Abstract
Aims/Introduction
We aimed to assess the distribution of transcription factor 7‐like 2 gene TCF7L2 (rs7903146) polymorphism and to find possible associations between TCF7L2 and the characteristics of type 1 diabetes.
Materials and Methods
We studied 190 newly diagnosed type 1 diabetes patients (median age 12.7 years, range 2.0–72.5) and 246 controls (median age 23.8 years, range 1.4–81.5) for TCF7L2 single nucleotide polymorphism. We determined anti‐islet autoantibodies, random C‐peptide levels, diabetes associated HLA DR/DQ haplotypes and genotypes in all patients.
Results
There were no differences in the distribution of TCF7L2 single nucleotide polymorphism between patients and controls. However, patients with in type 1 diabetes, after adjusting for age and sex, subjects carrying C allele were at risk for a C‐peptide level lower than 0.5 nmol/L (OR 5.65 [95% CI: 1.14–27.92]) and for zinc transporter 8 autoantibody positivity (5.22 [1.34–20.24]). Participants without T allele were associated with a higher level of islet antigen‐2 autoantibodies (3.51 [1.49–8.27]) and zinc transporter 8 autoantibodies (2.39 [1.14–4.99]).
Conclusions
The connection of TCF7L2 polymorphism with zinc transporter 8 and islet antigen‐2 autoantibodies and C‐peptide levels in patients supports the viewpoint that TCF7L2 is associated with the clinical signs and autoimmune characteristics of type 1 diabetes. The mechanisms of the interaction between the TCF7L2 risk genotype and anti‐islet autoantibodies need to be studied further.
Keywords: Anti‐islet autoantibodies, Transcription factor 7‐like 2 gene, Type 1 diabetes
We aimed to assess possible association between TCF7L2 genotypes (rs7903146) and characteristics of type 1 diabetes including autoantibodies and C‐peptide levels. Our findings about the association of TCF7L2 polymorphism with zinc transporter 8, islet antigen 2 autoantibodies and C‐peptide levels in patients support the view that TCF7L2 may contribute to development of type 1 diabetes. The interaction between the TCF7L2 risk genotype and autoantibodies needs to be studied further.
INTRODUCTION
Diabetes mellitus is a disease that is categorized according to the different mechanisms of pathogenesis, but shares a common ground with them: increased blood glucose levels, or hyperglycemia 1 . The type characterized by autoimmune destruction of beta‐cells, is referred to as type 1 diabetes. In contrast, type 2 diabetes derives from many different pathophysiological processes including lower ability of insulin to stimulate glucose uptake, lower ability of glucose to stimulate and suppress its own uptake, and suppression of insulin action 1 , 2 . Regardless of the mechanism of pathogenesis, both type 1 diabetes and type 2 diabetes share similar signs and symptoms, while their onset and development are different 2 .
The mechanisms underlying type 1 diabetes related beta‐cell destruction have been studied thoroughly and can be examined under environmental and genetic factors 3 , 4 , 5 . Concerning genetic triggers, associations between human leukocyte antigen (HLA) and type 1 diabetes were studied nearly half a century ago 6 . The results indicated that specific HLA regions were spotted more often in insulin dependent diabetes 6 . Overall, it was shown that HLA class II DQ‐DR loci on chromosome 6 were of great importance and haplotypes DR3‐DQ2 and DR4‐DQ8 had a synergistic effect in pathogenesis 7 , 8 , 9 .
According to a genome‐wide association study (GWAS), one single nucleotide polymorphism (SNP) of gene transcription factor 7‐like 2 (TCF7L2) was the most significantly associated signal in type 2 diabetes patient groups 10 , 11 . The TCF7L2 is a transcription factor that forms part of the lymphopoiesis influencing the Wnt signaling pathway 12 . It also plays an important role in the development of the intestinal epithelium, as TCF7L2 knock‐out mice cannot survive more than 1 day according to previous studies 13 . Supporting this, the loss of TCF7L2 impairs the morphology of cells as shown in another recent study 14 . Yet the mechanism is still being investigated. The TCF7L2 regulates the expression of proglucagon gene (glu), supporting the production of glucagon and glucagon like peptide‐1 (GLP‐1) 2 , 15 . The opposing glucagon, GLP‐1, stimulates insulin production, and lowers blood sugar level 2 , 15 . Recent studies have shown that even though TCF7L2 is a transcription factor known to be associated with type 2 diabetes, it can also have unexpected effects on the pathogenesis of type 1 diabetes. For example, it was found that TCF7L2 rs7903146 C‐to‐T polymorphism (T‐risk allele) decelerates single to multiple autoantibody progression in type 1 diabetes patients with low‐risk HLA loci, demonstrating a protective effect 16 .
Proceeding from previous knowledge and the above research, we aimed to assess the prevalence of the risk genotype in the Estonian population and its possible association with type 1 diabetes markers, including clinical and baseline characteristics and autoantibody presence, which influence the mechanism of pathogenesis of type 1 diabetes.
MATERIALS AND METHODS
Study population
This study included 190 recent‐onset type 1 diabetes patients (2.0–72.5 years old, median age 12.7 years, 105 males and 85 females). A total of 246 unrelated participants without anti‐islet autoantibodies were included in the study as controls (1.4–81.5 years old, median age 23.8 years, 90 males and 156 females) (Table 1). The study was approved by the Research Ethics Committee of the University of Tartu (protocols 163/T‐6 from 24.09.2007 and 275/M/15 from 20.11.2017). All participants or their parents (or guardians, if needed) signed a written informed consent form before participation.
Table 1.
Summarized characteristics of the study participants
| Characteristic | Type 1 diabetes patients (n = 190) | Controls (n = 246) |
|---|---|---|
| Male, % (n) | 55.3 (105/190) | 36.6 (90/246) |
| Age (years), median (IQR) | 12.7 (8.7–26.3) | 23.8 (16.6–32.2) |
| Duration of symptoms (days), median (IQR) | 30 (14–60) | – |
| BMI, % (n) | ||
| Underweight | 19.1 (18/94) | – |
| Normal weight | 50.0 (47/94) | |
| Overweight | 20.2 (19/94) | |
| Obesity | 10.6 (10/94) | |
| Ketonuria, % (n) | ||
| Yes | 73.0 (130/178) | – |
| No | 27.0 (48/178) | |
| Ketoacidosis, % (n) | ||
| Yes | 35.5 (66/186) | – |
| No | 64.5 (120/186) | |
| C‐peptide (nmol/L), median (IQR) | 0.23 (0.13–0.36) | – |
| C‐peptide, % (n) | ||
| <0.5 nmol/L | 87.7 (136/155) | – |
| ≥0.5 nmol/L | 12.3 (19/155) | |
| AAB positive persons, % (n) | ||
| GADA | 79.5 (151/190) | 0 (0/246) |
| IA2A | 57.9 (110/190) | 0 (0/246) |
| ZnT8A | 68.4 (130/190) | 0 (0/246) |
| AAB levels (U/mL), median (IQR) | ||
| GADA | 150 (35.5–1,900) | |
| IA2A | 380 (60–4,000) † | |
| ZnT8A | 370 (98.3–950) | |
| AAB count, % (n) | ||
| 0 | 8.4 (16/190) | 100 (246/246) |
| 1 | 17.4 (33/190) | 0 (0/246) |
| 2 | 34.2 (65/190) | 0 (0/246) |
| 3 | 40 (76/190) | 0 (0/246) |
| HLA risk, % (n) | ||
| DR3/DR4 | 25.8 (49/190) | 0.4 (1/246) |
| DR4/x | 26.8 (51/190) | 18.3 (45/246) |
| DR3/x | 24.7 (47/190) | 18.3 (45/246) |
| x/x | 22.6 (43/190) | 63.0 (155/246) |
AAB, (anti‐islet) autoantibodies; BMI, body mass index; GADA, autoantibodies against glutamic acid decarboxylase; HLA, human leukocyte antigen; IA2A, autoantibodies against islet antigen‐2; IQR, interquartile range; n, count; ZnT8A, autoantibodies against zinc transporter 8.
Results are presented for values detected after May 2015. Median values detected before May 2015 were 320 (IQR 92.5–1750).
Patients with type 1 diabetes were recruited from the Children's Clinic of Tartu University Hospital, from Tallinn Children's Hospital and from the Internal Medicine Clinic of Tartu University Hospital between October 2004 and March 2020. Patients with type 1 diabetes were diagnosed by using the internationally accepted diagnosis criteria 1 . Their peripheral blood samples were obtained up to 30 days after the diagnosis (1–26 days). The random C‐peptide levels of the patients were 0.01–2.78 nmol/L (median = 0.23 nmol/L, cut‐off value = 0.5 nmol/L were used 17 ). C‐peptide levels were lower than 0.5 nmol/L in more than half of the patients (87.7%). Two participants in the type 1 diabetes group were young children with C‐peptide levels higher than the reference range. For further calculations, they were added to the >0.5 nmol/L C‐peptide group. More details about the patient group and the control group are presented in Table 1.
The control group for this study consisted of non‐diabetic children visiting the Surgery Clinic of Tartu University Hospital and volunteers. All controls had normal HbA1c and/or normal glucose levels and had no anti‐islet autoantibodies. For the study, controls and patients were recruited synchronously.
Variables
Information on the baseline characteristics including age, sex, body mass index (BMI), family history of type 1 diabetes, and concomitant autoimmune diseases were obtained from patients and controls at the time of recruitment. Clinical variables, including duration of symptoms, random C‐peptide levels, and the presence of ketonuria and ketoacidosis, were obtained for the patient group.
Detection of antibodies
Autoantibodies against glutamic acid decarboxylase (GADA), islet antigen‐2 (IA2A), and zinc transporter 8 (ZnT8A) were detected in patients with type 1 diabetes and in controls using commercial ELISA kits in compliance with the manufacturer's instructions (RSR Ltd, Cardiff, UK). The cut‐off values indicating antibody positivity were ≥5 U/mL for GADA and ≥15 U/mL for ZnT8A. The cut‐off values indicating IA2A positivity were ≥15 U/mL (tests up to May 2015) or ≥7.5 U/mL (tests after May 2015). All tests are regularly validated as external quality assurance by the Islet Autoantibody Standardization Program 18 . For statistical analysis, the results for autoantibodies were divided into two groups: (1) lower level, i.e. values lower than median and (2) higher level, i.e. values equal to or higher than median (Table 1).
Genotyping
For HLA‐DR/DQ genotyping, PCR‐based lanthanide labeled oligonucleotide hybridization with time resolved fluorometry was used as portrayed in the literature 19 , 20 . Comparison groups were created regarding the following combinations of HLA‐DRB1‐DQA1‐DQB1 haplotypes: (1) DR3/DR4, (2) DR4/x, (3) DR3/x, and (4) x/x, where the haplotypes that were not DR3 or DR4 were denoted by x. Grouping was carried out according to Ilonen et al. 9
DNA was extracted from peripheral blood leukocytes using the salting out method. The TCF7L2 rs7903146 were genotyped using the TaqMan SNP genotyping assay (Applied Biosystems, Foster City, CA, USA) on the ABI 7000 instrument (Applied Biosystems). The SNPs were genotyped according to the manufacturer's protocol (Applied Biosystems). All investigated polymorphisms were in the Hardy–Weinberg equilibrium (P > 0.05).
Statistical analysis
Type 1 diabetes patients and controls were first evaluated using descriptive statistics. Continuous variables were expressed as median and interquartile range (IQR) and non‐parametric data were provided as count and percentage. For comparison of categorical variables, the χ2 test or Fisher's exact test was employed. For continuous variables, the Mann–Whitney U‐test or the Kruskal–Wallis test was used for comparison. Logistic regression models adjusted for age and gender were used to evaluate the influence of TCF7L2 on the outcome of interest. A linear regression model with log‐transformation, adjusted for age and gender, was used for variables with a normal distribution. The R: A Language and Environment for Statistical Computing (version 3.6.1; R core Team, Vienna, Austria) was used for statistical analyses. Values of P <0.05 were evaluated as statistically significant. The Hardy–Weinberg equilibrium of SNPs was verified with the χ2 test. There was no significant variation in the studied populations and their equilibrium was not significantly different.
RESULTS
Distribution of genotypes in the study populations
The results of genotyping TCF7L2 rs7903146 revealed that 11 of the 190 (5.8%) patients with type 1 diabetes and 11 of the 246 (4.5%) controls were carrying the type 2 diabetes risk genotype TT. Among the 190 patients with type 1 diabetes, 120 had the CC genotype (63.2%) and 59 had the CT genotype (31.1%). The distribution of the TCF7L2 genotypes and alleles did not reveal any significant difference between patients with type 1 diabetes and the controls. Additional information is presented in Table 2.
Table 2.
Distributions of the TCF7L2 (rs7903146) genotypes and alleles in the study population
| TCF7L2 | All (n = 436) | Type 1 diabetes patients (n = 190) | Controls (n = 246) | P‐value |
|---|---|---|---|---|
| Genotype, % (n) | ||||
| CC | 60.1 (262/436) | 63.2 (120/190) | 57.7 (142/246) | 0.3170 |
| CT | 34.9 (152/436) | 31.1 (59/190) | 37.8 (93/246) | |
| TT | 5.0 (22/436) | 5.8 (11/190) | 4.5 (11/246) | |
| CT + TT | 40.3 (174/436) | 36.8 (70/190) | 42.3 (104/246) | 0.2936 |
| Allele, % (n) | ||||
| C | 77.5 (676/872) | 78.7 (299/380) | 76.6 (377/492) | 0.5221 |
| T | 22.5 (196/872) | 21.3 (81/380) | 23.4 (115/492) | |
Associations with baseline characteristics of controls
Initially we evaluated the TCF7L2 rs7903146 genotypes and alleles in controls against the HLA risk genotypes, age and sex, however, we did not establish any significant association between them (see Tables S1 and S2).
Associations with baseline clinical characteristics of patients with type 1 diabetes
Further, the effects of TCF7L2 rs7903146 polymorphism were investigated in patients with type 1 diabetes. Additional information is presented in Table 3 and Table S3. Median C‐peptide levels were similar between the different genotypes (Figure S1). However, we found a significant association of genotypes with C‐peptide levels <0.5 nmol/L and ≥0.5 nmol/L (P = 0.0437; Table 3). A similar tendency was also found between these C‐peptide levels and genotype groups (P = 0.0591; Table S3). We did not establish significant association of genotype or genotype group with onset variables (ketoacidosis, ketonuria, or duration of symptoms), HLA risk genotypes or any other disease characteristics (Table 3 and Table S3).
Table 3.
Associations between the TCF7L2 genotypes and baseline clinical characteristics of type 1 diabetes patients
| Characteristic | CC genotype (n = 120) | CT genotype (n = 59) | TT genotype (n = 11) | P‐value |
|---|---|---|---|---|
| Male, % (n) | 54.2 (65/120) | 57.6 (34/59) | 54.5 (6/11) | 0.9392 |
| Age (years), median (IQR) | 13.9 (9.4–25.9) | 11.5 (6.9–25.9) | 12.2 (10.6–29.7) | 0.2659 |
| Duration of symptoms (days), median (IQR) | 30 (14–45) | 30 (14–60) | 14 (11–53) | 0.6168 |
| BMI, % (n) | ||||
| Underweight | 19.4 (12/62) | 8.5 (5/27) | 9.1 (1/5) | 0.9757 |
| Normal weight | 50.0 (31/62) | 23.7 (14/27) | 18.2 (2/5) | |
| Overweight | 16.1 (12/62) | 10.2 (6/27) | 9.1 (1/5) | |
| Obesity | 11.3 (7/62) | 3.4 (2/27) | 9.1 (1/5) | |
| Ketonuria, % (n) | 70.8 (80/113) | 75.0 (42/56) | 88.9 (8/9) | 0.5405 |
| Ketoacidosis, % (n) | 36.1 (43/119) | 31.0 (18/58) | 55.6 (5/9) | 0.3793 |
| C‐peptide (nmol/L), median (IQR) | 0.23 (0.13–0.33) | 0.21 (0.15–0.41) | 0.23 (0.11–0.65) | 0.7404 |
| C‐peptide, % (n) | ||||
| <0.5 nmol/L | 91.0 (91/100) | 85.1 (40/47) | 62.5 (5/8) | 0.0437 |
| ≥0.5 nmol/L | 9.0 (9/100) | 14.9 (7/47) | 37.5 (3/8) | |
| GADA positivity, % (n) | 83.3 (100/120) | 71.2 (42/59) | 81.8 (9/11) | 0.1767 |
| GADA (U/mL), median (IQR) | 135 (34–1825) | 150 (36–2000) | 160 (43–310) | 0.9254 |
| GADA level, % (n) | ||||
| Lower level | 51.0 (51/100) | 50.0 (21/42) | 44.4 (4/9) | 0.9629 |
| Higher level | 49.0 (49/100) | 50.0 (21/42) | 55.6 (5/9) | |
| IA2A positivity, % (n) | 58.3 (70/120) | 55.9 (33/59) | 63.6 (7/11) | 0.9107 |
| IA2A level, % (n) | ||||
| Lower level | 40.0 (28/70) | 63.6 (21/33) | 100.0 (7/7) | 0.0011 |
| Higher level | 60.0 (42/70) | 36.4 (12/33) | 0.0 (0/7) | |
| ZnT8A positivity, % (n) | 67.5 (81/120) | 76.3 (45/59) | 36.4 (4/11) | 0.0375 |
| ZnT8A (U/mL), median (IQR) | 450 (112–1,175) | 225 (68–775) | 175 (121–267) | 0.1897 |
| ZnT8A level, % (n) | ||||
| Lower level | 42.7 (35/82) | 60.9 (28/46) | 75.0 (3/4) | 0.0915 |
| Higher level | 57.3 (47/82) | 39.1 (18/46) | 25.0 (1/4) | |
| AAB count, % (n) | ||||
| Single AAB positive | 18.0 (20/111) | 18.9 (10/53) | 30.0 (3/10) | 0.6071 |
| Multiple AAB positive | 82.0 (91/111) | 81.1 (43/53) | 70.0 (7/10) | |
| AAB positivity, % (n) | ||||
| AAB negative | 7.5 (9/120) | 10.2 (6/59) | 9.1 (1/11) | 0.6772 |
| AAB positive | 92.5 (111/120) | 89.8 (53/59) | 90.9 (10/11) | |
| HLA risk group, % (n) | ||||
| DR3/DR4 | 23.3 (28/120) | 32.2 (19/59) | 18.2 (2/11) | 0.5984 |
| DR4/x | 26.7 (32/120) | 25.4 (15/59) | 36.4 (4/11) | |
| DR3/x | 23.3 (28/120) | 27.1 (16/59) | 27.3 (3/11) | |
| x/x | 26.7 (32/120) | 15.3 (9/59) | 18.2 (2/11) | |
P‐values in bold are statistically significant.
AAB, (anti‐islet) autoantibodies; BMI, body mass index; GADA, autoantibodies against glutamic acid decarboxylase; HLA, human leukocyte antigen; IA2A, autoantibodies against islet antigen‐2; IQR, interquartile range; n, count; ZnT8A, autoantibodies against zinc transporter 8.
As a next step, we attempted to find out if TCF7L2, age and sex have any effect on a C‐peptide level lower than <0.5 nmol/L in the logistic regression model. Logistic regression analysis, adjusted for age and sex, revealed that the CC/CT group was the risk factor vs the TT group (OR 5.65, P = 0.0336) for having a C‐peptide level of <0.5 nmol/L. Older age (OR 0.96, P = 0.0122) was a protective factor for a C‐peptide level of <0.5 nmol/L (Table 4). Further, we also examined the genotypes and found that the CC genotype was a risk factor for having a C‐peptide level of <0.5 nmol/L vs the TT genotype (OR 7.04, P = 0.0214) and older age (OR 0.96, P = 0.0122; Table S4).
Table 4.
Logistic regression analysis evaluating the association of the TCF7L2 genotype groups, age, and sex with a C‐peptide level lower than 0.5 nmol/L vs higher than 0.5 nmol/L at the time of the diagnosis of type 1 diabetes
| Characteristic | Unadjusted OR (95% CI) | P‐value | Adjusted OR † (95% CI) | P‐value |
|---|---|---|---|---|
| Genotype group | ||||
| TT | 1 ‡ | 1 ‡ | ||
| CC + CT | 4.91 (1.07–22.52) | 0.0405 | 5.65 (1.14–27.92) | 0.0336 |
| Age (years) | 0.97 (0.94–0.99) | 0.0153 | 0.96 (0.94–0.99) | 0.0122 |
| Sex | ||||
| Male | 1 ‡ | 1 ‡ | ||
| Female | 1.44 (0.53–3.87) | 0.4741 | 1.34 (0.48–3.74) | 0.5821 |
P‐values in bold are statistically significant.
CI, confidence interval; OR, odds ratio.
Adjusted for age and sex.
Reference.
Associations with autoantibodies in patients with type 1 diabetes
A significant association was found between the TCF7L2 genotypes and ZnT8A positivity (P = 0.0375; Table 3). Also, we noted a significant association between the genotype groups (CT/TT and CC) and higher and lower levels of ZnT8A (P = 0.0484) (data not shown). Regarding the genotype group, there was a significant association with ZnT8A positivity (P = 0.0384; Table S3). Although there was no significant association between IA2A positivity and TCF7L2 genotypes, we found a statistically significant association of TCF7L2 genotypes and genotype groups with lower and higher levels of IA2A (P = 0.0011 and P = 0.0129, respectively). We also examined the distribution of TCF7L2 genotypes (also with genotype groups) between patients with type 1 diabetes with single autoantibody positivity vs ≥2 autoantibody positivity and autoantibody negativity vs autoantibody positivity, but we failed to find any significant differences in this regard (Table 3 and Table S3).
As a next step, we attempted to find out whether TCF7L2, age, and sex have any effect on antibodies in the logistic regression model. First, we considered risk factors for ZnT8A positivity. For patients with C allele (CC/CT group), the likelihood of developing ZnT8A positivity was 5.22 times higher than for patients without C allele (P = 0.0170; Table 5). In the model of genotypes, the CC and CT genotypes were both significant risk factors for ZnT8A positivity (OR 4.69, P = 0.0277 and OR 6.65, P = 0.0115, respectively; Table S5). Older age was a protective factor for ZnT8A positivity in both models (OR 0.93, P < 0.0001 and OR 0.93, P < 0.0001, respectively) after adjusting for age and sex. Next, logistic regression analysis was performed for higher levels of ZnT8A. The CC group was at risk of having higher levels ZnT8A compared with the CT/TT group (OR 2.39, P = 0.0204) after age and sex adjustment (Table 6). There was no significant association with higher level of ZnT8A and genotypes in logistic regression analysis.
Table 5.
Logistic regression analysis evaluating the association of the TCF7L2 genotype groups, age, and sex with ZnT8A positivity at the time of the diagnosis of type 1 diabetes
| Characteristic | Unadjusted OR (95% CI) | P‐value | Adjusted OR † (95% CI) | P‐value |
|---|---|---|---|---|
| Genotype group | ||||
| TT | 1 ‡ | 1 ‡ | ||
| CC + CT | 4.16 (1.17–14.81) | 0.0278 | 5.22 (1.34–20.24) | 0.0170 |
| Age (years) | 0.93 (0.91–0.96) | <0.0001 | 0.93 (0.91–0.96) | <0.0001 |
| Sex | ||||
| Female | 1 ‡ | 1 ‡ | ||
| Male | 1.02 (0.55–1.88) | 0.9605 | 1.29 (0.64–2.61) | 0.4893 |
P‐values in bold are statistically significant.
CI, confidence interval; OR, odds ratio.
Adjusted for age and sex.
Reference.
Table 6.
Logistic regression analysis evaluating the association of the TCF7L2 genotype groups, age and sex with higher level of ZnT8A vs lower level of ZnT8A at the time of the diagnosis of type 1 diabetes
| Characteristic | Unadjusted OR (95% CI) | P‐value | Adjusted OR † (95% CI) | P‐value |
|---|---|---|---|---|
| Genotype group | ||||
| TT + CT | 1 ‡ | 1 ‡ | ||
| CC | 2.19 (1.07–4.50) | 0.0326 | 2.39 (1.14–4.99) | 0.0204 |
| Age (years) | 0.98 (0.96–1.02) | 0.3180 | 0.98 (0.95–1.01) | 0.1855 |
| Sex | ||||
| Male | 1 ‡ | 1 ‡ | ||
| Female | 1.06 (0.54–2.11) | 0.8610 | 1.04 (0.51–2.12) | 0.9147 |
P‐values in bold are statistically significant.
CI, confidence interval; OR, odds ratio.
Adjusted for age and sex.
Reference.
We continued with the assessment of risk factors for a higher level of IA2A, using logistic regression analysis. After adjustment for age and sex, the presence of the CC group was found to be a risk factor against having a higher level of IA2A compared with the CT/TT group (OR 3.51, P = 0.0042; Table 7).
Table 7.
Logistic regression analysis evaluating the association of the TCF7L2 genotype groups, age, and sex with higher level of IA2A vs lower level of IA2A at the time of the diagnosis of type 1 diabetes
| Characteristic | Unadjusted OR (95% CI) | P‐value | Adjusted OR † (95% CI) | P‐value |
|---|---|---|---|---|
| Genotype group | ||||
| TT + CT | 1 ‡ | 1 ‡ | ||
| CC | 3.50 (1.53–8.01) | 0.0326 | 3.51 (1.49–8.27) | 0.0042 |
| Age (years) | 0.99 (0.96–1.02) | 0.5191 | 0.98 (0.95–1.01) | 0.2575 |
| Sex | ||||
| Female | 1 ‡ | 1 ‡ | ||
| Male | 2.28 (1.06–4.92) | 0.0350 | 2.11 (0.94–4.73) | 0.0691 |
P‐values in bold are statistically significant.
CI, confidence interval; OR, odds ratio.
Adjusted for age and sex.
Reference.
DISCUSSION
In this study, our aim was to explore the prevalence of a type 2 diabetes marker, TCF7L2, in patients with type 1 diabetes in the Estonian population and to find out what other parameters could be linked with TCF7L2 rs7903146 polymorphism in these patients.
According to the inclusion criteria for our study group, the patients’ blood samples were taken within 30 days of diagnosis of type 1 diabetes. Using these criteria, we aimed to investigate early changes that occur prior to the diagnosis and attempted to exclude long‐term compensatory mechanisms for changes that occur due to the onset of insulin deficiency.
First, we did not find any association between the TCF7L2 rs7903146 TT genotype and the presence of type 1 diabetes (Table 2). This finding supports previous results obtained for TCF7L2 and type 1 diabetes, showing that the presence of the TCF7L2 rs7903146 TT genotype in type 1 diabetes patients is not related to autoimmunity mediated destruction in the beta‐cells of the pancreas 21 , 22 , 23 , 24 , 25 , 26 .
However, we found that carrying the TCF7L2 risk genotype has an effect on C‐peptide levels in patients diagnosed with type 1 diabetes. We showed that the frequency of patients with type 1 diabetes with random C‐peptide levels of <0.5 nmol/L were higher in carriers of the TCF7L2 rs7903146 C allele. A similar tendency was seen in the case of the genotypes. As C‐peptide levels indicate insulin production capacity and islet dysfunction, we can suggest that carrying the TCF7L2 rs7903146 T allele may result in preservation of beta‐cell function. Bakhtadze et al. have reported that the TCF7L2 rs7903146 CT/TT genotypes are associated with GADA negative diabetes in younger patients, but not in middle‐aged patients, making the risk allele an important indicator of milder islet autoimmunity 27 . However, that study included both autoimmune and non‐autoimmune diabetic patients and did not clearly differentiate between the possible effects of the risk genotypes on different types of diabetes conditions 27 . Nor did our study find a significant association between TCF7L2 and GADA positivity. Furthermore, Redondo et al. reported in several studies that TCF7L2 risk variants were associated with higher C‐peptide levels and single autoantibody positivity, which is consistent with the findings of our study 16 , 28 .
Also, we found an association between the presence of ZnT8A and TCF7L2 genotypes. Apart from patients having the TCF7L2 rs7903146 CC and CT genotypes, patients carrying the TCF7L2 rs7903146 risk genotype TT had less ZnT8A positivity. Therefore, we believe that the association between ZnT8A and TCF7L2 deserves special attention. ZnT8 is a self‐antigen, present in the pancreatic islet cells, which is related to insulin secretion. The more insulin is secreted, the more ZnT8 is expressed on the surface of pancreatic islet cell 29 . According to Zhou et al. 30 the TCF7L2‐regulated transcription network has been described to be connected with SLC30A8 gene encoding ZnT8. Silencing of TCF7L2 gene expression in donor human pancreatic islet cells led to a weaker expression of ZnT8 30 . Moreover, Dwivedi et al. 31 reported recently that the loss of function of ZnT8‐coding gene SLC30A8 is associated with better insulin secretion and protection against type 2 diabetes. It is worth mentioning that the above study used the gene silencing method, but the effects of SNPs of TCF7L2 on ZnT8 expression were not confirmed. Redondo et al. 28 reported previously in a cohort of children that a higher frequency of the TT genotype was linked with single autoantibody positivity and children with single autoantibody positivity were less likely to express ZnT8A. Confirming these findings, we found that carrying the type 2 diabetes risk TCF7L2 genotype (TT) was associated with ZnT8A negativity. Also subjects with T allele (CT/TT) had a lower level of ZnT8A, which may yield a potential protective pathogenetic mechanism in patients with type 1 diabetes, leading to milder islet autoimmunity. It is quite difficult to explain this association, since TCF7L2 may play an important role in different cellular signaling pathways while some of them may have an indirect effect on diabetes. However, to our knowledge, this is the first study that shows the presence of an association contributing to the above mentioned milder islet autoimmunity with the TCF7L2 rs7903146 risk genotype TT. Therefore, these associations need further research.
Additionally, another important association was found between the IA2A and TCF7L2 genotype groups. IA2 autoantigen is located at the insulin secretory granule membrane. It is thought that during granule exocytosis, the cytosolic fragment of IA2 is cleaved to have effects on the nucleus for transcription of granule genes (insulin, IA2) 32 . Its role in the insulin secretory granule has important consequences in the pathogenesis of type 1 diabetes. For instance, the presence of IA2A in first degree relatives can lead to a 50% cumulative risk for the development of type 1 diabetes within 10 years 33 . Moreover, it was previously shown that when IA2A or IAA is present as a single autoantibody in overweight/obese patients, the likelihood of progression to multiple autoantibody positivity in 5 years was 83% greater in TCF7L2 risk T allele carriers 16 . Redondo et al. have shown in several studies that associations of the TCF7L2 risk variant is more frequently established in distinct type 1 diabetes patient groups, such as individuals with fewer islet autoimmunity markers 28 , 34 , 35 . Supporting these findings, our study demonstrates that T allele (CT/TT) is associated with a lower titer of IA2A in type 1 diabetes patients, meaning that it can contribute to milder type 1 diabetes. Whether this is related to a change of the TCF7L2 function in antibody producing cells, thymus or elsewhere requires further investigations. It has been found that TCF7L2 plays an important role in different cells and tissues that are involved in the development of diabetes and changes in glucose metabolism 36 .
Among the limitations of this study is the relatively small sample size, which included 190 type 1 diabetes patients and 246 controls. The non‐match of the controls and the patients according to age and sex could have affected the findings. Additionally, data on fasting C‐peptide levels of the study participants were not available and random C‐peptide levels were used for analysis. The notable strengths of this study were the inclusion criteria and the detailed grouping of the study sample.
In terms of consistency with other studies, it is evident, as associations were found between type 2 diabetes associated gene TCF7L2 and some type 1 diabetes markers, C‐peptide levels, IA2A and ZnT8A, that research should be continued focusing on associations of TCF7L2 at the beta cell level. We believe that the establishment of cellular mechanisms connecting C‐peptide levels with autoantibodies and the TCF7L2 rs7903146 TT genotype would help the scientific community to better understand the nature of type 1 diabetes. In addition, we emphasize the prominent pathogenetic role of TCF7L2 in a wide range of other diseases, as pointed out also in a thorough recent overview by del Bosque‐Plata et al. 36 .
FUNDING
This study was supported by the Estonian Research Council [IUT20‐43 and PRG712].
DISCLOSURE
All authors declare no conflicts of interest.
This study was conducted in accordance with the Declaration of Helsinki.
Approval of the research protocol: The study was approved by the Research Ethics Committee of the University of Tartu (protocols 163/T‐6 from 24.09.2007 and 275/M/15 from 20.11.2017).
Informed consent: Written informed consent was obtained from all participants and controls (in the case of children from their parent).
Registry and the registration no. of the study/trial: N/A.
Animal studies: N/A.
Supporting information
Figure S1 | Median C‐peptide level with the interquartile range (IQR) between the TCF7L2 genotypes of patients with type 1 diabetes.
Table S1 | Associations between the TCF7L2 genotypes and baseline clinical characteristics of the control group.
Table S2 | Associations between the TCF7L2 CC/CT group and TT group and baseline clinical characteristics of the control group.
Table S3 | Associations between the TCF7L2 CC/CT group and TT group and baseline clinical characteristics of patients with type 1 diabetes.
Table S4 | Logistic regression analysis evaluating the association of the TCF7L2 genotypes, age, and sex with a C‐peptide level lower than 0.5 nmol/L, vs higher than 0.5 nmol/L, at the time of diagnosis of type 1 diabetes.
Table S5 | Logistic regression analysis evaluating the association of the TCF7L2 genotypes, age and sex with ZnT8A positivity at the time of diagnosis of type 1 diabetes.
ACKNOWLEDGMENTS
The authors would like to thank Professor Jorma Ilonen from the Immunogenetic Laboratory of the University of Turku, Finland, for performing HLA studies. The authors would like to acknowledge Prof. Margus Lember, Prof. Vallo Tillman, and Prof. Vallo Volke from the University of Tartu and Tartu University Hospital for their institutional support in these studies. We would also like to thank Drs Ingrid Reppo, Kaia Tammiksaar, Svetlana Matjus, Jekaterina Nerman, Lisette Lindam, Olga Gusseva, and Tarvo Rajasalu from Tartu University Hospital and Dr Ülle Einberg from Tallinn Children's Hospital for collecting clinical samples from the patients. In addition, Dr Tamara Vorobjova (University of Tartu), Dr Karin Varik, and Dr Margot Peetsalu from Tartu University Hospital supervised sample collection from the controls.
REFERENCES
- 1. American Diabetes Association . Classification and diagnosis of diabetes. Sec. 2. In Standards of Medical Care in Diabetes—2015. Diabetes Care 2015; 38: S8–S16. [DOI] [PubMed] [Google Scholar]
- 2. Vella A, Matveyenko A. Walking a fine line between ‐cell secretion and proliferation. J Biol Chem 2018; 293: 14190–14191. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Giwa AM, Ahmed R, Omidian Z, et al. Current understandings of the pathogenesis of type 1 diabetes: genetics to environment. World J Diabetes 2020; 11: 13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Felner EI, Klitz W, Ham M, et al. Genetic interaction among three genomic regions creates distinct contributions to early‐ and late‐onset type 1 diabetes mellitus. Pediatr Diabetes 2005; 6: 213–220. [DOI] [PubMed] [Google Scholar]
- 5. Ilonen J, Lempainen J, Veijola R. The heterogeneous pathogenesis of type 1 diabetes mellitus. Nat Rev Endocrinol 2019; 15: 635–650. [DOI] [PubMed] [Google Scholar]
- 6. Nerup J, Platz P, Andersen OO, et al. HL‐A antigens and diabetes mellitus. Lancet 1974; 304: 864–866. [DOI] [PubMed] [Google Scholar]
- 7. Gillespie KM. Type 1 diabetes: pathogenesis and prevention. CMAJ 2006; 175: 165–170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Bakay M, Pandey R, Grant SFA, et al. The genetic contribution to type 1 diabetes. Curr Diab Rep 2019; 19: 116. [DOI] [PubMed] [Google Scholar]
- 9. Ilonen J, Kiviniemi M, Lempainen J, et al. Genetic susceptibility to type 1 diabetes in childhood – estimation of HLA class II associated disease risk and class II effect in various phases of islet autoimmunity. Pediatr Diabetes 2016; 17: 8–16. [DOI] [PubMed] [Google Scholar]
- 10. Billings LK, Florez JC. The genetics of type 2 diabetes: what have we learned from GWAS? Ann N Y Acad Sci 2010; 1212: 59–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Grant SFA. The TCF7L2 locus: a genetic window into the pathogenesis of type 1 and type 2 diabetes. Diabetes Care 2019; 42: 1624–1629. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Weedon MN. The importance of TCF7L2. Diabet Med 2007; 24: 1062–1066. [DOI] [PubMed] [Google Scholar]
- 13. Korinek V, Barker N, Moerer P, et al. Depletion of epithelial stem‐cell compartments in the small intestine of mice lacking Tcf‐4. Nat Genet 1998; 19: 379–383. http://genetics.nature.com. [DOI] [PubMed] [Google Scholar]
- 14. Wenzel J, Rose K, Haghighi EB, et al. Loss of the nuclear Wnt pathway effector TCF7L2 promotes migration and invasion of human colorectal cancer cells. Oncogene 2020; 39: 3893–3909. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Yi F, Brubaker PL, Jin T. TCF‐4 mediates cell type‐specific regulation of proglucagon gene expression by beta‐catenin and glycogen synthase kinase‐3beta. J Biol Chem 2005; 280: 1457–1464. [DOI] [PubMed] [Google Scholar]
- 16. Redondo MJ, Steck AK, Sosenko J, et al. Transcription factor 7‐like 2 (TCF7L2) gene polymorphism and progression from single to multiple autoantibody positivity in individuals at risk for type 1 diabetes. Diabetes Care 2018; 41: 2480–2486. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Berger B, Stenström G, Sundkvist G. Random C‐peptide in the classification of diabetes. Scand J Clin Lab Invest 2000; 60: 687–693. [DOI] [PubMed] [Google Scholar]
- 18. Lampasona V, Pittman DL, Williams AJ, et al. Islet autoantibody standardization program 2018 workshop: interlaboratory comparison of glutamic acid decarboxylase autoantibody assay performance. Clin Chem 2019; 65: 1141–1152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Mikk ML, Kiviniemi M, Laine AP, et al. The HLA‐B*39 allele increases type 1 diabetes risk conferred by HLA‐DRB1*04:04‐DQB1*03:02 and HLA‐DRB1*08‐DQB1*04 class II haplotypes. Hum Immunol 2014; 75: 65–70. [DOI] [PubMed] [Google Scholar]
- 20. Kiviniemi M, Hermann R, Nurmi J, et al. A high‐throughput population screening system for the estimation of genetic risk for type 1 diabetes: an application for the TEDDY (the environmental determinants of diabetes in the young) study. Diabetes Technol Ther 2007; 9: 460–472. [DOI] [PubMed] [Google Scholar]
- 21. Field SF, Howson JMM, Smyth DJ, et al. Analysis of the type 2 diabetes gene, TCF7L2, in 13,795 type 1 diabetes cases and control subjects. Diabetologia 2007; 50: 212–213. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Cauchi S, Vaxillaire M, Choquet H, et al. No major contribution of TCF7L2 sequence variants to maturity onset of diabetes of the young (MODY) or neonatal diabetes mellitus in French white subjects. Diabetologia 2007; 50: 214–216. [DOI] [PubMed] [Google Scholar]
- 23. Qu HQ, Polychronakos C. The TCF7L2 locus and type 1 diabetes. BMC Med Genet 2007; 8: 1–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Cervin C, Lyssenko V, Bakhtadze E, et al. Genetic similarities between latent autoimmune diabetes in adults, type 1 diabetes, and type 2 diabetes. Diabetes 2008; 57: 1433–1437. [DOI] [PubMed] [Google Scholar]
- 25. Raj SM, Howson JMM, Walker NM, et al. No association of multiple type 2 diabetes loci with type 1 diabetes. Diabetologia 2009; 52: 109–116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Howson JMM, Rosinger S, Smyth DJ, et al. Genetic analysis of adult‐onset autoimmune diabetes. Diabetes 2011; 60: 2645–2653. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Bakhtadze E, Cervin C, Lindholm E, et al. Common variants in the TCF7L2 gene help to differentiate autoimmune from non‐autoimmune diabetes in young (15–34 years) but not in middle‐aged (40–59 years) diabetic patients. Diabetologia 2008; 51: 2224–2232. [DOI] [PubMed] [Google Scholar]
- 28. Redondo MJ, Muniz J, Rodriguez LM, et al. Association of TCF7L2 variation with single islet autoantibody expression in children with type 1 diabetes. BMJ Open Diabetes Res Care 2014; 2: e000008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Merriman C, Huang Q, Gu W, et al. A subclass of serum anti‐ZnT8 antibodies directed to the surface of live pancreatic β‐cells. J Biol Chem 2018; 293: 579–587. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Zhou Y, Park S‐Y, Su J, et al. TCF7L2 is a master regulator of insulin production and processing. Hum Mol Genet 2014; 23: 6419–6431. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Dwivedi OP, Lehtovirta M, Hastoy B, et al. Loss of ZnT8 function protects against diabetes by enhanced insulin secretion. Nat Genet 2019; 51: 1596–1606. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Arvan P, Pietropaolo M, Ostrov D, et al. Islet autoantigens: structure, function, localization, and regulation. Cold Spring Harb Perspect Med 2012; 2: a007658. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Morran MP, Casu A, Arena VC, et al. Humoral autoimmunity against the extracellular domain of the neuroendocrine autoantigen IA‐2 heightens the risk of type 1 diabetes. Endocrinology 2010; 151: 2528–2537. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Redondo MJ, Grant SFA, Davis A, et al. Dissecting heterogeneity in paediatric type 1 diabetes: association of TCF7L2 rs7903146 TT and low‐risk human leukocyte antigen (HLA) genotypes. Diabet Med 2017; 34: 286–290. [DOI] [PubMed] [Google Scholar]
- 35. Redondo MJ, Geyer S, Steck AK, et al. TCF7L2 genetic variants contribute to phenotypic heterogeneity of type 1 diabetes. Diabetes Care 2018; 41: 311–317. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. del Bosque‐Plata L, Hernández‐Cortés EP, Gragnoli C. The broad pathogenetic role of TCF7L2 in human diseases beyond type 2 diabetes. J Cell Physiol 2022; 237: 301–312. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1 | Median C‐peptide level with the interquartile range (IQR) between the TCF7L2 genotypes of patients with type 1 diabetes.
Table S1 | Associations between the TCF7L2 genotypes and baseline clinical characteristics of the control group.
Table S2 | Associations between the TCF7L2 CC/CT group and TT group and baseline clinical characteristics of the control group.
Table S3 | Associations between the TCF7L2 CC/CT group and TT group and baseline clinical characteristics of patients with type 1 diabetes.
Table S4 | Logistic regression analysis evaluating the association of the TCF7L2 genotypes, age, and sex with a C‐peptide level lower than 0.5 nmol/L, vs higher than 0.5 nmol/L, at the time of diagnosis of type 1 diabetes.
Table S5 | Logistic regression analysis evaluating the association of the TCF7L2 genotypes, age and sex with ZnT8A positivity at the time of diagnosis of type 1 diabetes.