| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Endocrinology, Barts and the London Queen Mary School of Medicine, University of London (L.B.J., M.O.S., A.J.L.C.), London, United Kingdom EC1A 7BE; Goteborg Pediatric Growth Research Center, Institute for the Health of Women and Children, Sahlgrenska Academy at Goteborg University (J.D., L.G., K.A.W.), 41685 Goteborg, Sweden; and Institut National de la Santé et de la Recherche Médicale, U-457, Hôpital Robert Debré (J.L., P.C.), 75019 Paris, France
Address all correspondence and requests for reprints to: Dr. Linda B. Johnston, Pediatric Endocrine Section, Department of Endocrinology Barts and the London Queen Mary School of Medicine, West Smithfield, London, United Kingdom EC1A 7BE. E-mail: l.b.johnston{at}qmul.ac.uk.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
2 analysis for SNPs and by ANOVA for serum IGF-I levels. IGF-I genotype was associated with the SGA phenotype, in particular with symmetrical SGA and low birth weight, and with IGF-I levels in SGA subjects. Association with postnatal growth was different in the two populations, which may reflect the power of the smaller subphenotype groups. Haplotype analysis in the Swedish short SGA subjects showed that the region of association lay between the promoter and intron 2 of the IGF-I gene. These studies validate the association of the IGF-I gene with birth size and refine the region of association in Swedish short SGA subjects. |
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Perinatal IGF-I levels, animal gene knockout, and transgenic studies demonstrate that in addition to insulin, IGF-I has a major influence on fetal and postnatal growth (12, 13, 14, 15, 16, 17, 18). Complete IGF-I deficiency as a result of IGF-I gene deletion in mice or humans causes severe intrauterine growth failure and postnatal growth failure (14, 15, 16). Recently published association and transmission equilibrium data in Dutch subjects suggest that IGF-I genotype associates with low birth weight and the subphenotype of SGA subjects who are both short and light at birth and have postnatal short stature (19, 20).
We have studied the IGF-I gene in two additional SGA populations, from Haguenau, France, and from Gothenburg, Sweden, in which children with both symmetrical [low birth weight and low birth length (SGALW)] and asymmetrical [low birth weight (SGAW) or low birth length (SGAL)] birth size were studied, and individuals with and without postnatal catch-up growth have been recruited. These populations provide the opportunity to confirm and validate the association of fetal and postnatal growth with IGF-I gene polymorphisms in addition to allowing comparisons between populations.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
To compare populations and investigate the association of IGF-I genotype with birth size, the SGA subjects were redefined as symmetrical (SGALW) or asymmetrical (i.e. either SGAW or SGAL) (24). Although the latter two groups are classed as asymmetrical SGA they were analyzed separately in view of their different phenotypes. All three groups were compared with the normal controls for each population.
Ethical approval was obtained from both centers’ research ethics committees, and individuals consented to the procedures. Auxological findings in subjects from Gothenburg and Haguenau are shown in Tables 1
and 2, respectively.
|
|
One hundred and seventy-four SGA subjects, 114 short SGA and 60 SGA catch-up (gestation, 29–43 wk), were studied. There were 292 appropriate for gestational age (AGA) controls defined by birth size alone, 114 of whom had normal childhood stature. The SGA catch-up and the AGA control groups had height after the age of 2 yr more than -2 SDS. Children with any identifiable syndrome (except Silver Russell syndrome), chronic illness, endocrine disorder, or other identifiable cause of their growth failure were excluded from the study. The short children were recruited through the Goteburg Pediatric Growth Research Center, and the normal AGA controls were either recruited by direct invitation from school children, aged 5–15 yr (25), or were followed yearly as the center control population. The SGA catch-up subjects were recruited from a longitudinal study population (24). Only Caucasians were allowed to participate in this study.
All subjects had serum IGF-I measured in childhood or in late puberty (controls) by RIA, and values were converted to SDS using the normative data established in the laboratory (26). IGF-I SDS and the measured genotypes were analyzed using ANOVA to assess any association. The IGF1/PCR1 and D12S318 markers were analyzed in the controls and the combined SGA group. The two SNPs, IGF1/737.738 and the two proximal haplotypes were analyzed in the controls and the short SGA groups. The dinucleotide repeat markers and haplotypes were reduced to three genotypes (homozygous for the common allele, heterozygous, or no common allele) to maximize the groups’ sizes.
Haguenau subjects
Two hundred and sixty-four term (>37 wk gestation) singletons from the Haguenau population study, collected originally to assess long-term consequences of intrauterine growth failure, were selected for investigation on the basis of DNA availability (27). There were 144 AGA controls and 120 SGA subjects. Direct comparison of this subgroup with those not studied from the original group confirms that there was no significant difference (P > 0.05) in auxological features. The subset was therefore representative of the group. This population had reached final height and was selected on birth size alone, so the majority were SGA catch-up subjects (n = 108), and only 10% were short SGA (n = 12). Ninety-nine percent of the subjects were Caucasian.
Genotyping
Genomic DNA extracted from peripheral blood leukocytes was used for genetic analysis of both dinucleotide repeat microsatellite markers and single nucleotide polymorphisms (SNPs).
Markers were selected on the basis of their genetic position relative to the IGF-I gene (Fig. 1). The intron 2 CT dinucleotide repeat marker (IGF1/PCR1) was studied in both populations (27). The 3' microsatellite marker (D12S318) was studied in the entire Gothenburg population (28). The promoter CA repeat (IGF1/737.738) and two SNPs (IGF1/T-1148C and IGF1/A+1771C) were studied in further analysis of the Gothenburg short SGA subjects. All molecular analysis was undertaken blind to the phenotypic data.
|
Samples were amplified by PCR using fluorescently labeled (HEX or FAM) forward primers. PCRs were carried out in 10-µl reaction mixtures containing 100 ng genomic DNA, 0.1 mM deoxy-NTP, 1.5 mM MgCl2, and 0.25 U Taq DNA polymerase (Sigma-Aldrich, Poole, UK) over 35 cycles with an annealing temperature of 58 C. Samples were then pooled, denatured, and electrophoresed with the ROX 350 fluorescent size marker on an ABI 377 DNA Sequencer (PE Applied Biosystems, Warrington, UK). Results were analyzed using GENESCAN and GENOTYPER software to determine the length of each allele (PE Applied Biosystems). All samples were run in duplicate.
SNP analysis
The IGF1/T-1148C and IGF1/T+1771C SNPs were analyzed using dynamic allele-specific hybridization (DASH; Hybaid, Ashford, UK) (29). This technique involved immobilization of short (50–60 bp), biotinylated target DNA on a streptavidin plate and hybridization of two probes, one complementary to each allele of the SNP. Nonbiotinylated strands were removed by alkali before the buffer, dye, and probe mix was applied. The probe was annealed by heating samples to 85 C and slow cooling to 25 C. The dye/probe mix was removed, discarded, and replaced by 50 µl dye/buffer mix. The SYBR green dye I (Molecular Probes, Inc., Eugene, OR) emitted fluorescence proportional to the amount of probe-target duplex present. The sample was heated from room temperature to 78 C, and the fall in fluorescence as the duplex denatured was measured from 45–78 C. A single base mismatch in the probe significantly altered the temperature at which this denaturation took place. The plates were stripped on completion, and the second probe was applied using the same protocol.
Analysis was performed automatically. All results were checked manually and compared with the results of the complementary probe. Ten percent samples were sequenced, and there was 100% agreement between sequence and DASH genotype data.
DNA sequencing
The results of the DASH experiments were confirmed by sequencing 10% samples. New forward PCR primers (IGF1/T-1148C, GTT GGG CAC ATA GTA GAG C; IGF1/T+1771C, AAA TAA CCC ATT CAT AGC) were used with the nonbiotinylated DASH reverse primer to amplify a fragment that was amenable to sequencing with the Big Dye Terminator sequencing protocol (PE Applied Biosystems).
Statistical methods
Hardy-Weinberg equilibrium.
The asymptotic 2 test was used to analyze the Hardy-Weinberg equilibrium for genetic markers in the control groups where the contingency table had more than five observations in each cell, as was the case with the SNPs studied. This approach was not valid, however, for the more polymorphic dinucleotide repeat markers and the haplotypes that generated larger, sparser contingency tables. Therefore, in the more polymorphic markers the observed and expected frequencies of homozygotes and heterozygotes were compared using a standard 2 x 2 2 test. This method tests whether the observed heterozygosity is significantly different from the expected values calculated from the observed allele frequencies.
Monte Carlo method of association analysis of multiallelic markers (CLUMP).
CLUMP software was used to compare dinucleotide marker allele frequencies between groups and calculate the 2 significance values using Monte Carlo simulations. The probability of the observed marginal totals occurring in a specified number of computer-generated simulations of the contingency table is determined, and the computer reports the number of times the 2 value is achieved by randomly associated data (30). This software generates four different 2 values, the value chosen (T1) is the significance value using the exact test and has n-1 degrees of freedom for each test, where n equals the number of marker alleles found. Sham and Curtis (30) recommend that sufficient simulations should be performed to obtain the real 2 value at least 20 times. Thus, the number of simulations was increased as necessary to fulfill this requirement.
Haplotype analysis. As the parents of these subjects were not recruited, haplotypes could not be identified by descent, i.e. allele transmission from parent to child. Haplotypes identical by state, i.e. the same genomic sequence, were therefore created using the available genotypes where individuals were homozygous for at least one marker. Subjects heterozygous for two markers had to be excluded because it was not possible to determine by PCR of genomic DNA which two of the four possible haplotypes they carried. Association analysis of the haplotypes was performed by CLUMP, and the results stated were generated from the T1 statistic.
Linkage disequilibrium (LD) calculation.
The LD between the two SNP markers studied was estimated using the equation below to calculate r2 (31). This equation is for biallelic loci on the same chromosome with alleles A and a at the first locus and with B and b at the second locus. The allele frequencies of A, B, a, and b individual alleles and the AB haplotype are represented by 
A, B, a, b, and AB, respectively.
 |
r2 value of 0.098 as equivalent to significant LD at the 0.05 level in a sample size of 400. |
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Significant positive association was detected in both populations when IGF-I genotype in SGA subjects was compared with normal birth weight controls (Haguenau IGF1/PCR1, P = 0.018; Gothenburg IGF1/PCR1, P = 0.0041; D12S318, P = 0.00009; Table 3).
|
). Analysis of postnatal growth in SGA subjects
The Haguenau population showed significant association between IGF-I genotype and SGA catch-up phenotype (P = 0.043), with no association with the short SGA group (Table 4). In the Gothenburg population there was significant association with both markers in the short SGA group (IGF1/PCR1, P = 0.001; D12S318, P = 0.046), and no association was found in the smaller SGA catch-up group (Table 4).
|
Three further intragenic markers (promoter SNP IGF1/T-1148C, promoter CA repeat IGF1/737.738, and 3'- untranslated region SNP IGF1/T+1771C) were studied to localize the region of association in these subjects (Fig. 1). The SNP marker genotypes generated by DASH were confirmed by all sequencing reactions performed. The genotypes for all three markers in the control group were found to be in Hardy Weinberg equilibrium. The three markers were not significantly associated on their own (IGF1/T-1148C, P = 0.377; IGF1/737.738, P = 0.123; IGF1/T+1771C, P = 0.139); therefore, haplotype analysis was undertaken.
The two proximal marker combinations were significantly associated. Twelve of the potential 26 IGF1/-1148:IGF1/737.738 haplotypes were observed in a total of 312 chromosomes (from 78 controls and 78 patients). CLUMP analysis revealed association of these haplotypes with the short SGA phenotype (P = 0.04; Table 5). Twenty-one of the potential 30 IGF1/-1148:IGF1/PCR1 haplotypes were observed in a total of 324 chromosomes from 79 controls and 83 patients, and they were significantly associated (P = 0.003) on CLUMP analysis (Table 5).
|
r2 = 0.053), suggesting that these markers are very unlikely to be in LD. There was no association when the two 3' markers IGF1/+1771 and D12S318 were combined to form haplotypes from 71 controls and 66 patients (Table 5). Association of serum IGF-I levels with IGF-I genotypes
Childhood serum IGF-I levels were significantly lower in both the short SGA and the SGA catch-up groups compared with controls (Table 6). There was a significant association between childhood serum IGF-I SDS and IGF1/PCR1 genotype in the SGA group (P = 0.0039). The IGF-I SDS median, mean, and SD were -0.679, -0.455, and 0.888, respectively, for individuals homozygous for the 189 allele; -0.292, -0.250, and 0.666 for the heterozygotes; and -0.250, -0.195, and 0.627 for those with no 189 allele. There was no association of IGF-I levels with this marker among the controls. No other single marker or haplotype was associated with IGF-I SDS in either the control or SGA groups.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In the Swedish and French populations studied here, IGF-I genotype was significantly associated with birth size. The intron 2 marker IGF1/PCR1 was associated with the total SGA groups, the SGALW and SGAW groups, but not consistently with the SGAL groups. The 3' D12S318 microsatellite was significantly associated with all SGA birth size phenotypes in the Swedish population. The SGALW, SGAW, and SGAL group analyses had reduced power, as subphenotyping led to a marked decrease in group numbers and a greater difference in group size compared with controls. Power calculations suggested that 100 subjects in each group were required to have 90% power to detect a 15% difference in allele frequency at the 5% level (36). However, given the consistent significance, these results confirm the IGF-I gene is associated with the SGA phenotype, in particular with the SGALW and SGAW subjects in both populations. This suggests that IGF-I may be more important in determining birth weight than birth length, which would concur with the observed association of IGF-I genotype with fetal macrosomia in diabetic pregnancies (37). However, given the limited power, we cannot exclude an influence of IGF-I on birth length. These results agree with the positive transmission disequilibrium and association studies of the IGF-I gene and birth size in two Dutch populations and are consistent with the fetal insulin hypothesis (19, 38, 39).
In the Swedish population, ANOVA showed that serum IGF-I levels in childhood were significantly lower in SGA subjects homozygous for the IGF1/PCR1 marker common allele, but the promoter marker was not associated. Studies of the promoter IGF1/737.738 marker in British and Dutch type 2 diabetics have shown association with IGF-I levels, as has a study of Australian men with osteoporosis (39, 40, 41). Association of the intron 2 marker genotype with childhood serum IGF-I levels in the SGA group supports the hypothesis that this marker is associated with a functional polymorphism. However, its lack of association in the controls suggests that there are other factors predisposing the SGA subjects to susceptibility at the IGF-I locus, and it is not known what the fetal IGF-I levels were in cases or controls.
Only 10% of the French SGA subjects remained short, and most SGA subjects had shown catch-up growth to reach normal adult heights. In this population there was association with the SGA catch-up group, but the short SGA group (n = 12) was too small. In the Swedish population, both IGF-I gene markers were associated with the short SGA phenotype (D12S318, P = 0.046; IGF1/PCR1, P = 0.001), but not the smaller SGA catch-up group (n = 60). These conflicting results may reflect the difference in group sizes and its consequent effect on power or may possibly relate to the different ages at which catch-up growth was defined. The IGF-I association in the Swedish short SGA group is consistent with results in the Dutch population and the fact that IGF-I is a major growth factor in both fetal and postnatal life in man and other mammals (14, 16, 19).
The IGF-I gene has many polymorphisms in the 5' and 3' regions, but no SNPs have been detected in the main coding exons (42). Many of the SNPs available are not very polymorphic. We have described and studied such an SNP in intron 2 (IGF1/A-23C) in the Haguenau population and found no association with birth size (42). None of the additional markers studied here in the Swedish short SGA group was significantly associated when analyzed individually.
Haplotypes offer increased power to detect genetic association (43). An understanding of the LD across the gene is necessary to generate meaningful haplotypes. However, the pattern of LD across the IGF-I gene has not been reported. In this population there was no significant LD between the IGF1/-1148 in the promoter and the 3'UTR SNP IGF1/+1771 (r2= 0.053), which are separated by a physical distance of about 82 kb (31). Therefore, haplotypes were generated for the 5' and 3' regions separately.
The two 5' haplotypes, one within the promoter/exon 1 region and the other spanning approximately 4 kb across the promoter and first two exons, were associated with the short SGA phenotype. Assuming short distances of LD around each marker, association of the proximal polymorphisms would suggest they influence gene expression or translatability by changing the leader sequences. In the absence of parents, haplotypes could not be identified by descent, and subjects for whom identity by state could not be determined (i.e. subjects were heterozygous for both markers) were excluded from analysis. This may have introduced a selection bias, so these results need to be verified in another population, ideally where parents are available.
Localization of the genetic association using marker haplotypes to 5' region and association with IGF-I levels provides a mechanism for the etiology of fetal growth failure. Further studies are needed to identify the etiological variants and describe their mechanism of influence.
These two studies provide validation for the association of IGF-I genotype with birth size and postnatal growth. All genotypes were generated blind to knowledge of the phenotype using the same technology, making them comparable and suitable for validation purposes (35). The study of the Gothenburg short SGA subjects shows that LD does not extend across the whole IGF-I gene. Analysis of the proximal haplotypes provides greater detail of the genetic association and will be helpful in designing future functional studies. It will be necessary to study this locus further to more clearly delineate the variants that influence fetal and postnatal growth. This strategy will increase understanding of the LD across the IGF-I gene region, which may be of benefit to the investigation of other disorders.
These IGF-I studies add to our understanding of the genetic predisposition to SGA. Larger studies are required to confirm or refute the role of the IGF-I gene in postnatal growth and insulin resistance in these SGA subjects. IGF-I polymorphisms join glucokinase and insulin variants as susceptibility factors for fetal growth failure (44, 45). These three genes all fulfill the requirements of the fetal insulin hypothesis that proposes that fetal genes can predispose to abnormal prenatal growth and susceptibility to insulin resistance in later life (38). Further studies of the interaction of these genotypes with more complex aspects of the SGA phenotype will be intriguing.
|
Footnotes |
|---|
Received April 4, 2003.
Accepted July 15, 2003.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  V. M. Chia, L. C. Sakoda, B. I. Graubard, M. V. Rubertone, S. J. Chanock, R. L. Erickson, and K. A. McGlynn Risk of Testicular Germ Cell Tumors and Polymorphisms in the Insulin-Like Growth Factor Genes Cancer Epidemiol. Biomarkers Prev., March 1, 2008; 17(3): 721 - 726. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. C. Coutinho, R. R. D. Coletta, E. M. F. Costa, P. R. Pachi, M. C. S. Boguszewski, D. Damiani, B. B. Mendonca, I. J. P. Arnhold, and A. A. L. Jorge Polymorphisms Identified in the Upstream Core Polyadenylation Signal of IGF1 Gene Exon 6 Do Not Cause Pre- and Postnatal Growth Impairment J. Clin. Endocrinol. Metab., December 1, 2007; 92(12): 4889 - 4892. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. S Schernhammer, S. S Tworoger, A H. Eliassen, S. A Missmer, J. M Holly, M. N Pollak, and S. E Hankinson Body shape throughout life and correlations with IGFs and GH Endocr. Relat. Cancer, September 1, 2007; 14(3): 721 - 732. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. M. N. Bannink, J. van Doorn, P. G. H. Mulder, and A. C. S. Hokken-Koelega Free/Dissociable Insulin-Like Growth Factor (IGF)-I, Not Total IGF-I, Correlates with Growth Response during Growth Hormone Treatment in Children Born Small for Gestational Age J. Clin. Endocrinol. Metab., August 1, 2007; 92(8): 2992 - 3000. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. B. Dunger, C. J. Petry, and K. K. Ong Genetics of Size at Birth Diabetes Care, July 1, 2007; 30(Supplement_2): S150 - S155. [Full Text] [PDF]
|
|
|
|
 |
|
 P. Saenger, P. Czernichow, I. Hughes, and E. O. Reiter Small for Gestational Age: Short Stature and Beyond Endocr. Rev., April 1, 2007; 28(2): 219 - 251. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Chellakooty, A. Juul, K. A. Boisen, I. N. Damgaard, C. M. Kai, I. M. Schmidt, J. H. Petersen, N. E. Skakkebaek, and K. M. Main A Prospective Study of Serum Insulin-Like Growth Factor I (IGF-I) and IGF-Binding Protein-3 in 942 Healthy Infants: Associations with Birth Weight, Gender, Growth Velocity, and Breastfeeding J. Clin. Endocrinol. Metab., March 1, 2006; 91(3): 820 - 826. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Balogh, A. Treszl, A. Vannay, and B. Vasarhelyi A Prevalent Functional Polymorphism of Insulin-Like Growth Factor System Is Not Associated With Perinatal Complications in Preterm Infants Pediatrics, February 1, 2006; 117(2): 591 - 592. [Full Text] [PDF]
|
|
|
|
|
|
R. Verkauskiene, D. Jaquet, S. Deghmoun, D. Chevenne, P. Czernichow, and C. Levy-Marchal Smallness for Gestational Age Is Associated with Persistent Change in Insulin-Like Growth Factor I (IGF-I) and the Ratio of IGF-I/IGF-Binding Protein-3 in Adulthood J. Clin. Endocrinol. Metab., October 1, 2005; 90(10): 5672 - 5676. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S Mohan and D J Baylink Impaired skeletal growth in mice with haploinsufficiency of IGF-I: genetic evidence that differences in IGF-I expression could contribute to peak bone mineral density differences J. Endocrinol., June 1, 2005; 185(3): 415 - 420. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. B. Terry, B. L. De Stavola, I. dos Santos Silva, M. J. Wadsworth, and M. Ahlgren Birth Weight and Breast Cancer N. Engl. J. Med., January 20, 2005; 352(3): 304 - 306. [Full Text] [PDF]
|
|
|
|
 |
|
 A R Hayati, F C Cheah, J F Yong, A E Tan, and W M Norizah The role of serum insulin-like growth factor I (IGF-I) in neonatal outcome J. Clin. Pathol., December 1, 2004; 57(12): 1299 - 1301. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
