Identification of Chromosomal Regions Linked to Diabetic Nephropathy: A Meta-Analysis of Genome-Wide Linkage Scans

Abstract

Aims:

Diabetic nephropathy (DN) has become a serious public health problem. Genetic factors are involved in the pathogenesis of DN, but the exact mode of inheritance is still unknown. Genome-wide linkage scans (GWLS) have produced inconclusive or inconsistent results. In an effort to test consistency and provide more conclusive results, we applied a heterogeneity-based genome search meta-analysis (HEGESMA) to GWLS regarding DN.

Materials and Methods:

We combined results from eight GWLS in the primary analysis and nine GWLS for a conditional analysis about DN for both diabetes types, as well as in each type of diabetes and ethnicity in subgroup analyses.

Results:

HEGESMA identified cytogenetic locations that rank highly on average in terms of linkage statistics across multiple genome scans, taking into consideration the magnitude of heterogeneity of the results between scans. Main analyses: Our meta-analysis identified 13 cytogenetic locations (bins) with statistical significance (P_rank ≤ 0.05), 11 of which were significant in both weighted and unweighted analyses located on chromosomes 1q, 3q, 4p, 5q, 7q, 15q, 16p, 17q, 19q, and 22p. In addition, four novel regions (5q11.2-5q14.3, 5q23.2-5q34, 17q24.3-17q25.3, and 22q12.3-22q13.3) were identified. Seven bins on chromosomes 4p, 5q, 7q, 15q, 22p, and 22q were common between both types of diabetes and in all subgroup analyses, in addition 5q14.3-5q23.2 was significant across all analyses. Conditional analyses: meta-analysis identified nine different cytogenetic locations, among which 7p22.3-7p15.3 was significant only in type 2 diabetes mellitus conditional analysis. Ethnicity subgroup analyses identified 11 different cytogenetic locations, 5 out of which are novel findings. However, none of the chromosomal regions reached genome-wide statistical significance (P_rank < 0.00042).

Discussion:

This meta-analysis provides evidence for linkage for nine novel cytogenetic regions that should be further investigated for genes that confer susceptibility to DN.

Introduction

Diabetic nephropathy (DN) is a serious microvascular complication of either type 1 (T1D) or type 2 diabetes (T2D) characterized by progressive clinical course with loss of kidney function. It constitutes the most frequent cause of end-stage renal disease (ESRD) requiring chronic renal replacement therapy (Saran et al., 2016). Main risk factors of DN are poor glycemic control and long duration of diabetes mellitus (Dronavalli et al., 2008). However, patients with good glycemic control may still develop nephropathy, implying the contribution of genetic background, as well as epigenetic mechanisms (Rich, 2006; Kato and Natarajan, 2014). This fact and a proven significant familial clustering of DN (Seaquist et al., 1989; Borch-Johnsen et al., 1992; Quinn et al., 1996) clearly implicate that genetically defined predisposition is involved in the pathogenesis of nephropathy in both DM types. However, the genetic component has not yet been deciphered (Zintzaras et al., 2007a, 2009; Stefanidis et al., 2009, 2014, 2018; Tziastoudi et al., 2017).

Genetic association and family-based studies have shown that a significant number of variants are associated with DN. However, both categories of studies have produced inconclusive results so far (Mooyaart et al., 2011; Palmer and Freedman, 2012). Genome-wide linkage scan (GWLS) is an approach for identifying chromosomal regions that may be linked to complex diseases, such as DN. GWLS in DN have shown that genetic components play a role in the disease, but the exact inheritance pattern remains to be established. The genome searches on DN have produced inconclusive inferences; as linkage signals tend to be rather low in each study, the number of families is relatively small and the individual genome scans have identified evidence for linkage in different chromosomal regions.

Genome-wide scan meta-analysis (GSMA) is a method for synthesizing the results of genome searches to identify loci linked to DN (Wise et al., 1999; Zintzaras and Ioannidis, 2005a, 2005b). Zintzaras and Ioannidis (2005a, 2005b) developed an extension of the GSMA, the heterogeneity-based genome search meta-analysis (HEGESMA), which also evaluates the heterogeneity between different genome scans. HEGESMA is the most established method for meta-analysis of genome scans and it has been applied to genome scans of several complex diseases (Zintzaras and Ioannidis, 2005a, 2005b, 2008; Zintzaras and Kitsios, 2006; Trikalinos et al., 2006; Zintzaras et al., 2006; Zintzaras et al., 2007b; Bouzigon et al., 2010; Rao et al., 2011). In the present article, we performed a HEGESMA to provide conclusive evidence on regions linked to DN as a discreet trait (Zintzaras and Ioannidis, 2005a, 2005b).

Materials and Methods

Eligible whole genome scans

All GWLS for DN published before March 2018 were considered in the meta-analysis. We searched the PubMed database for English articles referred to humans. As a search criterion, we used the combination of the following terms: “genome search,” “genome scan,” “genome screen,” “logarithm of odds (LOD) score,” “linkage,” “nonparametric linkage (NPL) score,” “genome-wide” or “genome-wide linkage analysis,” and “diabetic nephropathy.”

The meta-analysis included GWLS fulfilling the following criteria: (1) subjects were humans (2) with DN or ESRD due to diabetes; (3) the required linkage score data were available or were extractable from published graphs; and (4) there was no sample overlapping.

All potentially eligible articles were read to assess their appropriateness for meta-analysis and references of eligible articles were also perused. In studies with overlapping cases, the largest study or the most recent one, including extractable data, was included. Scans restricted to specific candidate regions or chromosomes were excluded.

Definition of outcome

DN was defined on the basis of a long-standing diabetes mellitus, either T1D or T2D, with macroalbuminuria (urinary albumin excretion rate ≥200 μg/min or ≥300 mg/24 h, equivalent to an overt glomerular proteinuria) (Eknoyan et al., 2003) and/or chronic renal insufficiency (stage 3-5 chronic kidney disease) in the absence of nondiabetic renal disease (KDIGO, 2013). Cases were defined similarly in the included studies. Studies in which diabetic probands were selected regardless of the degree of renal function or albuminuria were rejected. Excluded from the meta-analysis were also GWLS with quantitative surrogate markers for DN as estimated glomerular filtration rate (eGFR), albuminuria, and serum cystatin.

Data extraction

For each study, the following information was extracted: first author, year of publication, ethnicity of study population, type of diabetes, number of families, selection criteria, number of markers and intermarker interval, linkage statistic, type of statistical analysis, metric and software of linkage analysis, and finally, the locations with chromosomal regions corresponding to maximum LOD score, as proposed by Lander and Kruglyak (1995).

Genome scan data, if not available, were derived from published figures by digitizing graphs of linkage scores per distance in each chromosome using Engauge Digitizer (version 2.12, Mark Mitchell, 2002). Regarding included scans, the corresponding investigators were asked to provide whole genome results to avoid a special form of publication bias when results are presented only for regions that show some evidence for linkage.

Genome scan meta-analysis and heterogeneity testing

GSMA starts by splitting the chromosomes into genetic regions (bins) of approximately equal length; usually, each bin has a width of 30 cM giving 120 bins for the whole genome. The 120th bin refers to the highest rank. For nomenclature purposes, bin c.n is the nth bin on chromosome c. For each genome scan, the most significant result within the bin is recorded. Then, for each scan, the bins are ranked according to their significance and the ranks for each bin, weighted or not by a specific factor, are summed across scans. The significance of the average rank of each bin is assessed empirically against the distribution of average ranks.

When a bin has a high summed rank, this is considered evidence for linkage. Equal test statistics for several bins within a study were assigned as tied ranks and negative linkage scores were ranked as 0.

Heterogeneity between studies for each bin was assessed using Q statistic, which is defined as the sum of the squared deviations of each study's bin rank from the mean of the ranks. In GSMA, low between-study heterogeneity indicates consistency of study results for the same bin and the presence of low heterogeneity for a specific bin with high ranks can be interpreted as further supportive evidence for the significance of this bin. The statistical significance of the average rank and the Q metric were assessed using a Monte Carlo method (Zintzaras and Ioannidis, 2005a, 2005b). In this method, in a run, the ranks of each study are randomly permuted and the simulated average rank and Q metric are calculated. The procedure is repeated for 50,000 runs and a null distribution for the average rank, and for the Q metric, is constructed. The significance level (P_rank) of the average rank of bins against the null distribution of average ranks is the percentage of simulated average ranks greater or equal than the observed. The statistical significance level (P_Q) for low heterogeneity is the percentage of simulated metrics less or equal than the observed. Moreover, a Monte Carlo test has been performed, generating null distributions separately for each bin, considering only the simulated distributions of the Q metric (Q_adjusted) for bins with the neighboring simulated average rank (±2) as the bin being considered each time (Zintzaras and Ioannidis, 2005a).

In this study, GSMA (P_rank and P_order) and heterogeneity testing were performed unweighted and weighted (Zintzaras and Ioannidis, 2005a, 2005b).

Software and analytic details

Main analysis considered only the genome scans providing results of all bins, while conditional analysis considered one more genome scan, which had evaluated all chromosomes, but results were available only for 22 bins. We also conducted subgroup analyses by type of diabetes and ethnicity. The different ethnicities were classified as European Americans/Europeans/Whites/Caucasians, African Americans, and American Indians.

We performed both unweighted and weighted analyses. In weighted analyses, we weighted bin ranks by √(number of families × number of markers) and then, the weights were scaled to sum up to 1. Weighting ensures that smaller, lower powered studies will have a smaller influence on the overall results. As the optimal weighting for GSMA is unclear, we also performed unweighted analyses. Inferences were performed at the a = 0.05 level. Thus, we considered a p-value ≤0.05 of suggestive significance, and a p-value ≤0.00042 of genome-wide significance (using Bonferroni corrections for 120 comparisons), after adjustment for multiple comparisons. Graph digitizing was performed with Engauge Digitizer (version 2.12, Mark Mitchell, 2002). Genetic distances were obtained using chromosomal maps from the Marshfield Center for Medical Genetics (http://research.marshfieldclinic.org/genetics/GeneticResearch/compMaps.asp). The cytogenetic locations of bins were determined using the Map Viewer and the identification of candidate genes harbored in these regions using the HuGE Navigator and the term “diabetic nephropathies” (last database update on February 2018). The evaluation of the significance of average ranks and the significance of heterogeneity was performed using HEGESMA software (http://biomath.med.uth.gr) (Zintzaras and Ioannidis, 2005a, 2005b).

Results

Literature search of PubMed retrieved 228 articles, out of which 20 were linkage scans for DN or ESRD due to diabetes and 5 articles were retrieved from the references (Fig. 1). Finally, only six articles (nine studies) met the inclusion criteria (Imperatore et al., 1998; Bowden et al., 2004; Osterholm et al., 2007; Rogus et al., 2008; Igo et al., 2011; Wessman et al., 2011). Among the aforementioned scans, nine were partial linkage scans (Yu et al., 1996; Bowden et al., 1997; Freedman et al., 1997; Moczulski et al., 1998; Yu et al., 1999; Freedman et al., 2002; Vardarli et al., 2002; Iyengar, 2003; McDonough et al., 2009), whereas 10 articles referred to surrogate markers for DN as eGFR, albuminuria, and serum cystatin (Covic et al., 2001; Freedman et al., 2005; Krolewski et al., 2006; Placha et al., 2006; Chen et al., 2007; Iyengar et al., 2007; Puppala et al., 2007; Freedman et al., 2008; Schelling et al., 2008; Thameem et al., 2013). The study of Bowden et al. (2004), which presented raw data only for 22 bins, was considered only in conditional analyses. A more detailed justification for the exclusion of the remaining eleven GWLS is presented in Table 1.

FIG. 1.

Flowchart of retrieved studies for meta-analysis and studies excluded, with justification of reasons. MALD, mapping by admixture linkage disequilibrium; RCT, randomized controlled trial; TDT, transmission disequilibrium testing.

Table 1.

Genome-Wide Linkage Scans on Diabetic Nephropathy Retrieved from PubMed

Source (first author, publication year)	Study sample	Reasons for exclusion
Studies included
Imperatore et al. (1998)	Study in Gila River Indian Community	—
Osterholm et al. (2007)	Finnish hospital discharge registry database before 1996 and Finnish diabetes registry	—
Rogus et al. (2008)	Joslin Study on Genetics of Diabetic Nephropathy	—
Igo et al. (2011)	FIND Study African American Group	—
Igo et al. (2011)	FIND Study American Indian Group	—
Igo et al. (2011)	FIND Study European American Group	—
Igo et al. (2011)	FIND Study Mexican American Group	—
Wessman et al. (2011)	(FinnDiane) Study Group, Danish Registry of Diagnosis and Genesis France-Belgium Study	—
Bowden et al. (2004)	African Americans	—
Studies excluded
Thameem et al. (2013)	Mexican American pedigrees	Linkage to variation albuminuria, GFR in diabetes type 2
Covic et al. (2001)	American pedigrees	Subjects with ESRD not only by diabetes type 2
Freedman et al. (2005)	African American pedigrees	Subjects with ESRD not only by diabetes type 2
Krolewski et al. (2006)	Joslin Study on the Genetics of T2DM	Linkage to UAER in diabetes type 2, recruitment regardless of degree of kidney function or albuminuria
Placha et al. (2006)	Joslin Study on the Genetics of T2D	Linkage to variation in renal function in diabetes type 2, recruitment regardless of degree of kidney function or albuminuria
Chen et al. (2007)	African America Diabetes Mellitus Study	Linkage to renal function phenotypes in diabetes type 2, recruitment regardless of degree of kidney function or albuminuria
Iyengar et al. (2007)	FIND Study	Overlapping subjects with the larger study by Igo et al. (2011)
Puppala et al. (2007)	San Antonio Family Diabetes/Gallbladder Study	Linkage to eGFR in diabetes type 2, recruitment regardless of degree of kidney function or albuminuria
Freedman et al. (2008)	Diabetes Heart Study	Linkage to renal function and albuminuria in diabetes type 2, recruitment regardless of degree of kidney function or albuminuria
Schelling et al. (2008)	FIND Study	Overlapping subjects with the larger study by Igo et al. (2011)

Both included and excluded genome searches from meta-analysis and the reason of exclusion are shown (listed from earlier to most recent).

eGFR, estimated GFR; ESRD, end-stage renal disease; GFR, glomerular filtration rate; T2D, type 2 diabetes; T2DM, TD mellitus; UAER, urinary albumin excretion rate.

Details on the analyzed studies are shown in Table 2. All included studies used similar diagnosis criteria for DN. They were conducted in populations with different ethnicities and different type of diabetes. According to Igo et al. (2011), the probands were considered to have T2D, as 90-95% of probands were believed to have T2D. The linkage data of Igo et al. (2011) refer to four different studies. Upon request for linkage data, only Igo et al. (2011) provided the raw data of the four ethnic groups. Regarding the remaining studies, the presented graphs showing linkage scores per distance for all chromosomes were digitized for the calculation of linkage statistic. The unpublished data set was requested from Bowden et al. (2004), but it was not available. All studies analyzed only the autosomes.

Table 2.

Characteristics of the Genome-Wide Linkage Studies Included in the Meta-Analysis

First author (year)	Ethnicity	Type of diabetes	Families, n	Selection criteria	Markers (intermarker interval), n (cM)	Type of analysis/metric (software)	Chromosomal regions corresponding to maximum LOD score
Igo et al. (2011)	African Americans	Type 2	348	Probands with DN with diabetic siblings with or without DN as defined based on DN as a dichotomous trait	4702 (∼0.6)	Multipoint HE (nonparametric)/p-value (SIBPAL)	7p13-7q11.23, 16p12.3-16q12.2, 4q35.1-4q35.2, 12q24.31-12q24.33, 7q11.23-7q22.3
Igo et al. (2011)	American Indians	Type 2	254	Probands with DN with diabetic siblings with or without DN as defined based on DN as a dichotomous trait	3830 (∼0.6)	Multipoint HE (nonparametric)/p-value (SIBPAL)	15q14-15q22.32, 12q12-12q21.2, 10p15.3-10p14, 1q43-1q44, 18p11.32-18p11.22
Igo et al. (2011)	European Americans	Type 2	196	Probands with DN with diabetic siblings with or without DN as defined based on DN as a dichotomous trait	4562 (∼0.6)	Multipoint HE (nonparametric)/p-value (SIBPAL)	8q24.21-8q24.3, 15q11.2-15q14, 4p14-4q13.3, 1q31.1-1q32.3, 1q23.2-1q31.1
Igo et al. (2011)	Mexican Americans	Type 2	479	Probands with DN with diabetic siblings with or without DN as defined based on DN as a dichotomous trait	4546 (∼0.6)	Multipoint HE (nonparametric)/p-value (SIBPAL)	21q22.11-21q22.3, 2q35-2q37.1, 14q23.3-14q32.12, 2q32.1-2q35, 11p15.5-11p15.1
Wessman et al. (2011)	Whites	Type 1	175	Families with one or both parents and siblings concordant or discordant for DN or solely sib pairs concordant or discordant for DN	384 (9.5)	Multipoint ASP NPLpairs analysis/LOD score (MERLIN)	6p23-p22.3, 6p22.3, 6p21, 7q21.11, 22q11
Rogus et al. (2008)	95% Caucasians	Type 1	83	DSP families with DN (persistent proteinuria, ESRD or renal transplant)	5382 (0.62) (deCODE)	Multipoint nonparametric analysis/LOD (SPLAT)	1q (1q42.13), 2q, 3q, 5q, 10q, 17p, 19q, 20p
Osterholm et al. (2007)	Europeans	Type 1	73	DSP families with DN probands (defined based on UAER or ACRs)	900 (∼4)	Multipoint DSP analysis/LOD score (ASPEX, MAPMAKER/SIBS)	3q, 4p, 9q, 16q, 22p
Imperatore et al. (1998)	American Indians	Type 2	59	ASPs for DN	516 (6.4)	Multipoint nonparametric HE/LOD score (SIBPAL)	7q, 3q, 9q, 20p
Bowden et al. (2004)	African Americans	Type 2	166	Families with affected sib pairs concordant for DN	392 (8.9)	Multipoint single locus NPLR/LOD (Genehunter)	7p, 12p, 14q, 16p, 18q, 21q

ACR, albumin:creatinine ratio; ASP, affected sib pairs; DSP, discordant sib pairs; LOD, logarithm of odds; NPL, nonparametric linkage; NR, not reported.

All studies, except that of Igo et al. (2011), implemented the LOD score as a metric of linkage statistics. The study of Igo et al. (2011) regarding Mexican Americans was the most weighted study (w = 0.24), whereas the least weighted study was that of Imperatore et al. (1998) (w = 0.03). The regions with suggestive linkage identified from each scan are shown in Table 2. This meta-analysis provides the results of analysis with both DM types (Fig. 2) and subgroup analyses regarding the type of diabetes with and without the inclusion of the study of Bowden et al. (2004), as well as ethnicity subgroup analyses (Tables 3 and 4).

FIG. 2.

Unweighted (open circles) and weighted (filled circles) average ranks from eight diabetic nephropathy genome scans with 120 bins regarding the analysis with both types of diabetes. Bins with significant P_rank in unweighted or weighted analysis are above the line at p ≤ 0.05.

Table 3.

Diabetic Nephropathy Genome Scan Meta-Analysis Results—Main and Conditional—in Unweighted and (in Parentheses) Weighted Analyses

	Both DM types
		Ranks									Average rank					P for low heterogeneity
Bin	Cytogenetic location	Igo et al. (2011) AA	Igo et al. (2011) AI	Igo et al. (2011) EA	Igo et al. (2011) MA	Imperatore et al. (1998)	Osterholm et al. (2007)	Rogus et al. (2008)	Wessman et al. (2011)	Bowden et al. (2004)	Observed	P_rank	P_rank-condition	P_order	P_order condition	Q	Q condition	Q_adjusted	P_Ha	P_B
4.3	4p14-4q13.3	77	74	118	101	93	50	83	88	49.5	85.50 (89.23)	0.017 (0.020)	0.028 (0.022)	0.157 (0.441)	0.430 (0.516)	0.030 (0.036)	0.059 (0.057)	0.04 (0.048)	0.024 (0.041)	0.034 (0.053)
5.4	5q14.3-5q23.2	95	72	107	98	76	78	119	78	106	90.38 (94.51)	0.006 (0.006)	0.001 (0.004)	0.502 (0.539)	0.189 (0.383)	0.014 (0.014)	0.011 (0.010)	0.049 (0.039)	0.026 (0.023)	0.016 (0.021)
5.5	5q23.2-5q34	103	110	111	105	71.5	47	39.5	32	49.5	77.38 (92.53)	0.080 (0.009)	0.103 (0.012)	0.115 (0.342)	0.133 (0.402)	0.554 (0.500)	0.566 (0.585)	0.549 (0.773)	0.686 (0.674)	0.549 (0.175)
15.1	15q11.2-15p13	84	43	119	107	111	97	39.5	85.5	49.5	85.75 (84.87)	0.016 (0.042)	0.027 (0.048)	0.331 (0.805)	0.636 (0.860)	0.307 (0.399)	0.391 (0.418)	0.470 (0.510)	0.248 (0.354)	0.307 (0.463)
16.2	16p12.3-16q12.2	119	15	61	23	92	86	100	113	117	76.13 (64.99)	0.096 (0.380)	0.034 (0.322)	0.126 (0.514)	0.357 (0.260)	0.870 (0.963)	0.913 (0.981)	0.869 (0.950)	0.819 (0.956)	0.871 (0.966)
22.1	22p13-22q12.3	26	104	62	103	95	117	84	120	116	88.88 (79.45)	0.008 (0.096)	0.002 (0.071)	0.256 (0.874)	0.019 (0.958)	0.393 (0.628)	0.409 (0.671)	0.681 (0.682)	0.297 (0.488)	0.347 (0.591)
22.2	22q12.3-22q13.33	14	81	86	113	91	110	82	91	49.5	83.50 (77.47)	0.027 (0.122)	0.040 (0.127)	0.211 (0.776)	0.338 (0.856)	0.342 (0.739)	0.401 (0.724)	0.440 (0.766)	0.088 (0.372)	0.167 (0.664)
T1DM
1.10	1q43-1q44	—	—	—	—	—	63	114	106	—	94.33 (19.82)	0.042 (0.025)	na	0.38 (0.595)	na	0.419 (0.423)	na	0.858 (0.612)	0.437 (0.449)	0.411 (0.392)
3.6	3q21.2-3q25.32	—	—	—	—	—	119	97.5	114	—	110.17 (20.71)	0.003 (0.012)	na	0.3 (0.411)	na	0.181 (0.215)	na	0.764 (0.338)	0.183 (0.221)	0.179 (0.21)
5.3	5q11.2-5q14.3	—	—	—	—	—	84	112	67	—	87.67 (18.82)	0.076 (0.05)	na	0.430 (0.21)	na	0.332 (0.408)	na	0.289 (0.464)	0.321 (0.412)	0.348 (0.425)
5.4	5q14.3-5q23.2	—	—	—	—	—	78	119	78	—	91.67 (19.81)	0.054 (0.025)	na	0.278 (0.348)	na	0.346 (0.445)	na	0.543 (0.629)	0.368 (0.472)	0.311 (0.436)
16.2	16p12.3-16q12.2	—	—	—	—	—	86	100	113	—	99.67 (19.55)	0.024 (0.03)	na	0.296 (0.125)	na	0.203 (0.189)	na	0.367 (0.163)	0.192 (0.047)	0.21 (0.2)
17.4	17q24.3-17q25.3	—	—	—	—	—	80	106	108	—	98.00 (19.74)	0.029 (0.025)	na	0.242 (0.183)	na	0.229 (0.22)	na	0.471 (0.267)	0.235 (0.232)	0.218 (0.202)
19.4	19q13.33-19q13.43	—	—	—	—	—	96	120	111	—	109.00 (22.10)	0.005 (0.002)	na	0.084 (0.178)	na	0.19 (0.207)	na	0.761 (0.502)	0.189 (0.215)	0.196 (0.204)
22.1	22p13-22q12.3	—	—	—	—	—	117	84	120	—	107.00 (19.39)	0.006 (0.033)	na	0.039 (0.078)	na	0.283 (0.362)	na	0.964 (0.494)	0.298 (0.385)	0.273 (0.357)
22.2	22q12.3-22q13.33	—	—	—	—	—	110	82	91	—	94.33 (17.63)	0.041 (0.097)	na	0.209 (0.682)	na	0.211 (0.228)	na	0.324 (0.159)	0.211 (0.242)	0.219 (0.225)
T2DM
4.3	4p14-4q13.3	77	74	118	101	93	—	—	—	49.5	92.60 (74.16)	0.017 (0.030)	0.033 (0.033)	0.345 (0.731)	0.341 (0.794)	0.067 (0.087)	0.172 (0.121)	0.215 (0.168)	0.070 (0.122)	0.077 (0.126)
5.4	5q14.3-5q23.2	95	72	107	98	76	—	—	—	106	89.60 (74.70)	0.029 (0.027)	0.008 (0.018)	0.255 (0.867)	0.659 (0.676)	0.035 (0.031)	0.022 (0.016)	0.072 (0.067)	0.051 (0.044)	0.039 (0.038)
5.5	5q23.2-5q34	103	110	111	105	71.5	—	—	—	49.5	100.10 (84.80)	0.004 (0.002)	0.010 (0.002)	0.360 (0.202)	0.331 (0.248)	0.049 (0.010)	0.215 (0.089)	0.438 (0.126)	0.039 (0.010)	0.032 (0.002)
7.1	7p22.3-7p15.3	80	107	99	28	86	—	—	—	120	80.00 (59.19)	0.107 (0.229)	0.027 (0.178)	0.888 (0.903)	0.628 (0.908)	0.391 (0.605)	0.505 (0.684)	0.408 (0.561)	0.271 (0.479)	0.344 (0.621)
7.5	7q22.3-7q34	107	31	102	99	120	—	—	—	49.5	91.80 (71.28)	0.020 (0.050)	0.036 (0.053)	0.207 (0.763)	0.253 (0.827)	0.563 (0.495)	0.676 (0.504)	0.841 (0.723)	0.427 (0.328)	0.404 (0.357)
15.1	15q11.2-15p13	84	43	119	107	111	—	—	—	49.5	92.80 (72.73)	0.016 (0.039)	0.031 (0.042)	0.619 (0.727)	0.539 (0.793)	0.381 (0.395)	0.529 (0.414)	0.717 (0.667)	0.386 (0.366)	0.355 (0.431)
22.1	22p13-22q12.3	26	104	62	103	95	—	—	—	116	78.00 (60.05)	0.132 (0.211)	0.039 (0.169)	0.881 (0.942)	0.158 (0.986)	0.509 (0.635)	0.596 (0.689)	0.526 (0.612)	0.576 (0.682)	0.481 (0.648)

AA, African Americans; AI, American Indians; EA, European Americans; MA, Mexican Americans; na, non-applicable.

Table 4.

Diabetic Nephropathy Genome Scan Ethnicity Subgroup Meta-Analysis Results in Unweighted and (in Parentheses) Weighted Analyses

Bin	Cytogenetic location	Igo et al. (2011) EA	Osterholm et al. (2007)	Rogus et al. (2008)	Wessman et al. (2011)	Observed	P_rank	P_order	Q	Q_adjusted	P_Ha	P_B
Caucasians
1.7	1q23.2-1q31.1	116	105	89	32	85.50 (95.98)	0.068 (0.031)	0.170 (0.316)	0.703 (0.578)	0.745 (0.602)	0.617 (0.510)	0.689 (0.520)
1.10	1q43-1q44	94	63	114	106	94.25 (98.00)	0.022 (0.023)	0.099 (0.292)	0.217 (0.162)	0.522 (0.297)	0.197 (0.114)	0.220 (0.184)
3.6	3q21.2-3q25.32	54	119	97.5	114	96.13 (82.78)	0.017 (0.125	0.593 (0.437	0.405 0.630)	0.816 (0.486)	0.363 (0.531)	0.391 (0.630)
4.3	4p14-4q13.3	118	50	83	88	84.75 (95.18)	0.077 (0.035)	0.078 (0.097)	0.354 (0.444)	0.221 (0.431)	0.267 (0.369)	0.383 (0.470)
5.3	5q11.2-5q14.3	103	84	112	67	91.50 (99.16)	0.032 (0.019)	0.068 (0.410)	0.172 (0.172)	0.288 (0.342)	0.201 (0.218)	0.187 (0.149)
5.4	5q14.3-5q23.2	107	78	119	78	95.50 (103.75)	0.019 (0.008)	0.1562 (0.221)	0.187 (0.198)	0.511 (0.541)	0.241 (0.261)	0.186 (0.184)
11.4	11q13.3-11q22.1	106	109	99	32	86.50 (95.16)	0.060 (0.034)	0.165 (0.036)	0.674 (0.447)	0.680 (0.434)	0.643 (0.449)	0.524 (0.257)
16.2	16p12.3-16q12.2	61	86	100	113	90.00 (82.58)	0.038 (0.127)	0.027 (0.209)	0.212 (0.312)	0.310 (0.160)	0.215 0.350)	0.233 (0.387)
17.2	17p12-17q21.2	110	22	104	93	82.25 (95.46)	0.103 (0.034)	0.088 (0.195)	0.830 (0.663)	0.848 (0.736)	0.770 (0.661)	0.733 (0.390)
17.4	17q24.3-17q25.3	72	80	106	108	91.50 (88.02)	0.032 (0.077)	0.023 (0.185)	0.144 (0.180)	0.223 (0.144)	0.190 (0.252)	0.144 (0.220)
19.4	19q13.33-19q13.43	109	96	120	111	109.00 (111.13)	0.00102 (0.001)	0.107 (0.088)	0.041 (0.033)	0.537 (0.341)	0.026 (0.018)	0.040 (0.038)
22.1	22p13-22q12.3	62	117	84	120	95.75 (82.59)	0.017 (0.130)	0.339 (0.330)	0.350 (0.47668	0.730 (0.311)	0.435 (0.536)	0.371 (0.449)
22.2	22q12.3-22q13.33	86	110	82	91	92.25 (88.24)	0.028 (0.077)	0.109 (0.286)	0.068 (0.054)	0.091 (0.039)	0.063 (0.055)	0.061 (0.036)

Bin	Cytogenetic location	Igo et al. (2011) AA	Bowden et al. (2004)	Observed	P_rank	P_order	Q	Q_adjusted	P_Ha	P_B
African Americans
1.3	1p35.2-1p32.1	87	115	101.00 (91.66)	0.043 (0.186)	0.828 (0.175)	0.438 (0.442)	0.763 (0.136)	0.438 (0.442)	0.438 (0.442)
4.8	4q35.1-4q35.2	118	49.5	83.75 (106.61)	0.117 (0.042)	0.982 (1.000)	0.925 (0.923)	0.770 (0.670)	0.925 (0.923)	0.925 (0.923)
5.4	5q14.3-5q23.2	95	106	100.50 (96.83)	0.046 0.135)	0.682 (0.393)	0.182 (0.181)	0.329 (0.063)	0.182 (0.181)	0.182 (0.181)
7.1	7p22.3-7p15.3	80	120	100.00 (86.65)	0.046 (0.235)	0.511 (0.410)	0.625 (0.624)	0.970 (0.171)	0.625 (0.624)	0.625 (0.624)
7.3	7p13-7q11.23	120	49.5	84.75 (108.28)	0.101 (0.026)	0.947 (0.913)	0.942 (0.940)	0.985 (1.000)	0.942 (0.940)	0.942 (0.940)
12.2	12p13.2-12q12	99	118	108.50 (102.16)	0.021 (0.083)	0.751 (0.918)	0.303 (0.304)	0.857 (0.161)	0.303 0.304	0.303 (0.304)
13.3	13q22.1-13q33.1	96	107	101.50 (97.83)	0.043 (0.128)	0.929 (0.689)	0.182 (0.180)	0.378 (0.085)	0.182 (0.180)	0.182 (0.180)
16.2	16p12.3-16q12.2	119	117	118.00 (118.67)	0.001 (0.001)	0.123 (0.106)	0.0349 (0.034)	0.478 (0.303)	0.035 (0.034)	0.035 (0.034)

Bin	Cytogenetic location	Igo et al. (2011) AI	Imperatore et al. (1998)	Observed	P_rank	P_order	Q	Q_adjusted	P_Ha	P_B
American Indians
1.1	1p36.32-1p36.21	114	74.5	94.25 (108.06)	0.098 (0.039)	0.894 (0.998)	0.571 (0.570)	0.749 (0.659)	0.571 (0.570)	0.571 (0.570)
15.2	15q14-15q22.32	120	32	76.00 (106.77)	0.246 (0.047)	0.954 (0.998	0.967 (0.967)	0.993 (1.000)	0.967 (0.967)	0.967 (0.967)
20.1	20p13-20p12.2	86	119	102.50 (90.96)	0.044 (0.203)	0.999 (0.989)	0.505 (0.508)	0.901 (0.503)	0.505 (0.508)	0.505 (0.508)

Main analysis

In all analyses regarding the eight studies, 13 different bins were found significant (P_rank ≤ 0.05) with both DM types or subgroup analyses, weighted or unweighted analyses. More specifically, in analysis with both DM types, six cytogenetic locations (4p14-4q13.3, 5q14.3-5q23.2, 5q23.2-5q34, 15p13-15q11.2, 22p13-22q12.3, and 22q12.3-22q13.33) were significant in weighted or unweighted analyses (P_rank ≤ 0.05), out of which bins 4p14-4q13.3, 5q14.3-5q23.2, and 15p13-15q11.2 were significant in both weighed and unweighted analyses (P_rank ≤ 0.05). In subgroup analysis regarding T1DM type, nine cytogenetic locations (1q43-1q44, 3q21.2-3q25.32, 5q11.2-5q14.3, 5q14.3-5q23.2, 16p12.3-16q12.2, 17q24.3-17q25.3, 19q13.33-19q13.43, 22p13-22q12.3, and 22q12.3-22q13.3) were significant in weighted or unweighted analyses (P_rank ≤ 0.05), out of which bins on chromosomes 1q, 3q, 16p, 17q, 19q, and 22p were significant in both weighed and unweighted analyses (P_rank ≤ 0.05). In subgroup analysis regarding type 2 diabetes mellitus (T2DM), five cytogenetic locations (4p14-4q13.3, 5q14.3-5q23.2, 5q23.2-5q34, 7q22.3-7q34, and 15p13-15q11.2) were significant in both weighed and unweighted analyses (P_rank ≤ 0.05). Cytogenetic location 7q22.3-7q34 is identified significant only in T2DM analysis. However, no bin reached the significance threshold after adjustment for multiple comparisons in either analysis with both DM types or subgroup analyses (P_rank > 0.00042).

Regarding ordered analyses, in analyses with both DM types, no bin was significant, whereas in T1DM subgroup analysis, only 22p13-22q12.3 was significant (P_order < 0.05). In T2DM subgroup analysis, no bin was significant (P_order < 0.05) (Table 3).

In low heterogeneity testing, chromosomal regions 4p14-4q13.3 and 5q14.3-5q23.2 were statistically significant with left-sided P_Q/Ha/B < 0.05 (Table 3). In analyses with both DM types, 16p12.3-16q12.2 was significant with P_Ha < 0.05 in T1DM subgroup analyses and 5q14.3-5q23.2 and 5q23.2-5q34 in T2DM subgroup analyses with P_Q/Ha/B < 0.05, respectively. Chromosomal region 5q14.3-5q23.2 in analysis with both DM types and regions 5q14.3-5q23.2 and 5q23.2-5q34 were significant in both weighted and unweighted analyses regarding all metrics of heterogeneity (P_Q/Ha/B < 0.05).

Regarding bins repeatedly identified as significant across analyses with both DM types and subgroup analyses, only chromosomal location 5q14.3-5q23.2 was common across all analyses. However, between analysis with both DM types and T1DM subgroup analyses, chromosomal locations 5q14.3-5q23.2, 22p13-22q12.3, and 22q12.3-22q13.3 were common.

Conditional analysis

Meta-analysis of nine genome-wide scans identified nine different cytogenetic locations (4p14-4q13.3, 5q14.3-5q23.2, 5q23.2-5q34, 7p22.3-7p15.3, 7q22.3-7q34, 15p13-15q11.2, 16p12.3-16q12.2, 22p13-22q12.3, and 22q12.3-22q13.33). Two cytogenetic locations, 7p22.3-7p15.3 in T2DM analysis and 16p12.3-16q12.2 in both DM type analyses, were significant only in relevant conditional analyses.

Regarding ordered analyses, only cytogenetic location 5q14.3-5q23.2 in analysis with both DM types showed statistical significance (P_order < 0.05). Similarly, in low heterogeneity testing, this region was statistically significant in both weighted and unweighted analyses (P_Q < 0.05).

The genome scan meta-analysis and heterogeneity testing in DN revealed the significance of four novel locations, 5q11.2-5q14.3, 5q23.2-5q34, 17q24.3-17q25.3, and 22q12.3-22q13.3, and replicated the significance of the remaining cytogenetic locations, whereas conditional analysis showed one more cytogenetic region 7p22.3-7p15.3, indicative of linkage with DN.

Regarding the three ethnicity subgroup analyses, in total, 11 new cytogenetic locations were identified significant (P_rank ≤ 0.05). More specifically, in Caucasians, 13 cytogenetic locations were identified significant, 3 out of which (1q23.2-1q31.1, 11q13.3-11q22.1, and 17p12-17q21.2) were only in Caucasians. In African Americans, eight cytogenetic locations are identified significant, five out of which (1p35.2-1p32.1, 4q35.1-4q35.2, 7p13-7q11.23, 12p13.2-12q12, and 13q22.1-13q33.1) were only in African Americans. Last, but not least, in American Indians, all regions (1p36.32-1p36.21, 15q14-15q22.32, and 20p13-20p12.2) identified are significant only in American Indians.

Regarding ordered analyses based on ethnicity, only cytogenetic locations 16p12.3-16q12.2 and 17q24.3-17q25.3 in Caucasians showed statistical significance (P_order < 0.05). Similarly, in low heterogeneity testing, only regions 19q13.33-19q13.43 and 22q12.3-22q13.33 in Caucasians, as well as 16p12.3-16q12.2 in African Americans were statistical significant (P_Q/Ha/B < 0.05).

The subgroup analyses based on ethnicity revealed some additional novel cytogenetic locations (1p36.32-1p36.21, 1p35.2-1p32.1, 11q13.3-11q22.1, 13q22.1-13q33.1, and 17p12-17q21.2).

Discussion

The meta-analysis of 8 genome scans identified a total of 13 chromosomal regions with evidence for linkage from P_rank. Remarkably, 11 chromosomal regions were significant in both weighted and unweighted analyses, analysis with both DM types, or subgroup analysis based on type of diabetes. However, no bin revealed significance at genome-wide level (P_rank ≤ 0.00042). For this reason, we focused on bins that reached statistical significance (P_rank < 0.05) and were novel findings, or were common among different analyses. For purposes of completeness, meta-analysis was also conducted after the inclusion of the study of Bowden et al. (2004), as it was clearly eligible, but it presented results only for 22 bins. Furthermore, it has been shown that when results are not available for all bins, there is no increase in false-positive results (Forabosco et al., 2007). The conditional analyses were post hoc analyses that should be viewed as exploratory. Novel findings of this study constitute the chromosomal regions that were identified as significant by the meta-analysis, but not by the individual genome scans, as cytogenetic locations 5q11.2-5q14.3, 5q23.2-5q34, 17q24.3-17q25.3, and 22q12.3-22q13.3. Only chromosomal region 22p13-22q12.33 indicated evidence for linkage from P_order statistic.

Heterogeneity testing revealed low unadjusted heterogeneity at chromosomal regions 4p14-4q13.3 and 5q14.3-5q23.2 in analysis with both DM types, as well as regions 16p12.3-16q12.2 in T1DM and 5q14.3-5q23.2 and 5q23.2-5q34 in T2DM analysis, respectively. All of the aforementioned bins were significant in both weighted and unweighted analysis, except bin on chromosome 16p.

Chromosomal region 5q14.3-5q23.2 was the common finding between analysis with both DM types and subgroup analysis and constitutes the most significant region for DN in analysis with both DM types. However, among included studies, only Rogus et al. (2008) had shown evidence of linkage within this chromosomal region, whereas another genome-wide linkage study, which was excluded from the meta-analysis due to improper inclusion criteria, had detected evidence of linkage to serum creatinine in 5q23 (Chen et al., 2007). Chromosome 5q, as well as 7q and 22q, account for the majority of genetic variance of urinary albumin excretion in both diabetic and nondiabetic individuals (Krolewski et al., 2006). PCSK1 is the only gene located in this chromosomal region that has been examined in genetic association studies regarding DN (Jacobsen et al., 2002; De Cosmo et al., 2002).

The most significant region in T2DM-DN is the cytogenetic location 5q23.2-5q34, which constitutes a novel finding. Locations adjacent to this region have shown strong or suggestive evidence of linkage with DN. Rogus et al. (2008) detected evidence for linkage with DN in 5q at 118 cM. Also, Krolewski et al. (2006) found strong evidence for linkage to urinary albumin excretion on chromosome 5 at 69 cM (LOD = 3.4) both in diabetic and nondiabetic individuals. IL4, GPX3, SMAD5, IL12B, IL13, GALNT10, and MFAP3 are located in this region and have been previously examined in genetic association studies (Arababadi, 2010; Vieira et al., 2011; McKnight et al., 2009; Grzegorzewska et al., 2014; Ng et al., 2012).

Another novel cytogenetic location on chromosome 5, 5q11.2-5q14.3, showed marginal significance in subgroup analysis of T1DM (P_rank = 0.05). ITGA2 is the only gene located in this region, which has been examined in genetic association studies (Tsai et al., 2001). ITGA2 encodes the alpha subunit of one major platelet membrane receptor for collagen and related proteins exposed to subendothelial and subsequently, it is assumed to be implicated in the accumulation of extracellular matrix (ECM), one of the hallmarks of DN.

Cytogenetic location 17q24.3-17q25.3 is one more novel region detected from meta-analysis. Among the excluded GWLS, only Chen et al. (2007) identified suggestive linkage at 17q24, with CrCl and GFR with LOD score 1.36 and 1.04, respectively. FN3K is the only gene already examined in genetic association studies in this region (Tanhauserova et al., 2014). Advanced glycation end products (AGEs) have great impact on the development and progression of DN, and FN3K is assumed to play an important role in prevention of further glycation processes (Delpierrre et al., 2004; Ott et al., 2014).

Across the cytogenetic region 22q12.3-22q13.33, none of the included studies identified evidence of linkage with DN. Among candidate genes in this region are APOL1, APOL2, APOL3, MYH9, HMOX1, and MIOX. Although APOL1 explains a substantial proportion of nondiabetic end-stage kidney disease in African Americans, less is known about its contribution in DN (Freedman et al., 2010). Polymorphisms in MYH9 also exhibit strong association with nondiabetic kidney disease, but Genovese et al. (2010) found that the statistical association of APOL1 was stronger compared with MYH9 (Freedman et al., 2010).

The most important regions for T1DM-DN are on chromosomes 3q, 19q, and 22p in weighted or unweighted analyses. Chromosome 3q has been identified as a major locus for DN susceptibility genes from many different studies, either linkage scans or association studies. This meta-analysis replicates the significance of chromosomal region 3q21.2-3q25.32, as individual linkage scans had identified evidence of linkage in different locations across the whole region (Imperatore et al., 1998; Osterholm et al., 2007; Rogus et al., 2008; Moczulski et al., 1998; Placha et al., 2006; Krolewski et al., 2006). However, two other linkage scans identified evidence of linkage at neighboring regions (Bowden et al., 2004; Chen et al., 2007). Among candidate genes in this region are MRAS, NPHP3, ACAD11, TRPC1, and AGTR1, which is one of the most studied genes.

This meta-analysis is the first to search for chromosomal regions harboring genes responsible for DN, considering DN as a discreet trait. Therefore, linkage studies specifically for proteinuria or other quantitative surrogate markers of renal function were excluded. The most relevant meta-analysis to our study was the one by Rao et al. (2011), who combined data of GWLS on renal function traits with the difference that almost half of the studies were conducted in general population, whereas our study was conducted only in diseased population.

GWLS about quantitative kidney phenotypes were not included in this GSMA for many reasons. First, although proteinuria and impaired kidney function are major phenotypes of DN that aggregate in families of diabetics and nondiabetics as heritable traits, they are not genetically correlated even in nondiabetics (Placha et al., 2005). This could bias the results. Second, the consideration of DN as a discreet trait provides more consistent inclusion criteria for the meta-analysis, reducing the heterogeneity between studies, as GFR and albumin:creatinine ratio (ACR) (urinary albumin excretion rate) present high variability in their calculation and measurement. The eGFR values depend on the estimating equations and many factors, whereas in single ACR measurements, the results are not reliable. Both traits may not reveal the initial stages of DN, as GFR is high in the hyperfiltration stage, whereas microalbuminuria can be reversible (Glassok et al., 2017). Antihypertensive treatment could also minimize the variability of both phenotypes, undermining the genetic influences. It has been postulated that the development of DN consists of two distinct processes, one associated with glomerula and proteinuria and another with tubules, interstitium, and kidney function (Placha et al., 2005). Therefore, this integrating approach of this meta-analysis could reveal genes and biological processes common to both processes. More specifically, the novel and significant cytogenetic locations identified from this meta-analysis include genes involved in insulin resistance, immune system, transforming growth factor beta 1 and bone morphogenetic protein signaling pathways, oxidative stress, protein glycosylation and AGEs, EC accumulation, and lipid metabolism, as well as cytokinesis and angiogenesis.

The identified locations derived from analyses, including both DM types, could harbor susceptibility genes either of DN or diabetes (T1D or T2D). Similarly, statistically significant locations in subgroup analyses could reflect effects of either susceptibility genes of DN per se or specific genetic effects of the type of diabetes under investigation. However, any common genetic region between T1D or T2D subgroup analyses is more likely to harbor susceptibility genes of DN, not affected by the type of DM. These genes could constitute the common genetic architecture of all types of DN and the future therapeutic targets. More specifically, genes located in region 5q23.2-5q34 and in chromosome 22, as well as genes in linkage disequilibrium with them, could constitute the genetic determinants of DN regardless of the type of diabetes. It is noteworthy to mention that cytogenetic region 7p22.3-7p15.3, identified significant only in conditional analysis of T2DM, could be related with genetic loci that are specific for African Americans with T2DM, as the population sample in the study of Bowden et al. (2004) was from this ethnic group.

Similarly, ethnicity subgroup analyses could reveal susceptibility loci specific for different ethnicities. More specifically, cytogenetic locations 1q23.2-1q31.1, 11q13.3-11q22.1, and 17p12-17q21.2 are identified significant only in Caucasians and locations 1p35.2-1p32.1, 4q35.1-4q35.2, 7p13-7q11.23, 12p13.2-12q12, and 13q22.1-13q33.1 only in African Americans, while locations 1p36.32-1p36.21, 15q14-15q22.32 and 20p13-20p12.2 only in American Indians. Among the aforementioned locations, novel findings constitute the locations 11q13.3-11q22.1 and 17p12-17q21.2 in Caucasians, locations 1p35.2-1p32.1 and 13q22.1-13q33.1 in African Americans, as well as location 1p36.32-1p36.21 in American Indians. Cytogenetic locations 5q14.3-5q23.2 and 22p13-22q12.3 are identified in all analyses regarding diabetes type and ethnicity, except in American Indian subgroup analysis. Last, but not least, nine regions identified in T1DM analysis and three regions identified in T2DM analysis are also identified in Caucasians, suggesting that Caucasians share more common regions with T1DM rather than T2DM genetic architecture.

Although HEGESMA is based on bins with width of 30 cM, an analysis with bin width <30 cM could provide some more information. HEGESMA currently considers only autosomes. Therefore, no conclusion about possible linkage on X and Y chromosome can be derived. Nevertheless, the fact the HEGESMA calculates the heterogeneity between the included genome scans makes the results more robust.

Conclusions

In conclusion, given the modest results from GWASs on DN, the growing interest in rare variants, and a resurgence of family-based studies, this genome scan meta-analysis and heterogeneity testing in DN replicate the significance of several cytogenetic locations and especially emphasize the significance of locations 5q14.3-5q23.2 and 22p13-22q12.3, which are identified in all analyses regarding diabetes type and ethnicity, except in American Indian subgroup analysis. In addition, this meta-analysis revealed significance of nine novel locations, 1p35.2-1p32.1, 1p36.32-1p36.21, 5q11.2-5q14.3, 5q23.2-5q34, 11q13.3-11q22.1, 13q22.1-13q33.1, 17p12-17q21.2, 17q24.3-17q25.3, and 22q12.3-22q13.3. The conditional analysis revealed one more region, 7p22.3-7p15.3, indicative of linkage with DN. Although the meta-analysis provides some of the strongest evidence of linkage compared to the results of the individual genome scans, especially for cytogenetic locations statistically significant in more than one analyses, we should interpret the results with caution, as it is likely that some important regions have been missed, while some others may result from type I errors, given the aforementioned limitations of the method. However, this genome scan meta-analysis provides some evidence of linkage for several regions to prioritize follow-up studies in candidate locations.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Arababadi

(2010) Interleukin-4 gene polymorphisms in type 2 diabetic patients with nephropathy. Iran J Kidney Dis, 4:302.

Borch-Johnsen

, Norgaard

, Hommel

, et al. (1992) Is diabetic nephropathy an inherited complication?. Kidney Int, 41:719-722.

Bouzigon

, Forabosco

, Koppelman

, et al. (2010) Meta-analysis of 20 genome-wide linkage studies evidenced new regions linked to asthma and atopy. Eur J Hum Genet, 18:700.

Bowden

, Colicigno

, Langefeld

, et al. (2004) A genome scan for diabetic nephropathy in African Americans. Kidney Int, 66:1517-1526.

Bowden

, Sale

, Howard

, et al. (1997) Linkage of genetic markers on human chromosomes 20 and 12 to NIDDM in Caucasian sib pairs with a history of diabetic nephropathy. Diabetes, 46:882-886.

Chen

, Adeyemo

, Zhou

, et al. (2007) A genome-wide search for linkage to renal function phenotypes in West Africans with type 2 diabetes. Am J Kidney Dis, 49:394-400.

Covic

, Iyengar

, Olson

, et al. (2001) A family-based strategy to identify genes for diabetic nephropathy. Am J Kidney Dis, 37:638-647.

De Cosmo

, Miscio

, Zucaro

, et al. (2002) The role of PC‐1 and ACE genes in diabetic nephropathy in type 1 diabetic patients: evidence for a polygenic control of kidney disease progression. Nephrol Dial Transplant, 17:1402-1407.

Delpierrre

, Vertommen

, Communi

, et al. (2004) Identification of fructosamine residues deglycated by fructosamine-3-kinase in human hemoglobin. J Biol Chem, 279:27613-27620.

10.

Dronavalli

, Duka

, Bakris

(2008) The pathogenesis of diabetic nephropathy. Nat Rev Endocrinol, 4:444.

11.

Eknoyan

, Hostetter

, Bakris

, et al. (2003) Proteinuria and other markers of chronic kidney disease: a position statement of the national kidney foundation (NKF) and the national institute of diabetes and digestive and kidney diseases (NIDDK). Am J Kidney Dis, 42:617-622.

12.

Forabosco

, Ng

, Bouzigon

(2007) Data acquisition for meta-analysis of genome-wide linkage studies using the genome search meta-analysis method. Hum Hered, 64:74-81.

13.

Freedman

, Bowden

, Rich

, et al. (2005) A genome scan for all-cause end-stage renal disease in African Americans. Nephrol Dial Transplant, 20:712-718.

14.

Freedman

, Bowden

, Rich

, et al. (2008) Genome-wide linkage scans for renal function and albuminuria in type 2 diabetes mellitus: the Diabetes Heart Study. Diabet Med, 25:268-276.

15.

Freedman

, Kopp

, Langefeld

, et al. (2010) The apolipoprotein L1 (APOL1) gene and nondiabetic nephropathy in African Americans. J Am Soc Nephrol, 21:1422-1426.

16.

Freedman

, Rich

, Yu

, et al. (2002) Linkage heterogeneity of end-stage renal disease on human chromosome 10. Kidney Int, 62:770-774.

17.

Freedman

, Yu

, Spray

, et al. (1997) Genetic linkage analysis of growth factor loci and end-stage renal disease in African Americans. Kidney Int, 51:819-825.

18.

Genovese

, Friedman

, Ross

, et al. (2010) Association of trypanolytic ApoL1 variants with kidney disease in African Americans. Science, 329:841-845.

19.

Glassock

, Warnock

, Delanaye

(2017) The global burden of chronic kidney disease: estimates, variability and pitfalls. Nat Rev Nephrol, 13:104-114.

20.

Grzegorzewska

, Ostromecki

, Zielińska

, et al. (2014) T-cell cytokine gene polymorphisms and vitamin D pathway gene polymorphisms in end-stage renal disease due to type 2 diabetes mellitus nephropathy: comparisons with health status and other main causes of end-stage renal disease. J Diabetes Res, 2014:120317.

21.

Igo

, Iyengar

, Nicholas

, et al. (2011) Genomewide linkage scan for diabetic renal failure and albuminuria: the FIND study. Am J Nephrol, 33:381-389.

22.

Imperatore

, Hanson

, Pettitt

, et al. (1998) Sib-pair linkage analysis for susceptibility genes for microvascular complications among Pima Indians with type 2 diabetes. Pima Diabetes Genes Group. Diabetes, 47:821-830.

23.

Iyengar

(2003) Linkage analysis of candidate loci for end-stage renal disease due to diabetic nephropathy. J Am Soc Nephrol, 14:195S-201S.

24.

Iyengar

, Abboud

, Goddard

, et al. (2007) Genome-wide scans for diabetic nephropathy and albuminuria in multiethnic populations: the family investigation of nephropathy and diabetes (FIND). Diabetes, 56:1577-1585.

25.

Jacobsen

, Grarup

, Tarnow

, et al. (2002) PC‐1 amino acid variant (K121Q) has no impact on progression of diabetic nephropathy in type 1 diabetic patients. Nephrol Dial Transplant, 17:1408-1412.

26.

Kato

, Natarajan

(2014) Diabetic nephropathy—emerging epigenetic mechanisms. Nat Rev Nephrol, 10:517-530.

27.

Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group (2013) Chapter 1: definition and classification of CKD. Kidney Int Suppl, 3:19-62.

28.

Krolewski

, Poznik

, Placha

, et al. (2006) A genome-wide linkage scan for genes controlling variation in urinary albumin excretion in type II diabetes. Kidney Int, 69:129-136.

29.

Lander

, Kruglyak

(1995) Genetic dissection of complex traits: guidelines for interpreting and reporting linkage results. Nat Genet, 11:241-247.

30.

McDonough

, Bostrom

, Lu

, et al. (2009) Genetic analysis of diabetic nephropathy on chromosome 18 in African Americans: linkage analysis and dense SNP mapping. Hum Genet, 126:805-817.

31.

McKnight

, Woodman

, Parkkonen

, et al. (2009) Investigation of DNA polymorphisms in SMAD genes for genetic predisposition to diabetic nephropathy in patients with type 1 diabetes mellitus. Diabetologia, 52:844-849.

32.

Moczulski

, Rogus

, Antonellis

, et al. (1998) Major susceptibility locus for nephropathy in type 1 diabetes on chromosome 3q: results of novel discordant sib-pair analysis. Diabetes, 47:1164-1169.

33.

Mooyaart

, Valk

, van Es

, et al. (2011) Genetic associations in diabetic nephropathy: a meta-analysis. Diabetologia, 54:544-553.

34.

MCY

, Hester

, Wing

, et al. (2012) Genome-wide association of BMI in African Americans. Obesity (Silver Spring, MD), 20:622-627.

35.

Osterholm

, He

, Pitkaniemi

, et al. (2007) Genome-wide scan for type 1 diabetic nephropathy in the Finnish population reveals suggestive linkage to a single locus on chromosome 3q. Kidney Int, 71:140-145.

36.

Ott

, Jacobs

, Haucke

, et al. (2014) Role of advanced glycation end products in cellular signaling. Redox Biol, 2:411-429.

37.

Palmer

, Freedman

(2012) Insights into the genetic architecture of diabetic nephropathy. Curr Diab Rep, 12:423-431.

38.

Placha

, Canani

, Warram

, et al. (2005) Evidence for different susceptibility genes for proteinuria and ESRD in type 2 diabetes. Adv Chronic Kidney Dis, 12:155-169.

39.

Placha

, Poznik

, Dunn

, et al. (2006) A genome-wide linkage scan for genes controlling variation in renal function estimated by serum cystatin C levels in extended families with type 2 diabetes. Diabetes, 55:3358-3365.

40.

Puppala

, Arya

, Thameem

, et al. (2007) Genotype by diabetes interaction effects on the detection of linkage of glomerular filtration rate to a region on chromosome 2q in Mexican Americans. Diabetes, 56:2818-2828.

41.

Quinn

, Angelico

, Warram

, et al. (1996) Familial factors determine the development of diabetic nephropathy in patients with IDDM. Diabetologia, 39:940-945.

42.

Rao

, Mottl

, Cole

, et al. (2011) Meta-analysis of genome-wide linkage scans for renal function traits. Nephrol Dial Transplant, 27:647-656.

43.

Rich

(2006) Genetics of diabetes and its complications. J Am Soc Nephrol, 17:353-360.

44.

Rogus

, Poznik

, Pezzolesi

, et al. (2008) High-density single nucleotide polymorphism genome-wide linkage scan for susceptibility genes for diabetic nephropathy in type 1 diabetes discordant sibpair approach. Diabetes, 57:2519-2526.

45.

Saran

, Robinson

, Abbott

, et al. (2017) US Renal Data System 2016 Annual Data Report: epidemiology of kidney disease in the United States. Am J Kidney Dis, 69:S1-S688.

46.

Schelling

, Abboud

, Nicholas

, et al. (2008) Genome-wide scan for estimated glomerular filtration rate in multi-ethnic diabetic populations: the Family Investigation of Nephropathy and Diabetes (FIND). Diabetes, 57:235-243.

47.

Seaquist

, Goetz

, Rich

, et al. (1989) Familial clustering of diabetic kidney disease. Evidence for genetic susceptibility to diabetic nephropathy. N Engl J Med, 320:1161-1165.

48.

Stefanidis

, Kreuer

, Dardiotis

, et al. (2014) Association between the interleukin-1β Gene (IL1B) C-511T polymorphism and the risk of diabetic nephropathy in type 2 diabetes: a candidate-gene association study. DNA Cell Biol, 33:463-468.

49.

Stefanidis

, Kytoudis

, Papathanasiou

, et al. (2009) XbaI GLUT1 gene polymorphism and the risk of type 2 diabetes with nephropathy. Dis Markers, 27:29-35.

50.

Stefanidis

, Tziastoudi

, Tsironi

, et al. (2018) The contribution of genetic variants of SLC2A1 gene in T2DM and T2DM-nephropathy: Association study and meta-analysis. Ren Fail, 40:561-576.

51.

Tanhäuserová

, Kuricová

, Pácal

, et al. (2014) Genetic variability in enzymes of metabolic pathways conferring protection against non-enzymatic glycation versus diabetes-related morbidity and mortality. Clin Chem Lab Med, 52:77-83.

52.

Thameem

, Igo

Jr , Freedman

, et al. (2013) A genome-wide search for linkage of estimated glomerular filtration rate (eGFR) in the Family Investigation of Nephropathy and Diabetes (FIND). PLoS One, 8:e81888.

53.

Trikalinos

, Karvouni

, Zintzaras

, et al. (2006) A heterogeneity-based genome search meta-analysis for autism-spectrum disorders. Mol Psychiatry, 11:29-36.

54.

Tsai

, Jiang

, Wu

, et al. (2001) Platelet collagen receptor α2β1 integrin and glycoprotein IIIa PlA1/A2 polymorphisms are not associated with nephropathy in type 2 diabetes. Am J Kidney Dis, 38:1185-1190.

55.

Tziastoudi

, Stefanidis

, Hadjigeorgiou

, et al. (2017) A systematic review and meta-analysis of genetic association studies for the role of inflammation and the immune system in diabetic nephropathy. Clin Kidney J, 10:293-300.

56.

Vardarli

, Baier

, Hanson

, et al. (2002) Gene for susceptibility to diabetic nephropathy in type 2 diabetes maps to 18q22.3-23. Kidney Int, 62:2176-2183.

57.

Vieira

, Monteiro

, Marques

, et al. (2011). Association of genetic variants in the promoter region of genes encoding p22phox (CYBA) and glutamate cysteine ligase catalytic subunit (GCLC) and renal disease in patients with type 1 diabetes mellitus. BMC Med Genet, 12:129.

58.

Wessman

, Forsblom

, Kaunisto

, et al. (2011) Novel susceptibility locus at 22q11 for diabetic nephropathy in type 1 diabetes. PLoS One, 6:e24053.

59.

Wise

, Lanchbury

, Lewis

(1999) Meta-analysis of genome searches. Ann Hum Genet, 63:263-272.

60.

, Bowden

, Spray

, et al. (1996) Linkage analysis between loci in the renin-angiotensin axis and end-stage renal disease in African Americans. J Am Soc Nephrol, 7:2559-2564.

61.

, Sale

, Rich

, et al. (1999) Evaluation of markers on human chromosome 10, including the homologue of the rodent Rf-1 gene, for linkage to ESRD in black patients. Am J Kidney Dis, 33:294-300.

62.

Zintzaras

, Ioannidis

(2005a) HEGESMA: genome search meta-analysis and heterogeneity testing. Bioinformatics, 21:3672-3673.

63.

Zintzaras

, Ioannidis

(2005b) Heterogeneity testing in meta-analysis of genome searches. Genet Epidemiol, 28:123-137.

64.

Zintzaras

, Ioannidis

(2008) Meta-analysis for ranked discovery datasets: theoretical framework and empirical demonstration for microarrays. Comput Biol Chem, 32:38-46.

65.

Zintzaras

, Kitsios

(2006) Identification of chromosomal regions linked to premature myocardial infarction: a meta-analysis of whole-genome searches. J Hum Genet, 51:1015-1021.

66.

Zintzaras

, Kitsios

, Harrison

, et al. (2006) Heterogeneity-based genome search meta-analysis for preeclampsia. Hum Genet, 120:360-370.

67.

Zintzaras

, Kitsios

, Kent

, et al. (2007b) Genome-wide scans meta-analysis for pulse pressure. Hypertension, 50:557-564.

68.

Zintzaras

, Papathanasiou

, Stefanidis

(2009) Endothelial nitric oxide synthase gene polymorphisms and diabetic nephropathy: a HuGE review and meta-analysis. Genet Med, 11:695-706.

69.

Zintzaras

, Uhlig

, Koukoulis

, et al. (2007a) Methylenetetrahydrofolate reductase gene polymorphism as a risk factor for diabetic nephropathy: a meta-analysis. J Hum Genet, 52:881-890.