Anti-Fibrosis Therapy and Diabetic Nephropathy
- First Online:
- Cite this article as:
- Karihaloo, A. Curr Diab Rep (2012) 12: 414. doi:10.1007/s11892-012-0290-7
- 604 Views
Diabetes mellitus is rapidly becoming a global health issue that may overtake cancer during the next two decades as it covertly affects multiple organ systems that goes undiagnosed long after the onset. A number of complications are associated with poorly controlled hyperglycemia. Diabetic nephropathy is one of the most common complications of diabetes mellitus. Other than angiotensin-converting enzyme (ACE) inhibitors and angiotensin receptor blocker (ARB) there is not much in the armamentarium with which to treat patients with overt diabetic nephropathy. Research points towards a multifactorial etiology and complex interplay of several pathogenic pathways that can contribute to the declining kidney function in diabetes. Patients with diabetic nephropathy (and with any chronic kidney disease) eventually develop kidney fibrosis. Despite the financial and labor investment spent on determining the basic mechanism of fibrosis, not much progress has been made in terms of therapeutic targets available to us today. This may be in part due to paucity in the experimental animal models available. However, there now seems to be a concerted effort from several pharmaceutical companies to develop a drug that would halt/delay the process of fibrosis, if not reverse it. This review discusses the current state of research in the field while staying within the context of diabetic nephropathy.
Diabetes mellitus is rapidly gaining pandemic proportions. By 2030, the number of people with diabetes is projected to reach a staggering 350 million . Chronic poorly managed diabetes gives rise to complications such as neuropathy, retinopathy and nephropathy that are generally grouped as “microvascular” complications of diabetes mellitus. Roughly one-third of the diabetic population will develop diabetic nephropathy (DN) that manifests itself as microalbuminuria, macroalbuminuria and eventually renal failure . In the event of non-effective treatment/therapy, macroalbuminuria leads to gradual decline in glomerular filtration rate (GFR) and eventually end-stage renal disease (ESRD) . However, the extent of disease progression may not correlate with albuminuria as has been suggested by a recent study [4•]. Another recent global, cross-sectional study (11,573 patients) found that chronic kidney disease (CKD) was present in 17 % of subjects with normoalbuminuria [5•]. What leads to GFR decline in the absence of proteinuria (mostly prevalent in people with type 1 diabetes) is not known, although intrarenal vascular dysfunction may provide some explanation. The progression may occur typically within 15 years after the onset of diabetes . Structurally, diabetic nephropathy progresses in stages, starting with the thickening of the glomerular basement membrane, mesangial cell expansion and then gradually progressing into glomerulosclerosis and interstitial fibrosis eventually culminating in renal failure. Interstitial fibrosis is characterized by excessive deposition/turnover of extracellular matrix components (ECM) such as collagen type I and III as well other ECM components that play crucial role in the progression of interstitial fibrosis [7, 8].
The pattern of interstitial fibrosis may differ depending on the insult. However, the most commonly observed pattern in biopsies is that of diffused fibrosis associated with diseased glomeruli or atrophied tubules or the affected vasculature [9, 10]. Tubular atrophy is probably associated with microvascular changes or glomerular filtration rate changes. At the molecular level, several cell types have been implicated in bringing about these changes, at least in experimental models of fibrosis. Briefly, the following cell types have been shown to contribute to various aspects of the evolution of fibrosis. Fibroblasts are thought to acquire myofibroblastic phenotype that then produce collagen type III . Recently, fibrocytes, a cell population distinct from fibroblasts, have also been implicated in matrix production, mostly collagen type I. They have been also associated with systemic nephrogenic fibrosis . Renin/angiotensin system also has been suggested to modulate ECM production .
Clinical and experimental evidence suggests that several pathogenetic mechanisms can lead to DN. For example, glomerular hyperfiltration, accumulation of advanced glycation end products (AGEs), cytokine activation are among most commonly endorsed pathways. Related to AGEs, more recently the receptor for advanced glycation end products (RAGE) has been implicated as well. RAGE is a signal transducing receptor for the AGEs of proteins and lipids and its overall role in CKD has been discussed in detail elsewhere [14••]. AGEs have been shown to induce RAGE gene expression by increasing peroxisome proliferator-activated receptor gamma (PPARγ) activity and activate inflammatory pathways that in turn may contribute to fibrosis [15, 16]. The critical role of inflammation in fibrosis is discussed below. Another entity of interest is the nuclear hormone receptors that have become major pharmaceutical targets for a number of diseases. Eight nuclear receptors have been implicated in DN. This has been discussed in more detail in a recent review article . In addition, genetic susceptibility is thought to play a role in some patients [18, 19].
Role of MicroRNAs (miRNA)
miRNAs are short RNAs that are produced endogenously to repress and regulate the expression of genes and thereby control cell cycle, proliferation, etc. Evidence is now emerging to support their role in renal diseases including DN that is broadly discussed in a recent review article [20••]. While no details about the role of miRNAs will be discussed here, we will briefly mention some that might become important therapeutic targets in future. Inhibiting miRNA-192 was shown to be protective against fibrosis in DN  while suppressing miRNA-29 resulted in collagen production . miRNA-451 has been shown to modulate mesangial hypertrophy in early stages of DN . Similarly, mirRNA-29c ablation prevents progression of DN in db/db mice .
The extent of renal damage in diabetic patients cannot be solely explained by increased systemic and intraglomerular pressure. Over a decade of research points to the inflammation as one of the key pathophysiological mechanisms promoting DN. Long-term innate immune response that results in chronic inflammation promotes disease instead of repair . Renal tubular cells in response to proteinuria express proinflammatory cytokines like (IL-1, IL-6, IL-18, TNF-α) and numerous chemokines, reviewed elsewhere . This suggests that the inflammatory response is most likely associated with DN from very early stages. Hyperglycemia, hemodynamic alterations, formation of AGEs, increased levels of angiotensin-II are all associated with activation of inflammatory response via stimulation of nuclear factor κB (NFκB). NFκB is a ubiquitously expressed transcription factor that can be activated by a myriad of stimuli associated with DN such as proinflammatory cytokines.
Treatment Options for DN
As mentioned above, DN is characterized by inflammation, accumulation of mesangial matrix and tubulointerstitial fibrosis. Fibrosis and matrix deposition correlate with the progression of the disease to ESRD [30–32].
Renin-Angiotensin System (RAS)
The standard regimen of treatment has been tight glycemic control, inhibitors of the renin-angiotensin system (RAS) that includes angiotensin converting enzyme inhibitors (ACEIs) and angiotensin II receptor blockers (ARBs) [33–35]. RAS inhibitors potentially provide added protection as anti-fibrotic agents via inhibition of TGF-β. However, the RASS study showed no significant benefit associated with the use of ACE inhibitors or ARBs in participants with diabetes who had normal albumin levels or normal blood pressure . Blocking RAS thus may not be sufficient to prevent DN. In addition, the ONTARGET study demonstrated that combination therapy with an ACEI and an ARB is deleterious in people with diabetes who had normoalbuminuria or microalbuminuria. These patients had a higher rate of doubling of serum creatinine . Among the ARBs, telmisartan may have the best potential as an anti-fibrotic agent because of its added property of being a partial PPARΥ agonist imparting it anti-inflammatory properties. This is known to reduce production of monocyte chemoattractant protein-I (MCP-1) that would result in fewer macrophages and hence reduced fibrosis. However, evidence from patients with DN is lacking at this point. A detailed discussion about telmisartan’s potential has been recently discussed by Balakumar et al. [38•].
Several experimental in vitro and in vivo studies suggest that hyperlipidemia can lead to glomerulosclerosis and tubulointerstitial fibrosis [39–43]. These studies further suggest that hyperlipidemia can promote infiltration of monocytes, macrophages (intraglomerular) and cause podocyte damage. Hyperlipidemia stimulates the production of cytokines, chemokines and TGF-β in mesangial and tubulointerstitial cells. TGF-β promotes the extracellular matrix deposition and fibrosis. Nearly two decades ago, patients with type 2 diabetes were shown to have hyperlipidemia that associated with severity of nephropathy . Statins have renoprotective effects independent of their lipid-lowering properties. This is most likely due to the reduced macrophage infiltration in the glomeruli by attenuating the production of proinflammatory chemokines, such as macrophage chemoattractant protein 1 (MCP-1), muted cytokine production, such as TNF-α and IL-1 β [44–46]. Even though statins are associated with antifibrotic benefits in experimental DN, no hard evidence is available as yet in humans. Furthermore, meta-analysis of about 27 randomized, controlled trials demonstrated that statins slowed down the decline in eGFR; however, no significant benefit was evident in a diabetic population . However, recently the CARDS trial indicated that after 4 years of treatment with statin, only a modest beneficial effect on the declining renal disease (as indicated by eGFR) was observed in the participants who had diabetes and albuminuria .
Connective Tissue Growth Factor (CTGF)
CTGF too has been implicated in fibrosis and is activated during diabetes . It can act downstream of TGF-β or independent of it. During DN, CTGF is highly expressed in the podocytes. Administration of the blocking antibodies to CTGF in diabetic mice resulted in retardation of the DN and fibrosis in an allograft nephropathy model [50, 51]. While a randomized blinded trial is awaited, a phase-I trial indicates that anti-CTGF-antibodies are well tolerated and actually led to reduced albuminuria .
Bardoxolone is an antioxidant inflammation modulator that has been beneficial in animal models of acute kidney injury. Since DN is associated with both oxidative stress and inflammatory processes (both of which can lead to fibrosis) bardoxolone was thought to be potential drug of choice for treating DN in humans. The results of phase II double blind, randomized, placebo-controlled trial (BEAM) involving 227 patients showed an improvement in eGFR in advanced CKD patients with type 2 diabetes at 52 weeks [53, 54]. However, since bardoxolone significantly increased albuminuria, and side effects such as muscle spasms as well increased systolic blood pressure by 2–4 mmHg have collectively led to a reduced enthusiasm for the use of this drug in DN.
We will discuss this drug in more detail, as this seems to hold some promise in near future. Pirfenidone is a drug originally developed by InterMune for treating idiopathic pulmonary fibrosis (IPF) that has been approved in Europe, Japan and China for treating IPF [55•]. Various studies have established its antifibrotic and anti-inflammatory effects in animal models and negative modulation of TGF-β as well as matrix production by in vitro systems . This prompted the hypothesis that pirfenidone might be effective in patients with DN by slowing down the decline in eGFR and fibrosis. Accordingly, a double-blind, placebo-controlled, exploratory study was recently conducted in subjects with type 1 or type 2 diabetes and established nephropathy. Two doses (1,200 and 2,400 mg/day) were used based on the open label FSGS study . Seventy-seven subjects with established diabetes, proteinuria and reduced eGFR (20–75 ml/min per 1.73 m2) were enrolled and randomized. Fifty-two subjects completed the 54-week study. The mean eGFR in the group receiving pirfenidone, 1,200 mg/day, increased (+3.3 ± 8.5 ml/min per 1.73 m2), whereas it decreased in the placebo group (−2.2 ± 4.8 ml/min per 1.73 m2), P = 0.026 versus pirfenidone at 1,200 mg/day. However, pirfenidone at 2,400 mg/day was not significantly different than the placebo group and had a very high dropout rate due to adverse events at this high dose. There was no change in albuminuria among the three groups. (Of note, in diabetic db/db mice pirfenidone treatment improved glomerular histology and reduced matrix proteins without altering albuminuria ). The increase in eGFR in the pirfenidone group was observed after 3–6 months, suggesting that most likely it is independent of any effects on hemodynamics. Whether it is due to any structural improvements would require biopsies to check for fibrosis markers. Future studies would benefit from biopsies, the only definitive way for determining the anti-fibrotic effects (direct or indirect) of any drug being tested. Although this study determined a positive correlation between several fibrosis markers and the progression of DN, the drug itself did not have any effect. One needs to take into account however that these are serum markers and they may not reflect the true picture of the kidney where microenvironment is likely to be different. Hence, biopsy becomes critical. Even the urine TGF-β levels were not significantly affected by pirfenidone treatment although this could have been due to small sample size and variability in urine levels of TGF-β among patients. It is heartening that pirfenidone is being tested in two Phase III trials for idiopathic pulmonary fibrosis in North America, Europe and Australia although the drug is not yet approved in the USA. Thus, pirfenidone seems to have potential to be the first antifibrotic drug that might be useful for treating DN.
Other Potential Targets
As is evident, there are not many anti-fibrotic drug options available today. In this section, we will try to summarize pathways/targets that may have potential for attenuating fibrosis in general.
Transforming Growth Factor-β
TGF-β expression is activated by several stimuli including chronic hyperglycemia and in DN [57, 58]. Numerous experimental studies have associated TGF-β with CKD and the progression of fibrosis . Studies using neutralizing antibodies to TGF-β in mice/rat models of DN demonstrate significant effects on matrix production, renal hypertrophy [60, 61] without impacting albuminuria . Blocking TGF-β receptor type II with small interfering RNA in mice that had undergone ureteral obstruction to induce kidney fibrosis demonstrated reduced fibrosis . Some of the protective effects of the ACE inhibitors may in part be attributable to reduced TGF-β levels . Despite some promising results in animal models of DN no data is yet available from any human studies, although several pharmaceutical companies are pursuing TGF-β as a therapeutic target. Of interest is a report by Murphy and co-workers implicating induced in high glucose-1 (HIG-1) that is highly upregulated in biopsies of subjects with diabetic nephropathy and in experimental models of renal fibrosis in general. The authors conclude that HIG-1 can accentuate the effects of TGF-β1, thereby contributing towards interstitial fibrosis . In further support of TGF-β1, Brennan and co-workers recently identified TGF- β1-associated gene profiles in renal tubular cells that complimented the profiles established from kidney biopsies from the subjects of diabetic nephropathy .
Lysophosphatidic Acid (LPA1) Receptor Antagonist
LPA is a phospholipid that is known to regulate cell motility, proliferation, survival and cell differentiation via its specific G-protein coupled receptors (LPA1, LPA2, LPA3 and LPA4) . LPA levels were observed to be are upregulated in the CKD patients . Following this, Pradère et al.  demonstrated that LPA1 receptor activation could mediate renal interstitial fibrosis in a mouse model of unilateral ureteral obstruction (UUO). It was noticed that following development of fibrosis the expression of LPA1 receptor was highly upregulated and the mice lacking this receptor sub-type had attenuated fibrotic response. Notably, mRNA expression of certain fibrotic markers such as collagen III, α-smooth muscle actin (α-SMA) and F4/80 (macrophage marker) were all significantly lower in mice lacking the LPA1 receptor and had undergone UUO. Further, wild-type mice that underwent UUO and were treated with a LPA1 receptor antagonist had significantly muted fibrotic response and significantly attenuated TGF-β and CTGF mNA expression. What is encouraging is that a recent study in mice using orally active LPA1 receptor antagonist demonstrated attenuation of bleomycin-induced lung fibrosis . This raises the potential that LPA/ LPA1 receptor pathway could be a common pathway leading to fibrosis. Currently there are phase II trials going on for idiopathic lung fibrosis using two different LPA1 receptor antagonists, one by Bristol Myers Squibb and the other by Sanofi (no data available yet).
Ezetimibe is a potent cholesterol absorption inhibitor. A recent study in db/db mice demonstrated that mice treated with this drug had improved vascular endothelial function that was attributable to improved eNOS function and reduced oxidative stress. Importantly, these mice had significantly less cardiac interstitial fibrosis, lesser coronary arterial thickening and reduced macrophage infiltration . More recently, Tamura et al.  demonstrated that db/db mice treated with Ezetimibe ameliorates early DN. This protective effect was attributed to its lipid lowering properties that resulted in 50 % reduction in urinary excretion of albumin in mice receiving the therapy. However, reduced macrophage load too would inhibit fibrotic response. Mice had received treatment from 8 weeks of age for 8 weeks. Also, important to note is that it did not affect the blood pressure or glucose metabolism.
Activin-like Kinase 3 (Alk3)
Alk3 is one of the receptors for bone morphogenetic protein 7 (BMP7), a member of the TGF-β superfamily. Bmp7 has been shown to have anti-inflammatory functions and acts as a TGF-β antagonist [73, 74]. Sugimoto et al. [75••] now report that Alk3 functions to inhibit fibrosis by modulating inflammation. Using five different mouse models of renal injury (acute and chronic) this study demonstrated attenuation of fibrosis by using a small peptide agonist of BMP7 that functions through Alk3 receptor, thereby bringing Alk3 into the spotlight. This is interesting because a recent study suggests activation of the BMP-4 signaling mimics DN in non-diabetic mice and it utilizes Alk3. In diabetic mice BMP-4 was significantly upregulated and was responsible for ECM accumulation [76••]. It has also been reported earlier that there is loss of BMP-7 in experimental DN and neutralization of endogenous BMP7 raises the expression of fibronectin and collagen III mRNA levels in renal tubular cells . Thus, BMP7 pathway may provide some potential therapeutic targets for DN in future and the small peptide that has been developed could pave the way.
Targeting inflammation using pro-resolving eicosanoid lipoxinA4 and its synthetic analogue, benzo-LXA4, has shown promise at least as prophylactic. When administered intravenously prior to ureteral obstruction, it blunted the fibrotic response in rats. Specifically, collagen deposition and macrophage infiltration both were attenuated .
Although human gene therapy has had mixed results, Celec et al.  present convincing data in support of gene therapy as an option for treating DN using a rat model of DN. In this study, two therapeutic targets were used: angiogenic modulator, Angiomotin (Amot) to block angiogenesis and an inhibitor of MCP1 to inihibit macrophage infiltration, in streptozocin-induced diabetic rats. Three weeks after the induction of diabetes, rats received two injections of plasmid DNA expressing 7ND (MCP1 blocking sequence) or Amot (angiogenesis inhibitor). Rats were sacrificed after 4 months. The treatment resulted in reduced structural alterations of glomeruli and attenuated fibrotic markers and that of inflammation (TGF-β, TNF-α, IL-6, collagen I) and reduced vascular endothelial growth factor (VEGF) levels in the renal cortex, without affecting proteinuria. Along the similar lines, Flaquer et al.  report that hepatocyte growth factor (HGF) gene therapy too may provide a means to regenerate of a diabetic kidney. The authors argue that hepatocyte growth factor (HGF) enhances macrophage infiltration, which in turn was protective. Further studies would be required to determine how HGF is inhibiting fibrosis because macrophages (M2) themselves contribute to the process of fibrosis. Does HGF prevent alternate activation of marophages to the M2 state?
Rho-associated Kinases (ROCK)
The Rho GTPases and their downstream effectors Rho-associated kinases mediate several pathophysiological signals triggered by the diabetic milieu. Activation of ROCKs during experimental DN stimulates the production of extracellular matrix (ECM) and cytokine production. Studies using ROCK inhibitors have shown reduced fibrosis with or without antiproteinuric effects or any change in hyperglycemia or blood pressure. However, the specificity of ROCK inhibitors remains an issue, but overall evidence suggests that ROCK inhibitors might broaden our choice of treatment options for addressing DN. However, it may not be suitable for patients with overt DN, as in some of the studies inhibitors were effective only when started at the time of diabetes onset and not if started late in the disease process. Various studies using ROCK inhibitors have been discussed in two recent review articles [80, 81].
Activated Protein C
Activated Protein C (aPC) is a natural anti-coagulant that exhibits anti-inflammatory and cytoprotective properties. A recent study demonstrated attenuation of DN by the exogenous administration of aPC in streptozotocin-induced diabetic mice. Mice treated with aPC showed improved blood nitrogen urea-creatinine ratio, proteinuria, and had lesser fibrosis as indicated by collagen content. Furthermore, the total MCP-1 expression in kidney tissues was markedly lower, suggesting potentially less macrophage infiltration. In addition, aPC-treated mice expressed less VEGF, TGF-β and CTGF mRNA expression compared with vehicle-treated diabetic mice . The authors report that no effects on coagulation pathway were observed. Another study reports prevention of diabetes in NOD mice that was attributed to the suppression of inflammation and the activation of the regulatory T cells (Tregs). Interestingly, in contrast to study mentioned above, they reported increased TGF-β expression, (which would be necessary for T-cell activation). They reported reduced IL-1 β and interferon-Υ suggesting inhibition of pro-inflammatory pathway, in aPC-treated mice . Nonetheless, aPC could potentially provide an option for slowing down fibrosis associated with DN.
Macrophages are the first responders to any “perceived” tissue/organ injury regardless of the nature of the insult. Recently, there has been heightened interest in understanding their role in kidney injury and repair. Biopsies of diabetic patients (such as the one shown in Fig. 1) show immune cell infiltrates and the urine from DN patients has high levels of MCP-1 protein. Ample experimental evidence supports a clear role of macrophages in injury as well as the repair process via phenotypically two distinct subtypes, M1 (pro-injury) and M2 (pro-repair/fibrosis) [84–88]. In other models of renal fibrosis such as UUO, depletion of macrophages associates with the attenuation of fibrosis. Unpublished observations from our laboratory too support the idea that macrophages play an important role in promoting fibrosis. Although a number of studies clearly demonstrate a role of macrophages in promoting fibrosis, intervention (in DN patients) directed towards macrophages is not feasible. Systemic administration of any inhibitor directed towards macrophages will not be desirable because they perform tissue homeostatic functions and are important for normal host defense. Moreover, it is the microenvironment that will determine the phenotype of macrophages. So, we need to identify novel targets that for example regulate pro-fibrotic phenotype of macrophages so that the intervention, even if systemic, will be directed towards a sub-set of macrophages only. However, that will also depend on whether the “factor X” is normally present at high levels, whether it has a function elsewhere in other organ systems, etc. Moreover, it is likely that multiple factors in the microenvironment will influence the phenotypic switch. If that microenvironment is similar in different kidney diseases would have to be determined. Thus, while macrophages are clearly important players in fibrosis, we are still long ways off from utilizing the information towards a feasible therapeutic strategy in general.
There is ample evidence from the experimental models that many pathways are implicated in the pathogenesis of DN. However, despite our improved understanding about pathogenic mechanisms, there are no targeted therapies available as yet. Specifically, no remedies are in place for treating fibrosis. One stumbling block might be the inherent difficulty in creating an animal model that can adequately mimic human disease. In humans, it takes at about 5–10 years before any signs of DN are visible. Most of the animal models currently available are diabetic for relatively much shorter duration . A second challenge will be conducting the trials in these patients for any potential anti-fibrotic drug. For example, to definitively confirm a beneficial effect on fibrosis, it is imperative to biopsy the patients and that might not always be possible. Another point to consider is whether to take the patients off their ongoing medications (ACEIs) before conducting a trial. Should the drug being tested only have anti-fibrotic properties without affecting blood pressure? After all we do not necessarily need another blood pressure medication. Should the patients adhere to a uniform diet while on trial? But, before reaching that stage druggable targets that are kidney-specific must be identified. However, should we instead search for targets that are common to multiple organs such as lungs and kidney? Given the lengthy course of fibrosis in CKD (whether DN or otherwise), it is likely that any anti-fibrotic drug would have to be prescribed for a long period of time that raises the potential for long-term side effects. In experimental models, it is almost impossible to study/identify any long-term unintended side effects of any drug, and a similar problem may ensue in human studies. Can we have biomarkers to identify DN at a very early stage when no structural damage is obvious and no albuminuria is present so as to get a head start? Finally, given all the hurdles that we still need to cross, we should probably go back and rely on the age-old adage “prevention is the best cure” and try to achieve tighter glycemic control.
No potential conflicts of interest relevant to this article were reported.