Serum circulating microRNA profiling for identification of potential markers of diabetic nephropathy in black African South Africans with type 2 diabetes mellitus

Valentin, Stefan Drikus
Journal Title
Journal ISSN
Volume Title
Introduction: Type 2 diabetes mellitus (T2DM) is a metabolic disorder characterised by insulin resistance. The prevalence of T2DM is increasing with an estimated 14.2 million adults with T2DM expected in Africa by the year 2040. Poor glycaemic control, over a prolonged period of time, leads to diabetic micro- and macrovascular complications. Of the microvascular complications, diabetic nephropathy (DN) has been implicated in causing the majority of deaths associated with T2DM. DN is characterised by progressive decline of renal function, resulting in end-stage renal disease. With early intervention, kidney damage can be prevented or even reversed, however no test currently exists for early detection of DN. Currently, urine albumin-to creatinine ratios (UACR) are used to monitor kidney function and diagnose DN. UACR as a marker is unspecific and unreliable, and there is a need for a more sensitive and specific marker. Serum circulating microRNA (miRNA) have shown promise as potential novel biomarkers as they are stable and often differentially expressed in different pathological states. This study, therefore, aimed to detect differentially expressed serum miRNAs in black South African participants with T2DM with and without microalbuminuria (MA) to identify potential biomarkers for early detection of DN. In addition, the prevalence of polymorphisms in the HMGA2 and TGF-β1 gene and their associations with markers of kidney function were investigated. Further, the use of serum cystatin C (SCC) as an alternative to serum creatinine for estimation of glomerular filtration rate (GFR; used to measure kidney function and to classify the stage of kidney disease) was assessed. Methods: Black South African participants with T2DM (n = 238) were recruited from diabetes clinics at Charlotte Maxeke Johannesburg Academic Hospital and Chris Hani Baragwanath Hospital. A subset of participants (n = 14) were classified into two groups based on the absence (UACR < 2 mg/mmol; controls, n = 7) or presence (UACR 2-20 mg/mmol; cases, n = 7) of MA. The participants were matched for age, medication use, and anthropometry. miRNA expression profiles were analysed using the Affymetrix® GeneChip® 4.0 assay and transcriptome analysis console (TAC) software. Two differentially expressed miRNAs were selected for validation through real-time quantification polymerase chain reaction (PCR) in a subset of 40 participants (20 participants with the lowest UACR, and 20 participants with the highest UACR). Two genes that were targeted by the differentially expressed miRNA were selected for polymorphism analysis. The total cohort was genotyped for HMGA2 rs1114167319, and TGF-β1 V rs1800417 polymorphism by amplification refractory mutation systems (ARMS) PCR. In addition, participants were genotyped for the HMGA2 rs1114167320 polymorphism by PCR restriction fragment length polymorphism (RFLP). SCC concentrations were measured using a R&D enzyme-linked immunosorbent assay (ELISA). Estimated GFR (eGFR) were calculated through the use of two creatinine-based equations (modification of diet in renal disease study (MDRD) equation without ethnic factor correction and the chronic kidney disease epidemiology collaboration (CKD-EPI) equation), and three SCC-based equations. Passing-Bablok linear regression and Band-Altman plots were generated to evaluate correlation and method agreement between the eGFR equations, using the MDRD equation as the reference method. Results: The study population consisted of a 238 black South Africans with T2DM. The cohort had a mean age of 56.6 ± 9.88 years and a mean duration of disease of 11.5 ± 8.82 years. The majority of the participants were female (61.8 %), and they had poor glycaemic control (HbA1c of 8.68 ± 2.73 %). Differential miRNA expression profiling (> 2-fold difference between cases and controls) identified nine miRNAs. miR-455-3p and let-7b-5p were upregulated, and miR-4740-3p, miR 7704, miR-101-5p, miR-5189-3p, miR-1273g-3p, miR-16-5p and miR-6880-5p were downregulated in cases relative to controls. Real-time quantitative PCR validation of expression levels of miR-455-3p and let-7b-5p found that expression levels of both miRNAs were higher in the group with the lowest UACR compared to the group with the highest UACR (miR-455-3p: 40.5 [39.1; 42.8] vs 42.5 [41.3; 45.3] CT value, p = 0.043; and let-7b-5p: 26.3 ± 1.50 vs 27.6 ± 2.03 CT value, p = 0.022). The HMGA2 rs111416319 and rs111416320 major alleles were the C (0.51) and A allele (1.0), respectively. The major allele for the TGF-β1 rs1800471 polymorphism was the G allele (0.64). None of the polymorphisms were in Hardy-Weinberg equilibrium (HWE). Despite not being in HWE, we did perform additional analyses with rs111416319 and rs1800471. rs111416319 was not associated with any markers of kidney function or DM. Similarly, rs1800471 was not associated with markers of kidney function, however, participants with the GG genotype had significantly lower glucose concentrations compared to participants with the GC/CC genotype (6.60 [4.80; 8.80] vs. 8.10 [5.70; 11.3] mmol/l; p = 0.011). The classification of stages of kidney disease using the five different eGFR equations resulted in significantly different clinical classifications of kidney function (p < 0.001). The MDRD VI and CKD-EPI equation correlated strongly (r = 0.945; p < 0.001) however there was poor agreement between the equations with a bias of -39.7 %. There was good agreement between SCC equation 2 and equation 3 with the MDRD equation with a bias of -7.71 % and -7.10 %, respectively. Conclusion: Let-7b-5p and miR-455-3p protect against the development of DN. Let-7b-5p binds to the TGF-β1 receptor 1 inhibiting the TGF-β1 signalling pathway. This pathway has been shown to play a role in DN by enhancing glucose-induced cell hypertrophy and increased expression of extracellular matrix (ECM) genes (e.g. collagen and fibronectin). In addition, let-7b 5p targets the 3’ untranslated region of the HMGA2 gene, that has been implicated in activation of the TGF-β1 signalling pathway. Thus, increased expression of let-7b-5p leads to decreased renal fibrosis via inhibiting HMGA2 and TGF-β1 receptor expression. miR-455-3p overexpression decreases ECM synthesis and inflammatory cytokines resulting in protection from renal fibrosis. Furthermore, TGF-β1 inhibits the expression of miR-455-3p, thus activating renal fibrotic pathways, resulting in reduced kidney function and DN development. Further studies are required in the South African black population to confirm the role of these miRNAs in DN. Longitudinal studies need to be conducted to confirm the role of let-7b-5p and miR-455-3p as possible early markers of DN. The eGFR equations showed significant variation in terms of classification of kidney disease and cannot be used interchangeably
A dissertation submitted in fulfilment of the requirements for the degree of Master of Science in Medicine to the Faculty of Health Sciences, School of Pathology, University of the Witwatersrand, Johannesburg, 2020