Introduction

Diabetes is the leading cause of end-stage renal disease in the western world, affecting approximately 30% of type 1 diabetic patients [1]. Microalbuminuria is the earliest clinical marker of renal complications in diabetes [2]. Patients develop macroproteinuria as renal injury progresses and suffer from GFR reduction, which subsequently leads to renal failure.

In addition to vascular endothelium, components of the endothelin-1 system are located in many cell types in the kidney, including mesangial cells [3], podocytes [4] and tubular epithelium [5]. Aside from vasoconstriction [6], endothelin-1 is a potent proinflammatory and profibrotic mediator [7]. In diabetes, intrarenal endothelin-1 activity appears to be increased [8]. Glomerular endothelin-1 mRNA expression is elevated in rats with streptozotocin-induced diabetes [9]. Kidney endothelin receptor-A (ETA) expression is upregulated in rabbits with alloxan-induced diabetes [10]. Consistent with these observations is the finding that urinary endothelin-1 excretion (a marker of intrarenal endothelin-1 generation) is increased in diabetic patients with microalbuminuria [11]. Furthermore, our laboratory has demonstrated that endothelin-1 plasma levels were not changed, but endothelin-1 excretion was increased in rats after streptozotocin-induced hyperglycaemia [12].

Studies from several labs including our own have shown that chronic blockade of ETA reduces albuminuria and renal inflammation in the streptozotocin-induced rat model of diabetes, along with a modest decrease [12] or no significant change [13] in arterial pressure. However, the mechanism for this effect is not known. Inflammation is considered to be a contributing factor in the development of diabetic nephropathy and may contribute to proteinuria as well as interstitial fibrosis and cellular damage. Pro-inflammatory chemokines monocyte chemoattractant protein-1 (MCP-1) and soluble intercellular adhesion molecule-1 (ICAM-1) are important in attracting infiltrating cells and inducing their attachment to the endothelium, thus facilitating the early process of macrophage infiltration into the kidney [14]. Increased ICAM-1 expression leading to increased leucocyte infiltration has been demonstrated in experimental models of nephropathy [15].

Proteinuria represents an early sign of glomerular injury, the presence of which predicts not only an elevated risk of nephropathy, but also cardiovascular disease in general [16]. The mechanistic pathways of proteinuria in diabetes have not been fully resolved. One of the glomerular filtration barrier defects leading to glomerular injury and subsequent proteinuria is damage to the filtration-slit formed by glomerular podocytes. Nephrin, a 1241-residue transmembrane immunoglobulin protein, is an important filtration-slit molecule [17]. Levels and urinary excretion of nephrin show characteristic changes in diabetes and in other acquired proteinuric diseases [18].

The present study was undertaken to test the hypothesis that ETA contribute to glomerular inflammation and permeability defects in a rat model of hyperglycaemia by directly impairing glomerular permeability and promotion of glomerular inflammation. More specifically, we investigated whether endothelin-1 via the ETA has a direct effect on the glomerulus, thus inducing nephrin shedding and reducing glomerular permeability to albumin (Palb), which would provide a mechanism for diabetic proteinuria. We also investigated whether the anti-inflammatory actions of ETA blockade were due to changes in expression of early inflammatory response pathways such as MCP-1 and soluble ICAM-1. We used an in vivo model to specifically address these questions in the streptozotocin-induced rat model of type 1 diabetes.

Methods

Streptozotocin-induced hyperglycaemia

Experiments used male Sprague–Dawley rats (250–275 g) from Harlan Laboratories (Indianapolis, IN, USA). All protocols were approved by the Institutional Animal Care and Use Committee of the Medical College of Georgia and followed the NIH Public Health Service Policy on Humane Care and Use of Laboratory Animals. Rats were housed in constant temperature and humidity, and exposed to a 12 h light–dark cycle. We studied four groups of rats: (1) sham-injected controls; (2) hyperglycaemic; (3) sham + ABT-627 (5 mg kg−1 day−1, in drinking water); and (4) hyperglycaemic + ABT-627 (n = 8 in all groups). Hyperglycaemia was attained by injection of streptozotocin (Sigma-Aldrich, St Louis, MO, USA) at a dose of 65 mg/kg body weight, given intravenously through the penile vein under isoflurane anaesthesia; sham-injected animals received saline. ABT-627 is a selective ETA blocker (around 1000 times greater affinity for ETA vs endothelin receptor-B [ETB]) and provides maximum ETA blockade and selectivity at this dose in vivo [19]. Drug treatment was started 1 day after streptozotocin injection and after confirming that all rats had blood glucose levels >20 mmol/l. At the same time, insulin or palmitic acid (blank) implants (Linshin, Scarborough, ON, Canada) were inserted subcutaneously into the hyperglycaemic and sham-injected rats, respectively. Insulin implants maintained blood glucose levels at 20 to 25 mmol/l (Table 1). Glucose in whole blood (taken from a small incision on the tail) was measured with a glucometer (Accu-Chek; F. Hoffmann-La Roche AG, Basel, Switzerland). Glycaemia was monitored twice a week throughout the study. During the final 2 days of treatment, rats were placed in metabolism cages in order to collect urine for determination of protein excretion rates. At the end of all experiments, rats were anaesthetised using sodium pentobarbital (50 mg/kg; i.p.) and killed, and a blood sample immediately taken from the abdominal aorta and plasma stored at −80°C. Kidneys were removed and glomeruli isolated.

Table 1 Characteristics of the experimental rats after 3 and 6 weeks of treatment
Full size table

Glomerular isolation

Glomeruli were isolated by gradual sieving techniques [20]. Upon removal, kidneys were decapsulated and placed in ice-cold PBS (pH 7.4) containing phenylmethylsulfonylfluoride (PMSF) (1 mmol/l). The cortex was dissected and minced into small pieces. The cortical tissue was then passed through a 180 μm stainless steel sieve to separate glomeruli from larger fragments of renal tubules and vasculature. The resulting tissue was then passed through a 200 μm micro-cellulose filter. The filtrate was re-circulated on a smaller pore size micro-cellulose filter (70 μm). The glomeruli retained on top of the 70 μm sieve were washed with ice-cold PBS/PMSF into a 50 ml Eppendorf tube. The resulting decapsulated glomeruli devoid of afferent and efferent arterioles were re-suspended in ice-cold PBS buffer. Tubular contamination was verified to be less than 5% of the tissue as assessed under the light microscope. The glomerular suspension was then centrifuged (10,000 g, 10 min) and the pellet re-suspended in PBS. The glomeruli were washed one final time using the same centrifugation and the final pellet was re-suspended in 1 ml PBS.

For immunoassays, the final suspension of isolated glomeruli was snap-frozen in liquid nitrogen and stored at −80°C. The frozen glomeruli were re-suspended in lysis buffer (20 mmol/l HEPES, pH 7.4, 10 mmol/l NaCl, 5 mmol/l EDTA, 0.2% [vol./vol.] Triton X-100, 10 mmol/l sodium fluoride, 1 mmol/l sodium ortho-vanadate, 1 mmol/l PMSF, 1 μg/ml leupeptin and 1 μg/ml pepstatin) and homogenised by ultrasonic homogeniser (20 s). After centrifugation for 10 min at 10,000 g, the supernatant fraction was used for analysis and protein determined using the Bradford method (Bio-Rad Laboratories, Hercules, CA, USA) according to the manufacturer’s instructions.

Measurement of P alb

After isolation, glomeruli were re-suspended at room temperature in 5% (wt/vol.) BSA (containing 115 mmol/l NaCl, 5 mmol/l KCl, 10 mmol/l sodium acetate, 1.2 mmol/l dibasic sodium phosphate, 25 mmol/l sodium bicarbonate, 1.2 mmol/l magnesium sulphate, 1 mmol/l calcium chloride and 3.5 mmol/l glucose, pH 7.4).

The rationale and methodology for determination of albumin permeability has been described in detail previously [21]. In brief, images of 10–15 glomeruli per kidney preparation (i.e. per rat) were captured using a digital camera through an inverted microscope before and after a medium change to one containing 1% (wt/vol.) BSA. The medium exchange created an oncotic gradient across the basement membrane resulting in capillary expansion and a glomerular volume increase (ΔV = [V final–V initial]/V initial), which was measured off-line by an image analysis programme (Digimizer; MedCalc Software, Mariakerke, Belgium). The software determined the average radius of the glomerulus in two-dimensional space and the volume was then derived from the formula V = (4/3)π r3. The magnitude of ΔV was related to the albumin reflection coefficient, σalb, by the following equation: (σalb) experimental = (ΔV) experimental /(ΔV) control; the σalb of the control glomeruli was assumed to be equal to 1. Palb was defined as 1–σalb, and describes the movement of albumin subsequent to water flux. When σalb is zero, albumin moves across the membrane with the same velocity as water and Palb is 1.0. Conversely, when σalb is 1.0, albumin cannot cross the membrane with water and Palb is zero.

In additional experiments, we examined the role of ETA and ETB on Palb in glomeruli isolated from untreated hyperglycaemic rats (n = 6) using a selective ETA antagonist (BQ-123; Calbiochem, San Diego, CA, USA) and a selective ETB antagonist (BQ-788; Calbiochem). Glomeruli were pre-incubated with these antagonists at concentrations of 1 × 10−9 to 1 × 10−5 mol/l for 15 min at 37°C and the Palb response was determined. These experiments used a minimum of five glomeruli from each rat.

Biochemical analyses

Commercially available kits for soluble ICAM-1 (Quantikine sICAM-1 Immunoassay; R&D Systems, Minneapolis, MN, USA) and MCP-1 (RayBioTech, Norcross, GA, USA) were used to determine concentrations in plasma and glomerular homogenates. Nephrin concentration was determined in urine via an ELISA kit (Exocell, Philadelphia, PA, USA). Urinary protein concentrations were determined using the Bradford method.

Nephrin immunofluorescence

Isolated glomeruli taken from kidneys perfused with PBS were placed on engraved glass slides and allowed to dry at room temperature before freezing at −80°C. Glomeruli were then fixed using paraformaldehyde (2% vol./vol.) in a slide chamber (Antibody Amplifier; IHC World, LLC, Woodstock, MD, USA ). Slides were then washed with PBS and incubated for 1 h with normal goat serum (5% vol./vol.) in PBS and Triton-X (0.3% vol./vol.) before being incubated overnight on a shaker at 4°C with goat anti-human nephrin primary antibody (1:500 vol./vol.; sc-19000; Santa Cruz Biotechnology, Santa Cruz, CA, USA). On the second day, washing and blocking procedures were repeated before incubating slides for 1 h in the antibody amplifier chamber with a mixture of Alexa Fluor 488 chicken anti-goat IgG fluorescent-tagged secondary antibody (1:5,000 vol./vol.) and rhodamine phalloidin for F-actin staining (5 μmol/l). Both antibodies were purchased from Invitrogen (Carlsbad, CA, USA). Washing was repeated and slides mounted with glass coverslips. Images were acquired with the confocal/multiphoton system (FV-1000 MPE; Olympus, Center Valley, PA, USA) available at the Indiana Center for Biological Microscopy (Indianapolis, IN, USA). Confocal image stacks of whole glomeruli were collected with a 60 × W NA1.2 objective at 512 × 512 frame size and sequential mode using 488 and 559 nm laser lines. Nephrin and actin signals were collected with the emission filter at 520 and 612 nm, respectively. At least five glomeruli per rat were imaged. Metamorph software version 7.5 (Molecular Devices, Downingtown, PA, USA) was used for quantitative analysis. A sum of all planes intensities for both channels within a Z-stack was used to measure the total intensity values within the glomerulus. Background values from control stacks were subtracted from the total intensity values for both channels. Nephrin signals were then normalised to the corresponding actin signal by calculating the ratio from the background-corrected total intensity values.

Statistical analyses

All data are presented as mean ± SEM. Differences between data obtained from sham, sham + ABT-627, hyperglycaemic and hyperglycaemic + ABT-627 animals were compared using two-way ANOVA followed by Bonferroni post hoc tests. Differences between 3- and 6-week-old groups were compared by using the unpaired Student’s t test. A value of p < 0.05 was considered statistically significant. Analyses were performed using GraphPad Prism Version 5.0 software (GraphPad Software, La Jolla, CA, USA).

Results

As expected, blood glucose levels were elevated in hyperglycaemic rats with or without ETA blockade (Table 1). Similarly, blood glucose in sham-injected rats was not changed by treatment with ABT-627. Relative to sham and sham + ABT-627 rats, rats in the hyperglycaemic and hyperglycaemic + ABT-627 groups had lower body weights during the 3- and 6-week periods of study despite having hyperphagia, polydipsia and polyuria. Treatment with ABT-627 did not change any of these variables.

Glomerular permeability and proteinuria

Palb significantly increased after 3 weeks of hyperglycaemia when compared with sham. Treating animals with ABT-627 significantly decreased the elevated Palb value. Palb further increased significantly after 6 weeks of hyperglycaemia. Again, ABT-627 significantly decreased Palb after 6 weeks of treatment (Fig. 1a). Changes in urinary protein excretion followed a similar pattern to that of Palb. As shown in Fig. 1b, rats with hyperglycaemia exhibited significant proteinuria at 3 weeks compared with sham-injected rats. ABT-627 reduced proteinuria to levels comparable to sham. Proteinuria was further increased after 6 weeks of hyperglycaemia. Again, ABT-627 prevented the increase in proteinuria at the 6 week time-point.

Fig. 1
figure 1

Effect of streptozotocin-induced hyperglycaemia (HG), with or without concurrent treatment with ABT-627, on Palb (a) and daily protein excretion rate (b) at times indicated; n = 5–8 per group. Two-way ANOVA resulted in the following p values: Palb, 3 weeks, p HG = 0.0001, p ABT-627 = 0.0003, p HG×ABT-627 = 0.0002 and 6 weeks, p HG < 0.0001, p ABT-627 = 0.0001, p HG×ABT-627 < 0.0001; protein excretion rate, 3 and 6 weeks, p HG < 0.0001, p ABT-627 < 0.0001, p HG×ABT-627 < 0.0001. Post hoc analysis comparing individual means are denoted by *p < 0.05 vs sham group, † p < 0.05 vs hyperglycaemia, and ‡ p < 0.05 vs 3-week treatment

Full size image

To further determine the role of glomerular ETA and ETB in regulating Palb in hyperglycaemic rats, isolated glomeruli from hyperglycaemic rats were incubated for 15 min at 37°C with soluble endothelin-1 antagonist peptides. At 10−7 mol/l, the ETA antagonist, BQ123, significantly reduced the elevated Palb (Fig. 2a). Palb was further reduced at higher BQ123 concentrations. In contrast, the ETB antagonist, BQ788, had no effect on the elevated Palb of hyperglycaemic glomeruli at any concentration of that antagonist (Fig. 2b). Incubation of isolated glomeruli from hyperglycaemic rats with BQ123 and BQ788 reduced Palb in a fashion similar to BQ123 alone (Fig. 2c).

Fig. 2
figure 2

In vitro effect of selective ETA and ETB antagonists on Palb of isolated glomeruli from untreated hyperglycaemic rats after incubation with concentrations ranging from 1 × 10−9 to 1 × 10−5 mol for 15 min at 37°C. a BQ123 (1 × 10−5 to 1 × 10−9 mol/l) significantly reduced the elevated Palb of glomeruli isolated at 6 weeks of hyperglycaemia. b BQ788 had no effect on elevated Palb of hyperglycaemic glomeruli. c Addition of both antagonists produced the same effect as that of BQ123 alone. Hyperglycaemic rats used for treatment of glomeruli in vitro: n = 6; at least five glomeruli per rat were analysed. *p < 0.05, ***p < 0.001 vs zero concentration

Full size image

Systemic and glomerular inflammation

Soluble ICAM-1 and MCP-1 are two early pro-inflammatory molecules implicated in the pathogenesis of diabetic nephropathy. After 3 weeks of hyperglycaemia, we did not observe any changes in plasma or glomerular levels of soluble ICAM-1 and MCP-1 (Fig. 3a, b, Fig. 4a, b). After 6 weeks, however, plasma and glomerular soluble ICAM-1 and MCP-1 were significantly elevated compared with those in sham-injected animals. These increases were significantly attenuated by ABT-627.

Fig. 3
figure 3

Effect of streptozotocin-induced hyperglycaemia (HG), with or without concurrent treatment with ABT-627, on (a) plasma soluble ICAM-1 (sICAM-1) and (b) plasma MCP-1 concentrations at times indicated; n = 5–8 per group. Two-way ANOVA resulted in the following p values: plasma soluble ICAM-1, 6 weeks, p HG = 0.016, p ABT-627 = 0.033, p HG×ABT-627 < 0.0001; plasma MCP-1, 6 weeks, p HG = 0.0001, p ABT-627 = 0.0056, p HG×ABT-627 = 0.0008. Post hoc analysis comparing individual means are denoted by *p < 0.05 vs sham group, † p < 0.05 vs hyperglycaemia, and ‡ p < 0.05 vs 3-week treatment

Full size image
Fig. 4
figure 4

Effect of streptozotocin-induced hyperglycaemia (HG), with or without concurrent treatment with ABT-627, on (a) glomerular soluble ICAM-1 (sICAM-1) and (b) glomerular MCP-1 concentrations at times indicated; n = 5–8 per group. Two-way ANOVA resulted in the following p values: glomerular soluble ICAM-1, 6 weeks, p HG < 0.0001, p ABT-627 = 0.0153, p HG×ABT-627 = 0.0279; glomerular MCP-1, 6 weeks, p HG < 0.0001, p ABT-627 = 0.0097, p HG×ABT-627 = 0.0342. Post hoc analysis comparing individual means are denoted by *p < 0.05 vs sham, † p < 0.05 vs hyperglycaemia and ‡ p < 0.05 vs 3-week treatment

Full size image

Nephrinuria and glomerular nephrin expression

We monitored glomerular levels of nephrin by immunostaining using fluorescence microscopy. As depicted in Fig. 5, glomeruli isolated from 6-week hyperglycaemic rats exhibited significantly lower nephrin levels than those of sham (Fig. 6a). In addition, glomeruli isolated from ABT-627-treated hyperglycaemic rats had nephrin levels comparable to those of sham-injected animals (Fig. 6a). Nephrin is also excreted into urine in the early stages of diabetic nephropathy, which is an early sign of breakdown of the glomerular filtration barrier. Therefore, we determined whether nephrin loss in urine of hyperglycaemic rats is ETA-dependent. As expected, urinary excretion of nephrin was significantly increased after 6 weeks of hyperglycaemia, while ABT-627 administration completely prevented the hyperglycaemia-induced increase in nephrin excretion (Fig. 6b).

Fig. 5
figure 5

Representative images of localisation of nephrin protein in the glomeruli of sham and hyperglycaemic (HG) untreated and ABT-627-treated rats, as determined by immunofluorescence. All photographs were taken under the same conditions for the laser confocal microscope using a ×60 lens and quantified by calculating the ratio to actin intensity using phalloidin staining; n = 4–5 rats in each group. Negative controls for the primary nephrin antibody and the absence of phalloidin staining were evaluated as indicated above (Methods) and used to calculate the corrected intensity for nephrin and actin staining

Full size image
Fig. 6
figure 6

a Effect of streptozotocin-induced hyperglycaemia (HG), with or without concurrent treatment with ABT-627, on glomerular nephrin levels as evaluated by immunofluorescence (n = 4–5 rats in each group; n = 5–9 glomeruli per rat), as well as on (b) nephrin excretion rate (nephrinuria; n = 5–6 per group), both at the indicated times. a Two-way ANOVA resulted in the following p values: Nephrin to actin ratio, 6 weeks, p HG = 0.0002, p ABT-627 = 0.0328, p HG×ABT-627 = 0.0017. Post hoc analysis comparing individual means are denoted by *p < 0.05 vs sham and † p < 0.05 vs hyperglycaemia. b Two-way ANOVA resulted in the following p values: nephrinuria, 6 weeks, p HG = 0.0146, p ABT-627 = 0.0035, p HG×ABT-627 = 0.0078. Post hoc analysis comparing individual means are denoted by *p < 0.05 vs sham and † p < 0.05 vs hyperglycaemia

Full size image

Discussion

The current study provides new information on the mechanisms of endothelin effects in the glomerulus of the hyperglycaemic kidney and the potential use of endothelin antagonists in proteinuric renal disease. This includes evidence of a direct, ETA-dependent increase in glomerular permeability and nephrin loss that occurs in the hyperglycaemic kidney. In addition, ETA blockade provides anti-inflammatory actions by reducing hyperglycaemia-dependent increases in early inflammatory markers such as MCP-1 and soluble ICAM-1. With the increased incidence of chronic diabetic complications, strategies able to improve prevention of end-stage renal disease are urgently needed. Experimental data suggest that endothelin receptor blockade could be a novel therapeutic approach to end-stage renal disease [12], while results from clinical trials suggest that these drugs could, in addition, be used in conditions such as diabetic nephropathy [22]. The protective effects of endothelin receptor antagonists on renal injury appear to be independent of blood pressure lowering.

The potential role of direct glomerular actions of endothelin receptors has not been defined. We observed that treatment with ABT-627, administered orally at 5 mg kg−1 day−1 to hyperglycaemic rats, prevented increase in Palb in conjunction with reduction of proteinuria. The use of isolated glomeruli allowed us to measure the glomerular capillary Palb independently of the potential confounding effects of changes in mean arterial blood pressure and/or renal haemodynamics during measurement. Thus our findings demonstrate that a reduction in mean arterial BP or GFR is not a prerequisite for the anti-proteinuric effect of ETA blockade in the early phases of diabetic nephropathy. Collectively, these data are consistent with a non-haemodynamic effect of endothelin-1 on glomerular filtration barrier function.

In patients with diabetic nephropathy, Wenzel et al. showed a reduction in macroalbuminuria after 12 weeks of treatment with the moderately selective ETA antagonist avosentan [22]. This effect was observed without a change in BP, suggesting a pressure-independent anti-proteinuric effect of this antagonist [22]. Most of those participants were already on a combination of treatment with angiotensin receptor blockers and ACE inhibitors, providing evidence that the anti-proteinuric effect of endothelin antagonism was independent of any effect on the renin–angiotensin system. Opocenský et al. observed that late-onset ETA blockade reduces proteinuria in homozygous Ren-2 (also known as Ren2) rats despite severe hypertension [23]. Hocher and colleagues reported that the ETA antagonist, LU 135252, reduced proteinuria and completely normalised glomerular matrix protein expression in streptozotocin-induced hyperglycaemic rats [24].

In vitro experiments using isolated glomeruli from hyperglycaemic rats incubated with water-soluble endothelin peptide antagonists (BQ compounds) were conducted in order to investigate the specific role of glomerular ETA and ETB in permeability, independently of haemodynamic influence. ETA antagonism with BQ-123 reduced the Palb defect of glomeruli from hyperglycaemic rats. The rapid effect of the antagonist on Palb can be explained by the role of ETA in rearrangement of the actin cytoskeleton in podocytes. Morigi et al. reported that shigatoxin stimulates the synthesis of endothelin-1, which may regulate cytoskeleton remodelling and thus glomerular permeability in an autocrine manner. Cytoskeleton rearrangement is associated with an increase in albumin permeability detected by transepithelial passage of fluorescent albumin to the basolateral compartment of podocytes [25]. It has been postulated that endothelin-1 released from glomerular endothelial cells or podocytes might interact with signalling to filtration-slit components and directly influence podocyte function [26]. Selective blockade of ETB via BQ-788 had no effect on Palb of isolated glomeruli from hyperglycaemic rats. These results clearly suggest that ETB is unlikely to play a role in the regulation of glomerular permeability in hyperglycaemic glomeruli. We recently observed that exogenous endothelin-1 added to cultured podocytes increased Palb directly, an effect that can be blocked by an ETA, but not an ETB antagonist [27]. Thus, we speculate that any effect of ETB on modulation of proteinuria is via haemodynamic changes rather than direct effects on the filtration barrier.

Further evidence of a direct influence of endothelin-1 on glomerular filtration barrier function comes from our observations that ABT-627 reduced nephrin excretion and glomerular nephrin levels in treated hyperglycaemic rats. In several animal models of diabetes, nephrin excretion is increased while glomerular levels are decreased [28], which is consistent with the current study. Additional evidence that endothelin-1 is an important contributor to nephrin loss comes from a report that media taken from cultured endothelial cells conditioned with sera from pre-eclampsia patients induced nephrin loss from human cultured podocytes, an effect that can be blocked by ETA blockade [29]. Gagliardini et al. showed that avosentan prevented glomerular nephrin loss in a streptozotocin-induced model of diabetes, as detected via immunohistochemistry [13]. More recently, we reported that chronic endothelin-1 infusion in normoglycaemic rats for 2 weeks increased Palb without any change in BP and that incubation of isolated glomeruli with endothelin-1 for a little as 15 min significantly increased Palb [27]. Taken together, these studies all provide strong evidence that endothelin-1 directly facilitates increased Palb within the glomerulus independently of haemodynamic effects.

We previously observed that ETA blockade with ABT-627 reduced infiltration of macrophages and T cells in kidneys of hyperglycaemic rats after 10 weeks of treatment [12]. The current study used the same model to profile the relative time course of two early markers of inflammation, MCP-1 and soluble ICAM-1. After 6 weeks of hyperglycaemia induction, plasma and glomerular soluble ICAM-1 and MCP-1 concentrations were significantly increased, an increase largely prevented by ABT-627. There was no significant increase in inflammatory markers at the 3-week time-point. As we know from previous studies, infiltration of macrophages in the glomeruli and interstitium is one of the primary pathological features in models of diabetic nephropathy [30]. Leucocyte infiltration into inflammatory sites is mediated by coordinated actions of cell adhesion molecules and chemokines. Therefore our results support the hypothesis that endothelin-1, through the ETA, functions as an early pro-inflammatory signal. This hypothesis is further supported by previous studies in other model systems showing that endothelin-1 can increase pro-inflammatory signalling pathways, such as mitogen-activated protein kinase and nuclear factor-κB, which are known to increase production of chemoattractants and adhesion molecules such as MCP-1 and ICAM-1 [31].

ICAM-1 is a cell surface glycoprotein primarily involved in promoting leucocyte attachment to the endothelium and leucocyte transmigration through its expression on the vascular endothelium (diapedesis) and binding to β2 leucocyte integrins [32]. ICAM-1 production can be induced by hyperglycaemia, advanced glycation end-products, oxidative stress, hyperlipidaemia and hyperinsulinaemia [33]. Previous studies have shown that, in kidneys of patients with diabetic nephropathy, the accumulation of macrophages resulted from increased expression of cell adhesion molecules, such as ICAM-1 and selectins [34]. Several experimental models of diabetic nephropathy have investigated the correlation between ICAM-1 expression and disease activity. Sugimoto et al. reported that upregulation of ICAM-1 was associated with macrophage infiltration in early diabetic renal injury and was maintained during the period of study. The mechanism of ICAM-1 upregulation was mainly attributed to endothelial shear stress associated with elevated GFR in the hyperfiltration phase in hyperglycaemic rats [15]; these studies suggested that monoclonal antibodies against ICAM-1 abrogated the infiltration of macrophages in glomeruli in these hyperglycaemic rats. Similarly in ICAM-1-deficient db/db mice, Chow et al. showed that ICAM-1 deficiency reduced glomerular macrophage infiltration with subsequent amelioration of glomerular hypertrophy and interstitial fibrosis [35], indicating a deteriorative mechanism for ICAM-1 in experimental models of diabetic nephropathy.

MCP-1, also termed chemokine C-C motif ligand 2, is a potent chemoattractant for monocytes/macrophages [36]. Numerous cell types, including tubular epithelial cells and mesangial cells, are known to be capable of producing MCP-1. As reported previously, MCP-1 production was upregulated in association with renal macrophage infiltration in diabetic patients [37]. The mechanisms leading to upregulation of MCP-1 in various types of renal injury, including diabetic nephropathy, are not fully understood. However, human and rodent mesangial cells can synthesise MCP-1 in response to several factors thought to be involved in glomerular injury. These include interleukin-1, TNF-α and low-density lipoprotein [38]. In the case of human mesangial cells, a high concentration of glucose and glycated albumin have been reported to promote MCP-1 production [39]. An additional influence that may induce synthesis of MCP-1 is the generation of reactive oxygen species [40]. MCP-1-deficiency in mice markedly reduced kidney macrophage accumulation and prevented the development of diabetic nephropathy [14]. Furthermore, blocking of the MCP-1/chemokine C-C motif receptor 2 pathway ameliorated glomerulosclerosis, indicating that this pathway has an important role in progression of diabetic nephropathy [41].

Using a chronic infusion of endothelin-1 in the rat at doses that do not produce hypertension, we recently demonstrated a relationship between endothelin-1, soluble ICAM-1 and MCP-1 in the kidney [27]. Furthermore, endothelin-1 can increase expression of soluble ICAM-1 and MCP-1 in other tissue or cell types. In vitro studies suggest that endothelin-1 induces neutrophil adhesion to cardiac myocytes by increasing ICAM-1 expression, which is mediated via ETA on cardiac myocytes. ICAM-1 expression induced by activation of ETA appears to be mediated through the protein kinase C pathway [42]. Further studies using human brain-derived endothelial cells showed that in vitro endothelin-1 can directly stimulate MCP-1 mRNA expression and MCP-1 protein; and that this endothelin-1-induced MCP-1 production is mediated by the ETA [43].

Perspectives

Our results support the hypothesis that ETA contributes to glomerular permeability defects and inflammation in hyperglycaemic rats. In a model of diabetic nephropathy, we have now shown that endothelin-1 induces nephrin loss, providing a mechanism for how ETA antagonists may produce changes in Palb. These findings provide more thorough mechanistic support for the therapeutic potential of ETA antagonists in proteinuric renal disease as well as renal inflammation. Additional studies are needed to establish whether the beneficial effects of endothelin receptor blockade can reverse established changes in glomerular permeability, proteinuria and renal inflammation. Available endothelin-related antagonists include both ETA selective and combined ETA and ETB inhibitors. Given the often opposing actions of these subtypes, it will also be important to distinguish what, if any, receptor-specific effects occur on these clinically relevant variables.