Key Genetic Components of Fibrosis in Diabetic Nephropathy: An Updated Systematic Review and Meta-Analysis

Renal fibrosis (RF) constitutes the common end-point of all kinds of chronic kidney disease (CKD), regardless of the initial cause of disease. The aim of the present study was to identify the key players of fibrosis in the context of diabetic nephropathy (DN). A systematic review and meta-analysis of all available genetic association studies regarding the genes that are included in signaling pathways related to RF were performed. The evaluated studies were published in English and they were included in PubMed and the GWAS Catalog. After an extensive literature review and search of the Kyoto Encyclopedia of Genes and Genomes (KEGG) database, eight signaling pathways related to RF were selected and all available genetic association studies of these genes were meta-analyzed. ACE, AGT, EDN1, EPO, FLT4, GREM1, IL1B, IL6, IL10, IL12RB1, NOS3, TGFB1, IGF2/INS/TH cluster, and VEGFA were highlighted as the key genetic components driving the fibrosis process in DN. The present systematic review and meta-analysis indicate, as key players of fibrosis in DN, sixteen genes. However, the results should be interpreted with caution because the number of studies was relatively small.


Introduction
Diabetic nephropathy (DN) is a multifactorial condition that involves both metabolic and hemodynamic factors, but clear evidence (heritability estimates, familial clustering of the disease, and differential vulnerability according to race) also indicates that the genetic profile of individuals plays a major role [1][2][3]. In addition, differences in the risk of DN among ethnic populations indicate that differences in lifestyle and environment also play major roles in the selection of alleles contributing to the different risks of disease occurrence [4]. These observations suggest that some diabetic patients are programmed to develop DN. Given the continuous increase in both DM and its complications, the effort to prevent DM is still considered a global challenge.
Among the histological lesions that occur in DN [5,6], renal fibrosis (RF) is considered as a complex and irreversible process in the late stages of DN, further exacerbating the progression of the disease [6]. In RF, which takes place in response to injury and inflammation, there is excessive deposition of the extracellular matrix (ECM) and epithelial-mesenchymal transition (EMT), leading to the loss of differentiated epithelial cells and their vascular capillary bed, accumulation of myofibroblasts and inflammatory cells, and, ultimately, the formation of a scar [7]. As a result, the destruction of the normal architecture and function of the kidney occurs [6,8,9]. RF also constitutes the end point of all kinds of chronic kidney disease (CKD), regardless of the initial cause of disease [8].
It is noteworthy that, although fibrosis is considered a detrimental condition, recent studies suggest a protective role of this process, because it helps the maintenance of crosstalk with injured proximal tubular cells supporting their regeneration [10]. In autosomal-dominant polycystic kidney disease (ADPKD), for instance, fibrosis may have some protective roles, as the fibrogenic response is generally correlated with cystic disease regression [11].
In order to identify the key players of fibrosis in the context of DN, we performed a systematic review and meta-analysis of all available genetic association studies regarding genes that are involved in signaling pathways related to RF. In this study, we included all available genetic association studies regarding fibrosis-related genes that have been published within a time period of 35 months since our field synopsis of all available genetic association studies regarding DN [13] and presented an updated meta-analysis of the relevant polymorphisms.

Data Synthesis and Analysis
The genetic association between each polymorphism and DN was assessed using the generalized odds ratio (ORG) [24,25]. The threshold for the meta-analysis was the presence of two studies per genetic polymorphism. The pooled OR was estimated using the Der Simonian and Laird random-effects model [26]. The associations are presented with ORs and their corresponding 95% confidence intervals (Cis). The between-study heterogeneity was tested with Cochran's Q statistic (considered statistically significant at p < 0.10) and its extent was assessed with the I 2 statistic [27,28].
We also examined if controls confronted with Hardy-Weinberg equilibrium (HWE) predicted genotypes using Fisher's exact test for each study that provided genotype counts. We also tested for the 'small-study effect' with the Egger test [29].

Study Characteristics
The literature search of both PubMed and the GWAS Catalog retrieved 5058 papers after the exclusion of duplicate studies, whereas, in the meta-analysis, 180 articles were included. When an article provided data for different populations, each population was considered as a different study. Figure 2 presents a flowchart of the retrieved articles and the reasons for the exclusion of certain papers. The included studies were published between 1994 and 2021. The demographic characteristics of each study are shown in Supplementary  Table S9, which was updated from our previous study [13]. Supplementary Table S9 only presents the included studies of these polymorphisms that were updated from our previous study [13].

Meta-Analysis Results
Among the 884 different genes that were involved in eight pathways related to fibrosis, 134 genetic variants located in 45 different genes were meta-analyzed. In comparison with our previous study [13], the present study updated results for 24 genes. Table 2 shows the statistically significant results of the meta-analyses based on genotype counts, whereas Table 3 shows the statistically significant results of the meta-analyses based on allele counts. Tables 2 and 3 include also updated data from meta-analyses that were already published by our group in a previous study [13]. However, for the purpose of completeness, Supplementary Tables S10-S12 also present the non-significant results. Figures 3 and 4 are forest plots that show the pooled odds ratios of the significant results. Overall, sixteen genes provided significant results in all meta-analyses.    In meta-analyses based on genotype counts, statistically significant results were reported for angiotensin I-converting enzyme (ACE), angiotensinogen (AGT), erythropoietin (EPO), gremlin 1, DAN family BMP antagonist (GREM1), interleukin 1 beta (IL1B),    In meta-analyses based on genotype counts, statistically significant results were reported for angiotensin I-converting enzyme (ACE), angiotensinogen (AGT), erythropoietin (EPO), gremlin 1, DAN family BMP antagonist (GREM1), interleukin 1 beta (IL1B), interleukin 6 (IL6), interleukin 10 (IL10), nitric oxide synthase 3 (NOS3), and transforming growth factor beta 1 (TGFB1).
More specifically, in a comparison of diseased controls versus cases, EDN1 rs1794849 was significantly associated with DN, with a pooled OR of 1.16 (95% CI 1.02-1.31), FLT4 rs2242221 produced significant results with a pooled OR of 1.14 (95% CI 1.01-1.29), two variants in the IGF2/INS/TH cluster, rs1004446 and rs4320932, were also associated with DN, with pooled OR values of 1.16 (95% CI 1.03-1.31) and 0.84 (95% CI 0.73-0.96), respectively, and VEGFA rs2146323 also produced significant results, with a pooled OR of 0.85 (95% CI 0.76-0.95). In a comparison of healthy controls versus cases, only one variant in IL12RB1 (rs372889) was associated with DN, with a pooled OR value of 1.24 (95% CI 1.13-1.37). Figures 3-6 are forest plot representations of the genetic variants that are significantly associated with DN.
More specifically, in a comparison of diseased controls versus cases, EDN1 rs1794 was significantly associated with DN, with a pooled OR of 1.16 (95% CI 1.02-1.31), FL rs2242221 produced significant results with a pooled OR of 1.14 (95% CI 1.01-1.29), t variants in the IGF2/INS/TH cluster, rs1004446 and rs4320932, were also associated w DN, with pooled OR values of 1.16 (95% CI 1.03-1.31) and 0.84 (95% CI 0.73-0.96), resp tively, and VEGFA rs2146323 also produced significant results, with a pooled OR of 0 (95% CI 0.76-0.95). In a comparison of healthy controls versus cases, only one varian IL12RB1 (rs372889) was associated with DN, with a pooled OR value of 1.24 (95% CI 1. 1.37). Figures 3-6 are forest plot representations of the genetic variants that are sign cantly associated with DN.

Discussion
To the best of our knowledge, this review constitutes the most comprehensive stu of the genetic variants that are related to fibrosis in the context of DN. This study is extension of our previous work in the field of the genetic epidemiology of DN [13]

Discussion
To the best of our knowledge, this review constitutes the most comprehensive study of the genetic variants that are related to fibrosis in the context of DN. This study is an extension of our previous work in the field of the genetic epidemiology of DN [13], but, for the first time, we focused on fibrosis-related genes. Among the 134 genetic variants located in 45 different genes that were meta-analyzed, sixteen genes (ACE, AGT, EDN1, EPO, FLT4, GREM1, IL1B, IL6, IL10, IL12RB1, NOS3, TGFB1, IGF2/INS/TH cluster, and VEGFA) produced significant results in all meta-analyses.
The present systematic review confirmed the statistical significance of genes that are well known to have fibrotic effects, such as ACE, AGT, EDN1, GREM1, IL1B, IL6, and TGFB1, whereas some other genes are known for their anti-fibrotic effects, such as EPO and NOS3. The statistically significant contribution of IL12RB1 and TH constitutes a novel finding, as there are no available experimental results to clarify their contribution to fibrosis.
More specifically, ACE and AGT constitute members of the renin-angiotensin system (RAS), but this finding is no surprise because RAS inhibitors also have antifibrotic effects [30]. Many lines of evidence indicate that RAS is a major regulator of renal fibrosis, as angiotensin II (AngII) promotes the release of TGF-β and also activates the inflammatory process [31]. In addition, many publications indicate the involvement of TGFB1, which has been also characterized as the master regulator of fibrosis via the activation of both canonical and non-canonical pathways [32]. TGF-β is also regarded as the most important inducer of endothelial-to-mesenchymal transition (EndMT) in vitro and in vivo, a form of EMT [33], suggesting that targeting the TGF-β receptor signaling pathway could constitute a putative treatment for fibrosis [33].
Regarding the role of the innate proinflammatory cytokine IL1B, it has been reported that it induces a metabolic switch from oxidative phosphorylation to glycolysis in kidney stromal cells (SCs), promoting proximal tubule damage and fibrosis [34]. Increased expression of IL-6 and extensive and chronic activation of STAT3 were also associated with fibrosis [35]. The pathways of the anti-inflammatory mediator IL-10 have been found to be conserved in many diseases associated with fibrosis, although their underlying dissimilarity suggests that the IL-10 signaling pathway may have antifibrotic properties [36]. With regard to IL12RB1, to the best of our knowledge, this is a novel finding as no previous studies have found any association of IL12RB1 with kidney fibrosis.
A previous study reported that the disruption of eNOS and ApoE genes accelerates kidney fibrosis and senescence after injury [37], indicating a protective role of the normal function of these genes. In addition, a meta-analysis that investigated the association between the (eNOS) 4b/a gene polymorphism and renal interstitial fibrosis in patients with DN demonstrated that the frequency of eNOS4bb in DN renal interstitial patients was lower than that in non-nephropathy diabetic patients and normal controls [38]. Regarding EDN1, it has been found that the endothelin receptors in renal interstitial cells do not contribute to the development of fibrosis during experimental kidney disease [39]. However, a randomized controlled trial, SONAR, regarding the selective endothelin A receptor antagonist atrasentan showed promising results, as atrasentan reduced the risk of renal events in patients with diabetes and chronic kidney disease [40]. Regarding the EPO gene, it has been found that erythropoietin attenuates renal interstitial fibrosis via the inhibition of fibrocyte accumulation [41]. Gremlin1 (Grem1), which is an antagonist of bone morphogenetic proteins, plays a key role in kidney development and renal fibrosis, and a previous study demonstrated that its levels are increased in many diseases associated with fibrosis [42]. More specifically, the grem1 levels are increased in renal fibrosis, as well as in fibrosis of the heart and lungs. FLT4, a tyrosine kinase receptor for VEGFC and VEGFD, is involved in lymphangiogenesis, which is a condition that develops during the progression of fibrosis, indicating a fibrotic effect of this factor [43].
Regarding the IGF2/INS/TH cluster, previous results have shown that IGF2 stimulates the differentiation into myofibroblasts that produce large amounts of collagen and other extracellular matrix proteins (ECM) [44]. Regarding INS, a hormone that plays a key role in carbohydrate and lipid metabolism, it has been found that sodium-glucose cotransporter 2 (SGLT2) inhibitors suppressed kidney fibrosis in diabetic mice [45]. Regarding TH, which is involved in tyrosine-to-dopamine conversion, there are no results regarding its involvement in kidney fibrosis. Finally, VEGFA, a growth factor essential for both physiological and pathological angiogenesis, can inhibit the expression of Smad3 and miR192, thereby suppressing TGF-β-induced EMT and improving renal fibrosis [46].
Based on the statistically significant results of the present systematic review and meta-analysis, it should be noted that mast cells (MCs) are implicated in many fibrotic conditions [47][48][49][50]. ACE, IL1B, IL-6, and IL-10 are key mediators of MCs, and these genes have produced statistically significant results in the present systematic review and metaanalysis. The role of MCs in the fibrotic process is controversial, as studies in humans and in vitro data indicate a pro-fibrotic role of these cells, whereas animal studies have produced inconsistent results [51]. The reason may be the duration of the stimuli. Based on the fact that MCs are scarce in healthy human kidneys and are very rarely observed in glomeruli [52], these cells could serve as sensors of injury and could enhance the repair process. When the injury is short-lived, the MCs have an anti-fibrotic effect, but when the injury is chronic or repeated, the MCs have a pro-fibrotic effect [51]. It has been also found that an increase in the number of MCs is correlated negatively with renal function [53,54], but is correlated positively with the extent of fibrosis [54][55][56]. One more reason for the inconsistency is the fact that MCs produce both pro-fibrotic and anti-fibrotic mediators [49,[57][58][59][60].

Conclusions
In summary, the present systematic review and meta-analysis indicate, as key players of fibrosis in DN, sixteen genes (ACE, AGT, EDN1, EPO, FLT4, GREM1, IL1B, IL6, IL10, IL12RB1, NOS3, TGFB1, IGF2/INS/TH cluster, and VEGFA). However, the results should be interpreted with caution because the number of studies in most meta-analyses is relatively small.