Nitric Oxide Metabolism in Erythropoietin-Induced Hypertension
Effect of Calcium Channel Blockade
Abstract—Long-term administration of erythropoietin (EPO) frequently causes hypertension in humans and animals with chronic renal failure (CRF). We recently demonstrated that EPO-induced hypertension is hematocrit independent and accompanied by elevated cytosolic [Ca2+]i and nitric oxide (NO) resistance. This study was undertaken to examine the effects of therapy with EPO alone or together with calcium channel blockade on NO metabolism. Urinary excretion of NO metabolites (NOx) and thoracic aorta and kidney endothelial and inducible NO synthases (eNOS and iNOS) were studied in 4 groups of 6 nephrectomized rats treated with either placebo, EPO, the calcium channel blocker felodipine, or EPO plus felodipine for 6 weeks. A group of sham-operated placebo-treated animals served as control. The placebo-treated CRF group exhibited moderate hypertension, elevated basal and depressed stimulated platelet [Ca2+]i, reduced urinary NOx excretion, and diminished vascular and renal eNOS and iNOS proteins. EPO therapy further raised blood pressure and increased resting and stimulated [Ca2+]i but did not change NOx excretion or NOS proteins. Concurrent administration of felodipine abrogated EPO-induced hypertension, normalized resting and stimulated [Ca2+]i, and increased NOx excretion and eNOS and iNOS proteins. Thus, EPO therapy leads to marked increases in blood pressure and resting and stimulated [Ca2+]i. These abnormalities are ameliorated by calcium channel blockade, which restores [Ca2+]i to normal and increases vascular and renal NOS expression.
Introduction of recombinant erythropoietin (EPO) for clinical use has revolutionized the management of anemia in patients with end-stage renal disease. However, long-term administration of EPO frequently results in development of de novo hypertension or exacerbation of preexisting hypertension.1 2 3 Several factors have been implicated in the pathogenesis of EPO-induced hypertension. Chief among them is the resulting rise in hematocrit, which is thought to raise vascular resistance by increase in blood viscosity, loss of hypoxic vasodilation, and enhanced competition by hemoglobin for endothelium-derived nitric oxide (NO).4 5 6 7 Other proposed mechanisms include volume expansion, increased endothelin production,8 enhanced tissue renin-angiotensin activity,9 and direct vasopressor action of EPO.10
In a series of recent studies, we demonstrated that EPO-induced hypertension is not due to the associated increase in hematocrit or erythrocyte mass.11 This was based on the observation that hypertension occurred in both iron-deficient and iron-sufficient uremic rats receiving long-term EPO therapy despite divergent hematocrits. Moreover, repeated red blood cell transfusions did not raise arterial blood pressure despite complete correction of anemia in the uremic animals.11 Accordingly, these experiments provided convincing evidence that EPO-induced hypertension is not related to the erythropoietic action of the hormone. We further found that long-term EPO administration leads to a significant rise in resting cytosolic [Ca2+]i above the elevated values seen in untreated rats with chronic renal failure (CRF).11 Moreover, EPO therapy normalized the defective stimulated surge in [Ca2+]i seen in the untreated CRF animals. The study further showed that long-term EPO administration resulted in a marked reduction in hypotensive response in vivo and vasorelaxation response in vitro to administration of the NO donor sodium nitroprusside. These observations pointed to NO resistance as a possible cause of EPO-induced hypertension. Since the vasodilatory action of NO is mediated by cGMP-induced fall in [Ca2+]i, we hypothesized that the observed NO resistance may be due to the demonstrated rise in resting and/or stimulated [Ca2+]i with EPO therapy.11 The present study was designed to explore the effect of long-term EPO therapy with and without calcium channel blockade on NO production and NO synthase (NOS) expression.
Male Sprague-Dawley rats (Harlan Sprague Dawley Inc) with an average weight of 225 g were used. They were housed in a climate-controlled, light-regulated space with 12-hour light (500 lux) and 12-hour dark (<5 lux) cycles. Animals were allowed free access to water and a low-nitrate rat chow (Purina Rat Chow, Purina Mills Inc). The rats were randomized into the following 5 groups: (1) Placebo-treated CRF group: Animals assigned to the placebo-treated CRF group were subjected to surgical resection of the upper and lower thirds of the left kidney followed by right nephrectomy 4 days later. The procedures were performed under general anesthesia with intraperitoneal injection of sodium pentobarbital 50 mg/kg. Strict hemostasis and aseptic measures were observed. The procedures were performed extraperitoneally with the use of a dorsal incision. The animals were then observed for 6 weeks. (2) EPO-treated CRF group: Animals assigned to this group were subjected to renal mass reduction as described above. They were then treated with intraperitoneal injections of recombinant human EPO (Amgen Inc) 150 U/kg twice a week for 6 weeks. The given dosage was based on our preliminary experiments performed to discern the amount required to prevent or correct anemia of CRF in this model. (3) Felodipine and EPO–treated CRF group: CRF animals assigned to this group were simultaneously treated with EPO and the calcium channel blocker felodipine (Astra Merck Inc). Felodipine was administered by implanted osmotic pumps (Alza Inc) at 7 mg/kg per day. (4) Felodipine-treated CRF group: Animals assigned to this group were treated with felodipine alone. (5) Control group: Animals assigned to this group were sham operated and placebo treated and maintained under conditions identical to those in the other groups.
Body weight, tail arterial blood pressure, and hematocrit (microhematocrit method) were measured regularly. At the end of the 6-week observation period, the animals were placed in individual metabolic cages for 24-hour urine collections. Urine was collected in sterilized containers over dry ice. Animals were then killed by exsanguination by means of cardiac puncture performed with the animals under general anesthesia. Thoracic aorta and kidneys were removed immediately, snap-frozen in liquid nitrogen, and stored at −70°C until processed. In addition, plasma was separated and stored at −70°C.
Blood Pressure Measurement
Systolic blood pressure was measured weekly by a tail sphygmomanometer (Harvard Apparatus). On each occasion the rats were placed on a heated pad in a climate-controlled room. The rat’s tail was placed inside the cuff, and the cuff was inflated and released several times to condition the animal to the procedure. Thereafter, 3 separate measurements were performed, and the mean of the values obtained was used.
Measurements of NO2−/NO3−
The concentration of total nitrates and nitrites (NOx) in the test samples was determined with the use of the purge system of a Sievers model 270B Nitric Oxide Analyzer (NOA, Sievers Instruments Inc) as described previously.12
Standard curves were constructed with various concentrations of NO3− (5 to 100 μmol/L), relating the luminescence produced to the given NO3− concentrations of the standard solutions. The amount of NO2−/NO3− in the test sample was determined by interpolation of the result into the standard curve. All samples were run in triplicate, and the mean of the values obtained was used.
Thoracic aorta and remnant kidneys were homogenized (25% wt/vol) in 10 mmol/L HEPES buffer, pH 7.4, containing 320 mmol/L sucrose, 1 mmol/L EDTA, 1 mmol/L DTT, 10 μg/mL leupeptin, and 2 μg/mL aprotinin at 0°C to 4°C with a tissue grinder fitted with a motor-driven ground glass pestle. Homogenates were centrifuged at 12 000g for 5 minutes at 4°C to remove tissue debris without precipitating plasma membrane fragments.13 14 The supernatant was used for determination of NOS protein mass. Total protein concentration was determined with the use of a Bio-Rad kit (Bio-Rad Laboratories).
Western Blot Analysis
The endothelial and inducible NOS (eNOS and iNOS, respectively) proteins were measured with the use of anti-eNOS monoclonal antibody, anti-Mac NOS-I, human endothelial positive control, mouse macrophage positive control, and peroxidase-conjugated goat anti-mouse IgG antibody (Transduction Laboratories). Briefly, tissue preparations (50 μg of protein) were size-fractionated on 4% to 12% Tris-glycine gel (Novex) at 120 V for 3 hours. After electrophoresis, proteins were transferred onto hybond-ECL membrane (Amersham Life Science Inc) at 400 mA for 120 minutes with the Novex transfer system. In preliminary experiments we had found that the given protein concentrations were within the linear range of detection for our Western blot technique. The membrane was prehybridized in 10 mL of buffer A (10 mmol/L Tris-HCl, pH 7.5, 100 mmol/L NaCl, 0.1% Tween 20, and 10% nonfat milk powder) for 1 hour and then hybridized for an additional 1-hour period in the same buffer containing 10 μL of the given anti-NOS monoclonal antibody (1:1000). The membrane was then washed for 30 minutes in a shaking bath, and the wash buffer (buffer A without nonfat milk) was changed every 5 minutes before 1 hour of incubation in buffer A plus goat anti-mouse IgG/horseradish peroxidase at the final titer of 1:1000. Experiments were performed at room temperature. The washes were repeated before the membrane was developed with a light-emitting nonradioactive method with the use of ECL Western blot detection reagent (Amersham Life Science Inc). The membrane was then subjected to autoluminography for 10 seconds. The autoluminographs were scanned with a laser densitometer (model PD1211, Molecular Dynamics) to determine the relative optical densities of the bands. In all instances, the membranes were stained with Ponceau stain, which verified the uniformity of protein load and transfer efficiency across the test samples.
Platelet Cytosolic Calcium Measurements
At the end of the 6-week observation period a subgroup of rats was killed, and platelets were isolated by means of a procedure described previously.15 The isolated platelets were suspended in HEPES-buffered saline with the following composition (mmol/L): NaCl 145, KCl 5.0, Na2HPO4 0.8, KH2PO4 0.2, MgCl2 1.0, glucose 10, HEPES 10, pH 7.4. This resulted in a platelet suspension containing ≈2×108 cells/mL. [Ca2+]i was determined with the fluorescent calcium indicator fura 2-AM. The suspended platelets were loaded with 4 μmol fura 2-AM in the presence of 0.02% pluronic F-127 to facilitate entry of the indicator into the cells. In addition, 2 mmol/L probenecid was added to minimize leakage of fura 2 out of the platelets.15 Cells were incubated for 60 minutes at 37°C, then centrifuged at 400g for 20 minutes. The supernatant was decanted, and an equal volume of HEPES-buffered saline was added. The cells were incubated at 37°C for 30 minutes to allow complete hydrolysis of the AM group. The dye-loaded cells were suspended in HEPES-buffered saline containing 2 mmol CaCl2 and kept under constant magnetic stirring in a thermostatically controlled cuvette of a spectrofluorometer (DMX 1000, SLM Instruments Inc). Alternating excitation wavelengths of 340 and 380 nm were used, with an emission wavelength of 510 nm. Ratios of fluorescence (R=340/380 nm) were measured every second, automatically corrected for autofluorescence, and plotted graphically for each sample analyzed. Values of autofluorescence were <5% of the fluorescence of the dye-loaded cells and were measured for every experiment.
[Ca2+]i was calculated according to the following formula: [Ca2+]i=(Kd)(B)(R−Rmin)/(Rmax−R), where B is the ratio of 380 nm fluorescence in the absence and presence of saturating concentration of calcium, Kd is the dissociation constant for fura 2 (assumed to be 225 nmol/L), and R is the ratio of fluorescence as defined above. The cells were lysed with Triton (0.05%) to obtain Rmax in the presence of 2 mmol calcium, and Rmin was obtained by the addition of 10 mmol EGTA and sufficient NaOH to raise the pH to 8.5.
Data Presentation and Analysis
Data are presented as mean±SEM. ANOVA and Duncan multiple range test were used as appropriate. P values <0.05 were considered statistically significant.
Effects of CRF
As expected, the placebo-treated CRF group showed a significant increase in serum creatinine concentration and a significant fall in creatinine clearance compared with the normal control group. In addition, the placebo-treated CRF group exhibited a significantly lower weight gain than that seen in the control animals (Table⇓). CRF resulted in a significant fall in hematocrit and a significant rise in blood pressure during the observation period (Figures 1⇓ and 2⇓). This was associated with a significant increase in basal cytosolic [Ca2+]i and a significant reduction in stimulated cytosolic [Ca2+]i in the platelets (Figure 3⇓), consistent with our earlier observations.11 The CRF-associated hypertension was accompanied by a significant reduction in urinary NOx excretion together with depressed thoracic aorta and remnant kidney tissue eNOS and iNOS protein abundance (Figures 4 through 8⇓⇓⇓⇓⇓). These observations are in agreement with our recent studies demonstrating downregulation of l-arginine–NO pathway in this model.16
Effects of EPO
EPO therapy did not lead to a discernible change in either serum creatinine, creatinine clearance, or weight gain compared with values seen in the untreated CRF animals (Table⇑). EPO therapy completely prevented the CRF-associated anemia (Figure 1⇑) but led to a marked rise in arterial blood pressure, which began ≈1 week after the onset of EPO administration (Figure 2⇑).
The EPO-induced hypertension was accompanied by a significant rise in basal [Ca2+]i above the elevated value found in the CRF group. In addition, EPO therapy resulted in a significant increase in stimulated [Ca2+]i to a value that was similar to that of the control group. Thus, EPO therapy resulted in complete correction of the CRF-induced defective surge in platelet [Ca2+]i after thrombin stimulation (Figure 3⇑). This observation is consistent with our earlier studies in this model.11 However, EPO therapy failed to alter either urinary NOx excretion or thoracic aorta eNOS or iNOS protein abundance (Figures 4 through 6⇑⇑⇑). Likewise, EPO therapy had no effect on either eNOS or iNOS protein abundance of the remnant kidney in the CRF group treated with EPO alone (Figures 7⇑ and 8⇑).
Effects of Calcium Channel Blockade
Coadministration of felodipine with EPO completely abrogated the EPO-induced hypertension (Figure 2⇑). This was accompanied by a reduction in resting [Ca2+]i to a level that was similar to that seen in the normal control group (Figure 3⇑). Interestingly, felodipine administration did not affect the thrombin-stimulated rise in [Ca2+]i (Figure 3⇑). Thus, the combination of EPO therapy and calcium channel blockade resulted in normalization of both basal and stimulated [Ca2+]i. In addition, calcium channel blockade led to a significant rise in urinary NOx excretion, suggesting increased total body NO production (Figure 4⇑). The rise in urinary NOx excretion with calcium channel blockade was accompanied by a significant increase in thoracic aorta eNOS and iNOS proteins (Figures 5⇑ and 6⇑). Likewise, felodipine administration increased eNOS and iNOS protein abundance in the remnant kidneys of the CRF group treated with EPO and felodipine. Administration of felodipine in the CRF group resulted in amelioration of CRF-associated hypertension, reduction in basal [Ca2+]i, no change in stimulated [Ca2+]i, increased urinary NOx excretion, and enhanced renal and vascular eNOS and iNOS protein expression, confirming our earlier observations (Figures 7⇑ and 8⇑).16
The CRF animals included in the present study exhibited an expected fall in hematocrit and a modest rise in arterial blood pressure. This was accompanied by a significant decline in urinary excretion of total NOx together with depressed vascular and remnant kidney tissue eNOS and iNOS protein abundance. These observations confirm the results of our previous studies16 and point to downregulation of l-arginine–NO pathway in the CRF animals. Regular administration of EPO resulted in prevention of CRF-associated anemia and led to a further rise in arterial blood pressure as well as increased resting and stimulated [Ca2+]i, confirming our previous studies in this model.11 Interestingly, urinary NOx excretion was unaffected by EPO therapy and the resultant rise in hematocrit, arterial blood pressure, and [Ca2+]i. Likewise, EPO therapy failed to change either eNOS or iNOS protein abundance in the vascular and remnant kidney tissues. It thus appears that maintenance EPO therapy does not significantly affect NO production or NOS expression by either vascular or renal tissue in the CRF animals. The lack of a discernible effect of long-term EPO therapy on NOS shown here is consistent with the result of the in vitro study of cultured endothelial cells reported by Lopes Ongil et al.17 These investigators found that the addition of EPO (0.1 to 10 U/mL) to the medium did not alter either NOS mRNA, preproendothelin mRNA, endothelin release, or NO production in cultured bovine thoracic endothelial cells.17 On the basis of these observations, they concluded that EPO does not affect either NOS or endothelin-1 expressions.17 Thus, both our in vivo experiments with EPO replacement therapy gauged to prevent anemia in CRF animals and the in vitro studies using physiological concentration of EPO in cultured endothelial cells showed no discernible effect on NO production or NOS expression. In contrast, induction of erythrocytosis with regular EPO administration in normal animals has been shown to increase NO production.18 19 20 For instance, del Castillo et al19 showed a significant rise in urinary excretion of NOx in normal rats treated with EPO 150 U/kg thrice weekly for 3 weeks. This was associated with a significant rise in arterial blood pressure and severe erythrocytosis. In another study, induction of severe erythrocytosis with EPO therapy in normal rats resulted in hypertension and increased renal blood flow.18 Administration of NOS inhibitor abrogated the associated rise in renal blood flow and led to an exaggerated hypertensive response in animals with erythrocytosis. These findings were taken as indirect evidence that EPO-induced erythrocytosis results in a compensatory stimulation of l-arginine–NO system in normal rats.18
Several factors can account for the different effects of EPO therapy on NO production in CRF animals used here and normal animals used in the studies of Wilcox et al,18 del Costillo et al,19 and Tsukahara et al.20 First, the animals used here had CRF, which as shown earlier and confirmed here, results in marked downregulation of NOS expression and NO production.16 Second, the EPO dosage used here was gauged to correct and prevent anemia. Thus, our EPO-treated animals had normal hematocrit and erythrocyte mass. In contrast, normal animals used in the former studies exhibited severe erythrocytosis with long-term EPO administration. This was inevitably accompanied by a marked rise in blood viscosity, volume expansion, and increased shear stress, which can upregulate vascular and renal NOS expression.21 22 23 24 These factors were clearly absent in our CRF group treated with EPO alone. Similarly, the goal of EPO replacement in patients with renal disease is the amelioration of CRF anemia as opposed to the induction of erythrocytosis.
In a recent study we demonstrated that EPO-induced hypertension in rats with CRF is unrelated to changes in erythrocyte mass and hematocrit.11 This was based on the demonstrated occurrence of hypertension with regular EPO administration to CRF rats with iron deficiency in the face of persistent EPO-resistant anemia. Furthermore, repeated blood transfusions, sufficient to prevent anemia, in a separate group of CRF rats failed to raise arterial blood pressure. These observations pointed to the role of regular EPO administration as opposed to the correction of anemia per se in the genesis of EPO-induced hypertension. We further showed that EPO-induced hypertension was associated with resistance to the hypotensive action of the NO donors sodium nitroprusside and S-nitroso penicillamine in vivo and to their vasodilatory actions in vitro.11 We hypothesized that the resistance to the vasodilatory action of NO may be due to EPO-induced elevations of resting and stimulated [Ca2+]i. This hypothesis was based on the fact that the vasodilatory action of NO is mediated by the cGMP-induced reduction in [Ca2+]i. Hence, elevated [Ca2+]i can naturally contribute to NO resistance and hypertension. On the basis of these observations, in the present study we included a group of CRF animals receiving concurrent EPO treatment and the calcium channel blocker felodipine.
Concurrent administration of felodipine normalized resting [Ca2+]i and abrogated the EPO-induced hypertension. This was accompanied by a rise in urinary excretion of NOx, suggesting increased total body NO production. The rise in NO production was coupled with increased vascular and remnant kidney tissue NOS protein expression in EPO-treated CRF animals. Similarly, calcium channel blockade normalized blood pressure and reversed downregulation of l-arginine–NO pathway in the CRF group, confirming our earlier observations.16 The mechanism by which EPO raises [Ca2+]i in erythroid progenitor and contractile cells has been recently elucidated.25 26 It has been demonstrated that EPO binding to its receptor results in the activation of a cytosolic tyrosine kinase that, in turn, catalyzes the tyrosine phosphorylation and activation of phospholipase C-γ1. The latter causes hydrolysis of phosphatidylinositol 4,5-bisphosphate and generation of inositol 1,4,5-triphosphate. These events lead to a biphasic rise in [Ca2+]i associated with an initial Ca2+ release from intracellular stores, followed by influx of Ca2+ through EPO receptor-operated, voltage-independent channels.26 Interestingly, AT1 receptor stimulation by angiotensin II also involves tyrosine kinase–dependent phosphorylation/activation of phospholipase C-γ1.27 28 This is particularly noteworthy since administration of angiotensin-converting enzyme (ACE) inhibitor has been shown to prevent EPO-induced hypertension in rats with CRF. Given the dual role of ACE in catalyzing the production of angiotensin II and degradation of bradykinin (kininase activity), its inhibition by ACE inhibitors can contribute to the reduction of [Ca2+]i by 2 mechanisms: (1) by mitigating the angiotensin II–mediated activation of phospholipase C-γ1 and (2) by bradykinin-mediated activation of eNOS, leading to increased NO production. In addition, EPO has been shown to upregulate tissue renin-angiotensin system in the rat.9 These observations form the basis for the demonstrated benefit of ACE inhibition in animals with EPO-induced hypertension.9
In conclusion, the results of the present study point to the possible role of dysregulation of [Ca2+]i in the pathogenesis of depressed l-arginine–NO pathway and NO resistance in EPO-treated CRF animals. If true, calcium channel blockade may be a logical approach to treatment of hypertension in this setting.
Reprint requests to N.D. Vaziri, MD, Division of Nephrology and Hypertension, Department of Medicine, UCI Medical Center, 101 The City Drive, Orange, CA 92868.
- Received May 25, 1998.
- Revision received June 10, 1998.
- Accepted June 12, 1998.
Abraham PA, Macres MG. Blood pressure in hemodialysis patients during amelioration of anemia with erythropoietin. J Am Soc Nephrol. 1991;2:927–936.
Neff MS, Kim KE, Persoff M, Onesti G, Swartz C. Hemodynamics of uremic anemia. Circulation. 1971;43:876–883.
Carlini R, Obialo CI, Rothstein M. Intravenous erythropoietin (rHuEPO) administration increases plasma endothelin and blood pressure in hemodialysis patients. Am J Hypertens. 1993;6:103–107.
Eggena P, Willsey P, Jamgotchian N, Truckenbrod L, Hu MS, Barrett JD, Eggena MP, Clegg K, Nakhoul F, Lee DBN. Influence of recombinant human erythropoietin on blood pressure and renin-angiotensin systems. Am J Physiol. 1991;261:E642–E646.
Vaziri ND, Zhou XJ, Naqvi F, Smith J, Oveisi F, Wang ZQ, Purdy RE. Role of nitric oxide resistance in erythrocyte-induced hypertension in rats with chronic renal failure. Am J Physiol. 1996;271(Endocrinol Metab.):E113–E122.
Martin PY, Xu DL, Niederberger M, Weigert A, Tsai P, St. John J, Gines P, Schrier RW. Upregulation of endothelial constitutive NOS: a major role in the increased NO production in cirrhotic rats. Am J Physiol. 1996;270(Renal Fluid Electrolyte Physiol.):F494–F499.
Vaziri ND, Zhou XJ, Smith J, Oveisi F, Baldwin K, Purdy RE. In vivo and in vitro pressor effects of erythropoietin in rats. Am J Physiol. 1995;269(Renal Fluid Electrolyte Physiol.):F838–F845.
Vaziri ND, Ni Z, Wang XQ, Oveisi F, Zhou XJ. Down-regulation of nitric oxide synthase in chronic renal insufficiency: role of excess PTH. Am J Physiol. 1998;274(Renal Physiol.):642–649.
del Castillo D, Raij L, Shultz PJ, Tolins JP. The pressor effect of recombinant human erythropoietin is not due to decreased activity of the endogenous nitric oxide system. Nephrol Dial Transplant. 1995;10:505–508.
Tsukahara H, Hiraoka M, Hori C, Hata I, Okada T, Gejyo F, Sudo M. Chronic erythropoietin treatment enhances endogenous nitric oxide production in rat. Scand J Lab. Invest. 1997;57:487–494.
Awolesi MA, Sessa WC, Sumpio BE. Cyclic strain upregulates nitric oxide synthase in cultured bovine aortic endothelial cells. J Clin Invest. 1995;96:1445–1449.
Nava E, Luscher TF. Evidence for an increased production of nitric oxide in spontaneously hypertensive rats. Eur J Clin Invest. 1994;24(suppl 2):A264. Abstract.
Noris M, Morigi M, Donadelli R, Aiello S, Foppolo M, Todeschini M, Orisio S, Remuzzi G, Remuzzi A. Nitric oxide synthesis by cultured endothelial cells is modulated by flow conditions. Circ Res. 1995;76:536–543.
Pohl U, Holtz J, Busse R, Bassenge E. Crucial role of endothelium in the vasodilator response to increased flow in vivo. Hypertension. 1986;8:37–44.
Miller BA, Cheung JY. Mechanisms of erythropoietin signal transduction: involvement of calcium channels. Proc Soc Exp Biol Med. 1994;206:263–267.
Marrero MB, Schieffer B, Ma H, Bernstein KE, Ling BN. Ang II-induced tyrosine phosphorylation stimulates phospholipase C-gamma and Cl− channels in mesangial cells. Am J Physiol. 1996;270:C1834–C1842.