G I Bell

University of Chicago, Chicago, Illinois, United States

Are you G I Bell?

Claim your profile

Publications (300)2456.61 Total impact

  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Advances of modern sensing and sequencing technologies generate a deluge of high dimensional space-temporal physiological and next-generation sequencing (NGS) data. Physiological traits are observed either as continuous random functions, or on a dense grid and referred to as function-valued traits. Both physiological and NGS data are highly correlated data with their inherent order, spacing, and functional nature which are ignored by traditional summary-based univariate and multivariate regression methods designed for quantitative genetic analysis of scalar trait and common variants. To capture morphological and dynamic features of the data and utilize their dependent structure, we propose a functional linear model (FLM) in which a trait curve is modeled as a response function, the genetic variation in a genomic region or gene is modeled as a functional predictor, and the genetic effects are modeled as a function of both time and genomic position (FLMF) for genetic analysis of function-valued trait with both GWAS and NGS data. By extensive simulations, we demonstrate that the FLMF has the correct type 1 error rates and much higher power to detect association than the existing methods. The FLMF is applied to sleep data from Starr County health studies where oxygen saturation were measured in 22,670 seconds on average for 833 individuals. We found 65 genes that were significantly associated with oxygen saturation functional trait with P-values ranging from 2.40E-06 to 2.53E-21. The results clearly demonstrate that the FLMF substantially outperforms the traditional genetic models with scalar trait.
    Full-text · Article · Oct 2014
  • D Savic · S.Y. Park · K.A. Bailey · G.I. Bell · M.A. Nobrega
    [Show abstract] [Hide abstract]
    ABSTRACT: Aims/hypothesis: Recent functional characterisations of genome-wide association study (GWAS) loci suggest that cis-regulatory variation may be a common paradigm for complex disease susceptibility. Several studies point to a similar mechanism at the transcription factor 7-like 2 (TCF7L2) GWAS locus for type 2 diabetes. To address this possibility, we carried out an in vitro scan of this diabetes-associated locus to fine-map cis-regulatory sequences within this genomic interval. Methods: A systematic cell-based enhancer strategy was employed to interrogate all sequences within the 92 kb type-2-diabetes-association interval for cis-regulatory activity in a panel of cell lines (HCT-116, Neuro-2a, C2C12, U2OS, MIN6 and HepG2). We further evaluated chromatin state at a subset of these regions in HCT-116 and U2OS cells and examined allelic-specific enhancer properties at the type-2-diabetes-associated single nucleotide polymorphism (SNP) rs7903146. Results: In total, we assigned cis-regulatory activity to approximately 30% (9/28) of constructs tested. Notably, a subset of enhancers was active across multiple cell lines and overlapped with key epigenetic markers suggestive of cis-regulatory sequences. We further replicated the allelic-specific properties for SNP rs7903146 in pancreatic beta cells and additionally demonstrate identical allelic-specific enhancer effects in other cell lines. Conclusions: These results provide a detailed map of cis-regulatory elements within the TCF7L2 GWAS locus and support the hypothesis of cis-regulatory variation leading to type 2 diabetes predisposition. The detection of allelic-specific effects for SNP rs7903146 in multiple cell lines further alludes to the likelihood of a peripheral defect in disease aetiology.
    No preview · Article · Sep 2012 · Diabetologia
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We conducted genome-wide association studies (GWASs) and expression quantitative trait loci (eQTL) analyses to identify and characterise risk loci for type 2 diabetes in Mexican-Americans from Starr County, TX, USA. Using 1.8 million directly interrogated and imputed genotypes in 837 unrelated type 2 diabetes cases and 436 normoglycaemic controls, we conducted Armitage trend tests. To improve power in this population with high disease rates, we also performed ordinal regression including an intermediate class with impaired fasting glucose and/or glucose tolerance. These analyses were followed by meta-analysis with a study of 967 type 2 diabetes cases and 343 normoglycaemic controls from Mexico City, Mexico. The top signals (unadjusted p value <1 × 10(-5)) included 49 single nucleotide polymorphisms (SNPs) in eight gene regions (PER3, PARD3B, EPHA4, TOMM7, PTPRD, HNT [also known as RREB1], LOC729993 and IL34) and six intergenic regions. Among these was a missense polymorphism (rs10462020; Gly639Val) in the clock gene PER3, a system recently implicated in diabetes. We also report a second signal (minimum p value 1.52 × 10(-6)) within PTPRD, independent of the previously implicated SNP, in a population of Han Chinese. Top meta-analysis signals included known regions HNF1A and KCNQ1. Annotation of top association signals in both studies revealed a marked excess of trans-acting eQTL in both adipose and muscle tissues. In the largest study of type 2 diabetes in Mexican populations to date, we identified modest associations of novel and previously reported SNPs. In addition, in our top signals we report significant excess of SNPs that predict transcript levels in muscle and adipose tissues.
    Full-text · Article · Jun 2011 · Diabetologia
  • Source
    D F Steiner · S-Y Park · J Støy · L H Philipson · G I Bell
    [Show abstract] [Hide abstract]
    ABSTRACT: The preeminent role of the beta cell is to manufacture, store and release insulin. The mature insulin molecule is composed of two polypeptide chains designated as A and B that are joined by two pairs of disulfide bonds with an additional intramolecular disulfide bond in the A chain. However, the two chains of the insulin molecule are not synthesized as separate polypeptide chains but rather are generated by specific proteolytic processing of a larger precursor, proinsulin. This discovery in 1967 and the concept of prohormones changed our view of the biosynthesis of hormones and neuropeptides. It allowed studies of the regulation of insulin biosynthesis that highlighted the key role of glucose. In addition, the C-peptide, the polypeptide that joins the A and B chains in proinsulin and is stored with insulin in the secretory granules and secreted in equimolar amounts, allowed studies of pancreatic beta cell function in vivo including in patients with diabetes. Subsequent studies have identified the specific proteases, prohormone convertases 1/3 and 2 and carboxypeptidase E, that are involved in the conversion of proinsulin to proinsulin intermediates and then to insulin. Disorders of (pro)insulin biosynthesis continue to illuminate important aspects of this pathway, revealing important connections to diabetes pathogenesis. Recent studies of patients with insulin gene mutations that cause permanent neonatal diabetes have identified key residues affecting the folding and structural organization of the preproinsulin molecule and its subsequent processing. These findings have renewed interest in the key role of endoplasmic reticulum function in insulin biosynthesis and the maintainance of normal beta cell health.
    Full-text · Article · Nov 2009 · Diabetes Obesity and Metabolism
  • [Show abstract] [Hide abstract]
    ABSTRACT: We previously mapped an obesity quantitative trait locus (QTL), Adip1, to the central portion of murine chromosome 1 which contains the calpain-10 (Capn10) gene. Human studies have associated calpain-10 (CAPN10) variants with type 2 diabetes and various metabolic traits. We performed a quantitative hybrid complementation test to determine whether differences attributed to Adip1 are due to variant Capn10 alleles in LG/J and SM/J mice. We crossed LG/J and SM/J to wild-type (C57BL/6J) and Capn10 knockout (Capn10-/-) mice to form four F1 hybrid groups; LG/J by wild-type (N = 76), LG/J by Capn10-/- (N = 116), SM/J by wild-type (N = 79), and SM/J by Capn10-/-(N = 64). We performed a two-way ANOVA with experimental strain, tester strain, and their interaction as the factors. Significant interaction indicates a quantitative failure to complement. We found failure to complement for fat, organ, and body weights, and leptin, female free fatty acid and triglyceride levels. Capn10-/- resulted in heavier weights and higher serum levels in LG/J crosses but not in SM/J crosses. Failure to complement was also found for glucose tolerance and insulin response tests. Here, the Capn10-/- allele results in lower glucose levels in crosses with SM/J but had no effect in the LG/J crosses. Differences between the LG/J and SM/J Capn10 alleles are the likely source of some of the QTL effects mapped to Adip1 in the LG/J by SM/J cross and Capn10 plays an important role in regulating obesity and diabetes in mice.
    No preview · Article · May 2009 · The Journal of Lipid Research
  • Source
    S-Y Park · G I Bell
    [Show abstract] [Hide abstract]
    ABSTRACT: We have generated a transgenic mouse model (MIP- LUC) that allows real-time imaging of insulin-secreting pancreatic beta cells in living mice. The beta cells of MIP- LUC transgenic mice emit a light signal that can be visualized externally by bioluminescent imaging using specialized equipment. In order to determine whether the intensity of the bioluminescent signal accurately reflects changes in beta-cell mass rather than simply transcriptional modulation of the mouse insulin I promoter-luciferase transgene, we examined the correlation between the bioluminescent signal and the beta-cell mass in MIP- LUC mice fed a regular or high-fat Western diet. Male MIP- LUC mice were fed a standard rodent diet (5% of calories from fat) or a high-fat Western diet (42% from fat) beginning at 4 weeks of age. The bioluminescent signal and beta-cell mass were measured after 6 and 10 weeks on each diet. The body weight, beta-cell mass, and bioluminescent signal increased with age and increased further in mice fed a high-fat diet. There was a statistically significant correlation between beta-cell mass and bioluminescent signal (r (2)=0.660, p=0.00137). Thus, in vivo bioluminescent imaging can be used to noninvasively monitor changes in beta-cell mass in living MIP- LUC mice, and it complements other approaches for monitoring beta-cell mass in states of insulin resistance, obesity, and diabetes.
    Full-text · Article · Nov 2008 · Hormone and Metabolic Research
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Our objective was to investigate the effect of lipid-induced insulin resistance and type 2 diabetes on skeletal muscle calpain-10 mRNA and protein levels. In the first part of this study, 10 healthy subjects underwent hyperinsulinemic euglycemic (4.5 mmol/liter) clamps for 6 h with iv infusion of either saline or a 20% Intralipid emulsion (Fresenius Kabi AG, Bad Homburg, Germany). Skeletal muscle biopsies were taken before and after 3- and 6-h insulin infusion and analyzed for calpain-10 mRNA and protein expression. In the second part of the study, muscle samples obtained after an overnight fast in 10 long-standing, sedentary type 2 diabetes patients, 10 sedentary, weight-matched, normoglycemic controls, and 10 age-matched, endurance-trained cyclists were analyzed for calpain-10 mRNA and protein content. Intralipid infusion in healthy subjects reduced whole body glucose disposal by approximately 50% (P<0.001). Calpain-10 mRNA (P=0.01) but not protein content was reduced after 6-h insulin infusion in both the saline and Intralipid emulsion trials. Skeletal muscle calpain-10 mRNA and protein content did not differ between the type 2 diabetes patients and normoglycemic controls, but there was a strong trend for total calpain-10 protein to be greater in the endurance-trained athletes (P=0.06). These data indicate that skeletal muscle calpain-10 expression is not modified by insulin resistance per se and suggest that hyperinsulinemia and exercise training may modulate human skeletal muscle calpain-10 expression.
    Full-text · Article · Mar 2008 · Journal of Clinical Endocrinology & Metabolism
  • M Hara · J Shen · W Pugh · K S Polonsky · M M Le Beau · G I Bell
    [Show abstract] [Hide abstract]
    ABSTRACT: Hepatocyte nuclear factor-6 (HNF-6) is the ONECUT-homeodomain transcription factor that is enriched in liver and also present in pancreas and central nervous system. It is expressed in the pancreatic bud at E10.5. In adult pancreas, its expression is restricted to the exocrine pancreas and duct cells. Since duct cells are thought to be precursors of endocrine cells and HNF-6 is involved in the regulation of the expression of HNF-4alpha and -1beta, genes that cause maturity onset diabetes of the young (MODY), we hypothesized that the sustained expression of HNF-6 would affect beta-cell function. We generated transgenic mice over-expressing human HNF-6 using the mouse insulin I promoter (MIP). We obtained one female founder in which the transgene had been incorporated into two sites; the chromosome (Ch) 14 and the X chromosome. The integration site of the latter was within centromeric heterochromatin and the transgene was inactivated. Studies on mice in which the transgene was integrated into Ch14 showed beta-cell specific defects functionally and pathologically. The insulin secretory response to glucose and arginine in the in situ-perfused pancreas was also significantly impaired in these mice. Immunohistochemical analysis revealed that the islets were smaller and had an abnormal architecture with an inverted ratio of alpha- and beta-cells resulting from beta-cell loss to 30% by 6-wk of age. The decreased number of beta-cells was quantified first time by fluorescent activated cell sorting using entire pancreata from the transgenic mice crossed with MIP-green fluorescent protein (GFP) mice. This severe loss of beta-cells involved programmed cell death.
    No preview · Article · Dec 2007 · Experimental and Clinical Endocrinology & Diabetes
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The mechanisms by which the calpain-10 gene (CAPN10) affects the risk of type 2 diabetes are unclear. Therefore, we investigated the effects of four polymorphisms in CAPN10 (single nucleotide polymorphism [SNP]-43, SNP-44, Insertion/Deletion [Indel]-19 and SNP-63) on insulin secretion, insulin action and abdominal fat distribution in offspring of patients with type 2 diabetes. Insulin secretion was determined by an IVGTT, insulin action by the hyperinsulinaemic-euglycaemic clamp and abdominal fat distribution by computed tomography in 158 non-diabetic offspring (age 34.9+/-6.3 years [mean+/-SD], BMI 26.2+/-4.9 kg/m(2)) of type 2 diabetic patients. SNP-43 (p=0.009 over the three genotypes, adjusted for age, sex, BMI and family relationship) and haplotypes carrying the A allele of SNP-43 were associated with intra-abdominal fat area. The A allele of SNP-43 was associated with intra-abdominal fat area in men (p=0.014) but not in women. SNP-44, InDel-19 and SNP-63 were not associated with intra-abdominal fat area or insulin action. Furthermore, we demonstrated in a separate sample of middle-aged men (n=234) who had a history of type 2 diabetes in first-degree relatives that the A allele of SNP-43 was associated with a large waist circumference, and high insulin levels in an OGTT. SNP-43 of CAPN10 may contribute to the risk of diabetes by regulating abdominal obesity in subjects with high risk of type 2 diabetes.
    Full-text · Article · Aug 2006 · Diabetologia
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The transcription factor upstream stimulatory factor 1 (USF1) regulates the expression of genes involved in glucose and lipid metabolism and has been associated with familial combined hyperlipidaemia. USF1 is located on chromosome 1q22-23, a region with evidence for linkage to type 2 diabetes and various traits of the metabolic syndrome in Chinese and other populations. The aim of this study was to investigate the linkage and association of USF1 with type 2 diabetes and the metabolic syndrome in Chinese individuals. We genotyped three haplotype-tagging single nucleotide polymorphisms (SNPs) (rs3737787, rs2516841 and rs2516839) at USF1 in three samples of the Hong Kong Chinese population, including members of 179 families from the Hong Kong Family Diabetes Study, 1,383 hospital cases with type 2 diabetes and/or the metabolic syndrome and 454 normal control subjects. We found significant association of individual polymorphisms and haplotypes with type 2 diabetes and/or metabolic syndrome-related traits in the family samples using either family-based or unrelated normal control subjects. However, these variants could not explain much of the evidence for linkage in this region. Moreover, they were not associated with type 2 diabetes and/or the metabolic syndrome in the hospital cases. The results are consistent with the hypothesis that variation at USF1 contributes to the risk of type 2 diabetes and the metabolic syndrome in families with strong evidence for linkage in the chromosome 1q region. However, they provide little support for USF1 as the susceptibility locus that generates the observed evidence for linkage at 1q21-25 for type 2 diabetes and/or the metabolic syndrome, and USF1 does not appear to have a major contribution to these phenotypes in the general Chinese population.
    Full-text · Article · Nov 2005 · Diabetologia

  • No preview · Article · Dec 2004
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Studies of the genetic basis of type 2 diabetes suggest that variation in the calpain-10 gene affects susceptibility to this common disorder, raising the possibility that calpain-sensitive pathways may play a role in regulating insulin secretion and/or action. Calpains are ubiquitously expressed cysteine proteases that are thought to regulate a variety of normal cellular functions. Here, we report that short-term (4-h) exposure to the cell-permeable calpain inhibitors calpain inhibitor II and E-64-d increases the insulin secretory response to glucose in mouse pancreatic islets. This dose-dependent effect is observed at glucose concentrations above 8 mmol/l. This effect was also seen with other calpain inhibitors with different mechanisms of action but not with cathepsin inhibitors or other protease inhibitors. Enhancement of insulin secretion with short-term exposure to calpain inhibitors is not mediated by increased responses in intracellular Ca2+ or increased glucose metabolism in islets but by accelerated exocytosis of insulin granules. In muscle strips and adipocytes, exposure to both calpain inhibitor II and E-64-d reduced insulin-mediated glucose transport. Incorporation of glucose into glycogen in muscle also was reduced. These results are consistent with a role for calpains in the regulation of insulin secretion and insulin action.
    Preview · Article · Oct 2001 · Diabetes
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Mutations in transcription factors that play a role in the development of the endocrine pancreas, such as insulin promoter factor-1 and NeuroD1/BETA2, have been associated with diabetes. Cell type-specific members of the basic helix-loop-helix (bHLH) family of transcription factors play essential roles in the development and maintenance of many differentiated cell types, including pancreatic beta-cells. Neurogenin 3 is a bHLH transcription factor that is expressed in the developing central nervous system and the embryonic pancreas. Mice lacking this transcription factor fail to develop any islet endocrine cells and die postnatally from diabetes. Because neurogenin 3 is required for the development of beta-cells and other pancreatic islet cell types, we considered it a candidate diabetes gene. We screened the coding region of the human neurogenin 3 gene (NEUROG3) for mutations in a group of unrelated Japanese subjects with maturity-onset diabetes of the young (MODY). We found three sequence variants: a deletion of 2-bp in the 5'-untranslated region (NEUROG3-g.-44-45delCA), a G-to-A substitution in codon 167 (g.499G/ A), resulting in a Gly-to-Arg replacement (G/R167), and a T-to-C substitution in codon 199 (g.596T/C), resulting in a Phe/Ser polymorphism F/S199. These polymorphisms were not associated with MODY, thereby suggesting that mutations in NEUROG3 are not a common cause of MODY in Japanese patients.
    Full-text · Article · Apr 2001 · Diabetes

  • No preview · Article · Feb 2001 · Journal of Diabetes and its Complications
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The basic helix-loop-helix (bHLH) family of transcription factors plays an important role in the normal development and function of the endocrine pancreas. Heterozygous mutations in the gene encoding one member of this family, NeuroD1/BETA2, are associated with a monogenic form of diabetes that resembles maturity-onset diabetes of the young (MODY) in many respects. This result prompted us to screen the genes encoding related bHLH transcription factors that are also expressed in pancreatic islets for diabetes-associated mutations. We have screened 57 unrelated Japanese subjects with a clinical diagnosis of MODY for mutations in the NeuroD4/Math-3/ATH-3 gene (NEUROD4). This analysis revealed seven frequent polymorphisms that were not associated with MODY, including five in the 5'-untranslated region (UTR) (-477G/A, -436delA, -324delT, -107insTTTT, and -104T/C [cDNA sequences]) and two in the 3'-UTR (1027C/T and 1076C/A). A missense mutation, K68T (203A/C), was found in a heterozygous state in one MODY subject and two nondiabetic subjects. The results of our study suggest that genetic variation in NEUROD4 is not a common cause of MODY in Japanese.
    Full-text · Article · Dec 2000 · Diabetes
  • Source
    M Hara · X Wang · V P Paz · N J Cox · N Iwasaki · M Ogata · Y Iwamoto · G I Bell
    [Show abstract] [Hide abstract]
    ABSTRACT: Mutations in the transcription factor hepatocyte nuclear factor (HNF)-4alpha are the cause of one form of maturity-onset diabetes of the young, MODY1. The HNF-4gamma is structurally related to HNF-4alpha and is expressed together with HNF-4alpha in pancreatic islets. We therefore tested the hypothesis that genetic variation in the HNF-4gamma gene (HNF4G) is associated with MODY in Japanese subjects. We screened the protein coding region of HNF4G (exons 3-11) for mutations in 57 unrelated Japanese subjects with MODY by amplifying each exon and adjacent intron region using the polymerase chain reaction (PCR) and specific primers and then directly sequencing the PCR products. The frequency of each variant was compared between patients with MODY and a group of non-diabetic subjects. We found ten sequence variants, two of these were located in exons: exon 6, a silent substitution in codon 144, c.432A/G and exon 7, a G-to-A substitution in codon 190 (c.570G/A) resulting in a conservative Met-to-Ile substitution (M/I190) in the putative ligand-binding region of HNF-4gamma protein. The remaining eight variants were located in introns. There was no significant difference in the frequency of these polymorphisms between subjects with MODY and non-diabetic control subjects. Genetic variation in the coding region of HNF4G is unlikely to be a major cause of MODY in Japanese people.
    Full-text · Article · Sep 2000 · Diabetologia
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Mutations in the hepatocyte nuclear factor 4alpha (HNF-4alpha) gene are associated with one form of maturity-onset diabetes of the young (MODY1). The R154X mutation generates a protein lacking the E-domain which is required for normal HNF-4alpha functions. Since pancreatic beta-cell dysfunction is a feature of MODY1 patients, we compared the functional properties of the R154X mutant in insulin-secreting pancreatic beta-cells and non-beta-cells. The R154X mutation did not affect nuclear localisation in beta-cells and non-beta-cells. However, it did lead to a greater impairment of HNF-4a function in beta-cells compared to non-beta-cells, including a complete loss of transactivation activity and a dominant-negative behaviour. .
    Full-text · Article · Sep 2000 · FEBS Letters
  • Source

    Full-text · Article · Jul 2000 · Diabetologia

  • No preview · Article · Jul 2000 · Diabetologia
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Subjects with the Q268X mutation in the hepatocyte nuclear factor (HNF)-4alpha gene (RW pedigree/maturity-onset diabetes of the young [MODY]-1) have diminished insulin and glucagon secretory responses to arginine. To determine if pancreatic polypeptide (PP) secretion is likewise involved, we studied PP responses to insulin-induced hypoglycemia in 17 RW pedigree members: 6 nondiabetic mutation-negative [ND(-)], 4 nondiabetic mutation-positive [ND(+)], and 7 diabetic mutation-positive [D(+)]. Subjects received 0.08 U/kg body wt human regular insulin as an intravenous bolus to produce moderate self-limited hypoglycemia. PP areas under the curve (PP-AUCs) were compared among groups. With hypoglycemia, the PP-AUC was lower in the D(+) group (14,907 +/- 6,444 pg/ml, P = 0.03) and the ND(+) group (14,622 +/- 6,015 pg/ml, P = 0.04) compared with the ND(-) group (21,120 +/- 4,158 pg/ml). In addition, to determine if the beta-cell secretory defect in response to arginine involves amylin in addition to insulin secretion, we analyzed samples from 17 previously studied RW pedigree subjects. We compared the AUCs during arginine infusions for the 3 groups both at euglycemia and hyperglycemia as well as their C-peptide-to-amylin ratios. The D(+) and ND(+) groups had decreased amylin AUCs during both arginine infusions compared with the ND(-) group, but had similar C-peptide-to-amylin ratios. These results suggest that the HNF-4alpha mutation in the RW/MODY1 pedigree confers a generalized defect in islet cell function involving PP cells in addition to beta- and alpha-cells, and beta-cell impairment involving proportional deficits in insulin and amylin secretion.
    Preview · Article · Jul 2000 · Diabetes

Publication Stats

23k Citations
2,456.61 Total Impact Points

Institutions

  • 1987-2014
    • University of Chicago
      • • Department of Human Genetics
      • • Department of Medicine
      • • Department of Biochemistry & Molecular Biology
      Chicago, Illinois, United States
  • 2001
    • Concordia University–Ann Arbor
      Ann Arbor, Michigan, United States
  • 1987-2000
    • Howard Hughes Medical Institute
      • Molecular Biology Facility
      Ashburn, Virginia, United States
  • 1999
    • Institut de Biologie de Lille
      Lille, Nord-Pas-de-Calais, France
  • 1998
    • Haukeland University Hospital
      • Department of Pediatrics
      Bergen, Hordaland, Norway
  • 1997
    • Steno Diabetes Center
      • Epidemiology
      Gjentofte, Capital Region, Denmark
    • University of Iowa
      • Department of Pediatrics
      Iowa City, Iowa, United States
    • University of Illinois at Chicago
      Chicago, Illinois, United States
  • 1996
    • The Rockefeller University
      New York City, New York, United States
    • University of Texas Southwestern Medical Center
      Dallas, Texas, United States
  • 1995
    • Virginia Mason Medical Center
      Seattle, Washington, United States
    • Tulane University
      New Orleans, Louisiana, United States
    • Institut Louis Bachelier
      Lutetia Parisorum, Île-de-France, France
  • 1994
    • Hamamatsu University School Of Medicine
      Hamamatu, Shizuoka, Japan
  • 1993
    • Kyoto University
      • Graduate School of Medicine / Faculty of Medicine
      Kioto, Kyōto, Japan
    • University of Pennsylvania
      • Department of Pharmacology
      Filadelfia, Pennsylvania, United States
  • 1992
    • Oxford University Hospitals NHS Trust
      • Department of Haematology
      Oxford, England, United Kingdom
  • 1989
    • Wayne State University
      • School of Medicine
      Detroit, Michigan, United States
  • 1988
    • Vanderbilt University
      • Department of Biochemistry
      Nashville, MI, United States
  • 1985-1987
    • Roswell Park Cancer Institute
      • Department of Human Genetics
      Buffalo, New York, United States
    • Hokkaido University
      • Laboratory of Molecular Enzymology
      Sapporo, Hokkaidō, Japan
    • Washington University in St. Louis
      • Department of Medicine
      San Luis, Missouri, United States
  • 1986
    • Harvard University
      Cambridge, Massachusetts, United States
  • 1981-1986
    • University of California, San Francisco
      • Department of Biochemistry and Biophysics
      San Francisco, California, United States