Archives of Pharmacal Research

, Volume 36, Issue 2, pp 167–177

Genetics of type 2 diabetes and potential clinical implications


  • Soo Heon Kwak
    • Department of Internal MedicineSeoul National University Hospital
    • Department of Internal MedicineSeoul National University Hospital
    • Department of Internal MedicineSeoul National University College of Medicine
    • WCU Department of Molecular Medicine and Biopharmaceutical Sciences, Graduate School of Convergence Science and Technology and College of MedicineSeoul National University

DOI: 10.1007/s12272-013-0021-x

Cite this article as:
Kwak, S.H. & Park, K.S. Arch. Pharm. Res. (2013) 36: 167. doi:10.1007/s12272-013-0021-x


Type 2 diabetes (T2DM) is a common complex metabolic disorder that has a strong genetic component. Recent advances in genome-wide association studies have revolutionized our knowledge regarding the genetics of T2DM. There are at least 64 common genetic variants that are strongly associated with T2DM. However, the pathophysiologic roles of these variants are mostly unknown and require further functional characterization. The variants identified so far have a small effect size and their added effect explains less than 10 % of the T2DM heritability. The current ongoing whole exome and whole genome studies of T2DM are focused on identifying functionally important rare variants that have a stronger effect. Through these efforts, we will have a better understanding of the genetic architecture of T2DM and its pathophysiology. The potential clinical applications of genetic studies of T2DM include risk prediction, identification of novel therapeutic targets, genetic prediction of efficacy and toxicity of anti-diabetic medications, and eventually optimization of patient care through personalized genomic medicine. We hope further research in genetics of T2DM could aid patient care and improve outcomes of T2DM patients.


GeneticsGenomic medicineRisk predictionType 2 diabetes

Diabetes mellitus—definition and heritability

According to the World Health Organization’s definition, diabetes mellitus is a metabolic disorder of multiple etiologies characterized by chronic hyperglycemia with disturbance of carbohydrate, fat and protein metabolism resulting from defects in insulin secretion, insulin action, or both (World Health Organization 1999). Chronic hyperglycemia in diabetes can result in microvascular complications in eyes, kidneys, and neurons. It is also associated with increased risk of macrovascular complications, such as coronary heart disease and cerebrovascular accident. It is estimated that 366 million people worldwide have diabetes in 2011 and this figure will increase to 552 million in 2030 (International Diabetes Federation 2011). Diabetes is the leading cause of end stage renal disease, non-traumatic leg amputation and blindness. A total of 4.6 million deaths are attributable to diabetes which account for 8.2 % of global all-cause mortality (International Diabetes Federation 2011).

Type 2 diabetes (T2DM) is a common complex disease where multiple genetic and environmental factors have an intricate interplay (American Diabetes Association 2012). It is well known that a positive family history is a major risk factor for T2DM. In clinical practice, T2DM patient with strong positive family histories of diabetes are frequently seen. Subjects who have one parent with T2DM have approximately 30–40 % lifetime risk of developing T2DM and those who have both parents with T2DM have 70 % risk (Meigs et al. 2000). In addition, risk of developing T2DM in an individual who has a sibling with T2DM (λs) is increased about 2–4 fold compared to the normal population (Hemminki et al. 2010). Based on these findings, there has been an enormous effort to elucidatethe genetic risk factors of T2DM.

Hypothesis regarding the genetic architecture of T2DM

The inheritance of T2DM does not comply with Mendel’s law. T2DM is diagnosed based on a certain glucose threshold. The glucose concentration in a population follows a Gaussian distribution and it would be impossible for one or two genetic variants to explain the substantial variation of glucose concentrations. If a large number of independent variants were involved, then the distribution would mimic the Gaussian distribution. This was the basis of the polygenic theory that was explicitly presented by RA Fisher in 1918 (Fisher 1918). Following the polygenic theory of complex disease, there is an ongoing debate on the genetic architecture of T2DM (Schork et al. 2009). Is common T2DM caused by a large number of common variants with low penetrance (common disease, common variant hypothesis) or multiple rare variants with high penetrance (common disease, rare variant hypothesis)? There are several points that should be considered regarding these two hypotheses (Fig. 1). First, it is likely that both hypotheses are on the extreme end of a spectrum and there could be situations where both common and rare variants jointly predispose to T2DM. It is possible that some T2DM patients having rare highly penetrant variants will also have their disease risk modified through interaction with common less penetrant variants. Second, rare variants are assumed to have a larger effect size and T2DM associated with these rare variants will more likely exhibit earlier onset, more severe glucose excursion, and more familial clustering (Manolio et al. 2009). Common variants would have smaller effect sizes and their phenotypic presentation would be milder with less familial clustering. Third, the study design to identify genetic variants of T2DM depends on the hypothesis. Genome-wide association (GWA) study is more suitable for identifying common variants. Sequencing T2DM enriched families and subjects with extreme phenotypes are more suitable for rare variant identification. It is uncertain to what extent common variants and rare variants predispose to T2DM. Although there has been extensive research on the common genetic variants of T2DM, we are now looking forward to the results of the sequencing studies that are expected to reveal rare variants of T2DM.
Fig. 1

Genetic architecture of T2DM and related hypotheses. Rare variants are thought to have a stronger effect and are more enriched in familial cases with extreme phenotype. Common variants have small effect size and over several hundreds of variants are expected to modulate T2DM susceptibility (Morris et al. 2012). The rare variant hypothesis and common variant hypothesis are on the extreme ends of a spectrum and it is possible that both types of variants have joint effects through interaction. The size of the circle is relative to its effect size

Recent advances in T2DM genetics

The GWA study tries to identify common genetic variants across the entire human genome associated with T2DM using more than 100,000 genetic markers in a large sample size of 1,000 or more (Pearson and Manolio 2008). The GWA study of T2DM was first introduced in 2007 and has tremendously increased our knowledge in the genetics of T2DM (Sladek et al. 2007). There are more than 30 T2DM GWA studies listed in the Catalog of Published Genome-Wide Association Studies web site (, accessed on October 1st, 2012) and more than 64 genetic variants are identified as associated with T2DM in a genome-wide significance level of P < 5.0 × 10−8(Zeggini et al. 2007, Saxena et al. 2007, Scott et al. 2007, WTCCC 2007, Zeggini et al. 2008, Voight et al. 2010, Kooner et al. 2011, Cho et al. 2012a, Morris et al. 2012). In addition, 53 genetic variants are associated with glycemic traits of fasting glucose, fasting insulin, and post-challenge 2-hour glucose concentration (Dupuis et al. 2010, Scott et al. 2012).

Following the GWA studies in Europeans, there have been reports of GWA studies and GWA meta-analysis in different ethnicities including East Asians, South Asians, and Africans (Cho et al. 2012a, Kooner et al. 2011, Palmer et al. 2012). Although many genetic loci were common to different ethnic groups, some showed interethnic differences. For example, variants that were specific to East Asians include those located near PTPRD, SRR, CDC123/CAMK1D, PSMD6, MAEA, ZFAND3, KCNK16, GCC1/PAX4, GLIS3, and PEPD(Cho et al. 2012b, Morris et al. 2012). Some variants, such as rs7903146 in TCF7L2, differed significantly in risk allele frequency between East Asians (2 %) and Europeans (25 %). Therefore, the odds ratios(OR) and their resulting contribution to genetic susceptibility to T2DM were different between the two ethnic groups (Park 2011). Further research is required to contrast the differences in genetic risk factors for different ethnicities and their influence on phenotypic difference of T2DM.

There has been a plethora of T2DM genetic variants identified mainly through large scale GWA studies (Fig. 2). Most of the variants identified so far have not been implicated with T2DM previously and their function is mostly unknown. Among the variants with known function or results on physiologic analyses, many of them seem to result in pancreatic β-cell dysfunction. However, there are also variants that exert diabetogenic effect through obesity or insulin signaling defect. It is beyond our focus to review all the genetic variants of T2DM in this article, but we will highlight some of the variants whose function seems to have a major role in the pathogenesis of T2DM.
Fig. 2

Suggested function of genes associated with T2DM in GWA studies. References indicate the first association results for variants with unknown function or the reports for the physiologic functional analysis


The first two genetic variants significantly associated with T2DM were located in PPARG and KCNJ11, and were identified by candidate gene approach (Altshuler et al. 2000, Gloyn et al. 2003). PPARG encodes peroxisome proliferator-activated receptor gamma and it is a target for an anti-diabetic medication thiazolidinedione. The rs1801282 (P12A) variant of this gene has been consistently associated with risk of T2DM. In the most recent meta-analysis of the GWA studies, this variant increased the risk of T2DM with OR 1.13, 95 % confidence interval (CI) 1.09–1.17, and P = 1.1 × 10−12(Morris et al. 2012). KCNJ11 encodes potassium inwardly-rectifying channel, subfamily J, member 11 and consists a subunit of sulfonylurea receptor in pancreatic β-cell. One of the most widely used anti-diabetic medications, sulfonylurea, binds to this receptor and stimulates insulin secretion. The rs5215 (V250I) variant of this gene increases the risk of T2DM with OR 1.07 (95 % CI 1.05–1.10), and P = 8.5 × 10−10 (Morris et al. 2012).


The association between TCF7L2 and T2DM was first identified by a linkage analysis that mapped to the long arm of chromosome 10 (Reynisdottir et al. 2003). Fine mapping effort in this region showed that a microsatellite marker located in the intron of TCF7L2 was robustly associated with risk of T2DM (Grant et al. 2006). TCF7L2 encodes high mobility group box-containing transcription factor that plays a key role in the Wnt signaling pathway (Grant et al. 2006). The intronic variant of TCF7L2, rs7903146 showed the most significant association with the largest effect size among all the identified genetic variants of T2DM (OR 1.39, 95 % CI 1.35–1.42, and P = 1.2 × 10−139) (Morris et al. 2012). This variant is also associated with decreased β-cell insulin secretion and decreased incretin effect although the latter is debated (Villareal et al. 2010, Smushkin et al. 2012).


One of the novel genetic variants first identified through GWA studies included CDKAL1 (Scott et al. 2007, Saxena et al. 2007, Zeggini et al. 2007). The CDKAL1 encodes cyclin-dependent kinase (CDK) 5 regulatory subunit associated protein 1-like 1. The rs7756991 variant of CDKAL1 is associated with T2DM (OR 1.17, 95 % CI 1.14–1.20, and P = 7.0 × 10−30) and decreased β-cell insulin secretion (Morris et al. 2012, Pascoe et al. 2007). At first, it was thought that CDKAL1 would be associated with decreased pancreatic mass and compensation as it is homologous with CDK5 inhibitor CDK5RAP1(Zeggini et al. 2007). However, recent findings suggest that a defect in CDKAL1 could impair processing of proinsulin to insulin and thereby reduce insulin secretion (Wei et al. 2011).


Another interesting variant, rs10811661, that was newly identified through GWA study, is located within the intergenic region in the CDKN2A-CDKN2B gene cluster at chromosome 9p21 (Saxena et al. 2007, Scott et al. 2007, Zeggini et al. 2007). The rs10811661 variant has strong association with T2DM (OR 1.18, 95 % CI 1.15–1.22, and P = 3.7 × 10−27) (Morris et al. 2012). Although the exact functional consequence of this variant is not fully understood, it is speculated to be associated with pancreatic β-cell dysfunction (Grarup et al. 2007). An interesting thing about this 9p21 locus is that it is not only associated with T2DM, but also with risk of myocardial infarction, intracranial aneurysm, Alzheimer’s disease, malignant melanoma, and glaucoma (Shea et al. 2011). Although variants that are driving the risk of myocardial infarction are located in a different haplotype of T2DM, it would be interesting to find out whether and how these variants interact with T2DM associated variants.


Variants in the FTO gene are significantly associated with both T2DM and body mass index (BMI) (Scott et al. 2007, Frayling et al. 2007). FTO encodes fat mass and obesity associated gene and one of its variant rs9936385 increases the risk of T2DM (OR 1.13, 95 % CI 1.10–1.16, and P = 2.6 × 10−23) (Morris et al. 2012). This variant was first discovered to be associated with T2DM but subsequent analysis revealed that the association with T2DM was conferred by its effect on BMI (Frayling et al. 2007). In a recent meta-analysis including 96,551 Asians, rs9939609 variant in FTO gene was consistently associated with risk of obesity and T2DM (Li et al. 2012). Functional analysis in a mouse model showed that overexpression of FTO gene resulted in increased fat mass and body mass (Fischer et al. 2009).


The last variant that we would like to underscore is located in MTNR1B. MTNR1B encodes melatonin receptor 1B which is a G-protein coupled receptor expressed in pancreatic β-cell (Bouatia-Naji et al. 2009). The rs10830963 variant of MTNR1B increases risk of T2DM with OR 1.10, 95 % CI 1.07–1.13, and P = 5.3 × 10−13(Morris et al. 2012). We found that rs10830962, a variant that is in linkage disequilibrium with rs10830963 was significantly associated with risk of gestational diabetes, which is thought to share a similar genetic background with T2DM (Kwak et al. 2012). Recently, large-scaleexome sequencing of this gene showed that variants that impaired melatonin receptor signaling was specifically associated with T2DM (Bonnefond et al. 2012).

The shortcomings

Above listed are only part of the 64 confirmed T2DM associated genetic variants. The functions of the genes identified so far are mostly unknown. It is not clear whether the identified variants are causal variants per se or just markers in close linkage disequilibrium with the unidentified causal variants. As most of the variants are located in introns or intergenic regions, the functional consequences of the allelic variation are also very difficult to predict. Recent fine mapping efforts using Metabochips were not successful in identifying functional common variants with larger effect size (Voight et al. 2012, Morris et al. 2012). Criticismarose soon after the initial reports of the GWA studies, that the effect size of the common variants was too small and could not explain more than 10 % of the heritability T2DM poses (Manolio et al. 2009). There are several large-scaleexome sequencing studies currently underway. We will have to see whether these efforts will identify rare functional variants with large effect sizes or successfully pinpoint the functional causal variants that are marked by the variants that have been identified through GWA studies.

Potential clinical application of T2DM genetics

There are four potential clinical applications of the T2DM genetics (Fig. 3). First, it could be used for the prediction of T2DM. Second, it may reveal new therapeutic targets for drug development. Third, pharmacogenomic research will enable us to selectthe right drug for the right patient. Fourth, knowledge regarding genetics of T2DM will eventually lead us to personalized genomic medicine to optimize patient care.
Fig. 3

Potential clinical application of genetic information of T2DM

Prediction of T2DM

A major suggested clinical utility of genetic information that spurred the research on genetics of T2DM was improving risk prediction. After the outburst of initial discoveries of genetic loci that were associated with T2DM, several studies investigated the predictability of genetic information compared to known clinical risk factors (Vassy and Meigs 2012). Most of the studies used genotype risk score which is either the un-weighted or weighted sum of the total risk alleles a subject inherits (Lyssenko et al. 2008, Meigs et al. 2008, de Miguel-Yanes et al. 2011). When the area under the curve of the receiver operating characteristics (C-statistics) were compared, T2DM prediction model composed of clinical risk factors such as family history, BMI, blood pressure, and fasting glucose showed significantly higher C-statistics compared to the genotype risk score alone (Meigs et al. 2008). In addition, when genotype risk score was incorporated into the clinical prediction model, the added value was minimal (Meigs et al. 2008). There could be several explanations for this. First, it could be due to the small effect size of the individual variants. Second, many of the clinical risk factors would already reflect genetic effects of the variants. Third, most studies involved middle-aged populations and the genetic effect would be more pronounced in subjects who develop T2DM at earlier age. At the current state, routine screening of genetic information to predict T2DM is not recommended (Vassy and Meigs 2012). Nevertheless, if the exome sequencing studies identify functional variants with a higher effect size, the predictability would be improved. If genetic variants of a certain T2DM individual are determined, then the same variants can be tested in the first degree relatives to see whether they share the same variants and have similar risk of developing T2DM. The utility of the genetic prediction model in high risk groups such as women with previous history of gestational diabetes and early onset T2DM would be an interesting research subject.

Novel therapeutic targets of T2DM

As mentioned above, the biomolecular functions of the genetic variants are mostly unknown. Many research groups are making efforts to clarify their role in the pathogenesis of T2D and this will hopefully reveal novel therapeutic targets of T2DM. Among the variants discussed above, melatonin receptor 1B could be a potential target of T2DM. As genetic variants impairing this receptor signaling are associated with T2DM (Bonnefond et al. 2012) and as melatonin itself is a strong anti-oxidant (Hardeland 2005), there is a possibility that melatonin could be used as an anti-diabetic medication (Korkmaz et al. 2012). Another potential target of T2DM that has been revealed through genetic research is CREB binding protein (CREBBP). In a recent GWA study, pathway analysis and protein–protein interaction analysis revealed that transcriptional coactivator CREBBP was the major interacting protein for the T2DM associated variants (Morris et al. 2012). This suggests that CREBBP and its associated transcription factors are important in the pathogenesis of T2DM and should be investigated for potential modulation. We hope further functional investigation of the genes that have been associated with T2DM will provide us with opportunities to develop novel drugs targeting these genes.

Pharmacogenomics of T2DM

The four most widely used oral anti-diabetic medications include metformin, sulfonylurea, thiazolidinedione, and dipeptidylpeptidase-4 inhibitor. There are individual variations in response to these medications and genetic predisposition could be one of the factors influencing these variations. Regarding sulfonylurea, the rate-limiting enzyme of its metabolism is CYP2C9 (Kirchheiner et al. 2002). It has been reported that loss-of-function variants of CYP2C9 were associated with improved glycemic response presumably due to impaired metabolism of sulfonylurea (Zhou et al. 2010). Several other genetic variants in KCNJ11 and ABCC8 gene have been suggested to be associated with the glycemic response to sulfonylurea (Sesti et al. 2006, Feng et al. 2008). Metformin is transported to hepatocytes via organic cation transporter 1 (OCT1) (Wang et al. 2002) and excreted to bile through multidrug and toxin extrusion 1 protein (MATE1) (Tanihara et al. 2007). There are reports that variants in OCT1 and MATE1 influence the pharmacokinetics of metformin, affectingthe glucose lowering response (Shu et al. 2007, Becker et al. 2009). The most noticeable finding regarding pharmacogenetics of metformin was derived from a GWA study. In 3,920 European subjects, the rs11212617 variant in the ataxia telangiectasia mutated gene (ATM) was significantly associated with better glycemic response to metformin (OR 1.35, 95 % CI 1.22–1.49, and P = 2.9 × 10−9) (Zhou et al. 2011). It is suggested that activation of AMP-activated protein kinase (AMPK), which is a target of metformin, through ATM is required for the glycemic response to metformin. Pharmacogenetic information regarding thiazolidinedione and dipeptidylpeptidase-4 inhibitor is somewhat limited and further research is required.

Personalized genomic medicine

The final goal of genetic research on T2DM would be to tailor therapy on an individual basis and optimize patient care. There has been a study showing the possibility of incorporating the whole genome sequencing data of an individual into clinical decision making in terms of risk prediction, choosing an optimal medication with appropriate dosage (Ashley et al. 2010). This approach has been extended to multi-omics profiling combining genomics, transcriptomics, proteomics, and metabolomics in an individual in a time series of 14 months where the study subject experienced two viral infections and new onset of T2DM (Chen et al. 2012). This study suggests that multi-omics monitoring in an individual could help interpret the physical condition of a patient. However, these were only a proof-of-concept study and there are many hurdles before personalized genomic medicinesare realized. The accuracy of sequencing has to be improved for medical diagnostic purposes and more solid information regarding genotype-phenotype relationship should be accumulated. In addition, the clinical utility and cost-effectiveness of using a novel genomic tool in clinical practice should also be evaluated.

Concluding remarks

T2DM is a common complex disease where genetic predisposition has a major role in its development. Recent advancement in GWA studies has significantly improved our knowledge regarding the genetic architecture and pathophysiology of T2DM. The exome sequencing studies that are currently undergoing would further improve our insights on genetics of T2DM. Through these efforts we hope genetic information could be used to optimize patient care and improve outcomes in T2DM patients.


This work was supported by the National Project for Personalized Genomic Medicine, Ministry for Health & Welfare, Republic of Korea (grant no.A111218-12-GM01).

Copyright information

© The Pharmaceutical Society of Korea 2013