Association Between Transcription Factor 7-Like-2 Polymorphisms and Type 2 Diabetes Mellitus in a Ghanaian Population

Christian Obirikorang, Evans Asamoah Adu, Enoch Odame Anto, Emmanuel Acheampong, Lawrence Quaye, Brodrick Yeboah Amoah, Max Efui Annani-Akollor, Aaron Siaw Kwakye, Foster Fokuoh, Michael Appiah, Eric Nana Yaw Nyarko, Freeman Aidoo, Eric Adua, Ebenezer AfrifaYamoah, Lois Balmer, Wei Wang 1 Department of Molecular Medicine, School of Medicine and Dentistry, Kwame Nkrumah University of Science and Technology, Kumasi, Ghana


3
Background Diabetes continues to rise in prevalence across the globe. Alarmingly, 4 out of 5 people with diabetes live in low-middle-income countries (LMIC), where adequate resources for diabetes management are scarce [1]. According to the International Diabetes Federation (IDF) 2017, it is estimated that nearly half a billion of the world's population is living with diabetes [1]. Khan, Hashim et al., (2) reported that 462 million individuals are affected by diabetes, corresponding to 6.28% of the world's population.
This figure is projected to reach 7079 individuals per 100,000 by 2030, reflecting a continued rise across all regions of the world [2]. Due to the continues increase rate of diabetes worldwide, the 2030 projected rate is likely to be exceeded [2,3]. This leads to high mortality from diabetic complications including cardiovascular disease [4]. The alarming situation is that 174.8 million (45.8%) of all diabetic cases, in adults, are undiagnosed with the most affected individuals from LMIC [5]. Thus, T2DM remains a threat to the world's population, even though there is increasing life-expectancy, which requires both public health and clinical care to prevent what is already a worldwide health crisis.
In Ghana, T2DM is prevalent among 6.46% [6] of the population, which is significantly greater than the 2017 International Diabetes Federation (IDF) estimation [1]. A number of factors including physical inactivity, poor diet, rapid urbanisation and ageing are common among people living with diabetes [7].
Besides these factors, T2DM has a strong genetic predisposition, taking the combination of inequalities across different populations, familial aggregation and concordance among monozygotic twins [8,9].
Specifically, in T2DM, these factors are closely related to the metabolic abnormalities that underline the conditions of impaired insulin action, obesity, increased endogenous glucose output, and insulin secretory dysfunction [10]. It is well-known that genes control and regulate glucose and metabolic homeostasis, which influence the digestion, absorption and metabolic processes of lipids and carbohydrates can initiate the major mechanisms in T2DM [11].
Genome-wide association studies (GWAS) have identified genetic variants in these metabolic pathways which would provide new ways to understand the mechanism of T2DM [12,13]. Among these genetic risk variants, the transcription factor 7-like 2 (TCF7L2), represent the most identified and promising T2DM candidate gene to date [14]. Variants in this gene has been identified with the strongest association with T2DM [14][15][16][17][18][19]. Since a common microsatellite in the TCF7L2 gene region, DG10S478, (containing SNPs rs7903146 and rs12255372) was found to be associated with T2DM [20], the finding has been convincingly replicated in different population groups. Using non-coding SNPs in strong linkage disequilibrium with DG10S478 [16,21,22], an estimated population attributable risk of 17-28% has been established.
TCF7L2 encodes a high mobility group of box-containing transcription factors, which is involved in the Wnt signalling pathway [17]. The Wnt signalling pathway represents one of the pancreatic beta cell's important developmental and regulatory mechanism, which is critical in determining its fate, proliferation, polarity, and apoptosis during embryonic development and adult tissue homeostasis [20].
Risk alleles of TCF7L2 have been strongly associated with impaired insulin secretion/beta-cell function and T2DM [23][24][25]. Likewise, TCF7L2 has been shown to mediate regulation of glucagon-like peptide-1 expression, which represents a mechanism by which alteration of this gene influences susceptibility to T2DM [24]. There has been only a limited number of studies within the African Context, which makes it difficult to pool the effects and homogeneity of variants associated with T2DM [26]. Thus, replication studies are needed to understand the effect of these gene variants on T2DM. This study investigated the association between TCF7L2 gene polymorphisms and T2DM in a Ghanaian population.  Genotype and allele frequencies of TCF7L2 gene polymorphism

Characteristics of the study participants
The genotype and minor allele frequencies (MAF) of rs7903146 and rs12255372 of the TCF7L2 gene are shown in Table 2. The MAF of rs7903146 was 50.5% and 33.3% among cases and control participants, with statistically significant differences (p= 0.011). The homozygote recessive allele (TT) of rs7903146, was more prevalent among cases than control participants (13.2% and 4.5%, p= 0.024).
The MAF of rs12255372 was 24.1% among cases, which is significantly higher (p= 0.013) than that of controls (14.2%). The GG genotype (p=0.013) and GT genotype (p= 0.023) but not the TT genotype (p= 0.501) of rs12255372 was disproportionately distributed between the cases vs control participants.

Linkage disequilibrium between rs7903146 and rs12255372
The linkage disequilibrium (LD) and pairwise LD coefficients r 2 and D' between rs12255372 and rs7903146 demonstrated that the alleles for the rs7903146 (C>T) and rs12255372 (G>T) SNPs are weakly associated (D' = 0.021; r 2 = 0.0001).
In Table S3, rs7903146 was significantly associated with increased odds of high LDL among the controls (OR= 2.98, p=0.017). Among the case group, individuals with the minor allele of rs7903146 were significantly associated with an increased odds ratio of high TG (OR=2.95, p=0.023) (Table S3).
rs12255372 SNP was significantly associated with an increased odds ratio of high cholesterol (OR=14.0, p=0.023) and elevated LDC-C (OR=3.12, p=0.027) among the control group (Table S3).

Discussion
Variants of TCFL2 gene has been consistently associated with T2DM in several studies among multiple ethnic groups [16,21,22]. This study demonstrated the transferability of the TCF7L2 gene polymorphisms (rs12255372 and rs7903146) and their association with T2DM among a Ghanaian population. However, the associations were observed to be relevant only in a subgroup-women, patients that do not have a family history of diabetes and those with dyslipidaemia with or without elevated blood pressure and increased BMI.
This study observed that rs7903146 and rs12255372 polymorphisms have a significant effect on T2DM, however, the effect was mediated by other diabetes risk factors. In a systematic review and Metaanalysis among a Sub-Saharan Africa (SSA) population [26], the effect of rs7903146 was 1.75 in a recessive model and 2.27 in an additive model, which is similar to our present findings. These effects are also consistent among the Caucasian, East Asian, South Asian and other ethnicities [21]. To the best of our knowledge, this study is currently among the few studies in SSA, that have demonstrated the transferability of rs12255372 [27,28]. In contrast studies by Yako, Madubedube et al., (29) and Mandour, Darwish et al., (30), found the SNP rs12255372 to not associated with T2DM, in a South African mixed-ancestry and Egyptian population, respectively. The MAF of rs7903146 observed in this study is similar to the findings of Danquah, Othmer (15) among a Ghanaian population, where the MAF of rs7903146 was 30% in the control population and 36% in T2DM population. Among other SSA populations, the MAF of rs7903146 among T2DM vs controls has been reported at 9.5% vs 0% in Cameroonian population [31] and 32.2% vs 24.3% in South African mixed-ancestry [29]. Furthermore, from the Allele Frequency Aggregator (ALFA) HapMap and 1000 Genomes project [32] the MAF of rs7903146 were 29.2%, 24.1% and 22.8%, respectively, which is consistent with our findings. The race-specific risk allele frequency of rs7903146 is lower among the Asian population (range: 3.1%-4.7%) and higher among Europeans (29.2%), Africans (29.5%), African American (29.7%), Latin American (34.2%) and South Asians (31.8%) [32]. A wide population difference in the minor allele frequency of rs12255372 (Asians= 0.000 to 0.029, Europeans = 0.217-0.50 and SSA= 0.267-0.274) has been reported [17,30,33]. Moreover, from the results of the 1000 genomes project, the MAF of rs12255372 is 21.4% globally, 30.2% in Africa, 29.2% among Europeans and the Americans 21.9%, similar to our study.
From the findings of this study and related literature, the association between rs12255372 and rs7903146 with T2DM are similar among West African populations, Asians, Caucasians [34,35] and other population groups [16]. This suggests that the association of rs7903146 and rs12255372 with T2DM is moderately strong in multiple ethnicities which are likely to project a vital contribution of differences in environmental factors that confer susceptibility to T2DM.
It appears that rs12255372 and rs7903146 are not associated with T2DM among West African populations (Ghanaians and Yoruba), unlike other population groups. This finding is essential in understanding the different types and aetiology of T2DM present in SSA [36], and the wider population differences in T2DM prevalence. For example, as indicated in this study the association between rs12255372 with T2DM was relevant among women, obese/overweight individuals, individuals with negative family history, low HDL-C and individuals aged above 60 years. Similarly, rs7903146 was relevant among females, individuals with negative family history, elevated TG and low HDL-C but normal blood pressure. This indicated that TCF7L2 risk variants may be relevant to a subgroup of individuals with single or multiple diabetes risk factors, other than those with a positive family history This study observed that the risk allele rs7903146 was associated with elevated FPG and HbA1c in both the diabetic and controls. In addition to FPG and HbA1c, the risk alleles of rs12255372 were associated with elevated BMI, blood pressure, TG and reduced LDL in T2DM and controls. This finding reflects the pleiotropic role TCF7L2 polymorphisms play in T2DM (adipogenesis, incretin and insulin expression as well as beta-cell growth and development). The rs7903146 risk alleles are associated with reduced proportional control of β-cell function among newly diagnosed T2DM patients [44]. Loos, Franks et al., (38) reported that the T-allele of rs7903146 was associated with higher HbA1c concentrations and reduced β-cell function, assessed by the homeostasis model assessment of β-cell function. TCF7L2 protein was also found to be increased during adipogenesis in 3T3-L1 cells and primary adipocyte stem cells [46] and TCF7L2 expression was required for the regulation of Wnt [39].
High expression of TCF7L2 mRNA was identified with reduced TCF7L2 protein levels in individuals with the T-allele of rs7903146, which is associated with a downregulation of GIP-and GLP-1 receptors and impaired beta-cell function Shu, Matveyenko et al., (40).
These findings suggest that TCF7L2 polymorphisms rs7903146 and rs12255372 may confer T2DM susceptibility through the expression of adipocyte-insulin resistance, delayed insulin secretion (resulting from beta-cell dysfunction) and reduced intracellular energy homeostasis. Therefore, individuals with the risk alleles of TCF7L2 (rs7903146 and rs12255372) may present with obesity, elevated TG and reduced HDL-C. Also, the findings suggest that among individuals with multiple diabetes risk factors, those who do develop T2DM may have a genetic predisposition to beta-cell dysfunction which includes the SNPs rs12255372 and rs7903146 [18]. Some limitations of the study are worth mentioning when interpreting the findings of this study. The differences in population substructures may result in a different interpretation of risk alleles in candidate gene association studies. Also, the number of genetic roots underlying T2DM is still being investigated and could be numerous given its multifaceted nature. Therefore, the effect of the variants among the small population considered in this study may be unpredictably low, when we consider it from a larger population. Besides this limitation, the findings of this study are consistent with the recommended approach for conducting a candidate gene association study and present consistent findings with similar studies among other population groups.

Materials and Methods
This case-control study was conducted at the Begoro District Hospital (BDH) and two communities (Bosuso and Begoro) in the Eastern Region of Ghana. The study included participants aged 30-79 years.
Cases were defined as clinically diagnosed T2DM patients attending the diabetic clinic at the BDH.
Controls were defined as non-diabetic patients with normal fasting plasma glucose (FPG) and/or normal HbA1C that voluntarily participated in a health screening program conducted as part of the study in Bosuso or Begoro. Participants willing to have their body mass index (BMI), FPG, HbA1c and lipid profile checked and provided written informed consent during the health screening were recruited as controls. Control participants with a self-reported, previous or current, history of renal, liver, thyroid disorders or other endocrine diseases, chronic diseases, on hormone replacement therapy or corticosteroid therapy were excluded from the study. T2DM patients who self-managed or being managed for micro-vascular or macrovascular complications were excluded from the study.

Recruitment Process
The control participants were recruited during a health screening program conducted in the two communities. Additional controls were sorted through community members that were informed via the various religious organizations. Initially, the intent, preparation and benefits of the health screening program was announced in all the religious-based organisations. The health screening process took place every morning (Monday, Wednesday and Friday) from 8:00 am to 11:30 am from May 2019 to November 2019. T2DM patients were recruited using a total sampling method. All patients that visited the BDH hospital during the study period and met the inclusion criteria were considered as a sample.
Blood pressure measurement was taken using an automated sphygmomanometer (Omron BP7200, Omron Healthcare, Inc., USA). Three different measurements were taken at 5-minute intervals and the average was used as the final reading. All blood pressure recordings were done in mmHg.
The Omron B511 (Omron Healthcare, Inc., USA) was used for weight measurement, following the manufacturer's instruction. Briefly, participants were asked to stand in upright on the main unit of the scale and without shoes. Portable Height Rod Stadiometers (QuickMedical, Warwick) were used for body height; the subject stood upright, with feet together and flat on the ground. BMI was calculated as the ratio of weight to height-squared (Kg/m 2 ).
Six millilitres (ml) of venous blood were taken in a fasting state and two millilitres each were placed into sterile ethylene diamine tetraacetic acid (EDTA) (BD Vacutainer ® , USA), gel separator and Sodium Fluoride/Calcium oxalate tubes (BD Vacutainer ® , USA). Fasting was defined as no caloric intake for at least 8 h, overnight. The EDTA samples were used for HbA1c analysis using standard A1cCare System (SD BIOSENSOR, republic of Korea). The remaining EDTA sample was used for DNA extraction and consequent SNPs genotyping. The gel separator tubes samples were processed by centrifugation at 3,000 rpm for 10 minutes and the serum was collected into 1.5 ml eppendorf tubes. The obtained sera were used for lipid profiling. The plasma obtained from the Sodium Fluoride/Calcium oxalate tube samples was used for the determination of Fasting Plasma Glucose (FPG).

Genomic DNA Extraction
Genomic DNA was extracted using a modified, simplified non-enzymatic salting-out method (details are shown in supplementary material). Briefly, 300 ul of whole blood was used from each of the 24 samples resulting in good quality of DNA with the concentration ranging between 60 to >100 ng/µl being extracted. The results of 1% (w/v) Agarose gel and Thermo Scientific™ NanoDrop 2000 full spectrum of the product are presented in Figure S1 and S2.

Data Analysis
Analysis of the data was conducted using SPSS version 25. Normality of continues data was checked using the Kolmogorov Simonov test. Descriptive statistics were reported using frequencies for categorical data and mean for continues data. Chi-square test was used to evaluate the independence of the categorical data and t-test for the continues data between cases and controls. To test for associations between T2DM phenotype and all SNPs, logistic regression models were fitted, in which each SNP was presented as a predictor variable whose values were equal to the number of copies of the minor allele (0, 1, 2) in an additive model, or presence of at least one copy of the minor allele (0, 1) in a codominant model or presence of two copies of the minor allele (0, 1) in a recessive model. Sex, age and family history were included as covariates in the fitted model. The structure of the model was represented as: Logit [pr (D = 1)] = α + β 1 G + β 2 sex(male) + β 3 age (≥60 years) + β positive family history + β high cholesterol β overweight/obesity To examine the associations between T2DM phenotype and various allelic configurations of rs7903146 and rs12255372, haplotype association analysis was performed. The statistical analyses for linkage disequilibrium (LD), haplotype frequencies, and multiple test corrections per 1000 random permutations were performed using the Haploview 4.2 software [46,47]. The haplotype blocks were defined by the four-gamete test [48,49] as implemented in Haploview. A weak interaction between rs7903146 and rs12255372 was observed, data not shown. Univariate models adjusted for age and gender, these were used to evaluate the association between the SNPs and cardiometabolic markers. Pvalues <0.05 were considered statistically significant.

Conclusion
There is a relatively high prevalence of the risk alleles of TCF7L2 (rs7903146 and rs12255372) among the Ghana population with a modest association with T2DM. The association between rs7903146 and rs12255372 with T2DM may be relevant in a subgroup of individuals (women), particularly those with no family history of diabetes and those with cardiometabolic risk with or without elevated blood pressure. Therefore, genotyping of these SNPs is essential for future multi-ethnic association studies and clinically significant for classifying high-risk patients.  Table S1: Master mix recipe for genotyping quality control Table S2: Association of rs7903146 with Anthropometric and biochemical markers among the control and case groups Table S3: Association of rs7903146 with cardiometabolic risk factors among the control and case groups Table S4: Association of rs12255372 with Anthropometric and biochemical markers among the control and case groups