I27L Polymorphism in hepatocyte nuclear factor-1α gene and type 2 diabetes mellitus: A meta-analysis of studies about orient population (Chinese and Japanese)

1. Introduction 

Hepatocyte nuclear factor-1α (HNF-1α) is a liver-enriched transcription factor that plays a role in the transcription regulation of several genes involving glucose metabolism and islet development. It has been confirmed that some rare gene mutations of this factor led to type 3 maturity-onset diabetes of the young (MODY3) [1]. And it has been speculated that common variants of this gene might play a role in the pathogenesis of type 2 diabetes mellitus (T2DM). Recently, several genome-wide scans have found evidence of linkage of type 2 diabetes to chromosome 12q24, where the HNF-1α gene is located [2], [3], [4]. One of the commonest variants in HNF-1α gene, I27L/exon1, occurs in a completely conserved region among humans, rats, mice and hamsters, which is important for HNF-1α dimerization [5]. The replacement of I (Isoleucine) with L (Leucine) might give rise to a small conformational change and might thereby influence DNA binding. Some studies have indicated that the I27L variant was not only was associated with β cell dysfunction, but also played a modest role in the pathogenesis of insulin resistance [5], [6]. All of this evidence indicates that I27L might be one of the risk factors of T2DM.

A number of studies have investigated the relationship between I27L polymorphism and T2DM, mainly concerning Caucasian populations. The previous small size studies did not find any association [7], [8], [9], [10]. Subsequent large population studies, however, displayed inconsistent results. For instance, Winckler et al. [11] genotyped I27L polymorphism in >4000 diabetic patients and control subjects from Sweden, Finland and Canada, and found a nominally significant association of 27L with T2DM, but their subsequent studies in an additional 4470 Caucasian cases and matched controls of US and Polish ancestry, and the concurrent study by Weedon et al. [12] in 5307 Caucasian T2DM and controls, did not show any evidence of the association of 27L to T2DM. Holmkvist et al. further evaluated their relationship, based on the data of a case-control study (3736 Scandinavian cases and controls) and two prospective studies, that is, thew Botnia prospective study (2293 Finland non-diabetic subjects) and the Malmö Prevention Project (MPP, 4873 Sweden individuals), their results showed that the I27L polymorphism was associated with increased risk of T2DM, odds ratio (OR)=1.5 (p=0.002; multiple logistic regression), particularly in elderly (age >60years) and overweight (BMI >25kg/m2) patients (OR=2.3, p=0.002) [13]. Their findings proposed a hypothesis that the I27L variant might cause subtle defects in beta cell function which should become more easily manifested during conditions of enhanced demand, e.g. insulin resistance as in elderly overweight individuals.

People of oriental origin tended to be younger and less obese than the Caucasian population at the onset of T2DM, which suggested that oriental people have more vulnerable islet β cell function, or are more insulin resistant [14], [15], [16]. Therefore, oriental people with 27L polymorphism should, in theory, be at increased risk of T2DM, as is the case in older or overweight Caucasian populations. A small number of studies involving the orient population, however, failed to find any association between 27L and T2DM [9], [17], [18], [19], [20]. Due to limited sample size of these studies, a possible relationship might be still exist and need to be further investigated. The study here aimed to investigate the distribution of I27L gene variance in T2DM patients and normal glucose tolerance controls within Chinese Han nations from the Chengdu area, and to investigate the relationship between the I27L and T2DM in Orient population by pooling data for this study and other published reports.

2. Patients and methods 

Case-control study 149 unrelated type 2 diabetic patients diagnosed according to the 1999 World Health Organization criteria were recruited from Chendu city of China, which included 80 males and 69 females, with a mean age 61.8±10.1years. Patients suspected with other specific types of diabetes were excluded. Ninety-six unrelated non-diabetic individuals, including 54 males and 42 females, aged 60years or more and with mean age 70.72±6.71, served as controls. All participants gave informed consent. HbA1c was measured using an ion-exchange high-performance liquid chromatography method (Bio-Rad Diamat, Hercules, CA). Plasma glucose levels were measured by using glucose oxidase methods. Genomic DNA was extracted from peripheral leukocytes. The I27L polymorphism was genotyped using a PCR–RFLP method, as described elsewhere [10].

3. Results 

In our case-control study, the genotype distribution of I27L/exon1 polymorphism in the HNF-1α gene was in Hardy–Weinberg equilibrium (χ2=2.34, 0.05<P<0.1). The genotype frequencies were II 0.34, IL 0.52, LL 0.14 in T2DM and II 0.43, IL 0.49, LL 0.08 in non-diabetic controls respectively, without significantly statistical difference (χ2=2.54, P>0.05). The I allele and L allele were 0.60, 0.40 in T2DM group and 0.67, 0.33 in non-diabetic controls respectively, the difference was not significant also (χ2=2.84, P=0.24).

Our electronic searches identified 27 reports about HNF-1α gene variance and T2DM in orient population that were selected for full text review; the completed list of these reports is available on request. 23 (82.1%) papers were excluded because they did not detect I27L polymorphism or present the information of this polymorphism. Five studies (17.9%) and our present study providing information on a total of 1225 unique subjects (717 type 2 diabetic patients and 508 controls) were subjected to the following meta-analysis. Three published studies concerned Chinese population and two studies involved Japanese population. As shown in Table 1, the frequencies of 27L ranged from 0.39 to 0.53 in diabetic patients, and from 0.33 to 0.48 in controls. Total 27L allele frequencies were 0.43 in diabetic patients, a little higher than in controls (0.40). Heterogeneity was not found between the six trials (χ2=3.83, P=0.57, I2=0%). Thus, a fixed-effects model was performed. The result showed that 27L slightly increases the risk of diabetes (OR 1.22, 95% CI 1.03–1.44, P=0.02) (see Fig. 1).

Table 1. I27L allelic frequencies of HNF-1α gene in different studies of East Asian population.

	Author	Population	DM		Controls		P
			N	Freq. (I/L)	N	Freq. (I/L)
	Nishigori [17]	Japanese	60	0.56/0.44	100	0.58/0.42	NS
	Babaya [18]	Japanese	57	0.47/0.53	88	0.52/0.48	NS
	Rissanen [9]	Chinese	105	0.5/0.5	51	0.61/0.39	NS
	Fang [19]	Chinese	154	0.55/0.45	93	0.62/0.38	NS
	Yang Zhen [20]	Chinese	192	0.61/0.39	80	0.59/0.41	NS
	Present study	Chinese	149	0.60/0.40	96	0.67/0.33	NS

View Large Image Download to PowerPoint

Fig. 1. Meta-analysis result of the relationship between HNF-1α 27L/exon1 and T2DM in East Asian population.

4. Discussion 

Type 2 diabetes is a common human disease that is influenced by both genetic and environmental factors. As in most common diseases, very few variants have been rigorously proven to play a role in the common form of type 2 diabetes. Up to now, only a few gene variants, e.g. rs7903146 in TCF7L2, E23K in KCNJ11, and P12A in PPAR-γ, are established as major genetic contributors to T2DM in Caucasian population [21], [22], [23]. Their effects on T2DM in the oriental population, however, were quite different [24], [25], [26], [27], [28], [29]. For instance, rs7903146 in TCF7L2 was the top one major genetic risk factor for T2DM in population of European origin, but its allelic frequency was much lower in the East Asian population than in Europeans, and its association with T2DM were inconsistent among studies [24], [25], [30]. As for P12A in PPAR-γ gene, no association with T2DM was found in the Chinese and Japanese populations [28], [29]. What is more, the prevalence of T2DM was not uncommon, or was even more common in the oriental population, as compared with the Caucasian population [14]. All of this suggested that genetic background to T2DM varied from Caucasian populations to oriental populations, and there might be other potential risk factors responsible for T2DM in the oriental population.

I27L polymorphism is located within the dimerization domain of the HNF-1α gene, and the isoleucine amino acid is conserved among several species, indicating a potential functional importance of this residue. In vitro study, 27L polymorphisms in the HNF-1α gene showed decreased transcriptional activity on the target promoters glucose transporter 2 [13]. In vivo studies, I27L variance was also indicated to have an effect on glucose metabolism. For example, Ken et al. examined 52 healthy glucose-tolerant and normotensive subjects and found that the 27L polymorphism is associated with insulin resistance [6]. They then further investigated 78 healthy subjects and found individuals with 27LL genotype may also have a declined β cell function [5].

In our case-control study, we found that the 27L allelic frequency and 27IL/27LL genotypes frequencies were slightly higher in the T2DM group than in non-diabetes control group (0.44, 0.52, 0.14 vs. 0.30, 0.49, 0.08), but the differences were not statistically significant. The 27L allelic frequency was similar to that obtained from studies in Caucasian population, ranging from 0.21–0.47 in T2DM patients and 0.26–0.45 in controls [9], [17], [18], [19], [20], and was also similar to that of other studies in orient population, that is, 0.39–0.53 in T2DM patients and 0.33–0.48 in controls [9], [17], [18], [19], [20]. No relationship was found between 27L and T2DM in our case-control study, which was in line with other small sample studies in orient populations [9], [17], [18], [19], [20].

We then performed a meta-analysis, by pooling together previously published studies and our new study. The results show that 27L has a positive association with diabetes (OR 1.22, 95% CI 1.03–1.44). This result was in keeping with Holmkvist et al.’s recent findings that 27L variant increased the risk of T2DM, especially in elderly and overweight persons who were more insulin resistant and/or deteriorated beta cell function [13]. It is known that T2DM patients of Asian origin usually have a less BMI and a younger age at the onset or T2DM, which indicating orient population might be more genetically susceptible to T2DM [14], [16]. The causative mechanism is still undetermined, though it has been proposed that the islet beta cell of East Asians were more vulnerable in the setting of accumulating insulin resistance [15]. This meta-analysis results corresponding with Holmkvist et al.’s findings, which implied that a subtle defect of beta cell function caused by the common variant of 27L variant might have a modest effect on the onset of T2DM in a subset of high risk populations of T2DM.

The primary limitation of our study is the small sample size. Even when combining previous published studies with our present study, there are still only 1225 individuals. Publication bias is another potential limitation. Small sample studies with negative results might not be published. We tried to minimize this effect by search meeting abstracts via ISI proceedings (http://apps.isiknowledge.com), but there still might be studies with negative result could not be retrieved. Finally, the studied populations are mainly Chinese and Japanese; no studies about Korea population and other populations of East Asian origin were found.

From the present study, may we conclude that the common amino acid polymorphism of the HNF-1a gene, I27L does seemingly play a role in type 2 diabetes mellitus in oriental populations (Chinese and Japanese). For the sake of potential limitations, this result still needs to be further confirmed by large population-based studies.

5. Declaration of competing interests 

None to declare.