Linkage and association studies of non-HLA susceptibility genes for insulin-dependent diabetes mellitus (IDDM)

MISSING IMAGE

Material Information

Title:
Linkage and association studies of non-HLA susceptibility genes for insulin-dependent diabetes mellitus (IDDM)
Physical Description:
viii, 84 leaves : ill. ; 29 cm.
Language:
English
Creator:
Bui, Marilyn Yuanxin Ma, 1962-
Publication Date:

Subjects

Subjects / Keywords:
Diabetes Mellitus, Type I -- genetics   ( mesh )
Insulin -- genetics   ( mesh )
Genes, Structural -- genetics   ( mesh )
Linkage (Genetics)   ( mesh )
Chromosomes, Human, Pair 11   ( mesh )
Chromosome Mapping   ( mesh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1995.
Bibliography:
Includes bibliographical references (leaves 72-83).
Statement of Responsibility:
by Marilyn Yuanxin Ma Bui.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 49849436
ocm49849436
System ID:
AA00011224:00001

Full Text














LINKAGE AND ASSOCIATION STUDIES OF NON-HLA SUSCEPTIBILITY
GENES FOR INSULIN-DEPENDENT DIABETES MELLITUS (IDDM)

















By

MARILYN YUANXIN MA BUI


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1995




































A dedication to Grandma and Katie:

My own secret inspiration
















ACKNOWLEDGMENTS


I would like to express my sincere appreciation to

the chairman of my supervisory committee, Dr. Jin-Xiong

She, for his most valuable scientific guidance and

financial support. I thank the members of my committee,

Drs. Noel K. Maclaren, Edward K. Wakeland, Margaret R.

Wallace, William E. Winter, and Thomas P. Yang, for their

generous consultation not only on my dissertation

research but also on my growth as a scientist. I thank

Ms. Nana Tian and De-Fang Luo for their technical

assistance. I thank Ms. Christy Myrick for her critique

on my manuscript. I am also very grateful to all my

friends at the Department of Pathology and Laboratory

Medicine, whose friendship meant a lot to me throughout

these five years.

Last but not the least, a very special appreciation

goes to my family, especially my husband Vinh Q. Bui Jr.

and my parents, for their unconditional love and support.


















TABLE OF CONTENTS


ACKNOWLEDGMENTS.........................................iii

ABSTRACT ........................ ........ ............. vii

CHAPTERS

1 INSULIN-DEPENDENT DIABETES MELLITUS (IDDM) IS
AN AUTOIMMUNE DISEASE OF INSULIN-PRODUCING
PANCREATIC BETA CELLS, AND IS INFLUENCED BY
MULTIPLE GENETIC AS WELL AS ENVIRONMENTAL
FACTORS ............... .... ..... ...... .... .......1

Insulin-Dependent Diabetes Mellitus...............1
Autoimmune Mechanisms ............................. 2
Environmental Factors ............................... 4
Genetic Susceptibility..............................4
The Role of the MHC............................. 5
The Role of the Insulin Gene (INS) Region........ 8
Significance of Genetic Studies of IDDM .......... 9
Difficulties in Mapping IDDM
Susceptibility Gene........................10
Strategies for Gene Mapping Studies..............10
Mapping IDDM Susceptibility Genes
by Association Studies .......................12
Mapping IDDM Susceptibility Genes
by Linkage Studies ...........................14
Microsatellite Genetic Markers ...................15
Specific Aims of This Research...................18

2 ANALYSIS OF THE INSULIN GENE
(INS) REGION .. ................................ 19

Introduction ............ .......................... 19
Materials and Methods............................21
Patients and Controls for










Association Study .............................21
Samples for Linkage Study ......................21
DNA Preparation ................................ 22
PCR Amplification .............................. 22
Genotyping of Polymorphisms in
the INS Region................................22
RNA Extraction and RT-PCR Analysis .............24
Association Analysis ........................... 25
Affected Sibpair Analysis......................25
Transmission/disequilibrium Test (TDT).........26
Results. .........................................26
There is Association Between
INS and IDDM ................................. 26
A 6.5 Kb Genomic Interval on lip
Confers IDDM Susceptibility..................28
There Is No Interaction Between
HLA and INS. .................... ............. 28
Affected Sibpair Analysis Reveals
Weak Linkage Between INS and IDDM
in Male Meioses ................. ..............32
TDT Reveals Sex Difference of
INS Transmission. .............................34
There Is No Segregation Distortion
of INS Transmitted to Unaffected
Children ..................................... 37
INS Is Biallelically Expressed in
Human Pancreatic Tissue......................39
Discussion ....................................... 39

3 MAPPING OF TWO NOVEL IDDM SUSCEPTIBILITY
INTERVALS (4q AND 6q) BY AFFECTED
SIBPAIR ANALYSIS. ............................. 46

Introduction.. ..................... ............. 46
Materials and Methods............................ 47
Affected Sibpair Families...................... 47
Microsatellite Markers .........................48
Genotyping ..................................... 48
Data Analysis .................................. 51
Results. .........................................52
Screen for Linkage on Several
Chromosomal Regions..........................52
Fine Mapping of Chromosome 4q Region...........53
Fine Mapping of IDDM8 on Chromosome 6q.........55
Genetic Heterogeneity According










to HLA-DR/DQ Status of the
Affected Sibpairs .............. .............. 60
Discussion .................. ..................... 61

4 DISCUSSION ...................................... 68

REFERENCE LIST.......................................... 72

BIOGRAPHICAL SKETCH .....................................84
















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of
the Requirements for the Degree of Doctor of Philosophy

LINKAGE AND ASSOCIATION STUDIES OF NON-HLA SUSCEPTIBILITY
GENES FOR INSULIN-DEPENDENT DIABETES MELLITUS (IDDM)

By

Marilyn Yuanxin Ma Bui

August, 1995

Chairman: Jin-Xiong She
Major Department: Pathology and Laboratory Medicine


Insulin-dependent diabetes mellitus (IDDM) is an

autoimmune disease of the insulin-producing pancreatic P

cells. Susceptibility to IDDM is influenced by a number

of genetic as well as environmental factors. Previous

studies have indicated that IDDM1 is located in the HLA

Class II region on chromosome 6p, and IDDM2 is in the

insulin gene (INS) region on llp15. These two regions

together explain less than 50% of the total familial

clustering of IDDM, suggesting the existence of other

susceptibility factors.

In this study, the insulin gene region was further

investigated as a candidate susceptibility factor by

association and linkage studies. The susceptibility

interval on llp15 was narrowed to within a 6.5 Kb region,


vii










which contains the INS gene and its associated VNTR.

Linkage between INS and IDDM was detected only in male

meioses using the affected sibpair method.

Transmission/disequilibrium test further confirmed the

gender-related bias with respect to linkage with INS.

Even though maternal imprinting was a very attractive

hypothesis to explain the observed bias, biallelic

expression of the INS gene in human fetal pancreatic

tissue suggested that the INS locus was not imprinted.

In order to search for additional susceptibility

genes, several chromosomal regions were screened with 50

highly polymorphic microsatellite markers in up to 25

affected sibpair families. Preliminary linkage evidence

was obtained for two chromosomal regions (4q and 6q).

Analysis of 104 affected sibpairs confirmed our initial

observation. These two regions were then mapped with

additional microsatallite markers spaced at 1-5 cM.

Linkage evidence for the 4q region (p=0.028) was weak in

the total data set. In contrast, strong linkage evidence

(p=0.001) was obtained for the 6q region in the vicinity

of D6S264. Together with the UK 96 data set, linkage

with the 6q region was established and the disease locus

has now been designated as IDDM8.


viii
















CHAPTER 1
INSULIN-DEPENDENT DIABETES MELLITUS (IDDM) IS AN
AUTOIMMUNE DISEASE OF INSULIN-PRODUCING PANCREATIC BETA
CELLS, AND IS INFLUENCED BY MULTIPLE GENETIC AS WELL AS
ENVIRONMENTAL FACTORS


Insulin Dependent Diabetes Mellitus


Insulin dependent diabetes mellitus (IDDM, or Type I

diabetes), is characterized by a prolonged, selective and

irreversible destruction of insulin-producing pancreatic

p cells; an absolute requirement for exogenous insulin;

and a young age of onset. IDDM is generally considered

to be a disorder of the developed world. Indeed, after

asthma, IDDM is the second most common chronic childhood

illness in industrialized countries.1 In the United

States, the prevalence of IDDM by the age of 20 years is

about 0.26 percent, the lifetime prevalence approaches

0.4 percent,2 and the average annual incidence of IDDM

between 1970 to 1988 under age 15 years was 13.8 per

100,000.3 Overall, it is estimated that with a

population of 250 million, one million Americans have

IDDM.4

Patients with IDDM depend on a lifelong supply of

insulin and medical attention. Although insulin

replacement increases life expectancy, the disease is










associated with severe macrovascular and microvascular

complications that include blindness and kidney failure.

For these reasons, both the quality and quantity of life

can be dramatically reduced for IDDM patients. A huge

economic burden is placed on the patients, their families

and society.5

IDDM is also a serious medical problem in the

developing world. Although the incidence of the disease

is lower in third-world countries, life expectancy is

substantially less. One of the main reasons for the

reduced life expectancy may be the lack of an insulin

supply. Essentially, IDDM is a lethal disease in third-

world countries.6

Although IDDM is an ancient and worldwide disorder,

the etiology and pathogenic mechanisms of P cell

destruction are not yet completely understood.

Significant progress has been made in the past decade

that has advanced our knowledge of the etiopathogenesis

of IDDM.


Autoimmune Mechanisms


The guidelines7 generally accepted for establishing

the diagnosis of an autoimmune disease are the following:

(1) The disease state can be transferred by the patients'

antibodies or T-cells. (2) The disease course can be

slowed or prevented by immunosuppressive therapy. (3) The










disease is associated with manifestations of humoral or

cell-mediated autoimmunity directed against the target

organ. (4) The disease can be experimentally induced by

sensitization to an autoantigen present in the target

organ, which presupposes knowledge of the target

autoantigen. According to these guidelines, there is

plentiful evidence8-10 demonstrating that the destruction

of p cells in humans is autoimmune in nature: (1) After

allogeneic bone marrow transplantation with a diabetic

donor, the recipient acquired diabetes.11 Similarly,

diabetes was observed after pancreas transplantation

between identical twins.12 (2) There are examples of

immunosuppressant-dependent survival of pancreatic grafts

in diabetic recipients12 and immunosuppressant

augmentation of the length of remission in new-onset

IDDM.13,14 (3) There is immune cells infiltration in the

pancreas (called insulitis).15 There are multiple

abnormalities of the immune system,16 such as changes in

the ratios of T-cell subsets,17 and the appearance of

autoantibodies to islet cell components.18 In spite of

the fact that the autoantigens of IDDM remain elusive,

because other evidence is overwhelming, it is generally

accepted that IDDM is a classic organ-specific autoimmune

disease. In this disorder, 0 cells are destroyed by T-

cell mediated mechanisms, and circulating autoantibodies

are markers of the ongoing disease process.19 There is

also evidence indicating that, well before the T-cell










mediated amplification and perpetuation phase of P cell

destruction, a series of events takes place in a non-

lymphocyte-dependent initial phase.20,21 It remains

possible that other pathogenic mechanisms, including

direct lysis of p cells by cytokines22 and macrophage-

mediated killing,23 may participate.


Environmental Factors


Although the environmental factors that may trigger

the development of 0 cell immunity are poorly defined,

the importance of the environment has been clearly

demonstrated by the following facts: (1) Genetically

identical twins are only 36% concordant.24 (2) There is

an increase in IDDM incidence in several countries where

there are important changes in the environmental

factors25-27 and among ethnic groups immigrated from lower

incidence countries.28 It remains unclear how

environmental factors contribute to IDDM susceptibility.

It is speculated that the environmental factors are

somehow required in the anti-P cell autoimmunity and

allow the expression of IDDM predisposing genes.27,29


Genetic Susceptibility


The basic concept of genetic susceptibility is that

our body's response to environmental factors triggering

the autoimmune process leading to diabetes is genetically










controlled. IDDM has long been known to be a hereditary

disease because of its familial clustering: (1) Up to 15%

of IDDM patients have a first-degree relative with the

disease.30 (2) The disease concordance rate is 36% in

identical twins.24 (3) The risk for siblings (6%) is

much greater than the population prevalence (0.4%). The

familial clustering ratio, defined by Risch31 as ks, has

been calculated to be 15 for IDDM (average lifetime

sibling risk of 6% divided by the population prevalence

of 0.4%).


The Role of the MHC


The human major histocompatibility complex (MHC) on

chromosome 6p encodes HLA class I molecules that are

present on the surface of all nucleated cells. The

function of class I molecules is to present antigenic

peptides to CD8 (cytotoxic or suppresser) T-cells. The

MHC also encodes three HLA class II molecules: HLA-DP,

DQ, and DR, that are expressed on the surface of antigen-

presenting cells. The function of class II molecules is

to present antigens to CD4 (helper) T-cells. Both CD4

and CD8 cells have unique T-cell receptors for antigens

on their surface, which are specific for particular

complexes of peptide antigens and HLA molecules. Given

the major role of MHC molecules in antigen presentation

to T cells, MHC genes are obvious candidate










predisposition genes for autoimmune diseases such as

IDDM. In fact, genes in the HLA class II complex are by

far the most important factors in determining genetic

susceptibility or resistance to IDDM.32 The HLA class II

susceptibility was first found associated with DR3 and

DR4.33 Recent studies have demonstrated that IDDM

susceptibility is most strongly associated with DQBl*0201

and DQB1*0302, while protection from IDDM is strongly

associated with DQB1*0602.32,34,35 Although trans-racial

studies have shown that the susceptible molecules and the

strength of their susceptibility appear to be different

in various populations,37,38 DQA1*0301 is found to be

significantly associated with IDDM in all ethnic groups

and has been considered a candidate susceptibility

factor.36

Attention has been drawn to the nature of the

residue at position 57 of the HLA DQP-chain.32,39,40 The

Asp residue is rarely found in diabetic patients as

compared to the general population and almost never in

homozygous state (double copy). This observation is

particularly interesting with respect to MHC-peptide

interactions. It was hypothesized that the DQ molecules

associated with IDDM susceptibility may preferentially

bind and present P cell derived peptides to trigger

otherwise energized T-cells, causing P cell

destruction.32










The DQa/P cis and/or trans heterodimeric

complementation hypothesis has been proposed to account

for the synergistic effects observed in DR3/4 and DR3/9

heterozygous genotypes.35,41,42 Individuals who are

homozygous for the DR3 or DR4 are at a much higher risk

than those who have only one copy of the susceptibility

alleles (eg. DR3/X and DR4/X heterozygotes). This

phenomenon suggests that the dose of susceptibility

antigens may influence the degree of disease

susceptibility.41 However the above DQ hypothesis is not

able to explain the complexity of HLA associations with

IDDM. Recently, Huang et al. suggested a unified

hypothesis for HLA associations and disease prevalence.43

This hypothesis was based upon the fact that HLA-encoded

susceptibility to IDDM is determined by the combined

effects of both DR and DQ molecules (i.e. by both

genotypic combinations and linkage disequilibria of DR

and DQ genes). So far, this hypothesis can explain the

majority (if not all) of the observed associations

between HLA and IDDM, and is fully consistent with the

known IDDM incidence rates across ethnic populations.

While the HLA genes seem to be the most important

susceptibility factors (ks = 3.1-4.5),44 they obviously

cannot account for the total genetic contribution to the

disease (,s = 15).31 This observation suggests that other

susceptibility factors must exist. In fact, previous










studies have indicated that the INS gene region may be an

IDDM susceptibility factor.45-48,72-75


The Role of the Insulin Gene (INS) Region


The insulin gene (INS) region on chromosome llp15

has received considerable attention as a candidate region

for IDDM. The contribution of INS region to IDDM

susceptibility was initially demonstrated as association

using a VNTR polymorphism at the 5' end of the INS

gene.45 Others have since confirmed this

association.46,48,72,73 However, the exact locus that may

be responsible for disease susceptibility remains

unknown. In addition, the linkage of INS to IDDM has

been a controversial issue.46-49 Julier et al.46 reported

that in a French population the polymorphisms in the INS

region were linked to IDDM only in HLA-DR-positive

individuals, especially in paternal meioses. However,

using the same analytical methods described by Julier,

different results were obtained in a British

population.48,75 Further studies are required to

investigate whether there is a gender-related bias of INS

in respect to linkage between the INS and IDDM.

The total number of loci contributing to IDDM

susceptibility is unknown. A theoretical calculation

indicates that HLA (Xs 3.1-4.5)44 may account for less

than one-third of the familial clustering of IDDM (Xs =










15);31 while INS (Xs 1.3-1.5)50 and HLA together (As "

4.4-6.0) can only explain less than 50% of the total

genetic influence. It appears that genetic factors

unlinked to the HLA and the INS are required to fully

account for the total familial clustering of the

disease.51 In fact, the p cell destruction in NOD mice

(a model of human IDDM) is controlled by at least ten

genes not linked to the MHC H-2 region.52,53 This

provides further support for the speculation of

additional susceptibility loci outside the HLA and INS

regions.


Significance of Genetic Studies of IDDM


Identification of the IDDM susceptibility genes is

extremely important, because it might lead better

prediction, prevention and treatment. If doctors were

able to identify people at risk for IDDM according to

their genetic profiles, they could possibly modify the

patients' exposure to environmental factors to prevent or

delay the onset of the disease. They could closely

monitor the patients and treat them at the first sign of

disease to postpone the progression to full-blown

diabetes so that the quality and quantity of the

patients' life could be improved.









Difficulties in Mapping IDDM Susceptibility Genes


A simple genetic disease is genetically controlled

by one gene, and is inherited according to Mendelian

Laws. In contrast, IDDM is clinically very heterogeneous

and is a complex and multifactorial disease which does

not follow Mendelian inheritance patterns. Factors that

contribute to the difficulties in mapping IDDM genes are:

(1) Substantial genetic heterogeneity (identical clinical

symptoms are caused by defects at two or more genetic

loci). (2) Unknown mode of inheritance and incomplete

penetrance of the disease. (3) Lack of large pedigrees

with multiple affected members. Finally, mapping of the

remaining polygenic susceptibility factors is difficult

because each has a small effect and requires the

development of more effective mapping strategies.


Strategies for Gene Mapping Studies


One strategy is to first study an analogous form of

IDDM in an animal model. Comparative mapping has

demonstrated that there are some regions of synteny (two

or more homologous genes are located on the same

chromosome region in two different species) in mouse and

humans. However, because of large differences in the

biology of mouse and humans, the effectiveness of gene

mapping based on syntenic regions is limited. Recently,

Todd and colleagues54 demonstrated that the magnitude of










the gene effect in an experimental backcross of NOD is

likely to correlate only weakly, at best, with the

expected magnitude of effect in humans. The reason is

that in humans the gene effect will depend more heavily

on disease allele frequencies than on the observed

penetrance ratios, while such allele frequencies are

variable.54 Hence, the major benefit from animal studies

may be a better understanding of the disease process

itself, rather than identification of susceptibility

regions through comparative mapping.

The second is a candidate gene strategy, in which

one selects candidate genes to seek association and

linkage between their polymorphisms and the disease.

When a candidate gene is implicated in the disease, the

coding sequences can be characterized and functional

studies can be carried out to shed light on the

pathological mechanism. Virtually any gene that affects

0 cell function or the operation of the immune system is

a potential candidate, such as the T-cell receptor, MHC

molecules, insulin, and cytokines. Other regions in the

human genome that may hold candidate genes are those

chromosomal segments homologous to IDDM regions of the

mouse genome.55 Historically, the candidate gene strategy

has been extremely successful in the study of the

genetics of diabetes. In fact, the involvement in IDDM

of both the HLA and INS genes were discovered using this

strategy. Another successful example was the discovery










of linkage of the glucokinase gene with early-onset non-

insulin-dependent diabetes mellitus (MODY) in several

European pedigrees.56'57 In at least one family, a

nonsense mutation in the glucokinase gene causes

disease.58

The third strategy is positional cloning. The

location of a disease gene is first identified by

association and linkage analyses using anonymous genetic

markers. Then, attempts to clone the gene can be

followed without any knowledge of the function of the

disease gene. Several disease genes, such as, the

Huntington's disease gene on chromosome 4,59 the cystic

fibrosis gene on chromosome 760 and the neurofibromatosis

1 gene on chromosome 1761 were successfully mapped using

positional cloning. These successes have a major impact

on risk prediction, counseling for prevention, and

ultimately gene therapy. Positional cloning thus has

great potential in identifying genes contributing to IDDM

susceptibility.


Mapping IDDM Susceptibility Genes by Association Studies


Association studies identify genetic markers close

to the disease genes. They are also important for

investigating the interactions between the disease genes

and for assessing the relative risks of various genotypic

combinations of disease genes in human populations.










There are two kinds of generally-applied association

studies. One is case-control analysis, and the other is

family-based linkage disequilibrium analysis. The

principle of a case-control association study involves

the comparison of the frequency of a genetic marker in

patients (cases) with the frequency of that marker in

normal controls from the same ethnic population. If an

association between a marker and a disease exists, the

genotypic frequencies will differ between the two study

groups.62 However, the marker should not have a

selective effect on the individual, which is an spurious

association between the disease and the marker.63

Candidate genes (by their nature of having some

importance in the pathway of disease) may have selective

effect. In this case, it is important to differentiate a

true association from a spurious association.

The transmission/disequilibrium test (TDT) evaluates

the transmission of presumably disease-associated alleles

from heterozygous unaffected parents to affected

children. The statistical properties of the family-based

TDT have been investigated by Spielman et al.64 This

analysis has been used in several studies46'48 and has

proven to be more sensitive than the affected sibpair

method for detecting linkage.so TDT has the advantage of

not requiring families with multiple affected members.

Thus, simplex families can be included in a study. Since

a case-control association study may give a false










positive result due to population stratification, TDT is

often used as an alternative association analysis. This

analysis can narrow the genetic intervals that contain

the susceptibility genes identified by linkage studies.


Mapping IDDM Susceptibility Genes by Linkage Studies


A linkage study maps genes by analyzing the

cosegregation of a genetic marker with the disease. The

principle of the approach is simple: in an affected

family, if the disease locus and another polymorphic

locus (often called the marker locus) are closely located

on the same chromosome, they are preferentially passed on

together rather than independently assorted at meiosis.

However, the application of this principle is

complicated.

The statistical techniques used in current linkage

analysis are mostly based on maximum likelihood

estimation and likelihood ratio testing, which requires

extended affected families, known mode of inheritance,

known penetrance values and disease frequency.

Unfortunately, for IDDM most of these parameters are

unknown and only few large pedigrees are available. Due

to the obvious heterogeneity of IDDM, it would be

impossible to attempt a classic linkage study by adding

together numerous small families. Thus the affected

sibpair method becomes a practical alternative. This










analysis only requires nuclear families of at least two

affected children and unaffected parents. It reflects

the idea that if two affected siblings share a given

allele more often than expected by chance, it supports

the hypothesis that the disease is linked to that

particular locus. This method has been widely used in

family-based epidemiological studies for detecting

linkage in non-Mendelian disorders.65 In fact, it was

successful in detecting linkage of the HLA region to

IDDM.66

The affected sibpair analysis can identify linkage

between a marker and a disease (or a disease trait) even

if the recombination distance is as large as 10-15 cM.

It thus allows us to localize genomic intervals that

contain susceptibility genes. Association studies can

then further narrow the susceptibility intervals. Once

one or more markers are found at a distance of less than

1 cM of the disease gene, they can be used as starting

points for positional cloning of the gene, or for

identification of candidate genes found in that interval.


Microsatellite Genetic Markers


An essential requirement for mapping IDDM

susceptibility genes is the availability of highly

polymorphic genetic markers. In general, the most useful

markers should be maximally informative and easiest to










genotype. Before 1988, DNA polymorphisms were limited to

restriction-fragment-length polymorphism (RFLPs) which

are based on nucleotide substitution. RFLPs are not very

informative, because they usually have a small number of

alleles67 and their polymorphism information content

(PIC) value is low. In addition, RFLPs are routinely

genotyped using restriction enzyme digestion, blotting,

and hybridization. This process is tedious, expensive,

labor intensive, uses a lot of DNA, and is time

consuming. The introduction of the polymerase chain

reaction (PCR) using thermostable DNA polymerase,

provided entirely new means of analyzing polymorphisms

and made practical the analysis of highly polymorphic

length variations in simple-sequence tandemly repeated

DNA. Because simple sequence repeats (SSRs) occur

frequently and randomly throughout the human genome and

are polymorphic, these elements have shown great utility

as genomic markers for genetic mapping. SSRs include

minisatellites/variable number tandem repeats (VNTRs) and

microsatellites. Microsatellites are oligonucleotide

tandem repeats, such as CA repeats and CT repeats. The

repeated unit of VNTRs is relatively longer than in

microsatellites. The informativeness of microsatellites

and VNTRs are very similar. The average PIC value for a

CA marker is 0.61, which is about twice the average PIC

for RFLPs.69,70 Microsatellites, however, have more

important advantages than VNTRs: (1) They are abundant










and uniformly distributed throughout the human genome.69

For example, there are an estimated to be 50,000 copies

of (TG)n repeat (n=10-60) sequences interspersed through

the human genome.69 Because of the advances in the Human

Genome Project, an international effort to first map and

eventually sequence the entire human genomes,

microsatellites of very high heterozygosity (70-90%) are

easily accessible. (2) They are usually less than 100 bp

in length and, therefore are easy to clone, sequence and

develop into a PCR assay. In genotyping these by PCR,

typically the forward primer is labeled using kinase; the

PCR products are detected on a polyacrylamide gel after

electrophoresis and radiographed. The potential of

automating the entire microsatellite typing process,

including data analysis, has made it feasible to analyze

the human genome to map IDDM susceptibility genes. (3)

microsatellite PCR primers are commercially available.

For example, Research Genetics currently offers over

4,000 markers and new markers are constantly being added.

These primers are ready to use, come with recommendations

for reaction conditions, and are reasonably priced. For

the above reasons, PCR-based highly polymorphic

microsatellites are obviously the markers of choice for

gene mapping.









Specific Aims of This Research


The aim of this research is to map non-HLA genomic

intervals containing IDDM susceptibility genes by

association and linkage studies. Previous studies have

demonstrated that genes in the human major

histocompatibity complex appear to have the greatest

effect on diabetogenesis. The literature suggests that

other promising loci are present on chromosome lip in the

vicinity of the insulin gene. My study was designed to

achieve the following aims:

1. To identify the susceptibility locus on

chromosome llp15 using case-control association analysis.

2. To investigate whether there is a gender-related

difference with respect to the linkage between the INS

region and IDDM, and if so, what is the molecular basis.

3. To perform a limited genome-wide search for IDDM

genes with highly polymorphic microsatellite markers

using affected sibpair analysis.

4. To confirm and replicate potential linkages with

a large number of affected sibpair families as well as

additional microsatellite markers.














CHAPTER 2
ANALYSIS OF THE INSULIN GENE (INS) REGION


Introduction


The INS region on chromosome llp15 is a 19 kb

interval spanning the tyrosine hydroxylase gene (TH), the

insulin gene (INS) and the insulin-like growth factor II

gene (IGF-2). Association between the INS region and

IDDM was first demonstrated using a VNTR polymorphism at

the 5' of the INS gene.45 The association was then

confirmed in many populations using additional

polymorphisms in the INS region.46,48,72,73 However, the

exact locus responsible for IDDM susceptibility remains

unknown.

Linkage of INS to IDDM has been demonstrated using

the affected sibpair analysis and the TDT test.46,48,74

Julier et al.46 studied a French population and first

reported that the polymorphisms in the INS region were

linked to IDDM only in HLA-DR-positive individuals,

suggesting an interaction between HLA and INS. This

effect was strongest in paternal meioses, suggesting a

possible role for maternal imprinting. However, using

the same analytical methods described by Julier,

transmission distortion (linkage) was observed in both










maternal and paternal meioses in a British

population.48,75

Therefore, in order to assess the strength of

association and potential interactions between the INS

and the HLA-DQBI loci, I studied five polymorphisms in

the INS gene and surrounding loci in a Caucasian diabetic

population ascertained from the South-Eastern United

States. My results indicate that the risks conferred by

INS are not significantly different according to HLA

genotypes, suggesting that there is no interaction

between the two genetic systems in my study group.

Furthermore, my analyses of the polymorphisms around the

INS gene region suggest that a 6.5 Kb interval on llp,

which contains the INS gene and its associated VNTR, is

responsible for IDDM susceptibility.

In order to investigate the controversy of the

gender-specific effect, I analyzed the INS Pst I +1127

polymorphism46 in 123 multiplex families. Linkage was

only detected in male meioses using either the affected

sibpair analysis or the TDT test. In order to test the

maternal imprinting hypothesis, RT-PCR analysis was used

to reveal the expression of the INS gene in human fetal

pancreatic tissues. The biallelic expression, found by

this study, indicated that INS is not imprinted in the

human pancreas, suggesting that the observed gender-

related effect cannot be accounted for by maternal

imprinting.









Materials and Methods



Patients and Controls for Association Study


All patients and controls used in the association

study were unrelated US Caucasians of Northern European

descent. The patients had IDDM clinically confirmed

using the criteria of the National Diabetes Data Group.76

They were phenotyped for autoimmune endocrine diseases

and the associated relevant autoantibodies. The healthy

control subjects were negative for islet cell

autoantibodies (ICA) and had no immediate family history

of diabetes.


Samples for Linkage Study.


A total of 123 Caucasian families with two or more

affected sibs were used for haplotype sharing analysis.

In this data set, 53 families were from the Human

Biological Data Interchange (HBDI), 8 were from Dr.

Spielman at the University of Pennsylvania and 62 were

from the South-Eastern USA (mostly Florida). These

multiplex families and 15 additional simplex families

from North-Central Florida were used for the

transmission/disequilibrium test.









DNA Preparation


Lymphocytes were purified from 10-20 ml of whole

blood using Ficoll-Hypaque. DNA was purified using

proteinase K digestion, phenol/chloroform extraction, and

isopropanol precipitation.


PCR Amplification


All PCR amplifications were performed with a

template of 50-100ng of genomic DNA in a 25-50 gl

reaction volume containing 50 mM KC1, 10 mM Tris-Cl pH

8.3, 1.5 mM MgC12 and 60 JM of all four dNTPs, 0.2 ng of

each primers and 0.5 u of Taq polymorase (Boeheringer)

Samples were subjected to 35 cycles of 30 seconds at 94

C for denaturing, 30 seconds at optimum temperatures for

annealing and 30 seconds at 72 *C for extension, using an

automated thermal cycler (9600 Perkin-Elmer-Cetus,

California). An additional 2 minutes were added to the

denaturing step of the first cycle as well as the

extension step of the last cycle.


Genotyping of Polymorphisms in the INS Region


The five primers used to analyze polymorphisms in

the INS region are listed in Table 2-1. These










Table 2-1. List of PCR primers used in association
study.



Polymorphisms Detection Method Primers Tm (OC)



-4217 (T,C) Pst I TH5/TH6 66

+1127 (C,T) Pst I INS3/INS2 64

+1428 Fok I INS3/INS2 64

+2336 (5bp del) 6% acrylamide INS55/INS41 66

+3580 Msp I IGF2-1/IGF2-2 64


Primer
TH5:
TH6:
INS3:
INS2:
INS55:
INS44:
IGF2-1:
IGF2-2:


sequences
GTG ACG
ACC CAG
GGA ACC
AGC CCA
ACC TTT
GGT GAG
CCC CAT
GGG AGA


(5'-3') :
CCA AGG
CAG CCC
TGC TCT
GCC TCC
CCT GAG
CTC CTG
GTG AGC
CTT GGG










polymorphisms were detected using restriction digestion

with appropriate enzymes, followed by agarose gel

electrophoresis and staining with ethidium bromide.


RNA Extraction and RT-PCR analysis


RNA was extracted from pancreatic tissue of 4

aborted human fetuses between the ages of 55 and 113 days

using a protocol modified from Chomczynski and Sacchi.67

The tissues were briefly homogenized in solution D (4M

guanidinium isothiocyanate, 0.75M Na citrate pH 7, 0.5%

sarcosyl). RNA was then purified with phenol/chloroform

extraction and precipitated with isopropanol. Total RNA

(2 gg) was used for cDNA synthesis using reverse

transcriptase and oligo-dT priming. An aliquot of cDNA

(2 pl, 1/20 volume) was then used as template for PCR

amplification of the insulin cDNA. The forward primer

(INS7: 5'- CTACACACCCAAGACCCGC-3') is located at the 3'

end of exon 1 and the reverse primer (INS8: 5'-

TGCAGGAGGCGGCGGGTGT-3') is located in the 3' untranslated

region. PCR was done using conditions described above.

The optimum annealing temperature was 60 "C. These two

primers amplify a fragment of 227 bp from cDNA and a

fragment of 1003 bp from genomic DNA (including 786 bp of

intron 1 sequences). Thus, the 227 bp product amplified

from cDNA should not contain any contamination from

amplified genomic DNA, if any was present in the RNA










preparations. Since the amplified fragment contains the

Pst I +1127 polymorphic site, digestion of RT-PCR

products allowed me to distinguish the two INS alleles.


Association Analysis
2
X tests were used to reveal the statistical

significance of the observed genotypic frequency

differences between patient and control groups. A p

value of less than or equal to 0.05, indicates

significant association between the marker and the

disease of interest. Relative risks (RR) were calculated

by the method of Woolf.71


Affected Sibpair Analysis


The inheritance of different alleles at a given

locus by affected children from their heterozygous

parents was analyzed using identity by descent (IBD).

One ibd was scored when the same alleles were shared by

the affected sibs. Zero ibd was counted when different

alleles were inherited by the affected siblings. Under

the hypothesis of no linkage, the random expectation
2
should be 50% for 1 ibd and 0 ibd respectively. A X

test was performed by comparing the observed sharing of

the INS alleles in affected sibs with random expectation.

When deviation from random expectation is statistically










significant, linkage of the INS polymorphism and the

disease is indicated.


Transmission/disequilibrium Test (TDT)


TDT evaluates the transmission of the presumably

disease-associated INS allele from heterozygous parents

to their affected offspring. If there is linkage of INS

with IDDM, statistically more disease-associated INS

alleles should be transmitted.


Results



There Is Association Between INS and IDDM


A total of 343 IDDM patients (220 sporadic cases and

123 probands in multiplex families) and 272 normal

controls were genotyped for the Pst I +1127 polymorphism

3' of the INS gene. The frequencies of the INS +/+

homozygous genotype were found to be significantly

increased in both sporadic patients and probands of

multiplex families above controls (Table 2-2). These

results confirmed association between INS and IDDM. The

disease-associated allele is the INS + allele.

The relative risk (RR) conferred by the INS gene was

2.1, suggesting that individuals with the INS +/+ are

twice as likely to develop the disease as those with the

INS +/- or -/- genotypes.






27


Table 2-2. Genotypic frequencies of the Pst I +1127
polymorphism and relatives risks conferred by the INS +/+
genotype in sporadic patients and probands.



INS genotypes RR 2 p

+/+ +/-,-/-



Controls 167 (61.4%) 105 (38.6%)



Sporadics 167 (75.9%) 53 (24.1%) 2.0 11.7 0.0006



Probands 98 (79.7%) 25 (20.3%) 2.5 12.8 0.0004



Combined 265 (77.3%) 78 (22.7%) 2.1 18.3 0.00002









A 6.5 Kb Genomic Interval on lip Confers IDDM
Susceptibility


Five distinct genomic polymorphisms within the INS

gene and the surrounding region were analyzed (Table 2-3)

to define the susceptibility interval on chromosome

llp15. 159 normal controls and 197 unrelated diabetic

patients were genotyped using the polymerase chain

reaction and restriction enzyme digestion. Two

polymorphisms within INS (+1127 Pst I and +1428 Fok I)

were in complete linkage disequilibrium and demonstrated

significant associations with IDDM (RR = 2.0, P < 0.005).

However, the -4217 Pst I polymorphism in the TH gene (5'

of the INS VNTR) was not significantly associated with

IDDM, defining the 5' boundary of the susceptibility

interval on chromosome llp. Similarly, the +2336 5 bp

deletion and + 3580 Msp I polymorphisms were also not

significantly associated with IDDM, thus defining the 3'

boundary of the susceptibility interval. The -4217 Pst I

site and the +2336 5 bp deletion site encompass a genomic

region of 6.5 Kb including the INS gene and its

associated VNTR (Figure 2-1), but excluding the TH and

the IGF2.


There Is No Interaction Between HLA and INS


To investigate the possible interactions between the

INS and HLA genes, the relative risks conferred by INS

were calculated according to their DQB1 genotypes.

























om
o L
o o




om
0 CO








0) OH O O

O NN HO


In v OD r
LA' m H









OLD'D OLD









a) 0 CO w



W m m 0n




SLn nU)








o! 0
HH H




CO O 03
a h inS

r, co wo 0
r- 1 m n
H H mN

++ ++















cci



I -H


rE
a)
)0 >4J
4J r-




E"
HU












jm


u0
0 0


> 0

(15




00


o n a)






-0 au )
(a41






- a) a0



a c n

(a
Sn m
- H !(
04
H >a


N' 0 -H


;J H (a) 04
3H a) a)
r- 4 u H













co




< 1




4+
N


v ri


14 Q




41
0 3 0









-in
-u I








CHR



C I0
~J s- b

m.









14










When 197 diabetic patients were subdivided into four DQ01

genotype categories (*0201/0302, *0302/0302 or *0302/X,

*0201/0201 or *0201/X, and X/X), the relative risks of

the INS +/+ homozygotes ranged from 1.6 to 2.4 (Table 2-

4). These results are very similar for the entire

patient population (RR = 2.1). Since IDDM susceptibility

is most strongly associated with *0201 and *0302 (the

relative risks conferred by *0201/0302 and *0303/0302 are

20.9 and 12.9, respectively),73 these results suggest

that there is no interaction between the HLA and the INS

loci.


Affected Sibpair Analysis Reveals Weak Linkage Between
INS and IDDM in Male Meioses


The Pst I + 1127 polymorphism was analyzed in 123

families containing at least two affected siblings

(ASPs). There were 42 informative parents (22 fathers

and 20 mothers) who were heterozygous for INS and whose

transmission of INS alleles to their affected children

can be unambiguously determined. In this data set, 27

affected sibpairs inherited identical INS alleles (scored

as 1 ibd) and 19 inherited different alleles. Under the

hypothesis of no linkage, 1 ibd and 0 ibd should be equal

(i.e. 23). In fact, there was no significant difference
2
in observed and expected ibd values in total meioses X =

(27-19)2/(27+19)=1.4. However, there were significantly







33


Table 2-4. Relative risks in diabetic patients conferred
by INS according to their HLA-DQBI status.


INS Status RR4 p

HLA-DQBI Status +/+ +/-,-/-

0201/0302 49 20 1.6 ns

0302/0302 or 0302/X 41 13 2.0 0.05

0201/0201 or 0201/X 41 11 2.4 0.05

X/X 17 5 2.2 ns

All 148 49 2.0 0.005


* The relative risks were computed using 97 (61%)
controls with the INS +/+ and 62 (39%) controls with
the INS +/- or -/-.










more (p=0.01) affected sibpairs that inherited identical

alleles than different alleles from their heterozygous

fathers (19 one ibd versus 6 zero ibd) (Table 2-5).

Thus, a weak linkage in male meioses was confirmed using

conventional haplotype sharing analysis among affected

sibpairs.


TDT Reveals Sex Difference of INS Transmission


All 123 multiplex and 15 simplex families were

combined for TDT. There were 56 informative heterozygous

parents for INS (31 fathers and 25 mothers). These

parents transmitted 103 alleles (69 allele + and 34

allele -) to their diabetic offspring (Table 2-6). Under

the hypothesis of no linkage, the expected number of +

and alleles transmitted is equal (i.e. 51.5). The X

was calculated using the formula (x-y) /(x+y), where x is

the number of the + alleles and y is the number of

alleles that are transmitted. The difference observed
2 2
was significant, X =(69-34) /(69+34)=11.9, p=0.0006

supporting linkage. In the case of INS, this study again

demonstrated that TDT is more sensitive than affected

sibpair analysis in detecting linkage.

To test whether there is sex difference in INS

transmission, paternal and maternal transmissions were

counted separately. Among 31 fathers heterozygous






35


Table 2-5. Affected sibpairs analysis at the INS locus.



Fathers Mothers Combined

IBD (1 : 0) IBD (1 : 0) IBD (1 : 0)



Observed 19 : 6 8 : 13 27 : 19

Expected 12.5 : 12.5 10.5 : 10.5 23 : 23

X2 6.7 1.2 1.4

p 0.01 ns ns










Table 2-6. Transmission-disequilibrium test
- alleles transmitted from heterozygous (+/-
mothers to affected children.


of INS + and
) fathers or


Fathers Mothers Combined
+ + -+ -

Observed 44 12 25 22 69 34

Expected 28 28 23.5 23.5 51.5 51.5

2 18.3 0.2 11.9

p 0.00002 ns 0.0006










for INS, 37 + alleles and 12 alleles were transmitted

to their diabetic children. This is significantly
2
different from random expectation: X =(44-

12)2/(44+12)=18.3, p <0.00002. Among 25 mothers

heterozygous for INS, 25 + alleles and 22 alleles were

transmitted. This difference is not significant from

random expectation. These results suggest that there is

a transmission distortion of INS from fathers to diabetic

children.


There Is No Segregation Distortion of INS Transmitted to
Unaffected Children


The difference found with the TDT could be due to an

"artifact" of meiotic segregation distortion. If it was

an artifact, one would expect to see such distortion in

both affected and unaffected offspring. The INS

transmissions from heterozygous parents to unaffected

sibs within diabetic families, as well as to normal

children in non-diabetic families were analyzed. As

shown in Table 2-7, among the 29 informative individuals

who inherited INS alleles from heterozygous fathers, 13

were unaffected children in diabetic families (6 +

alleles and 7 alleles) and 16 were children in normal

families (11 + alleles and 5 alleles). Among the 32

informative individuals who inherited INS alleles from

heterozygous mothers, 11 were unaffected children in

diabetic families (3 + alleles and 8 alleles) and 21







38


Table 2-7. Observed and expected number of INS + and -
alleles transmitted from heterozygous fathers or mothers
(+/-) to normal children.


Fathers Mothers Combined

Alleles + + +



Observed 17 12 14 18 31 30



Expected 14.5 14.5 16 16 30.5 30.5



p ns ns ns










were children in normal families (11 + alleles and 10 -

alleles). The observed numbers of INS alleles

transmitted to non-diabetic children were not

significantly different from random expectation in male

or female meioses. These results do not support the

speculation of segregation distortion.


INS Is Biallelically Expressed in Human Pancreatic Tissue


Pancreatic tissue was obtained from four aborted

human fetuses. Their genomic DNAs were used as templates

to amplify the INS Pst I +1127 polymorphism site. Pst I

digestion of these PCR products revealed that two samples

(p8 and p9) were heterozygous for the INS + allele, while

the other two samples (p5 and p7) were homozygous for -

or + alleles respectively. RT-PCR analysis from p8 and

p9 mRNA revealed that both INS alleles were expressed, at

apparently equal level. This biallelic expression of INS

(Fig. 2-2) suggests that INS is not imprinted in human

pancreatic tissues.


Discussion


Both association and linkage studies have shown that

the genomic region on chromosome llp spanning the insulin

gene contains a susceptibility locus for

IDDM.45,46,48,72,75,77,78 There have been attempts to









Figure 2-2. Genomic polymorphism and expression of INS
in human pancreas. Genomic PCR: A fragment of 338 bp
which contains the Pst I +1127 polymorphism was amplified
from genomic DNA using primers INS3 and INS6. The
products were digested with Pst I restriction enzyme and
then eletrophoresed in a 3% agarose gel. The + alleles
only contain a monomorphic Pst I site and were digested
into two fragments (163 bp and 75 bp). The alleles which
contain a monomorphic site and the polymorphic Pst I +
1127 site were digested into three fragments (112, 51 and
75 bp). The samples P8 and P9 were heterozygous for INS,
as shown in the left panel. RT-PCR: A fragment of 227 bp
which contains the Pst I +1127 polymorphic site was
amplified from cDNA (derived from total RNA of human
pancreas) using the primers INS7 and INS8. RT-PCR
products were digested with Pst I. Digested products of
the alleles produced two fragments (197 bp and 30 bp
respectively). Products of + alleles were not digested
(227 bp). The samples P8 and P9 were biallelically
expressed as shown in the right panel.
















genomic PCR


M ) CO 0. I_
M a- CL CL CL


-/- +/- +/- +/+


-/- +/- +/- +/+


RT-PCR


1. wO .)
Q. Q. Q. a. M










the IDDM susceptibility factor on lip. In this study,

significant associations with IDDM were found for two

polymorphisms within the INS gene, while no significant

associations were found for the polymorphisms flanking

INS. A 6.5 Kb genomic region was defined by the Pst I -

4217 polymorphism in the TH gene and the +2336 deletion

polymorphism in the IGF2 gene. Similar observations were

obtained by Lucassen et al.78 After analyzing ten

polymorphisms in a 4.1 kb region extending from the INS

5' VNTR and across the insulin gene, they found

significant associations with IDDM. However, it is not

possible to specifically identify the IDDM susceptibility

site(s) since all of these polymorphisms are in strong

linkage disequilibrium. In addition, they were not able

to detect associations with IDDM at the INS flanking

regions, as in this study.

Both Lucassen's and my studies indicate that the

susceptibility interval on llp contains the INS gene and

its associated VNTR. However, the mechanism by which the

INS gene and/or its associated VNTR contribute to IDDM

susceptibility is unknown.

The possible interaction between HLA and INS has

been a controversial issue. Analyses of the French

population by Julier and Lucassen have suggested that the

association of INS with IDDM may be stronger in HLA*DR4

positive individuals, indicating interactive effects

between the INS and the HLA susceptibility loci.










However, my analyses showed that the risk conferred by

INS was similar in all HLA genotypes. Similar results

have also been reported in three other studies.48,72,75

These observations suggest that there are no interactions

between HLA and INS.

Risk assessment is an important aspect of genetic

studies of IDDM. At the INS locus, the absolute risk for

general population is 0.0084, which is calculated by the

relative risk (2.1) multiplied by the disease presence

(0.004). It seems that the INS gene has very minor

effect in IDDM susceptibility. In addition, the

predictibility of such assessment is limited in IDDM,

because the concordance of the disease in identical twin

pairs is as low as 36%.24 Therefore, it may be more

feasible to exclude the people who are not at risk rather

than to identify the people at risk to IDDM.

Two of the most important issues with respect to

linkage of INS and IDDM are: (1) is there a gender-

related bias, (2) if there is, what is the molecular

mechanism responsible for the sex difference. It appears

that a sex difference exists in most ethnically

heterogeneous populations, such as the French population

and the US populations. However, it does not exist in

ethnically more homogeneous populations, such as the

British population.48 There are several possible

explanations for the sex difference in transmission.

Random transmission of INS in non-diabetic families is










not consistent with the hypothesis of segregation

distortion and thus provides further evidence for

linkage. Since the maternal gene did not seem to be

important in IDDM susceptibility, the maternal gene may

not be expressed, in another word, may be imprinted.

Maternal imprinting could account for the observation,

and was an very attractive hypothesis because of

previously documented maternally imprinted genes in this

region.68,79-83 The IGF2 gene located 3' of INS is known

to be imprinted in the mouse68 and human.81-83 INS is

also known to be imprinted in the mouse yolk sac although

not in the pancreas.84 However, our RT-PCR analysis

revealed biallelic expression of INS in the pancreas of

human fetus. Similar results were also obtained from

adult pancreas.85 These results indicate that INS is not

imprinted in the pancreatic islets. Therefore, other

potential mechanisms must be responsible for the observed

sex difference.

It remains possible that the INS gene may be

maternally imprinted in human yolk sac. Another possible

mechanism could be mother-fetal interactions. This

hypothesis implies that maternal insulin would have an

impact on IDDM susceptibility, probably through its

effects on p cell mass of the fetus during the early

developmental stage. The third possibility is that the

neighboring locus IGF2 could be a candidate gene for

IDDM. Supporting evidence for this hypothesis is that










82,83
IGF2 is maternally imprinted.823 In addition, IGF2

encodes insulin-like growth factor 2 which is important

in embryogenesis and in p cell development. However,

two polymorphisms in the IGF2 gene (+2336 5 bp del and

+3580 Msp I) were not associated with IDDM in our

population and in a French population. These results

did not support the IGF2 hypothesis. Nevertheless, there

may exist other polymorphisms in the IGF2 gene that are

in linkage disequilibrium with the disease-associated INS

polymorphisms. Alternatively, the polymorphisms in INS

may affect the expression of the IGF2 gene, since these

two regions are only separated by a few kilo-base pairs.

Thus, further studies are required to understand which

gene in the INS-IGF2 region on lip is involved in IDDM

susceptibility, and by what mechanism this gene acts.















Chapter 3
MAPPING OF TWO NOVEL IDDM SUSCEPTIBILITY INTERVALS (4q
AND 6q) BY AFFECTED SIBPAIR ANALYSIS


Introduction


As mentioned above, the HLA class II genes and INS

gene together can only explain a portion of the total

genetic influence, suggesting that other IDDM

susceptibility factors exist. Indeed, linkage studies

have suggested that at least 10 genes are involved in the

expression of insulitis and/or diabetes in the nonobese

diabetic (NOD) mouse.52,86 Given the ethnic and genetic

heterogeneities of IDDM in humans, the number of

susceptibility genes is probably even higher. The

candidate gene approach has been successful in limited

cases such as INS. In the case of the majority of the

susceptibility genes, which are likely scattered

throughout the genome, linkage studies seem to be more

feasible. In fact, several groups have recently reported

localization of at least four other non-HLA IDDM

susceptibility regions44,87 using genome-wide linkage

mapping. In my mapping studies, a two-stage approach has

been applied. The first stage involved an initial

genome-wide screen using a subset of 25 Florida affected

sibpair families and 50 microsatellite markers located










throughout several chromosomal regions to obtain

preliminary linkage evidence. The second stage was to

replicate the linkages with 104 affected sibpair families

and additional microsatellite markers in those regions.

My study demonstrated that there is some evidence for

linkage in a novel region on chromosome 4q in the

vicinity of marker D4S1566 (p=0.028). Most importantly,

strong linkage evidence for the 6q25-q27 region was

obtained. Together with results from a UK data set,44

linkage to this second region was confirmed. This

disease locus has now been designated as IDDM8.


Materials and Methods



Affected Sibpair Families


Genomic DNA from a total of 104 American Caucasian

families was obtained. Each family had two affected

siblings and normal parents. In this set, forty-seven of

the samples were collected and ascertained in our hands

from the South-Eastern United States, mostly from North-

Central Florida (Florida data set). Forty-nine other

families were obtained from the Human Biological Data

Interchange (HBDI data set). Eight more were provided

generously by Dr. Richard Spielman at the University of

Pennsylvania.









Microsatellite Markers


Microsatellite markers were purchased from Research

Genetics. Distances between markers are measured in

centimorgans (cM). For markers that did not meet our

technical specifications, new markers were redesigned and

synthesized based on published sequence.


Genotyping


Highly polymorphic microsatellite markers were

genotyped using radioactive labeling of PCR primers and

denaturing polyacrylamide gel electrophoresis (Figure 3-

1). One of the PCR primers was end-labeled using y32P-ATP

and T4 polynucleotide kinase. PCR amplifications were

performed on 40 ng of genomic DNA (prealiquoted into a

96-well microtitre plate) in a 12 gl reaction volume

containing 50 mM KCL, 10 mM Tris-CL pH 8.3, 1.5 mM MgCl2,

and 60 iM of all four dNTPs, 0.2 ng of each primers and

0.5 u of Taq polymorase (Boehringer). Samples were

subjected to 27-30 cycles of 30 seconds at 94C for

denaturing, 30 seconds at the optimum annealing

temperature, and 30 seconds at 720C for extension using a

Perkin-Elmer-Cetus 9600 thermal cycler. After PCR

amplification, two volumes of sequencing loading solution

(0.3% xylene cyanol, 0.3% bromophenol blue, 10 mM EDTA pH

8.0 and 90% volume of formamide) were added. The samples









Figure 3-1. An example of genotyping D4S243 using
radioactive labeling of PCR primers and denaturing
polyacrylamide gel electrophoresis. Eleven affected
sibpair families were analyzed. F: Father. M: Mother.
Sl, S2: Affected siblings.


















D4S243


FM S1 S2
--
p U u

e










were then heated at 950C for 10 min to denature the DNA,

and 2-4 ul were immediately loaded onto a 6.5 %

polyacrylamide DNA sequencing gel. PCR products from 3-4

different markers with non-overlapping allele sizes

(amplified in separate reactions) were combined together

before loading to genotype multiple markers

simultaneously. Alternatively, in some cases products of

the same marker (but different samples) were loaded four

times (each separated by 30-60 min). Multiplexing of

different markers or multiple loading of products from

the same marker greatly increased the efficiency of

genotyping.


Data Analysis


A 2 test was used to determine the statistical

significance of the excess of gene sharing by affected

sibpairs. The X2 was calculated using (1 ibd-0 ibd)2/(1

ibd + 0 ibd), with one degree of freedom. A p value less

than or equal to 0.05 suggests linkage. In order to

detect potential linkages, correction for multiple

comparisons was not performed.

The maximum lod score (MLS) statistic T was

calculated according to Risch88 using the following

equation: T= (Ni) [loglo (Ni/0.5N)] + (NO) [loglo(No/0.5N)] .

Where N is the total number of informative meioses

(Ni+N) N1 and No are the observed number of affected










sibpairs sharing 1 or 0 alleles respectively. The random

expectation for 1 ibd and 0 ibd is 50% respectively. A

MLS value of 1.0 indicates linkage.

To increase the informativeness of these families,

informative flanking markers were used to deduce the

transmission of alleles from homozygous parents (referred

to as haplotyping). Haplotyping analyses were performed

using markers spaced at less than 5 cM to minimize the

possibility of double recombinations. Percent of gene

sharing (PGS) was calculated by the formula 1 ibd/(l ibd

+ 0 ibd).


Results



Screen for Linkage on Several Chromosomal Regions


Initially, up to twenty-five of the Florida families

were analyzed for 50 microsatellite markers randomly

chosen throughout several chromosomal regions. Among

these regions, some were syntenic to IDDM genes in NOD

mouse, some encampass candidate disease genes in humans.

As expected, the ibd values drawn from the 25 sibpairs

were not sufficient to claim linkage. For example, IL2RB

on 22q had a p value of 0.01 in the first 25 affected

sibpairs, but linkage disappeared when all 104 affected

sibpairs were analyzed. However some positive

preliminary data were obtained on two markers, D4S1566 on










4q and D6S264 on 6q. In addition to the linkage

evidence, these markers are in candidate gene regions.

The 4q region is syntenic to a mouse chromosome 3 region

which contains a IDDM gene (Iddm3) in the NOD mouse. The

6q region is in the neighborhood of the SOD2 and IGF2R

genes in human. It was obvious that these two regions

were worthy of further investigation.

The rest of 104 affected sibpairs were then

genotyped at D4S1566. Weak evidence of linkage was

obtained in the Florida data set (p=0.026) and the total

data set (p=0.028) (Table 3-1). The affected sibpairs in

the HBDI families had increased gene sharing compared to

random expectation but the excess of gene sharing was not

statistically significant. For D6S264, linkage evidence

was obtained in the Florida data set (p=0.03) and HBDI

families (p=0.0073). The combined data set gave a p

value of 0.0013 (Table 3-1). At this point, I proceeded

to more closely map the 4q and 6q regions to localize the

potential IDDM susceptibility genes.


Fine Mapping of Chromosome 4q Region


Seven additional microsatellite markers were

analyzed. They are D4S393, D4S1603, D4S349, D4S1596,

D4S243, D4S1545 and D4S622 (Table 3-2). Linkage evidence

was strongest at D4S1566 (p=0.028). Since this region






54


Table 3-1. Linkage Evidence from


Markers Data sets IBD (1:0)


D4S1566 FL 47 46 : 27
HBDI 49 48 : 40
UF 104 102 : 73

D6S264 FL 47 35 : 19
HBDI 49 52 : 28
UF 104 89 : 51


I


Genome-wide Screen.


PGS p MLS


63.0% 0.026 1.1
54.5% ns
58.3% 0.028 1.1

64.8% 0.030 1.1
65.0% 0.0073 1.6
63.6% 0.0013 2.3










has not been previously reported and is in the vicinity

of a candidate region, further studies in other

independent data sets will be necessary to confirm this

linkage.


Fine Mapping of IDDM8 on Chromosome 6a


As shown in Figure 3-2, twenty-one markers were

analyzed to localize the susceptibility gene on 6q. to be

within 1-2 cM of the given locus are flagged with "a".

The first six markers are in the interval of IDDM5 a

These markers encompass a region of 43 cM with an average

distance of 3-5 cM. The markers that are estimated round

ESR, which was first identified by Davies and

colleagues.44 In my study, the linkage at ESR was

surprisingly weak (MLS=0.9, which was only slightly

higher than its flanking markers). The strongest linkage

evidence was detected at D6S446, which gave a MLS value

of 2.8 (1 ibd = 116 and 0 ibd = 68). Since this marker

was more than 30 cM telemetric to ESR, it was speculated

that there may exist another IDDM predisposition gene in

the 6q region.

In order to verify this speculation, combining the

result from the 96 UK data set44 with ours, the total

MLS values were recalculated (Table 3-3). For ESR, the

combined results were (95 1 ibd, 59 0 ibd and MLS=1.8),










Table 3-2.

Marker

D4S393

D4S1603

D4S349

D4S1566

D4S1596

D4S243

D4S1545

D4S622


Fine mapping of the

D (cM) 1 ibd

0 34

1 79

2 93

5 102

6 83

7 101

12 74

13 63


region

0 ibd

38

59

66

73

70

84

71

48


around D4S1566.


X p

ns

2.9 0.09

4.6 0.03

4.8 0.028

1.1 ns

1.6 ns

ns

ns











Figure 3-2. Schematic presentation of the locations of

IDDM5 and IDDM8. The plot was based on the data in Table

3-4.























IDDM8


IDDM5
2.0 -


40
Distance (cM)


I klI 41










Table 3-3. Fine mapping of IDDM8 on chromosome 6q.


Markers D (cM) 1 ibd 0 ibd MLS MLS(+UK)


D6S311 0 88 85 0.0

D6S476 2 101 88 0.2

ESR 4 110 82 0.9 2.5

D6S440 6 109 90 0.4

D6S290 7 108 90 0.4

D6S442 10 110 89 0.5

D6S415 13 107 89 0.4

D6S437 15 107 82 0.7

D6S253 22 112 81 1.1 1.8

IGF2R =22 111 78 1.3

D6S220 = 23 111 76 1.4

D6S1008 2 25 108 81 0.8

D6S980 a 27 109 81 0.9

D6S396 a 29 109 83 0.8

D6S392 a 30 112 85 0.8

D6S264 32 117 75 2.0 3.4

D6S297 35 113 76 1.6

D6S503 a 37 112 76 1.5

D6S446 41 116 68 2.8

D6S281 42 107 63 2.5 2.0

TBP a 43 79 48 1.7










the total MLS was 2.5. For D6S264, a MLS value of 3.4

was achieved. In addition, for this marker, a p value of

0.001 was initially demonstrated in our data set.

Together with additional linkage evidence (p=0.01)

obtained in the independent UK 96 data set, it was very

clear that 6q encompassed another IDDM susceptibility

locus besides IDDM5. This second disease locus, near

D6S264, has been officially designated as IDDM8.


Genetic Heterogeneity According to HLA-DR/DO Status of
the Affected Sibpairs


To test HLA-associated heterogeneity, the identity

by decent (ibd) data of affected families were subdivided

according to HLA-DR/DQ haplotypes: sibpairs who shared 2

identical HLA haplotypes (HLA 2) and sibpairs who shared

1 or 0 HLA haplotype (HLA 1, 0). There were variations

in the proportions of genes shared by affected sibpairs

between the HLA 2 and HLA 1,0 categories for most marker

loci in this study. There were also variations of ibd

values in data subsets with different HLA-DR. However,

none of the comparisons reached statistical significance.

Therefore, the differences in ibd values between

different HLA categories in most cases is likely due to

random chance, or HLA's effect is too weak to be

detected.









Discussion


Mapping genes predisposing to complex disorders such

as IDDM is a difficult task. Suarez and colleagues89

have shown by computer simulation that if a number of

loci (each with a moderately small effect on disease) are

implicated, then linkage will be difficult to detect and

to replicate. The difficulty is due to heterogeneity

expected between data sets, or even within studies. In

monogenic diseases, the generally accepted norm for

linkage is a LOD score of 3 (p<0.001). Previous

studies44,87,90 have shown that this norm can not be

effectively achieved in studies of diseases with

substantial genetic heterogeneity. The reason is that

weak linkages could easily be missed even with 100 or

more affected sibpairs. Lander and Schork have suggested

that a p value of 3x10-5 (or MLS=3.6) is required to claim

a true linkage (confident at the 5% level) when the human

genome is examined.91 Such criteria may be difficult to

apply to complex diseases such as IDDM, because pooling

of different data sets in light of substantial genetic

heterogeneity may create serious problems.

Alternatively, Davies and colleagues have suggested

guidelines for statistic significance: 1) to obtain a p

value of 0.001 in the initial data set. 2) to replicate

this linkage in another independent data set with a p

value of 0.05.44 However, the false positive rate such










criteria is not yet known. In general, it is accepted

that less stringent criteria should be applied for the

initial establishment of linkage for complex diseases and

more stringent criteria should then be applied to confirm

the susceptibility genes. Therefore, I have reported any

linkage evidence when p is less or equal to 0.05. Even

though such evidence is not strong considering the number

of markers tested, any marker that indicates linkage in

one data set should be further investigated.

The linkage evidence for D4S1566 was novel and

warrants further studies in other independent families.

Linkage evidence for IDDM8 in my data set (MLS=2.8 for

D6S446 and MLS=2.0, p=0.001 for D6S264) and the weak

evidence in the UK data set (MLS=1.4, p=0.01 for D6S264)

together establish the presence of a disease locus in the

6q region using the criteria of Davies et al. When the

UK data set and my data set were combined, linkage

evidence for D6S264 (MLS=3.4) almost reached the

stringent criteria (MLS=3.6) suggested by Lander and

Schork. Since D6S264 is 28 cM more telomeric than ESR

(IDDM5), this study suggests that there are probably two

distinct IDDM genes on 6q (IDDM5 near ESR and IDDM8 near

D6S264-D6S446). This conclusion is also supported by the

UK data set. Since a 95% confidence interval is defined

as the region that contains all markers having a MLS

value greater than or equal to MLS,, 1.4,92 all markers

that have a MLS of 1.4 (i.e. 2.8-1.4=1.4) are in the 95%










confidence interval of IDDM8. Thus, IDDM8 is probably

located in the interval telomeric to D6S220.

There are two observations worthy of notice. First,

there was a fluctuation of MLS values along the 6q

region. This observation is consistent with the allele-

sharing of a complex genetic trait.92 In the situation

of a complex trait, the MLS follows a random walk in the

neighborhood of its peak, with steps occurring at

transitions between sharing and nonsharing. Second, the

percentage of genes shared by affected sibpairs was 62.5%

in the UK data set, which is very similar to that

observed in my USA families (62.2%) If these

observations can be confirmed in other independent

families, IDDM8 may be one of the most important

susceptibility genes for IDDM in addition to the HLA

class II genes. The contribution of a single disease

locus to the total X. can be estimated from the ratio of

the expected proportion of affected sibpairs sharing no
so
alleles (0 ibd=0.25) and the observed proportion.o In

fact, the X, conferred by IDDM8 was estimated to be 1.8,

which was higher than other non-HLA susceptibility genes

(ks = 1.5, 1.4, 1.6, 1.2 and 1.3 for IDDM2, IDDM3, IDDM4,
96
IDDM5, and IDDM7). IDDM8 is thus the most important

IDDM susceptibility factor other than HLA.

In order to investigate the characteristics of the

potential IDDM8, the evidence of linkage for IDDM8 was

analyzed according to parent-of-origin status. As shown






64



in Table 3-4 and Figure 3-3, it appeared that linkage for

IDDM8 was only detected in maternal meioses but not in

paternal meioses. Since the paternal gene did not seem

to be important in IDDM susceptibility, the paternal gene

may not be expressed, suggesting a possible role for

paternal imprinting.










Table 3-4. Evidence of paternal imprinting at IDDM8.

Markers D (cM) Paternal Meioses Maternal Meioses
1 ibd 0 ibd MLS 1 ibd 0 ibd MLS

ESR 4 54 42 0.3 56 40 0.6

D6S437 15 55 40 0.5 52 42 0.2

D6S253 22 53 44 0.2 59 37 1.1

IGF2R 22 53 45 0.1 58 33 1.5

D6S220 23 53 44 0.2 58 32 1.7

D6S1008 25 52 46 0.1 56 35 1.1

D6S980 27 53 46 0.1 56 35 1.1

D6S396 29 52 46 0.1 57 37 0.9

D6S392 30 51 50 0.0 61 35 1.6

D6S264 32 53 43 0.2 64 32 2.4

D6S297 35 53 43 0.2 60 33 1.7

D6S503 37 53 42 0.3 59 34 1.5

D6S446 41 58 36 1.1 58 32 1.7

D6S281 42 54 31 1.4 53 32 1.1

TBP 43 38 24 0.7 41 24 1.0









Figure 3-3. Schematic presentation of evidence for
paternal imprinting at IDDM8. The plot was based on the
data in Table 3-4.










Maternal Meloses


IDDM8


40
Distance (cM)


Paternal Meloses


0 10 20 30 40
Distance (cM)


MLS


MLS

3.0

2.5

2.0

1.5
















CHAPTER 4
DISCUSSION

Three years ago, I set out to answer three

questions: (1) How many genes may contribute to IDDM

susceptibility? (2) Where are they located? (3) How can

they be identified? To date, most of these questions

have been at least preliminary answered.

Genetic susceptibility to IDDM is complex, with HLA

class II genes on chromosome 6p21 (IDDMI) as the major

locus, with the insulin (INS) gene on chromosome llp15

(IDDM2) as a minor locus, and with at least five

additional minor loci on chromosomes 15q (IDDM3),90 llq

(IDDM4),87 6q (IDDM5),44 2q (IDDM7)50,93 and 6q (IDDM8).96

For IDDM1, the genetic determinants are the

polymorphisms within the peptide-binding sites of the

HLA-DQ and -DR molecules, but the identity of other

disease-predisposing mutations remain to be identified.

For IDDM2, the locus was mapped by this and

Lucassen's78 study to the INS gene and its associated

VNTR. However, the exact identity of IDDM2 remained

unknown until recently. Bennett et al.85 revealed that

IDDM2 is determined by the VNTR at the 5' of the INS gene

using a cross-match haplotype analysis. This notion is

now generally accepted. Since this polymorphism does not










encode any known gene products so that it must exert its

effect on IDDM susceptibility by regulating the

expression of other genes. I hypothesize that the VNTR

may regulate the transcription of its downstream genes,

such as INS and IGF2.

The INS-associated VNTR is a 14 bp repeat sequence

located in the promoter of the INS gene and is 365 bp

from the INS's transcription initiation site. This

interesting location suggests that VNTR might be

essential in regulating the INS gene expression. Since

the INS gene encodes insulin (which may be an autoantigen

in the process of disease development), the effect of the

INS gene may be derived from increased insulin secretion

and thereby lead to an augmentation of the targeted

autoantigens expressed on pancreatic beta cells. There

is evidence to support this hypothesis. Recently,

Kennedy et al.94 demonstrated that the INS-associated

VNTR could be bound and activated by a transcription

factor Pur-1 in vitro. The same study was also able to

present preliminary evidence that the transcriptional

levels of reporter genes are correlated with allelic

variation within the VNTR. However, the VNTR-INS

hypothesis cannot explain the observed gender-related

transmission bias of IDDM2.

The next downstream gene to the INS is the IGF2 gene

which encodes a protein (insulin-like growth factor) that

is important in 0 cell development.83 In addition, this










gene is known to be maternally imprinted.82,83 This

evidence suggests the potential role for the IGF2 gene in

IDDM pathogenesis. Nevertheless, it remains possible

that both the INS and IGF2 genes are involved in the

VNTR's effects in IDDM.

The identity of IDDM8 is still unknown. In this

study, the paternal imprinting characteristic of IDDM8

was first identified. Recently, evidence suggests that

an imprinted gene on chromosome 6 may be involved in

transient neonatal diabetes mellitus (TNDM).95 This gene

appears to be important for pancreatic 0 cell

development. It remains to be seen whether the TNDM gene

is identical or related to IDDM8 on 6q. Another

candidate gene for IDDM8 is IGF2R. Since IGF2R exhibits

paternal imprinting in mice and in humans, it may be the

paternally imprinted factor on 6q. Intriguingly, IGF2, a

candidate gene for IDDM2 on llq15, is maternally

imprinted. The above information together suggests that

the IGF2-IGF2R hypothesis is a very attractive mechanism

for IDDM susceptibility and deserves further

investigation. In our lab, a microsatellite marker

located in the 3'-untranslated region of IGF2R was

examined by other colleagues using linkage disequilibrium

analysis. Although linkage disequilibrium was not

demonstrated, this does not exclude IGF2R as a candidate

for IDDM8. Further mutation analysis, especially in the

regulatory region, is of great importance.










Thus far seven susceptibility loci (IDDMI, IDDM2,

IDDM3, IDDM4, IDDM5, IDDM7 and IDDM8) have been

identified. What is their combined effect on the total

familial clustering of IDDM (?s=15)? IDDM1 is the major

locus for IDDM susceptibility, with a k. of 3.1-4.5.44

The X. for IDDM2 and IDDM7 are both 1.3.50 IDDM4 and

IDDM5 both have X,=1.1.44,87 For IDDM3, the Xs is 1.4.96

Finally, the X, for IDDM8 is 1.8. Therefore, the total X,

is 11.1-12.5, which is about 80% of the total familial

clustering of IDDM. Three conclusions can be drawn from

this calculation. First, IDDM is definitely

polygenically controlled. Second, it seems that most of

the IDDM susceptibility genes, if not all, have been

localized. The next logical step will be to reveal the

identities of these genes and to investigate how they

interact with one another and the environment to cause

disease. Third, since k, for IDDM8 is 1.8, which

accounts for a higher proportion of the familial

clustering of IDDM (i.e. higher X, value) than other non-

HLA susceptibility genes, IDDM8 may be the most important

non-HLA susceptibility factor.

The success in the localization of polygenic factors

of IDDM is a big leap for mankind in the journey of

conquering this ancient and worldwide disease. The

genetic studies of IDDM will ultimately have a great

impact on the prediction, prevention and treatment of the

disease.
















REFERENCE LIST


1. R. David, G. Ledlie, and B. Elliott. Early
environmental events as cause of IDDM. Diabetes 43:843-
850 (1994).

2. R.E. LaPorte and K.J. Cruickshanks. Incidence and
risk factors for insulin-dependent diabetes. In:
"National Diabetes Data Group's diabetes data compiled
1984", Department of Health and Human Services., NIH,
Bethesda, MD. pp. III-1-III-12 (1985).

3. M. Karvonen, J. Tuomilehto, I. Libman, and R.F.
LaPorte. A review of the recent epidemiological data on
the worldwide incidence of Type 1 (insulin-dependent)
diabetes mellitus. Diabetologia 36:883-892 (1993).

4. M.A. Atkinson and N.K. Maclaren. The pathogenesis
of insulin-dependent diabetes mellitus. N Engl J Med
331:1428-1436 (1994).

5. A.L. Notkins. The causes of diabetes. Sci Am
241:62-73 (1979).

6. K.G. Alberti. Insulin dependent diabetes mellitus:
A lethal disease in the developing world. BMJ 309:754-
755 (1994).

7. H.S. Jun and J.W. Yoon. Initiation of autoimmune
type 1 diabetes and molecular cloning of a gene encoding
for islet cell-specific 37 kd autoantigen. Adv Exp Md
Biol 347:207-220 (1994).

8. J.A. Todd. Genetic control of autoimmunity in type
1 diabetes. Immunology Today 11:122-129 (1990).











9. G.S. Eisenbarth. Type 1 diabetes mellitus: a
chronic autoimmune disease. N Enal J Med 314:1360-1368
(1986).

10. R. Wassmuth and A. Lernmark. The genetics of
susceptibility to diabetes. Clin Immunol Immunopathol
53:358-399 (1989).

11. E.F. Lampeter, M. Homberg, K. Quabeck, U.W.
Schaefer, P. Wernet, J. Bertrams, H. Grosse-Wilde, F.A.
Gries, and H. Kolb. Transfer of insulin-dependent
diabetes between HLA-identical siblings by bone marrow
transplantation. Lancet 341:1243-1244 (1993).

12. R.K. Sibley, D.E. Sutherland, F. Goetz, and A.F.
Michael. Recurrent diabetes mellitus in the pancreas
iso- and allograft: a light and electron microscopic and
immunohistochemical analysis of four cases. Lab Invest
53:132-144 (1985).

13. Canadian-European Randomized Control Trail Group.
Cyclosporin-induced remission of IDDM after early
intervention: association of 1 year of cyclosporin
treatment with enhanced insulin secretion. Diabetes
37:1574-1582 (1988).

14. G. Feutren, L. Papoz, and R.E. Assan. Cyclosporin
increases the rate and length of remissions in insulin-
dependent diabetes of recent onset: results of a
multicenter double-blind trial. Lancet 2:119-124 (1986).

15. A.K. Foulis and A. Clark. "Joslin's Diabetes
Mellitus" Lea & Febiger, Baltimore, MD., (1994).

16. D.W. Drell and A.L. Notkins. Multiple immunological
abnormalities in patients with type 1 (insulin-dependent)
diabetes mellitus. Diabetologia 30:132-143 (1987).

17. D. Faustman, G. Eisenbarth, J. Daley, and J.
Breitmeyer. Abnormal T-lymphocyte subsets in type 1
diabetes. Diabetes 38:1462-1468 (1989).











18. Y. Naparstek and P.H. Plotz. The role of
autoantibodies in autoimmune disease. Annu Rev Immunol
11:79-104 (1993).

19. J. Nerup, T. Mandrup-Poulsen, S. Helqvist, H.U.
Andersen, and T.E. Lorenzen. On the pathogenesis of
IDDM. Diabetologia 37:S82-S89 (1994).

20. T. Mandrup-Poulsen. On the pathogenesis of insulin-
dependent diabetes mellitus. Dan Med Bull 39:509-542
(1992).

21. J. Molvig. A model of the pathogenesis of insulin-
dependent diabetes mellitus. Dan Med Bull 39:509-541
(1992).

22. J. Nerup, T. Mandrup-Poulsen, and J.E. Molvig.
Insulin resistance and insulin-dependent diabetes
mellitus. Diabetes Care 17:S16-S23 (1987).

23. B. Appels, V. Burkart, and G.E. Kantwerk-Funke.
Spontaneous cytotoxicity of macrophages against
pancreatic islet cells. J Immunol 142:3803-3808 (1989).

24. P. Olmos, R. A'hern, D. Heaton, B.A. Millward, D.
Risley, D.A. Pyke, and R.D.G. Leslie. The significance
of the concordance rate for type 1 (insulin-dependent)
diabetes in identical twins. Diabetologia 31:747-750
(1988).

25. M.A. Metcalfe and J.D. Baum. Incidence of insulin
dependent diabetes in children aged under 15 years in the
British Isles during 1988. Br Med J 302:443-447 (1991).

26. J. Tuomilehto, M. Rewers, A. Reunanen, P. Lounamaa,
R. Lounamaa, E. Tuomilehto-Wolf, and H.K. Akerblom.
Increasing trend in type 1 (insulin-dependent ) diabetes
mellitus in childhood in Finland. Diabetologia 34:282-
287 (1991).











27. C.E. Drykoningen, A.L. Mulder, G.J. Vandrager, R.E.
LaPorte, and G.J. Bruining. The incidence of male
childhood type 1 (insulin-dependent) diabetes mellitus is
rising rapidly in the Netherlands. Diabetologia 35:139-
142 (1992).

28. D.P. Singal and M.A. Blajchman. Histocompatibility
(HLA) antigens, lymphocytotoxic antibodies and tissue
antibodies in patients with diabetes mellitus. Diabetes
22:429-432 (1973).

29. J. Bach. Insulin-dependent diabetes mellitus as an
autoimmune disease. Endocrine Reviews 15:516-542 (1994).

30. E. Wolf, K.M. Spencer, and A.G. Cudworth. The
genetic susceptibility to type 1 (insulin-dependent)
diabetes: Analysis of the HLA-DR association.
Diabetologia 24:224-230 (1983).

31. N. Risch. Assessing the role of HLA-linked and
unlinked determinants of disease. Am J Hum Genet 40:1-14
(1987).

32. E. Thorsby and K.S. Ronningen. Particular HLA-DQ
molecules play a dominant role in determining
susceptibility or resistance to type 1 (insulin-
dependent) diabetes mellitus. Diabetologia 36:371-377
(1993).

33. J.L. Tiwari and P.I. Terasaki. HLA and disease
associations. In: "Endocrinology", J.L. Tiwari,
Springer-Verlag, New York, pp. 183-231 (1985).

34. D.G. Owerbach and K.H. Gabbay. Primary association
of HLA-DQw8 with type I diabetes in DR4 patients.
Diabetes 38:942-945 (1989).

35. E. Thorsby and K.S. Ronningen. Role of HLA genes in
predisposition to develop insulin-dependent diabetes
mellitus. Ann Med 24:523-531 (1992).











36. J.A. Todd, C. Mijovic, J. Fletcher, D. Jenkins, A.R.
Bradwell, and A.H. Barnett. Identification of
susceptibility loci for insulin-dependent diabetes
mellitus by trans-racial gene mapping. Nature 338:587-
589 (1989).

37. T. Awata, T. Kuzuya, A. Matsuda, Y. Iwamoto, and Y.
Kanazawa. Genetic analysis of HLA class II alleles and
susceptibility to type I (insulin-dependent) diabetes
mellitus in Japanese subjects. Diabetologia 35:419-424
(1992).

38. C. Mijovic, D. Jenkins, K.H. Jacobs, M.A. Penny,
J.A. Fletcher, and A.H. Barnett. HLA-DQA1 and -DQB1
alleles associated with genetic susceptibility to IDDM in
a black population. Diabetes 40:748-753 (1991).

39. J.A. Todd, J.I. Bell, and H.O. McDevitt. HLA-DQ
beta gene contributes to susceptibility evidence for gene
trans-complementation. Nature 329:599-604 (1987).

40. D.J. Charron, V. Lotteau, and P. Turmel. Hybrid
HLA-DC antignes provide molecular evidence for gene
trans-complementation. Nature 312:157-159 (1984).

41. I. Khalil, I. Deschamps, V. Lepage, R. Al-Daccak, L.
Degos, and J. Hors. Dose effect of cis and trans-encoded
HLA-DQap heterodimers in IDDM susceptibility. Diabetes
41:378-384 (1992).

42. R. Buzzetti, L. Nistico, J. Osborn, C. Giovannini,
A. Chersi, and R. Sorrentino. HLA-DQA1 and DQB1 gene
polymorphisms in type I diabetic patients from central
Italy and their use for risk prediction. Diabetes
42:1173-1178 (1993).

43 H.S. Huang, J.T. Peng, J.Y. She, L.P. Zhang, and
J.X. She. HLA-encoded susceptibility to IDDM is
determined by DR and DQ genes as well as their linkage
disequilibria: a unified hypothesis for HLA association
and disease prevalence. Human Immunology (in press).









44. J.L. Davies, K. Yoshihiko, S. Bennett, J.B. Copeman,
H.J. Cordell, L.E. Pritchard, P.W. Reed, S.C.L. Gough, C.
Jenkins, S.M. Palmer, K.M. Balfour, B.R. Rowe, M.
Farrall, A.H. Barnett, S.C. Bain, and J.A. Todd. A
genome-wide search for human type 1 diabetes
susceptibility genes. Nature 371:130-136 (1994).

45. G.I. Bell, S. Horita, and J.H. Karam. A polymorphic
locus near the human insulin gene is associated with
insulin-dependent diabetes mellitus. Diabetes 33:176-183
(1983).

46. C. Julier, R.N. Hyer, J. Davies, F. Merlin, P.
Soularue, L. Briant, G. Cathelineau, I. Deschamps, J.I.
Rotter, P. Froguel, C. Boitard, J.I. Bell and G.M.
Lathrop. Insulin-IGF2 region on chromosome lip encodes a
gene implicated in HLA-DR4-dependent diabetes
susceptibility. Nature 354:155-159 (1991).

47. N.J. Cox, L. Barker, and R.S. Spielman. Insulin
gene sharing in sub-pairs with insulin-dependent diabetes
mellitus: No evidence for linkage. Am J Hum Genet
42:167-? (1988).

48. S.C. Bain, J.B. Prins, C.M. Hearne, N.R. Rodrigues,
B.R. Rowe, L.E. Pritchard, R.J. Richie, J.R. Hall, D.E.
Undlien, K.S. Ronningen, D.B. Dunger, A.H. Barnet and
J.A. Todd. Insulin gene region-encoded susceptibility to
type 1 diabetes is not restricted to HLA-DR4-positive
individuals. Nat Genet 2:212-215 (1992).

49. W.E. Winter, T. Chihara, and D. Schatz. The
genetics of autoimmune diabetes: Approaching a solution
to the problem. Am J Dis Child 147:1282-1290 (1993).

50. J.B. Copeman, F. Cucca, C.M. Hearne, R.J. Cornall,
P.W. Reed, K.S. Ronningen, D.E. Undlien, L. Nistico, R.
Buzzetti, R. Tosi, F. Pociot, J. Nerup, F. Cornelis, A.H.
Barnett, S.C. Bain, and J.A. Todd. Linkage
disequilibrium mapping of a type 1 diabetes
susceptibility gene (IDDM7) to chromosome 2q31-q33.
Nature Genetics 9:80-85 (1995).









51. J.A. Todd and S.C. Bain. A practical approach to
identification of susceptibility genes for IDDM.
Diabetes 41:1029-1034 (1992).

52. J.A. Todd, T.J. Aitman, R.J. Cornall, S. Ghosh,
Hall, C.M. Hearne, A.M. knight, J.M. Love, M.A. McAleer,
J. Prins, N. Rodrigues, G.M. Lathrop, A. Pressey, N.
DeLarato, L.B. Peterson, and L.S. Wicker. Genetic
analysis of autoimmune type 1 diabetes mellitus in mice.
Nature (Lond) 351:542-547 (1991).

53. M. Prochazka, E.H. Leiter, D.V. Serreze, and D.L.
Coleman. Three recessive loci required for insulin-
dependent diabetes in nonobese diabetic mice. Science
237:286-289 (1987).

54. N. Risch, S. Ghosh, and J.A. Todd. Statistical
evaluation of multiple -locus linkage data in
experimental species and its relevance to human studies:
application to nonobese diabetic (NOD) mouse and human
insulin-dependent diabetes mellitus (IDDM). Am J Hum
Genet 53:702-714 (1993).

55. J.H. Nadeau. Maps of linkage and synteny homologies
between mouse and man. Trends Genet 5:82-86 (1989).

56. P. Froguel, M. Vaxillaire, F. Sun, G. Velho, H.
Zouali, M.O. Butel, S. Lesage, N. Vionnet, K. Clement, F.
Fougerousse, Y. Tanizawa, J. Weissenbach, J.S. Beckmann,
G.M. Lathrop, P. Passa, M.A. Permutt, and D. Cohen.
Close linkage of glucokinase locus on chromosome 7p to
early-onset non-insulin-dependent diabetes mellitus.
Nature (Lond) 356:162-164 (1992).

57. A. Hattersley, R.C. Turner, A.M. Permutt, P. Patel,
Y. Tanizawa, K.C. Chiu, S. O'Rahilly, P. Watkins, and
J.S. Wainscoat. Type 2 diabetes is linked to the
glucokinase gene in a large pedigree. Lancet 340:54-55
(1992).










58. N. Vionnet, M. Stoffel, J. Takeda, K. Yasuda, G.I.
Bell, H. Zouali, S. Lesage, G. Velho, F. Iris, P. Passa,
and P.D.C. Froguel. Nonsense mutation in the glucokinase
gene causes early-onset non-insulin-dependent diabetes
mellitus. Nature (Lond) 356:721-722 (1992).

59. J.F. Gusella, N.S. Wexler, P.M. Conneally, S.L.
Naylor, M.A. Anderson, R.E. Tanzi, P.C. Watkins, K.
Ottina, M.R. Wallace, A.Y. Sakaguchi, A.B. Young, I.
Shoulson, E. Bonilla, and J.B. Martin. A polymorphic DNA
marker genetically linked to Huntington's disease.
Nature (Lond) 306:234-238 (1983).

60. L.C. Tsui, M. Buchwald, D. Barker, J.C. Braman, R.
Knowlton, J.W. Schumm, H. Eiberg, J. Mohr, D. Kennedy,
and N. Plavsic. Cystic fibrosis locus defined by a
genetically linked polymorphic DNA marker. Science
230:1054-1057 (1985).

61. M.R. Wallace, D.A. Marchuk, L.B. Anderson, R.
Letcher, H.M. Odeh, A.M. Saulino, J.W. Fountain, A.
Brereton, J. Nicholson, A.L. Mitchell, B.H. Brownstein,
and F.S. Collins. Type 1 neurofibromatosis gene:
identification of a large transcript disrupted in three
NF1 patients. Science 249:181-186 (1990).

62. S.S. Rich. Mapping genes in diabetes-genetic
epidemiological perspective. Diabetes 39:1315-1319
(1990).

63. R.C. Elston. The use of polymorphic markers to
detect genetic variability. In: "Phenotypic Variation in
Populations: Relevance to Risk Assessment", A.D.
Woodhead, Plenum, New York, pp. 105-112 (1988).

64. R.S. Spielman, R.E. McGinnis, and W.J. Ewens.
Transmission test for linkage disequilibrium: The insulin
gene region and insulin-dependent diabetes mellitus
(IDDM). Am J Hum Genet 52:506-516 (1993).

65. J. Ott. Methods of linkage analysis. In: "Analysis
of Human Genetic Linkage", The Johns Hopkins University
Press, Baltimore and London, pp. 54-80 (1991).









66. J.A. Todd. Genetic control of autoimmunity in type
1 diabetes. Immunol Today 11:122-129 (1990).

67. P. Chomoczynski and N. Sacchi. Single-step method
of RNA isolation by acid guanidium thiocyanate-phenol-
chloroform extraction. Anal Biochem 162:156-159 (1987).

68. T.M. DeChiara, E. Robertson, J., and A.
Efstratiadis. Parental imprinting of the mouse insulin-
like growth factor II gene. Cell 64:849-859 (1991).

69. J.L. Weber. Human DNA polymorphisms based on length
variations in simple sequence tandem repeats. In:
"Genome Analysis", K.E. Davies, CSHL, pp. 159-181 (1990).

70. J.L. Weber. Informativeness of human (dC-dA)n.(dG-
dT)n polymorphisms. Genomics 7:524-530 (1990).

71. J.L. Tiwari and P.I. Terasaki. The data and
statistical analysis. In: "HLA and disease
associations", J.L. Tiwari and P.I. Terasaki, Springer
Verlag, New York, pp. 18-27 (1985)

72. D. Owerbach and K.H. Gabbay. Localization of a type
I diabetes susceptibility locus to the variable tandem
repeat region flanking the insulin gene. Diabetes
42:1708-1714 (1993).

73. J.X. She, M.M. Bui, X.H. Tian, A. Muir, E.K.
Wakeland, B. Zorovich, L.P. Zhang, M.C. Liu, G. Thomson,
and N.K. Maclaren. Additive susceptibility to insulin-
dependent diabetes conferred by HLA-DQB1 and Insulin
genes. Autoimmunity 18:195-203 (1994).

74. D. Owerbach and K.H. Gabbay. Linkage of the
VNTR/insulin-gene and type I diabetes mellitus: Increased
gene sharing in affected sibling pairs. Am J Hum Genet
54:909-912 (1994).

75. B.J. Van-der-Auwera, H. Heimberg, A.F. Schrevens, C.
Van-Waeyenberge, J. Flament, and F.C. Schuit. 5' insulin
gene polymorphism confers risk to IDDM independently of
HLA class II susceptibility. Diabetes 42:851-854 (1993).









76. National Diabetes Data Group. Classification and
diagnosis of diabetes mellitus. Diabetes 28:1039-1057
(1979) .

77. G. Thomson, W.P. Robinson, M.K. Kuhner, S. Joe, and
W. Klitz. HLA and insulin gene associations with IDDM.
Genet Epidemiol 6:155-160 (1989).

78. A.M. Lucassen, C. Julier, J.P. Beressi, C. Boitard,
P. Froguel, M. Lathrop, and J.I. Bell. Susceptibility to
insulin dependent diabetes mellitus maps to a 4.1 kb
segment of DNA spanning the insulin gene and associated
VNTR. Nature Genetics 4:305-310 (1993).

79. P. Heutink, A.G.L. Van Der Mey, L.A. Sandkuijl,
A.P.G. Van Gils, A. Bardoel, G.J. Breedveld, M. Van
Vliet, G.J.B. Van Ommen, C.J. Cornelisse, B.A. Oostra,
J.L. Weber, and P. Devilee. A gene subject to genomic
imprinting and responsible for paragangliomas maps to
chromosome 11q23. Human Molec Genet 1:7-10 (1992).

80. Y. Zhang and B. Tycko. Monoalleleic expression of
the human H19 gene. Nature Genetics 1:40-44 (1992).

81. S. Rainier, L.A. Johnson, C.J. Dobry, A.J. Ping,
P.E. Grundy, and A.P. Feinberg. Relaxation of imprinted
genes in human cancer. Nature 362:747-749 (1993).

82. N. Giannoukakis, N. Deal, J. Paquette, C.G. Goodyer,
and C. Polychronakos. Paternal genomic imprinting of the
human IGF2 gene. Nature Genetics 4:94-97 (1993).

83. R. Ohlsson, A. Nystrom, S. Pfeifer-Ohlsson, V.
Tohonen, F. Hedborg, p. Schofield, F. Flam, and T.J.
Ekstrom. IGF2 is parentally imprinted during human
embryogenesis and in the Beckwith-Wiedemann syndrome.
Nature Genetics 4:94-97 (1993).

84. S.J. Giddings, C.D. King, K.W. Harman, J.F. Flood,
and L.R. Carnaghi. Allele specific inactivation of
insulin 1 and 2, in the mouse yolk sac, indicates
imprinting. Nature Genetics 6:310-313 (1994).









85. S.T. Bennett, A.M. Lucassen, S.C.L. Gough, E.E.
Powell, D.E. Undlien, L.E. Pritchard, M.E. Merriman, M.J.
Kawaguchi, F. Dronsfield, J. Pociot, N. Nerup, A.
Bouzekri, A. Cambon-Thomsen, K.S. Ronningen, A.H.
Barnett, S.C. Bain, and J.A. Todd. Susceptibility to
human type 1 diabetes at IDDM2 is determined by tandem
repeat variation at the insulin gene minisatellite locus.
Nature Genetics 9:284-292 (1995).

86. S. Ghosh, S.M. Palmer, N.R. Rodrigues, H.J. Cordell,
C.M. Hearne, R. Cornall, J., J.B. Prins, P. McShane, G.M.
Lathrop, L.B. Peterson, L.S. Wicker, and J.A. Todd.
Polygenic control of autoimmune diabetes in nonobese
diabetic mice. Nature Genetics 4:404-409 (1993).

87. L. Hashimoto, C. Habita, J.P. Beressi, M. Delepine,
C. Besse, A. Cambon-Thomsen, I. Deschamps, J.I. Rotter,
S. Djoulah, M.R. James, P. Froguel, J. Weissenbach, G.M.
Lathrop, and C. Julier. Genetic mapping of
susceptibility locus for insulin-dependent diabetes
mellitus on chromosome 11q. Nature 371:161-164 (1994).

88. N. Risch. Linkage strategies for genetically
complex traits. Am J Hum Genet 46:229-241 (1990).

89. B.K. Suarez, C.L. Hampe, and P. Van Eerdewegh.
"Genetic Approaches to Mental Disorders." American
Psychiatric Press, Washington, D.C. (1994).

90. L.L. Field, R. Tobias, and T. Magnus. A locus on
chromosome 15q26 (IDDM3) produces susceptibility to
insulin-dependent diabetes mellitus. Nature Genetics
8:189-194 (1994).

91. E.S. Lander and N.J. Schork. Genetic dissection of
complex traits. Science 265:2307-2048 (1994).

92. L. Kruglyak and E. Lander. High-resolution genetic
mapping of complex traits. Am J Hum Genet 56:1212-1223
(1995).






83


93. D. Owerbach and K. Gabbay. The HOXD8 locus (2q31)
is linked to type 1 diabetes. Diabetes 44:132-136
(1995).

94. G.C. Kennedy, M.S. German, and W.J. Rutter. The
minisatellite in the diabetes susceptibility locus IDDM2
regulates insulin transcription. Nature Genetics 9:293-
298 (1995).

95. R.S. James, J.A. Crolla, F.L. Sitch, P.A. Jacobs,
W.M. Howell, P. Betts, J.D. Baum, and J.P.H. Shield. An
imprinted genes) for diabetes. Nature Genetics 9:110-
112 (1995).

96. D.F. Luo, M.M. Bui, A. Muir, N.K. Maclaren, G.
Thomson, and J.X. She. Affected sibpair mapping of a
novel susceptibility gene to insulin-dependent diabetes
mellitus (IDDM8) on chromosome 6q25-q27. Am J Hum Genet
(in press) (1995).











BIOGRAPHICAL SKETCH


Marilyn Yuanxin Ma Bui was born in Beijing, China,

on August 10, 1962. She grew up with two brothers in a

family of career diplomats. After earning her medical

doctor degree in 1986, she joined the faculty of the same

medical school--Capital Institute of Medicine in Beijing,

China. In 1989, she came to the States. In 1990, she

entered the graduate program in the Department of

Pathology and Laboratory Medicine at the University of

Florida College of Medicine. Her Ph.D. degree was

conferred on August 12, 1995.











I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.


Jin-Xiong She, Chair
Assistant Professor of
Pathology and Laboratory
Medicine

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.


Noel K. Maclaren
Professor of Pathology and
Laboratory Medicine

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.


Edward K. Wakeland
Professor of Pathology and
Laboratory Medicine

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.


Margaret R. Wallace
Assistant Professor of
Biochemistry and Molecular
Biology










I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.


William E. Winter
Associate Professor of
Pathology and Laboratory
Medicine

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of philosophy.


Thomas P. Ya g
Associate Profes 4o of
Biochemistry an molecular
Biology

This dissertation was submitted to the Graduate Faculty
of the College of Education and to the Graduate School and
was accepted as partial fulfillment of the requirements for
the degree of Doctor of Philosophy.

August, 1995 / 'a L-ea


6ean, College of Medicine


Dean, Graduate School





































S11 ii 5ll 1111 1 i 1 ill111 1
3 1262 08554 6298