| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
McGill University-Montreal Children’s Hospital Research Institute and the Department of Pediatrics, Division of Endocrinology, McGill University (P.V., R.G., C.P.), Montreal, Quebec, Canada H3H 1P3; and the Wellcome Trust Center for Human Genetics, Nuffield Department of Surgery, University of Oxford (S.T.B., J.A.T.), Oxford, United Kingdom OX3 7BN
Address all correspondence and requests for reprints to: Constantin Polychronakos, M.D., F.R.C.P.(C), Montreal Children’s Hospital, 2300 Tupper Street, Montreal, Quebec, Canada H3H 1P3. E-mail: mc97{at}musica.mcgill.ca
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). Class I VNTR alleles (26–63 repeat
units) predispose to type 1 diabetes, whereas class III VNTR alleles
(140–210 repeat units) are dominantly protective (14). Within class
III, there are two size modes identifiable by linkage disequilibrium
with neighboring polymorphisms, one of which is more protective [very
protective haplotype (VPH)] than the other [protective haplotype
(PH)].
|
A functional effect of VNTR alleles on INS does not preclude the possibility that other genes, such as IGF2, are also targets of the IDDM2 effect. It has been hypothesized that the IDDM2 target gene is IGF2 (18, 19). Several studies report parent of origin effects in the genetics of the IDDM2 locus (3, 8, 14, 20, 21, 22, 23). This suggests involvement of a gene subject to parental imprinting, the phenomenon by which a gene is expressed only or mostly from the gene copy inherited from the parent of a specific sex. IGF2 is an imprinted gene with exclusive expression from the paternal allele in most tissues (24). In leukocytes, there is a genotype-dependent variable relaxation of IGF2 imprinting (25), with expression of the maternal gene copy ranging from complete repression to expression equal to that of the paternal copy (15). This variable reactivation of the maternal allele in leukocytes could result in a biologic effect of maternal haplotypes on expression levels of IGF2. Depending on haplotype frequencies in different populations, such an effect could appear stronger or weaker than the paternal one in genetic studies of end-point biologic effects such as diabetes susceptibility.
A role for IGF2 in type 1 diabetes has also been proposed based on its biological properties as a mitogenic peptide, active in thymus, pancreas, and activated lymphocytes (reviewed in Ref. 19) and its potential role as a selecting peptide in the thymus, which may help induce immune tolerance to proinsulin, with which it shares significant homology (18, 26). In addition, there are less direct mechanisms by which IGF2 could affect susceptibility. For example, class I VNTR alleles are associated with somewhat increased IGF2 expression in placenta (27), which may affect intrauterine growth and birth size. Birth weight and early growth patterns were found to be associated with type 1 diabetes risk in some studies (28, 29, 30, 31).
The three tissues known to be critical to the pathogenesis of type 1 diabetes are thymus, where immune tolerance develops, pancreas, where the autoimmune assault occurs, and leukocytes, which mediate ß-cell destruction. Evaluation of the effects of VNTR on IGF mRNA levels in thymus and pancreas revealed no differential effect of VNTR allele class (32). We now focus on leukocytes.
T lymphocytes and their thymic precursors express both IGF-II and the type I IGF receptor (19, 33). Transgenic mice overexpressing IGF2 in the thymus under the class II MHC promoter had a large increase in T lymphocyte number (34, 35). The same effect could not be achieved by IGF-II administration, and despite plasma IGF-II levels that were 2–3 times higher in transgenic mice than in wild-type, only tissues expressing the IGF2 transgene had increased growth. Autocrine and paracrine signaling mechanisms play a major role in the biological action of the IGF-II peptide. Thus, alleles that increase IGF2 mRNA levels would increase local IGF-II levels and therefore increase the survival of lymphocytes and their resistance to apoptosis, including potentially autoreactive and activated lymphocytes. Such alleles would be predisposing to type 1 diabetes.
In this report, we present studies in leukocytes showing a divergence between genetic determinants of IGF2 expression in leukocytes, on the one hand, and diabetes, on the other; the VNTR allele classes determining diabetes susceptibility have no differential effects on IGF2 expression levels, whereas an ApaI restriction fragment length polymorphism (RFLP) in the 3'-untranslated region (3'-UTR) of IGF2, a polymorphism strongly associated with IGF2 expression levels, shows no association with type 1 diabetes.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Peripheral blood samples were obtained from type 1 diabetes patients and their parents from the Montreal Children’s Hospital diabetic clinic, with signed informed consent and approval by the institutional review board of the Montreal Children’s Hospital and the Quebec Ministry of Health. Nucleic acids were purified by phenol-chloroform extraction. Lymphocytes were separated by Ficoll-Hypaque (Pharmacia, Piscataway, NJ) and stimulated with 5 mg/mL concanavalin A for 48 h before collection.
Genotyping
All samples were genotyped for VNTR status as described previously (8, 14, 15, 16). Class I alleles are designated by size in mobility units (8) or repeat number (36), following previously established terminology. Stable transmission of the VNTR alleles was confirmed by typing available parental DNA.
Amplification of a 236-bp segment of the IGF2 3'-UTR containing an ApaI RFLP common to transcripts from all four promoters was carried out as previously described (37). The [32P]deoxy-ATP internally labeled PCR product was digested with ApaI and electrophoresed in a polyacrylamide gel. Heterozygous samples were identified as those with a 236-bp undigested band and two digestion fragments of 173 and 63 bp. Complete digestion was ensured by the inclusion of a sample homozygous for the ApaI cutting site with all digest runs.
Linkage disequilibrium measurements
The extent of linkage disequilibrium in the Canadian-UK data set (n = 676) is described by the coefficient of linkage disequilibrium D' (38, 39) calculated as follows: D' = D/Dmax with D = pab - papb, where pab = frequency of haplotypes with allele a at one locus and allele b at another locus, pa = frequency of allele a, pb = frequency of allele b, and Dmax = pa(1 - pb) is the maximum linkage disequilibrium possible.
Quantitation of IGF2 in leukocytes and lymphocytes
We compared the relative expression levels of the two IGF2 gene copies within each sample, using a previously described method (15). Briefly, for VNTR and ApaI we determined the VNTR/ApaI haplotypes and their parental origin in the child. Individuals who were VNTR class I/III and ApaI +/- double heterozygotes, with haplotypes of known parental origin, were selected for further study. To distinguish the transcripts originating from each chromosome, RNA was reverse transcribed to complementary DNA, and the ApaI fragment was amplified by PCR. The PCR products, labeled internally with [32P]deoxy-ATP, were digested with ApaI, resolved by PAGE, and quantified using a phosphorimager (Fuji, Tokyo, Japan) to obtain a ratio of the 173-bp band representing the digested ApaI(+) allele to the 236-bp undigested band representing the ApaI(-) allele. The same procedure was performed for DNA, where the two alleles are known to be present in equal abundance. Each ratio shown is the average of two to seven separate measurements. If the two alleles are present in equal abundance, then the RNA ratio will equal the DNA ratio. Any differences between the two ratios represent a difference in the relative abundance of the mRNA of the two alleles. To control for the imprinting effect, the ratio of the paternal over maternal allele was compared as follows: RNA/DNA = [ApaI(+)/ApaI(-) RNA ratio]/[ApaI(+)/ApaI(-) DNA ratio] (Eq. 1). If ApaI(+) is the paternal allele, take the ratio from Eq. 1 as: RNA/DNA (Eq. 2A). If ApaI(-) is the paternal allele, take the ratio from Eq. 1 as: 1/(RNA/DNA) (Eq. 2B).
Absolute IGF2 mRNA levels in leukocytes were quantified using a quantitative competitive RT-PCR method described in detail previously (32). Each subject was assayed from one to three separate blood samples drawn on different days. Each sample was assayed using one to four different competitor concentrations. Some samples were assayed in duplicate using the same RNA sample in two different RT reactions and assaying both reactions in parallel. As both IGF2 alleles are expressed in this tissue, the absolute quantitation represents the sum of the transcripts from both alleles. From the ratio of paternal/maternal allele levels in each RNA sample, we were then able to calculate which fraction of the total IGF2 mRNA came from the chromosome inherited from each parent.
mRNA stability study
To determine whether the effect of the ApaI polymorphism on IGF2 mRNA levels is due to differential RNA stability, lymphocytes from an ApaI(±) individual known to equally express both copies of IGF2 were stimulated for 72 h with concanavalin A, followed by addition of the transcription blocker actinomycin D. The cells were collected at various time points, and RNA was extracted using Trizol (Life Technologies, Gaithersburg, MD). The ApaI fragment was amplified from RNA by RT-PCR and digested with ApaI, and a ratio of the ApaI(+) to the ApaI(-) allele was obtained at each time point by quantitation of each band using a phosphorimager.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In leukocytes, there is a variable relaxation of IGF2
imprinting between individuals, ranging from complete repression of the
maternal gene copy to equal expression of both copies (15). We have
shown preliminary evidence that the VNTR is not the locus controlling
the variable expression level of the maternal allele in leukocytes
(15). However, the VNTR could modulate IGF2 expression
independently of parental origin. Using an IGF2 3'-UTR
ApaI RFLP to distinguish between alleles, the relative
expressions of the paternal to maternal IGF2 gene copies
were compared between individuals with a paternal class III-maternal
class I VNTR status and those with a paternal class I-maternal class
III status. The ratio of paternal to maternal IGF2 mRNA was
not significantly different between these two groups [4.2 ± 1.5
(n = 4) vs. 3.1 ± 1.3 (n = 4)] in
leukocytes (Fig. 2 and Table 1), indicating that class III alleles do
not affect IGF2 expression levels differently from class I
alleles.
|
|
). To confirm this association, we compared
the paternal to maternal IGF2 ratios in the larger dataset
consisting of ApaI(±) heterozygous and VNTR class I/I
individuals (Table 1). Once again the paternal to maternal ratio was
higher with a paternal ApaI(+) than with a paternal
ApaI(-) [mean, 7.9 (n = 13) vs. 0.6
(n = 10), respectively]. In samples with a paternal/maternal
ratio significantly less than 1, the ApaI effect obviously
overrides the parental imprinting effect. Because of recent evidence that, unlike all other class I alleles, the allele designated 814 is not predisposing to diabetes when inherited from 814/III heterozygous fathers (23), we examined such alleles specifically. Of the 22 paternal alleles, only 2 were 814 that were inherited from 814 class III fathers. Neither could be distinguished from the rest of the class I alleles, either by ApaI haplotype (both were + as most 814 are) or by paternal/maternal ratio (4.0 and 5.5, values close to the median for ApaI heterozygotes with a paternal +). This may be taken as additional evidence that IGF2 levels in leukocytes are not relevant to the IDDM2 effect.
The class I VNTR alleles can be subdivided into a large and a small
mode based on size and linkage disequilibrium with neighboring
polymorphisms (38). Two hundred and seventy-six of 348 class
I/ApaI(-) haplotypes analyzed contained a small class I
allele (smaller than or equal to the 742 allele, 37 repeat units), and
970 of 1034 class I/ApaI(+) haplotypes had a large class I
allele (larger than or equal to the 756 allele, 38 repeat units).
Interestingly, only about 50% of the 728 and 742 class I alleles are
in cis with ApaI(-), indicating that these 2
alleles are in linkage equilibrium with ApaI. Overall, the
ApaI(-) allele is in strong linkage disequilibrium with
small class I alleles (up to allele 714), whereas ApaI(+) is
in strong linkage disequilibrium with large class I alleles (greater
than allele 756, inclusive; D' = 0.82; Fig. 3). This suggested the possibility that
ApaI could be a marker for a transcriptional effect of VNTR
class I allele subgroups. When we examined mismatched haplotypes
containing the ApaI(+) allele with a short class I VNTR or
ApaI(-) with a long class I, the ApaI(+) allele
was still expressed more strongly than the (-) allele, suggesting that
the effect on IGF2 mRNA levels does not functionally reside
in differential effects of the two class I size categories. Linkage
disequilibrium of ApaI was also seen with the VPH class III
alleles, as 15 of 18 VPH class III alleles analyzed were in
cis with an ApaI(+) allele.
|
]. There was a significant correlation between those ratios in
leukocytes and stimulated lymphocytes from the same individual (r
= 0.71). Differential mRNA stability
Association of ApaI(+) alleles with higher
IGF2 steady state mRNA levels than ApaI(-) may
be a transcriptional effect, or it may indicate a difference in
stability between the two transcripts. To investigate this, we arrested
transcription in cultured lymphocytes using actinomycin D and compared
the relative abundance of the two transcripts at various time points.
The ratio of one allele to the other did not change significantly over
this time period (Fig. 4); thus, the two
alleles are equally stable under these conditions. This result
indicates that both alleles have the unusually high stability of
IGF2 mRNA reported by Sussenbach et al. (40),
making it unlikely that the ApaI effect is due to
differential RNA degradation.
|
As ApaI does not affect mRNA stability, the only other
explanation for the observed association between ApaI and
IGF2 mRNA levels is that ApaI is associated with
a transcriptional effect on IGF2. We hypothesized that the
role of ApaI or, more likely, the sequence variant for which
it is a marker is to act as an imprinting modifier locus that affects
the extent to which the repressed maternal gene copy is activated, as
imprinting of IGF2 has been shown to be genotype dependent
(25). To determine whether parental origin affected the ApaI
association with IGF2 mRNA levels, we used quantitative
competitive RT-PCR to quantitate absolute levels of IGF2
mRNA in the samples for which the paternal to maternal ratio was known
(Fig. 2). Combining the two, we determined the paternal and maternal
contributions to the absolute amount of IGF2 mRNA. The
maternal ApaI(+) IGF2 mRNA level (mean, 8.9
arbitrary units; n = 10) was still higher than the maternal
ApaI(-) level (mean, 3.4; n = 14), and the paternal
ApaI(+) mRNA level (mean, 17.0; n = 14) was higher than
the paternal ApaI(-) level (mean, 4.4; n = 10; Table 1). Thus, the ApaI(+) allele is associated with higher
IGF2 mRNA levels regardless of parental origin, indicating
that the transcriptional effect does not involve modifying the extent
to which imprinting is relaxed.
VNTR/ApaI genetic analysis
If genetic variation in IGF2 expression levels in
leukocytes has a functional role in the development of type 1 diabetes,
then the association of ApaI alleles with very different
IGF2 mRNA levels would make ApaI a marker for
diabetes susceptibility, either through linkage disequilibrium with
specific VNTR alleles or independently of the VNTR. To investigate this
possibility, we compared the number of transmitted to nontransmitted
ApaI(+) and ApaI(-) alleles to diabetic children
from unaffected parents. There was no significant difference in the
frequency with which the alleles were transmitted to diabetic offspring
from ApaI(±) heterozygous fathers or mothers, indicating
the absence of an association between ApaI and type 1
diabetes (Table 2), and, consequently, no
evidence for an association between ApaI RFLP-regulated
IGF2 expression in leukocytes and type 1 diabetes.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Despite its lack of association with diabetes, the ability of the ApaI polymorphism or a polymorphism in tight linkage disequilibrium with it to modulate IGF2 mRNA levels may be important in other human diseases, such as cancer and growth disorders in early life. The mechanism of this effect may involve either differential mRNA stability or effects on transcription levels of the gene. Our finding that blocking transcription had no effect on the ratio of the two alleles over time is evidence against differential RNA degradation as the mechanism of the ApaI effect. We also explored the possibility that the ApaI polymorphism was responsible for modifying the derepression of maternal gene expression. The ApaI(+) allele was associated with higher IGF2 mRNA levels regardless of parental origin, ruling out the possibility that ApaI or something in linkage disequilibrium with it acts as a modifier locus for IGF2 imprinting.
Alleles within each VNTR class are heterogeneous in terms of both size and sequence of variants of the consensus repeat unit (9). Although the majority of the reported contribution of the VNTR to the genetics of type 1 diabetes is between class I and class III as a whole, there is evidence that alleles within classes may differ in the degree of diabetes susceptibility or protection that they confer (8, 23). Therefore, although the main IDDM2 effect (class I vs. class III) cannot be accounted for by IGF2 modulation in the tissues examined, we cannot rule out that such modulation may play a role in the differences between alleles of the same class. Specifically, the ApaI effect could be due to an as yet unknown linkage disequilibrium with specific VNTR alleles.
We have demonstrated an absence of association between the VNTR locus responsible for the IDDM2 effect on susceptibility and expression levels of the nearby IGF2 gene in leukocytes, suggesting that IGF2 expression in leukocytes does not play a role in the autoimmune pathogenesis of type 1 diabetes. However, it is still possible that IGF2 modulation by the VNTR in other tissues, such as the small effect seen in placenta, could affect diabetes susceptibility through nonimmune mechanisms, such as influencing early growth patterns. The demonstration of the absence of modulation of IGF2 in tissues with the most obvious relevance to the immunology of type 1 diabetes weakens the position of IGF2 as a candidate target and reinforces the hypothesis that INS expression in the thymus is a primary mode of action of INS VNTR allelic variation in type 1 diabetes pathogenesis.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Present address: Oxagen Ltd., 91 Milton Park, Abingdon, Oxon,
United Kingdom OX14 4RY.
Received March 11, 1998.
Accepted May 7, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  R. Bachner-Melman, A. H. Zohar, L. Nemanov, U. Heresco-Levy, I. Gritsenko, and R. P. Ebstein Association Between the Insulin-Like Growth Factor 2 Gene (IGF2) and Scores on the Eating Attitudes Test in Nonclinical Subjects: A Family-Based Study Am J Psychiatry, December 1, 2005; 162(12): 2256 - 2262. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Durinovic-Bello, E. Jelinek, M. Schlosser, T. Eiermann, B. O. Boehm, W. Karges, L. Marchand, and C. Polychronakos Class III Alleles at the Insulin VNTR Polymorphism Are Associated With Regulatory T-Cell Responses to Proinsulin Epitopes in HLA-DR4, DQ8 Individuals Diabetes, December 1, 2005; 54(suppl_2): S18 - S24. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. S. Sandhu, B. Heude, E. H. Young, R. Luben, J. Luan, K.-T. Khaw, J. Todd, and N. J. Wareham INS VNTR Class Genotype and Indexes of Body Size and Obesity: Population-Based Studies of 7,999 Middle-Aged Men and Women Diabetes, September 1, 2005; 54(9): 2812 - 2815. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. F. Escobar-Morreale, M. Luque-Ramirez, and J. L. San Millan The Molecular-Genetic Basis of Functional Hyperandrogenism and the Polycystic Ovary Syndrome Endocr. Rev., April 1, 2005; 26(2): 251 - 282. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. Schrager, S. M. Roth, R. E. Ferrell, E. J. Metter, E. Russek-Cohen, N. A. Lynch, R. S. Lindle, and B. F. Hurley Insulin-like growth factor-2 genotype, fat-free mass, and muscle performance across the adult life span J Appl Physiol, December 1, 2004; 97(6): 2176 - 2183. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. A. McCann, Y. Q. Xu, R. Frechette, L. Guazzarotti, and C. Polychronakos The Insulin-Like Growth Factor-II Receptor Gene Is Associated with Type 1 Diabetes: Evidence of a Maternal Effect J. Clin. Endocrinol. Metab., November 1, 2004; 89(11): 5700 - 5706. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. San Millan, M. Corton, G. Villuendas, J. Sancho, B. Peral, and H. F. Escobar-Morreale Association of the Polycystic Ovary Syndrome with Genomic Variants Related to Insulin Resistance, Type 2 Diabetes Mellitus, and Obesity J. Clin. Endocrinol. Metab., June 1, 2004; 89(6): 2640 - 2646. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M A Kelly, M L Rayner, C H Mijovic, and A H Barnett Molecular aspects of type 1 diabetes Mol. Pathol., February 1, 2003; 56(1): 1 - 10. [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 |
