Background
Diabetic kidney disease (DKD) is currently one of the main causes of end-stage renal disease (ESRD) [
1‐
5], with an estimated incidence of 26 % in patients starting dialysis [
6]. In addition to ESRD, DKD also causes an important increase in cardiovascular-related morbidity and mortality [
6,
7].
An increasing body of evidence supports a genetic basis for DKD [
8,
9]. However, little is known about the mode of transmission, which is probably polygenic, or about the mechanisms of gene-environment interaction. Until this moment, genome wide-association study (GWAS) and candidate gene studies have produced heterogeneous results [
8‐
13].
A GWAS of 360,000 single nucleotide polymorphisms (SNPs) has identified 13 SNPs associated with DKD in two independent populations of patients with type 1 diabetes mellitus (T1DM) [
14]. The strongest association was observed at the
FRMD3 (4.1 protein ezrin, radixin, moesin [FERM] domain containing 3) locus. Moreover, other relevant association was observed at
CARS (cysteinyl-tRNA synthetase) locus. Confirmation of implicated SNPs was obtained in participants of the Diabetes Control and Complications Trial (DCCT)/Epidemiology of Diabetes Interventions and Complications (EDIC) study [
15‐
17].
Because patients with T1DM and type 2 diabetes mellitus (T2DM) might share common genes related to DKD, the present study was designed to (1) investigate whether the SNPs found to be associated with DKD in T1DM are also associated with DKD in T2DM; (2) evaluate the association of these SNPs in early (microalbuminuria) and advanced stages of DKD (macroalbuminuria or ESRD); and (3) evaluate gene and protein expression of the genes associated with DKD in human kidney biopsies.
Methods
Patients analyzed in the association study
A case–control study was conducted with 1098 white T2DM patients selected from a cross-sectional study performed in the state of Rio Grande do Sul, Brazil [
11]. T2DM was defined according to World Health Organization criteria [
18]: diagnosis of diabetes after the age of 35 years, no use of insulin during the first year after diagnosis, and no episodes of diabetic ketoacidosis. Control T2DM patients were those with known diabetes duration of at least 5 years and normoalbuminuria (albuminuria <30 mg/24 h, n = 380). Cases were divided into two categories: early DKD (albuminuria 30–299 mg/24 h, n = 323) or advanced DKD (albuminuria ≥300 mg/24 h or ESRD, n = 395). The protocol was approved by the ethics committee, and all patients gave their written informed consent.
All patients underwent an evaluation that included a standardized questionnaire and physical examination, as previously described [
19]. Briefly, information was collected about age, age at T2DM diagnosis, drug treatment and smoking. They were weighed bare feet, wearing light outdoor clothes and their height was measured. Body mass index (BMI) was calculated as weight (kg)/height
2 (meters). For patients on dialysis, the mean of three weights measured after dialysis sessions were used. Hypertension was defined as blood pressure (BP) ≥140/90 mmHg or use of any antihypertensive medication. BP was defined based on the mean of two measurements in the sitting position using a standard mercury sphygmomanometer (phases I and V of Korotkoff). Diabetic retinopathy (DR) was assessed by fundus examination after mydriasis by an experienced ophthalmologist and graded as absent, non-proliferative or proliferative diabetic retinopathy.
Fasting blood samples were collected for laboratory and molecular analyses. Fasting plasma glucose was determined by a glucose oxidase method and HbA1c by an ion-exchange high performance liquid chromatography procedure (Merck-Hitachi L-9100 Glycated Hemoglobin Analyzer, Tokyo, Japan) with inter- and intra-assay coefficient of variation (CV) of 2.4 and 0.5 %, respectively (reference interval 4.1–6.0 %) [
20]. Serum creatinine was determined by the Jaffé reaction [
21]. Triglycerides and cholesterol levels were measured by enzymatic methods. LDL-cholesterol was calculated using the Friedewald equation. Urinary albumin excretion (UAE) was measured in 24 h-urine samples by immunoturbidimetry (Bayer, TarryTown, NY, USA), with intra- and interassay CV of 4.5 and 11 % respectively [
22]. Use of angiotensin-converting enzyme inhibitors or angiotensin receptor antagonists was interrupted at least 1 week before albuminuria measurement.
DNA isolation and genotyping
DNA was extracted from peripheral blood leukocytes by a standardized salting-out procedure. Seven out of 13 SNPs initially associated with DKD in T1DM in the GoKinD Study [
15] were chosen to be included in the present study (rs1888747, rs9521445, rs39075, rs451041, rs1411766, rs6492208, rs1041466). Not all 13 SNPs were genotyped because some of them were in linkage disequilibrium (LD): rs1888747 and rs10868025 (
r
2 = 0.81), both located at chromosome 9q; rs39059 and rs39075 (7p,
r
2 > 0.96); rs451041 and rs739401 (11p,
r
2 > 0.87), rs9521445 and 7989848 (13q,
r
2 > 0.87), rs1411766 and 17412858 (13q,
r
2 = 1), and rs6492208 and rs2391777 (13q,
r
2 = 1) [
15]. In LD pairs, the SNP having the strongest association with DKD was chosen for analysis. LD data was obtained from the HapMap Project (
http://www.hapmap.org) for the Caucasian population. Polymorphism loci analyzed are shown in the supplementary material (Additional file
1: Table S1).
All polymorphisms were genotyped using primers and probes contained in the Human Custom TaqMan Genotyping Assay (40×) (Assays-By-Design Service, Life Technologies, Foster City, CA; USA). Primer and probe sequences used for genotyping these SNPs are described in Additional file
1: Table S2. Real-time (RT) PCR reactions were conducted in 96-well plates, in 5 µl total reaction volume using 2 ng of genomic DNA, TaqMan Genotyping Master Mix (1×) (Life Technologies), and Custom TaqMan Genotyping Assay (1×). Plates were positioned in a RT PCR thermal cycler (7500 Fast Real PCR System; Life Technologies) and heated for 10 min at 95 °C, followed by 40–50 cycles at 95 °C for 15 s and 60–62 °C for 1 min. The percentage of duplicates included in genotyping was 10 %. Genotyping success was more than 95 %, with a calculated error rate based on PCR duplicates of <0.01 %.
FRMD3 expression in kidney tissue samples
For gene expression analysis, kidney tissue samples were collected from 91 patients undergoing therapeutic nephrectomy over a period of 30 months at Hospital de Clínicas de Porto Alegre. A standard questionnaire was used to collect information regarding age, gender, presence of arterial hypertension, diabetes mellitus, and smoking habit. A peripheral blood sample was collected from each subject for DNA extraction and genotyping. Excised normal kidney tissue was divided into aliquots for mRNA expression analyses and evaluation of protein expression and localization by western blot (WB) and immunohistochemistry (IHC), respectively. From 91 kidney samples analyzed for gene expression, only 48 kidney samples were utilized for protein expression analyzes, since 30 samples lacked enough material for proper protein extraction and the protein extraction was unsuccessful in 13 samples due to technical difficulties. The study protocol was approved by the Hospital’s ethics committees, and all patients gave their written informed consent.
RNA isolation and quantification of FRMD3 gene expression by real-time reverse transcription PCR
Kidney biopsies were homogenized in phenol/guanidine isothiocyanate (Invitrogen Life Technologies, Carlsbad, CA). RNA was extracted with chloroform and precipitated with isopropanol by centrifugation (12,000×
g) at 4 °C. RNA pellet was washed twice with 75 % ethanol and resuspended in 10–50 μl of diethylpyrocarbonate treated water. Concentrations of isolated RNAs were assessed using NANODROP 2000 spectrophotometer (Thermo Scientific Inc., DE, USA). Only total RNA samples achieving adequate purity ratios (A260/A280 = 1.9–2.1) were used for subsequent analyses [
23]. In addition, RNA integrity and purity were also checked on agarose gel containing Gel Red Nucleic Acid Gel Stain (Biotium Inc, Hayward, CA). The mean (± SD) concentration of isolated RNA was 2.68 ± 1.69 µg/250 mg kidney.
Real-time reverse transcription-PCR (RT-qPCR) was performed in two separate reactions: first, RNA was reverse transcribed into cDNA, and subsequently cDNA was amplified by RT-qPCR. Reverse transcription of 5 µg of RNA into cDNA was carried out using the Super Script Vilo Master Mix Kit (Synthesis System for RT-PCR; Invitrogen). RT-qPCR experiments were performed in a 7500 Real Time PCR System (Life Technologies). Experiments were performed by real-time monitoring of the increase in fluorescence of SYBER Green dye [
24]. Primers were designed using Primer Express 3.0 Software (Life Technologies) and are depicted in Additional file
1: Table S2.
PCR reactions were performed using 10 µl of SYBER Green (1×) (Life Technologies), 1 µl of specific primers, FRMD3 or cyclophilin A (Invitrogen), 7 µl of water and 1 μl of cDNA (0.2 μg/μl) in a total volume of 20 μl. Each sample was assayed in triplicate and a negative control was included in each experiment. The thermocycling conditions for these genes were as follows: an initial cycle of 95 °C for 10 min, followed by 50 cycles of 95 °C for 15 s and 60 °C for 1:30 min.
Quantification of
FRMD3 cDNA was performed using the ∆∆Cq method [
23,
25] and expressed relative to the reference gene (
cyclophilin A) in 91 kidney tissue samples: 76 carrying the G allele (GG or GC) of the rs1888747 SNP, and 15 with the CC genotype. Validation assays were done by separate amplification of the target (
FRMD3) and reference (
cyclophilin A) genes using serial dilutions of a cDNA sample. As a requirement of this method, both target and reference genes exhibited equal amplification efficiencies (E = 95–105 %) in all experiments. The ∆∆Cq method calculates changes in gene expression as relative fold differences (n-fold changes) between an experimental and an external calibrator sample [
23]. RT-qPCR specificity was determined using melting curve analyses and all primers generated amplicons that produced a single sharp peak during the analyses [
26].
Determination of FRMD3 protein distribution and concentration in kidney
FRMD3 protein distribution and concentrations were determined in only 48 kidney sections, due to insufficient material for protein extraction. Thirty-nine of these sections belonged to subjects carrying the rs1888747 SNP G allele (GG = 23 and CG = 16). FRMD3 protein distribution was evaluated by IHC of formalin-fixed, paraffin-embedded kidney sections, while FRMD3 protein quantities were determined using WB. For IHC analyses, a goat rabbit anti-FRMD3 polyclonal antibody (Abcam, Cambridge, MA, USA) was used to detect FRMD3 protein distribution in human kidney tissue. IHC analyses were performed on 4 μm-thick kidney sections [
27]. The routine IHC technique comprised: deparaffinization and rehydration, antigenic recovery, inactivation of endogenous peroxidase and blocking of nonspecific reactions. Slides were incubated with primary antibody and then incubated again with a biotinylated secondary antibody, streptavidin horseradish peroxidase conjugate (LSA; Dako Cytomation Inc, Carpinteria, CA, USA), and diamino benzidine tetra hydrochloride (Kit DAB Dako Cytomation Inc). Images were visualized through a Zeiss microscope (model AXIOSKOP-40; Carl Zeiss, Oberkochen, Germany) and captured using the Cool Snap-Pro CS (Media Cybernetics, Rockville, USA) camera.
For WB analyses, proteins from human kidneys were extracted using RIPA buffer and quantified using a colorimetric BCA™ Protein Assay Kit (Thermo Scientific). Protein extracts (20 µg) were resolved on 10 % polyacrylamide gels, transferred to Immobilon
®-P
SQ membranes (Millipore, Billerica, MA, USA), and incubated with monoclonal antibodies to FRMD3 (Bioss, Woburn, MA, USA) or β-Actin (Millipore) overnight. Secondary antibodies consisted of horseradish peroxidase conjugated goat anti-rabbit antibodies (Millipore). Detection was performed using Immobilon Western Chemiluminescent HRP Substrate (Millipore), and images were acquired and quantified in an Image Quant LAS 500 (GE Healthcare, Pittsburgh, USA) digital imaging system. Initially, the process was applied for FRMD3 antibody. After stripping, the membrane was re-blotted with β-Actin antibody. Data are expressed as arbitrary units (AU). Three isoforms of FRMD3 protein (
http://www.uniprot.org/uniprot/A2A2Y4) were quantified separately (data not shown) and grouped by WB.
Statistical analyses
In the association study, we compared controls vs. cases (defined as any degree of DKD) and controls vs. different DKD stages (microalbuminuria, macroalbuminuria/ESRD). Data were analyzed assuming dominant, recessive and additive inheritance models to determine the best fit. Allele frequencies of all SNPs were determined by gene counting, and departures from Hardy-Weinberg equilibrium (HWE) were verified using the Chi square (χ2) test. Allele and genotype frequencies were compared between groups using the χ2 test. Continuous data with normal distribution are presented as mean ± SD, while continuous variables with skewed distribution were log-transformed before analysis and are presented as medians (minimum–maximum values), with exception of FRMD3 protein concentration (WB) for which log-transformation did not normalize the variable, thus we used the non-log variable and non-parametric tests. Categorical data are expressed as number of cases and percent of individuals affected. One-way analysis of variance (ANOVA), χ2 or Student’s t test were used to compare the groups in terms of clinical and laboratory characteristics or FRMD3 gene/protein expression. The magnitude of the association was estimated using odds ratios (ORs) with 95 % confidence interval (95 % CI). Multiple logistic regression analysis was used to evaluate the independence of possible SNP associations with DKD, adjusting for covariables. Pearson’s correlation test was used to evaluate correlation between quantitative variables. Multiple linear regression analyses were performed with FRMD3 gene or protein expression as dependent variables and age and sex as independent variables. Results for which P < 0.05 were considered statistically significant. All statistical analyses were performed using SPSS 18.0 (Chicago, IL, USA).
Discussion
In the present study, we found that rs1888747 SNP is associated with DKD in T2DM, as previously described for T1DM patients [
15]. The CC genotype conferred protection against the development of DKD. Additionally to previous data from T1DM [
15], a subgroup with microalbuminuria was evaluated. The confirmation of this SNP with DKD in our population should be considered carefully, since frequently genetic association studies failed to be replicate in additional populations [
13,
28,
29]. DKD is probably a multifactorial disorder resulting from an interaction between environmental and genetic factors. GWAS are capable of identifying unknown chromosomal regions that may be involved in the pathogenesis of DKD. However, replication of the findings in distinct populations is essential to confirm the associations observed in such studies.
The rs1888747 SNP (C/G) is an intergenic polymorphism, located near the promoter region of the
FRMD3 gene [
30]. This gene encodes a structural/cytoskeletal protein involved in maintaining cellular shape [
31,
32], but whose function remains otherwise unknown. Studies focusing on the relation between rs1888747 and DKD have been performed in T1DM and T2DM individuals of various ethnic backgrounds, such as Caucasians, Japanese, African Americans, and Chinese, with conflicting results [
15,
33‐
37].
A second SNP found in the GoKinD sample collection and replicated in DCCT/EDIC subjects is the rs451041 SNP, located in chromosome 11p. The nearest gene to rs451041 is
CARS, which is expressed in mesangial and proximal tubule cells. Mutations in this gene have been associated with neurodegenerative diseases and cystinosis [
38,
39]. However, neither this nor the other SNPs originally associated with DKD in T1DM patients were associated with DKD in the present study with patients with T2DM. Although this finding requires further validation, if confirmed it might suggest that different genes are involved in the development of DKD in T1DM and T2DM.
The evidence to infer susceptibility genes contributing to the pathogenesis of DKD remains elusive. Advances in genotyping technologies have allowed the identification of several chromosomal regions potentially associated with DKD. Nevertheless, methodological limitations such as lack of statistical power, differences regarding the population studied and phenotypes analyzed make it difficult to compare some of the study findings and build up on results.
In the present study, we also aimed to evaluate if the rs1888747 risk allele was influencing FRMD3 gene and/or protein expression. Unfortunately, we were not able to demonstrate an association of these variables with the different genotypes. The
FRMD3 gene (ID: 257019) is located on chromosome 9q21.32, size of 2.282 pb; 21 exons; 81.4 Mb position (
http://www.ncbi.nlm.nih.gov/gene/257019). Members of the protein 4.1 family have well characterized functions in a variety of cell types, including rat nephron [
31,
32].
FRMD3 is present in adult ovaries as well as fetal skeletal muscle, brain, and thymus [
40]. Our findings confirm that FRMD3 is expressed in proximal renal tubular cells and human podocytes, as previously shown [
14,
35], suggesting that this protein is involved in maintaining the function and integrity of the slit diaphragm. Therefore, it is a good candidate gene for albuminuria development. Furthermore, FRMD3 protein concentrations negatively correlated with
FRMD3 gene expression. Although this needs further confirmation, it suggests that
FRMD3 mRNA levels may not reflect the expression of the protein itself, probably due to important translational regulation mechanisms.
Recently, changes in
FRMD3 expression have been linked to progression of DKD in a group of 22 Pima Indians with T2DM [
41]. The authors proposed an influence of the rs1888747 polymorphism in the
FRMD3 promoter on transcriptional regulation within the bone morphogenetic protein (BMP)-signaling pathway [
41], suggesting that the transcriptional coregulation of BMP pathway members and
FRMD3 might be mediated by the four transcription factor binding site (TFBS) promoter modules in the functional context of DKD [
42]. The mechanism mediating the connection between
FRMD3 and BMP pathway members remains unknown. No evidence at the protein, RNA, or microRNA levels was found [
42]. The present findings, derived from genotyping, gene and protein expression studies in human kidney tissue, are important to elucidate the role of the rs1888747 polymorphism. Even so, further research is necessary to confirm our observations and to unveil the mechanisms underlying the association of the rs1888747 polymorphism with DKD.
A limitation of case–control studies of DKD such as the present one is survival bias. However, we do not believe that this had an impact on the results we report, because minor allele frequency was similar in mild and severe stages of DKD, although data from the prospective DCCT/EDIC cohort [
16] suggest that the rs1888747 G allele had a faster progression to severe DKD. Another possible limitation of our study is the fact that the samples for the expression studies were not obtained from patients with DKD, but instead from patients undergoing therapeutic nephrectomy. Thus, the present gene and protein expression findings cannot be directly translated to patients with DKD. Further studies are necessary to determine if the rs1888747 SNP is or not related to changes in FRMD3 gene and protein expression in DKD.
In conclusion, the present study replicates the protective effect of rs1888747 SNP against established DKD previously described for T1DM and T2DM. Moreover, in agreement with previous data, it also shows that this effect is already present in less advanced stages of DKD, such as microalbuminuria. Outside the context of diabetes, this study was not able to show an influence of rs1888747 SNP on FRMD3 gene and/or protein expression.
Authors’ contributions
MPB conceived and designed the study, collected and analyzed data, and wrote the manuscript; DC conceived and designed the study, analyzed data, and wrote the manuscript; AS collected and analyzed data and contributed to the discussion; RC analyzed data, contributed to the discussion, and reviewed the manuscript; BMS collected and analyzed data and contributed to the discussion, MPC collected data and contributed to the discussion; MIE analyzed data and contributed to the discussion; MB analyzed data and contributed to the discussion; to the discussion; LHC conceived and designed the study, analyzed data, and wrote the manuscript. All authors read and approved the final manuscript.