Differential Effects of Cyclosporin A and Tacrolimus on Magnesium Influx in Caco2 Cells

- PURPOSE: Hypomagnesemia with urinary magnesium (Mg) wasting is a well acknowledged side effect of cyclosporin A (CsA) and tacrolimus (FK506) treatments. TRPM6, TRPM7 and MagT1 are involved in the active transcellular Mg transport processes in intestine and kidney. Since Mg homeostasis is tightly controlled by the dynamic action of intestinal absorption of dietary Mg and renal excretion of Mg, we investigated whether CsA and FK506 in commercially available solutions for clinical use decrease the expression and the function of TRPM6, TRPM7 or MagT1 in the intestinal epithelial cell line Caco2. METHODS: Changes of intracellular free Mg concentrations were measured by Mag-fura-2 imaging in Mg-free medium after the addition of 1 mM MgCl2. TRPM6, TRPM7 and MagT1 were evidenced in cells by immunofluorescence. Proteins and mRNAs were quantified after 18 hours of treatment with CsA or FK506 by western-blot and real-time RT-PCR analyses, respectively. RESULTS: TRPM6 and MagT1 were evidenced on all cell membranes, TRPM7 only on the inner membranes. CsA was responsible for a profound decrease in Mg 2+ influx in intestinal epithelial cells, which may result in a decrease of intestinal Mg absorption, whereas FK506 was responsible for a marked increase in Mg 2+ influx. Neither CsA nor FK506 altered TRPM6, TRPM7 or MagT1 mRNA levels or MagT1 protein level. CONCLUSIONS: In Caco2 cells, Mg 2+ influx was inhibited by CsA solutions whereas enhanced by FK506 solutions, without alteration of MagT1, TRPM6 and TRPM7 expression, leading to the conclusion that CsA and FK506 have opposite effects in the functional activity of the Mg transporters herein examined. In clinical use, FK506 should be preferred for patients at risk for hypomagnesemia.


INTRODUCTION
Cyclosporin A (CsA) and tacrolimus (FK506) are potent immunosuppressants widely used to prevent organ allograft rejection and for patients with allergies and immune-mediated diseases. Both drugs inhibit the phosphatase calcineurin, but through different mechanisms (1). Hypomagnesemia with urinary magnesium (Mg) wasting is a well acknowledged side effect of CsA and FK506 treatments (2)(3)(4)(5)(6). Severe Mg deficiency causes tetany, hypertension, seizures, and cardiac arrhythmia (7). Mg homeostasis is tightly controlled by the dynamic action of intestinal absorption of available dietary Mg, exchange with bone, and renal excretion (8). TRPM6 (Transient Receptor Potential Melastatin) and TRPM7 are channel kinases involved in the active transcellular Mg transport processes in intestine and kidney. These ion channels have a permeation profile of magnesium > calcium (9,10). They have been associated with Mg 2+ influx and homeostasis (11,12). While TRPM6 is mainly restricted to the absorptive epithelial intestine and kidney where it regulates total body Mg levels, TRPM7 is ubiquitously expressed and may be more important in regulating intracellular Mg homeostasis and free intracellular levels of magnesium ([Mg 2+ ] i ) (9,(13)(14)(15)(16). MagT1, a selective Mg 2+ transport protein with very little permeability to other divalent cations, is ubiquitously expressed (17,18). It is involved in maintaining intracellular Mg 2+ levels (19).
The expression of TRPM6, the gatekeeper of the body's Mg balance, was shown to be regulated by extracellular Mg concentration, several hormones, and drug treatments, among which CsA and FK506, whereas its channel activity is modified by [Mg 2+ ] i and ATP (20)(21)(22). ________________________________________ CsA decreases TRPM6 expression and function in renal tubular epithelial cells, which could result in a decreased renal Mg reabsorption (23). However, we have previously shown that CsA treatment of mice induced moderate hypomagnesemia without urinary wasting or renal dysfunction, which could not be ameliorated by oral supplementation. Plasma Mg levels improved rapidly upon CsAtreatment cessation (24). So far, calcineurin inhibitors influence on MagT1 function and expression has not been investigated. Since Mg homeostasis is a balance of epithelial intestinal absorption and renal excretion, we investigated whether CsA and/or FK506 decrease the expression and/or function of TRPM6, TRPM7 or MagT1 in Caco2 cells taken as an epithelial intestinal cell model.

Cell culture
Caco2 cells were grown in Dulbecco's modified Eagle's medium (D-MEM, Invitrogen-GIBCO) high glucose (4,5 g/L) containing sodium pyruvate (110 mg/L) and 3.97 mM L-glutamine, and supplemented with 1% non-essential amino acids (Invitrogen-GIBCO), 20% fetal calf serum (Eurobio, Les Ulis, France), penicillin (100 U/mL) and streptomycin (100 µg/mL). After 7 days, cells are confluent as a monolayer. All experiments were performed after 7 days of culture. On the day before given experiments with CsA or FK506, cells culture in dishes or flasks were washed with warm media then provided with fresh media containing CsA or FK506.
After fixation with 4% paraformaldehyde in PBS for 30 min, cells were washed 3 times in PBS. WGA (Wheat Germ Agglutinin) Alexa Fluor ® 594 conjugate (Invitrogen), a plasma membrane marker, was applied at 5 µg/ml for 10 mins at room temperature. Cells were washed 3 times in PBS. For TRPM6 and TRPM7, cells were permeabilized in 0.1% Triton X-100 at 4°C for 5 min. After a rinse in PBS, non specific binding sites were blocked by incubation in 10% normal rabbit or goat IgG (R&D Systems, Lille, France) for TRPM6/7 and MagT1, respectively, for 1 h. Subsequently, cells were incubated with primary anti-TRPM6, anti-TRPM7 or anti-MagT1 antibody in a solution reducing background signal (Antibody diluent, Dako Real ™ , Dako France, Trappes) at 4°C overnight. Then secondary Alexa Fluor ® antibody was applied for 30 min. Negative control coverslips were processed following the same protocol, but normal goat or rabbit IgG (R&D Systems) was applied instead of primary antibody to check for non specific binding. DAPI (SIGMA D8417, 0.001 µg/mL) was applied for 15 min to counterstain nuclei in blue. After a brief rinse in PBS, coverslips were mounted on slides with Fluorescent Mounting reagent (Dako). Immunolabeled cells were visualized on LSM 700 confocal microscope (Carl Zeiss).

Total Mg quantification
Caco2 cells were grown in 25 cm 2 flasks (500 000 cells/5 mL) and cultured for 6 days. The confluent cells were either untreated, or treated with CsA (10 µM) or FK506 (0.1 µM) for 24h. Cells were washed with PBS, trypsinized then homogenized in cold PBS. 2x10 6 cells were transferred to a microfuge tube and placed on ice, then sonicated in 100 µL PBS for five pulses at a 40% duty cycle using an ultrasonic processor (VibraCellTM, Sonics and materials, Danbury, USA). After centrifugation for 10 min at 10 000 rpm, the supernatant was collected and total Mg 2+ concentration was measured using a xylidyl blue method by routine procedure in a clinical chemistry analyzer (AU400, Olympus, Rungis, France) at a wavelength of 520 nm. Results were expressed in nmol/10 6 cells.  (25). The minimum (Rmin) and maximum (Rmax) ratios were determined using 20 µM digitonin. The Rmax for Mag-fura-2 was determined by the addition of 50 mM MgCl 2 , and the Rmin was obtained by removal of the Mg 2+ and the addition of 100 mM EDTA (26). The change in [Mg 2+ ] i with time (d([Mg 2+ ] i )/dt) was determined by linear regression analysis of the fluorescence ratio over the initial 600 s and expressed in nmol/s.

RNA quantification by real-time reversetranscription-polymerase chain reaction (RT-PCR)
Total RNA was isolated from cell homogenates according to the Agilent Total RNA Isolation protocol (Agilent Technologies, Massy, France). Reverse transcription into cDNA with elimination of genomic DNA was performed using the QuantiTect ® Reverse Transcription protocol (Qiagen, Les Ulis, France). Real-time PCRs were performed using the SteponePlus ™ Real-Time PCR System according to the TaqMan ® Gene Expression Assays protocol (Applied Biosystems, Villebon-sur-Yvette, France) in 96-wells plates in a total volume of 20 µl. The TRPM6, TRPM7, MagT1 and β-actin primer pairs amplify fragments of 68, 88, 67, and 171 bp, respectively. Relative gene expression was calculated using the delta-delta Ct method with normalization to βactin (A.U.).

Western blot analysis
Caco2 were scraped into cold PBS and precipitated by centrifugation at 1,400 g for 5 min at 4°C. Then the cells were lysed in a buffer containing 150 mM NaCl, 1 mM EDTA, 1% NP-40, 1 mM phenylmethylsulfonyl fluoride, 1 mM NaVO4, 1 mM NaF, 50 mM Tris-HCl (pH 8.0), and the protease inhibitors leupeptine 10 µg/ml and aprotinine 20 µg/ml. After sonication for 20 s, the whole cell lysate contained plasma membrane, cytoplasm and nuclear fraction. Protein concentration was measured using the BC Assay kit (Interchim).

STATISTICS
Results were presented as mean ± SEM. Differences between groups were analyzed by one-way analysis of variance, and individual comparisons were made using the parametric Tukey's multiple comparison test (PRISM ® 5 software, GraphPad, San Diego, USA). Significant differences were assumed at p<0.05.

RESULTS
Since in clinical use, CsA and FK506 may be formulated with various vehicles (solution for injection or tablets), in this study we compared the effects of the two immunosuppressant molecules to their corresponding vehicle.

Cell evidence of TRPM6, TRPM7 and MagT1
Immunocytofluorescence showed that TRPM6, TRPM7 and MagT1 are expressed in the cell membranes of Caco2 cells after 7 days of culture ( Figure 1). However, TRPM7 is weakly expressed and only on the inner membranes.

Differential effects of CsA and FK506 on [Mg 2+ ] i elevation
In Mg-free conditions, the addition of 1, 2.5, and 5 mM MgCl 2 caused a dose-dependent elevation of [Mg 2+ ] i (Figure 3). In given experiments, cells were pretreated with 10 µM of the TRPMinhibitor ruthenium red (RR) or with nitrendipine, an organic blocker of L-type Ca 2+ channel for 10 min before adding MgCl 2 . Nitrendipine was shown to be an effective inhibitor of Mg 2+ entry in MagT1 expressed oocytes and renal epithelial cells (18,29). The elevation of [Mg 2+ ] i caused by the addition of 1 mM MgCl 2 was significantly inhibited by RR or nitrendipine (p<0.001), indicating that it reflects a Mg 2+ influx from the extracellular compartment driven by the TRPM and MagT1 transporters (Figure 3). Therefore, a concentration of 1 mM MgCl 2 was used in all the experiments.

Effects of CsA and FK506 on TRPM6, TRPM7 and MagT1 expression
RT-PCR showed that the only effect of the CsA vehicle was a dramatic decrease in TRPM6 mRNA level at the dose of 100 µM. (Figure 5A, C and E). CsA at 100 µM significantly increased MagT1 mRNA level compared with both untreated cells and vehicle (p<0.001) ( Figure 5A). At the same dose, CsA significantly decreased TRPM6 mRNA level (p<0.001 versus untreated cells) but similarly to the vehicle ( Figure 5C). By contrast, FK506 vehicle did not significantly alter MagT1, TRPM6 and TRPM7 mRNA levels ( Figure 5B, D and F). FK506 at 0.01 µM significantly enhanced MagT1 mRNA level (p<0.001 versus untreated cells) but similarly to the vehicle ( Figure 5B). FK506 did not significantly alter TRPM6 mRNA level whatever the dose was ( Figure 5D). TRPM7 expression was not affected by the drugs ( Figure  5E and F).
Western-blot analysis of TRPM6 and TRPM7 proteins could not be performed since they were very weakly expressed in Caco2 cells after 7 days of culture. By contrast, MagT1 was highly expressed and neither CsA nor FK506 altered its expression ( Figure 6).

DISCUSSION
Hypomagnesemia, a frequent side-effect of calcineurin inhibitors, has been ascribed to an inhibition of renal tubular magnesium reabsorption, due to a decreased expression of TRPM6 (5,23). In contrast to CsA, FK506 was shown to induce only a moderate and transient hypomagnesemia (30). In the present study, we demonstrate that those two calcineurin inhibitors display opposite effects on Mg 2+ influx. So far, TRPM6 is the best known component mediating transepithelial Mg 2+ transport, but TRPM7 and MagT1 might be involved in renal transepithelial Mg 2+ transport (31). However, the disturbance of magnesium homeostasis may involve a decrease of intestinal absorption. Therefore, we investigated the effects of CsA and FK506 on the Mg transporters MagT1, TRPM6 and TRPM7 in intestinal epithelial cells. So far, these transporters have been widely investigated in kidney cells, and also in rumen and intestinal cells (13-16, 18, 32). The selective Mg-transporter MagT1 is a 38 kDa protein that is widely expressed among tissues, mainly colon and intestine, and in all cell membranes including endoplasmic reticulum, early and late endosomes or apical and basolateral plasma membrane fractions (18). It is upregulated by extracellular Mg 2+ (32). In the present study, we used the Caco2 cell line since colon is one of the Mg 2+ -absorptive epithelia. These cells expressed both TRPM6 and TRPM7 at low levels, and MagT1 at a high level on plasma membranes. Therefore MagT1 might, in part, be responsible for intestinal Mg absorption. Both RR and nitrendipine inhibited Mg 2+ influx in Caco2 cells. The influx of Mg 2+ mediated by TRPM6 and TRPM7 is inhibited by RR, a potent blocker of inward monovalent currents while leaving outward currents unaltered, probably through a competition between Mg 2+ for binding sites within the channel pore (9). Our findings support the notions that Mg 2+ entry in Caco2 cells is via channel-like proteins and that both TRPMs and MagT1 may be involved.
We compared the effects of the immunosuppressant molecules to those of their vehicle to evaluate the intrinsic effect of the molecules since the vehicles vary with the pharmaceutical formulation. Therefore, in clinical use, the effect of the treatment on Mg 2+ absorption is expected to be at least in part dependent on the drug formulation. CsA, which is usually used in in vitro experimental designs in the range of 1 to 100 µM, induced a significant but not complete These findings support the notion that CsA by itself (compared to the vehicle) is responsible for a profound decrease in Mg 2+ influx in intestinal epithelial cells, which may result in a decrease of intestinal Mg absorption, whereas FK506 by itself (compared to the vehicle) is responsible for a marked increase in Mg 2+ influx. Moreover, they emphasize the role of the vehicle in the overall pharmacological efficacy and side effects of these drugs.
The mRNAs of MagT1, TRPM6 and TRPM7 were expressed in the Caco2 cells, but only the MagT1 protein was highly expressed. Indeed, the expression of Mg transporters varies with the tissue, the cell type and the step of the cell cycle. Therefore, we could not investigate the effect of the drugs on TRPM6 and TRPM7 protein expression. By contrast with the modifications exhibited in Mg 2+ influx, no decrease of TRPM6, TRPM7 or MagT1 mRNA levels and of MagT1 protein level was observed with CsA, and no significant increase of the mRNA and MagT1 protein level with FK506. It is worthy to note that the highest concentration of CsA vehicle induced damages in Caco2 cells and a dramatic decrease of TRPM6 mRNA level without affecting TRPM7 and MagT1 mRNAs. However, TRPM6 expression was shown to be decreased by CsA in normal rat kidney cells, but TRPM7 expression was not (23). We therefore hypothesize that both CsA and FK506 alter the channel function but not the mRNAs and proteins of those Mg transporters, unless other non-identified transporters are affected by these drugs. In