Gene and protein markers of diabetic nephropathy

Diabetic nephropathy (DN) develops in a large number of type 1 (T1D) and type 2 diabetic (T2D) patients after a variable latency period. DN absolute number is increasing because of greater longevity and the epidemic of diabetes [ 1 ]. Although this disease is progressively becoming a heavier burden for the health care system, it is as yet poorly understood in many aspects. The pathogenesis of DN is clearly multifactorial and several genes, proteins and environmental factors are likely to contribute to the onset of the disease. Several metabolic, haemodynamic and intracellular causes have been proposed to play a role in the pathogenesis of DN [ 2 ]. Acquired risks also participate in the development of kidney hyperfiltration [ 3 ]. As a primary initiator of DN, high glucose level is associated with an increased synthesis of cytokines and growth factors and with the diversion of glucose metabolism into at least three metabolic pathways, the polyol, the protein kinase C and the hexosamine pathways, as shown in Figure 1 (reviewed in [ 4 ]). Pathways involved in the development of diabetic nephropathy. AGE: advanced glycation endproducts, DAG: diacyl glycerol, SLC2A1: glucose transporter 1, PPARγ: peroxisome proliferator-activated receptor gamma, ROS: reactive oxygen species and TGFβ: transforming growth factor beta. Due to the growing burden of the management of diabetes and its complications, it is important to identify DN predictors, in order to facilitate its diagnosis and treatment [ 5,6 ]. In this review, we first examine the genetic factors potentially associated with the disease, and then survey the literature on early urinary protein indicators of DN. It is thought that specific genetic backgrounds might influence DN development. Indeed, only 30% of T1D and 40% of T2D patients will develop DN irrespective of their diabetic treatment [ 7 ], and DN often shows a familial clustering in siblings with diabetes. The observation of a familial clustering of the disease strongly suggests that genetic factors are involved in the development of DN, whereas segregation analyses point to the existence of susceptibility genes [ 8,9 ] and have established that the onset and progression of DN are genetically determined [ 10,11 ]. The prevalence rates of DN also show marked ethnic variations. The reason for such inter-racial differences in incidence is unclear; nevertheless, genetic predisposition seems to constitute a major component in the development and/or progression of DN. Two strategies can be used to identify DN susceptibility loci: the linkage analysis (i.e. family-based studies) or the association analysis (i.e. case-control studies). Both have led to the discovery of many chromosomal and gene regions that may confer susceptibility to DN. Genome-wide scans are valuable tools for the identification of loci that could include major genetic components of a disease. Genome-wide scans have identified several DN susceptibility loci. Two complete genome scans for DN have been performed in T2D sib pairs. In Pima Indians, four loci with multipoint LOD >1.0 on chromosomes 3, 7, 9 and 20 were identified [ 12 ] and in African Americans, evidence for linkage was reported for chromosomes 3, 7 and 18 [ 13 ]. A partial genome-wide scan in Turkish families with type 2 DN reported linkage on chromosome 18 [ 14 ], a finding that was weakly replicated in Pima Indians (LOD = 0.3). Another genome-wide scan for type 1 DN in the Finnish population revealed a linkage to a single locus on chromosome 3q with a maximum LOD of 2.67 [ 15 ]. This locus has also been reported by others in T1D patients [ 16,17 ]. Table 1 recapitulates these studies. Chromosomal regions linked to DN Chromosomal regions linked to DN Association studies are commonly used to test for association between traits (disease) and highly correlated single nucleotide polymorphisms (SNPs), within a candidate gene or region, selected from the literature and from the HapMap database (www.hapmap.org). Several candidate genes have been identified for association with DN using case-control studies. They were selected regarding their positional and/or functional characteristics and the contribution of the corresponding proteins in the pathophysiological axes. Several pathways may contribute to the development and/or progression of DN, such as the renin-angiotensin system (RAS) or the peroxisome proliferator-activated receptor (PPAR), as suggested by clinical trials using RAS blockers or PPAR gamma agonists ( Figure 1 ). Table 2 includes the HUGO symbol and the OMIM reference of the discussed genes. HUGO gene symbol and OMIM reference of the discussed genes HUGO gene symbol and OMIM reference of the discussed genes The angiotensin-converting enzyme (ACE) converts angiotensin I into the biologically active angiotensin II. Recently the DCCT/EDIC study reported a relative protective effect of the II genotype (insertion/insertion polymorphism: rs1799752) over the deletion allele (D) on the development of microalbuminuria and progression to overt proteinuria in 1300 T1D patients, followed for 17 years [ 10 ]. The II genotype is also associated with a better response to ACE inhibitors [ 18 ]. Interestingly, microalbuminuric T2D patients with the DD genotype present more severe glomerular lesions than patients with the I allele [ 19 ]. A meta-analysis including 47 studies (14 727 subjects) [ 20 ] showed that subjects with the II genotype had a 22% lower risk of DN than carriers of the D allele [OR = 0.78, CI 95% (0.69–0.88)]. In the Boright et al. study [ 10 ], the haplotype TIC/TIC (which corresponded to the rs1800764, rs1799752 and rs9896208 alleles) was associated with a lower risk for development of persistent microalbuminuria ( P = 0.0009) and severe nephropathy ( P = 0.006), whereas in the Hadjadj et al. study [ 21 ], the haplotype analysis showed that DGG (rs1799752, rs4366 and rs12449782) was associated with an increased risk for DN, compared with the ICA. In summary, it appears that polymorphisms in the ACE gene may have a role in the progression of DN, rather than in the susceptibility to it. Aldose reductase (AKR1B1) is a cytosolic enzyme that, in the presence of NADPH, catalyzes the rate-limiting step of the polyol pathway converting glucose into sorbitol. Several mechanisms have been proposed to explain how AKR1B1 activity leads to hyperglycaemia-induced lesions in different tissues [ 22 ]. Ko et al. [ 23 ] were the first to identify seven alleles at the locus of the (AC)n dinucleotide repeat sequence upstream of AKR1B1. The most common allele contains 24 (AC) repeats and was named Z. Several studies have demonstrated a correlation between the Z-2 allele (23 repeats) and susceptibility to an increased risk of DN in both T1D and T2D [ 24–28 ]. However, other studies have reported a lack of association [ 29–31 ]. Interestingly, diabetic patients, who are homozygous for the XbaI (−) allele of the glucose transporter 1 gene and carry the Z-2 allele, present a 9-fold increased risk of DN, suggesting that this is a particularly high-risk genetic combination [ 32 ]. A second AKR1B1 polymorphism has been observed at position −106 of its promoter region. This C106T polymorphism was identified in both Caucasian and Asian subjects with T1D or T2D, and association with DN has been observed. Liu et al. [ 28 ] have suggested that the C106T polymorphism might be associated with an increased expression of AKR1B1 in the Chinese population. Neamat-Allah et al. [ 25 ] conducted a meta-analysis including seven studies and showed that the carriage of the −106 T allele was significantly associated with DN [OR = 2.2 CI 95% (1.69–2.94)]. Sivenius et al. [ 33 ] and Gosek et al. [ 34 ] suggested that this polymorphism could be involved in the early development of microalbuminuria in Finnish T2D patients and is a risk factor for development of DN in T2D patients with poor glycaemic control, respectively. In the Makiishi et al. study [ 35 ], the TT genotype presented a risk for DN [OR = 4.7 CI 95% (1.3–17)] in T2D Japanese subjects. All these results suggest that AKR1B1 polymorphisms play a role in DN development. PPARG2 is the predominant adipose isoform of this receptor and is expressed selectively in the adipose tissue where it modulates the expression of target genes implicated in adipocyte differentiation and glucose homeostasis. PPARG2 is considered, therefore, a major candidate gene for T2D and/or obesity and, recently, for type 2 DN. Insulin resistance may be a DN risk factor in type 2 diabetes [ 36 ] and improved insulin sensitivity could be the link between the Pro12Ala polymorphism of PPARG2 and the decreased risk of DN. Three studies have evaluated its association with type 2 DN. In the study by Herrmann et al . [ 37 ], the Pro12Ala polymorphism was associated with lower albumin excretion rates among Ala12 carriers with type 2 DN, which may indicate a protective effect of this allele. These findings were confirmed by Caramori et al . [ 38 ]. More recently, Pollex et al . [ 39 ] showed that the Ala12 allele carriers have reduced occurrence of microalbuminuria (1.5-fold reduction of the albumin/creatinine ratio). All these results indicate a protective correlation between the Ala12 polymorphism and the albumin excretion rate. The mechanism underlying the protective effect of the Ala12 allele is yet unknown. The glucose transporter (SLC2A1) is the major representative of the family of facilitative glucose transporters that are expressed in glomerular, mesangial, endothelial cells and podocytes. SLC2A1 is likely to be pivotal in raising intracellular glucose levels by activating pathogenic pathways [ 40–42 ]. Several works have tried to determine whether SLC2A1 might be a candidate gene conferring susceptibility to DN. These investigations have focused on testing whether the SLC2A1 XbaI SNP (rs841853) is associated with DN using a case-control study design. Although this XbaI SNP was found not to be associated with DN [ 43,44 ], heterozygosity for the XbaI (−) allele has been linked with a 1.9-fold higher susceptibility to nephropathy in Chinese T2D patients [ 45 ]. However, homozygosity for the XbaI (−) allele was also associated with an increased DN risk in other studies [ 32 , 46 , 47 ]. Inconsistent results have thus so far been reported regarding the possible role of the SLC2A1 XbaI SNP in DN susceptibility. A recent meta-analysis concluded that there is, indeed, a significant association between the SLC2A1 XbaI polymorphic site and DN, but larger studies are needed [ 48 ]. Four SNPs have also been identified in SLC2A1 putative enhancer elements, i.e. enhancer-2 SNP1 (rs841847) and enhancer-2 SNP2 (rs841848), one in exon 2 (rs1385129) [ 47 ] and one in the 5′ promoter region (rs710218) [ 49 ]. The authors showed that the AA genotype of rs841847 is a ‘risk genotype’ (OR = 2.38) and that the TT genotype of rs710218 is associated with DN ( P < 10 −6 ). Moreover, they showed that patients with the AG haplotype (rs841847–rs841853) have an increased risk of DN with an OR = 2.4; also, the TT haplotype (rs710218–rs841853) is more frequent in nephropathic patients ( P < 7 × 10 −7 ) and might be associated with an enhanced glucose uptake rate via SLC2A1 in mesangial cells. Our recent findings showed that two haplotypes (composed of rs1385129–rs841847–rs841848) are associated with a 4.4- and 2.6-fold increased risk of nephropathy in the Tunisian T2D patients (to be published). In conclusion, the SLC2A1 gene seems to play an important role in the development of DN. Carnosine dipeptidase 1, a member of metallopeptidase M20 family, degrades the dipeptide carnosine. It reduces the synthesis of matrix components, fibronectin, collagen IV and TGFβ2 (transforming growth factor beta 2) in renal cell lines [ 50 ]. The carnosinase gene, CNDP1, both a functional and a positional candidate gene, resides in the 18q linkage region, as previously reported [ 13,14 ]. CDNP1 encodes for the secreted serum carnosinase. A trinucleotide repeat in exon 2, coding for a leucine repeat in the leader peptide of the carnosinase-1 precursor, has been associated with DN [ 50,51 ]. Since this polymorphism lies in the 5′ coding part of the transcript, the number of trinucleotide repeats is directly related to the number of leucine residues in the leader peptide of the carnosinase precursor, i.e. five, six or seven leucines. The five-leucine allele, the shortest allelic form, is more common in subjects without nephropathy, with an OR of 2.56, and is associated with lower serum carnosinase levels [ 50 ]. This is expected to result in higher renal concentrations of the protective dipeptide carnosine. It would be interesting to evaluate whether carnosine administration or carnosinase activity inhibition could prevent the development of nephropathy in susceptible diabetic individuals. It is currently impossible to reliably predict which, and when, diabetic patients will develop nephropathy and progress to kidney failure. Clinically, DN can be diagnosed when albumin is detected in the urine (albumin excretion 30–300 mg/day). The presence of albumin in the urine is considered predictive of the subsequent development and clinical progression of DN [ 52 ]. However, microalbuminuria could also be regarded as an indicator of an established glomerular damage. Moreover, microalbuminuria is not specific for DN, since albumin can also be detected in some other pathological conditions and the prevalence of microalbuminuria in the general population is rather high. Finally, microalbuminuria is a poor predictor of DN, since only 30–45% of microalbuminuric patients develop overt proteinuria over more than 10 years [ 53 ]. However, it is well established that regression of microalbuminuria, obtained by renin–angiotensin system blockers, is associated with protection from the development of overt DN. The conclusions of the Benedict study, where ACE inhibitors prevent development of microalbuminuria [ 54 ], and Diabiopsies study, where ACE inhibitors reduce renal interstitial expansion [ 55 ], support this view. If microalbuminuria is a useful indicator of developing DN and regression of renal lesions by treatment, it is far from meeting the criteria of a predictive marker. As a consequence, the identification of biological markers appearing earlier than albumin would be highly desirable. We review here some of the proteins that have recently been considered of interest to complement or even replace the dosage of albumin in the urine. Standard methods for measuring albuminuria are based on the immunochemical detection of serum albumin [radioimmunoassay (RIA), immunonephelometry (IN) and immunoturbidimetry (IT)]. HPLC methods can also be used to identify non-immunoreactive albumin, allowing the detection of microalbuminuria 3.9 and 2.4 years earlier than conventional measurement in T1D and T2D patients, respectively [ 56 ]. In the large Australian diabetes, obesity and lifestyle cohort, the prevalence of microalbuminuria was found to be four times higher using HPLC than by IN (20 versus 5.5%); also, 17.4% of individuals classified as normoalbuminuric with IN were reclassified as microalbuminuric by HPLC [ 57 ]. Another study showed that the mean urinary albumin concentration was 6.8 ± 4.3 mg/L by nephelometry in 998 subjects versus ± mg/L with HPLC [ ]. conventional may the albumin concentration because albumin in urine might in two and non-immunoreactive [ ]. However, the higher of albumin detected by HPLC has been to by proteins that with albumin [ ]. the HPLC results it is to the that is A system has been recently used to both and non-immunoreactive This was to more albumin than the reference conditions where of the albumin by several proteins was [ ]. such a system has over conventional but to be evaluated in a large of IV collagen is the major component of the glomerular matrix and the level of type IV collagen in the urine might the rate of matrix in by showed that type IV collagen is found in the urine of diabetic patients with microalbuminuria and not in the urine of patients from in which albuminuria can be present [ ]. These findings indicate that type IV collagen has a for DN that albumin has but since urinary collagen IV is only when microalbuminuria is present [ ], this the of that is for an indicator of DN an urinary type IV collagen is one of the criteria for early diagnosis of DN that the Japanese Diabetic to [ ]. is an of levels are linked to insulin resistance and type 2 diabetes. et al. in the urine of T2D patients with a protein an [ ]. they demonstrated that the concentration is significantly correlated with both the urinary albumin level and the serum observed ± The in urinary may result from the of the serum concentration and enhanced the findings were more recently that both serum and urinary concentrations are in T2D patients with nephropathy, as compared to patients with microalbuminuria [ ]. are of the glomerular and proteins can be potentially interesting urinary markers for the early diagnosis of an of the [ ]. Indeed, urinary have been detected only in the urine of diabetic patients with and whereas they were in T2D patients without microalbuminuria or patients from renal [ ]. is one of the many proteins that have been in the et al. have for the presence of in of T1D They found in 30% of normoalbuminuric of of and of patients who recently microalbuminuric [ ]. was in the urine of The finding that of normoalbuminuric T1D patients is of a of Since et al. not in the urine of normoalbuminuric T1D patients [ ], it is possible that in the of the disease, is as a be to the possible interest of proteins and their since are of the kidney glycation are a of proteins and to which residues are The of the that to the development of diabetic is not yet Since are thought to be one of the for the of the occurrence of in the urine and its putative interest have been [ ]. The in urine of type 2 diabetic patients with microalbuminuria, or persistent shows significant differences between and microalbuminuric However, since has also been detected in of individuals and is in of diabetic patients with microalbuminuria, this in its present form, not the criteria of a valuable Finally, proteins can be with a In patients with T2D, serum are higher than in and their concentration is linked to the glomerular rate [ ], that this test may have clinical Two recent point to a possible of the transforming growth protein urinary is an matrix protein that is by transforming growth factor a growth factor implicated in the pathogenesis of DN. et al. [ ] first showed that the to is significantly higher in T2D patients than in subjects. A correlation was found between urinary and albumin excretion and even normoalbuminuric diabetic patients have a higher than In study [ ], it was also observed that the urinary level of is higher in diabetic patients ± than in individuals ± these studies results for a of in the early detection of renal in to methods at and the complete of proteins present in a biological at a In et al. compared the urinary of type 1 and type 2 DN patients proteinuria to the urinary of [ ]. a of they found that several in the diabetic were or This study early markers of DN, because patients with proteinuria were of urine from T2D patients with microalbuminuria showed that four proteins and [ ]. It is not yet whether the excretion of some of proteins may albumin in the of DN. It was reported that a of seven urinary proteins increased in concentration with increasing whereas four other proteins were [ ]. Another recent used a to for specific for early DN. were found to significant differences between patients with early DN and patients treatment of the led to the of possible risk factors for DN [ ]. the of the was not on et al. found a in the urine of T2D patients who DN [ ]. The reported sensitivity and is in of a Finally, in a recently [ ], the authors found and as selectively in the urine of type 2 with DN. gene analysis the study of both major and gene but The observed could be by differences in the particularly ethnic backgrounds and genetic and by their The of SNPs in association studies of has most of the the single gene with a large effect has been identified and only of a of polymorphisms in a number of genes have been linkage analyses association studies performed support the of major gene polymorphisms involved in DN development and/or The maximum LOD was linked to chromosome 18 [ 14 ] and the OR found was for carriers of XbaI (−) and Z-2 alleles of SLC2A1 and AKR1B1 genes, respectively [ 32 ]. in and will into the mechanisms involved in DN development. This will the early identification of patients at risk or DN and will to specific strategies [ , ]. As for urinary protein the protein to be detected earlier than albumin in the of DN. or their could such candidate since of and of the are of albumin from the matrix could also be as DN is by a of this is a biological that has only recently to be by [ ]. As many as different proteins have been recently identified from a of [ ]. This of the urinary is both and for one it appears that the urine the protein is high and not yet the other more that identified urinary proteins a useful database for the identification of that would not to the protein or would to of of interest .