Ablation of the N-type calcium channel ameliorates diabetic nephropathy with improved glycemic control and reduced blood pressure

Pharmacological blockade of the N- and L-type calcium channel lessens renal injury in kidney disease patients. The significance of specific blockade of α1 subunit of N-type calcium channel, Cav2.2, in diabetic nephropathy, however, remains to be clarified. To examine functional roles, we mated Cav2.2−/− mice with db/db (diabetic) mice on the C57BLKS background. Cav2.2 was localized in glomeruli including podocytes and in distal tubular cells. Diabetic Cav2.2−/− mice significantly reduced urinary albumin excretion, glomerular hyperfiltration, blood glucose levels, histological deterioration and systolic blood pressure (SBP) with decreased urinary catecholamine compared to diabetic Cav2.2+/+ mice. Interestingly, diabetic heterozygous Cav2.2+/− mice also decreased albuminuria, although they exhibited comparable systolic blood pressure, sympathetic nerve activity and creatinine clearance to diabetic Cav2.2+/+ mice. Consistently, diabetic mice with cilnidipine, an N-/L-type calcium channel blocker, showed a reduction in albuminuria and improvement of glomerular changes compared to diabetic mice with nitrendipine. In cultured podocytes, depolarization-dependent calcium responses were decreased by ω-conotoxin, a Cav2.2-specific inhibitor. Furthermore, reduction of nephrin by transforming growth factor-β (TGF-β) in podocytes was abolished with ω-conotoxin, cilnidipine or mitogen-activated protein kinase kinase inhibitor. In conclusion, Cav2.2 inhibition exerts renoprotective effects against the progression of diabetic nephropathy, partly by protecting podocytes.


Improvement of Glucose Metabolism in Ca v 2.2 −/− Mice. To examine glucose metabolism in db/db
Ca v 2.2 −/− mice, we analyzed 6-h fasting blood glucose levels every 2 weeks during the experimental period. All db/db mouse groups showed hyperglycemia at 8 weeks of age, whereas all db/+ mouse groups had normal glucose levels (Fig. 2a). Notably, db/db Ca v 2.2 −/− mice had or tended to have lower levels of blood glucose than db/db Ca v 2.2 +/+ mice during the experimental period (Fig. 2a). Sixteen-hour fasting serum insulin levels were not different among db/db mouse groups at 16 weeks of age (Fig. 2b). Intraperitoneal glucose tolerance tests (IPGTTs) were performed to further evaluate glucose metabolism at 15 weeks of age. The blood glucose levels in GTTs peaked at 30 min were 250 mg/dL in db/+ Ca v 2.2 −/− mice and 400 mg/dL in db/+ Ca v 2.2 +/+ mice, indicating better glucose tolerance in db/+ Ca v 2.2 −/− mice than that in db/+ Ca v 2.2 +/+ mice (Fig. 2c). db/db Ca v 2.2 +/+ mice developed severe glucose intolerance, whereas db/db Ca v 2.2 −/− mice exhibited significantly reduced blood glucose levels compared with those in db/db Ca v 2.2 +/+ mice (Fig. 2d). The level of HbA1c was also reduced in db/db Ca v 2.2 −/− mice compared with db/db Ca v 2.2 +/+ mice (Fig. 2e). Serum insulin levels of db/db mouse groups in GTTs were increased compared with those of db/+ mouse groups (Fig. 2f). There was no significant difference among db/+ mouse groups. On the other hand, the insulin levels in db/db Ca v 2.2 −/− and db/db Ca v 2.2 +/− mice were significantly higher than those of db/db Ca v 2.2 +/+ mice, suggesting that insulin secretion increased in mice with Ca v 2.2 gene deletion.
Insulin tolerance tests (ITTs) were performed to determine whether the improved glucose tolerance observed in Ca v 2.2 −/− mice was associated with increased insulin sensitivity. Ca v 2.2 −/− mice showed lower glucose levels than Ca v 2.2 +/+ mice in both db/+ and db/db genotypes after insulin injection (Fig. 2g,h). These results indicate that deficiency of Ca v 2.2 improves insulin secretion and insulin sensitivity in diabetic conditions.

Reduced Urinary Albumin Excretion and Improved Hyperfiltration in Diabetic Ca v 2.2 −/−
Mice. To evaluate the functional alterations in the kidney of diabetic Ca v 2.2 −/− mice, we examined urinary albumin excretion and serum creatinine level and calculated creatinine clearance (CCr). At baseline, there were no significant differences in urinary albumin excretion between db/+ Ca v 2.2 +/+ and db/+ Ca v 2.2 −/− mice (Fig. 3a). Urinary albumin excretion markedly increased in db/db Ca v 2.2 +/+ at 8 weeks of age. In contrast, db/db Ca v 2.2 −/− mice exhibited approximately 70% lower urinary albumin excretion than db/db Ca v 2.2 +/+ mice. Interestingly, db/db Ca v 2.2 +/− mice also exhibited decreased albuminuria to the level comparable to that of db/db Ca v 2.2 −/− mice (Fig. 3a). These results suggest that even a partial ablation of the N-type calcium channel leads to reduction in urinary albumin excretion.
exhibited CCr elevation because of hyperfiltration, the increase in CCr was almost completely abolished in db/db Ca v 2.2 −/− mice, suggesting that hyperfiltration was normalized by deletion of the Ca v 2.2 gene (Fig. 3c).

N-type Calcium Channel Expression in Glomeruli of Control Mice and Renal
Histological Improvement in Diabetic Ca v 2.2 −/− Mice. N-type calcium channel localization was examined by an immunohistochemical study. Ca v 2.2 was expressed in tubules and glomerular cells in the kidney of both db/+ Ca v 2.2 +/+ and db/db Ca v 2.2 +/+ mice (Fig. 3d). Double immunohistochemical staining showed that cells positive for Ca v 2.2 in a glomerulus were also positive for WT1, a podocyte marker, indicating that Ca v 2.2 is expressed in podocytes (Fig. 3e).  We examined renal histology at 16 weeks of age. We observed mesangial expansion with glomerular hypertrophy in db/db Ca v 2.2 +/+ mice, which was consistent with diabetic alterations (Fig. 4a). In contrast, db/db Ca v 2.2 −/− mice exhibited reduced glomerular mesangial expansion and inhibited glomerular hypertrophy compared with those seen in db/db Ca v 2.2 +/+ mice (Fig. 4a). db/db Ca v 2.2 +/− mice also showed ameliorated glomerular changes. Morphometric analysis revealed that the mesangial area was increased in db/db Ca v 2.2 +/+ mice, whereas this increase was significantly suppressed in both db/db Ca v 2.2 −/− and db/db Ca v 2.2 +/− mice (Fig. 4b). These results suggest that N-type calcium channel ablation can limit the progression of diabetic nephropathy. Next, we evaluated podocyte injury in these mice. Immunostaining of nephrin and podocin, expressed predominantly in podocytes, was markedly decreased in db/db Ca v 2.2 +/+ mice compared with that in db/+ Ca v 2.2 +/+ mice (Fig. 4c). In contrast, db/db Ca v 2.2 −/− mice maintained the expression of nephrin and podocin to the same level as db/+ mice, indicating the amelioration of podocyte injury (Fig. 4c). In electron microscopic analysis, db/db Ca v 2.2 +/+ mice showed thickening of the glomerular basement membrane (GBM) with slightly widened podocyte foot processes (Fig. 4d,e). GBM thickening was significantly ameliorated in db/db Ca v 2.2 −/− mice (Fig. 4e).

Glomerular Gene Expression and Phosphorylation of Extracellular Signal-Regulated Kinase (ERK) in Diabetic Ca
Analyses of the glomerular expression of extracellular matrix (ECM)-related genes revealed that TGF-β 1 (Tgfb1) mRNA as well as connective tissue growth factor (Ctgf) mRNA were increased in db/db Ca v 2.2 +/+ mice, whose increase was significantly reduced in db/db Ca v 2.2 −/− mice (Fig. 5a,b). Expression of pro-alpha 3 chain of collagen IV (Col4a3) mRNA was also significantly reduced in db/db Ca v 2.2 −/− mice compared with db/db Ca v 2.2 +/+ mice (Fig. 5c). Gene expression of fibronectin (Fn1) and pro-alpha 1 chain of collagen I (Col1a1) tended to decrease in db/db Ca v 2.2 −/− mice ( Supplementary Fig. S1a,b). Glomerular gene expression of Cacna1b, Ca v 2.2, was not altered in diabetic mice, was reduced in Ca v 2.2 +/− mice and was not detected in Ca v 2.2 −/− mice (Fig. 5d). Ca v 2.2 +/− or Ca v 2.2 −/− mice exhibited similar expression of Cacna1c, α 1 subunit of L-type calcium channel as Ca v 2.2 +/+ mice (Fig. 5e). Glomerular expression of Cana1g, and Cacna1g (f) were shown. Gapdh was used as control. * p < 0.05, * * p < 0.01.
Activation of extracellular signal-regulated kinase (ERK) has been shown to mediate TGF-β -induced accumulation of ECM protein in diabetic nephropathy 18 . We found that ERK phosphorylation was increased in glomeruli, including mesangial cells and podocytes, of db/db Ca v 2.2 +/+ mice compared with that of db/+ Ca v 2.2 +/+ mice (Fig. 6a). Phosphorylation of ERK was significantly lower in the glomeruli of db/db Ca v 2.2 −/− mice than db/ db Ca v 2.2 +/+ mice (Fig. 6a). Macrophages also play a critical role in the progression of diabetic nephropathy 19 . The immunohistochemical study showed that macrophage antigen-2 (Mac2)-positive cells in glomeruli increased by 3.0-fold in db/db Ca v 2.2 +/+ mice compared with those in db/+ Ca v 2.2 +/+ mice (Fig. 6b,c). This increase was significantly suppressed in db/db Ca v 2.2 −/− mice (Fig. 6b,c).

Pharmacological Inhibition of L-or N-type Calcium Channel Ameliorates Diabetic
Nephropathy. To evaluate the pharmacological effect of N-type CCBs, we administered the N-/L-type CCB cilnidipine or the L-type CCB nitrendipine to db/db Ca v 2.2 +/+ mice and compared with db/db Ca v 2.2 +/− and db/db Ca v 2.2 −/− mice. There was no significant difference in body weight among vehicle-, nitrendipine-, cilnidipine-treated, db/db Ca v 2.2 +/− and db/db Ca v 2.2 −/− mouse groups (Supplementary Fig. S2a). Diabetic db/db Ca v 2.2 −/− mice exhibited lower blood glucose level than db/db Ca v 2.2 +/− mouse as shown previously (Fig. 2a, Supplementary Table S1). Diabetic mice with nitrendipine exhibited high urinary noradrenaline and adrenaline excretion, however, diabetic mice with cilnidipine did not change urinary catecholamine levels compared with those with nitrendipine ( Supplementary Fig. S2b,c). Administration of nitrendipine or cilnidipine showed SBP almost similar to that in the vehicle and db/db Ca v 2.2 +/− mouse groups (Fig. 1d, Supplementary  Fig. S3a). Diabetic Ca v 2.2 −/− mice showed lower SBP than vehicle, nitrendipine-, cilnidipine-treated and db/db Ca v 2.2 +/− mice (Fig. 1d, Supplementary Fig. S3a). Urinary albumin excretion was suppressed in the cilnidipine-treated group and not in the nitrendipine-treated group, however, urinary albumin excretion in cilnidipine-treated mice was still higher than that in db/db Ca v 2.2 +/− mice (Supplementary Fig. S3b). Renal histology showed that treatment with cilnidipine, but not with nitrendipine, inhibited mesangial expansion in diabetic mice to the comparable extent of db/db Ca v 2.2 +/− mice ( Supplementary Fig. S3c,d). In electron microscopic analysis, footprocesses widening and GBM thickening in diabetic mice was ameliorated only by cilnidipine treatment, which was comparable to db/db Ca v 2.2 −/− mice ( Supplementary Fig. S3e,f). Glomerular expression of TGF-β 1 (Tgfb1) and connective tissue growth factor (Ctgf) mRNA tended to show a reduction in cilnidipine-treated mice, but the difference was not significant (Supplementary Fig. S4a,b). Accumulation of macrophages was decreased in cilnidipine-treated mice compared with vehicle-treated mice ( Supplementary Fig.  S4c,d).

The Functional Role of N-type Calcium Channel on Cultured Podocytes.
First of all, to examine the role of the N-type calcium channel on podocytes, we measured intracellular Ca 2+ ([Ca 2+ ] i ) concentration in cultured human podocytes. When podocytes were stimulated with 107 mM KCl, depolarization-dependent [Ca 2+ ] i increase was observed, and this [Ca 2+ ] i concentration was partially abolished by treatment with the N-type calcium channel blocker, ω -conotoxin (Fig. 7a,b). We also found that nifedipine, cilnidipine, or ω -conotoxin plus nifedipine inhibited depolarization-induced [Ca 2+ ] i in cultured podocytes (Fig. 7c,d). These results suggest that both N-type and L-type calcium channels are expressed in cultured human podocytes and are relevant to depolarization-induced [Ca 2+ ] i increase.
To confirm the effect of N-type calcium channel blockade on podocytes, we examined the changes in nephrin expression in cultured human podocytes treated with ω -conotoxin. Administration of exogenous TGF-β resulted in a decreased expression of nephrin in podocytes (Fig. 7e). This decrease was significantly reversed by pre-incubation with ω -conotoxin (Fig. 7e). Nitrendipine did not change the TGF-β -induced reduction of nephrin expression, but cilnidipine upregulated nephrin expression in podocytes (Fig. 7f).

Discussion
Investigation of diabetic nephropathy in rodents is rendered difficult partly by the lack of adequate animal models displaying typical diabetic nephropathy 20 . Among these limited mouse models of diabetic nephropathy, db/db mice are one of the most frequently used disease models. Nevertheless, the genetic background plays an important role in developing diabetic nephropathy; e.g., db/db mice on the C57BLKS background exhibit massive proteinuria and mesangial expansion, whereas db/db mice on the C57BL/6J show less severe glomerular and glycemic changes 20,21 . Knockout mice are mostly generated on the C57BL/6J or 129/SvJ backgrounds 22 . In order to overcome these situations, we backcrossed Ca v 2.2 knockout mice on the C57BL/6J background with C57BLKS to explore the role of the N-type calcium channel in diabetic nephropathy.
The present study demonstrated that glycemic control was improved with enhanced insulin secretion in diabetic mice by ablation of the N-type calcium channel. In a previous study, Ca v 2.2 −/− mice showed lower fasting glucose levels and better glucose tolerance than wild-type mice without any change in insulin sensitivity upon GTTs 23 . The same study reported that, after 10 weeks of high-fat diet feeding, Ca v 2.2 −/− mice still showed lower fasting glucose levels and better glucose tolerance than Ca v 2.2 +/+ and Ca v 2.2 +/− mice. The mechanisms how Ca v 2.2 deletion resulted in better glycemic control have not been clarified yet, but another report has shown that the N-type calcium channel is present on pancreatic α cells and that GLP-1 inhibits glucagon release by selectively suppressing this channel 24 . In our study, db/db Ca v 2.2 +/− mice showed a marginally improved glucose tolerance; however, the reduction in urinary albumin excretion was much larger than expected from the degree of glycemic control, suggesting that mechanisms other than the amelioration in glucose metabolism would contribute to renoprotective effects, particularly in db/db Ca v 2.2 +/− mice.
Diabetic Ca v 2.2 −/− mice showed lower SBP with a marked reduction in urinary catecholamine levels. Similarly, a previous report showed that Ca v 2.2 −/− mice exhibited lower SBP than control mice because of vasodilatation, reduction of heart contractile activity, and inhibition of sympathetic nerve activity 25 . Most notably in our study, urinary albumin excretion was reduced by 70% and renal hyperfiltration was normalized in db/db Ca v 2.2 −/− mice. In addition, even diabetic Ca v 2.2 +/− mice, in which the expression of the N-type calcium channel is ~50% less than wild-type mice, showed a decrease in albuminuria to the level comparable to that in diabetic Ca v 2.2 −/− mice. The former denied reduction in sympathetic nerve activity, suggesting that partial inhibition of the N-type calcium channel could ameliorate albuminuria without affecting sympathetic nerve function.
Cilnidipine has been shown to ameliorate glomerular hypertrophy by dilating both afferent and efferent arterioles in the kidney 12,16 , and L-type calcium channel blockade does not improve glomerular hypertension because the L-type calcium channel blockade mainly dilates afferent arterioles 15 . Because N-type calcium channels exist at synaptic nerve endings 26 in both the afferent and efferent arterioles, and the blockade of the N-type calcium channel inhibits norepinephrine release 26 , such sympatholytic effect of the N-type calcium channel ablation may have worked to ameliorate glomerular injury in our study. Cilnidipine showed similar antihypertensive effects and suppression of proteinuria both in innervated and denervated spontaneous hypertensive rat (SHR) 27 , thus suggesting that renal sympathetic nerves may have a limited contribution to its renoprotective effects. In our study, albuminuria as well as glomerular histological changes were significantly alleviated in db/db Ca v 2.2 +/− mice and cilnidipine-treated diabetic mice without reduction of urinary catecholamine levels, further providing a possibility for mechanisms independent of the sympathetic nerve activity. In addition, we showed no compensatory expression of α 1 subunits of L-and T-type calcium channel in Ca v 2.2 heterozygous or knockout mice. Further study is necessary to distinguish the effects of N-and T-type calcium channel blocker, because blockade of T-/L-type calcium channel by manidipine 28 or efonidipine 29 exhibit similar effects as N-/L-type calcium channel blocker, cilnidipine, in terms of amelioration of glomerular hypertension.
As diabetic N-type calcium channel knockout mice improved metabolic parameters including glycemic control and blood pressure, the amelioration of diabetic renal injury in N-type calcium channel knockout mice is partly due to simultaneous improvements of metabolic parameters. However, in addition to these actions, we showed possible functional roles of the N-type calcium channel in podocytes. Previous reports showed that glomerular podocytes express the N-type calcium channel 13,27 . Fan et al. demonstrated the immunoreactivity of N-type calcium channels in kidney vascular walls, possibly in the nerves in adventitia, distal tubules, and podocytes, and that the N-type calcium channel in cultured podocytes was involved in the angiotensin II-induced production of reactive oxygen species 13 . In the present study, we also found that the N-type calcium channel is expressed in glomeruli, presumably podocytes, in control mice. The genetic inhibition of N-type calcium channels reduced podocyte injury in diabetic mice. In vitro analysis using calcium imaging revealed that N-type calcium channels as well as L-type calcium channels are functional in depolarization-induced [Ca 2+ ] i increase in cultured human podocytes. Furthermore, we examined the changes in nephrin expression induced by TGF-β stimulation in cultured human podocytes and revealed that the decrease caused by the administration of exogenous TGF-β was canceled by pre-incubation with ω -conotoxin, cilnidipine or MEK inhibitor. High glucose-induced ERK activation in podocytes is closely associated with diabetic nephropathy through the protein kinase C pathway 30,31 , suggesting that ERK plays an important role in TGF-β -induced podocyte injury.
In conclusion, we have demonstrated that the ablation or blockade of the N-type calcium channel in diabetic mice exerts renoprotective effects, which effects may be brought about by both improvement of metabolic parameters and protection from podocyte injury. These results indicate that the N-type calcium channel works as an aggravating factor in a mouse model of diabetic nephropathy, suggesting a possibility that the N-type calcium channel should provide a promising therapeutic target for preventing the progression of diabetic nephropathy in humans.

Animals and drug treatment. All animal experiments were approved by the Animal Experimentation
Committee of Kyoto University Graduate School of Medicine and were carried out in accordance with the approved guidelines. Mice deficient in the α 1B subunit of the N-type calcium channel (Ca v 2.2 −/− mice) were produced on the 129/SvJ background 17 and then backcrossed with C57BL/6J mice more than 10 times. BKS. Dg-Dock7 m + /+ Lepr db /J mice (db/m) mice on the C57BLKS/J background were purchased from Clea Japan Co. Ltd (Tokyo, Japan). Ca v 2.2 −/− mice were backcrossed with db/m mice on C57BLKS/J more than six times. We prepared six groups as follows: db/+ Ca v 2.2 +/+ mice, db/+ Ca v 2.2 +/− mice, db/+ Ca v 2.2 −/− mice, db/db Ca v 2.2 +/+ mice, db/db Ca v 2.2 +/− mice, and db/db Ca v 2.2 −/− mice. Male diabetic db/db mice and their non-diabetic db/+ mice (control) were used in this study. Blood pressure was measured by the tail-cuff method (MK-2000ST; Muromachi Kikai, Tokyo, Japan) every 4 weeks 32 . Urine samples were collected using metabolic cages every 2 weeks for measurement of creatinine and albumin 28 . Thereafter, mice were sacrificed at 16 weeks of age.
Nitrendipine (15 mg/kg/day; a gift from Tanabe Mitsubishi pharmaceutical company), cilnidipine (15 mg/kg/ day; a gift from Mochida pharmaceutical company) or vehicle were mixed with powdered food at 150-187.5 μ g/g, the concentration of which is adjusted by weekly food intake. CCBs were administered to db/db Ca v 2.2 +/+ mice from 9 weeks to 16 weeks of age.
Calcium imaging. Calcium imaging was performed as described previously with some modification 39 .
Human podocyte cells were plated onto poly-L-lysine-coated glass coverslips and subjected to measurement 3-16 h after plating on the coverslips. The cells on coverslips were loaded with fura-2 in RPMI 1640 containing 1 μ M fura-2-acetoxymethyl ester (fura-2-AM; Dojindo Laboratories, Kumamoto, Japan), 10% fetal bovine serum (Sigma), 10 μ g/mL insulin, 5.5 μ g/mL transferrin, 5 ng/mL selenium (Sigma), 30 units/mL penicillin, and 30 μg/ mL streptomycin at 37 °C for 30 min. The coverslips were then plated in a perfusion chamber mounted on the stage of the microscope. The fura-2 fluorescence images of the cells were recorded in HEPES-buffered saline (HBS, 2 mM Ca 2+ ) (in mM): 107 NaCl, 6 KCl, 1.2 MgSO 4 , 2 CaCl 2 , 1.2 KH 2 PO 4 , 11.5 glucose, and 20 HEPES (pH 7.4 adjusted with NaOH). Four minutes after image acquisition, the cells were stimulated with 107 mM K + solution (in mM): 107 KCl, 6 NaCl, 1.2 MgSO 4 , 2 CaCl 2 , 1.2 KH 2 PO 4 , 11.5 glucose, and 20 HEPES (pH 7.4 adjusted with NaOH). All the reagents dissolved in water or dimethylsulfoxide were diluted to their final concentrations and applied to the cells by perfusion. Fluorescence images of the cells were recorded and analyzed with a video image analysis system (AQUACOSMOS; Hamamatsu Photonics, Shizuoka, Japan). The ratio of the fluorescence intensity at 340 nm to the intensity at 380 nm was calculated to evaluate the change in intracellular calcium levels.
Statistical analysis. Data are expressed as the mean ± SEM. Statistical analysis was performed using one-way ANOVA. P < 0.05 was considered statistically significant.