A single injection of anti-thy1 antibody into normal or uni-nephrectomized rats causes a sequence of distinct phases of “injury”, “matrix expansion” and subsequent “progression”, which enable us to study the NO-cGMP pathway in different processes separately, while most renal disease models show no distinct injury phase and involve overlapping cycles of tissue injury and matrix expansion, such as the model of hypertension and diabetes. The injury phase is characterized by marked inducible NO production and mesangial cell lysis, which has also been found in IgA nephropathy, Wegener’s granulomatosis, lupus nephritis, acute tubular necrosis and kidney transplant rejection. The subsequent fibrosis is characterized by matrix protein accumulation, which is similar to human mesangioproliferative glomerulonephritis and IgA nephropathy. All these histological features can be found in many experimental and human renal diseases [65]. Hence the model was chosen to interpret the relevant role of the NO-cGMP pathway and its modification in human renal disease. Renal mesangial cells are specialized smooth muscle cells that play an important role in the structural support of the glomerulus and in the control of glomerular filtration rate [66]. In the same way as shown in our model, they contribute also to the pathological processes of the glomerulus, particularly in glomerulonephritis and glomerulosclerosis. The NO-cGMP pathway plays an important role in mesangial cells [31], which might indicate an important role of NO-cGMP signal transduction in kidney diseases, such as mesangial proliferative glomerulonephritis. Therefore, we tested the hypothesis in anti-thy1 glomerulonephritis, an animal model for mesangial proliferative glomerulonephritis, which is ideal for studying the NO-cGMP pathway in vivo in comparison to the in vitro studies on the mesangial cell.

The common final pathway of chronic renal disease, independent of the primary disorder, suggests that a self-progressing mechanism and a common intrarenally determined program are operating, since the ongoing deterioration of organ function occurs even if the initial disease has been removed [9]. In the present study, anti-thy1-induced chronic glomerulosclerosis was produced through an injection of anti-thy1 antibody and a persistent hyperfiltration injury of the remaining kidney, which is consistent with a self-perpetuating and intrarenal common final pathway in many chronic renal diseases, while the ongoing and extrarenal injuries obliterate the decrease in renal function in the diabetic and hypertensive nephropathy models. Therefore, anti-thy1-induced glomerulosclerosis provides a good in vivo method to characterize the intrarenally self-progressing pathway contributing to chronic renal fibrosis. Because the molecular mechanisms of matrix expansion in various renal diseases, such as diabetic and hypertensive nephropathy, are rather common, the findings of this study may be relevant for fibrotic renal disease in general.

In the anti-thy1 glomerulonephritis animal model, the progressive renal disease is glomerular in origin. Thus, the measurements of sclerotic parameters in glomeruli are very important for this research. However, glomeruli occupy only 3-4% of the kidney volume compared to the >90% comprised by tubules and the interstitium [67]. Therefore, the metal sieves technique was used to yield glomeruli. Many studies have shown that tubulointerstitial injury and fibrosis correlate best with renal disease progression and insufficiency and develop independently no matter whether they originate from primarily glomerular or vascular lesions [68]. In anti-thy1-induced chronic glomerulosclerosis, the TGF-beta overexpression and matrix expansion spread from the glomeruli into the tubulointerstitium. Thus, the parameters of tubulointerstitial fibrosis are very important in the evaluation of chronic progression. Since renal cortex consists mainly of tubulointerstitial tissue (>95%), it was used as representative of the tubulointerstitium. The separation of glomeruli and cortical cell enables us to study the fibrogenic responses of the glomeruli and the tubulointerstitium of every individual animal separately, including the production of TGF-beta, fibronectin and PAI-1 at protein level in cell culture and at mRNA level.

There were preliminary studies in our work group, which showed that the production of TGF-beta1, fibronectin and PAI-1 by isolated glomeruli and cortical cells is constant for 48 hours in culture. The 48-hour harvesting of culture supernatant was chosen on the basis of these data and the fact that the quantities of these molecules are sufficient to get accurate ELISA measurements and correlate to ECM accumulation in vivo [69]. As shown in the results, there is a big difference in disease parameters, such as glomerular and cortical TGF-beta1, fibronectin and PAI-1 expression, between untreated diseased rats and normal controls, and these differences enable detection of small differences in therapeutic efficacy. ELISA data obtained from individual rats presented the fibrotic severity precisely, in contrast to quantization from PAS staining, which is variable in tissue sectioning, staining intensity and subjective scores.

Determination of the key fibrosis mediator TGF-beta1 is a valid sensitive marker of renal matrix expansion [56, 70]. This concept is proved in the rat model of anti-thy1 glomerulonephritis. In animals with two kidneys, anti-thy1 antibody injection leads to acute and reversible mesangioproliferative glomerulonephritis, in which TGF-beta returns to normal after the resolution of mesangial cell injury; whereas in uni-nephrectomized rats, injection of anti-thy1 antibody results in persisting TGF-beta overexpression and progressive glomerulosclerosis and tubulointerstitial fibrosis. The analysis of PAI-1 and fibronectin levels can also be an important evaluation of the anti-fibrotic effect of a new therapy independent of TGF-beta, by reflecting the imbalance of ECM synthesis and degradation. In this study, it is shown that TGF-beta production had a close relation with fibronectin and PAI-1 production and ECM accumulation, indicating that the increase in matrix accumulation is likely to result from both elevated matrix synthesis and inhibited matrix degradation.

Due to the importance of NO-cGMP signal transduction, modulation of this signaling cascade has attracted much attention. L-arginine and NO donor administration have long been used to modulate the NO-cGMP pathway. However, experimental studies have shown both therapeutic and detrimental consequences of treatment with L-arginine and NO donor, since the effector molecular effects of NO have also contributed to it. The use of L-arginine is not only obstructed by the potential synthesis of cytotoxic amounts of NO via iNOS, but also by the generation of agmatine, proline and polyamines via alternative L-arginine-metabolizing pathways, which are involved in renal tissue repair through different, partly opposing mechanisms [5]. The use of pharmacological NO donors leads to a rather unspecific release of NO across the whole body, is prone to tachyphylaxis and has been linked to an unwanted pro-oxidant activity [71]. In contrast, the new class of sGC modulators, such as Bay 41-2272, has attracted particular interest as the target of the L-arginine-NO-cGMP signal transduction without the formation of by-products, NO oxidative effects and tolerance, since they amplify the cGMP levels exactly at the downstream effector of NO, where cGMP is naturally generated. Therefore, sGC stimulators represent useful pharmacological tools to differentiate between signal molecular and effector molecular effects of NO.

In order to further define the role of the NO-cGMP pathway in renal disease, the rat model with anti-thy1 glomerulonephritis was first used to study the in vivo effects of Bay 41-2272 on renal function, pathology and the expression of pro-inflammatory and pro-fibrogenic factors. Second, in vitro cultures of glomeruli were used to study the possible mechanisms for the renoprotective effect of Bay 41-2272. Finally, the effects of Bay 41-2272 on the renal disease progression in this rat model were compared with those of a vasodilator, hydralazine.

It was interesting to find that there was about 90% decrease in glomerular alpha1 sGC and beta1 sGC expression of nephritic rats shortly after anti-thy1 antibody-induced mesangial cell injury had occurred. Corresponding to the gene expression, nephritic rats had significantly lower basal glomerular cGMP, which additionally could hardly be stimulated by DEA/NO. Bay 41-2272 pretreatment failed to stimulate sGC expression and activity to exogenous NO stimulation, probably because of the loss of glomerular sGC. The disruption of the NO-cGMP cascade paralleled the mesangial cell lysis, consistent with a high degree of sGC distribution in mesangial cells [31]. It has been reported that pro-inflammatory cytokines, such as interleukin-1beta and TNF-alpha, decrease the half-life of the sGC subunit [72]. Prolonged incubation of smooth muscle cells with NO donors has been shown to decrease sGC activity, protein and mRNA [73]. This suggests that high-output NO derived from iNOS and cytokines may also modulate sGC subunit expression. In contrast to the traditional meaning of “NO deficiency”, resulting from reduction in L-arginine and eNOS, decreased cGMP was present in the injury model with high-output NO production and normal eNOS expression. These results have opened a new explanation for “NO deficiency”.

In rat nephrotoxic nephritis, macrophage infiltration was decreased by iNOS inhibition, accompanied by an increase in eNOS [74]. It has been shown that in human mesangial cells, iNOS mRNA half-life was reduced by the NO-cGMP pathway to regulate the toxic effects of high levels of NO [66]. Then I inquired if Bay 41-2272 might activate the L-arginine-eNOS-cGMP pathway to depress the L-arginine-iNOS-NO pathway in the injury phase. As shown in the results, the animals in protocol 1 displayed a typical manifestation of the injury phase in acute anti-thy1 glomerulonephritis: up- regulation of iNOS mRNA expression with a high-output of NO and mesangial cell lysis. Pretreatment with Bay 41-2272 could not increase cGMP synthesis, perhaps due to the lack of the receptor sGC. With this in mind, it is not surprising that Bay 41-2272 had no effect on the degree of anti-thy1-induced mesangial cell injury, including proteinuria, mesangial cell counts, and NO output and iNOS mRNA expression. But it should be emphasized that unlike L-Arginine, which can increase mesangial cell injury due to pro-fibrogenic by-products and increased oxidative stress, Bay 41-2272 has no negative effects on the injury phase. It even slightly decreased proteinuria in our study. Hence, the deficiency of target receptor sGC in the injury phase is the reason for the lack of Bay 41-2272 effects in this phase of the disease.

In contrast to the injury phase, a dramatic change of sGC was shown in the matrix expansion phase. Alpha1 sGC and beta1 sGC expression increased 5.6-fold and 3.3-fold in untreated nephritic rats. Although already up-regulated by the disease itself, Bay 41-2272 treatment further elevated the expression of alpha1 sGC and beta1 sGC 17.6-fold and 12.2-fold. The increase in sGC expression in the nephritic group is parallel to mesangial cell proliferation, indicating that sGC expresses significantly in mesangial cells, which is consistent with the disruption of sGC expression in the injury phase accompanied by mesangiolysis. Apart from increasing sGC expression, sGC activity to generate cGMP was also increased by the disease itself and significantly further amplified by Bay 41-2272, which is parallel to sGC mRNA expression. Of interest in this context is the finding that sGC enhancer treatment not only significantly increased glomerular NO-induced cGMP generation, but also the mRNA expression of both subunits of the sGC. This may represent a self-reinforcement mechanism through which Bay 41-2272 can influence the expression of its effectors. However, it might as well be just a consequence of the improved glomerular wound repair in the treatment group. The increased sGC expression by Bay 41-2272 treatment, on the other hand, is a novel observation that has not been previously reported. It becomes important in pathophysiologic conditions, where deficient sGC expression is characterized. However, this is in contrast to the data from many cardiovascular studies, which have shown decreases in sGC expression and activity in hypertensive animal models [75, 76]. This may represent divergent regulatory mechanisms in the two different organs. In acute anti-thy1 glomerulonephritis, glomerular sGC expression and activity are compensatorily increased and relatively impaired, which may not suffice to counteract the fibrotic response, whereas Bay 41-2272 further elevated sGC expression and cGMP levels to limit matrix expansion. In addition, similar to the results in the injury phase, eNOS expression did not decrease in nephritic animals, which is consistent with no decrease in renal eNOS expression found in hypertensive rats [53], suggesting no eNOS-NO deficiency in the matrix expansion phase of acute anti-thy1 glomerulonephritis. Moreover, Bay 41-2272 increased eNOS expression transcriptionally. Further studies will be done to explore the mechanism of the up-regulation of eNOS expression by Bay 41-2272.

In the present study, increases in glomerular production of 10.5-fold for TGF-beta1, 9-fold for fibronectin and 3.4-fold for PAI-1 were seen in the untreated nephritic group when compared to normal controls. These significant differences make it possible to observe the antifibrotic effects of Bay 41-2272. The cGMP levels further amplified by Bay 41-2272 significantly limited the fibrotic response of the matrix expansion phase, as shown by reductions in proteinuria, glomerular histological matrix accumulation and expression of TGF-beta1, fibronectin and PAI-1. Although all disease parameters prove the antifibrotic effects of Bay 41-2272, it remains unclear how enhanced NO-cGMP signal transduction results in the decrease in TGF-beta1 overexpression and the antifibrotic effect, which will be discussed later. It has been previously proved that L-arginine has antifibrotic action in anti-thy1 glomerulonephritis mediated by endogenous production of NO through limiting TGF-beta overexpression [16]. Our findings further prove that the antifibrotic action of NO in renal fibrosis is mediated by cGMP through sGC modification.

Very little is known about the long-term regulation of sGC in the kidney. sGC expression and activity showed contrary change in the glomeruli and the tubulointerstitium of the chronic model. In parallel to increased glomerular expression and activity of sGC in acute anti-thy1 glomerulonephritis, tubulointerstitial sGC mRNA expression and NO-dependent cGMP production were significantly up-regulated in the presence of marked tubulointerstitial fibrosis 16 weeks after induction of progressive anti-thy1 glomerulosclerosis. However, glomerular sGC expression and activity were significantly inhibited, in contrast to tubulointerstitial NO-cGMP signaling in the same model at the same time, but also to the strong up-regulated glomerular sGC activity in the matrix expansion phase of the acute model. Although Bay 41-2272 could not enhance sGC expression, sGC activity in response to NO stimulation in both depressed glomerular and elevated tubulointerstitial sGC expression was significantly enhanced by Bay 41-2272 treatment. Therefore, subsensitivity of sGC to NO stimulation, as observed in our results, could attenuate the efficacy of NO-cGMP signal transduction, contributing to the progression of renal disease. Compared to the relatively impaired tubulointerstitial sGC, the glomerular sGC was completely impaired. The up-regulation of tubulointerstitial sGC expression in diseased rats might also be a self-compensatory effect to counteract the renal fibrosis.

Endothelium-independent NO-cGMP signaling by exogenous NO donor was blunted, indicating that dysfunction of sGC may be responsible for the blunted endothelium-independent NO-cGMP pathway. Considering slightly increased eNOS mRNA in diseased rats in the study, these alterations in the NO-cGMP system, namely, the enhanced eNOS-NO pathway and impaired sGC-cGMP pathway, seem to be the phenotypic character of renal fibrosis rats. Dysfunction of sGC-cGMP rather than eNOS-NO deficiency contributes to the impairment of NO-cGMP signal transduction in our animal model.

Bay 41-2272 treatment increased cGMP generation and subsequently markedly limited the glomerular and tubulointerstitial progression in this model, as shown by reduced histological glomerular and tubulointerstitial fibrosis and expression of TGF-beta1, fibronectin and PAI-1. The inhibitory effects on TGF-beta1, fibronectin and PAI-1 protein and mRNA expression lead to a decrease in ECM synthesis and an increase in ECM protein degradation resulting in limiting renal fibrosis. In addition, administration of the sGC stimulator lowered blood creatinine and urea concentrations and increased creatinine clearance and hematocrit levels, underscoring that the morphological and molecular glomerular and tubulointerstitial benefits achieved went along with improved renal function. The present study in progressive anti-thy1-induced chronic glomerulosclerosis confirms the critical role of sGC and NO-dependent cGMP production in pathological renal matrix accumulation revealed in the acute anti-thy1 model. The important role of sGC in fibrotic kidney disease is thereby now expanded from glomerular to tubulointerstitial TGF-beta overexpression and matrix expansion, as well as from acute, reversible to chronic progressive renal disease.

However, at the glomerular level, Bay 41-2272 reduced TGF-beta1, fibronectin and PAI-1 as well, but the benefits are more moderate than those at the tubulointerstitial level. In addition, Bay 41-2272 treatment has no statistically significant effect on proteinuria, which mainly reflects glomerular barrier impairment. It seems that Bay 41-2272 could reduce tubulointerstitial impairment more effectively than glomerular impairment. I think the possible reason is that Bay 41-2272 treatment was started as late as 7 days after induction of anti-thy1 glomerulosclerosis; at this point in time, the glomerular disease was already established. This view is supported by the results in the matrix expansion phase of acute anti-thy1 glomerulonephritis, in which Bay 41-2272 has significant effects on reducing glomerular TGF-beta1 overexpression, fibronectin, PAI-1 and proteinuria when begun as early as 24 hours after anti-thy1 glomerulonephritis induction. On the other hand, a late start of the therapy may result in more serious sGC impairment in the glomeruli than in the tubulointerstitium, which also leads to a less efficiency of Bay 41-2272 on proteinuria. In addition, it has been pointed out that proteinuria in anti-thy1-induced renal disease does not always follow the severity of the fibrotic response [77]. Thus, the finding that proteinuria is not significantly reduced in the chronic model is of interest, yet the reasons for it remain unclear.

Bay 41-2272 can significantly decrease systolic blood pressure by increasing cGMP, which leads to the question if the antifibrotic action of Bay 41-2272 is mediated by reduced glomerular blood pressure. However, there was a Bay 41-2272 dose-dependent reduction in TGF-beta1 protein synthesis in cultured normal and nephritic glomeruli, an environment where blood pressure plays no role. In addition, acute anti-thy1 glomerulonephritis is a normotensive model, and it has been found that systemic hypertension and treatment with high doses of β-blocker does not influence the excess of the mesangial matrix deposition in this model [92], and renovascular hypertension did not influence healing of the glomerular lesions in the acute anti-thymocyte serum nephritis [93]. These findings suggest that sGC stimulation to increase cGMP can inhibit TGF-beta1 expression via a direct and blood pressure-independent mechanism in the model with acute glomerulonephritis.

In the chronic glomerulosclerosis model, the sole vasodilator hydralazine did not significantly alter the degrees of glomerulosclerosis, tubulointerstitial fibrosis and renal insufficiency, despite a comparable blood pressure decrease having been achieved, while only Bay 41-2272 significantly reduced pro-fibrotic factors, such as TGF-beta1, fibronectin and PAI-1, inflammatory cell infiltration and histological signs of glomerulosclerosis. This suggests that Bay 41-2272 slowed the progressive course of anti-thy1-induced glomerulosclerosis and appeared to be at least in part independent of blood pressure reduction. Renal TGF-beta1 overexpression and matrix accumulation were also reduced by blood pressure-independent mechanisms. The superiority of Bay 41-2272 over conventional antihypertensive therapies may result from elevated cGMP levels. Therefore, enhancement of cGMP production probably has antifibrotic effects that are independent of the reduction in blood pressure, which includes the following three mechanisms.