Advertisement

Diabetologia

, Volume 51, Issue 5, pp 811–815 | Cite as

Association analysis of the IGF1 gene with childhood growth, IGF-1 concentrations and type 1 diabetes

  • A. Vella
  • N. Bouatia-Naji
  • B. Heude
  • J. D. Cooper
  • C. E. Lowe
  • C. Petry
  • S. M. Ring
  • D. B. Dunger
  • J. A. Todd
  • K. K. OngEmail author
Open Access
Short Communication

Abstract

Aims/hypothesis

Insulin-like growth factor-1 is a major childhood growth factor and promotes pancreatic islet cell survival and growth in vitro. We hypothesised that genetic variation in IGF1 might be associated with childhood growth, glucose metabolism and type 1 diabetes risk. We therefore examined the association between common genetic variation in IGF1 and predisposition to type 1 diabetes, childhood growth and metabolism.

Materials and methods

Variants in IGF1 were identified by direct resequencing of the exons, exon–intron boundaries and 5′ and 3′ regions in 32 unrelated type 1 diabetes patients. A tagging subset of these variants was genotyped in a collection of type 1 diabetes families (3,121 parent–child trios). We also genotyped a previously reported CA repeat in the region 5′ to IGF1. A subset of seven tag single nucleotide polymorphism (SNPs) that captured variants with minor allele frequency (MAF) ≥0.05 was genotyped in 902 children from the Avon Longitudinal Study of Parents And Children with data on growth, IGF-1 concentrations, insulin secretion and insulin action.

Results

Resequencing detected 27 SNPs in IGF1, of which 11 had a MAF > 0.05 and were novel. Variants with MAF ≥ 0.10 were captured by a set of four tag-SNPs. These SNPs showed no association with type 1 diabetes. In children, global variation in IGF1 was weakly associated with IGF-1 concentrations, but not with other phenotypes. The CA repeat in the region 5′ to IGF1 showed no association with any phenotype.

Conclusions/interpretation

Common genetic variation in IGF1 alters IGF-1 concentrations but is not associated with growth, glucose metabolism or type 1 diabetes.

Keywords

Children Genetic variation Growth IGF-1 Insulin-like growth factor-1 Microsatellite Single nucleotide polymorphism SNP Tag-SNPs Type 1 diabetes 

Abbreviations

ALSPAC

Avon Longitudinal Study of Parents And Children

MAF

minor allele frequency

SNP

single nucleotide polymorphism

WTCCC

Wellcome Trust Case Control Consortium

Introduction

Type 1 diabetes is a common autoimmune disorder that arises by an interaction between genes and the environment. To date, ten loci have been identified including the HLA class II genes, the cytotoxic T-lymphocyte associated protein 4 (CTLA4) locus, protein tyrosine phosphatase, non-receptor type 22 (lymphoid) (PTPN22), IL-2 receptor, alpha (IL2RA), interferon induced with helicase C domain 1 (IFIH1) and four novel loci identified by a genome-wide association study [1].

The protein products of these genes play important roles in antigen presentation and the cellular immune response, highlighting the importance of the immune system in the pathogenesis of type 1 diabetes. However, not all patients with evidence of islet autoimmunity develop complete beta cell failure [2]. Given the evidence to suggest that early weight gain and growth in childhood are associated with type 1 diabetes [3], it is possible that factors influencing insulin action (body weight and fat mass) and islet function may predispose to or influence the presentation of type 1 diabetes.

Lower IGF-1 concentrations predict the development of glucose intolerance in adults [4]. George et al. reported that islet overexpression of IGF1 in diabetic mice enabled islet regeneration and gradual correction of hyperglycaemia and hypoinsulinaemia [5]. A microsatellite in the region 5′ to IGF1 has been associated with adult height, cardiovascular risk, osteoporosis and type 2 diabetes in some [6] but not all [7] studies.

Since common variation in IGF1 could alter circulating IGF-1 concentrations and therefore body habitus, insulin action and beta cell secretion, we sought to determine the associations, if any, between IGF1 and circulating IGF-1 concentrations, size at birth, both fasting and postprandial insulin concentrations, and type 1 diabetes.

Materials and methods

Polymorphism identification

DNA samples from 32 randomly selected type 1 diabetic probands were amplified using specifically designed forward and reverse primers. This provided 88% probability of detecting single nucleotide polymorphisms (SNPs) with minor allele frequencies (MAF) of 0.033, 96% probability for MAF 0.05 and 99.8% for MAF 0.10 [8]. Resequencing included the 3 kb region 5′ to the gene, all exons and exon–intron boundaries and 3′ untranslated region. Four 1 kb segments in the second intron, ~10 kb intervals apart, were also resequenced.

Tag-SNP selection

The resequencing genotype data were used to select SNP subsets that predicted the genotypes of the remainder, using a coefficient of determination, R 2, which measures the ability to predict each known SNP genotype by linear regression on the tag-SNP genotypes [9]. We considered only SNPs with a MAF ≥ 0.10 in the type 1 diabetes collection and ≥0.05 in the Avon Longitudinal Study of Parents And Children (ALSPAC) cohort (for details see below and Electronic supplementary material [ESM]), using a minimum R 2 of 0.8.

Genotyping

All genotyping data were double-scored to minimise error. All genotypes were in Hardy–Weinberg equilibrium (p > 0.05).

Populations studied

All DNA samples were collected after ethics approval and informed consent had been obtained.

Type 1 diabetes family collection

Type 1 diabetes families were of white European descent, with two parents and at least one affected child. The populations studied have been described previously [8].

ALSPAC

Details of this birth cohort are available on the ALSPAC website (www.alspac.bris.ac.uk). The children in this study are from a 10% ‘Children in Focus’ sub-cohort (1,335 full-term singleton infants) randomly selected from the last 6 months of recruitment for more detailed measurements of growth. Birthweight was noted from hospital records, and length and head circumference were measured after birth. At age 7 years (mean age 7.5 ± 0.1 years), body weight and height were measured. Body composition was assessed at age 9 years by whole-body dual-energy X-ray absorptiometry. Internal SD scores were calculated for all parameters of growth to adjust for age and sex. IGF-1 concentrations were measured in cord blood samples (birth) and in venous blood (7 or 8 years of age). Fasting insulin sensitivity was assessed by the homeostatic model assessment index (www.dtu.ox.ac.uk/homa/) and insulin secretion at 30 min post oral glucose by the insulinogenic index at age 8 years.

Statistical analysis

All statistical analyses to test the association between IGF1 and type 1 diabetes were performed in either Stata (www.stata.com/) or R (www.r-project.org/) statistical systems. Additional routines may be downloaded (www-gene.cimr.cam.ac.uk/clayton/software/). Missing tag-SNP genotypes were imputed under the null hypothesis and were analysed using a multilocus test [9]. The microsatellite, 5′ IGF1 CA repeat, was analysed using TRANSMIT [10]. The global effect of the IGF1 tag-SNP set on each outcome variable was entered into a multi-locus regression model. R 2 change was taken as the contribution to the total variation in each outcome. Associations of IGF1 SNPs with growth and metabolic phenotypes were analysed using univariate ANOVA (general linear models). The association of the 5′ IGF1 CA repeat was analysed by comparing the wild-type allele (192 bp) to all other alleles [6, 7].

For further details on Materials and methods section, see ESM.

Results

Genetic variation in IGF1

Resequencing identified 27 novel polymorphisms consisting of 24 SNPs, one rare non-synonymous SNP in exon 3 (Ala to Thr) and two deletion/insertion polymorphisms. Eleven SNPs had a MAF > 0.05. Four tag-SNPs (Table 1) were selected and genotyped in the type 1 diabetes families studied. In ALSPAC, seven tag-SNPs were selected (ESM Table 1).
Table 1

A summary of IGF1 tag-SNPs genotyped in 2,396 families

Parameter

Tag-SNP

rs35140968a

rs6214

rs3730220

s6219

Sequence

AAATAACT[-/AT]CTCAAATA

GACTTAAC[A/G]TGTTTTCT

CAGGTTGG[A/T]CTCAAACT

AACCTCAA[A/G]CTGTCTAC

Allele coding, 1 (minor)/2

Ins/Del

A/G

T/A

A/G

MAF in parents

0.277

0.443

0.493

0.0769

Minor allele transmissions

Transmitted

815

1309

915

446

Untransmittedb

790

1386

942

446

Genotype 2/2

Transmitted

1,150 (53.1)

947 (33.3)

798 (33.0)

2,501 (84.8)

Untransmittedb

3,478 (53.6)

2,724 (31.9)

2,352 (32.4

7,489 (84.6)

Genotype 2/1

Transmitted

849 (39.2)

1,382 (48.5)

885 (36.6)

423 (14.3)

Untransmittedb

2,541 (39.1)

4,226 (49.5)

2,685 (37.0)

1,297 (14.7)

Genotype 1/1

Transmitted

166 (7.7)

519 (18.2)

737 (30.5)

27 (0.91)

Untransmittedb

476 (7.3)

1,594 (18.7)

2,223 (30.6)

67 (0.8)

OR for minor allele (95% CI)

1.03 (0.93–1.14)

0.94 (0.87–1.01)

0.97 (0.89–1.06)

1.00 (0.88–1.14)

aDeletion/insertion polymorphism

bUntransmitted (pseudocontrol) genotypes are estimated as previously described [8]

Multilocus test p = 0.425

IGF1 and type 1 diabetes

The selected tag-SNPs were genotyped in 2,439 families (3,121 parent–child trios). The multilocus test of the IGF1 tag-SNPs provided no evidence of an association with type 1 diabetes (χ 4 2 = 4.4; p = 0.356; Table 1). The microsatellite, 5′ IGF1 CA repeat, was genotyped in 2,109 families and analysed using TRANSMIT [10]. There was no evidence of association (p = 0.358) (ESM Table 2).

Wellcome Trust Case Control Consortium data

We used data from the Wellcome Trust Case Control Consortium (WTCCC) genome-wide association study [1] to test association of the extended IGF1 region with type 1 diabetes. The region contained 22 SNPs. When the WTCCC data were analysed using a logistic regression model adjusted for variation in allele frequencies across Great Britain, no evidence of association of these SNPs within the region was found (ESM Table 3).

Association with IGF-1 concentrations

In the ALSPAC children, multilocus regression models showed that the IGF1 tag-SNP set was associated with IGF-1 protein levels at birth (R 2 = 0.063; p = 0.029) and weakly associated at age 7 to 8 years (R 2 = 0.030; p = 0.055; ESM Table 4). Results of IGF-1 protein level associations with individual SNPs are shown in Table 2. However, the multilocus IGF1 tag-SNP set did not associate with any other childhood growth or metabolic phenotype (ESM Table 4). No associations were observed between the IGF1 CA repeat and IGF-1 concentrations or any growth or metabolic phenotype (ESM Table 5).
Table 2

IGF-1 protein levels (ng/ml) at birth and at age 7 or 8 years by genotypes of seven IGF1 tag-SNPs in a representative birth cohort

Parameter

n

Genotype (%)

p value

rs35140968

 

I/D (58)

I/I (35)

D/D (7)

 

 At birth

315

88.1 (2.8)

91.1 (3.6)

105.0 (8.9)

0.179

 7 or 8 years

547

147.8 (2.9)

147.3 (3.9)

157.4 (9.1)

0.581

rs6214

 

G/G (34)

G/A (50)

A/A (16)

 

 At birth

305

88.6 (3.6)

87.7 (3.2)

102.43 (5.30)

0.048

 7 or 8 years

556

144.1 (4.0)

148.9 (3.1)

155.83 (5.70)

0.239

rs3730220

 

A/A (38)

A/T (54)

T/T (9)

 

 Birth

299

94.1 (3.8)

88.6 (2.9)

89.6 (7.4)

0.509

 7 or 8 years

548

152.0 (3.8)

145.7 (3.0)

141.8 (8.1)

0.331

rs6219

 

G/G (84)

*/A (16)

 

 Birth

300

91.0 (2.4)

85.0 (5.7)

0.329

 7 or 8 years

547

150.1 (2.5)

138.8 (5.8)

0.072

rs2946831

 

T/T (89)

*/G (11)

 

 Birth

294

90.0 (2.3)

95.2 (7.2)

0.501

 7 or 8 years

541

148.0 (2.4)

153.9 (7.1)

0.432

rs3730204

 

T/T (97)

*/C (3)

 

 Birth

294

111.3 (13.5)

90.3 (2.3)

0.126

 7 or 8 years

548

173.8 (12.9)

147.5 (2.3)

0.046

rs12579108

 

C/C (97)

*/A (3)

 

 Birth

301

90.0 (2.2)

101.4 (12.1)

0.358

 7 or 8 years

554

147.6 (2.3)

176.8 (15.9)

0.017

Values are mean (SE) for IGF-1 concentration and per cent for genotype

*Additive models, adjusted for age and sex

Discussion

We used a UK birth cohort, ALSPAC, and a large family collection of type 1 diabetes to explore association with a subset of SNPs generated by in-depth resequencing of the IGF1 locus. A subset of SNPs that effectively predicted the genotype of the remainder was subsequently genotyped in these cohorts [9]. We report no evidence to support the hypothesis that genetic variation in IGF1 is associated with major susceptibility to type 1 diabetes. However, genetic variation in this locus may modestly influence circulating IGF-1 concentrations, at least at birth and during childhood.

Prior studies examining the role of genetic variation in IGF1 in predisposition to common human disease have focused on a microsatellite in the region 5′ to IGF1. This variant has been inconsistently associated with adult height and type 2 diabetes [6, 7]. Perplexingly, the same allele has been associated with high [6] and low [7] IGF-1 concentrations. It is possible that other common genetic variants, in different degrees of linkage disequilibrium with this variant, alter IGF-1 concentrations and explain the discrepancy in the published studies.

We found no association between this microsatellite and circulating IGF-1 concentrations, growth or type 1 diabetes. To exclude the possibility that other polymorphisms in the gene predispose to disease, we undertook a systematic analysis of IGF1 gene variation using tagging SNPs [8]. Our relatively large white type 1 diabetes family collection of European ancestry provided ~84% power to detect a causal allele for type 1 diabetes predisposition with MAF = 0.1 and OR of 1.2 at the 5% significance level. The study was therefore adequately powered to detect weak predisposition to type 1 diabetes conferred by IGF1. We did not directly examine whether these variants alter the age at diagnosis of type 1 diabetes, since there was no primary evidence of association.

The possibility remains that variation in IGF1 could alter insulin secretion or action as well as childhood growth. Therefore, to examine these associations, we adopted the same strategy in an established birth cohort. While we demonstrate that common genetic variation in IGF1 modestly influences circulating IGF-1 concentrations at birth and during childhood, we did not find a major contribution of IGF1 to birthweight, growth or response to oral glucose. It remains possible that maternal IGF1 genotypes that influence maternal metabolism might alter birth size, but they are unlikely to significantly impact on childhood growth.

In summary, this large systematic study shows that common IGF1 variants modestly influence circulating IGF-1 concentrations at birth and during childhood. However, these variants are not associated with birthweight, childhood growth and insulin secretion or action. Moreover they do not alter type 1 diabetes risk.

Notes

Acknowledgements

We gratefully acknowledge the participation of all of the type 1 diabetic patients and family members. We also thank the Human Biological Data Interchange and Diabetes UK for USA and UK multiplex families, respectively, as well as C. Ionescu-Tirgoviste and C. Guja for the Romanian families, D. Undlien for the Norwegian families and D. Savage for the Northern Ireland families. Thanks also to GET1FIN (J. Tuomilehto, L. Kinnunen, E. Tuomilehto-Wolf, V. Harjutsalo and T. Valle) for the Finnish families and to the Academy of Finland, the Sigrid Juselius Foundation and the JDRF for funding. We are extremely grateful to all the families who took part in the ALSPAC study, the midwives for their help in recruiting them and the whole ALSPAC team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists and nurses. The UK Medical Research Council, the Wellcome Trust and the University of Bristol provide core support for ALSPAC. We thank H. Rance, T. Hassanali, J. Hutchings, G. Coleman, S. Field, T. Mistry, K. Bourget, S. Clayton, M. Hardy, J. Keylock, P. Lauder, M. Maisuria, W. Meadows, M. Sebastian and S. Wood for preparing DNA samples. This study makes use of data generated by the WTCCC. A full list of the investigators who contributed to the generation of those data is available from www.wtccc.org.uk. Funding for the project was provided by the Wellcome Trust under award 076113. A. Vella is supported by the Mayo Clinic and by the National Institutes of Health DK078646-01. B. Heude was funded by the Danone Institute.

Duality of interest

The authors declare that there is no duality of interest associated with this manuscript.

Open Access

This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.

Supplementary material

125_2008_970_MOESM1_ESM.pdf (16 kb)
ESM Table 1 IGF1 SNPs genotyped in ALSPAC children (PDF 16.2 KB)
125_2008_970_MOESM2_ESM.pdf (63 kb)
ESM Table 2 TRANSMIT analysis for association of the microsatellite in the 5′ region of IGF1 with type 1 diabetes (PDF 62.5 KB)
125_2008_970_MOESM3_ESM.pdf (29 kb)
ESM Table 3 Summary of 23 WTCCC SNPs in the extended IGF1 region (PDF 28.6 KB)
125_2008_970_MOESM4_ESM.pdf (17 kb)
ESM Table 4 Multilocus test for associations with IGF-1 protein levels, childhood growth and glucose metabolism in a representative birth cohort (PDF 17.1 KB)
125_2008_970_MOESM5_ESM.pdf (72 kb)
ESM Table 5 Association testing for the 5′ IGF1 CA repeat with childhood growth and glucose metabolism in a representative birth cohort (PDF 72.3 KB)
125_2008_970_MOESM6_ESM.pdf (81 kb)
ESM Table 6 ESM Text (PDF 81.3 KB)

References

  1. 1.
    Todd JA, Walker NM, Cooper JD et al (2007) Robust associations of four new chromosome regions from genome-wide analyses of type 1 diabetes. Nat Genet 39:857–864PubMedCrossRefGoogle Scholar
  2. 2.
    Borg H, Gottsater A, Fernlund P, Sundkvist G (2002) A 12-year prospective study of the relationship between islet antibodies and beta-cell function at and after the diagnosis in patients with adult-onset diabetes. Diabetes 51:1754–1762PubMedCrossRefGoogle Scholar
  3. 3.
    Johansson C, Samuelsson U, Ludvigsson J (1994) A high weight gain early in life is associated with an increased risk of type 1 (insulin-dependent) diabetes mellitus. Diabetologia 37:91–94PubMedCrossRefGoogle Scholar
  4. 4.
    Yakar S, Wu Y, Setser J, Rosen CJ (2002) The role of circulating IGF-I: lessons from human and animal models. Endocrine 19:239–248PubMedCrossRefGoogle Scholar
  5. 5.
    George M, Ayuso E, Casellas A, Costa C, Devedjian JC, Bosch F (2002) Beta cell expression of IGF-I leads to recovery from type 1 diabetes. J Clin Invest 109:1153–1163PubMedGoogle Scholar
  6. 6.
    Vaessen N, Heutink P, Janssen JA et al (2001) A polymorphism in the gene for IGF-I: functional properties and risk for type 2 diabetes and myocardial infarction. Diabetes 50:637–642PubMedCrossRefGoogle Scholar
  7. 7.
    Frayling TM, Hattersley AT, Smith GD, Ben-Shlomo Y (2002) Conflicting results on variation in the IGFI gene highlight methodological considerations in the design of genetic association studies. Diabetologia 45:1605–1606PubMedCrossRefGoogle Scholar
  8. 8.
    Vella A, Howson JM, Barratt BJ et al (2004) Lack of association of the Ala(45)Thr polymorphism and other common variants of the NeuroD gene with type 1 diabetes. Diabetes 53:1158–1161PubMedCrossRefGoogle Scholar
  9. 9.
    Vella A, Cooper JD, Lowe CE et al (2005) Localization of a type 1 diabetes locus in the IL2RA/CD25 region by use of tag single-nucleotide polymorphisms. Am J Hum Genet 76:773–779PubMedCrossRefGoogle Scholar
  10. 10.
    Clayton D (1999) A generalization of the transmission/disequilibrium test for uncertain-haplotype transmission. Am J Hum Genet 65:1170–1177PubMedCrossRefGoogle Scholar

Copyright information

© The Author(s) 2008

Authors and Affiliations

  • A. Vella
    • 1
    • 2
  • N. Bouatia-Naji
    • 3
  • B. Heude
    • 3
    • 4
    • 5
  • J. D. Cooper
    • 1
  • C. E. Lowe
    • 1
  • C. Petry
    • 6
  • S. M. Ring
    • 7
  • D. B. Dunger
    • 6
  • J. A. Todd
    • 1
  • K. K. Ong
    • 3
    • 6
    Email author
  1. 1.Juvenile Diabetes Research Foundation/Wellcome Trust Diabetes and Inflammation Laboratory, Department of Medical Genetics, Cambridge Institute for Medical ResearchUniversity of CambridgeCambridgeUK
  2. 2.Mayo ClinicRochesterUSA
  3. 3.MRC Epidemiology UnitInstitute of Metabolic ScienceCambridgeUK
  4. 4.INSERM Unit 780, IFR69VillejuifFrance
  5. 5.Faculty of MedicineUniversity Paris-SudOrsayFrance
  6. 6.Department of PaediatricsUniversity of CambridgeCambridgeUK
  7. 7.Department of Community-Based MedicineUniversity of BristolBristolUK

Personalised recommendations