Main

The prevalence of diabetes mellitus has reached pandemic proportions and continues to increase rapidly across the globe. In 2015 an estimated 415 million individuals had diabetes mellitus, and this number is expected to rise to 642 million by 2040 (Ref. 1). Although the prevalence of both type 1 diabetes mellitus (T1DM) and type 2 DM (T2DM) is increasing, the most pronounced increase is seen among individuals with T2DM1. Diabetes mellitus has a substantial impact on health, and is responsible for considerable individual and socioeconomic costs1. The most serious complication of diabetes, diabetic angiopathy, can be divided into two subtypes: microangiopathy, which affects the capillaries and arterioles in the retina (retinopathy), nerves (neuropathy) and the kidney (nephropathy); and macroangiopathy, which affects arteries in the brain (cerebrovascular disease), heart (for example, ischaemic heart disease and congestive heart failure) and the lower extremities (peripheral artery disease)1,2.

The development of diabetes mellitus and the late diabetic complications are intimately linked to low-grade inflammation. This association is evidenced by the presence of high levels of inflammatory markers such as high-sensitivity C-reactive protein (hsCRP)3,4,5. These data suggest that late diabetic complications might be induced by inflammation; however, the underlying mechanisms that initiate low-grade inflammation are poorly understood. Substantial amounts of data support a role for the innate immune system in inflammatory processes associated with diabetes mellitus6, and roles for pattern recognition receptors, such as membrane-bound Toll-like receptors and nucleotide-binding oligomerization domain (NOD)-like receptors, have been described in association with diabetic nephropathy6. The complement system has a key role in innate immune defence through interactions with different pattern recognition receptors, but is also an important driver of inflammation. Increasing evidence points toward a role for the complement system in the pathogenesis of diabetic nephropathy, and suggests that complement activation might not only identify patients at risk of this complication but might also be targeted therapeutically.

This Review describes the experimental and clinical evidence supporting a role for the complement system in the pathogenesis of diabetic nephropathy, and discusses how this association might facilitate the identification of new biomarkers of disease progression and targets for therapeutic inhibition.

Diabetic nephropathy: an overview

Epidemiology

Diabetic nephropathy is a chronic condition that develops over many years, and is typically characterized by a gradual increase in urinary albumin excretion (UAE), blood pressure and risk of cardiovascular disease (CVD), and decreasing glomerular filtration rate (GFR) eventually leading to the development of end-stage renal disease (ESRD), and need for renal replacement therapy (RRT)7. Diabetic nephropathy can occur in patients with either T1DM or T2DM, and all patients with diabetes mellitus should therefore be examined regularly for renal dysfunction7.

Although intensive management of diabetes mellitus through concurrent control of glucose, lipids and blood pressure, can slow the progression of diabetic nephropathy, diabetes mellitus remains the most common cause of ESRD7. The proportion of individuals commencing RRT as a result of diabetes mellitus varies enormously across countries, ranging from 12% in Ukraine to 66% in Singapore8. In the USA, the incidence of ESRD among patients with diabetes increased by 50% between 1996 and 2006, but interestingly has decreased slowly since then, most likely due to more aggressive and efficient therapy8. The decrease in the incidence of ESRD among patients with diabetes mellitus over the past decade seems to be mainly driven by changes in patients with T1DM, with incidence rates of ESRD in these patients remaining stable or declining over the past decade, while remaining steady or slowly increasing among patients with T2DM9,10,11. Accordingly, diabetic nephropathy remains a substantial clinical issue for which greater understanding of the underlying mechanisms will hopefully lead to the development of new therapeutic strategies.

Pathogenesis

In general terms, diabetic nephropathy is caused by prolonged exposure to high glucose levels, as has been established by large-scale prospective studies of patients with T1DM12 and T2DM13,14.

Increasing evidence from the past two decades indicates that a number of different pathways mediate the effects of hyperglycaemia on the vasculature, and contribute to the development of diabetic nephropathy. The number of pathways and interaction between various factors are complex but have previously been grouped into four main categories: metabolic factors, haemodynamic factors, growth factors and/or cytokines and intracellular factors2. In addition to these four pathways, an increasing body of evidence supports an important role for the innate immune system10, including the complement system9, in the development of diabetic nephropathy.

The complement system

The complement system is an important component of the immune system and has a key role in facilitating the clearance of microorganisms and damaged cells by antibodies and phagocytic cells. Activation of the complement system is typically thought to occur through three pathways: the lectin pathway, which is triggered by carbohydrates on cell surfaces; the classical pathway, which involves antibody-mediated activation via the C1 complex; and the alternative pathway, which involves direct activation of C3 through surface-binding or tick-over5. Activation of any one of these pathways leads to the production of complement C3 convertase, which activates complement component C3, leading to generation of the opsonic C3b and eventually generation of the membrane attack complex (MAC), which lyses, damages, or activates target cells (Fig. 1).

Figure 1: The complement system and its targets.
figure 1

The complement system is typically thought to involve three main pathways. The classical pathway is activated by binding of the C1 complex, comprising C1q, C1r and C1s, to antibodies, which in turn cleave complement components C4 and C2. The lectin pathway is activated through the recognition of carbohydrates by mannose-binding lectin (MBL) or recognition of N-acetylglucosamine residues by ficolins. Upon binding, MBL associated serine proteases (MASPs) become activated and cleave C4 and C2. The alternative pathway is activated through the spontaneous activation of C3b. All three pathways lead to cleavage of C3, resulting in generation of the C5b−9 complex, known as the membrane attack complex (MAC). Approaches to inhibit the complement system can be categorized into five main groups. (1), agents that disrupt the initiating complexes or inhibit initiating enzymes of the lectin and classical pathways (for example, neutralizing monoclonal antibodies against MBL, C1q, C1s or MASP) or enzymes that inhibit complement initiation (for example, the C1 inhibitor, C1INH, and nafamostat mesilate). (2), agents that inhibit activating enzymes of the alternative pathway (for example, humanized immunoglobulin G1 (lampalizumab) and small-molecule protease inhibitors). (3), agents that inhibit C3 convertases and/or C3 activity (for example, C1INH). (4), agents that inhibit C5 convertases and/or C5 activity (for example, eculizumab, coversin and C5-specific aptamers). (5), agents that inhibit MAC function (for example, soluble recombinant CD59).

PowerPoint slide

Full size image

Mannose-binding lectin (MBL) is considered the classic activator of the lectin pathway, although studies from the past decade have shown that a group of pattern recognition molecules (PRMs) called ficolins (including H-ficolin, L-ficolin and M-ficolin) can also activate the lectin pathway9,15,16,17,18,19 (Fig. 1). Initiation of the lectin pathway occurs when MBL binds to glycan-associated mannose, or ficolins bind to N-acetylglucosamine, N-acetylgalactosamine or N-acetyl-neuraminic acid residues, on microbial surfaces15. Such binding activates MBL-associated serine proteases (MASPs), leading to cleavage of the complement components C2 and C4 and formation of the C4b2b C3 convertase16. Activation of the classical pathway occurs through the formation of an antigen–antibody complex via the C1q domain of complement C1 (Fig. 1). C1 is a non-covalent complex comprised of three different proteins, C1q, C1r and C1s. After formation of the complex, C1q activates C1r and C1s, which then cleave C2 and C4, leading to formation of the C4b2b C3 convertase. The alternative complement pathway is activated in an antibody-independent manner on foreign surfaces, such as those of viruses, bacteria, and biomaterials, and leads to the subsequent production of a C3bBb C3 convertase (Fig. 1). The C3b fragment generated by any of the complement pathways can bind factor B (FB) and, facilitated by factor D (FD), can form the alternative pathway C3bBb C3 convertase, thereby amplifying the complement cascade17. Finally, the alternative pathway is constitutively activated at low level by the continuous hydrolysis of C3, producing an intermediate that forms alternative pathway-initiating C3 convertases with FB and FD via a mechanism known as tick-over.

Following cleavage of C3 to C3a and C3b, the three pathways merge to follow a single pathway. The C3b fragments bind to C3 convertases to produce a C5 convertase, which in turn divides C5 further into C5a and C5b. Thus, C6–9 binds to C5b to form MAC9,18 (Fig. 1).

Complement in diabetic nephropathy

The evidence for a connection between the complement system and renal dysfunction spans decades, and is supported by findings from experimental and clinical studies5,20,21,22,23. Two main mechanisms are thought to explain the involvement of complement in the development of diabetic nephropathy. First, activation of the lectin pathway occurs following binding of PRMs to proteins that are glycated as a result of exposure to sugars24,25. Second, hyperglycaemia is thought to induce glycation of complement regulatory proteins26,27, leading to dysfunction of their regulatory capacity, which in turn might facilitate complement auto-attack, through overactivation of complement pathways25,27. These pathophysiologic mechanisms are described in further detail below.

Dysregulation of the lectin pathway

Findings from a number of in vitro studies suggest that diabetes-induced alterations in glycoproteins might stimulate complement activation through binding of MBL to neo-epitopes24,25. These diabetes-induced changes might develop through altered enzymatic protein glycosylation or by a non-enzymatic formation of advanced glycation end-products (AGEs)24,25 (Box 1).

MBL. A number of experimental studies have demonstrated a causal relationship between MBL and diabetic nephropathy. Compared to wild-type mice with streptozotocin-induced diabetes (a model of T1DM), diabetic MBL-knockout mice have less renal damage, with reduced renal hypertrophy, UAE and collagen IV expression, suggesting a potential pathological role of MBL on the kidney under diabetic conditions28.

The adverse renal effects of MBL signalling under diabetic conditions have since been confirmed by other studies in mice with streptozotocin-induced diabetes29,30,31. One study that examined MBL levels after the induction of diabetes as well as the half-life of injected recombinant human MBL showed not only a rise in endogenous MBL levels after the induction of diabetes, but also an increase in the half-life of recombinant human MBL, indicating that the elevation in MBL levels with diabetes is a result of increased MBL production and decreased MBL turnover30. A 2016 study demonstrated a twofold increase in MBL levels in glomeruli of mice with established streptozotocin-induced diabetes compared to levels in non-diabetic controls31, indicative of a direct effect of MBL in the diabetic kidney31.

The clinical relevance of these experimental studies has been confirmed in clinical studies of patients with T1DM and T2DM. A cross-sectional study of normoalbuminuric patients with T1DM reported elevated levels of circulating MBL compared to those of healthy controls32; moreover, MBL levels were positively associated with UAE levels, even within the normal UAE range32. Another study reported higher serum MBL levels among individuals with T1DM and microalbuminuria or macroalbuminuria than in patients with normoalbuminuria33. In addition, genotypes associated with high circulating levels of MBL have been associated with the development of diabetic nephropathy and increased mortality among patients with T1DM34. In a cohort of individuals with T1DM, patients with genotypes associated with high MBL levels had a 50% increased risk of developing nephropathy32 and significantly greater mortality over 10 years of follow-up than patients with a 'low expression' MBL genotype35. The findings from this study contrast with those from an earlier study that found no association between 19 MBL single nucleotide polymorphisms (SNPs) associated with MBL concentrations and the presence of T1DM or diabetic nephropathy36. The reason for this discrepancy is unknown. In a study of patients with newly diagnosed T1DM, serum MBL levels measured shortly after diabetes onset were robust predictors of incident microalbuminuria over 18 years of follow-up37. Additional studies support an association between MBL and diabetic nephropathy among patients with T1DM36,37,38,39, including a large prospective multicentre study, which reported an association between concentrations of MBL and progression of renal disease from macroalbuminuria to ESRD in individuals with T1DM38.

Finally, in patients with T1DM and renal failure who received a simultaneous pancreas–kidney (SPK) transplant, preoperative serum MBL levels predicted graft and patient survival, with low MBL levels associated with improved outcomes40. Moreover, a 2016 study of patients with T1DM and diabetic nephropathy who underwent SPK transplantation or kidney transplantation alone41, found that circulating MBL levels normalized in patients who had undergone successful SPK transplantation, and that normalization mostly occurred in patients with a polymorphism in the MBL2 gene41. By contrast, circulating MBL levels remained elevated among patients who received a kidney transplant alone, indicating that glycaemic control, and not the reversal of ESRD, is associated with decreased MBL levels. In support of this proposal, levels of glucose and glycated haemoglobin (HbA1c), but not serum creatinine levels or estimated glomerular filtration rate correlated with MBL levels41.

As in T1DM, MBL is strongly predictive of worsening UAE and mortality among patients with T2DM42. Several studies from the past few years have confirmed the utility of circulating MBL to predict the development of diabetic nephropathy in T2DM43,44,45. One study that assessed the association of MBL2 polymorphisms with T2DM and diabetic nephropathy among a Northern Chinese population showed an association of GA and AA genotypes of rs1800450 with T2DM, but no association between MBL2 polymorphisms and diabetic nephropathy44. Patients with the GG genotype of rs1800450 and the CC genotype of rs11003125 had high serum MBL levels, however, and the researchers reported an association between elevated serum MBL and diabetic nephropathy44.

Together the above-mentioned studies confirm the value of circulating MBL levels as a robust predictor of diabetic nephropathy in T1DM and T2DM, although findings relating to MBL polymorphisms have not always been consistent. Of note, several of the above-mentioned studies in patients with T1DM and T2DM have studied the value of MBL and hsCRP, alone and in combination, for predicting the development of diabetic nephropathy38,42, and have demonstrated that levels of MBL and hsCRP are not correlated, which is most likely due to the strong influence of genetic factors on circulating levels of MBL. This lack of correlation between MBL and hsCRP probably explains the prognostic value of combining the two biomarkers for predicting progression of diabetic nephropathy.

Ficolins. A few studies from the past couple of years have examined whether ficolins, like MBL, also have a role in the development of diabetic nephropathy. In one study, deletion of the M-ficolin orthologue, B-ficolin, had no effect on the development of diabetes-induced renal damage in mice with streptozotocin-induced diabetes46. This finding suggests that unlike MBL, B-ficolin (and potentially M-ficolin in humans) does not have a role in the pathogenesis of diabetic nephropathy. As suggested by the authors of that study, the different outcomes associated with the deletion of B-ficolin and MBL in experimental models of diabetes might relate to differences in their specific carbohydrate-binding properties46.

A prospective 18 year observational follow-up study examined the association between H-ficolin and the risk of diabetic nephropathy among an inception cohort of 270 individuals with newly diagnosed T1DM47. Following adjustment for HbA1c, systolic blood pressure, smoking status and baseline UAE, patients in the highest quartile for baseline H-ficolin levels had a twofold greater risk of developing worsening UAE than patients in the lowest quartile. Thus H-ficolin levels seem to be robustly associated with risk of progression to microalbuminuria or macroalbuminuria47. So far no human studies have been published on the potential association between L-ficolin or M-ficolin and the risk of developing diabetic nephropathy.

MASP. In contrast to the substantial number of studies that have demonstrated a role for MBL in diabetic nephropathy, only one published study has investigated the role of MASPs in diabetes. The investigators compared levels of MASP-1, MASP-2 and MASP-3 in 30 children with T1DM and 45 adults with T1DM with levels in non-diabetic controls48. Compared to age-matched controls, levels of MASP-1 and MASP-2 were significantly higher among adults and children with T1DM, whereas levels of MASP-3 did not differ between patients and controls. HbA1c level correlated with levels of both MASP-1 and MASP-2, with these levels decreasing as a result of improved glycaemic control. So far no studies have been published on MASP levels in T2DM. Although the elevated levels of MASP-1 and MASP-2 might be indicative of increased complement activity in T1DM, the potential relevance of MASP-1 and MASP-2 to the development of diabetic nephropathy requires further investigation in patients with T1DM and T2DM.

Complement components C3, C4 and C5. A number of studies support a role for downstream components of the complement system — C3, C4 and C5 — in diabetic nephropathy. Some, but not all experimental studies have reported an accumulation of complement component C3 in animal models of diabetes. In studies from the 1970s, the induction of T1DM in rats by administration of streptozotocin led to an increase in glomerular C3 levels, which was normalized by transplantation of pancreas islets49,50. Furthermore, transplantation of kidneys from rats with T1DM into non-diabetic rats led to decreased expression of glomerular C3 and decreased mesangial volume in 50% of transplanted kidneys51. However, a 2016 study that demonstrated an increase in glomerular MBL levels in streptozotocin-induced diabetic mice, did not find a difference in the glomerular immunofluorescence intensity of complement factors C3, C4, or C9 compared to that of non-diabetic controls, although circulating levels of the complement activation product C3a were increased in diabetic compared to control mice31. Whether the apparent disparity in levels of glomerular MBL and glomerular C3 levels are real or due to methodological issues in that study is not known31. In contrast to the above-described 2016 study, elevated levels of C3 have been detected in the kidney of two other mouse models of T1DM — the non-obese diabetic mouse52 and the OVE26 diabetic mouse53.

Deposition of C3 in glomeruli and glomerular capillaries has also been observed in KK mice — a model of T2DM54. Similarly, a 2015 study of Zucker rats, which are a hypertensive, dyslipidaemic, obese and hyperglycaemic model of T2DM, found that a single episode of renal ischaemia was followed by a significant increase in levels of renal mRNA encoding C3, C4, C5, C6, C8 and C9, along with notable elevations in levels of renal mRNA encoding C3a and C5a55.

The role of downstream complement components in diabetic nephropathy has also been assessed in clinical studies. Transcriptome and immunohistochemical analyses found a sixfold increase in glomerular levels of C3 in kidney biopsy samples from patients with overt diabetic kidney disease56. In addition, a 2013 study showed that plasma levels of C3 were significantly higher in patients with T2DM and macroalbuminuria than in those with normoalbuminuria57.

The relationship between complement component C4 and diabetic nephropathy has also been described in a number of clinical studies. Among 64 children with T1DM, 25% of patients had low circulating complement C4 levels, with no association with circulating levels of complement C3 fragments58. Another study showed that patients with T1DM and microangiopathy had lower circulating levels of C4 than did patients with normoalbuminuria and that C4 levels were inversely correlated with the degree of complement activation59. This inverse association was not confirmed, however, by another study, which might suggest that low C4 levels might be a result of the diabetes, rather than have a pathogenic role in the disease60. Evaluations of C4 allotypes in patients with diabetes with or without microvascular complications have not shown consistent results61,62.

Membrane attack complex. A series of immunohistochemical studies using antibodies directed against the C9 component of MAC63,64,65 have localized MAC to the glomerular basement membrane, tubuli and Bowman capsule in renal tissue from patients with T1DM63,64,65. Evaluation of renal tissue from patients with varying degrees of renal function demonstrated a correlation between levels of MAC and the magnitude of mesangial expansion among patients with T1DM. The deposition of MAC in diabetic glomeruli has been confirmed in more recent studies27,66. These results are interesting as MAC is the end product of the complement system and elevated MAC levels in T1DM seems to highlight the universal activation of the complement system from MBL, MASP, and C3 to MAC. In the following section the potential mechanistic effects of an overactive complement system are discussed.

Mechanistic considerations

As described above, available evidence suggests that activation of the complement system has a pathogenic role in hyperglycaemia-induced renal injury. Under normal conditions MBL and ficolins do not bind to their receptors on cell surfaces; however, advanced glycation (Box 1) of proteins (resulting in the production of AGEs) in response to hyperglycaemia generates neo-epitopes to which lectin pathway PRMs bind24,25,27,67,68. In addition, accumulating data indicate increased levels of glycation in diabetes leads to glycation-induced dysfunction or inactivation of complement regulatory proteins such as CD59, which normally prevents the deleterious effects of complement overactivation by inhibiting MAC26,27,68. The combined effects of glycation-induced inactivation of CD59 and hyperglycaemia-induced activation of complement signalling have been proposed to increase tissue deposition of MAC, thereby activating intracellular signalling pathways, which in turn release proinflammatory cytokines and growth factors68. Three different pathways have been proposed to lead to the late complications of diabetes: those that are dependent on the complement system and activation of MAC (for example, activation of the lectin pathway through glycation of receptor carbohydrates in addition to pathways involving CD59); those that are independent of complement (for example, the polyol pathway or the hexosamine pathway); and those that are triggered by either hyperglycaemia and/or MAC (such as pathways involving reactive oxidative species (ROS), protein kinase C (PKC) and nuclear factor-κB (NF-κB)68,69 (Fig. 2).

Figure 2: The pathogenic role of complement in hyperglycaemia-induced renal injury.
figure 2

Under normal conditions mannose-binding lectin (MBL) or ficolins do not bind to their receptors on cell surfaces. However, advanced glycation of proteins in response to hyperglycaemia generates neo-epitopes to which lectin pathway pattern recognition molecules bind. In addition, accumulating data indicate that increased levels of glycation in diabetes leads to glycation-induced dysfunction or inactivation of complement regulatory proteins such as CD59 (not shown), which normally prevents complement overactivation by inhibiting the membrane attack complex (MAC). The combined effects of glycation-induced inactivation of CD59 and hyperglycaemia-induced activation of complement signalling have been proposed to increase tissue deposition of MAC, thereby activating intracellular signalling pathways, which in turn release proinflammatory cytokines and growth factors. Three different pathways have been proposed to lead to the late complications of diabetes: those that are dependent on the complement system and activation of MAC (that is, activation of the lectin pathway through glycation of receptor carbohydrates in addition to pathways involving CD59); those that are independent of complement (for example, the polyol pathway or the hexosamine pathway); and those pathways that are triggered by either hyperglycaemia and/or MAC (for example, actions involving reactive oxidative species (ROS), protein kinase C (PKC) and nuclear factor-κB (NF-κB). DAG, diacylglycerol; DHAP, dihydroxyacetone phosphate; Gln, glutamine; Glu, glutamate; NADP+, oxidized NADPH; NADPH, nicotinamide adenine dinucleotide phosphate; P, phosphate; UDP-GlcNAc, uridine diphosphate-N-acetylglucosamine.

PowerPoint slide

Full size image

Targeting the complement system

Increasing evidence of a role for the complement system in various diseases has stimulated interest in the development of antibodies, small molecules and biologics to target this pathway. Although attempts to block complement span more than 50 years, only a few drugs have made it into clinical trials, and even fewer have passed beyond phase 1 clinical trials70. Nevertheless, complement therapeutics remains an area of active interest, with new insights from molecular, genomic, and structural studies, as well as the approval and success of some complement inhibitors, such as the C5 inhibitor, eculizumab, for diseases such as atypical haemolytic uraemic syndrome, encouraging further research in this field68,70.

Approaches to modulate complement signalling can be categorized into five main groups. First, those that disrupt initiation of the lectin and classical pathways by inhibiting initiating complexes (for example, neutralizing monoclonal antibodies against MBL, C1q, C1s or MASP) or inhibiting initiating enzymes (for example, serine protease inhibitors such as the C1 inhibitor, C1INH, and nafamostat mesilate). Second, those that inhibit activating enzymes of the alternative pathway (for example, the humanized immunoglobulin G1, lampalizumab, and small-molecule protease inhibitors). Third, those that inhibit C3 convertases and/or C3 activity (such as C1INH). Fourth, those that inhibit C5 convertases and/or C5 activity (for example, eculizumab, coversin, and C5-specific aptamers) Finally, those that inhibit MAC function (for example soluble recombinant CD59)70 (Fig. 1).

Only a few experimental studies have assessed the therapeutic effect of blocking complement signalling in diabetic nephropathy. One study examined the effect of thrombomodulin on the development of renal changes in a diabetic mouse model of T1DM71. Thrombomodulin inhibits coagulation but also prevents complement activation via its lectin-like domain. Diabetic mice lacking the lectin-like domain of thrombomodulin had increased complement activation and more severe diabetic nephropathy than diabetic wild-type mice. Inhibition of complement by administration of a low molecular weight heparin (enoxaparin) reduced albuminuria and podocyte injury in the mutant mice. In vitro studies demonstrated that the lectin-like domain of thrombomodulin prevented glucose-induced complement injury of podocytes, suggesting that the lectin-like domain of thrombomodulin protects against diabetic renal damage by limiting glucose-induced complement activation71.

Two studies have investigated the effects of blocking complement C3a receptors72,73 and/or C5a receptors72 in Sprague–Dawley rats with streptozotocin-induced T1DM72 and a model of T2DM, induced by exposing Sprague–Dawley-rats to high fat diet and low-dose streptozotocin73. Blockade of C3a and C5a receptors in the T1DM model ameliorated endothelial-to-myofibroblast transition through the Wnt/β-catenin signalling pathway, indicating a potential protective effect of complement blockade on renal fibrotic changes72. Similarly, blockade of C3a receptors in rats with T2DM improved renal morphology and function by inhibiting cytokine release and TGFβ/Smad3 signalling, again supporting a role for C3a in diabetic renal changes73.

A study in another rat model of T2DM (the Otsuka Long–Evans Tokushima fatty rat), demonstrated that blockade of complement component C5 with the inhibitor K-76COONa diminished the severity of albuminuria and mesangial expansion74. Glomerular deposition of C3 was more pronounced in untreated diabetic rats than in diabetic animals treated with K-76COONa74, supporting a pathogenic role for C5 in diabetes-induced renal damage74.

The best approach to target the complement system to prevent the development of diabetic nephropathy is not currently clear. A number of compounds that modify the complement system are already in clinical use in diseases other than diabetic nephropathy (Fig. 1). The most promising of these agents should be tested for their ability to prevent or slow the progression of diabetic nephropathy. However, it is important to note that the complement system has an important role in host defence; despite the potential benefits of blocking the complement system on diabetes-induced renal damage, any approach to modulate complement signalling carries considerable risks. Individuals who are deficient in components of the lectin pathway or the classical pathway (for example, MBL, MASP-1 and MASP-2, C1 or C4) typically present with recurrent bacterial infections. Patients with deficient classical pathway signalling can have an impaired ability to clear immune complexes, which in turn has been shown to increase the risk of lupus nephritis in patients with systemic lupus erythematosus70. Individuals who have a deficiency in a component of the alternative pathway are at increased risk of developing bacterial infections of Gram-negative origin70. Insufficiencies in the terminal paths of the complement system increase the risk of meningococcal sepsis and/or meningitis70. As proposed previously, the best way to minimize the risk of iatrogenic complications is to minimally affect the physiological effects of complement signalling and to block the relevant pathway as downstream as possible69. Whether chronic complement inhibition, as would be required for a disease such as diabetes mellitus, is associated with unacceptable adverse effects remains to be determined69.

Conclusions and perspectives

Strict metabolic control and appropriate renoprotective and antihypertensive treatment are crucial to prevent and treat nephropathy among patients with either T1DM or T2DM. The increasing incidence and prevalence of diabetes worldwide is leading to an increase in the number of patients with late diabetic complications, including nephropathy. Although numerous therapeutic interventions are available that can postpone the development and progression of diabetic nephropathy, many patients continue to progress to ESRD and an urgent need exists to develop new biomarkers to identify patients at risk of disease progression and new therapeutic agents to treat this devastating disease.

A growing body of experimental and clinical evidence supports a role for various components of the complement system in the pathogenesis of diabetic nephropathy. Moreover, circulating MBL and H-ficolin seem to be reliable and robust biomarkers of the development and progression of diabetic nephropathy in humans.

The past few years have witnessed increasing interest in the development of therapeutic modulators of the complement system64. Some compounds are already in clinical use for diseases other than diabetic nephropathy, and these should be tested for their ability to slow or halt the progression of diabetic nephropathy. Ongoing work will also likely lead to the development of new complement inhibitors that benefit patients with diabetic nephropathy, although the effect of any complement inhibitor on susceptibility to infections or immune-complex disease will need to be closely monitored.