| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Vascular Biology Research Group, Robarts Research Institute, and Department of Medicine (H.C., E.v.d.V., M.R.B., J.G.P., R.A.H.), and Thames Valley Family Practice Research Unit (S.B.H.), University of Western Ontario, London, Ontario, Canada N6A 5K8; and Department of Medicine and Samuel Lunenfeld Research Institute, Mount Sinai Hospital and University of Toronto (A.J.G.H., B.Z.), and Department of Public Health Sciences, University of Toronto (T.K.Y.), Toronto, Ontario, Canada M5S 1A8
Address all correspondence and requests for reprints to: Dr. Robert A. Hegele, Blackburn Cardiovascular Genetics Laboratory, Robarts Research Institute, 100 Perth Drive, Room 406, London, Ontario, Canada N6A 5K8. E-mail: hegele{at}robarts.ca.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
G promoter SNP was within a cis-acting element required for basal and cAMP-mediated PCK1 gene transcription. The expression of a luciferase reporter construct containing -232G in three different cell lines showed significantly increased basal expression with no down-regulation by insulin compared with a construct containing -232C. The odds ratios for type 2 DM among subjects with one or two copies of -232G compared with -232C/C homozygotes were 1.9 (95% confidence interval, 1.2–3.0) in a Canadian aboriginal sample and 2.8 (95% confidence interval, 1.7–4.7) in a Caucasian sample. Thus, we report a promoter SNP in PCK1 that was resistant to down-regulation by insulin in vitro and was associated with type 2 DM in two independent study samples. |
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The study received approval from the University of Western Ontario ethics review panel. PCK1 genomic sequence was studied in 12 unrelated Caucasians with type 2 DM diagnosed between age 30–54 yr (median, 44 yr), and mean body mass index (BMI) of 30.8 kg/m2 (range, 27.0–40.3 kg/m2). This sample size provided 100% power to detect alleles with a frequency greater than 0.15, and 70% power to detect alleles with a frequency greater than 0.05. Disease association was further evaluated in 115 clinic-based Caucasian type 2 DM subjects (42% female) with mean ± SE age and BMI of 51.6 ± 5.6 yr and 31.2 ± 3.5 kg/m2, respectively. All type 2 DM subjects were treated with oral hypoglycemic agents and/or insulin. We also studied 260 clinic-based normoglycemic Caucasian controls (52% female) with age and BMI of 50.6 ± 6.5 yr and 25.8 ± 3.0 kg/m2, respectively. Control subjects had normal fasting glucose according to 1997 American Diabetes Association diagnostic criteria. PCK1 SNP allele frequencies were evaluated in population-based normal African, European, Chinese, East Indian, Oji-Cree, and Inuit subjects (n = 50 from each ethnic group). Disease association of PCK1 -232CG SNP genotype was evaluated in a well characterized, community-based sample of 170 Oji-Cree aged 18 yr or older either with type 2 DM or glucose intolerance and 392 normoglycemic controls (5, 9, 10). All Oji-Cree subjects were diagnosed based on a standardized glucose challenge (5, 9, 10).
DNA analysis
Primers to amplify the PCK1 promoter and exons 1–9, including at least 50 bp of intron sequence at each boundary (Table 1
) were designed using reference sequences from the National Center for Biotechnology Information (GenBank accession no. L_12760 and NM_002591). Amplification conditions were: denaturation at 94 C for 5 min; followed by 30 cycles comprised of 30 sec each at 94, 59, and 72 C; followed by a final extension step at 72 C for 10 min. Amplification products were run on 1.5% agarose and purified using QIAEX II (Qiagen, Mississauga, Canada). DNA was bidirectionally sequenced using amplification primers, sequences were detected on a 377 PRISM Automated DNA Sequencer (PE Applied Biosystems, Mississauga, Canada), and results were analyzed with Sequence Navigator software (PE Applied Biosystems). Each PCK1 mutation was seen in both directions and was confirmed using a second, independent amplification and resequencing on another day.
|
T and -1107CT SNPs was performed using direct genomic DNA sequencing. Exon 3 574GA and 610AG SNPs were genotyped using DNA amplification, followed by purification using serum alkaline phosphatase and exonuclease I (ExoI; PE Applied Biosystems). The treatment reaction mixture included 2.0 µl serum alkaline phosphatase (1.0 U/µl), 0.2 µl ExoI, 6.0 µl deionized water, and 4.0 µl amplification product. This was incubated at 37 C for 1 h and then at 72 C for 15 min to inactivate the enzyme. Additional primers were designed to terminate one nucleotide before the SNP. For the 574GA and 610AG SNPs, these primers were 5'-ATG GGG CCG CTG GGC TC-3' and 5'-ATC GAG CTG ACG GAT TC-3', respectively. The PCR template and primer were treated with a SNaPshot ddNTP Primer Extension Kit and analyzed on a PRISM 377 DNA Sequencer (PE Applied Biosystems). Altered restriction endonuclease recognition sites permitted genotyping of the remaining SNPs using 2% agarose gels. Digestion of the 1640-bp promoter product with MaeIII yielded fragments of 494, 483, 393, 180, 38, 34, and 18 bp for the -232C allele and fragments of 674, 483, 393, 38, 34, and 18 bp for the -232G allele. Digestion of the 341-bp exon 1 product with CacI yielded fragments of 195, 71, and 71 bp for the 190G allele and 270 and 71 bp for the 190A allele. Digestion of the 590-bp product that contained exons 3 and 4 with BsrI yielded fragments of 199, 160, 120, 85, and 17 bp for the L184 allele and 289, 199, 85, and 17 bp for the V184 allele. Digestion of the same product with TaqI yielded fragments of 484, 93, and 13 bp for the 583G allele and 484 and 106 bp for the 583A allele. Digestion of the 554-bp exon 6 product with BsaHI yielded fragments of 334 and 220 bp for the 1261C allele and 554 bp for the 1261T allele. Expression assays
The PCK1 promoter was mutagenized at position -232 by the two-primer pair method with primers: 5'-GTT GTG TCA AAA GTC ACT ATG GTT GGT-3' and 5'-ACC AAC CAT AGT GAC TTT TGA CAC AAC-3' (mutated nucleotide in underlined bold). A DNA fragment with 1.6 kb of PCK1 promoter was amplified using primers: 5'-TCT AAG TGA GTT TGG TCG GAG G-3' and 5'-CTG CAG AGT GCT GCT AAG GG-3'. The amplified promoter fragments were ligated into a TA cloning vector (Invitrogen, Carlsbad, CA). The promoter fragment was subcloned into the luciferase reporter PGL3 basic vector (Promega, Madison, WI) after SacI and XhoI digestion. The fidelity of amplification, replication and the introduction of -232CG were confirmed by DNA sequencing. Both wild-type and mutant constructs contained -1237C and -1107C as the genomic backbone.
The human hepatoma HepG2 and the mouse preadipocyte NIH-3T3 cell lines were obtained from American Type Culture Collection (Manassas, VA). Mouse dermal fibroblasts were also studied (11). Cells were grown in DMEM (Invitrogen, Carlsbad, CA) supplemented with 100 U/ml penicillin, 100 mg/ml streptomycin and 10% (vol/vol) heated-inactivated fetal calf serum (Invitrogen). Cell cultures grew to 70% confluence in six-well culture plates and then were transfected using calcium phosphate (www.stanford.edu/group/nolan/index.html). For each transfection, 4 µg DNA and 6.7 mg CaCl2 in HEPES-buffered saline were added. For each transfection with the PCK1-containing vector, 2 µg pSV-ß-galactosidase vector (ProFection Mammalian Transfection System, Promega, Madison, WI) was cotransfected. The transfection solution was removed after 20 h at 37 C and replaced by fresh medium for further culture. Insulin was added to achieve final concentrations of 0, 10, 25, 50, and 100 nmol/liter. Cells were harvested 48 h posttransfection. Luciferase activity was assayed (Promega Corp., Madison, WI), and luminescence was determined using a Lumat LB9507 luminometer (Berthold Systems, Pittsburgh, PA). Luciferase activities were normalized per unit of activity of ß-galactosidase [reporter luciferase units (RLU)]. The ß-galactosidase assay (Promega Corp.) used 150 µl cell lysate mixed with 150 µl 2x assay buffer and was incubated at 37 C for 3 h. The reaction was stopped by 500 µl 1 M sodium carbonate. Spectrophotometric absorbance at 420 nm was read. A blank was prepared using the lysate from nontransfected cells. Each data point was the mean of three triplicate assays from two experiments performed on different days.
Statistical analysis
SAS version 8 (SAS Institute, Cary, NC) was used for statistical analyses. Deviation of genotype frequencies from those predicted by Hardy-Weinberg equilibrium for each locus in each group were evaluated using the conditional 
2 goodness of fit test using 1 df (12), with adjustment for small numbers as described (12). Pairwise linkage disequilibrium (LD) between SNP genotypes was estimated as described using the maximal likelihood estimate of Hill and Robertson (12, 13). Significance and confidence intervals for type 2 DM odds ratios (ORs) were calculated based on case-control 2 analysis (with 1 df). For comparisons of data from luciferase expression experiments, paired t tests were used to compare individual insulin concentrations, and repeated measures ANOVA in SAS was used to compare expression curves over the dose range of insulin. A nominal value of P < 0.05 was used for all statistical comparisons.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Genomic DNA sequencing in 12 Caucasian subjects with type 2 DM uncovered nine PCK1 polymorphisms (Table 2). The observed genotype frequencies in all subsequent experiments and sample subgroups followed Hardy-Weinberg expectations (data not shown). Our primers captured the full PCK1 promoter (14), and we found three promoter SNPs, namely, -1237CT, -1107CT, and -232CG; one nonsynonymous coding SNP, namely, 671CG (trivial name L184V); and five synonymous coding SNPs, namely, 190AG, 574GA, 583AG, 610AG, and 1261CT. Although the promoter SNPs were not present in the National Center for Biotechnology Information database (db), the 190AG, 574GA, 583AG, 610AG, 671CG, and 1261CT SNPs were already assigned dbSNP reference numbers 1042521, 1062600, 1042523, 1062601, 707555, and 2070756, respectively. All SNPs were present with varying frequencies in different ethnic groups (Table 2). Pairwise linkage disequilibrium relationships differed according to ethnicity (Tables 3–5).
|
|
|
|
G SNP genotype with diabetes
Genotype and allele frequencies of the PCK1 -232CG genotype in Oji-Cree and Caucasian type 2 DM case-control samples are shown in Table 6. For both samples, the distribution of genotypes was significantly different between cases and controls, with ahigher number of -232G/G homozygotes in each diabetic subgroup. For both samples, the -232G allele frequency was significantly increased in subjects with type 2 DM. Compared with -232C/C homozygotes, Oji-Cree and Caucasian subjects with at least one -232G allele (dominant model, -232G/G plus -232G/C subjects) had type 2 DM OR of 1.9 [95% confidence interval (CI), 1.2 - 3.0; P = 0.009] and 2.8 (95% CI, 1.7–4.7; P = 0.00004), respectively. Compared with -232C/C homozygotes, Oji-Cree and Caucasian subjects with two -232G alleles (recessive model) had type 2 DM OR of 2.3 (95% CI, 1.4–4.0; P = 0.002) and 3.1 (95% CI, 1.6–6.2; P = 0.001), respectively. In Oji-Cree, we previously showed that the HNF1A G319S genotype was associated with increased risk of type 2 DM (5). In Oji-Cree more than 40 yr of age, subjects with one or more HNF1A S319 allele had type 2 DM OR of 3.0 (95% CI, 1.2–7.5; P = 0.009), but subjects with one or more HNF1A S319 allele and one or more PCK1 -232G allele had type 2 DM OR of 3.2 (95% CI, 1.5–6.7; P = 0.002) compared with a reference group homozygous for the wild-type allele for each polymorphism.
|
G SNP
DNA sequencing confirmed that all constructs contained both -1237C and -1107C, and that the only sequence variation was either C or G at nucleotide (nt) -232. The negative control construct contained the wild-type sequence in the reverse orientation, with no significant difference in activity compared with the blank (data not shown). Normalized luciferase activities (RLU) from three triplicate experiments performed on 2 separate d are shown in Fig. 1. Basal RLU (insulin concentration, 0 nmol/liter) was between 1.5- and 15-fold higher for the -232G-containing construct in HepG2, 3T3L1 preadipocytes and mouse dermal fibroblasts. An elevated basal RLU of the -232G-containing construct was confirmed in three additional transfection experiments using an independent fusion gene preparation, with an overall mean 1.8-fold increase compared with the -232C-containing construct (data not shown; P < 0.0001).
|
). Paired t tests showed that at 50 nmol/liter insulin, reporter activity was reduced from baseline for the -232C-containing construct in HepG2 and 3T3L1 cells, but not in fibroblasts. These differences were not significant for the -232G-containing construct in any cell line. These results indicate that the -232G-containing construct had both increased basal activity and failure of suppression of activity with insulin compared with the -232C-containing construct. |
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
As with any association study, a positive result may not necessarily be due to the variant in question being a direct determinant of risk, but might arise from population stratification or LD with a nearby unmeasured risk determining polymorphism. The results summarized in Tables 3–5 suggest that there are distinctive LD relationships between PCK1 SNPs in different ethnic groups. For instance, of all pairwise LD estimates, approximately one sixth were significant in African and Caucasian subjects, about one third were significant in East Indian and Chinese subjects, and about half were significant in aboriginal (Oji-Cree and Inuit) populations, strongly suggesting interethnic differences in the genetic architecture of PCK1. Furthermore, the block of alleles most commonly associated with -232CG differed between ethnic groups; only 583AG showed significant LD with -232CG in all groups. These varying LD relationships taken together with the clear in vitro functional impact of -232CG increase the likelihood that this marker is itself a contributor to the observed genetic associations.
The -232CG SNP occurs close to a region analogous to the P3(I) domain of the rat PCK1 promoter (14). This domain is trans-activated when CCAAT/enhancer-binding protein (C/EBP) isoforms C/EBP
and C/EBPß form heterodimers and bind. C/EBP is centrally involved in regulating the multiple hormone response of PCK1 gene expression. Engagement of the C/EBP and C/EBPß complex to the P3(I) regulatory element is required along with thyroid hormone (T3) and the retinoid X receptor binding in a heterodimer to the thyroid response element for T3 stimulation of PCK1 (21). cAMP stimulation of PCK1 transcription requires both cAMP response elements to bind cAMP response element-binding protein or C/EBP (with similar affinity) and C/EBP to bind to the P3(I) regulatory element (22). Alterations of such protein-DNA interactions can now be studied using PCK1 promoter -232C- and -232G-containing constructs.
How might the -232GC SNP alter insulin suppression of PCK1 expression? In transgenic mice, mutation of the accessory factor 2-binding region in PCK1 (between nt -451 and -353) abolished glucocorticoid-dependent hepatic transcription, with no inhibition of expression by insulin (23). Furthermore, the induced mutation had opposite effects in liver and kidney, indicating that tissue specificity compounds the complexity of PCK1 expression and demonstrating that insulin does not exert its negative effect on the PCK1 promoter through accessory factor 2 (23). Full PCK1 promoter activation also requires the coactivator PPAR
coactivator-1 (PGC-1) (24). The dependence of insulin suppression of PCK1 expression on PGC-1 was also suggested by induction of PGC-1 expression in both type 2 DM and mice with a disrupted insulin receptor gene (25).
The region of chromosome 20q in proximity to PCK1 was linked with type 2 diabetes in at least two genome-wide scans (26, 27), implicating PCK1 as a candidate gene based on both chromosomal position and function. Although earlier analysis of the PCK1 promoter using SSCP found no genomic variants in type 2 DM (4), a more recent study using SSCP identified PCK1 -232CG and found it to be associated with type 2 diabetes in Japanese subjects (28). Thus, SSCP might have variable sensitivity for detecting certain SNPs. There have been no systematic studies comparing the sensitivity and specificity of SSCP with direct DNA sequencing to detect mutations in different genomic regions using different protocols. In our facility, the increment in cost for screening by direct genomic DNA sequencing compared with SSCP is about 2.5-fold. Discoveries of diabetes-causing gene mutations using sequencing as a screening method (5, 6, 7, 8) have lowered our threshold for using this technology. Given the importance of minimizing false negative results when screening for DNA variants with potential pathogenic implications, it may be advantageous to include direct genomic sequencing in a screening strategy despite the increment in cost.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
S.B.H. is a Career Scientist with the Ontario Ministry of Health.
T.K.Y. is a Senior Investigator with the Canadian Institutes of Health Research.
J.G.P. and R.A.H. are Career Investigators with the Heart and Stroke Foundation of Ontario.
R.A.H. holds a Canada Research Chair (Tier I) in Human Genetics.
Abbreviations: BMI, Body mass index; C/EBP, CCAAT/enhancer-binding protein; CI, confidence interval; db, National Center for Biotechnology Information database; DM, diabetes mellitus; IRS, insulin response sequence; LD, linkage disequilibrium; nt, nucleotide; OR, odds ratio; PCK1, phosphoenolpyruvate carboxykinase gene; PGC-1, PPAR coactivator-1; RLU, reporter luciferase units; SNP, single nucleotide polymorphism; SSCP, single strand conformation polymorphism.
Received August 4, 2003.
Accepted October 27, 2003.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
G319S variant is associated with early-onset type 2 diabetes in Canadian Oji-Cree. J Clin Endocrinol Metab 84:1077–1082 G319S, a transactivation-deficient mutant, is associated with altered dynamics of diabetes onset in an Oji-Cree community. Proc Natl Acad Sci USA 99:4614–4619 (C/EBP) is required for the thyroid hormone but not the retinoic acid induction of phosphoenolpyruvate carboxykinase (PEPCK) gene transcription. Biochem J 322:343–349
This article has been cited by other articles:
 |
|
 |
|
  A. G. Gomez-Valades, A. Mendez-Lucas, A. Vidal-Alabro, F. X. Blasco, M. Chillon, R. Bartrons, J. Bermudez, and J. C. Perales Pck1 Gene Silencing in the Liver Improves Glycemia Control, Insulin Sensitivity, and Dyslipidemia in db/db Mice Diabetes, August 1, 2008; 57(8): 2199 - 2210. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Ghanaat-Pour, Z. Huang, M. Lehtihet, and A. Sjoholm Global expression profiling of glucose-regulated genes in pancreatic islets of spontaneously diabetic Goto-Kakizaki rats J. Mol. Endocrinol., August 1, 2007; 39(2): 135 - 150. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Grupe, R. Abraham, Y. Li, C. Rowland, P. Hollingworth, A. Morgan, L. Jehu, R. Segurado, D. Stone, E. Schadt, et al. Evidence for novel susceptibility genes for late-onset Alzheimer's disease from a genome-wide association study of putative functional variants Hum. Mol. Genet., April 15, 2007; 16(8): 865 - 873. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. J. Willer, L. L. Bonnycastle, K. N. Conneely, W. L. Duren, A. U. Jackson, L. J. Scott, N. Narisu, P. S. Chines, A. Skol, H. M. Stringham, et al. Screening of 134 Single Nucleotide Polymorphisms (SNPs) Previously Associated With Type 2 Diabetes Replicates Association With 12 SNPs in Nine Genes Diabetes, January 1, 2007; 56(1): 256 - 264. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. A. Hegele, K. Z. Al-Shali, A. A. House, A. J.G. Hanley, S. B. Harris, M. Mamakeesick, A. Fenster, B. Zinman, H. Cao, and J. D. Spence Disparate Associations of a Functional Promoter Polymorphism in PCK1 With Carotid Wall Ultrasound Traits Stroke, December 1, 2005; 36(12): 2566 - 2570. [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 |
). The upper, middle, and lower panels show results for transfections in HepG2 cells, 3T3L1 cells, and dermal fibroblasts, respectively. Concentrations of insulin are shown on the abscissa. Normalized RLU per unit of ß-galactosidase activity are shown on the ordinate. The mean ± SE of reporter activity for each insulin concentration is shown for three sets of triplicate experiments performed on 2 separate days. The normalized basal (insulin concentration, 0 nmol/liter) luciferase activity was significantly higher for HepG2 cells, 3T3L1 cells, and dermal fibroblasts transfected with the -232G-containing construct than in those transfected with the -232C-containing construct (P = 0.02, P = 0.05, and P < 0.0001, respectively). Repeated measures analysis of reporter activity across the range of insulin concentrations showed significant differences between the -232G- and -232C-containing constructs transfected in HepG2 (P = 0.0003), 3T3L1 (P = 0.0008), and dermal fibroblast (P < 0.0001) cell lines. Paired t tests showed that for an insulin concentration of 50 nmol/liter, reporter activity was significantly reduced from baseline for the -232C-containing construct transfected in both HepG2 (P = 0.0003) and 3T3L1 (P = 0.02) cell lines, but not in fibroblasts. Paired t tests showed that for an insulin concentration of 100 nmol/liter, reporter activity was significantly reduced from baseline for the -232C-containing construct transfected in both HepG2 (P = 0.05) and 3T3L1 (P = 0.02) cell lines, but not in fibroblasts. In contrast, these differences were not significant for the -232G-containing construct in any cell line (all P > 0.30).