Nisoldipine improves the impaired erythrocyte deformability correlating with elevated intracellular free calcium-ion concentration and poor glycaemic control in NIDDM

Abstract

Aims

To explore the mechanisms underlying the impaired erythrocyte deformability (RBC-df) in diabetic patients, the relationship between erythrocyte intracellular free calcium-ion concentration ([Ca²⁺]i) and RBC-df, and the effects of Ca²⁺-channel blocker on [Ca²⁺]i and RBC-df were evaluated.

Methods

Forty-eight patients with NIDDM and 24 control subjects were enrolled in this study. [Ca²⁺]i was determined using fura-2, and RBC-df by filtration method expressed as Deformability Index (DI). Erythrocytes were treated with nisoldipine to evaluate the effects of a Ca²⁺-channel blocker.

Results

[Ca²⁺]i was significantly higher (82.6 (78.0–87.2) vs 76.6 (74.3–81.2) nmol lRBC⁻¹, P<0.001), and DI was significantly lower (0.14 (0.09–0.28) vs 0.22 (0.16–0.28), P<0.01) in NIDDM than in controls. There was a significant correlation between HbA_1c and [Ca²⁺]i (r=0.38, P<0.01), between HbA_1c and DI (r=−0.51, P<0.01), and between [Ca²⁺]iand DI (r=−0.42, P<0.01). Stepwise multiple regression analysis revealed HbA_1cand [Ca²⁺]i as independent determinants for the impaired RBC-df. Nisoldipine treatment in vitro significantly decreased [Ca²⁺]i, and significantly improved RBC-df.

Conclusions

These data indicate that the impaired RBC-df in NIDDM may at least partly be attributed to the elevated [Ca²⁺]i and poor glycaemic control. In addition, favorable effects of a Ca²⁺-channel blocker on both [Ca²⁺]i and RBC-df have been demonstrated.

Introduction

Erythrocyte deformability (RBC-df) has been proposed to play an important role in the pathogenesis of diabetic vascular complications [1, 2]. Many studies have shown that RBC-df is decreased in patients with diabetes mellitus [3–12].

While various mechanisms have been attributed to the reduced RBC-df observed in diabetes, the nature of the primary defects remains elusive. A number of reports have revealed a critical role for maintaining low erythrocyte intracellular free calcium-ion concentration ([Ca²⁺]i) levels to maintain normal erythrocyte mechanical properties [13–16]. In addition, [Ca²⁺]i has been reported to be elevated in diabetic patients [17, 18]. Accordingly, one additional engaging working hypothesis revolves around the dysfunction of erythrocyte calcium homoeostasis in diabetes.

In this context, it is important to note that Ca²⁺-channel blockers suppress Ca²⁺ influx into erythrocytes on the one hand [19, 20], and improve RBC-df on the other hand [21–24]. Thus, patients with disorders associated with reduced RBC-df, including diabetic patients, might benefit clinically not only from the antihypertensive effects of Ca²⁺ channel blockers, but also from favourable effects on RBC-df.

Therefore, to explore the mechanism underlying the impaired RBC-df in diabetic patients, we have evaluated the relationship between [Ca²⁺]i and RBC-df in patients with NIDDM. Furthermore, the effects of nisoldipine, a 1,4-dihydropyridine Ca²⁺-channel blocker, on [Ca²⁺]i and RBC-df have also been investigated.

Methods

Subjects

Forty-eight patients with NIDDM (23 M; 25 F; aged 30 to 64 years), randomly selected from the patients attending our outpatient clinic at Kyoto University Hospital, were included in this study (DM group). The criteria of inclusion in this study were as follows: 1) NIDDM diagnosed by WHO criteria (1985), 2) serum creatinine level of less than 130 mmol l⁻¹, 3) no clinical or laboratory evidence of other kidney or renal tract diseases. None of the patients had a past history of cardiovascular disease. Fifteen patients were taking Ca²⁺-channel blockers, ten patients ACE inhibitors, seven patients β-adrenoceptor blockers, and four patients diuretics. None of the antihypertensive agents was discontinued or changed in its dosage throughout the study. Twenty-four healthy, normotensive volunteers (12 M; and 12 F; aged 28 to 60 years) were enrolled as normal controls (control group). This study was performed in accordance with the principles of the Helsinki declaration and approval for the study was given by the Ethics Committee of Kyoto University. Written informed consent was given by all subjects.

Classification of nephropathy was as follows. Three consecutive, sterile overnight urine collections were performed for measurement of urinary albumin excretion rate (AER), using radioimmunoassay. The median value of three specimens was used for classifying the patients into three categories: normoalbuminuria; AER <20 μg min⁻¹, microalbuminuria; AER 20–200 μg min⁻¹, macroalbuminuria; AER >200 μg min⁻¹. The patients were diagnosed to have diabetic nephropathy when they had microalbuminuria or macroalbuminuria.

Fundus examination was performed by the ophthalmologist after mydriasis once a year. The findings were graded as: 1) no signs of diabetic retinopathy; 2) simple retinopathy; or 3) proliferative retinopathy.

All subjects were weighed in indoor clothing without shoes at every visit, and height was also recorded. Blood pressure (phase I and V) was measured at every visit three times in a sitting position after at least 15 min rest by a standard mercury sphygmomanometer, with cuffs adapted to arm circumference. The median value of the three readings was reported as the value of the visit. Hypertension was diagnosed when the median value of the three consecutive office visits was above 140 mmHg (systolic blood pressure) and/or 90 mmHg (diastolic blood pressure), or the patient was already on antihypertensive medication.

Venous blood was collected after an overnight fast into heparin-treated tubes. Blood glucose, triglycerides, total cholesterol, HDL-cholesterol, and HbA_1c were measured. HbA_1c was determined by high-pressure liquid chromatography with a normal range of 4.0–5.8%, and other measurements were performed by routine laboratory methods by an auto-analyzer.

Erythrocyte count, haemoglobin concentration, haematocrit, MCV, MCH, MCHC, WBC count, and platelet count were determined by Coulter counter. There was no difference in all of these haematological parameters between control group and DM group (data not shown). An aliquot of the blood sample was processed for the measurements of [Ca²⁺]i and RBC-df.

[Ca²⁺]i measurements

Measurement of [Ca²⁺]i was performed using fura-2 according to the method reported in detail in our previous report [25], based on the method described by David-Dufilho et al. [26], with slight modifications. Briefly, erythrocytes were suspended at 1% haematocrit in buffer A (123 mmol l⁻¹ NaCl, 5 mmol l⁻¹ KCl, 1mmol l⁻¹ MgCl₂,1 mmol l⁻¹ CaCl₂, 10 mmol l⁻¹ glucose, 25 mmol l⁻¹ HEPES, pH 7.4 at 37° C), and then incubated for 60 min at 37° C with or without 0.5 μmol l⁻¹ fura-2-acetoxymethylester. Erythrocytes were washed three times with buffer A, and then diluted to 0.1% haematocrit in buffer B (buffer A contains additional 1 mmol l⁻¹ MnCl₂ to quench the extracellular fura-2 signals). Fluorescence was measured by fluorescent spectrophotometer (Shimadzu RF-5000, Kyoto, Japan) in quartz cuvettes. The excitation wavelengths were 340 nm and 380 nm with 3 nm bandwidth and the emission wavelength was 500 nm with 10 nm bandwidth. [Ca²⁺]i was calculated according to the equation described by Grinkiewicz et al. [27]. Fluorescence intensities were calculated by subtraction of the fluorescence of fura-2 unloaded erythrocytes (intrinsic fluorescence) measured at 340 and 380 nm from those of fura-2 loaded erythrocytes. The parameters used in the equation were determined from fura-2-calcium fluorescence calibration experiments with EGTA-calcium buffers similar to the intracellular conditions (10 mmol l⁻¹ NaCl, 120 mmol l⁻¹ KCl, 0.4 mmol l⁻¹MgCl₂, 10 mmol l⁻¹ glucose, 25 mmol l⁻¹ HEPES, 21 g l⁻¹ polyvinylpyrrolidone, and variable Ca²⁺ concentrations).

Deformability measurements

Determination of RBC-df was performed according to the method described by Brown et al. [28] based on the guidelines set by the International Committee for Standardization in Haematology, Expert Panel on Blood Rheology [29] with slight modifications. At first, whole blood was filtered through cotton wool removed from a leukocyte filter (Imugard 500, Terumo, Tokyo, Japan) [30]. After high-speed centrifugation, plasma and uppermost layer of erythrocytes were aspirated. The remaining erythrocytes were washed three times with isotonic PBS (NaCl 8.0 g, KCl 0.2 g, Na₂HPO₄.12H₂O 2.9 g, KH₂PO₄ 0.2 g, and albumin 5 g up to 1 l, pH=7.4). Washed erythrocytes, aspirated from the middle of the packed erythrocyte column, were resuspended in isotonic PBS to a final concentration of 5%. Virtually no contamination of WBC was observed by this method. RBC-df was expressed as a deformability index (DI), defined as the time required for 5 ml of PBS to pass through a 5 μm pore filter (Nucleopore, Pleasanton, USA) under a constant negative pressure of −20 cm H₂O at 37° C, divided by the time required for 5 ml of erythrocyte suspension. In this definition, higher DI means better deformability. The DI was reported as the average of three repeated measurements.

Effects of nisoldipine

Nisoldipine treatment in vitro was performed by incubating the fura-2 loaded erythrocytes, or 5% erythrocyte suspension in the respective buffer containing 10⁻⁷ mol l⁻¹ of nisoldipine (from 10⁻⁴ mol l⁻¹ stock solution dissolved in ethanol) for 10 minutes at 37° C. Erythrocytes treated with 0.01% ethanol alone were used as controls. Treatment with ethanol alone did not influence the [Ca²⁺]i and RBC-df. Then the respective erythrocyte suspension was used for the determination of [Ca²⁺]i and DI.

Statistical analysis

Data were expressed as median and (range). Statistical analyses were performed using Wilcoxon’s rank-sum test (non-matched pairs), Wilcoxon’s signed-ranks test (matched pairs), Spearman’s correlation coefficient, and stepwise multiple regression analysis. P values less than 0.05 were considered to be statistically significant.

Results

Clinical characteristics of the subjects are shown in Table 1. Plasma triglycerides and sBP were significantly higher, and HDL-cholesterol was significantly lower in the DM group than in the control group.

Table 1.

Clinical characteristics of the control subjects and NIDDM patients.

[Ca²⁺]i and erythrocyte deformability

[Ca²⁺]i was significantly higher in the DM group than in the control group (Figure 1a; 82.6 (78.0–87.2) nmol lRBC⁻¹vs 76.6 (74.3–81.2) nmol lRBC⁻¹; P<0.001). In contrast, DI was significantly lower in the DM group than in the control group (Figure 1b; 0.14 (0.09–0.28) vs 0.22 (0.16–0.28); P<0.01). In the DM group, there was no difference of [Ca²⁺]i and DI among the subgroups divided on the basis of nephropathy, retinopathy, hypertension, or taking a Ca²⁺-channel blocker (data not shown).

Figure 1.

a) Comparison of [Ca²⁺]i between control group and DM group and b) comparison of RBC-df (DI) between control group and DM group. The difference between two groups was analyzed using Wilcoxon’s rank-sum test. In brackets number of cases. Horizontal bars represent median.

Erythrocyte deformability and glycaemic control

Univariate regression analyses were applied to the data obtained in the DM group. There was a significant positive correlation between HbA_1c and [Ca²⁺]i (Figure 2a; n=48, r=0.38, P<0.01). In contrast, there were significant inverse correlations between HbA_1c and DI (Figure 2b; n=48, r=−0.51, P<0.01), and between [Ca²⁺]iand DI (Figure 2c; n=48, r=−0.42, P<0.01). There is no significant relationship between blood pressure, RBC-df, and [Ca²⁺]i.

Figure 2.

a) Correlation between HbA_1c and [Ca²⁺]i, b) correlation between HbA_1c and RBC-df (DI), c) correlation between [Ca²⁺]i and RBC-df (DI). The relationship between two variables was analyzed using Spearman’s correlation coefficient.

To elucidate the independently contributing factors to decreased DI in the DM group, stepwise multiple regression analysis was carried out (Table 2). HbA_1cappeared as the first significant determinant for DI, and [Ca²⁺]i appeared as the second. All the other factors evaluated (gender, age, known duration of diabetes, existence of microalbuminuria or macroalbuminuria, existence of simple or proliferative retinopathy, BMI, sBP, dBP, triglycerides, total- and HDL-cholesterol) did not appear as significant determinants. In this model, 41% of the variance in DI could be explained by these two factors (P<0.0001).

Table 2.

Stepwise multiple regression analysis with RBC-df (DI) as dependent variable.

Effects of nisoldipine in vitro

After the in vitro nisoldipine treatment, [Ca²⁺]i was significantly decreased from 82.6 (78.0–82.2) nmol l RBC⁻¹ to 80.4 (77.0–85.3) nmol l RBC⁻¹ (Figure 3a; n=48, P<0.01), and DI was significantly improved from 0.15 (0.09–0.28) to 0.18 (0.09–0.31) in the DM group (Figure 3b; n=48, P<0.01). In addition, there was a significant positive correlation between the degree of [Ca²⁺]i decrease and the degree of DI increase (n=48, r=0.52, P<0.001). In contrast, there was no significant change in either [Ca²⁺]i or DI in the control group (n=24, [Ca²⁺]i: from 76.6 (74.3–81.2) nmol l RBC⁻¹ to 76.7 (74.3–81.4) nmol l RBC⁻¹; DI: from 0.22 (0.16–0.28) to 0.24 (0.17–0.30)).

Figure 3.

Effects of nisoldipine on [Ca²⁺]i and b) effects of nisoldipine on RBC-df. [Ca²⁺]i and DI were measured after the treatment with 10⁻⁷ mol l⁻¹ of nisoldipine or with the solvent (0.01% ethanol) alone for 10 min at 37° C. The difference between two groups was analyzed using Wilcoxon’s signed-ranks test. Horizontal bars represent median.

Discussion

This study is the first to evaluate [Ca²⁺]i and RBC-df simultaneously in diabetic patients, and has revealed significant correlations of impaired RBC-df with elevated [Ca²⁺]i and poor glycaemic control. In addition, favourable effects of nisoldipine, a dihydropyridine Ca²⁺-channel blocker, on both [Ca²⁺]i and RBC-df have concurrently been demonstrated for the first time.

It has been proposed that impaired RBC-df may play an important role in the pathogenesis of diabetic microangiopathy and macroangiopathy [1, 2]. The hypothesis proposes that stiffened erythrocytes would require raised perfusion pressure to overcome their resistance to flow. In the renal circulation, intra-glomerular hypertension would ensue as a result, which is thought to be one of the main pathogenic mechanisms of diabetic nephropathy [31]. Moreover, it is generally recognized that haemorheological factors, especially the mechanical property of erythrocytes, can play a major role in governing nutritive tissue perfusion at the level of the microcirculation [32]. Likewise, impaired RBC-df has been found in conditions associated with an increased risk for atherosclerosis, including diabetes mellitus [33, 34], peripheral vascular disease [35], cardiovascular diseases [24], and also in cerebrovascular disease [36]. Thus, reduced RBC-df appears to be one factor of importance also in the pathogenesis of diabetic macroangiopathy.

Accumulating evidence has shown that RBC-df is decreased in patients with diabetes mellitus [3–12], and in animal models of diabetes mellitus [28, 37]. Previously suggested mechanisms underlying reduced RBC-df observed in diabetes include hypoinsulinaemia [5], increased sorbitol concentration [6], increased erythrocyte membrane rigidity caused by glycation of the membrane itself [38], oxidation of spectrin [11], formation of AGEs in erythrocyte [28], and alterations in membrane lipid composition [12]. In accordance with these ideas, several studies, including the present study, have revealed a significant correlation between poor glycaemic control and decreased RBC-df [3, 10, 11].

Although the nature of the primary defects that lead to decreased RBC-df in diabetes mellitus remains elusive, one additional engaging working hypothesis revolves around the dysregulation of erythrocyte calcium homoeostasis. The report by Weed et al. [13] was the first to show increased [Ca²⁺]i to be associated with decreased erythrocyte membrane deformability. Since then, a number of reports have appeared establishing a critical role for maintaining low [Ca²⁺]i levels in normal erythrocyte mechanical properties [14–16]. It has also been reported that high [Ca²⁺]i values were associated with reduced RBC-df in hypertensive patients [19]. Until now, only one group has examined the relationship between [Ca²⁺]i and erythrocyte rheology in diabetic patients [15]. They reported a significant reverse correlation between [Ca²⁺]i and erythrocyte membrane fluidity determined by membrane protein lateral mobility. Although some differences in the methodology exist, our present results agree with them in this regard.

One hypothesis for the pathogenesis of hypertension relates to an alteration in intracellular ion metabolism [39]. Hypertensive diabetic patients have been shown to have increased intracellular sodium, pH and calcium. This alteration of intracellular milieu may lead to increased vasoconstriction and cellular proliferation which can contribute to an increase in blood pressure and vascular complications. However, there was no significant relationship between blood pressure and RBC-df and [Ca²⁺]i in our study. This lack of relationship might be due to the antihypertensive treatment.

[Ca²⁺]i has been reported to be elevated in diabetic patients [17, 18]. On the other hand, a decrease in erythrocyte Ca²⁺-ATPase activity has been found in patients with IDDM [40] or NIDDM [41, 42]. It has been suggested that this decreased Ca²⁺-ATPase activity would be due to glycation of the protein itself [43]. Since the plasma membrane Ca²⁺-ATPase is the only system for extruding Ca²⁺ from human erythrocytes [44], the elevated [Ca²⁺]i in diabetic patients might at least partly be attributed to the decreased Ca²⁺-ATPase activity. The significant positive correlation between HbA_1cand [Ca²⁺]i presented in our study supports thisidea.

Apart from the effects of Ca²⁺-ATPase, close relationships between intracellular calcium and sodium concentrations have been reported [39]. Unfortunately, we did not measure intracellular sodium concentrations in the present study, and further study will be needed in this regard.

It has been shown that Ca²⁺-channel blockers can suppress Ca²⁺ influx into erythrocytes [19, 20]. We also have reported that nisoldipine blocks the increase of [Ca²⁺]i in diabetic patients other than those enrolled in this study [45]. On the other hand, it has also been reported that Ca²⁺-channel blockers can improve RBC-df [21–24, 46]. Dihydropyridine binding sites, although somewhat different from those of excitable cells, have been found on human erythrocyte membranes [47]. The existence of a voltage-dependent and dihydropyridine-sensitive calcium influx pathway has also been shown in human erythrocytes [48, 49]. Taking these reported results into consideration, the improved RBC-df caused by Ca²⁺-channel blockers can at least partly be attributed to the effect of decreasing [Ca²⁺]i. Our present study is the first that has examined the effects of Ca²⁺-channel blockers on [Ca²⁺]i and RBC-df simultaneously, and that has demonstrated the results supporting this hypothesis.

It is of interest that Ca²⁺-channel blockers reduce [Ca²⁺]i only when the influx of Ca²⁺ and [Ca²⁺]i levels are increased [19, 20]. Furthermore, Ca²⁺-channel blockers are unable to influence the deformability of normal erythrocytes [23, 50]. Our results also agree with these reports. Therefore, only the patients suffering from disorders associated with elevated [Ca²⁺]i and reduced RBC-df, including diabetic patients, might benefit clinically not only from the antihypertensive effects of the drugs but also from improvement in RBC-df [24]. However, the concentration of nisoldipine used in this study was several to ten times as high as clinically attainable levels, that is around 10⁻⁸ mol l⁻¹ [23]. As the clinical dosage of a Ca²⁺-channel blocker did not seem to alter the [Ca²⁺]i and RBC-df levels in this study, the result is somewhat different from those reported previously [21–23]. These discrepancies might be due to the differences of the Ca²⁺-channel blockers used in the experiments, or of the blood levels of these drugs at the time of blood sampling. Further investigations are necessary before conclusions can be drawn about the clinical relevance of the effects of Ca²⁺-channel blockers on RBC-df.

Some studies have found significant correlations between the severity of vascular complications and decreased RBC-df [3, 4, 7], but other studies, including our present study, have not found such correlations [8, 9]. Since all of these studies are cross-sectional ones, prospective studies would be required to estimate the clinical significance of decreased RBC-df in the pathogenesis of diabetic complications. If diabetic vascular complications were caused at least partly by abnormal haemorheology, then the possibility exists that these might be alleviated or prevented by the agents, including Ca²⁺-channel blockers, that enhance RBC-df and improve blood rheology. Longitudinal studies are needed also in this regard.