Introduction

In recent years, the number of patients with end-stage renal disease has dramatically increased worldwide [1]. The progression of chronic renal disease represents one of the biggest challenges in nephrology. Regardless of whether the underlying disease is glomerulonephritis, tubulointerstitial disease, hypertensive or diabetic nephropathy, the histological picture of chronic renal disease is uniformly characterized by a progressive accumulation of extracellular matrix (ECM) proteins that obliterates renal function and leads to organ failure [2]. Enhanced understanding of the diverse pathogenetic mechanisms (vascular, metabolic, or immunologic disorders) has resulted in therapeutic advances. However, efforts to halt or even slow the progression of chronic renal disease have been largely unsuccessful. For this reason, there is still a great need for a better understanding of the molecular and cellular mechanisms involved in renal fibrosis.

A number of previous studies have shown that the small molecule nitric oxide (NO) is critically involved in pathological matrix production and accumulation of the kidney. NO has been found to act as a protective signal molecule in blood pressure regulation, platelet deposition and cell infiltration and as a detrimental effector molecule and free radical in the immune response [3, 4, 5, 6, 7]. The enzyme soluble guanylate cyclase (sGC) represents the main signaling pathway of NO via generation of cyclic guanosine monophosphate (cGMP). So far, the role of sGC in renal disease has not been paid much attention to, since it has generally been assumed that this pathway is constitutively expressed and passes on the NO signal without any further modification of its own. The present study was designed to characterize the activity and expression of the NO-sGC-cGMP cascade in the injury, matrix expansion and progression phases of anti-thy1 antibody-induced renal disease of the rat. In addition, the novel pharmacological sGC stimulator Bay 41-2272 was used to enhance NO-sGC-cGMP signaling and to address the question of whether the beneficial and detrimental actions of NO in kidney disease can be therapeutically separated by this approach.

Pathogenesis and therapy of chronic progressive renal disease Histological and molecular characteristics

The progression of end-stage renal disease is often owing to the progressive scarring and fibrosis of the kidney, with associated glomerulosclerosis, tubulointerstitial fibrosis and vascular sclerosis, resulting in the loss of filtration surface and renal function. As a fact of clinical relevance, the rate of decline of the renal function in patients with chronic renal disease correlates strongly with the extent of tubulointerstitial fibrosis [8]. Interactions among renal intrinsic cells, inflammatory cells and platelets lead to activation and proliferation of the cells to release inflammatory factors and also stimulate the accumulation of ECM, which is a hallmark of both acute and chronic renal disease [9]. With the expanding of our understanding of the pathogenesis of renal fibrosis, three mechanisms have been found to play an important role in the matrix expansion at molecular level [2, 10]: (1) an excessive synthesis of ECM components such as fibronectin, laminin, proteoglycans and collagens type I, type III and type IV; (2) an inhibited ECM degradation via increased inhibitors of proteinases, such as tissue inhibitors of matrix metalloproteinases (TIMPs) and plasminogen activator inhibitor type 1 (PAI-1), and decreased matrix proteinases, such as matrix metalloproteinases (MMPs), collagenase (MMP-1), gelatinases (MMP-2/9), elastase and serin-protease; (3) increasing synthesis and expression on the cell membrane of a group of cell-matrix receptors called integrins, which interact with matrix components and infiltrative cells.

The central role of TGF-beta in renal fibrosis

Data from numerous studies of experimental and human diseases have shown that the persistent overexpression of the cytokine TGF-beta plays an important role in progressive renal disease from initial tissue injury to tissue fibrosis [10]. A number of factors known to be injurious to the kidney at the cellular level have now been shown to directly induce TGF-beta overexpression, including angiotensin II, mesangial cell stretch, fluid shear stress, high glucose levels, hypoxia, immune-complexes, protein trafficking, platelet derived growth factor and TGF-beta itself [11]. This increase of TGF-beta directly leads to increased pathological matrix accumulation, independent of the initial injury, via the three molecular mechanisms introduced above: increased matrix production, decreased matrix degradation and increased integrin synthesis and expression [2]. Thus the idea to normalize TGF-beta expression is a reasonable therapeutic target for renal fibrosis. Neutralizing the actions of TGF-beta with either an antibody or the proteoglycan decorin has been shown to prevent excessive matrix accumulation after tissue injury [12, 13]. In this study, TGF-beta expression is a very important parameter for evaluating the effectiveness of the new therapeutic method.

The sequence of phases leading to progressive renal fibrosis

Keeping in mind studies in dermal wounding, matrix accumulation in the kidney can be understood as renal wound healing. In this concept, various kinds of pathological stimuli such as hypertension, hyperglycemia or inflammation are uniformly perceived as injurious by the kidney (injury phase). Renal tissue injury, in turn, then leads to a uniform response with early cell proliferation and subsequent TGF-beta overexpression and matrix accumulation (matrix expansion phase). In acute renal disease, a single injurious stimulus leads to a transient TGF-beta overexpression and matrix accumulation that resolves by itself without over time intervention (normal wound healing). In contrast, in progressive chronic renal disease tissue injury is constant or repetitive, and subsequent TGF-beta overexpression becomes constant (pathological wound healing). As a result, the renal matrix expands continuously and impairs renal function progressively (progression phase).

The concept of normal and pathological renal wound healing is exemplarily reflected in the rat model of anti-thy1-induced renal disease. While in animals with two kidneys, anti-thy1 antibody injection leads to acute and reversible ECM accumulation, injection of anti-thy1 antibody into uni-nephrectomized rats results in progressive renal fibrosis and insufficiency. The injury, matrix expansion and progression phases of this tandem model are characterized as follows:

Injury phase. Injury in anti-thy1 renal disease is immunologically induced. Injection and binding of the antibody to the thy1 antigen on the mesangial cell surface, which is a glycosyl-phosphatidyl-inositol-anchored membrane protein [14], causes an acute complement-dependent and iNOS-mediated loss of mesangial cells (mesangiolysis) in the glomeruli [15]. This phase starts minutes after antibody injection and lasts approximately 24 hours.
Matrix expansion phase. Following mesangial cell injury, the matrix expansion phase is characterized by early mesangial cell proliferation and marked glomerular TGF-beta overexpression and ECM accumulation, which peaks at 7 days after antibody injection [16].
Progression phase. Following the initial mesangial injury and matrix expansion, consistent with a decrease in TGF-beta production [2], the accumulation of ECM resolves by 6-8 weeks in animals with two kidneys. However, injection of monoclonal antibody into uni-nephrectomy animals demonstrates persistent proteinuria, marked TGF-beta expression and typical progressive fibrotic changes, first in the glomeruli, then in the tubulointerstitium. Over 16-20 weeks, the whole organ becomes fibrotic and the animals develop progressive renal insufficiency. This course is highly analogous to human chronic renal disease of glomerular origin. A continuous hyperfiltration injury of the remaining kidney may relate to the chronic progression.

Therapeutic approaches to chronic progressive renal disease

Progressive tissue fibrosis results from ongoing tissue “injury” and “matrix expansion”, so treatment of tissue fibrosis can aim at limiting tissue injury or at interfering directly with the molecular mechanisms mediating the matrix expansion, i.e. the overexpression of TGF-beta. Conventionally, therapies have aimed at reducing glomerular capillary pressure and the injuries that result, such as blood pressure and glycemic control in patients with hypertension and diabetes, but many patients still show progressive renal fibrosis with well controlled blood pressure and glycemia. Many studies have shown that angiotensin II blockade or low dietary protein have blood pressure-independent effects on the overexpression of TGF-beta and the ECM accumulation, besides reducing glomerular capillary pressure and thus decreasing tissue injury [10]. Future therapies should be focused on the mediators of renal fibrosis, such as TGF-beta overexpression, since the tissue “matrix expansion” and subsequent renal failure might be independent of the primary injury. Up until now, no therapies have been able to reduce TGF-beta expression to normal. New insights into second messengers and intracellular transduction pathways may suggest new interventions aimed at neutralizing fibrogenic mediators.

The L-arginine-NO pathway in renal disease

The amino acid L-arginine is semi-essential and provides molecular substrate for the generation of NO, polyamines, L-proline and agmatine, all of which have been reported to be involved in renal pathology [3, 6, 17, 18]. Endogenous L-arginine synthesis occurs primarily in the proximal tubule of the kidney, using the non-protein amino acid L-citrulline as a precursor [19]. Minor endogenous L-arginine synthesis has been observed in endothelial cells and macrophages [6, 18].

Two faces of the L-arginine-NO pathway

L-arginine is the main endogenous source for the generation of NO by a highly regulated family of enzymes called NO synthase (NOS) that utilize O ₂ and NADPH as co-substrates and generate citrulline as a co-product [3, 6, 17]. NO is a diatomic and highly reactive radical gas and plays an important role in numerous biological processes ranging from neurotransmission to vasodilatation, and from inflammation to cell phenotype regulation. Once NO is produced, there exist two main pathways for it to interfere with renal physiology and pathophysiology: as an indirectly acting signaling and as a directly acting effector molecule [5]. At low physiological concentrations, NO acts locally as an important signaling molecule, because it diffuses freely across cell membranes with a diffusion coefficient 1.4 times that of oxygen in vivo, without being utilized by or reacting with intracellular molecules. The electronic structure of NO makes it an excellent ligand for heme-containing enzymes, which include cytochrome P450, NOS and sGC, allowing it to bind to the sGC heme at a low, non-toxic concentration [20]. Therefore, the signaling of NO is mainly mediated by the activation of sGC to increase cGMP levels and modify many of the biological effects of NO, including vasodilatation, inhibition of platelet activation and leukocyte migration [21].

At high concentrations, NO becomes a relatively non-specific destructive effector molecule of the immune system with a key role in mediating host defense, autoimmunity and self-destruction. Most of NO’s cytostatic and cytotoxic potential is induced via the irreversible reaction with superoxide radical (O ₂ ^- ) and formation of strong oxidant peroxynitrite (ONOO ^- ) [22], which results in nitration of a tyrosine residue in proteins and subsequent damage, inhibition of key enzymes of the respiratory chain and DNA synthesis by nitrolysation of their iron-containing catalytic elements and fragmentation of DNA. A high concentration of NO can result in autooxidation and the production of dinitrogen trioxide, which is injurious to biological tissues and causes chemical DNA alteration [23].

NO synthases

The glomerulus is a unique vascular network with three specialized cell types: endothelial cells, mesangial cells and visceral epithelial cells to express three potential nitric oxide synthase enzymes, endothelial, neuronal and inducible NOS (eNOS, nNOS and iNOS). Therefore, it has probably been more difficult to define roles for NO in glomerulonephritis than in most other immune and inflammatory diseases, despite a wealth of data [24]. It is assumed that different activities of NOS isoforms are essential for understanding the effects of in vivo NO.

iNOS can be induced by inflammatory cytokines, especially LPS, interferon γ , interleukin-1, and TNF-alpha, elevated in macrophages, renal mesangial and tubular cells in the diseased kidney, and it produces high levels of NO in a calcium-independent manner [25]. In the presence of iNOS, NO is produced in high concentration and becomes an effector molecule, which plays a key role in the period of rapid initial response to immune injury in glomerulonephritis [24]. Its NO production depends on the extracellular L-arginine supply.

eNOS and nNOS also called constitutive NOS (cNOS), are principally constitutively expressed and regulated by calcium [26]. Cell types containing cNOS generate low fluxes of NO for short periods of time, and signaling effects of NO are predominant. Its NO generation is not dependent on the extracellular L-arginine supply. eNOS is expressed by the endothelium. eNOS-dependent NO regulates vascular tone by promoting cGMP-dependent vascular smooth muscle cell relaxation, inhibits platelet aggregation and adhesion and prevents leukocyte adhesion [27]. nNOS also forms a class of cNOS and is localized in the macula densa and the efferent arterioles in the healthy human kidney [28]. nNOS is involved in the regulation of glomerular hemodynamics via tubulo-glomerular feedback and the release of renin [5].

Effects of L-arginine supplementation on renal disease

Over the last few years, L-arginine supplementation has been used as a novel therapy to modify the L-arginine-NO pathway and to slow the progression of renal fibrosis. However, the results of L-arginine treatment are controversial. Enhanced endothelial signaling NO production appears to mediate protective effects, while NO metabolized through the high-output iNOS seems to be a critical effector molecule in immune-mediated tissue damage [5]. In the injury phase of anti-thy1 glomerulonephritis, lupus nephritis, transplant rejection and acute tubular necrosis, where iNOS is increased, L-arginine intake results in the production of high concentrations of NO, which can worsen tissue injury and subsequent fibrosis. Yet in the matrix expansion phase of anti-thy1 glomerulonephritis, 5/6 nephrectomy, ureteral obstruction and nephropathy to diabetes and hypertension, L-arginine supplementation overcomes chronic NO deficiency by enhancing signaling molecule NO production via eNOS, which can reduce TGF-beta overexpression and matrix expansion. Additionally, there are several by-products in the pathway of L-arginine-NO, such as ornithine, polyamines and proline, which are essential for cellular proliferation and ECM deposits [5]. Interestingly, it has been found that L-arginine restriction limited TGF-beta and matrix accumulation, due to the limitation of substrate for polyamine and L-proline production [29]. These data suggest that both dietary L-arginine restriction and supplementation significantly limit matrix expansion, in contrast to normal L-arginine intake, which produces a more severe disease. The therapeutic and detrimental data of L-arginine supplementation reflect the two faces of the L-arginine-NO pathway. Investigating the question of how to enhance the signaling molecular effect and prevent the effector molecular effect is an aim of this study.

NO-cGMP signaling in renal disease

The most physiologically relevant action of NO mainly produced by eNOS is the activation of sGC and the conversion of GTP into the intracellular second messenger cGMP [30], a cyclic nucleotide, which regulates various cGMP effector systems such as protein kinases (PKG), phosphodiesterases (PDEs) and ion channels, to modify many physiological processes.

Distribution of NO-cGMP signaling in the glomeruli and the tubulointerstitium

The NO-sGC-cGMP signaling pathway is spatially distributed in the glomerular and tubulointerstitial cells of the kidney to a great extent and explained in the following:

- In the glomeruli

As shown in Figure 1, eNOS distributes in endothelial cells of the glomerular vasculature, and nNOS is found primarily in the macula densa, both of which contribute to the production of NO signaling molecules in the glomeruli. Then NO diffuses freely to sGC, which expresses significantly in glomerular arteriolar walls and mesangial cells of the intra- and extraglomerular mesangium [31]. The activated sGC increases cGMP levels, which mediates its main downstream effects via PKG. PKG is a serine/threonine kinase including the soluble PKGI and the membrane-bound PKGII. PKGI is expressed in most tissues. In the glomeruli, PKGI is localized in arteriolar smooth muscle cells, including afferent and efferent arterioles and mesangial cells [32].

Figure 1: Distribution of NO-cGMP signaling in the glomeruli. L-Arginine (L-Arg), NO (Nitric oxide), Endothelial NO synthase (eNOS), Neuronal NOS (nNOS), cGMP (Cyclic guanosine monophosphate), GTP (Guanosine 5`-triphosphate), sGC (Soluble guanylate cyclase).

- In the tubulointerstitium

As shown in Figure 2, eNOS expresses in endothelial cells of the tubulointerstitial vasculature. Signaling molecule NO generated from eNOS activates sGC in cortical and medullary interstitial fibroblasts and along the renal vasculature outside the glomeruli to increase cGMP levels [31]. The cGMP stimulates PKGI, which has been found in microvascular pericytes and myofibroblasts [32].

Figure 2: Distribution of NO-cGMP signaling in the tubulointerstitium. L-Arginine (L-Arg), NO (Nitric oxide), Endothelial NO synthase (eNOS), Neuronal NOS (nNOS), cGMP (Cyclic guanosine monophosphate), GTP (Guanosine 5`-triphosphate), sGC (Soluble guanylate cyclase).

sGC has not been found in endothelial cells. A heterogeneous distribution of NOS and sGC in the renal tissues suggests that NO acts through paracrine diffusion to its target for cGMP generation, rather than by intracellular signaling [31]. The anatomic features of these cells in renal tissue have also proved this, since in the kidney the endothelial cell is closely linked to the vasculature and the mesangium, and the macula densa is adjacent to the extraglomerular mesangium. This is supported by the coincident histochemical localization of PKG and sGC, implying that cGMP is a second messenger acting through intracellular signaling.

The physiology of NO-cGMP signaling

It was established by the mid-1970s that GC exists in membrane-bound and soluble forms in most cells [33]. In general, particulate GC does not respond to NO, since it has no heme prosthetic group in its structure. Purification of GC from the cytosolic compartment revealed the soluble isoform was a heterodimer composed of alpha- and beta-subunits and containing heme as a prosthetic group [34]. The most abundant subunits are alpha1 and beta1, which are found in many tissues [35]. NO binds to the enzyme’s prosthetic heme group that is linked to His-105 residue of the beta1-subunit induces conformational changes leading to an up to 400-fold increase in catalytic activity [36]. In vivo, sGC is activated at a low level of NO in the range of 5-100 nM; for NO >200 nM, nearly all of the binding sites for sGC are bound with NO [37]. Activated sGC catalyzes the conversion of GTP to cGMP and pyrophosphate [38].

PKG, PDE and the ion channel are the three main downstream effectors of the NO-cGMP pathway. PKGI can decrease cytosolic calcium and lead to smooth muscle relaxation [39]. Activation of PKG phosphorylates the thromboxane A2 receptor of platelet and then decreases platelets aggregation [40]. Furthermore, PKGI can phosphorylate vimentin, which is involved in neutrophil activation [41]. In contrast to PKGI, PKGII recently has been found in juxtaglomerular renin-secreting cells and in several nephron segments, contributing to renin release and ion transport in the tubule [42]. Therefore, the NO-cGMP pathway plays an important role in vasodilatation, inhibition of platelet aggregation and neutrophil infiltration.

PDEs are enzymes which hydrolyze the 3´-phosphoester bond of cGMP to their biologically inactive noncyclic nucleotides 5´-GMP [43]. The cellular levels of cGMP are defined by the balance between the activities of the synthesizing enzyme (sGC) and the catabolizing enzymes (PDEs). Investigations during recent years have revealed that PDEs are a critically important component of the NO-cGMP signal transduction. But the distribution of PDEs in the kidney is unclear. Studies of cultured rat mesangial cells show that atrial natriuretic peptide-induced cGMP accumulation is markedly decreased in the presence of angiotensin II, and this inhibition is abolished by IBMX [44]. Thus the PDE inhibitor might represent a novel class of pharmacologic agents suited for “signaling transduction pharmacotherapy” [45].

The action of cation channels regulated by cellular cGMP levels is involved in the molecular basis of sensory systems of visualization and olfaction [32].

The pathophysiology of NO-cGMP signaling

Only little is known about the regulation and expression of NO-cGMP signaling in renal fibrosis. Some data have shown that mechanisms behind the well known fibrogenic systems, such as the angiotensin system, the endothelin and the renal kallikrein-kinin system, are related to NO-cGMP signal transduction [46, 47, 48]. Angiotensin II inhibition could stimulate cGMP production and potentially contribute to renal protection in chronic nephritis [46]; and NO decreases secreted endothelin-1 due to the activation of sGC and increased cGMP generation [48]. In addition, nephrotoxic effects of cyclosporin A may result from the disruption of the bradykinin/sGC pathway [47]. It was shown that dexamethasone, one of several known selective inhibitors of iNOS, had a stimulatory effect on sGC in the glomeruli of rat kidney [49]. The contractile apparatus in the podocyte foot processes was proved to be a target of the cGMP signaling, which correlates with detachment of the cells from the glomerular basement membrane and is closely associated with many nephropathies [50]. Therefore, NO-cGMP signal transduction might take part in the regulation of the kidney functions and the progression of kidney diseases.

Pharmacological stimulators of sGC activity

In this study, the transcriptional regulation of the NO-cGMP pathway was investigated and the pathway was modified to further define its role in progressive renal fibrosis. From the above information, we know that sGC is the most important receptor for the ubiquitous biological messenger NO and is intimately involved in many signal transduction pathways. Activators of sGC are therefore very desirable as both pharmacological tools and potential therapeutics to probe the NO-cGMP pathway. In the 1990s, YC-1, chemically 3-(5’-hydroxymethyl-2’-furyl)-1-benzylindazole, was identified as a moderate and ~tenfold sGC activator. The actions of YC-1 are heme-dependent, synergistic with NO and with the inhibition of PDE activity [51]. Recently, Stasch and colleagues reported on the discovery of a more potent sGC stimulator devoid of inhibiting PDE activity, Bay 41-2272, chemically 5-cyclopropyl-2-[1-(2-fluoro-benzyl)-1Hpyrazolo(3,4-b)pyridin-3-yl]-pyrimidin-4-ylamine, which stimulates sGC up to 30-fold in an NO-independent manner by interacting with cysteines 238 or 243 of the alpha1 subunit, which is different from the binding site of NO [52]. Although Bay 41-2272 alone is not as strong a stimulator of sGC as NO, the specific activity of sGC could be 416-fold above the baseline through the combination of Bay 41-2272 and the NO donor [52]. These effects are probably mediated both by NO-independent stimulation of the enzyme and NO-dependent stimulation by sensitization of sGC towards endogenous NO. It has been demonstrated that, in a rat model of renin transgenic hypertension and spontaneously hypertensive rats (SHRs), oral administration of Bay 41-2272 lowered mean arterial pressure and showed antiplatelet activity, and enhanced survival in a low-NO rat model of hypertension [52]. In comparison to L-arginine, sGC stimulation with Bay 41-2272 avoids the potentially deleterious actions of the effector pathway of NO. A decrease in cGMP levels due to reduced sGC protein was found in the smooth muscle cells of SHRs, despite enhanced eNOS/NO production in the endothelium [53]. In this model, the sGC-cGMP pathway but not the eNOS/NO pathway was impaired, resulting in a decrease in cGMP production. So Bay 41-2272 was chosen to specifically enhance NO-cGMP signaling and further characterize the role of the sGC in renal pathological matrix accumulation.

Aim of the study

We tested the hypothesis that the NO signal transduction via sGC plays its own, transcriptionally regulated role in pathological renal matrix expansion and that a selective enhancement of the sGC activity by Bay 41-2272 will limit kidney fibrosis in a pressure-independent manner. To test these hypotheses, the following protocols were carried out in the model of anti-thy1 glomerulonephritis.

How is the glomerular NO-cGMP signal transduction regulated in the injury phase of acute anti-thy1 glomerulonephritis one day after antibody injection? How does this correlate with anti-thy1 antibody-induced inducible NO production and mesangial cell lysis? How will a pretreatment with Bay 41-2272 affect NO-induced glomerular injury? Bay 41-2272 treatment was started 6 days before disease induction.
How is the glomerular NO-cGMP signal transduction altered in the matrix expansion phase of acute anti-thy1 glomerulonephritis seven days after antibody injection? How do the activity and expression of the NO-cGMP signaling pathway relate to glomerular TGF-beta overexpression and matrix accumulation? Will the administration of Bay 41-2272 decrease glomerular matrix expansion? In this protocol, Bay 41-2272 treatment was started 24 hours after disease induction, so that the therapy began after the mesangial cell lysis had occurred.
How is the tubulointerstitial and glomerular NO-cGMP signal transduction changed in the progression phase of chronic anti-thy1-induced glomerulosclerosis 16 weeks after antibody injection? Is the signal pathway concordantly or discordantly expressed at the glomerular and tubulointerstitial level? How do the activity and expression of the NO-cGMP signaling pathway correlate with glomerulosclerosis, tubulointerstitial fibrosis and renal insufficiency? Will enhancement of sGC activity through Bay 41-2272 slow the progressive course of chronic anti-thy1-induced glomerulosclerosis? The actions of Bay 41-2272 were compared side-by-side to the sole vasodilator hydralazine. In this protocol, treatment with Bay 41-2272 or hydralazine was started 1 week after disease induction.