 0.004) and after 9 weeks (p = 0.003) versus controls. The serum Mg 2+ concentration did not differ significantly between the cisplatin-treated groups and the controls. The FE Mg 2+ was significantly higher in the cisplatin-treated group after 4 weeks (p,0.001) and was increased after 9 weeks versus the controls, although not significantly (p = 0.082). The serum Na + , serum K + and the FE Na + were comparable in all groups.

Morphological assessment of cisplatin-induced lesions (Figure 1)
In each cisplatin-treated group, one rat showed signs of increased interstitial fibrosis on PAS-stained sections. However, the experimental groups showed no difference in interstitial fibrosis: the mean score of interstitial fibrosis was 13.87 6 5.25% in cisplatin-treated rats at 4 weeks versus 11.58 6 1.44% in the control group (p = 0.343). At 9 weeks, the mean score of interstitial fibrosis was 15.10 6 7.52% in the cisplatin-treated rats versus 10.74 6 1.81% in the control group (p = 0.221).

Quantification of inflammatory cells (Figure 2)
The control rats showed very few ED1-positive cells in the cortex. A significant infiltration of ED1-positive cells was seen in the cortex of the rats treated with cisplatin at 4 weeks (p = 0.016). This was not longer the case at week 9 (p = 0.084).

Real-time RT-PCR
The renal mRNA expression of TRPM6 ( Figure 3 The renal mRNA expression of EGF correlated significantly with the TRPM6 mRNA expression (p = 0.001). As shown in    The groups showed no significant differences in renal mRNA expression of TRPM7 (data not shown) and claudin-16, whereas claudin-19 showed a tendency to increase although no significance was reached (p = 0.144 at 4 weeks and p = 0.530 at 9 weeks, Figure 5). The renal mRNA expression of EGFR did not differ between the cisplatin-treated and the controls (table 2).
Pi3, Akt and Rac1 mRNA expression showed no differences at week 4 while their expression was significantly upregulated in the cisplatin-treated rats versus controls at week 9 (p,0.001, p,0.001 and p = 0.028 respectively, table 2).

Western blotting
Western blot analysis for TRPM6 and TRPM7 could not be executed due to the unavailability of the appropriate specific primary anti-rat antibodies.
The EGFR protein expression did not differ at 4 weeks, but was significantly increased at 9 weeks (p = 0.013, table 2). Pi3 was significantly upregulated at 4 weeks versus controls (p = 0.006), but not at 9 weeks (p = 0.661, table 2). Akt1 was significantly upregulated in the cisplatin groups at 4 weeks (p,0.001) and at 9 weeks (p = 0.017, table 2). Rac1 protein expression was stable in all groups at 4 weeks and at 9 weeks (table 2).

Discussion
This in vivo study shows that cisplatin downregulates the EGF and TRPM6 mRNA levels while the FE Mg 2+ increases and therefore suggests that cisplatin treatment results in a diminished renal Mg 2+ re-absorption via the downregulation of the Mg 2+ channel TRPM6.
The rat model showed an increased serum creatinine and an increased FE Mg 2+ after 4 weeks of treatment, conform to what has been seen in other studies [2,17]. Our rat model showed the pathogenic characteristics of cisplatin nephrotoxicity: the profibrotic factor TGF-b was upregulated as previously shown [18]. The same was true for PAI-1, which also plays a role in the progression to fibrosis [19]. In addition to this pro-fibrotic environment, a pro-inflammatory environment with an infiltration of ED1-positive cells in the interstitium was seen at week 4. However, the measurement of the total amount of collagen showed no increase in interstitial fibrosis after cisplatin treatment. Therefore, the kidney damage remained limited to a proinflammatory pro-fibrotic condition.
This study focused on the role of TRPM6 and EGF in cisplatininduced renal Mg 2+ loss. A simultaneous downregulation of the renal TRPM6 and EGF mRNA expression was found together  Data are presented in 4 groups: control animals receiving vehicle only, cisplatin-treated animals (2.5 mg/kg/w during 3 weeks) at 4 and 9 weeks after treatment with vehicle or cisplatin. Protein expression levels are presented as % 6 SD compared to the control group at 4 weeks and at 9 weeks. mRNA expression levels are presented as mean normalized ratio (NR) 6 SD. Statistics were performed using generalized linear models. a p,0.05 versus controls at 4 weeks, b p,0.05 versus controls at 9 weeks. doi:10.1371/journal.pone.0057016.t002  with an increased FE Mg 2+ . The renal EGF expression correlated well with the renal TRPM6 mRNA expression. Cisplatin administration is known to result in a decreased renal EGF expression and urinary EGF excretion in the rat [14]. However, the link between renal EGF expression and renal Mg 2+ loss has not yet been established in this cisplatin rat model. Previously, our group proved the relation between CNI-induced renal Mg 2+ loss and TRPM6 downregulation in a cyclosporine (CsA) ratmodel [9].
In the CsA rat model, we found the simultaneous downregulation of the renal EGF mRNA expression. We hypothesize that the mechanism leading to hypomagnesemia is similar in rats treated with cisplatin and rats treated with CNI and is related to a decreased expression of the renal Mg 2+ channel TRPM6. The reduced expression of EGF in rats treated with cisplatin is in accordance with a mechanism, which was previously revealed in vitro. EGF stimulates the Mg 2+ re-absorption in the DCT via its receptor on the basolateral membrane and via activation of TRPM6 in the apical membrane in vitro [12,13]. However, the results from the present study together with the results of our previous study of CsA-induced nephrotoxicity suggest that an additional EGF-induced mechanism is involved in TRPM6 regulation. Both studies show simultaneous TRPM6 and EGF mRNA downregulation, which might indicate that EGF also influences the TRPM6 mRNA synthesis [9]. Also the finding of an upregulated TRPM6 mRNA expression in rats treated with EGF in our previous study further supports this hypothesis [9]. To further elucidate the relationship between EGF and TRPM6 in vivo, the effect of cisplatin on the EGFR pathway was studied. Based on the activation mechanism of TRPM6 via EGF, as established in vitro, Pi3, Akt and Rac1 expression profiles were evaluated. Activation of these EGFR pathway members leads to TRPM6 activation and redistribution [13]. However, the present study could not establish a straightforward association between the cisplatin-induced EGF downregulation and the down-stream activation of the EGFR pathway members. This could be related to the fact that the investigated intermediates are activated in many other pathways involved in kidney injury such as the insulindependent Pi3/Akt activation, the platelet-derived growth factorinduced Pi3/Akt pathway and the TGF-b1-increased collagen1 expression through the Pi3-PDK1-AKT pathway [20,21,22]. All these pathways are potentially upregulated in the cisplatin-induced nephrotoxicity rat model. Moreover, in the present study, the total cortex protein and mRNA levels were analyzed, which might be a too crude extract to detect specific DCT related changes in the Pi3/Akt pathway.
However, there is evidence for the EGFR pathway-mediated activation of TRPM6 in vivo. Dimke et al. studied the effect of an EGFR inhibitor on renal Mg 2+ handling. Erlotinib-injected mice failed to reduce the FE Mg 2+ in response to a decreased serum Mg 2+ concentration. TRPM6 mRNA was downregulated but not the TRPM6 protein level, indicating that the hypomagnesemia is due to the inactivation of TRPM6 at protein level [23]. Furthermore, several clinical studies reported hypomagnesemia due to renal Mg 2+ wasting after treatment with EGFR-inhibiting antibodies [24,25,26].
The mRNA expression levels of the EGFR remained unchanged. EGFR protein expression remained unchanged after 4 weeks but increased after 9 weeks. This up-regulation is probably a reaction in order to repair cisplatin-induced tubular damage [27].
TRPM7 is very homologue to TRPM6 (<50% homology) and also responsible for the cellular Mg 2+ homeostasis [28,29]. TRPM6 specifically interacts with TRPM7 to form a functional ion channel complex at the cell surface of human embryonic kidney 293 cells [30,31]. In our study, the cisplatin-treated groups showed no TRPM7 downregulation. At this point, it is unclear whether TRPM6 and TRPM7 are both expressed in the DCT of rat kidneys, which is the case for mice and humans [32,33]. Immunocytochemical stainings for both channels were unsuccesful due to the unavailability of appropriate primary anti-rat antibodies.
Since NCC is only expressed in de DCT, renal NCC mRNA expression and FE Na + were measured to evaluate DCT damage [34]. These parameters did not differ between the control group and the cisplatin-treated group after 4 weeks, suggesting that the TRPM6 downregulation is not due to damaged DCTs. This is in contrast to a recent report which describes the downregulation of 3 DCT markers (TRPM6, NCC and Parvalbumine) after cisplatintreatment in mice and concluded that cisplatin affects the entire DCT, leading to renal Mg 2+ , K + , Na + and Ca 2+ wasting [35]. The main difference with our rat model is that we induced low grade functional nephrotoxicity, leading to a selective down regulation of TRPM6, while Van Angelen et al. induced in mice tubular necrosis using a higher dose of cisplatin and less time between the consecutive cisplatin administrations, leading to acute DCT necrosis with downregulation of all DCT markers. Our data indicate a time-related negative effect of cisplatin on the NCC, with a decreased NCC expression level at 9 weeks. However, the FE Na + was stable which reflects the maintenance of the sodium re-absorption by compensatory regulatory mechanisms. Moreover, at 9 weeks, the TRPM6 mRNA expression and the FE Mg 2+ partially recovered compared to the cisplatin-treated group at 4 weeks. In our rat model, the cisplatin-induced Mg 2+ loss is DCTspecific. The mRNA expression of the tight junction proteins claudin-16 (also known as paracellin-1) and claudin-19, which are the key players of the paracellular Mg 2+ transport of the TAL, did not differ between the cisplatin-treated rats and the control rats [5,6]. Claudin-19 protein expression increased at 4 and at 9 weeks, which suggests the activation of a compensatory mechanism for the TRPM6 downregulation.
In conclusion, this study shows that cisplatin treatment results in EGF and TRPM6 downregulation in the rat kidney, causing renal Mg 2+ loss. Our results are in line with the hypothesis that EGF influences the renal expression or activation of TRPM6 and plays a significant role in Mg 2+ loss in medication-induced nephropathy.