Interaction of cyclosporine and FK506 with diuretics in transplant patients

Kidney International, Vol. 58 (2000), pp. 325–330

DIALYSIS – TRANSPLANTATION

Interaction of cyclosporine and FK506 with diuretics in transplant patients JOHN M. ARTHUR and SHAZIA SHAMIM Departments of Medicine and Biochemistry and Molecular Biology, The University of Louisville, Louisville, Kentucky, USA

Interaction of cyclosporine and FK506 with diuretics in transplant patients. Background. The calcineurin inhibitors cyclosporine and FK506 are widely used for immunosuppression in solid organ transplantation. One of the side effects of these agents is renal magnesium wasting. The site of action and molecular mechanism of this effect are not known. We hypothesized that agents such as diuretics that cause renal magnesium wasting through a similar action would not have an additive effect on magnesium deficiency with calcineurin inhibitors. Methods. The records of 50 heart transplant patients on calcineurin inhibitors were reviewed to determine levels of serum magnesium and required replacement dose of magnesium, diuretic usage, and other laboratory values. Results. Loop diuretics did not change either the magnesium level or magnesium replacement requirements in patients on calcineurin inhibitors. In contrast, the thiazide diuretic resulted in an increase in serum magnesium and a decrease in magnesium replacement. Results were similar when the cyclosporine or FK506 groups were evaluated alone. Patients taking FK506 had lower serum magnesium values and higher requirements for magnesium replacement compared with patients taking cyclosporine. Conclusion. We conclude that calcineurin inhibitors and loop diuretics have a similar site of action.

nesium wasting in animal models [3] and humans [4, 5]. The mechanism of the effect of calcineurin inhibitors on magnesium wasting is not clear. Unlike other ions that are primarily reabsorbed in the proximal tubule, the majority of magnesium reabsorption occurs more distally. Approximately 60% of filtered magnesium is reabsorbed in the loop of Henle and another 5% in the distal tubule [6]. The relatively small amount of magnesium reabsorbed in the distal tubule accounts for 60% of the amount delivered to this segment and is thus physiologically important in the regulation of magnesium homeostasis. Since the thick limb of the loop of Henle provides a large fraction of renal magnesium reabsorption and the distal tubule plays an important part in the final regulation of magnesium excretion, one of these segments is likely to be the portion of the tubule in which cyclosporine and FK506 work to inhibit renal magnesium reabsorption. Interactions between diuretics, which act in specific portions of the tubule, and calcineurin inhibitors could help to sort out the tubular sight of action of these drugs. Loop diuretics, like furosemide and bumetanide, are known to work in the thick ascending limb of the loop of Henle [7] and are associated with magnesium loss [8]. Thiazide diuretics exert their actions in the distal convoluted tubule [9]. They have also been associated with Mg wasting with prolonged treatment, although the direct effects of thiazides on the tubule are controversial [10–12]. We reviewed the charts of 50 heart transplant patients to determine what interactions occurred between cyclosporine or FK506 and diuretics on serum magnesium levels and the required magnesium dose.

Cyclosporine and FK506 are commonly used to prevent rejection after organ transplantation. Both inhibit the phosphatase calcineurin, although through different mechanisms. Cyclosporine binds cyclophilin and FK506 binds to FK-binding protein. These complexes inhibit the Ca2⫹- and calmodulin-dependent serine/threonine phosphatase calcineurin [1]. In T lymphocytes, calcineurin is a key rate-limiting step in the activation of transcription factors important for the expression of interleukin-2 (IL-2) [2]. Both calcineurin inhibitors have been associated with hypomagnesemia and urinary mag-

METHODS We reviewed the records of patients who had received a heart transplant at Jewish Hospital in Louisville, Kentucky, USA, between July 1, 1986, and November 30, 1998. Transplant coordinators kept flow sheets with laboratory values and medicines for each patient. We obtained every measurement of magnesium from these flow charts for 50 patients. The total number of magnesium

Key words: magnesium deficiency, immunosuppression, heart transplant, loop diuretics, calcineurin inhibition, thiazide diuretic. Received for publication October 25, 1999 and in revised form January 5, 2000 Accepted for publication January 18, 2000

 2000 by the International Society of Nephrology

325

326

Arthur and Shamim: Magnesium in transplant patients Table 1. Characteristics of heart transplant patients on calcineurin inhibitors

Cyclosporine FK506

Number of measurements 1755 726

Age

Mean time since transplant days

Median time since transplant days

Male %

48.8 ⫾ 0.3 49.8 ⫾ 0.4a

859 ⫾ 24 786 ⫾ 31

433 511

79 66

Cyclosporine level

FK506 level

mEq/L 214 ⫾ 2.9

Diabetic 10.2 ⫾ 0.2

66% 70%

P ⬍ 0.05 compared to cyclosporine

a

Fig. 1. Serum magnesium level and dose of magnesium in patients on calcineurin inhibitors. Patients on FK506 had significantly lower serum magnesium and were on a higher dose of magnesium oxide compared with patients on cyclosporine (*P ⬍ 0.01). Symbols are: ( ) cyclosporine; (䊏) FK506.

measurements was 2496. Where available, the following information was obtained for each date of magnesium measurement: serum creatinine, potassium, cyclosporine, and FK506 levels. Doses of magnesium oxide, cyclosporine, FK506, and diuretics, and whether the patient was diabetic were also recorded. Diabetes was defined as currently taking either insulin or a hypoglycemic agent. The age of patient and the length of time since transplantation were calculated for each measurement. Values are reported as mean and SEM. Comparisons between groups were made using one-way analysis of variance followed by Dunnett’s test. RESULTS The charts of 50 heart transplant patients on either cyclosporine or FK506 were reviewed. Characteristics of patients on cyclosporine and FK506 are shown in Table 1. Patients on cyclosporine were slightly younger (P ⬍ 0.05), and there was a trend toward a longer mean time since transplant, although the median time since transplant was longer in the FK506 group. The discrepancy between mean and median reflects the fact that three individuals in this study had been transplanted for longer than 10 years and all were on cyclosporine. Both groups had more males than females and a similar proportion of diabetic patients. A comparison of magnesium values and magnesium dose between the two groups is seen in Figure 1. Patients on FK506 had serum magnesium levels

that were slightly below the reference range (1.4 to 1.9 mEq/L) and significantly less than the cyclosporine group (FK 1.39 ⫾ 0.01, CsA 1.55 ⫾ 0.006, P ⬍ 0.001). The FK506 group also required a higher replacement dose of magnesium (P ⬍ 0.001). Mean serum creatinine exceeded reference values in both groups but was significantly higher in the FK506 group (1.67 ⫾ 0.02 mg/dL) compared with cyclosporine (1.53 ⫾ 0.02, P ⬍ 0.001; data not shown). The relationship between cyclosporine or FK506 levels and serum magnesium or magnesium dose was analyzed by linear regression. No statistically significant correlation between drug level and hypomagnesemia was found. We evaluated the effect of diuretics on serum magnesium level and required magnesium dose for patients on either calcineurin inhibitor (Table 2). Our primary goal was to determine the effect of loop and thiazide diuretics on serum magnesium levels and required magnesium oxide dose. Mean values were determined for serum magnesium levels obtained from patients on each class or combination of classes of diuretics. Data were not analyzed for the potassium-sparing diuretics (N ⫽ 3) and loop plus potassium-sparing diuretic (N ⫽ 23) groups because of the small numbers in these groups. There was no difference in magnesium or magnesium dose between patients not on diuretics and those on loop diuretics (Fig. 2). This demonstrates that loop diuretics do not produce an additional magnesuric effect for patients already on calcineurin inhibitors. In contrast, patients on thiazide diuretics had higher serum magnesium levels (control 1.47 ⫾ 0.01, thiazide 1.54 ⫾ 0.02 mEq/L, P ⬍ 0.05) and lower required doses of magnesium (1186 ⫾ 38 vs. 643 ⫾ 98 mg magnesium oxide/day, P ⬍ 0.05) compared with patients who were not on diuretics. Patients on loop plus thiazide diuretics had a higher serum magnesium level but no difference in the magnesium dose, while patients on thiazide plus potassium-sparing diuretics had higher magnesium levels and a lower magnesium dose than the control subjects. We next examined the results of the subgroup of patients on cyclosporine. When the serum magnesium levels and magnesium requirement for patients on cyclosporine were evaluated, the results were similar to those seen in the entire group of patients on calcineurin inhibitors (Table 3). Loop diuretics had no effect on the serum

327

Arthur and Shamim: Magnesium in transplant patients Table 2. Effect of diuretics in heart transplant patients on calcineurin inhibitors

Magnesium mEq/L Mean N SEM Potassium mEq/L Mean N SEM Magnesium dose mEq/L Mean N SEM a

Loop plus thiazide

Thiazide plus potassium-sparing

Loop plus thiazide plus potassium-sparing

Control

Loop

Thiazide

1.47 862 0.01

1.49 1315 0.01

1.54a 87 0.02

1.76a 91 0.03

1.75a 66 0.03

1.58 35 0.04

4.17 858 0.02

4.11 1284 0.01

4.07 87 0.05

3.78a 90 0.05

3.83a 65 0.05

4.05 35 0.09

1186 861 38

1102 1309 29

643a 87 98

1156 91 124

848a 66 101

2206 35 155

P ⬍ 0.05 compared to control

than the controls (1.41 ⫾ 0.02 vs. 1.72 ⫾ 0.1 mEq/L, P ⬍ 0.05). There was a trend toward a higher dose of magnesium in the loop plus thiazide group, but it did not reach statistical significance. DISCUSSION

Fig. 2. Comparison of serum magnesium level and dose of magnesium for patients on calcineurin inhibitors. Patients on either cyclosporine or FK506 were included in the analysis. Patients on loop diuretics did not differ from those on no diuretics. Patients on thiazide diuretics had a significantly higher serum magnesium concentration and were on a lower dose of magnesium replacement than those on no diuretics (*P ⬍ 0.05). Symbols are: (䊏) control; ( ) loop diuretic; ( ) thiazide diuretic; ( ) loop and thiazide diuretics; ( ) thiazide and potassium-sparing diuretics; ( ) loop, thiazide and potassium-sparing diuretics.

magnesium level or magnesium dose, although there was a significant decrease in serum potassium. The thiazide group had a significantly decreased magnesium requirement, although there was no difference in serum magnesium levels. Results for the groups taking combinations of diuretics were essentially the same as for the analysis of calcineurin inhibitors as a whole. Among patients on FK506, there were only three groups based on diuretics. These groups were those on no diuretic, loop diuretics, and loop plus thiazide diuretics. There were no patients on thiazide diuretics alone among the patients on FK506. The results in this group were similar to those seen for the calcineurin inhibitor group as a whole (Table 4). As in the cyclosporine group and the total calcineurin inhibitor group, patients on loop diuretics did not differ from the patients on no diuretics in terms of serum magnesium or dose of magnesium among patients on FK506. Patients on loop plus thiazide diuretics had a higher serum magnesium concentration

Calcineurin inhibitors have previously been shown to cause urinary magnesium wasting, but the renal site of action of these agents has not been known. We hypothesized that the addition of other promagnesuric agents working at the same renal tubular site or through a similar mechanism would not produce a further increase in magnesium loss beyond that seen with the calcineurin inhibitors alone. Loop diuretics produce renal magnesium wasting, have a well-characterized site and mechanism of action, and are commonly given to transplant patients in combination with calcineurin inhibitors. Thiazide diuretics also have a well-characterized site and mechanism of action for inhibition of sodium reabsorption, but their effect on magnesium reabsorption is less clear. We have examined surrogates for magnesium wasting in heart transplant patients on combinations of calcineurin inhibitors and diuretics. We have attempted to determine interaction of diuretics with calcineurin inhibitors to produce renal magnesium wasting and magnesium deficiency. Since only a small fraction of total magnesium is in the serum, the serum magnesium level is not necessarily reflective of magnesium deficiency. A number of methods have been used to try to assess more accurately magnesium deficiency, including skeletal muscle magnesium, red blood cell (RBC) magnesium, platelet magnesium, and urinary magnesium excretion in response to magnesium loading [13]. In this retrospective study, we have used the serum magnesium level and required magnesium dose to assess magnesium loss. In combination, these two factors give a relatively good indication of the amount of magnesium

328

Arthur and Shamim: Magnesium in transplant patients Table 3. Effect of diuretics in heart transplant patients on cyclosporine

Magnesium mEq/L Mean N SEM Potassium mEq/L Mean N SEM Magnesium dose mEq/L Mean N SEM

Loop plus thiazide

Thiazide plus potassium-sparing

Loop plus thiazide plus potassium-sparing

Control

Loop

Thiazide

1.51 545 0.01

1.54 916 0.01

1.54 87 0.02

1.77a 80 0.04

1.76a 66 0.03

1.58 35 0.04

4.18 541 0.02

4.09a 909 0.02

4.07 87 0.05

3.79a 80 0.05

3.85a 65 0.05

4.05 35 0.09

1110 545 44

970 916 32

644a 87 98

1020 80 118

848 66 101

2206a 35 155

P ⬍ 0.05 compared to control

a

Table 4. Effect of diuretics in heart transplant patients on FK506

Magnesium mEq/L Mean N SEM Potassium mEq/L Mean N SEM Magnesium dose mEq/L Mean N SEM

Loop plus thiazide

Control

Loop

1.41 317 0.02

1.38 398 0.02

1.72a 11 0.1

4.17 317 0.03

4.16 374 0.03

3.78 10 0.2

1312 317 69

1416 398 59

2145 11 468

P ⬍ 0.05 compared with control

a

losses, which presumably occurs primarily through the urine. In the current study, patients on FK506 had a greater degree of magnesium wasting, as demonstrated by lower serum magnesium and higher daily dose of oral magnesium compared with patients on cyclosporine. Among all of the patients on calcineurin inhibitors, neither the serum magnesium level nor required dose of magnesium was different for patients on loop diuretics compared with those who were not on diuretics. In contrast, patients on thiazide diuretics had higher serum magnesium levels and lower requirements for magnesium replacement. Potassium was in the normal range and was not significantly different between groups. The data for magnesium were essentially the same when the cyclosporine or FK506 groups were examined alone. Magnesium reabsorption in the renal tubule occurs by distinct mechanisms in different tubular segments. In the thick ascending limb, where the bulk of magnesium reabsorption occurs, much of the reabsorption occurs through a paracellular pathway, and is driven by the electrical lumen-positive gradient generated during reabsorption of NaCl with leak of K⫹ back into the tubular

lumen. This pathway requires the apical surface Na⫹/ K⫹/2 Cl⫺ cotransporter and ROMK potassium channels [14] and Na⫹,K⫹-ATPase and CLCNKB Cl⫺ channels [15, 16] on the basal surface (Fig. 3). The mechanism of the renal magnesium wasting caused by cyclosporine and FK506 is unclear. Both agents inhibit Na⫹,K⫹-ATPase activity in the medullary thick ascending limb and cortical collecting duct [17, 18]. Two other inhibitors of calcineurin, pentafluorophenol and peptide 412, also inhibited Na⫹,K⫹-ATPase activity. Rapamycin, which binds to the same immunophilin as FK506 without inhibiting calcineurin, and SDZ 220-384, which binds to cyclophilin without inhibiting calcineurin, do not inhibit Na⫹,K⫹-ATPase activity [18]. Phosphorylation of Na⫹,K⫹-ATPase by protein kinase C results in inhibition of its activity [19]. Therefore, inhibition of calcineurin may result in increased phosphorylation and decreased activity of Na⫹,K⫹-ATPase in the thick ascending limb. In the thick ascending limb, decreased Na⫹,K⫹-ATPase activity would result in a decrease in the lumen positive gradient, which is the driving force for magnesium reabsorption. This potential mechanism for calcineurin inhibitors to decrease magnesium reabsorption is outlined in Figure 3. Loop diuretics bind to the chloride-binding sight of the Na⫹/K⫹/2 Cl⫺ cotransporter (NKCC) on the apical surface of the thick ascending limb and inhibit its function. Inhibition of this transporter results in a loss of the ability to produce the lumen-positive gradient that is the driving force for magnesium reabsorption. If loop diuretics and calcineurin inhibitors work through the same general mechanism to inhibit magnesium reabsorption, we would expect that the effects of these agents would not be additive. This is, in fact, what we have seen in our study. Loop diuretics did not cause a further decrease in serum magnesium or an increase in magnesium requirement compared with calcineurin inhibitors alone. This supports the hypothesis that calcineurin inhibitors increase urine magnesium loss through an effect

Arthur and Shamim: Magnesium in transplant patients

329

and potassium depletion inhibits uptake [20]. The heart transplant patients in this study were very closely followed by a team of transplant coordinators and had frequent laboratory assessment. In spite of the diuretic use, hypokalemia was rare. Because of the close follow-up and replacement of potassium and the tendency of calcineurin inhibitors to increase potassium, hypokalemia was eliminated as a potential cause of magnesium wasting. This likely accounts for an unmasking of the beneficial effect of thiazide diuretics on magnesium metabolism. We have demonstrated that loop diuretics do not exacerbate the hypomagnesemic effects of calcineurin inhibitors and that thiazide diuretics may ameliorate these effects. When feasible, thiazide diuretics may be a better choice for treatment of volume excess in patients on calcineurin inhibitors in order to improve hypomagnesemia and to decrease the requirement for magnesium supplementation. Fig. 3. Schematic representation of thick ascending limb transport mechanisms responsible for magnesium reabsorption. Electroneutral movement of Na⫹, K⫹, and Cl⫺ across the apical membrane followed by leak of K⫹ back into the lumen generates the lumen-positive gradient that is responsible for paracellular magnesium reabsorption. Inhibition of any of the proteins responsible for generation of this gradient interferes with reabsorption of magnesium. The hypothesized mechanism of action of calcineurin inhibitors to decrease magnesium reabsorption is shown. Normally, calcineurin inhibits protein kinase C (PKC) and prevents it from inhibiting the Na⫹,K⫹-ATPase. However, when cyclosporine or FK506 inhibits calcineurin, PKC is free to phosphorylate the Na⫹,K⫹-ATPase and decrease its activity. This results in decreased net activity of the loop that generates the lumen-positive gradient. Loop diuretics inhibit the NKCC cotransporter, which also results in a decrease in the lumen-positive gradient. Thus, both loop diuretics and calcineurin inhibitors act through the same pathway to inhibit magnesium reabsorption. Abbreviations are: ClCNKB, chloride channel; ROMK, K⫹ channel; NKCC, Na⫹/K⫹/2 Cl cotransporter.

on Na⫹,K⫹-ATPase activity in the thick ascending limb of the loop of Henle. It is important to note that the calcineurin inhibitors do not have any known direct effect on the function of the NKCC cotransporter. In the proposed mechanism, both loop diuretics and calcineurin inhibitors decrease magnesium reabsorption by inhibiting the generation of a lumen-positive gradient. Loop diuretics do so through a known effect on the NKCC, while calcineurin inhibitors indirectly inhibit Na⫹,K⫹ATPase activity in the thick ascending limb. Patients on thiazide diuretics actually had less magnesium loss compared with those on no diuretics. This is consistent with the acute effect of thiazides to increase magnesium reabsorption. In previous studies, chronic use of thiazide diuretics has been associated with magnesium wasting. Since thiazide use is often associated with hypokalemia, which can cause magnesium wasting, thiazide use by itself may not lead to hypomagnesemia when potassium levels are adequate. This is supported by data obtained from distal convoluted tubule cells showing that thiazide diuretics increase magnesium uptake into cells

Reprint requests to John Arthur, M.D., Ph.D., Kidney Disease Program, 570 South Preston Street, 1st floor South, Donald E. Baxter Building, University of Louisville, Louisville, Kentucky 40202-1764, USA. E-mail: [email protected]

REFERENCES 1. Liu J, Farmer JD, Lane WS, Friedman J, Weissman I, Schrieber SL: Calcineurin is a common target of cyclophilin-cyclosporin A and FKBP-FK506 complexes. Cell 66:807–815, 1991 2. Clipston NA, Crabtree GR: Identification of calcineurin as a key signalling enzyme in T-lymphocyte activation. Nature 357:695–697, 1992 3. Andoh TF, Burdmann EA, Fransechini N, Houghton DC, Bennett WM: Comparison of acute rapamycin nephrotoxicity with cyclosporine and FK506. Kidney Int 50:1110–1117, 1996 4. McDiarmid SV, Colonna JO, Shaked A, Ament ME, Busuttil RW: A comparison of renal function in cyclosporine- and FK506-treated patients after primary orthotopic liver transplantation. Transplantation 56:847–853, 1999 5. Niederstadt C, Steinhoff J, Erbsloh-Moller B, Sack RK, Rob PM: Effect of FK506 on magnesium homeostasis after renal transplantation. Transplant Proc 29:3161–3162, 1997 6. Rouffignac CD, Quamme GA: Renal magnesium handling: Its hormonal control. Physiol Rev 74:305–322, 1994 7. Ellison DH: Diuretic drugs and the treatment of edema: From clinic to bench and back again. Am J Kidney Dis 23:623–643, 1994 8. Quamme GA: Effect of furosemide on calcium and magnesium transport in the rat nephron. Am J Physiol 241:F340–F347, 1981 9. Ryan MP: Diuretics: Potassium/magnesium depletion. Am J Med 82:38–47, 1987 10. Kroenke MK, Wood DR, Hanley JF: The value of serum magnesium determination in hypertensive patients receiving diuretics. Arch Intern Med 147:1553–1556, 1987 11. Hollifield JW: Thiazide treatment of systemic hypertension: Effects on serum magnesium and ventricular ectopic activity. Am J Cardiol 63:22G–25G, 1989 12. Dyckner T, Wester PO, Widman L: Amiloride prevents thiazideinduced intracellular potassium and magnesium losses. Acta Med Scand 224:25–30, 1988 13. Ryan MF, Barbour H: Magnesium measurement in routine clinical practice. Ann Clin Biochem 35:449–459, 1998 14. Hebert SC: An ATP-regulated, inwardly rectifying potassium channel from rat kidney (ROMK). Kidney Int 48:1010–1016, 1995 15. Kieferle S, Fong P, Bens M, van de Walle A, Jentsch TJ: Two highly homologous members of the C1C chloride channel family in both rat and human kidney. Proc Natl Acad Sci USA 91:6943– 6947, 1994

330

Arthur and Shamim: Magnesium in transplant patients

16. Simon DB, Rodrigueo-Soriano J, Mansfield TA, Nelson-Williams C, Mendonca E, Stone R, Schurman S, Nyir A, Alpay H, Bakkaloglu A, Rodriguez-Soriano J, Morales JM, Lifton RP, Taylor MC, Pilz D, Brem A, Trachtman H, Griswold W, Richard GA, John E: Mutations in the chloride channel gene, CLCNKB, cause Bartter’s syndrome type III. Nat Genet 17:171– 178, 1997 17. Tumlin JA, Sands JM: Nephron segment-specific inhibition of Na⫹,K⫹-ATPase activity by cyclosporine A. Kidney Int 43:246–251, 1993

18. Lea JP, Sands JM, McMahon SJ, Tumlin JA: Evidence that inhibition of Na⫹,K⫹-ATPase activity by FK506 involves calcineurin. Kidney Int 46:647–653, 1994 19. Middleton JP, Khan WA, Collinsworth GP, Hannun YA, Medford RM: Heterogeneity of protein kinases C-mediated rapid regulation of Na/K-ATPase in kidney epithelial cells. J Biol Chem 268:15958–15964, 1993 20. Dai LJ, Friedman PA, Quamme GA: Cellular mechanisms of chlorothizaide and cellular potassium depletion on Mg2⫹ uptake in mouse distal convoluted tubule cells. Kidney Int 51:1008–1017, 1997