Autoantibodies related to type 1 diabetes in childrenliu.diva-portal.org/smash/get/diva2:392919/FULLTEXT01.pdfIn conclusion, measurement of autoantibodies related to type 1 diabetes

Linköping University Medical Dissertations

No. 1218

Autoantibodies related to type 1 diabetes

in children

Camilla Skoglund

Division of Pediatrics

Department of Clinical and Experimental Medicine

Faculty of Health Sciences, Linköping University

SE-581 85 Linköping, Sweden

Linköping 2011

© Camilla Skoglund 2011

ISBN: 978-91-7393-276-9

ISSN: 0345-0082

Paper I has been reprinted with permission of John Wiley and Sons.

Paper II and IV have been reprinted with permission of Elsevier Ltd.

During the course of the research underlying this thesis, Camilla Skoglund (formerly

Gullstrand) was enrolled in Forum Scientium, a multidisciplinary doctoral programme at

Linköping University, Sweden.

Printed in Sweden by Liu-Tryck, Linköping 2011

To my family

Abstract

Type 1 diabetes is an autoimmune disease resulting from destruction of the insulin producing

beta cells in the pancreas. The patients need life-long heavy treatment and still complications,

both acute and later in life, are common. The incidence of type 1 diabetes has increased

rapidly during the last decades, especially among young children. The disease can be

predicted by genes predisposing type 1 diabetes, mainly human leukocyte antigen (HLA)

genes, together with presence of autoantibodies to beta-cell antigens, where multiple

autoantibodies confer the highest risk. A number of immune system intervention trials are

now ongoing aiming to halt the progression of the inflammatory process in the beta cells.

This thesis aimed to investigate the prevalence and levels of autoantibodies in healthy

children and in children with type 1 diabetes. Another aim was to study different properties of

one of these autoantibodies, such as to which epitopes the antibodies bind and the distribution

of immunoglobulin (Ig)-G subclasses, after immunomodulatory treatment in children with

type 1 diabetes.

We found that positivity to autoantibodies against glutamic acid decarboxylase (GADA) and

tyrosine phosphatase like protein islet antigen-2 (IA-2A) was associated with HLA risk

genotypes in 5-year old children from the general population. HLA risk genotypes seemed

important for persistence of autoantibodies and for development of type 1 diabetes, while

emergence of autoantibodies, especially transient autoantibodies, seemed to be more

influenced by environmental factors. Improved methods for detection of autoantibodies are

needed, for prediction of diabetes and for identification of high-risk individuals suitable for

prevention treatments. Therefore, an assay for measurement of insulin autoantibodies (IAA),

based on surface plasmon resonance (SPR), was developed. The main advantages of this

method are that there is no need for labelling and that it is time-saving compared to the

traditionally used radioimmunoassay (RIA), but further development of the method is needed.

Treatment with GAD-alum (Diamyd) in children with type 1 diabetes has shown to preserve

residual insulin secretion. This clinical effect was accompanied by an increase in GADA

levels. We investigated the epitope reactivity of GADA in both GAD-alum and placebo

treated children, and found that binding to one of the tested epitopes was temporarily

increased after injection of GAD-alum. This result suggests that the quality of GADA was, to

some extent, transiently affected by the treatment. On the other hand, no changes in binding to

epitopes associated with stiff person syndrome (SPS) were observed, which together with the

lack of change in GAD65 enzyme activity further strengthens the safety of the treatment. We

also observed that the distribution of IgG subclasses was changed by GAD-alum treatment,

with a lower proportion of IgG1 and higher IgG3 and IgG4. Lower IgG1 and higher IgG4

suggest a temporary switch towards a protective Th2 immune response, which has previously

been observed in the same individuals for other immunological markers.

In conclusion, measurement of autoantibodies related to type 1 diabetes is an important tool

for studying the autoimmune process in pre-diabetic and type 1 diabetic children. In addition

to the use as markers of disease progression, the autoantibodies may be used for studying the

effects of immunomodulatory treatments on the humoral immune response.

Sammanfattning

Typ 1 diabetes är en allvarlig sjukdom som uppkommer när de insulinproducerande

betacellerna i bukspottskörteln förstörs via en autoimmun process. Patienterna måste

behandlas hela livet, med bl a dagliga insulininjektioner, och komplikationer är vanliga.

Förekomsten av typ 1 diabetes har ökat den senaste tiden, och den kraftigaste ökningen har

setts hos små barn. Det går att förutspå vilka individer som har störst risk att utveckla typ 1

diabetes, genom att studera vissa gener, framförallt de som kodar för humant leukocyt-antigen

(HLA), i kombination med mätning av autoantikroppar mot olika ämnen som finns i

betacellerna. Autoantikroppar kallas de antikroppar som reagerar mot kroppsegna ämnen.

Förekomst av flera olika autoantikroppar mot de insulinproducerande cellerna medför högst

risk att utveckla typ 1 diabetes. Ett antal studier pågår, som genom att påverka

immunsystemet med specifika ämnen, har som mål att stoppa den inflammatoriska processen

som dödar betacellerna hos de barn som håller på att få diabetes och hos de som redan har fått

diabetes.

Målet med den här avhandlingen var att undersöka förekomst och nivåer av diabetesrelaterade

autoantikroppar hos friska barn och hos barn med typ 1 diabetes. Därutöver studera olika

egenskaper hos en av dessa autoantikroppar, t e x olika bindningsställen (epitoper) på ett

kroppseget protein (antigen) dit antikroppar kan binda, och subklasser av immunoglobulin G

(IgG)-antikroppar, hos barn med typ 1 diabetes som genomgått en immunmodulerande

behandling.

Resultaten visade att förekomst av autoantikroppar mot glutaminsyredekarboxylas (GADA)

och mot det tyrosinfosfatas-liknande proteinet IA-2 (IA-2A) var associerat med HLA-

riskgener hos 5-åriga barn från den allmäna befolkningen. HLA-riskgener verkade viktiga för

bestående autoantikroppar samt för utveckling av typ 1 diabetes. Uppkomsten av

autoantikroppar verkade däremot påverkas mer av miljöfaktorer än av HLA-riskgener, vilket

gällde särskilt för de så kallade övergående (transienta) autoantikropparna, dvs de som senare

försvann.

Det behövs bättre metoder för att mäta autoantikroppar, både för att kunna förutspå typ 1

diabetes och för att hitta personer med ökad risk för sjukdomen som skulle kunna delta i

förebyggande behandlingar. Därför utvecklade vi en metod för mätning av autoantikroppar

mot insulin (IAA) som bygger på en metod som kallas ytplasmonresonans (SPR). Fördelarna

med vår nyutvecklade metod är att provet inte behöver märkas in med något radioaktivt ämne

före analys samt att det går åt mindre tid att köra ett enskilt prov jämfört med den traditionellt

använda metoden immunoprecipitation (RIA), men metoden behöver utvecklas ytterligare.

Barn som nyligen fått typ 1 diabetes och som har behandlats med GAD bundet till aluminium

hydroxid som vaccin-adjuvans (GAD-alum; Diamyd), har visats kunna behålla sin egen

insulinproduktion bättre än de som inte fått denna behandling. Förutom denna kliniska effekt

observerades också att GADA-nivåerna ökade. Vi undersökte vilka epitoper GADA band till

både hos de barn som fått behandling och de utan behandling. GADA-bindning till en av de

testade epitoperna visade sig öka efter GAD-alumbehandling, vilket skulle kunna tyda på att

antikropparnas kvalitet till viss del har ändrats av behandlingen. Däremot var det ingen

förändring i GADA-bindning till de epitoper som är kopplade till sjukdomen stiff person

syndrome (SPS), vilket tillsammans med oförändrad enzymatisk aktivitet hos GAD65

ytterligare bekräftar behandlingens säkerhet. Fördelningen av IgG-subklasser ändrades efter

behandling med GAD-alum, med en lägre andel IgG1 och högre IgG3 och IgG4. IgG1 och

troligtvis även IgG3 är kopplade till så kallade T-hjälpar 1 (Th1)-cellers immunsvar medan

IgG4 är kopplat till Th2-svar. Den lägre andelen IgG1 och högre IgG4 som observerades i vår

studie skulle kunna tyda på en tillfällig förändring av immunsvaret till ett skyddande Th2-

svar, vilket också har setts hos dessa individer för andra immunologiska markörer.

Sammanfattningsvis, att mäta autoantikroppar som är relaterade till typ 1 diabetes är viktigt

för att studera den autoimmuna processen hos barn som håller på att få diabetes och hos de

som redan har diabetes. Utöver att använda autoantikropparna som markörer för utveckling av

typ 1 diabetes kan de även användas för att studera immunmodulerande behandlingars

effekter på det humorala immunsvaret.

Contents

Abbreviations ................................................................................................................... 5

Original publications ......................................................................................................... 7

Review of the literature .................................................................................................... 9

INTRODUCTION TO THE IMMUNE SYSTEM ................................................................................ 9

T cells ........................................................................................................................................................... 10

B cells ........................................................................................................................................................... 11

Antibodies.................................................................................................................................................... 12

INTRODUCTION TO DIABETES MELLITUS ................................................................................. 16

Definition and diagnosis .............................................................................................................................. 16

Classification ................................................................................................................................................ 17

Type 1 diabetes ....................................................................................................................................... 17

Type 2 diabetes ....................................................................................................................................... 17

Gestational diabetes ............................................................................................................................... 18

Other specific types of diabetes ............................................................................................................. 18

Epidemiology of type 1 diabetes ................................................................................................................. 18

PATHOGENESIS OF TYPE 1 DIABETES ....................................................................................... 19

Genetic risk .................................................................................................................................................. 20

HLA .......................................................................................................................................................... 21

Other genetic factors .............................................................................................................................. 22

Environmental factors ................................................................................................................................. 23

Viral infections ........................................................................................................................................ 24

Dietary factors ........................................................................................................................................ 24

Beta-cell stress ........................................................................................................................................ 25

Beta-cell autoantigens and autoantibodies ................................................................................................. 26

Islet cell antibodies (ICA) ........................................................................................................................ 26

Glutamic acid decarboxylase (GAD) ........................................................................................................ 27

GAD autoantibodies (GADA) ................................................................................................................... 28

The tyrosine phosphatase like protein islet antigen-2 (IA-2) .................................................................. 29

IA-2 autoantibodies (IA-2A) .................................................................................................................... 30

Insulin ..................................................................................................................................................... 31

Insulin autoantibodies (IAA) ................................................................................................................... 32

Zink transporter 8 (ZnT8) and ZnT8 autoantibodies (ZnT8A) ................................................................. 33

PREDICTION AND PREVENTION OF TYPE 1 DIABETES ............................................................... 34

Prediction and prevention studies .............................................................................................................. 35

DETECTION OF AUTOANTIBODIES ........................................................................................... 40

Aims of the thesis ........................................................................................................... 43

Subjects and methods ..................................................................................................... 45

STUDY POPULATIONS ............................................................................................................. 45

The ABIS study ........................................................................................................................................ 45

The phase II clinical trial with GAD-alum ................................................................................................ 46

Type 1 diabetic patients .......................................................................................................................... 47

METHODS .............................................................................................................................. 47

RIA for detection of GADA and IA-2A (Paper I and IV) ................................................................................ 47

HLA genotyping (Paper I) ............................................................................................................................. 49

RIA for detection of IAA (Paper II) ............................................................................................................... 50

Surface plasmon resonance (SPR) (Paper II) ................................................................................................ 51

Competitive epitope-specific RIA for determination of GADA epitopes (Paper III) .................................... 53

A modified RIA for detection of IgG subclasses of GADA (Paper IV) ........................................................... 54

GAD65 enzymatic activity assay (Paper IV) ................................................................................................. 55

Analysis of tetanus toxoid antibodies (Paper IV) ......................................................................................... 56

Analysis of IgE (Paper IV) ............................................................................................................................. 56

C-peptide measurement (Paper IV) ............................................................................................................. 56

STATISTICS ............................................................................................................................. 57

ETHICAL CONSIDERATIONS ..................................................................................................... 58

Results and discussion ..................................................................................................... 61

AUTOANTIBODY POSITIVITY AND RELATION TO HLA GENOTYPES (PAPER I) ............................. 61

GADA and IA-2A in children at three different ages .................................................................................... 62

Permanent and transient autoantibodies ................................................................................................... 63

Children with type 1 diabetes...................................................................................................................... 65

DEVELOPMENT OF AN IMMUNOASSAY BASED ON SPR (PAPER II) ............................................ 66

Attempts to develop an SPR-based immunoassay for GADA ...................................................................... 67

Development of the SPR-based immunoassay for IAA ................................................................................ 68

Efforts to reduce non-specific binding of serum proteins to the sensor surface ................................... 69

Detection of IAA in serum using the 3-day and 5-min methods ............................................................. 70

High-affinity IAA ...................................................................................................................................... 75

The 5-min method compared to RIA ...................................................................................................... 75

Type 1 diabetes-related analytes measured with SPR: Results from other studies ............................... 77

Further and future experiments with the 5-min method ....................................................................... 78

EFFECTS OF THE GAD-ALUM TREATMENT ON THE HUMORAL IMMUNE RESPONSE (PAPER III AND

IV) .......................................................................................................................................... 78

Increased levels of GADA ............................................................................................................................. 78

Temporary increase in binding to one rFab-defined epitope ...................................................................... 79

Changes in IgG subclasses of GADA ............................................................................................................. 84

No changes in SPS-related epitope binding or inhibition of GAD65 enzymatic activity .............................. 88

Antigen-specific effect ................................................................................................................................. 90

The role of GADA in the effect of GAD-alum ............................................................................................... 91

Summary and conclusions ............................................................................................... 93

Acknowledgements ........................................................................................................ 95

References ...................................................................................................................... 97

5

Abbreviations

ABIS all babies in southeast of Sweden

alum aluminum hydroxide

CM carboxymethylated

C-peptide connecting peptide

cpm counts per minute

CTLA-4 cytotoxic T lymphocyte antigen-4

DAISY the diabetes autoimmunity study in the young

DASP diabetes autoantibody standardization program

DIPP the diabetes prediction and prevention study

DPT-1 the diabetes prevention trial

ELISA enzyme-linked immunosorbent assay

ENDIT the European nicotinamide diabetes intervention trial

GABA gamma-aminobutyric acid

GAD glutamic acid decarboxylase

GAD65 65 kDa isoform of GAD

GADA autoantibodies to GAD65

HLA human leukocyte antigen

IA insulin antibodies

IA-2 tyrosine phosphatase like protein islet antigen-2

IA-2A autoantibodies to IA-2

IAA autoantibodies to insulin

ICA islet cell antibodies

IDS immunology of diabetes society

IFN interferon

Ig immunoglobulin

IL interleukin

INS insulin gene

LADA latent autoimmune diabetes in adults

LPS lipopolysacharide

mAb monoclonal antibody

MHC major histocompatibility complex

Abbreviations ’

6

MODY maturity-onset diabetes of the young

mRNA messenger RNA

NOD non-obese diabetic

OGTT oral glucose tolerance test

PCR polymerase chain reaction

PEG poly(ethylene glycol)

PLP pyridoxal 5-phosphate

PPV positive predictive value

PTP protein tyrosine phosphatase

PTPN22 protein tyrosine phosphatase N22

rFab recombinant Fab

RIA radioimmunoassay

RU resonance units

SAM self-assembled monolayer

SD standard deviation

SPR surface plasmon resonance

SPS stiff person syndrome

SPSS statistical package for the social sciences

TCR T cell receptor

TEDDY the environmental determinants in diabetes of the young

Th T-helper

Treg naturally occurring regulatory T cells

TRIGR the trial to reduce type 1 diabetes in the genetically at risk

U/ml units/ml

VNTR variable number of tandem repeats

WHO the world health organization

ZnT8 zink transporter 8

ZnT8A autoantibodies to ZnT8

7

Original publications

This thesis is based on the following four papers, which are referred to in the text by their

roman numerals:

I. Camilla Gullstrand, Jeanette Wahlberg, Jorma Ilonen, Outi Vaarala and Johnny

Ludvigsson

Progression to type 1 diabetes and autoantibody positivity in relation to HLA-risk

genotypes in children participating in the ABIS study.

Pediatric Diabetes 2008; 9(Part I): 182-190

II. Jenny Carlsson, Camilla Gullstrand, Gunilla T. Westermark, Johnny Ludvigsson,

Karin Enander and Bo Liedberg

An indirect competitive immunoassay for insulin autoantibodies based on surface

plasmon resonance.

Biosensors and Bioelectronics 2008; 24: 882-887

III. Camilla Skoglund, Mikael Chéramy, Rosaura Casas, Johnny Ludvigsson and

Christiane S. Hampe

GAD-alum treatment-induced increase in GAD autoantibody (GADA) titers of type 1

diabetes children and adolescents is accompanied by a transient change in GADA

epitope pattern.

Submitted

IV. Mikael Chéramy, Camilla Skoglund, Ingela Johansson, Johnny Ludvigsson,

Christiane S. Hampe and Rosaura Casas

GAD-alum treatment in patients with type 1 diabetes and the subsequent effect on

GADA IgG subclass distribution, GAD65 enzyme activity and humoral response.

Clinical Immunology 2010; 137(1): 31-40

8

9

Review of the literature

INTRODUCTION TO THE IMMUNE SYSTEM

The human body is continuously surrounded by microorganisms, where some of them are

good and others harmful (Janeway et al. 2005; Mölne et al. 2007). The disease-causing

microorganisms have been divided into four categories: viruses, bacteria, pathogenic fungi

and parasites. The body has to defend itself against these pathogens. A well functioning

immune system has been developed, consisting of a wide range of different white blood cells.

These cells perform different tasks in the immune system. Communication between immune

cells occurs via receptors on the surfaces of the cells and via different signaling molecules,

such as cytokines and chemokines.

The immune system consists of three lines of defense (Janeway et al. 2005; Mölne et al.

2007). The first line of defense comprises mechanical and chemical barriers protecting the

skin and mucosa, together with a microbiological barrier of non pathogenic bacteria.

Epithelial cells are joined by tight junctions, mucus is transported from the lungs by cilia,

fatty acids and antibacterial peptides are produced etc. These barriers are very effective and

most of the microorganisms are therefore unable to enter the body, but those who manage to

break through encounter the second line of defense, namely the innate immune system. This

non-specific immune system consists of granulocytes (neutrophils, eosinophils, basophils and

mast cells) and macrophages. Macrophages have a wide variety of cell-surface receptors that

recognize microbial structures on the pathogen, such as lipopolysacharide (LPS), peptide

glucane and virus-RNA. The pathogen is attached to a receptor and then engulfed by

phagocytosis and eliminated by the macrophage. This is followed by activation of the

macrophage with release of cytokines, chemokines and other mediators that initiate

inflammation in the tissue and bring neutrophils and plasma proteins to the site of an

infection, aiming at elimination of the microorganisms.

The third line of defense is the adaptive immune system, which can be activated by almost

any microbial structure (Janeway et al. 2005; Mölne et al. 2007). The adaptive immune

system consists of lymphocytes (T cells and B cells) and provides a specific response. Mature

lymphocytes circulate between blood vessels, lymph vessels and lymph nodes. The antigen is

Review of the literature ’

10

transported to the lymph nodes by antigen presenting cells, primarily macrophages. The

environment of the lymph node is ideal for interaction between antigen and lymphocyte. The

lymphocytes interact with antigen presenting cells through their antigen-specific receptors.

Upon contact with the specific antigen, the lymphocytes are activated, mature and divide

creating a clone of active cells, which leave the lymph node and migrate to places in the body

where they can have their effect (MacLennan et al. 1997). The T cells mature into antigen-

specific effector cells, often producing cytokines, and the B cells into antibody-secreting

plasma cells.

T cells

Precursors of the T cells are developed in the bone marrow (Janeway et al. 2005; Mölne et al.

2007). They migrate to the thymus, where they mature into T cells and get their unique

specificity. The T cells are selected based on their binding affinity to major histocompatibility

complex (MHC)-peptide complexes. If T cell receptors (TCRs) on a T cell recognize and bind

antigen presented by MHC molecules the cell will survive positive selection, while the other

T cells will die. However, if the T cell reacts strongly with self antigens, it will instead die by

negative selection, thereby maintaining tolerance to self antigens.

T cells can be divided into CD4-positive T cells, which react with the antigen bound to MHC-

II molecules, and CD8-positive T cells, which recognize the antigen on MHC-I molecules

(Janeway et al. 2005; Mölne et al. 2007). MHC-I molecules are present on all kinds of cells in

the body, except red blood cells, while MHC-II are mainly expressed on antigen presenting

cells. Human MHC molecules are referred to as human leukocyte antigen (HLA). When CD4-

positive T cells are activated they become either helper T cells or regulatory T cells (Janeway

et al. 2005; Mölne et al. 2007). Helper T cells can be divided into T-helper 1 (Th1), T-helper

2 (Th2) and T-helper 17 (Th17) cells, depending on their function in the immune system

(Mosmann et al. 1996; Rautajoki et al. 2008; Korn et al. 2009).

Activation of helper T cells is central for all immune reactions, by producing cytokines and by

cell-to-cell interactions. Th1 cells are important in cell-mediated immunity against

intracellular pathogens, by activating and attracting macrophages and cytotoxic cells to the

site of infection (Szabo et al. 2003; Rautajoki et al. 2008). In addition, Th1 cells stimulate the

production of immunoglobulin (Ig)-G antibodies that are involved in opsonization and


11

phagocytosis. Further, cytokines produced by Th1 cells, such as interferon-gamma (IFN-γ),

activate macrophages to increase their production of bactericidal substances, such as enzymes

and free radicals, to eliminate those microorganisms that survived inside the macrophage

(Mölne et al. 2007). Th2 cells are instead important in the humoral immune response (Mowen

et al. 2004). Cytokines produced by Th2 cells, such as interleukin-4 (IL-4), IL-5, IL-9 and IL-

13, assist B cells in their maturation and antibody production for elimination of extracellular

pathogens. Further, a predominant Th2 response is associated with atopic diseases and

allergies. The recently described Th17 cells produce cytokines including IL-17, IL-21 and IL-

22, and seem to have a proinflammatory role (Korn et al. 2009). In addition, IL-17 has been

shown to be important for host defense against some pathogens. Further, both Th17 and Th1

cells are associated with autoimmune diseases, including type 1 diabetes (Szabo et al. 2003).

Regulatory T cells control and down regulate various immune responses. Naturally occurring

regulatory T cells (Treg) constitutely express CD25 (a subunit of the IL-2 receptor) on their

surface and transcribe the FOXP3 gene (Sakaguchi 2004). Expression of the transcription

factor FOXP3 is vital to the development and function of Treg (Fontenot et al. 2003; Hori et

al. 2003). Defects in the function of Tregs have been hypothesized to be involved in the

pathogenesis of numerous autoimmune diseases, including type 1 diabetes (Sakaguchi 2004).

Activated CD8-positive T cells mature into cytotoxic T cells, which can eliminate infected

cells (Janeway et al. 2005; Mölne et al. 2007). Maturation of cytotoxic T cells is both

dependent on cytokines, such as IL-2 and IFN-γ, and on cell-to-cell contact with the helper T

cell.

B cells

B cells are developed in the bone marrow, where they also get their unique specificity

(Janeway et al. 2005; Mölne et al. 2007). The B cell receptor for recognition of an antigen

consists of a membrane bound antibody. When an antigen has bound to the B cell receptor,

the antigen is internalized and processed into peptides (Figure 1). The B cell presents the

peptides on MHC-II molecules and expresses cytokine receptors and the co-stimulatory

molecule CD40 on its surface. This enables a helper T cell that recognizes the antigen, to

attach to the B cell and then produce cytokines. The cell-to-cell contact between the B and the

T cell, together with cytokines binding to the receptors, activates the B cell to mature into a


12

plasma cell that produces antibodies. These antibodies have the same specificity as the

membrane bound antibody used as a receptor.

Figure 1. Schematic illustration of the process for the maturation of a B cell into an antibody

producing plasma cell. The antigen is bound to the B cell receptor followed by internalization and

processing into peptides, which are then presented on MHC-II molecules on the surface of the B cell.

The interaction between an antigen-specific T cell and the B cell, together with cytokines produced by

the T cell, leads to activation of the B cell and maturation into a plasma cell that produces antibodies.

TCR = T cell receptor, CD40L = CD40 ligand, MHC-II = major histocompatibility complex-II.

Antibodies

Antibodies are gamma globulin proteins that are found in blood or other body fluids, and are

used by the immune system to identify and neutralize foreign objects, such as bacteria and

viruses (Cohen 1963; Janeway et al. 2005). The immune system creates billions of different

antibodies with a limited number of genes by rearranging DNA segments during B cell

development, prior to antigen exposure (Janeway et al. 2005). Mutation can also increase

genetic variation in antibodies. Antibodies can react with almost any chemical structure in

nature, including our own proteins. Antibodies that recognize self antigens are called

autoantibodies (Milgrom et al. 1963).

Antibodies consist of two heavy chains and two light chains of amino acids, in a Y shaped

form, with a molecular mass of approximately 150 kDa (Figure 2) (Janeway et al. 2005;

Schroeder et al. 2010). The two ends of the antibody, linked via flexible hinge regions, have

different functions. The Fc region determines the mechanisms used to destroy antigen, such as

activation of complement and binding to Fc receptors, and the two Fab regions bind to a

B cellB cell B cellB cell B cellT cell

CD4

MHC-II

peptideTCR

CD40 CD40Lcytokinereceptors

cytokines

B cellPlasma cell

B cell receptor

antigen

antibodies


13

specific epitope on the antigen. Differences in antigenic and structural properties of the heavy

chains determine the class and subclass of the molecules. There are five antibody isotypes

known as immunoglobulin A (IgA), IgD, IgE, IgG and IgM, which have different roles in the

immune system.

Figure 2. A schematic presentation of an antibody. It is composed of two heavy chains and two light

chains. The Fab regions, containing the antigen-binding sites, are linked by hinge regions to the Fc

part of the antibody. Adapted from (Alberts et al. 2002).

IgM are relatively low-affinity antibodies that are associated with a primary immune response

and are found in blood and sometimes in extracellular fluids (Schroeder et al. 2010). IgM

consists of five antibody units in a pentamer structure, which makes it very efficient in

opsonizing antigens for destruction and fixing complement.

IgD antibodies are found at very low levels in serum and the function of IgD is unclear

(Schroeder et al. 2010).

IgG antibodies, which usually are of higher affinity, are produced later in the immune

response and this isotype is predominant in blood and extracellular fluids (Schroeder et al.

2010). IgG can be divided into four subclasses; IgG1, IgG2, IgG3 and IgG4, which are

numbered according to the rank order (IgG1>IgG2>IgG3>IgG4) of the serum levels of these

antibodies in the blood of healthy individuals (Figure 3). The major structural differences

between the IgG subclasses occur in the hinge region. IgG1 has a freely flexible hinge region

consisting of 15 amino acids (Brekke et al. 1995). IgG3 has an elongated hinge region of 62

amino acids, giving this subclass the greatest flexibility. The hinge region of IgG4 is shorter

FabFab

Fc

antigen-binding site

antigen-binding site

hingeregions

light chain

heavy chain heavy chain

light chain

HOOC COOH

HOOC COOH

H2N

H2N NH2

NH2


14

than IgG1, and IgG2 has the shortest, making this molecule rigid (Roux et al. 1997). The

flexibility of the molecule may be important for the function of the antibody (Brekke et al.

1995). IgG can activate the complement system, neutralize toxins, viruses and bacteria, and

opsonize them for phagocytosis (Schroeder et al. 2010). They bind Fc-receptors on

neutrophils, monocytes and macrophages, which facilitates opsonization of particles that are

covered with IgG antibodies. The Fc regions of IgG, as well as IgM, can bind complement

factor C1 and thereby activate complement via the classical way, leading to that also

encapsuled bacteria can be opsonized by the complement system. Not all IgG subclasses can

bind complement; IgG1 and IgG3 are effective complement activators, IgG2 is a weak

complement activator and IgG4 is unable to activate complement.

IgG1

IgG2

IgG3

IgG4

Property

Proportion of total 67

22

7

4

IgG in serum (%)

Serum half life (days) 21

21

7

21

Complement activation +++

+

+++

-

Figure 3. Schematic structure and properties of IgG subclasses. IgG1 is predominant in serum, IgG3 is

the subclass with the longest hinge region and the shortest half life and IgG4 is unable to fix

complement. Adapted from (Hamilton 1987; Lin et al. 2010; Schroeder et al. 2010).

IgA antibodies are found as monomers in blood and extracellular fluids and as dimeric

molecules at mucosal surfaces and in secretions, such as saliva and breast milk (Schroeder et

al. 2010). IgA has two subclasses, which differ mainly in their hinge regions; IgA1 and IgA2,

where IgA2 is more resistant to proteolysis. IgA provides a first line of defense against a wide

variety of pathogens by protecting mucosal surfaces from toxins, viruses and bacteria, via

direct neutralization or prevention of binding to the mucosal surface.


15

IgE is a very potent immunoglobulin, although it is present at the lowest serum concentration

of all isotypes (Schroeder et al. 2010). It binds with high affinity to FcεRI receptors on mast

cells, basophils, Langerhans cells and eosinophils. IgE antibodies are associated with

hypersensitivity and allergic reactions and with responses to parasitic worm infections.

During an antibody response the isotype and subclass of antibodies can be shifted without

changing the specificity (Janeway et al. 2005). The dominance of IgM in the beginning of an

antibody response is after re-exposure of the antigen switched into other classes of antibodies,

for example IgG in serum. Cytokines regulate the generation of different Ig isotypes and IgG

subclasses (Purkerson et al. 1992; Snapper et al. 1993). The distribution of various isotype-

specific antibodies may therefore reflect whether the immune response is Th1- or Th2-biased.

The Th2 cytokine IL-4 induces the synthesis of IgG4 and IgE (Lundgren et al. 1989; Gascan

et al. 1991), and the Th1 cytokine IFN-γ stimulates IgG1 and seems to stimulate IgG3

production (Widhe et al. 1998), but further studies are needed to clearly define the antibody

subclass association with Th1 or Th2 response in humans.

Antibodies can be monoclonal or polyclonal (Janeway et al. 2005). Polyclonal antibodies are

produced by different cells and are therefore immunochemically different from each other,

while monoclonal antibodies are the product of an individual clone of plasma cells and thus

immunochemically identical. Antibodies that bind to the same antigen but to different

epitopes, or that bind to similar antigens, have different binding capacity, so called specificity.

Affinity defines the strength of binding of the antibody to its antigen in terms of a single

antigen-binding site binding to a monovalent antigen. The total binding strength of a molecule

with more than one binding site is called its avidity. Anti-idiotypic antibodies recognise the

antigen-binding site of a specific antibody and can thereby interfere with the binding to the

corresponding antigen (Geha 1985). It has been suggested that there is a lack of anti-idiotypic

antibodies in type 1 diabetes (Oak et al. 2008).

In conclusion, T cells and B cells are important in the adaptive immune system. Activated T

cells become helper T cells or regulatory T cells, while activated B cells mature into plasma

cells that produce antibodies. The five antibody isotypes IgM, IgD, IgG, IgA and IgE have

different roles in the immune system.

http://www.ncbi.nlm.nih.gov/books/bv.fcgi?rid=imm.glossary.2528#2612


16

INTRODUCTION TO DIABETES MELLITUS

Definition and diagnosis

Diabetes mellitus is a group of metabolic disorders characterized by chronic hyperglycemia

with disturbances of carbohydrate, fat and protein metabolism resulting from defects in

insulin secretion, insulin action, or both (WHO 1999; American Diabetes Association 2008).

Several pathogenic processes are involved in the development of diabetes. These include

destruction of the beta cells of the pancreas with consequent insulin deficiency and

abnormalities that result in resistance to insulin action.

Symptoms indicating diabetes mellitus are signs of hyperglycemia, such as thirst, polyuria,

glucosuria, weight loss and others such as fatigue, blurred vision and recurrent infections

(WHO 1999). In severe forms, ketoacidosis or a non-ketotic hyperosmolar state may develop

and lead to drowsiness and coma, and in absence of effective treatment to death. Effects of

diabetes mellitus are long-term damage, dysfunction and failure of various organs, especially

the eyes, nerves, kidney, heart and blood vessels (Glastras et al. 2005; Melendez-Ramirez et

al. 2010; Skyler 2010).

The criteria for the diagnosis of diabetes are symptoms of hyperglycemia and a plasma

glucose value of ≥11.1 mmol/l or fasting plasma glucose of ≥7.0 mmol/l (in whole blood 6.1

mmol/l) or 2-h plasma glucose ≥11.1 mmol/l during an oral glucose tolerance test (OGTT)

(American Diabetes Association 2008). In the absence of symptoms, the diagnosis must be

confirmed another day by one of the three methods above. Diabetes may present differently in

different individuals, ranging from severe symptoms and gross hyperglycemia to a lack of

symptoms and with blood glucose values just above the diagnostic cut-off value. In children,

diabetes often presents with severe symptoms, very high blood glucose levels, marked

glycosuria and ketonuria (WHO 1999). Diagnosis is usually confirmed with blood glucose

measurements and treatment is initiated immediately. Rarely, children and adolescents lack

symptoms and are then diagnosed with a fasting blood glucose measurement and/or an

OGTT. This happens, for example, when patients are found by screening of autoantibodies to

identify individuals with increased risk of developing type 1 diabetes.


17

Classification

Diabetes mellitus can be classified into four types based on etiology; type 1 diabetes, type 2

diabetes, gestational diabetes and other types of diabetes. The majority of cases of diabetes

fall into type 1 diabetes or type 2 diabetes (American Diabetes Association 2008).

Type 1 diabetes

Type 1 diabetes is characterized by an absolute deficiency of insulin secretion (Notkins et al.

2001; American Diabetes Association 2008). The most common form of type 1 diabetes is

immune-mediated diabetes and accounts for 5-10% of all individuals with diabetes. This

disease is considered autoimmune (Rose et al. 1993; Bach 1994) and results from a cellular-

mediated autoimmune destruction of the beta cells of the pancreas (Knip 1997; Knip et al.

2008). Autoimmune destruction of beta cells has multiple genetic predispositions, mainly

strong HLA associations, and is influenced by environmental factors (Knip 1997; Akerblom

et al. 2002; Ilonen et al. 2002; Achenbach et al. 2005). Autoantibodies against beta-cell

proteins, such as glutamic acid decarboxylase (GAD) (Baekkeskov et al. 1990), the tyrosine

phosphatase like protein islet antigen-2 (IA-2) (Lan et al. 1996; Notkins et al. 1996), insulin

(Palmer et al. 1983) and zink transporter 8 (ZnT8) (Wenzlau et al. 2007), are produced during

the autoimmune destruction of the beta cells. One or more of these autoantibodies are present

in 90-95% of individuals when hyperglycemia is initially detected (Notkins et al. 2001). Type

1 diabetes occurs predominantly in children and adolescents, usually as a rapidly progressive

form, but may also occur in adults, often as a slowly progressive form, referred to as latent

autoimmune diabetes in adults (LADA). A minority of patients with type 1 diabetes fall into

the category named idiopathic diabetes (American Diabetes Association 2008). This form of

diabetes is most common in non-Caucasians and is strongly inherited, lacks immunological

evidence for beta-cell autoimmunity, and is not HLA associated.

Type 2 diabetes

In type 2 diabetes, the cause is a combination of resistance to insulin action and defects in

insulin secretion (American Diabetes Association 2008). In adults, type 2 diabetes is much

more prevalent than type 1 diabetes, but it is rare in children and adolescents in Sweden.

Individuals with type 2 diabetes do initially not need insulin treatment to survive, and often

not later in life either. Most patients are obese, which causes some degree of insulin

resistance. Type 2 diabetes is often associated with a strong genetic predisposition and the risk


18

of developing this disease increases with age, obesity and lack of physical activity.

Individuals with this form of diabetes are often undiagnosed for many years because their

hyperglycemia is not severe enough to present symptoms, but they have increased risk to

develop macrovascular and microvascular complications. Insulin sensitivity may be increased,

but not restored to normal, by weight reduction, increased physical activity and/or

pharmacological treatment of hyperglycemia.

Gestational diabetes

Any degree of glucose intolerance with onset or first recognition during pregnancy is referred

to as gestational diabetes (American Diabetes Association 2008).

Other specific types of diabetes

Some forms of diabetes may be associated with genetic defects of the beta cell, which are

often referred to as maturity-onset diabetes of the young (MODY) (Stride et al. 2002). These

individuals have impaired insulin secretion with minimal or no defects in insulin action. Other

types of diabetes are associated with genetic defects in insulin action or diseases of the

exocrine pancreas, endocrinopathies, drug- or chemical-induced diabetes, infections, other

genetic syndromes sometimes associated with diabetes, and uncommon forms of immune-

mediated diabetes (American Diabetes Association 2008). The last category includes patients

with stiff person syndrome (SPS) who also have developed diabetes. SPS is a rare neurologic

disorder characterised by muscle rigidity and episodic spasms involving axial and limb

musculature, and the patient has usually very high levels of GADA (Levy et al. 1999). About

one third of patients with SPS will develop diabetes.

Epidemiology of type 1 diabetes

The International Diabetes Federation has estimated the number of children globally aged 0-

14 years with type 1 diabetes to be 480 000 in 2010, with 76 000 newly diagnosed cases a

year (IDF 2009). One quarter of the cases come from South East Asia and more than one fifth

from Europe. The incidence of childhood onset type 1 diabetes is increasing worldwide, with

an overall annual increase of about 3% (Onkamo et al. 1999; The DIAMOND Project Group

2006; IDF 2009; Patterson et al. 2009). The DIAMOND project group has examined

incidence and trends of type 1 diabetes worldwide for the period of 1990-1999, giving an

annual increase in incidence of 2.8% (The DIAMOND Project Group 2006), and the


19

EURODIAB study group has done the same for Europe during 1989-2003, resulting in an

annual increase of 3.9% (Patterson et al. 2009).

The increase in incidence has been observed in countries with both high and low prevalence,

with a more steeply increase in some of the low prevalence countries, such as those in central

and eastern Europe (IDF 2009). The highest increase was found in children younger than 5

years of age, particularly in European populations (Karvonen et al. 1999; Green et al. 2001;

2006; Patterson et al. 2009). Finland has the highest incidence in the world; in 2005 the

annual incidence was 64.2/100 000 per year in children before 15 years of age (Harjutsalo et

al. 2008). Sweden has the second highest incidence, with an incidence in 2009 of 44.0/100

000 per year in children 0-14.9 years of age (SWEDIABKIDS 2009).

In general, the incidence of type 1 diabetes increases with age, peaking at puberty (Soltesz et

al. 2007). The overall sex ratio for type 1 diabetes is roughly equal in children, with a minor

male excess in incidence in Europe. A seasonality of onset has been reported, with a peak

occurring in winter, and it is more pronounced in countries with marked differences between

summer and winter temperatures (Dahlquist et al. 1994; Soltesz et al. 2007).

In conclusion, type 1 diabetes is a serious disease resulting from destruction of the beta cells

in the pancreas. The incidence of type 1 diabetes is increasing worldwide and Sweden has the

second highest incidence in the world.

PATHOGENESIS OF TYPE 1 DIABETES

The etiology of type 1 diabetes is largely unknown, but a combination of genetic

predisposition, environmental factors and a dysregulated immune system is believed to play

an important role for development of the autoimmune process leading to the disease (Figure

4). The genetic susceptibility of type 1 diabetes is mainly dependent of HLA class II genes,

but other genes are also involved (Rich et al. 2009). Environmental factors that are suggested

to influence the development of type 1 diabetes include viral infections, early infant diet,

toxins and psychological stress (Peng et al. 2006). The autoimmune process, causing an

inflammation in the beta cells (insulitis), is characterized by infiltration of CD8-positive

cytotoxic T cells, CD4-positive T cells, B cells and macrophages (Imagawa et al. 1999;

Moriwaki et al. 1999). This process, which may be initiated several years before clinical onset


20

of type 1 diabetes, leads to decreased beta-cell mass, reduced insulin production and finally to

clinical type 1 diabetes. During the pre-clinical period, autoantibodies against beta-cell

antigens often circulate in the peripheral blood, and measurement of these autoantibodies can

be used for identification of individuals at risk for the disease.

Figure 4. Schematic illustration of the development of type 1 diabetes. Interaction between genetic

predisposition and environmental triggers, together with a dysregulated immune system, may induce

an autoimmune response with autoantibody production, leading to loss of beta cells and progression to

type 1 diabetes. Modified from (Atkinson et al. 2001).

Genetic risk

Most of the children who develop type 1 diabetes have genetic predisposition for the disease.

The HLA gene complex is responsible for about 50% of the genetic risk for type 1 diabetes

and the remaining genetic susceptibility is conferred by a large number of loci, where most of

them have minor effects (Ilonen et al. 2002; Rich et al. 2009). Other genes known to have

effect on the risk of type 1 diabetes include insulin (INS), cytotoxic T lymphocyte antigen-4

(CTLA-4) and protein tyrosine phosphatase N22 (PTPN22) (Redondo et al. 2002; Onengut-

Gumuscu et al. 2004).

Pre-diabetic phase Overt diabetes

Time

Bet

a-c

ellm

ass

Geneticpredisposition

Environmental triggers

Immune dysregulation

Insulitis

Beta-cell autoimmunity

Glucose intolerance

Loss of C-peptide


21

HLA

The HLA gene complex is located on the short arm of chromosome 6 (6p21) and contains

genes encoding HLA class I (HLA-A, -B and -C) and class II (DR, DQ and DP) molecules,

which are peptide presenting molecules for T cells, consisting of an α- and a β-chain (She

1996; Undlien et al. 2001). These genes are the most polymorphic genes known in humans

and there is a strong linkage disequilibrum between the genes, which means that the alleles at

one HLA locus are non-randomly associated with alleles at other HLA loci. Polymorphisms

in the HLA genes affect the conformation of the molecule and especially the peptide binding

groove, and define which peptide that will be bound and presented to T cells, thus shaping the

T cell repertoire (Ilonen et al. 2002).

Type 1 diabetes has strong HLA associations, with linkage to the DQA and DQB genes, and

is influenced by the DRB genes (She 1996; Ilonen et al. 2002). The strongest determinant of

genetic risk is the presence of risk associated HLA class II haplotypes DR3-DQ2

(DRB1*0301-DQA1*0501-DQB1*0201) and DR4-DQ8 (DRB1*04-DQA1*0301-

DQB1*0302) or especially the combination of them both (Table 1) (Sanjeevi et al. 1995; She

1996; Ilonen et al. 2002; Redondo et al. 2002). About 90% of individuals with type 1 diabetes

have at least one of these two high-risk haplotypes compared to approximately 20% of the

general population (Redondo et al. 2002). The DRB1 allele modifies the risk conferred by the

DQ8 molecule; the DRB1*0401, *0402 and *0405 are associated with high susceptibility, the

DRB1*0404 with moderate susceptibility and DRB1*0403 with protection against type 1

diabetes.

Other class II haplotypes are protective, such as DR2-DQ6 (DRB1*15-DQA1*0102-

DQB1*0602), which is the most significant in the Northern European population (Table 1)

(Baisch et al. 1990; Redondo et al. 2002). This haplotype seems to confer dominant

protection, since it is protective also in the presence of a high-risk haplotype in the same

individual (Baisch et al. 1990; Pugliese et al. 1995; Redondo et al. 2002). About 20% of

Europeans have DR2-DQ6 while less than 1% of children with type 1 diabetes have this allele

(Redondo et al. 2002). Besides the susceptible and protective haplotypes, there are also

neutral HLA class II haplotypes, which do not affect the risk of type 1 diabetes (Table 1).


22

Table 1. Susceptibility, protective and neutral associated haplotypes (according to Jorma Ilonen,

personal communication 2005).

Susceptibility DRB1*0401/2/5-DQB1*0302 (DR4-DQ8) DRB1*0404-DQB1*0302 (DR4-DQ8) DRB1*0301-DQA1*0501-DQB1*0201 (DR3-DQ2)

Protective DRB1*15-DQA1*0102-DQB1*0602 (DR2-DQ6) (DR5)-DQA1*05-DQB1*0301 (DR7)-DQA1*0201-DQB1*0303 (DR14)-DQA1*0101-DQB1*0503 DRB1*0403-DQB1*0302 (DR1301)-DQB1*0603

Neutral (DR4)-DQA1*0301-DQB1*0301 (DR4-DQ7) (DR1)-DQB1*0501 (DR7)-DQA1*0201-DQB1*02 (DR1302)-DQB1*0604 (DR9)-DQA1*03-DQB1*0303 (DR8)-DQB1*04 (DR7)-DQA1*02-DQB1*02 (DR4)-DQA1*03-DQB1*0301 (DR2)-DQB1*0601 (DR16)-DQB1*0502 DQB1*0609

Other genetic factors

INS: The insulin gene (INS) is located on chromosome 11p15 and includes a non-coding

region with a variable number of tandem repeats (VNTR) (Bennett et al. 1995; Redondo et al.

2002; Ziegler et al. 2010). Polymorphisms in this region are associated with risk of diabetes

and influence thymic insulin messenger RNA (mRNA). There are three main types of the

insulin VNTR defined by the number of repeats, class I, class II and class III, where class III

has the highest number. The class I VNTRs are most common in Caucasians and the class II

alleles are rare (Stead et al. 2002). Homozygosity for class I alleles is associated with high

risk for diabetes, while class III alleles confer dominant protection (Redondo et al. 2002).

Class III alleles are associated with higher expression of insulin mRNA within the thymus and

high concentration of thymic insulin might lead to negative selection of high-avidity

autoreactive T cells and thus to the development of tolerance.


23

CTLA-4: The cytotoxic T lymphocyte antigen 4 (CTLA-4) gene is located on chromosome

2q33 and encodes a receptor expressed by activated T cells (Redondo et al. 2002). This

receptor can limit the proliferative response of activated T cells, some of which could be

autoreactive, upon binding to B7 molecules, and can also mediate T cell apoptosis.

Polymorphisms or mutations that alter the activity of CTLA-4 are believed to play a role in

the risk for developing autoimmunity. A polymorphism in the first exon of CTLA-4 (49 A/G)

results in an amino acid change (threonine/alanine) in the leader peptide of the expressed

protein. The presence of an alanine at codon 17 of CTLA-4 has been associated with genetic

susceptibility to type 1 diabetes (Nistico et al. 1996; Donner et al. 1997).

PTPN22: The protein tyrosine phosphatase N22 (PTPN22) gene, located on chromosome

1p13, encodes a lymphoid-specific phosphatase known as Lyp (Onengut-Gumuscu et al.

2004; Ladner et al. 2005). This protein is a negative regulator of T cell activation by

dephosphorylating T cell receptor activation-dependent kinases. The single nucleotide

polymorphism C1858T of the PTPN22 gene have been associated with type 1 diabetes.

Individuals lacking the C allele of PTPN22 may have reduced capacity to downregulate T cell

responses and may therefore be more susceptible to autoimmunity.

Environmental factors

The rapid increase in incidence of type 1 diabetes cannot be explained by changes in genetic

predisposition, but rather by environmental factors. In addition, the relatively low

concordance (with both twins affected) in monozygotic twins, 21-70%, and 6% of siblings to

type 1 diabetic patients that develops the disease (Redondo et al. 1999; Redondo et al. 2002;

Aly et al. 2006), further emphasize a role of environmental factors in the etiology of type 1

diabetes. Several environmental factors, including viral infections, cow´s milk, gluten and

psychological stress, have been suggested to trigger the autoimmune response and the

development of type 1 diabetes, as reviewed in (Akerblom et al. 2002; Peng et al. 2006;

Soltesz et al. 2007). A number of studies are ongoing aiming to further investigate the role of

different environmental factors on the development of type 1 diabetes, see the section

“Prediction and prevention studies”.


24

Viral infections

Viral infections have been suggested to trigger autoimmunity and type 1 diabetes (van der

Werf et al. 2007). Enterovirus infections, including Coxsackie B4, are the most studied, and

have been associated with seroconversion to islet autoantibody positivity and with diabetes

onset (Hyoty et al. 1995; Salminen et al. 2003; Moya-Suri et al. 2005). Congenital rubella

infection has been found to associate with a high rate of subsequent type 1 diabetes, but

effective immunization programs have eliminated this virus in most Western countries

(Menser et al. 1978). Enteroviral infections during pregnancy have been associated with

increased risk for type 1 diabetes in the offspring (Dahlquist et al. 1995; Hyoty et al. 1995).

Interestingly, it has been observed that in populations with high incidence of type 1 diabetes,

the frequency of enterovirus infections has tended to decrease over the last decades and is

lower than in populations with low incidence of type 1 diabetes (Viskari et al. 2000; Viskari

et al. 2004). This is in line with the polio hypothesis, which suggests that the complications of

enterovirus infections become more common in an environment with a decreased rate of

infections leading to a lack of immunity in the population.

The hygiene hypothesis suggests that the increased hygiene, leading to a lack of normal

background infections, predisposes the immune system to autoimmunity, including type 1

diabetes (Ludvigsson 2006). Probably, changes in the gut bacterial flora influence the

maturation of the immune system, leading to imbalance and thereby autoimmune reactions in

genetically predisposed individuals (Vaarala 1999b). Further, the increased hygiene might

lead to a lower immunity against certain viruses (Viskari et al. 2005).

Dietary factors

A number of dietary factors have been suggested to influence the development of beta-cell

autoimmunity and type 1 diabetes, as reviewed in (Virtanen et al. 2003). Factors that seem to

decrease development of autoimmunity include greater intake of breast milk, nicotinamide,

zinc, and vitamins C, D and E. On the other hand, increased islet autoimmunity seems to be

influenced by factors as early introduction of cow´s milk, gluten, nitrate, nitrite, and increased

calories causing increased linear growth and weight. Nitrate and nitrite are mainly found in

food, but may also originate from cigarettes, car interiors and cosmetics.

Breast-feeding has been suggested to have a protective effect; high frequency of breast-

feeding has been associated with low incidence of type 1 diabetes and short duration of


25

breast-feeding has been reported to increase the risk of type 1 diabetes (Borch-Johnsen et al.

1984; Virtanen et al. 1993; Virtanen et al. 2003). In a Finnish study, the effects of duration of

breast-feeding and the age at introduction of supplementary milk were studied (Virtanen et al.

1993). The results indicated that early introduction of cow´s milk increased the risk for type 1

diabetes, and that this factor may overcome the protective effects of breast-feeding. Further,

Wahlberg et al found that early introduction of cow´s milk based formula and high

consumption of cow´s milk were associated with higher levels of beta-cell autoantibodies

(Wahlberg et al. 2006). It has been proposed that an early exposure to cow´s milk formula

may result in an immune response to bovine insulin, and that this could trigger an immune

response to human insulin, which has a very similar amino acid sequence as bovine insulin

(Vaarala et al. 1999a). In addition, feeding with cow´s milk formula during infancy has been

associated with increased weight gain (Johansson et al. 1994), which might induce beta-cell

stress, see section below. Further, introduction of gluten too early (before 3 months) (Ziegler

et al. 2003), or too late (after 6 months) (Wahlberg et al. 2006), seems to be a risk factor for

induction of autoantibodies.

Beta-cell stress

Beta-cell stress has been suggested as a risk factor for the development of type 1 diabetes

(Ludvigsson 2006). During psychological stress and periods of rapid growth, such as infancy

and puberty, the beta cells have to work hard to produce insulin. This increased insulin

production may result in beta-cell stress and stimulation of the autoimmune process, leading

to overt diabetes. This beta-cell stress hypothesis is an extension of the accelerator hypothesis

(Wilkin 2001), which states that increased insulin resistance, associated with childhood

overweight and obesity, creates greater insulin secretory demand on the islets, leading to

acceleration of beta-cell destruction and type 1 diabetes. The accelerator hypothesis suggests

that the increased incidence of type 1 diabetes may be caused by an accelerated progression

rather than by an increase in the absolute lifetime risk.

Psychological stress may, in certain individuals, cause insulin resistance and thereby beta-cell

stress and diabetes-associated autoimmunity. Psychological factors that have been found to

influence the development of autoantibodies, in one or two and a half year old children of the

general population, include high parental stress and serious life events (Sepa et al. 2005; Sepa

et al. 2005).


26

In conclusion, genetic risk factors, mainly HLA, together with environmental factors, such as

viral infections and dietary factors, seem to play an important role for the development of

type 1 diabetes.

Beta-cell autoantigens and autoantibodies

Pancreatic beta-cell autoantigens are the targets of immune-mediated destruction of beta cells

(Yoon et al. 2005). One of the most common immunologic markers of individuals with

autoimmune diabetes is the presence of autoantibodies against beta-cell autoantigens. These

autoantibodies can also be used for prediction of type 1 diabetes, both in high-risk individuals

and in the general population, where positivity to multiple autoantibodies confer the highest

risk (Bingley et al. 1994; Bingley et al. 1997; Verge et al. 1998; Kulmala et al. 2001; LaGasse

et al. 2002). The autoantibodies most commonly studied are directed against glutamic acid

decarboxylase (GADA) (Baekkeskov et al. 1990), the tyrosine phosphatase like protein islet

antigen-2 (IA-2A) (Lan et al. 1996; Notkins et al. 1996) and insulin (IAA) (Palmer et al.

1983). Islet cell antibodies (ICA) (Bottazzo et al. 1974) were previously widely used to study

the clinical course and pathogenesis of type 1 diabetes, but were to a large extent replaced by

GADA and IA-2A, when methods for detection of these autoantibodies were developed.

Recently, autoantibodies against ZnT8 (ZnT8A) were discovered as an additional marker for

type 1 diabetes (Wenzlau et al. 2007).

Islet cell antibodies (ICA)

Islet cell antibodies (ICA), recognizing islet cytoplasmic antigens, were detected many years

ago in newly-diagnosed type 1 diabetic patients (Bottazzo et al. 1974) and comprise

autoantibodies to a number of antigens, with a predominance of GADA and IA-2A (Notkins

et al. 1996; Notkins et al. 2001). The presence of organ-specific pancreatic antibodies

provides evidence for type 1 diabetes as an autoimmune disease (Bottazzo et al. 1974;

MacCuish et al. 1974). ICA is measured using immunofluorescence, by incubation of sera

from type 1 diabetic patients with frozen tissue sections of normal blood group 0 pancreas,

which leads to staining of the pancreatic islets (Notkins et al. 2001). IAA is not recognized in

the ICA test, because insulin and c-peptide leach out from the unfixed frozen tissue sections

during sample preparation.


27

Glutamic acid decarboxylase (GAD)

Glutamic acid decarboxylase (GAD) is a biosynthesizing enzyme of the inhibitory

neurotransmitter gamma-aminobutyric acid (GABA) and has been the most extensively

studied beta-cell autoantigen (Baekkeskov et al. 1990; Yoon et al. 2005). This antigen was

first discovered in serum from children with type 1 diabetes in Linköping, and was described

as an antigen with the weight of 64kDa (Baekkeskov et al. 1982). Later, Baekkeskov et al

discovered that 64 kDa actually was GAD (Baekkeskov et al. 1990). This enzyme is

expressed at high levels by pancreatic beta cells and a subpopulation of central nervous

system neurons. There are two isoforms of GAD, GAD65 and GAD67, with molecular

masses of 65kDa and 67kDa (Bu et al. 1992). In human pancreatic islet cells only the GAD65

isoform is expressed (Hagopian et al. 1993). GAD65 is an intracellular membrane anchored

protein consisting of 585 amino acids (Figure 5) and the GAD65 gene is located on

chromosome 10p11 (Bu et al. 1992; Christgau et al. 1992). The protein is synthesized within

the cytoplasm as a soluble hydrophilic molecule, but becomes membrane anchored after a

two-step modification at the NH2-terminal domain. GAD65 is located to the membrane of

small synaptic-like microvesicles in the islet beta cells.

Autoantibodies in type 1 diabetes primarily recognize the GAD65 isoform (Hagopian et al.

1993). In addition to autoantibodies, GAD-specific CD4-positive T cells and HLA-A*0201-

restricted CD8-positive cytotoxic T cells reactive against GAD have been observed in

recently diagnosed type 1 diabetic patients and in high-risk individuals (Yoon et al. 2005).

These results indicate that GAD may be an important target antigen and GAD-reactive T cells

may play a pathogenic role in the destruction of pancreatic beta cells. Further, there is

molecular mimicry between GAD (amino acids 247-279) and Coxsackie B4 virus (amino

acids 32-47 of the P2-C protein of Coxsackie B virus), which might be a link between

enterovirus infections and development of type 1 diabetes (Atkinson et al. 1994).


28

Figure 5. Schematic representation of the GAD65 protein. The regions of membrane anchoring,

maximium divergence with GAD67, Coxsackie virus similarity and PLP binding are shown in the

figure, as well as the cysteine rich regions (CYS) and the major conformational epitope regions

recognized by GADA. Adapted from (Leslie et al. 1999).

GAD autoantibodies (GADA)

GADA has been found in approximately 50-80% of newly diagnosed type 1 diabetic patients

(Bonifacio et al. 1995; Sabbah et al. 1999; Strebelow et al. 1999; Winter et al. 2002;

Holmberg et al. 2006), and often persists in sera for many years after the diagnosis (Savola et

al. 1998). Besides the presence of GADA in type 1 diabetes, this autoantibody has also been

found in other autoimmune diseases, such as SPS (Levy et al. 1999), where approximately

80% of the patients have GADA (Rakocevic et al. 2004).

Epitopes: Diabetes-related GADA bind mainly to conformation dependent epitopes on GAD

(Figure 5), where the middle region of the protein is the most important (Padoa et al. 2003;

Ronkainen et al. 2004; Schlosser et al. 2005a). However, the GADA response in the

preclinical stage of type 1 diabetes is dynamic, with epitope spreading accompanied by an

increase in the number of epitopes recognized (Schlosser et al. 2005a). Initial reactivity of

GADA appears to target mainly the middle region of GAD (Bonifacio et al. 2000) and

spreads often rapidly to the C-terminal region (Ronkainen et al. 2006). Children with high risk

of developing type 1 diabetes, defined by having more than one autoantibody, often show

GADA reactivity both to epitopes in the middle and C-terminal part of the GAD molecule

(Hoppu et al. 2004a; Schlosser et al. 2005a). In children with both autoantibodies and genetic

risk for the disease this GADA reactivity generally spreads towards epitopes on the N-

terminal part and other epitopes located in the middle (Schlosser et al. 2005a). However, the


29

most important epitope region on GAD at diagnosis and the following years after diagnosis is

the middle region, often in combination with the C-terminal region (Falorni et al. 1996;

Ronkainen et al. 2004). After disease onset the GADA epitope binding pattern seems to be

quite stable (Hampe et al. 2002).

Subclasses: IgG1 is the dominant IgG subclass of GADA during the early antibody response

in prediabetic individuals and in newly diagnosed type 1 diabetic individuals (Bonifacio et al.

1999; Hoppu et al. 2004a; Ronkainen et al. 2006). Antibodies of subclasses IgG2 and IgG3

often appear together with IgG1 or soon after the initial IgG1 response, while IgG4 is the last

subclass to appear (Ronkainen et al. 2006). In addition, a broad initial response with three or

four IgG subclasses and the lack of an emerging IgG4 response during follow-up have been

associated with increased risk for progression to type 1 diabetes.

The tyrosine phosphatase like protein islet antigen-2 (IA-2)

The tyrosine phosphatase like protein islet antigen-2 (IA-2) is a transmembrane protein

consisting of 979 amino acids with a molecular mass of 106 kDa and is a major autoantigen in

type 1 diabetes (Lan et al. 1996). IA-2 belongs to the protein tyrosine phosphatase (PTP)

family and is expressed in pancreatic islets and brain tissues (Notkins et al. 1996). The IA-2

gene is located on chromosome 2q35 and encodes a protein consisting of a signal peptide and

an extracellular, a transmembrane and an intracellular domain (Figure 6) (Lan et al. 1994;

Notkins et al. 2001). The PTP core sequence, a highly conserved region of 11 amino acids

located within the intracellular domain of IA-2, differs from other PTPs primarily in the

substitution of aspartic acid for alanine, which might be the cause of the lack of enzyme

activity for IA-2 (Notkins et al. 1996). IA-2β, which is closely related to IA-2, is another

member of the PTP family and does also act as an autoantigen (Lu et al. 1996). ICA 512

represents a fragment of IA-2, 453 amino acids shorter, and covers the region from amino

acid 389 to amino acid 914 (Notkins et al. 1996). The extracellular domain of ICA 512

appears to reside within secretory granules and the intracellular domain is located in the

cytoplasm (Solimena et al. 1996).


30

Figure 6. Schematic representation of the IA-2 protein. The signal peptide and the extracellular,

transmembrane and intracellular domains are shown in the figure, as well as the cysteine rich regions

(CYS), the PTP core sequence and the region containing epitopes recognized by IA-2A. Adapted from

(Leslie et al. 1999).

IA-2 autoantibodies (IA-2A)

IA-2A is found in about 55-80% of newly diagnosed type 1 diabetic patients (Bonifacio et al.

1995; Sabbah et al. 1999; Strebelow et al. 1999; Winter et al. 2002). The frequency of IA-2A

varies with age and HLA genotype. In young children and in patients with HLA DR4-

DQA1*0301-DQB1*0302 genotype, the frequency and/or level of IA-2A is highest

(Bonifacio et al. 1995; Genovese et al. 1996; Gorus et al. 1997; Savola et al. 1998). The close

association between IA-2A and HLA DR4 indicates that IA-2A may be a more specific

marker of beta-cell destruction than GADA. IA-2A seems to be a strong predictor for

development of type 1 diabetes (Achenbach et al. 2004b). Further, high levels of IA-2A seem

to be correlated with rapid progression of disease (Bingley et al. 1994; Christie et al. 1994;

Kulmala et al. 1998).

Autoantibodies to IA-2β are found in 35-50% of patients with type 1 diabetes. Since more

than 95% of the patients who have autoantibodies to IA-2β also have IA-2A, screening for

IA-2beta autoantibodies for diagnosis is often omitted (Leslie et al. 1999). However, it has

been shown that presence of IA-2β autoantibodies in IA-2A positive individuals can increase

prediction of type 1 diabetes (Achenbach et al. 2004b).

Epitopes: IA-2A bind exclusively to epitopes located in the cytoplasmic domain of the

molecule (amino acids 601-979) (Figure 6) (Zhang et al. 1997; Dromey et al. 2004). Two

linear epitopes within the juxtamembrane domain (amino acids 611-620 and 621-630,

Epitope recognition

Extracellular Transmembrane Intracellular(cytoplasmic)

PTP core

907-917979 COOH750653500

577-60044838625026NH2 1

IA-2

Signal peptide

CYS


31

respectively) and conformational epitopes in the PTP domain toward the C terminus of the

molecule (aa 931-979) and within the central region (aa 795-889) of the PTP domain have

been identified (Dromey et al. 2004). The disulphide bonds within the intracellular domain

seems important for maintaining the antigenic structure of IA-2, since loss of these bonds

result in almost total loss of reactivity with type 1 diabetic autoantibodies (Xie et al. 1997).

IA-2A present early in the disease process are often directed to juxtramembrane domain

epitopes (Dromey et al. 2004), and IA-2A reactivity to these epitopes have been associated

with an increased risk of progression to type 1 diabetes (Hoppu et al. 2004b). By the time of

diabetes onset, IA-2A reactivity has spread to epitopes predominantly in the PTP domain and

reactivity to the juxtamembrane domain epitopes are less frequent (Dromey et al. 2004).

Subclasses: IA-2A of the IgG1 subclass dominates the autoantibody response in prediabetic

individuals and in newly diagnosed type 1 diabetic individuals (Bonifacio et al. 1999;

Achenbach et al. 2004b; Hoppu et al. 2004b). One study showed a correlation between IA-2A

response of IgG4 subclass and protection from diabetes (Seissler et al. 2002), but others do

not (Bonifacio et al. 1999; Achenbach et al. 2004b; Hoppu et al. 2004b). High risk of

progression to type 1 diabetes was in one of these studies associated with IgG2, IgG3 and/or

IgG4 subclasses of IA-2A (Achenbach et al. 2004b).

Insulin

Insulin, one of the major autoantigens in type 1 diabetes, is a hormone that regulates energy

and glucose metabolism in the body. The blood glucose level is lowered by insulin, by

accelerating the transport of glucose into cells. Insulin is secreted by the beta cells, which

constitute about 70% of the pancreatic islet cells. The protein consists of 51 amino acids and

is encoded on chromosome 11p15 (Notkins et al. 2001). The insulin molecule is a

heterodimer consisting of an A and a B chain of 21 and 30 amino acids, respectively, linked

by two disulfide bridges (Figure 7). Insulin is synthesized as pre-proinsulin by the beta cells

of the pancreas, and after cleavage of an NH2-terminal sequence, proinsulin is formed

(Eisenbarth 2008). After folding of proinsulin into its correct secondary and tertiary structure,

the 5.8 kDa insulin protein is formed by proteolytic cleavage and removal of the connecting

peptide (C-peptide), which is a large peptide in the middle of the proinsulin molecule.


32

Figure 7. The structure of proinsulin and insulin. Proinsulin is formed from pre-proinsulin and after

removal of the connecting peptide (C-peptide) the insulin protein is formed, consisting of an A and a B

chain of 21 and 30 amino acids (aa), respectively.

Insulin autoantibodies (IAA)

In 1983, IAA was identified in newly diagnosed untreated diabetic patients (Palmer et al.

1983). This is often the first autoantibody to appear in young children as a sign of beta-cell

autoimmunity and the levels correlate inversely with age (Vardi et al. 1988; Ziegler et al.

1999). IAA is found in about 40-70% of newly diagnosed type 1 diabetic children (Sabbah et

al. 1999; Strebelow et al. 1999; Winter et al. 2002; Williams et al. 2003; Holmberg et al.

2006).

Epitopes: The epitopes on insulin recognized by IAA are not well characterized. However, it

has been observed that IAA specific to human insulin seems to bind epitopes located on the B

chain of insulin, while antibodies cross-reactive to insulin of other species recognize both the

B chain and conformational epitopes involving both the A and B chains (Potter et al. 2000).

Further, IAA that bind epitopes dependent on threonine B30 seem not to be associated with

type 1 diabetes, while those who bind conformational epitopes incorporating the A and B

chains frequently are. Epitope analysis of IAA using a recombinant Fab (rFab) revealed that

IAA in type 1 diabetic patients can bind an epitope located predominantly on the A chain

(Padoa et al. 2005).

Subclasses: IAA of the IgG1 subclass is often predominant in both pre-diabetic and type 1

diabetic individuals, but the IgG3 subclass is also frequently occurring in pre-diabetic

individuals (Bonifacio et al. 1999; Potter et al. 2000; Hoppu et al. 2004c). In one study,

A chain

B chain

C-peptide

Proinsulin Insulin

(110 aa) (51 aa)

Pre-proinsulin A chain

B chain

B1 B30

A1A21


33

genetically susceptible young children who progressed rapidly to clinical type 1 diabetes were

characterized by strong IgG1 and IgG3 responses to insulin, whereas a weak or absent IgG3

response was associated with relative protection from disease (Hoppu et al. 2004c). In another

study, the risk of progression to type 1 diabetes was higher in individuals with IAA of

subclasses IgG2, IgG3 or IgG4 than in those without these IgG subclasses (Achenbach et al.

2004b).

Affinity: In IAA positive children from the general population, antibody affinity can identify

those at high and low risk (Schlosser et al. 2005b). High affinity has been associated with

HLA DRB1*04, young age of IAA appearance and progression to multiple autoantibodies or

type 1 diabetes (Achenbach et al. 2004a). In addition, the A8-A13 region on insulin is

important for binding of high-affinity IAA.

Zink transporter 8 (ZnT8) and ZnT8 autoantibodies (ZnT8A)

ZnT8, a multispanning transmembrane protein belonging to a large cation efflux family, was

recently discovered to be a major autoantigen in type 1 diabetes (Wenzlau et al. 2007).

Approximately half of the ZnT8 molecule comprises six membrane-spanning regions, which

might hinder the folding of the protein in an aqueous environment. Therefore, C-terminal, N-

terminal and N- and C-terminal fusion proteins have been constructed for use in

radioimmunoassays to detect ZnT8A (Wenzlau et al. 2007; Achenbach et al. 2009).

Autoantibodies to the C-terminal part of ZnT8 have shown a strong relation to disease

development. ZnT8A has been found in about 60% of children with new onset type 1

diabetes. In the first report, ZnT8A was found in 26% of type 1 diabetic patients previously

classified as autoantibody negative (Wenzlau et al. 2007). Both the levels and the prevalence

of ZnT8A increase with age, and the antibodies appear frequently in children by 3 years of

age. ZnT8A usually precedes the disease by many years and emerges often after GADA and

IAA.

In conclusion, autoantibodies against beta-cell proteins can be used as markers for the

autoimmune process, where GADA, IA-2A, IAA and ZnT8A are the most commonly used

today.


34

PREDICTION AND PREVENTION OF TYPE 1 DIABETES

The risk of developing type 1 diabetes is increased in first degree relatives of type 1 diabetic

patients compared to the general population. The risk of type 1 diabetes in the general

population is approximately 0.5%, or even higher in high incidence countries, while the risk

for siblings is about 6%, the risk for offspring of a diabetic mother is 1-4% and of a diabetic

father 6-9% (Hawa et al. 2002; Redondo et al. 2002). For prediction of type 1 diabetes, a

combination of genetic and immunological markers is useful.

Presence of HLA susceptibility genes in an individual increases the risk of developing type 1

diabetes, but only to a small extent, as reviewed in (Ziegler et al. 2010). For example, risk of

type 1 diabetes in children without a family history of this disease can vary from about 0.01%

up to more than 5%, depending on their HLA class II genotypes. Further, some of the HLA

genotypes confer extreme protection, such as DQB1*0602. The risk of type 1 diabetes in a

child carrying this allele and with a family history of type 1 diabetes is reduced to about 1%

of the risk in children with similar family history but without this allele.

The most important increase in type 1 diabetes risk of a child occurs when beta-cell

autoantibodies develop. Autoantibodies to insulin, GAD, IA-2 and ZnT8 are good markers for

prediction of type 1 diabetes in children and young adults (Bingley et al. 1994; Bingley et al.

1997; Kulmala et al. 1998; Verge et al. 1998; LaGasse et al. 2002; Wenzlau et al. 2007).

Presence of only one of these four autoantibodies is associated with a marginally increased

risk, both in individuals with and without a family history of type 1 diabetes. However,

positivity to two or more autoantibodies is highly predictive of type 1 diabetes (Verge et al.

1996; Bingley et al. 1997; Eising et al. 2010; Knip et al. 2010). Further, the risk of type 1

diabetes varies dependent on which of the antibodies that is present, and IA-2A has been

associated with the highest risk (Achenbach et al. 2004b).

In addition to autoantibody positivity, characteristics of the autoantibodies such as greater titer

and affinity, broadness of epitope reactivity and certain distribution of IgG subclasses are also

associated with high risk of type 1 diabetes (Achenbach et al. 2004b; Achenbach et al. 2006;

Achenbach et al. 2007; Mayr et al. 2007). For example, in one study, the risk of progression

to type 1 diabetes was higher in individuals with autoantibodies binding multiple epitopes

than single epitope reactivity, and in individuals with IA-2β autoantibodies than in those


35

without these autoantibodies (Achenbach et al. 2004b). The risk of developing type 1 diabetes

can be stratified from <1% up to >70% by using various combinations of risk markers (Figure

8) (Ziegler et al. 2010).

Figure 8. Stratification of type 1 diabetes risk by islet autoantibody properties together with other risk

markers, such as HLA genotype and age. Characteristics of the antibody response can increase

prediction for disease progression. Ab = antibody. Adapted from (Ziegler et al. 2010).

Prediction and prevention studies

Accurate assessment of risk of type 1 diabetes in non-diabetic individuals is necessary to

identify those individuals suitable for recruitment into intervention trials aiming to prevent the

clinical onset of diabetes. A number of intervention trials have recently been performed and

others are ongoing (Ludvigsson 2010a), e.g. in pre-diabetic individuals; the trial to reduce

type 1 diabetes in the genetically at risk (TRIGR) (Knip et al. 2010), the European

nicotinamide diabetes intervention trial (ENDIT) (Gale et al. 2004), the Diabetes Prevention

Trial (DPT-1) (DPT-1 2002) and The Diabetes Prediction and Prevention study (DIPP)

(Kupila et al. 2001), and in newly diagnosed type 1 diabetic patients; DiaPep277 (Schloot et

<5

10

40

30

20

80

70

60

50

Risk categories

Low Very highHighIntermediate

Single islet AbLow-affinity

Older ageNonsusceptibleHLA genotype

Single islet Abwith high affinity

Proinsulin reactive IAA, middleor C-terminal reactive GAD65

Young ageHLA DR3 or DR4

Two or three islet Abs(IA-2A, ZnT8A)

Young ageLow first-phase insulin(HLA risk genotypes)

Four islet AbsAbs to IA-2β epitopes

Higher titerMultiple IgG subclasses responses

Young age at initiationImpaired glucose tolerance

(HLA risk genotype)


36

al. 2007), anti-CD3 (Herold et al. 2009), anti-CD20 (Pescovitz et al. 2009), IFN-α (Rother et

al. 2009) and GAD-alum (Ludvigsson et al. 2008). Some of these studies are described below.

In several prediction studies, children at increased genetic risk of type 1 diabetes are followed

with frequent sampling and autoantibody analysis. In the All Babies in Southeast Sweden

(ABIS) study (Ludvigsson et al. 2001), children from the general population are followed

without selection for genetic risk, see the section “Study populations” for more information

about this study. Ongoing studies, such as the Diabetes Autoimmunity Study in the Young

(DAISY) (Barker et al. 2004), the Environmental Determinants in Diabetes of the Young

(TEDDY) (The TEDDY Study Group 2008) and ABIS, investigate the role of different

environmental factors on the development of autoimmunity and type 1 diabetes.

DAISY

The Diabetes Autoimmunity Study in the Young (DAISY) is a prospective study in Denver,

Colorado, USA that follows young first-degree relatives (siblings or offspring) of persons

with type 1 diabetes and newborn children with high or moderate risk HLA genotypes. It was

shown that the predictive value of the autoantibody measurements could be increased by

blinded duplicate and independent sample retesting together with the level of the

autoantibody. Approximately one third of the autoantibody positive individuals were false

positive, one third transiently positive and one third persistently positive (Barker et al. 2004).

Further, early or late introduction of cereal seemed to increase the risk of islet cell

autoimmunity in children younger than 3 years of age (Norris et al. 2007), while no

association was found for early introduction of cow´s milk protein (Norris et al. 1996).

TEDDY

The Environmental Determinants in Diabetes of the Young (TEDDY) study is an

observational clinical trial that started 2004 and involves six international centres; Seattle,

Denver and Augusta in North America, and Germany, Finland and Sweden in Europe

(Hagopian et al. 2006; The TEDDY Study Group 2007). The aim is to identify environmental

factors, such as infectious agents, dietary factors and psychosocial factors, which contribute to

or protect from islet autoimmunity and risk of type 1 diabetes. Newborns with HLA risk

genotypes or who are first-degree relatives of type 1 diabetic patients were recruited and will

be followed regularly from birth to the age of 15 years. By the age of 15, it is estimated that


37

800 children will develop islet autoimmunity and 400 will progress to type 1 diabetes (The

TEDDY Study Group 2008).

BABYDIAB

The BABYDIAB study is a German prospective study in which newborn offspring of parents

with type 1 diabetes were followed for development of autoantibodies and diabetes (Hummel

et al. 2004). GADA, IA-2A and IAA were measured at 9 months and at 2, 5 and 8 years of

age. By the age of 5 years, 24 (1.5%) children had developed type 1 diabetes, and all but one

had multiple autoantibodies before the onset of the disease. In addition, the frequency of

autoantibodies at this age was 5.9% and the frequency of multiple autoantibodies was 3.5%.

Children with multiple autoantibodies before the age of 2 years were at the greatest risk of

developing type 1 diabetes.

DIPP

The Diabetes Prediction and Prevention study (DIPP) is a Finnish trial in which genetically

predisposed children from the general population were closely followed for signs of diabetes-

associated autoimmunity (Kupila et al. 2001). Children who developed signs of beta-cell

autoimmunity, with at least two autoantibodies in two consecutive samples, were invited to a

separate prevention trial, with administration of intranasal insulin. The first autoantibody that

emerged was predominantly IAA, and in the majority of the children autoantibodies appeared

in the fall and winter, indicating that infectious agents may play a role in the induction of

beta-cell autoimmunity (Kimpimaki et al. 2001). By the age of 5 years, 13 (1.3%) children

had presented with type 1 diabetes, and all of them had developed at least two autoantibodies

before onset (Kukko et al. 2005). Further, it was found that nasal insulin did not prevent or

delay development of type 1 diabetes (Nanto-Salonen et al. 2008).

TRIGR

In the trial to reduce type 1 diabetes in the genetically at risk (TRIGR) study, which is a

multinational, randomized, double-blind, controlled trial, newborns with HLA-conferred

susceptibility to type 1 diabetes and with at least one family member with type 1 diabetes

were recruited. If exclusive breast-feeding was not possible during the first 6-8 months of life,

the child was randomly assigned either hydrolyzed cow´s milk formula or a standard cow´s

milk based formula. The participants will be monitored up to the age of 10 years for the

appearance of autoantibodies and clinical type 1 diabetes. Final results of the study will be


38

available in 2017 (TRIGR 2010). Autoantibodies against GAD, IA-2, IAA, ZnT8 and islet

cells are analyzed to investigate if the elimination of cow´s milk proteins during infancy can

decrease the development of beta-cell autoimmunity during the first years of life. Results

from a median observation of 10 years have shown that the hazard ratio for positivity for one

or more autoantibodies was reduced in the group that received hydrolyzed formula (Knip et

al. 2010).

ENDIT

The European nicotinamide diabetes intervention trial (ENDIT) was a double blind placebo-

controlled trial that aimed to assess whether high dose nicotinamide during 5 years could

prevent or delay clinical onset of diabetes in ICA positive individuals with a first-degree

family history of type 1 diabetes (Gale et al. 2004). The results showed that there were no

differences in development of diabetes between the nicotinamide and placebo treated

individuals.

DPT-1

The Diabetes Prevention Trial (DPT-1) was started in 1994 with the aim of determining

whether antigen based treatment with insulin would prevent or delay diabetes. Relatives to

type 1 diabetic patients were treated with oral or parenteral insulin. The results showed that

these treatments did not overall slow the progression to type 1 diabetes (DPT-1 2002; Skyler

et al. 2005). However, a beneficial effect of treatment with oral insulin was observed in a

subgroup of the treated individuals with higher IAA levels.

Anti-CD3 monoclonal antibody intervention studies

The humanized Fc mutated anti-CD3 monoclonal antibody OKT3γ (Teplizumab) was tested

in a clinical trial, in which newly diagnosed type 1 diabetic patients were randomly assigned

the antibody or placebo for 14 days (Herold et al. 2002; Herold et al. 2005). Still after two

years, a majority of the treated patients had a maintained or improved C-peptide response, as

well as reduced usage of insulin and improved HbA1c levels. Adverse events were common,

often mild, but some patients had serious adverse events. Further, a study with high-dose

Teplizumab in newly diagnosed type 1 diabetic patients showed to preserve the endogenous

insulin secretion still 5 years after intervention (Herold et al. 2009). However, the recruitment

of patients in that study was stopped after treatment of only 6 patients, due to more adverse

events, probably as a result of the high dose. Very recently, a phase II/III randomized double-


39

blind controlled multicenter study with intravenous administration of Teplizumab for 14 days

in two rounds was performed, in children and adults with recent-onset type 1 diabetes

(http://clinicaltrials.gov 2010a). This trial was a 4-armed dose-ranging study, with one group

of patients receiving placebo and the other three groups receiving different doses of

Teplizumab. The aims were to assess the efficacy, tolerability and safety of Teplizumab.

Efficacy was defined primarily by the capacity of Teplizumab to markedly reduce insulin

requirements while maintaining relatively normal blood sugar levels. In October 2010,

MacroGenics and Lilly announced that the analysis of one-year safety and efficacy data of the

phase III trial of Teplizumab was completed with the result that the primary efficacy endpoint

of the study was not met, indicating a lack of efficacy of Teplizumab (Press release

Macrogenics and Lilly 2010). As a consequence, the companies have decided to suspend

further recruitment and dosing of patients in two ongoing trials of Teplizumab in type 1

diabetes.

GAD-alum intervention studies

In non-obese diabetic (NOD) mice, one of the animal models for type 1 diabetes, several

studies have shown that diabetes can be delayed or prevented via administration of GAD

(Tisch et al. 1993; Tian et al. 1996; Tisch et al. 1999; Jun et al. 2002). This preventive effect

is probably due to restoration of self-tolerance, which seems not to be restricted to the

tolerogen, but spreads to other beta-cell autoantigens (Bach et al. 2001). Tian et al showed

that injection of GAD65 in an adjuvant that induced Th2 responses to GAD65 could greatly

inhibit the progression of the autoimmune process in NOD mice that already had an

established autoimmune process (Tian et al. 1996). This and other experiments led to the

GAD65 vaccine (GAD-alum; Diamyd) that DIAMYD is using in clinical trials on humans.

Aluminum hydroxide (alum) is a commonly used vaccine adjuvant that preferentially induces

a humoral Th2 response and a Th2-driven antibody isotype production against co-injected

antigens, rather than a cellular immune response (McKee et al. 2008).

The safety of GAD-alum was first tested in a phase I trial in healthy humans (Diamyd

Medical, unpublished). Then, the efficacy of GAD-alum treatment as well as further

evaluation of the safety and investigation of the immunological effect was tested in a phase

IIa dose-finding study in a group of LADA patients (Agardh et al. 2005). This study was

followed by a phase IIb clinical trial in newly diagnosed type 1 diabetic children (further

described in the “Study populations” section) (Ludvigsson et al. 2008). Results from the latter


40

study showed a better preservation of fasting and stimulated C-peptide in the individuals with

less than six months duration of disease at inclusion. The effect persisted still 4 years after

treatment (Ludvigsson et al. 2010b). At present, phase III clinical trials with GAD-alum are

ongoing in both Europe and the USA. Results from the European study will be available in

the spring or summer of 2011 (www.diamyd.com). The study comprises 334 type 1 diabetic

patients (10-20 years of age) diagnosed within the last three months. They have GADA

positivity and fasting C-peptide levels above 0.1 pmol/ml (http://clinicaltrials.gov 2010b).

This is a three-armed trial with one group receiving four subcutaneous injections with

placebo, another group GAD-alum at two occasions and placebo at the other two visits, and

the last group four injections of GAD-alum. It will be interesting to follow the results from

the GAD-alum phase III clinical trial.

In conclusion, improved prediction strategies for type 1 diabetes facilitate identification of

individuals suitable for intervention studies. Unfortunately, the ENDIT, DPT-1 and anti-CD3

intervention studies had no effect on the development of type 1 diabetes, but other treatments,

such as GAD-alum, are still promising.

DETECTION OF AUTOANTIBODIES

Autoantibodies related to type 1 diabetes are traditionally detected using radioimmunoassays

(RIAs). The first radioassays for detection of IAA in prediabetic and new onset diabetic

patients used 125

I-labelled insulin in fluid phase assays, with poly(ethylene glycol) (PEG) for

precipitation and relatively large volumes of serum were needed. Antibodies induced by

insulin injection were generally measured with enzyme-linked immunosorbent assays

(ELISAs). However, the detection limit of the ELISAs was higher than of the radioassays,

and the levels of the insulin antibodies induced by insulin injections are approximately 10-

fold higher than the commonly detected levels of IAA in individuals at diagnosis of type 1

diabetes (Winter et al. 2002). In 1992, Greenbaum et al showed that IAA measured by RIA

were more related to type 1 diabetes than those measured by ELISA, and thus recommending

RIA as the method of choice for measuring IAA associated with type 1 diabetes (Greenbaum

et al. 1992). This could be explained by the fact that in RIA, the ligand (insulin) is limiting

and the antibodies are in excess, favoring the binding of high-affinity IAA, and it has been

observed that IAA of high affinity is disease specific (Achenbach et al. 2004a; Schlosser et al.

2005b). On the other hand, in ELISA, an excess of the antigen is fixed to the plate, enabling


41

binding of low-affinity antibodies as well (Sodoyez-Goffaux et al. 1988; Wilkin 1990).

Further, almost all of the epitopes on the insulin molecule would be available for binding in

the fluid phase RIA, while in ELISA some epitopes might be blocked for antibody binding

due to binding of insulin to the solid surface (Greenbaum et al. 1992). To improve sensitivity

and specificity for detection of IAA with RIA, competition with unlabelled insulin was

introduced and micro-assays using smaller volumes of serum and protein A sepharose for

precipitation were developed (Williams et al. 1997) and this method is the standard assay for

IAA today.

When RIA had been established as the most suitable method for detection of IAA related to

type 1 diabetes, RIAs were also developed for the detection of GADA and IA-2A. These

immunoassays, which are widely used today, are based on in vitro synthesized radioactively

labelled (35

S-methionine) antigen and protein A sepharose for precipitation of antibody-

antigen complexes (Grubin et al. 1994). These fluid phase RIA’s have generally demonstrated

higher disease sensitivity compared to ELISAs, which are solid phase assays (Schmidli et al.

1995; Bingley et al. 2003). More recently the ELISAs have improved and some assays with

the same or better sensitivity and specificity than the conventional RIAs have been developed

(Ankelo et al. 2003; Brooking et al. 2003; Luo et al. 2004; Mueller 2007). Two ELISAs are

based on a divalent action of GADA that forms a bridge, between immobilized GAD65 and

liquid-phase GAD65-biotin in one assay (Brooking et al. 2003) and between immobilized

GAD65-biotin and liquid-phase europium-labelled GAD65 in the other (Ankelo et al. 2003).

Another assay uses a biotin-GAD65 fusion protein, which is free in the liquid-phase during

the incubation with antibody-containing serum, and is then bound to the surface for detection

of GADA (Luo et al. 2004). In The Diabetes Autoantibody Standardization Program (DASP)

2007, one laboratory using ELISA for detection of GADA and IA-2A had the highest

sensitivities and specificities among all participating laboratories (Mueller 2007). The high

sensitivity and specificity of these assays opens the question whether new ELISA assays

might replace RIAs for determination of GADA and IA-2A in the future.

RIAs for GADA, IA-2A and IAA are today widely used in laboratories around the world. To

improve and standardize the measurement of type 1 diabetes related autoantibodies, a series

of international workshops have been held (Greenbaum et al. 1992; Schmidli et al. 1994;

Schmidli et al. 1995; Verge et al. 1998) and were followed by the establishment of The

Diabetes Autoantibody Standardization Program (DASP). DASP is a collaboration of the


42

Immunology of Diabetes Society (IDS) and the US Centers for Disease Control and

Prevention. The major goals of DASP are to help laboratories to improve methods for

detection of autoantibodies by providing technical support, training and information, to

organize workshops that evaluate laboratory performance of the assays and to provide

reference material for the development of new measurement technologies (Bingley et al.

2003; Torn et al. 2008). The aim of the first DASP evaluation was to improve comparability

of the autoantibody measurements between laboratories and to evaluate the new WHO

international reference reagent for GADA and IA-2A (Mire-Sluis et al. 2000; Bingley et al.

2003).

There are some disadvantages with the RIAs for detection of GADA, IA-2A and IAA,

including the use of radioactively labelled material, low sensitivity for the IAA assay and long

assay times for a sample. Of the three RIAs, the assay for IAA has been and is still the most

difficult for new laboratories to perform and to standardize (Liu et al. 2007; Schlosser et al.

2010). This could be due to the very low levels of IAA that are present in serum of both

prediabetic and newly diagnosed individuals with type 1 diabetes, and the fact that there is a

very small difference in concentration between autoantibody positive individuals and normal

control subjects. Thus, there is a need for better methods to measure autoantibodies against

GAD, IA-2 and insulin. Surface plasmon resonance (SPR) is an interesting candidate

technique for this purpose (Shankaran et al. 2007). Biacore’s SPR technology is an efficient

method that investigates biomolecular interactions in real-time and uses label-free detection

(Liedberg et al. 1995) (www.biacore.com). This technique can identify binding partners to

target molecules, but also provide quantitative information on specificity of the binding

between two molecules, the concentration of the molecule, rate of association and dissociation

and strength of the binding. Comparison of SPR with two conventional immunoassays, i.e.

RIA and ELISA, has shown that SPR is a more powerful technique with several advantages,

such as the low quantities of reagents needed, no labelling of interactants and measurements

in real time (Revoltella et al. 1998).

In conclusion, RIAs are commonly used for detection of GADA, IA-2A and IAA. There is a

need for improved methods for detection of these autoantibodies, and SPR is an interesting

candidate.

43

Aims of the thesis

The general aims of this thesis were to investigate prevalence and levels of autoantibodies

related to type 1 diabetes, in healthy children and in children with type 1 diabetes. In addition,

we wanted to study different properties of GADA, such as binding epitopes and IgG

subclasses, after immunomodulatory treatment of children with type 1 diabetes.

The specific aims were:

I. To elucidate the influence of HLA genotypes on the development of diabetes-related

autoantibodies and manifest type 1 diabetes in children participating in the ABIS

study. We therefore decided to study the longitudinal changes of GADA and IA-2A

and their relation to HLA genotypes, both in healthy children and in children with

type 1 diabetes.

II. To develop a sensitive and specific method for detection of diabetes-related

autoantibodies based on the surface plasmon resonance (SPR) technique. We thought

that SPR might be a suitable technique due to its label-free detection, the short

analysis time for each sample and the possibility to simultaneously measure multiple

antibodies.

III. To study how treatment with GAD-alum in children with recent onset of type 1

diabetes influences GADA epitope binding. We hypothesized that the epitope

binding pattern would not be influenced by the treatment, since unaltered epitope

binding was previously observed after GAD-alum treatment in LADA patients.

IV. To study if treatment with GAD-alum in children with recent onset of type 1 diabetes

influences IgG subclasses of GADA and GAD65 enzyme activity. We hypothesized

that the IgG subclass distribution would be altered towards a Th2-like phenotype

with more IgG4, since an early GAD65 specific Th2 cytokine response has

previously been observed following this treatment.

44

45

Subjects and methods

STUDY POPULATIONS

The ABIS study

The All Babies In Southeast of Sweden (ABIS) study is a prospective population-based

follow-up study, in which 17055 out of 21700 infants (78.6%) born between 1 October 1997

and 1 October 1999 in Southeast of Sweden were included (Ludvigsson et al. 2001)

(www.abis-studien.se). The aim was to study the role of environmental factors for

development of type 1 diabetes, but also for other diseases, such as celiac disease,

inflammatory bowel disease and allergy. A second aim was to increase the knowledge of how

to identify individuals with risk of developing type 1 diabetes in the general population, using

markers such as autoantibodies and HLA. In ABIS I, 1997-2005, children were followed from

birth up to 5 years of age. The parents were asked to fill in comprehensive questionnaires and

biological samples, such as blood and hair samples, were taken from the children at birth and

at 1, 2.5 and 5 years of age. In the follow-up study, ABIS II, which started 2006, all children

participating in ABIS I were asked to participate. The children were asked to fill in

questionnaires and to contribute with biological samples at 8, 11 and 14 years of age. The

prevalence of diabetes in the ABIS cohort will be investigated, by registration of the diagnosis

from the clinics, until the end of 2017. Results from the study so far include a number of

factors that have been related to development of beta-cell autoantibodies, such as virus

infections early in life (Sepa et al. 2005), short duration of breast-feeding (Wahlberg et al.

2006; Holmberg et al. 2007), early introduction of cow’s milk and late introduction of gluten

during the first year of life (Wahlberg et al. 2005; Wahlberg et al. 2006) and psychological

stress (Sepa et al. 2005; Sepa et al. 2005).

In paper I, samples from subgroups of children who participated in the ABIS study were used

for analyses. Samples were grouped as follows:

1) GADA and IA-2A were investigated in all children who had given at least one serum or

whole blood sample at any age (1, 2.5 or 5 years of age). The relationship between HLA

alleles and GADA and IA-2A was investigated.

Subjects and methods ’

46

2) The occurrence of permanent vs transient autoantibodies was studied in 2329 children from

whom samples at all three ages were available, and the association to HLA genotype was

investigated. In this group, children who had developed type 1 diabetes were excluded.

3) Analyses of GADA, IA-2A and HLA genotype were performed in samples from 1, 2.5

and/or 5 years of age in children who developed type 1 diabetes in the ABIS study.

In paper II, serum samples from children in the ABIS study who had not developed type 1

diabetes were used for measurement of IAA in an indirect competitive immunoassay based on

SPR.

The phase II clinical trial with GAD-alum

A phase II clinical trial with injections of recombinant aluminum formulated human GAD65

(GAD-alum) has been performed in type 1 diabetic patients (Ludvigsson et al. 2008). The aim

of the study was to investigate if GAD-alum treatment could preserve beta cell function in

children with type 1 diabetes. The participants (n=70) were 10-18 years old at diagnosis of

type 1 diabetes and had disease durations of <18 months, fasting c-peptide levels of >0.1

nmol/l and tested positive for GADA at the initial screening. In this randomized, double-

blinded placebo-controlled trial, recent onset type 1 diabetic children and adolescents were

treated with either 20 μg GAD-alum (Diamyd®, Diamyd Medical, Stockholm, Sweden)

(n=35) or placebo (alum-formula alone) (n=35) subcutaneously in a prime and booster

injection four weeks apart. At day 1 and at 3, 9, 15, 21 and 30 months, the patients underwent

a mixed-meal tolerance test to stimulate residual insulin secretion (measured as C-peptide

level). Adverse events were regularly registered throughout the study, and neurological

examination was performed by a pediatrician at regular intervals. The outcome of the study

demonstrated that GAD-alum treatment had a clinical effect on the preservation of beta-cell

function in children with shorter disease duration, and immunomodulated the humoral

immune response, inducing GADA levels in GAD-alum treated individuals (Ludvigsson et al.

2008).

In addition, the effect of GAD-alum on the immune system was studied. Blood samples from

69 patients (one placebo patient dropped out after the first injection) were collected before the

first injection (day 0) and after 1, 3, 9, 15, 21 and 30 months. All samples were transported to

Linköping within 24 hours, and serum samples were stored at −70°C for simultaneous

analysis to avoid inter-assay variation.


47

In paper III, serum samples taken before the first injection (baseline) and at 1, 3, 9, and 15

months after the initial injection were included. Nine patients (7 from the placebo group and 2

from the GAD-alum group) with baseline GADA levels <60 units/ml (U/ml), were excluded

because measurement of competition with rFab in samples with low GADA levels is

unreliable. One additional patient in the placebo group was excluded, since this individual

dropped out after the first injection. Thus, GADA epitope specificity was analyzed in 27

patients from the placebo group and 33 from the GAD-alum group using rFab fragments in an

epitope-specific radioligand binding assay.

In paper IV, the IA-2A and tetanus toxoid titers were measured at baseline, 3 and 9 months.

GAD65 enzyme activity and IgG subclass measurements were performed at baseline, 3 and

15 months. Total IgE was analysed at baseline and 21 months.

Type 1 diabetic patients

In paper II, samples taken from children with manifest type 1 diabetes were studied. These

children have been recruited at the Pediatric Clinic at Linköping University hospital and the

diagnosis was based on the WHO criteria from 1999. All samples used were taken within 10

days after diagnosis, to be sure that the IAA measured were antibodies against endogenous

insulin and not antibodies produced after encounter of exogenous insulin.

METHODS

RIA for detection of GADA and IA-2A (Paper I and IV)

GADA and IA-2A were determined using a RIA with 35

S-labelled human recombinant

GAD65 and IA-2ic, respectively. These antigens were produced by in vitro transcription and

translation of complementary DNAs of GAD65 and IA-2ic, cloned into vectors (pcDNA2 and

pSP64 poly (A), respectively; kindly supplied by Prof. Å Lernmark, Seattle), in the presence

of 35

S -labelled methionine. Serum or whole blood samples diluted 1:25 were incubated in

duplicates with GAD or IA-2 antigen at +4°C over night. This was followed by precipitation

with protein A sepharose (Amersham Biosciencies, Uppsala, Sweden) to separate free

labelled antigen from antibody bound labelled antigen. A standard curve was included on

each plate, as well as a blank (containing only buffer) and positive and negative controls. The


48

immunoprecipitated radioactivity was measured in a Wallac 1450 Microbeta Liquid

Scintillation Counter (Perkin Elmer Life and Analytical Sciences), and the results were

expressed as U/ml in relation to the standard curve.

In paper I, cut-off for positivity at the 98th

percentile for GADA and IA-2A was used.

However, at 5 years of age, the 99th

percentile was used for IA-2A since less than 2% of the

samples had values above the detection limit. In the DASP workshop 2005, our GADA assay

had, at the 98th

percentile, a specificity of 95% and a sensitivity of 76% (Table 2). For IA-2A,

at the 99th

percentile, the specificity was 100% and the sensitivity 72%. Interassay variations

for negative and positive controls were 10% and 8% for GADA and 11% and 12% for IA-2A,

respectively.

Table 2. Specificities and sensitivities for our autoantibody assays from the Diabetes Autoantibody

Standardization Program (DASP) during the years 2003-2010. Measurements of GADA, IA-2A and

IAA were included all time points, while the three variants of ZnT8A; ZnT8(Arg)A, ZNT8(Gln)A and

ZnT8(Trp)A, were included for the first time in DASP 2010. The percentiles used as cut-off for

positivity varied between antibodies and years and are shown in the table. * = In 2007, the cut-off

value used for IA-2A was 9.9 U/ml, which exceeded the 99th percentile.

Year GADA IA-2A IAA ZnT8(Arg)A ZnT8(Gln)A ZnT8(Trp)A

Sensitivity 74% 48% 24% - - -

Specificity 98% 100% 100% - - -

Percentile 98th 98th 99th - - -

Sensitivity 76% 72% 28% - - -

Specificity 95% 100% 100% - - -


Sensitivity 82% 65% 25% - - -

Specificity 94% 98% 99% - - -

Percentile 98th >99th* 98th - - -

Sensitivity 66% 64% 40% - - -

Specificity 99% 99% 97% - - -


Sensitivity 80% 66% 34% 52% 28% 46%

Specificity 100% 100% 98% 97% 100% 91%

Percentile 98th 99th 98th 99th 99th 99th

2003

2005

2007

2009

2010


49

In paper IV, the cut-off for positivity determined at the 95th

percentile was regarded as 67.3

U/ml (corresponding to 23.1 WHO units) based on measurements from 1700 children, aged 5-

6 years, participating in the ABIS study. Validation of our assay in DASP 2007 showed 94%

specificity and 82% sensitivity for GADA and 98% specificity and 65% sensitivity for IA-2A

(Table 2).

HLA genotyping (Paper I)

HLA DQB1, DQA1 and DRB1 alleles were determined with a time-resolved fluorometry

based sandwich hybridization assay, and performed by the research group of Professor Jorma

Ilonen in Turku, Finland. Details of the method have been described earlier (Sjoroos et al.

1998; Nejentsev et al. 1999; Laaksonen et al. 2002; Hermann et al. 2003). Briefly, the

polymerase chain reaction (PCR) was carried out using a 3 mm in diameter disc from whole

blood spots on a filter paper punched directly into the PCR reagent mixture or using DNA

extracted from whole blood. For amplification of the genes, specific HLA-DQA1, DQB1 or

DRB1 primer pairs, of which one was biotinylated at the 5’-end, were used. The PCR product

was transferred into a streptavidin coated microtiter plate and heat denatured, followed by

addition of detection probes labelled with various lanthanide chelates (Europium, Samarium

or Terbium). After an incubation step, allowing hybridization of the PCR product and the

probes, the time-resolved fluorescence was detected in a VictorTM

1420 Multi label Counter

(Wallac Oy, Turku, Finland) (Figure 9). Differences in the emission wave lengths of the

lanthanides and the usage of a time delay before measurement enable the detection of the

three lanthanide labels.

For genotype analysis, HLA haplotypes were categorized into susceptibility associated (S),

neutral (N) and protective (P) ones. This classification was done based on the results of

haplotype-conferred risk in 622 Finnish families with one child affected by type 1 diabetes

(Hermann et al. 2003), but essentially similar results have been described in other European

populations (Koeleman et al. 2004; Lambert et al. 2004). Susceptibility associated haplotypes

included DR4-DQ8 (DRB1*0401/2/4/5-DQB1*0302) and DR3-DQ2 (DQA1*05-DQB1*02),

protective haplotypes included DR2-DQ6 (DQB1*0602), DR11/12/1303-DQ7 (DQA1*05-

DQB1*0301), DR7-DQ3 (DQA1*0201-DQB1*0303), DR14-DQ5 (DQB1*0503), DR403-

DQ8 (DRB1*0403-DQB1*0302) and DR1301-DQ6 (DQB1*0603) (Table 1). Other

haplotypes were defined as neutral.


50

Figure 9. Principle of time-resolved fluorometry based sandwich hybridization assay. PCR

amplification of DNA with a biotinylated primer was followed by hybridization with specific

detection probes on streptavidin coated microtiter plates and thereafter measurement of time-resolved

fluorescence. SA = streptavidin, bio = biotin.

RIA for detection of IAA (Paper II)

A competitive RIA, based on the method developed by Williams et al (Williams et al. 1997)

with some modifications (Holmberg et al. 2006), was used for detection of IAA. Serum

samples in quadruplicates (7.5µl/well) were incubated in a 96 deep-well plate with either 125

I-

insulin alone (2 wells) (25 000-30 000 cpm/well of human recombinant insulin, mono-iodated

at thyrosine position A14, Amersham Biosciences, Buckinghamshire, Great Britain) or with

125I-insulin together with an excess of unlabelled insulin (2 wells) (human recombinant

insulin, Sigma-Aldrich, St. Louis, MO, USA) shaking for 72 h at +4°C. A standard curve as

well as positive and negative controls were included on each plate. After incubation, the

insulin-antibody complexes were precipitated with Protein A Sepharose (Zymed, San

Francisco, CA, USA) and washed five times by centrifugation at +4°C, followed by

measurement of the radioactivity in a gamma counter (1282 CompuGamma, Wallac, Turku,

Finland). Specific antibody binding were calculated by subtraction of mean counts per minute

(cpm) of duplicate samples with an excess of unlabelled insulin from mean cpm of duplicate

samples with labelled insulin. The results were expressed as U/ml calculated in relation to a

standard curve. The cut-off limit for positivity corresponding to the 98th

percentile of 114

healthy Swedish children was 6.45 U/ml. Validation of our IAA assay in DASP 2003 showed

100% specificity and 24% sensitivity (Table 2).

SA

bio

Europium

SASASA SASA

probe

SASASA

bio

Samarium

bio

Terbium


51

Surface plasmon resonance (SPR) (Paper II)

Surface plasmon resonance (SPR) is an optical surface phenomenon that can be used to

monitor biomolecular interactions in real-time without the need for labelling (Liedberg et al.

1995; Shankaran et al. 2007) (www.biacore.com). The instrument used in our study, Biacore

X, is one of the commonly used instruments for SPR-based assays. One of the interacting

molecules, the ligand, is immobilized on a sensor chip consisting of a glass surface coated

with a thin layer of gold usually covered with a dextran hydrogel. The sample containing the

other interacting molecule, the analyte, is then injected over the surface in a controlled flow.

The SPR phenomenon occurs when light, under conditions of total internal reflection, strikes

a thin electrically conducting gold layer (Figure 10). Here, this gold layer is at the interface

between a glass slide and the flow channel with buffer, which have different refractive

indexes. When light strikes the gold film through a glass prism docked to the slide, an electric

field is generated, which extends from the opposite film interface into the buffer. At a certain

angle of incidence, the resonance angle, this so-called evanescent field excites the surface

plasmon, which is a collective oscillation (wave) of electrons that propagates along the gold

interface. This excitation is called surface plasmon resonance. When monitoring the intensity

of the reflected light as a function of the angle of incidence, the resonance angle can be

identified from a minimum. This is because most of the light energy has coupled into the

surface plasmon upon excitation at this angle. The resonance angle is dependent upon the

refractive index of the medium (the buffer or analyte solution) close to the gold layer. In this

way, changes of the refractive index at or close to the surface, for example due to binding of

biomolecules, results in displacement of the reflectance minimum (Figure 10a), which is

monitored as a function of time in a so-called sensorgram (Figure 10b). The displacement is

expressed in resonance units (RU, 1 RU ~1 pg/mm2 of any biomolecule). The contribution to

the refractive index change by any biomolecule binding to or dissociating from the surface is

usually proportional to the molecular mass. After the analysis, analyte molecules can be

removed using an appropriate regeneration solution that, ideally, does not affect the activity of

the immobilized ligand.


52

Figure 10. The principle of surface plasmon resonance, with a protein immobilized on the sensor

surface and a sample containing antibodies injected over the surface. Polarized light that strikes the

surface at a certain angle, the resonance angle, generates an electric field that excites a surface

plasmon, resulting in that the intensity of reflected light is at a minimum (I). When antibodies bind to

the protein, the refractive index close to the surface is changed, resulting in displacement of the

reflectance minimum (from I to II in graph a), which is displayed in a sensorgram (b), expressed as

resonance units (RU).

The SPR technique can identify binding partners to molecules of interest and monitor binding

specificity, but also provide quantitative information on analyte concentration as well as

kinetic and thermodynamic parameters describing the interaction.

The design of the sensor chip differs depending on the application. The carboxymethylated

(CM) dextran chip CM5 is a commonly used commercially available general-purpose chip for

interaction analysis involving all types of biomolecules (Biacore, Uppsala, Sweden). The

sensor chip is coated with gold and covered with a CM dextran layer (Figure 11). It is suitable

for concentration measurements, and is generally the first choice for protein immobilization

via amine coupling.

Sensor chip withgold film

Flow channel

Light-source

Polarizedlight

Prism

Opticaldetection

unit

III

Reflectedlight

I II

Intensity

Angle

Time

Resonancesignal

I

II

Sensorgram

a)

b)


53

Figure 11. The structure of a CM5 chip. At the bottom, the sensor chip coated with gold. On top of

that the dextran layer, which has been carboxylated.

Prior to immobilization of a ligand, a process called pre-concentration is often performed, to

get high absorption of the protein into the dextran matrix, and is optimized by adjusting the

pH-value. When using a CM5 sensor chip, the pH of the solution should be acidic to pre-

concentrate the protein in the negatively charged CM dextran matrix. This is usually obtained

at a pH of the antigen solution of approximately 1 pH unit below the isoelectric point of the

antigen.

Competitive epitope-specific RIA for determination of GADA epitopes

(Paper III)

The capacity of different rFab fragments to inhibit GADA binding to GAD65 was tested in a

competitive epitope-specific RIA. Serum samples were initially screened for GADA using the

traditional RIA, at a dilution of 1/25. Samples with a GADA titer exceeding 1000 U/ml were

diluted to <1000 U/ml and reanalyzed. This dilution was then used in the epitope-specific

RIA. Six GAD65-specific rFab were used to determine the GADA epitope specificity in

serum samples. The rFab were cloned, expressed and purified as described previously (Padoa

et al. 2003). Four of these have been related to type 1 diabetes (b96.11, DPA, DPD and

MICA3) (Richter et al. 1993; Padoa et al. 2003) and the other two (b78 and N-GAD65 mAb)

have been found in SPS patients (Raju et al. 2005). These rFab recognize epitopes on different

regions of GAD65, ranging from the N-terminal to the C-terminal part of the molecule

(Fenalti et al. 2008) (Figure 12). DPA and MICA3 are derived from two different patients

with type 1 diabetes, and recognize epitopes within the same epitope region on GAD65. The

insulin-specific rFab CG7C7 (Padoa et al. 2005) was included on each plate to determine non-

specific competition. The rFab were added at a concentration sufficient to compete binding of

the originating intact monoclonal antibody (mAb) to GAD65 by at least 80%, resulting in 4,

Sensor Chip CM5


54

0.6, 1.7, 1, 14, 7 and 2 μg/ml for rFab DPA, b96.11, DPD, MICA-3, N-GAD65 mAb, b78 and

CG7C7, respectively.

Figure 12. Illustration of the six rFab used in the competitive epitope-specific RIA, showing the

epitope regions on GAD65 recognized by each of them. The rFab recognize epitopes located at the N-

terminus (N-GAD65 mAb and DPD), middle region (b96.11) and C-terminus (DPA, MICA3 and

b78). DPA and MICA3 are derived from two different patients with type 1 diabetes, and recognize

epitopes within the same epitope region on GAD65.

Serum samples were incubated with rFab to inhibit GAD65 binding, or without rFab for non-

competed measurement, and the levels of GADA were determined by precipitation with

Protein A Sepharose. For each sample, non-competed GADA as well as GADA competed

with a GAD65-specific rFab, were measured in triplicates, from which a mean value was

calculated. This was done for each timepoint: baseline and month 1, 3, 9 and 15. The intra-

assay average coefficient of variation was 4%, with the highest value of 13% and the lowest

being 0.03%. Binding of GADA to GAD65 in the presence of rFab was expressed as follows:

ratio = GADA cpm in the presence of rFab (competed) / GADA cpm in the absence of rFab

(non-competed). A higher binding to GAD65 in the presence of an rFab indicates a lower

proportion of GADA binding to the respective epitope. The cutoff for specific competition

was determined as >15% by using the negative control rFab CG7C7.

A modified RIA for detection of IgG subclasses of GADA (Paper IV)

GADA of the subclasses IgG1, 2, 3 and 4 were measured using a modification of the

conventional RIA for GADA, based on a protocol kindly provided by Shilpa Oak (University

of Washington, Seattle, USA), with some further modifications in our lab. Briefly, serum

samples (5 µl) were incubated with 60 µl human recombinant 35

S-labelled GAD65 in a 96-

N-GAD65 mAb DPD DPAb96.11

96 173 308 365 483-499 556-585

b78

GAD65

483-499 556-585

532 540

MICA3

N-terminal C-terminal

224


55

well plate (Nunc 96 MicroWell plate, Nunc A/S, Roskilde, Denmark) under vigorous shaking

at 4°C over night. In parallel, biotinylated monoclonal mouse anti-human IgG1, IgG2 and

IgG4 (BD PharMingen, San Diego, CA, USA) and IgG3 (Southern Biotech, Birmingham,

AL, USA) were incubated with streptavidin agarose beads (Thermo Scientific, Rockford, IL,

USA), under vigorous shaking at 4°C over night. Thereafter, 50 µl of the mixture with 35

S-

labelled GAD65-GADA complexes and 50 µl of the biotinylated anti-human IgG antibody

coupled to streptavidin agarose beads were transferred into a 96-well filtration plate

(Millipore, Bedford, MA, USA), and incubated under vigorous shaking at 4°C for 2 h,

followed by washing 8 times with 150 µl/well assay buffer using a vacuum device

(Millipore). Then, scintillation liquid (OptiPhase Supermix, Perkin-Elmer Life Sciences

Wallac) was added to the wells, and the activity was measured in a liquid scintillation counter

(1450 Microbeta Trilux, Perkin-Elmer Life Sciences, Wallac). All samples were run in

duplicates. The cut-off value for each subclass was determined using a GADA negative

control, which was included in each assay. Results were expressed as cpm, and positivity of

each sample was calculated by subtraction of the mean cpm value plus three times the

standard deviation (SD) obtained for the negative control.

GAD65 enzymatic activity assay (Paper IV)

GAD65 enzyme activity was measured by a 14

CO2-trapping method based on the enzymatic

reaction of glutamate to GABA. GAD65 works as an enzyme together with the coenzyme

pyridoxal 5-phosphate (PLP). GADA-positive serum from SPS patients have shown to inhibit

this reaction (Raju et al. 2005). First, 150 l of a solution containing K2HPO4 (stock solution

50 mM), PLP (stock solution 3 mM, Fisher Scientific, Göteborg, Sweden) and GAD65 (stock

solution 0.2 mg/ml; Diamyd Diagnostics AB, Stockholm, Sweden) was mixed with 15 l

serum and incubated for 1 hour at room temperature. Thereafter, 28 l of a mix containing

K2HPO4 (stock solution 50 mM), L-glutamic acid (stock solution 5 mM, Fisher Scientific,

Göteborg, Sweden) and L-[14

C-(U)]-glutamic acid (0.4 Ci; Perkin Elmer, Boston, USA) was

added to each tube. All tubes were sealed with a rubber stopper attached with a center well

(Kimble Chase Kontes, Vineland NJ, USA), in which a NaOH-soaked (50 l of 1 M NaOH)

filter paper (Camlab Ltd, Cambridge, UK) was placed, and the reaction was carried out in a

water bath at 37°C for 1 hour with gentle agitation. The reaction was stopped by placing the

tubes on ice for 10 minutes. The radioactivity of the 14

CO2 captured on the filter papers was

measured in a Wallac Microbeta Liquid Scintillation Counter (Perkin Elmer Life and


56

Analytical Sciences, Inc, Boston, MA, USA). Serum from one SPS patient was included in all

assays as a positive control for inhibition.

Analysis of tetanus toxoid antibodies (Paper IV)

Determination of tetanus toxoid antibodies in serum was performed using an Immunozym

Tetanus Ab ELISA according to the manufacturer´s instructions (IBL, Hamburg, Germany).

The optical density was measured at 450 nm using a VersaMax microplate reader (MDS, Inc.,

Sunnyvale, CA, USA). Cut-off levels were determined at >0.070 IU/ml.

Analysis of IgE (Paper IV)

Total serum IgE was quantified using the ImmunoCap100 system (Phadia AB, Uppsala,

Sweden). In short, the principle of the method is binding of total IgE to anti-IgE covalently

coupled to ImmunoCap, followed by addition of enzyme-labelled antibodies against IgE,

addition of a developing agent and finally measurement of the fluorescence of the eluate. The

fluorescence is directly proportional to the concentration of IgE in the serum sample. The

measuring range for the assay was 2-50 000 kU/l, and calibrators were run in duplicates to

obtain a full calibration curve. Levels of total IgE ≥ 85 kU/l were regarded as positive.

C-peptide measurement (Paper IV)

C-peptide levels were measured in serum samples using a time-resolved fluoroimmunoassay

(AutoDELPHIATM

C-peptide kit, Wallac, Turku, Finland). Briefly, 25 µl of standards,

controls and patient serum samples were incubated together with anti-C-peptide-Eu tracer

solution in a microtiter plate coated with antibodies directed against C-peptide. After washing

and addition of enhancement solution, automatic measurement and result calculation was

performed using a 1224 MultiCalc® program (Wallac), with results expressed in pmol/ml.

The results in each assay were validated by inclusion of a C-peptide control module

containing a high, a medium and a low-level control (Immulite, DPC, UK).


57

STATISTICS

To elucidate if the immunological markers were normally distributed, the D’Agostino &

Pearson omnibus normality test was performed. One can also evaluate if a material seems to

be normally distributed by examining the distribution of the material in a histogram by

looking if the material has a similar shape as a normally distribution curve, or by comparing

mean and median values, which should be close to each other to support normally

distribution. When the material is large, the outcome data does not necessarily have to be

normally distributed for using parametric tests based on normally distributions (Lumley et al.

2002). A sufficiently large material in this context could be less than 100, but for

consideration of parametric tests, other parameters than actual size of the material are also

important, for example the shape of the distribution in the groups to be compared.

As the immunological markers in paper III and IV were not normally distributed, non-

parametric tests corrected for ties were used. In paper III, to investigate differences in

variability of the ratios over time, the sums of absolute differences between consecutive

measurements were compared using Mann-Whitney U-test. Unpaired analysis was performed

using Mann-Whitney U-test and correlations with Spearman’s rank order correlation test. Dot

plots and longitudinal plots were used to visualize the data. In paper IV, unpaired analyses

were performed using Kruskal-Wallis followed by the Mann-Whitney U-test. Differences

within the groups were analyzed by Friedman’s test followed by Wilcoxon-signed rank test.

In paper I, Fisher´s exact test was used to analyze frequencies of GADA and IA-2A, and the

relationship of HLA genotypes to autoantibodies and type 1 diabetes. Standard curves for RIA

were performed using GraphPad Prism 4 (GraphPad, Sotfware Inc., San Diego, California,

USA). Positive predictive value (PPV) in % was calculated from the formula: (antibody

positive who developed type 1 diabetes) / (antibody positive who developed type 1 diabetes +

antibody positive healthy individuals) * 100

In paper II, a cut-off for positivity for the 5-min method was calculated from the mean value

of a number of low IAA positive samples subtracted by 3 SD. In the standard curve plot with

error bars, the mean values are presented and the error bars show the range.


58

All statistical calculations were performed using the statistical package for the social sciences

(SPSS) for Windows (SPSS Inc., Chicago, Illinois, USA). In paper I version 13.0 was used, in

paper IV version 16.0 was used and in paper III the version named PASW Statistics 18 was

used. A two-tailed p-value of <0.05 was considered statistically significant.

ETHICAL CONSIDERATIONS

Since type 1 diabetes often develops in childhood, it is important to study environmental and

immunological factors influencing the development of type 1 diabetes in children, and not

only in adult individuals. It is very important to get more knowledge about why this disease

arises, to be able to prevent children from developing type 1 diabetes in the future.

Autoantibodies are used to predict the development of type 1 diabetes, which is crucial for the

identification of risk individuals to be included in prevention trials when candidate drugs are

available.

The protocol of the ABIS study was approved by the Research Ethics Committees of the

Medical Faculties of Linköping University, Linköping and Lund University, Lund. Parents

gave their informed consent to enter the study, and they were informed that the participation

was totally voluntary and could be discontinued at any time point without any specific reason.

The parents were told that they, upon active request, could receive information about genetic

risk genotypes or presence of autoantibodies, but that they were not going to be informed

about this automatically. The data from questionnaires and laboratory analyses were stored

without personal identifications, with only an ABIS code that identified the children in the

study. Ethical questions in the ABIS study have been studied separately (Stolt et al. 2002;

Stolt et al. 2005) as well as the maternal attitude to participation (Ludvigsson et al. 2001).

Collection of samples from children with manifest type 1 diabetes was approved by the

Research Ethics Committee of the Medical Faculty of Linköping University, Linköping and

parents gave their informed consent.

Approval for the phase II GAD-alum trial was obtained from the Swedish National

Regulatory authorities and the Research Ethics Committee of the Medical Faculty of

Linköping University, Linköping. Written informed consent was obtained from participants

and their parents according to the Declaration of Helsinki. The study was randomized and


59

double blinded. Since GAD-alum treatment seems to be safe, is easy to administrate and has

shown to be effective for preservation of endogenous insulin production compared to a

control group, we believe that the ethical benefits of the GAD-alum treatment exceed the

risks.

60

61

Results and discussion

AUTOANTIBODY POSITIVITY AND RELATION TO HLA GENOTYPES

(PAPER I)

The emergence of autoantibodies against beta-cell antigens is a dynamic process (Yu et al.

1996; Knip 1997). So far, there is no consensus on how often and at what age children in the

general population should be tested to achieve the best possible prediction value for clinical

type 1 diabetes. The emergence of autoantibodies has been investigated in different study

populations, including families with first-degree relatives of individuals having type 1

diabetes, and individuals with increased risk for type 1 diabetes determined by disease-

associated HLA genotypes. First-degree relatives of patients with type 1 diabetes have for

example been studied in the BABYDIAB (Hummel et al. 2004), DPT-1 (Krischer et al. 2003),

Childhood Diabetes in Finland (DiMe) (Savola et al. 2001) and the Bart´s Oxford and the

Munich family studies (Achenbach et al. 2004b). The DIPP study (Kimpimaki et al. 2001;

Kupila et al. 2001; Kukko et al. 2005) investigated instead populations with disease-

associated HLA genotypes. In the TEDDY (The TEDDY Study Group 2007) and DAISY

(Barker et al. 2004) studies, both individuals with HLA-risk genotypes and those with first-

degree relatives were included.

The rapid increase in the incidence of type 1 diabetes during the last decades cannot be

explained by genetic factors, but is probably instead related to changes in the environment.

Several environmental factors that seem to be related to development of type 1 diabetes have

been found, including viral infections, introduction of cow´s milk and gluten, beta-cell stress

and breast-feeding (Akerblom et al. 2002; Peng et al. 2006). To elucidate the influence of

environmental factors in the development of beta-cell autoimmunity and type 1 diabetes, it

would be appropriate to follow a general population that has not been selected for genetic

risk. Therefore, we studied the emergence of autoantibodies in children from the ABIS study,

and related the presence of autoantibodies to HLA-risk and -protective genotypes.

Results and discussion ’

62

GADA and IA-2A in children at three different ages

We started to investigate GADA and IA-2A in children from the ABIS study, from whom

samples were collected at 1, 2.5 and/or 5 years of age. Positivity to an autoantibody was then

related to HLA genotypes. We found that children positive for GADA or IA-2A at 5 years of

age had the HLA risk haplotype DR4-DQ8 or the high-risk genotype DR3-DQ2/DR4-DQ8

more often than autoantibody negative (GADA and IA-2A negative) children. In addition,

children at 5 years of age positive for IA-2A had the protective haplotype DR2-DQ6 more

seldom and the risk associated genotype DR4-DQ8/DR4-DQ8 more often than autoantibody

negative children. At 2.5 years of age, children positive for IA-2A had the protective genotype

DR2-DQ6/DR2-DQ6 more seldom than autoantibody negative children.

Similar relations between autoantibodies and HLA-risk and -protective alleles have been

found in other studies, as presented in Table 3. For example, GADA has been associated with

the HLA DR3-DQ2/DR4-DQ8 genotype (Kulmala et al. 2001; Bakhtadze et al. 2006) and the

DR4-DQ8 haplotype (Kulmala et al. 2001), which was also found in our study, as well as

associations for IA-2A with the DR4-DQ8 haplotype (Bakhtadze et al. 2006). An explanation

for differences in associations between autoantibodies and HLA alleles could be that in some

studies the autoantibodies were measured in individuals at onset of type 1 diabetes (Hagopian

et al. 1995; Genovese et al. 1996; Savola et al. 1998; Bakhtadze et al. 2006) while in other

studies in pre-diabetic individuals (Kulmala et al. 2000; Kulmala et al. 2001). Not all of the

healthy individuals with autoantibodies will develop diabetes, and therefore the HLA

genotypes in non-progressors might differ from the ones observed in individuals recently

diagnosed with type 1 diabetes. In our study, associations between beta-cell autoantibodies

and HLA-risk alleles were found only in children at 5 years of age. The lack of association at

1 and 2.5 years of age suggest that HLA-risk alleles related to type 1 diabetes are not

necessarily needed for presentation of beta-cell antigens and for induction of beta-cell

autoantibodies.


63

Table 3. Associations for GADA or IA-2A positive children with HLA alleles, found in our study

and/or in previous studies. aab = autoantibody.

Permanent and transient autoantibodies

To further evaluate the associations of HLA genotypes with the presence of autoantibodies,

we investigated samples collected at all three time points (1, 2.5 and 5 years of age) from a

group of non-diabetic children in the ABIS study (Figure 13). Eight of 2329 (0.3%) children

were antibody positive at least at 2.5 and 5 years of age, i.e. had permanent autoantibodies. In

addition, 143 of 2329 (6%) children were autoantibody positive only at one time point

followed by autoantibody negativity, i.e. had transient autoantibodies. Children with

permanent autoantibodies had more often the risk associated DR4-DQ8 haplotype than

autoantibody negative children, while no associations with HLA risk or protective genotypes

were found for transient autoantibodies.

A previous study has shown that more than 50% of the children with permanent

autoantibodies had the high-risk genotype DR3-DQ2/DR4-DQ8, while no children with

transient autoantibodies had this genotype (Yu et al. 2000). That study demonstrated the

occurrence of permanent antibodies in 0.9% of children from the general population, which

was slightly higher than the percentage observed in our study (0.3%). However, it should be

noted that the study by Yu et al. differed in some aspects from our study. That study included

measurements of three autoantibodies; GADA, IA-2A and IAA, the age interval of the

children was 0-13 years, and the consecutive measurements were performed more often;

annually in autoantibody negative children and every 3-6 months after autoantibody positivity

was found. Further, in our study the group of children with permanent autoantibodies was

small, with only eight individuals, which makes it difficult to draw conclusions. In a recent

Autoantibody HLA alleles associated with aab's References

DR3 Genovese et al. 1996; Kulmala et al. 2000

DR3-DQ2/DR4-DQ8 Bakhtadze et al. 2006; Kulmala et al. 2001+our study

GADA DR4-DQ8 Kulmala et al. 2001+our study

DR3-DQ2 Savola et al. 1998; Hagopian et al. 1995

DR2-DQ6 (negatively associated) Kulmala et al. 2001

DR4 Genovese et al. 1996; Savola et al. 1998

DR4-DQ8 Bakhtadze et al. 2006; Kulmala et al. 2001+our study

IA-2A DR4-DQ8/DR4-DQ8 our study

DR2-DQ6 (negatively associated) our study

DR2-DQ6/DR2-DQ6 (negatively associated) our study


64

study, both HLA risk genotypes and the T allele of the PTPN22 gene were shown to influence

progression to permanent islet autoimmunity (Steck et al. 2009).

Figure 13. Selection of children from the ABIS study for the follow-up of GADA and IA-2A at 1, 2.5

and 5 years of age. Thirty-two children had developed type 1 diabetes by the age of 6-7 years. Of the

2329 children from whom samples were available at all three ages, permanent and transient

autoantibodies were studied and related to HLA genotypes. A subgroup of autoantibody negative

children were randomly selected for HLA genotyping.

We found that children with transient autoantibodies and autoantibody-negative children had

a very similar pattern of HLA genotypes (Figure 14). This suggests that HLA-risk alleles are

not necessarily needed for the induction of beta-cell autoantibodies, but is likely rather

influenced by environmental factors or other genetic factors than HLA.

sample available

the occasion

before

diagnosis

All children in the ABIS study

n=17055

n=32developed

diabetes n=2329

n=22 n=10

GADA IA-2A

5

samples available

at all 3 ages

Yes No

54

n=8

aab

negative

n=14

n=678 n=143 n=8

aab

negative

transient

aab´s

permanent

aab´s

GADA and/or

IA-2A positive

n=2178 n=143 n=8

aab

negative

transient

aab´s

permanent

aab´s

HLA genotyped


65

Figure 14. Percentages of individuals with risk, protective, a combination of risk and protective and

neutral HLA genotypes in children with type 1 diabetes, in healthy children with permanent

autoantibodies or transient autoantibodies and in healthy autoantibody-negative children.

Children with type 1 diabetes

By the age of 6-7 years we identified 32 children with type 1 diabetes among the 17055

participants in the ABIS study. HLA risk genotypes associated with type 1 diabetes were

frequently found in these children, whereas protective alleles were rare (Figure 14). The

percentages of HLA-risk and -protective haplotypes corresponded well with other studies,

showing that nowadays the contribution of high-risk HLA haplotypes is lower and the

proportion of protective haplotypes is higher among type 1 diabetes patients compared to 50

years ago (Hermann et al. 2003; Gillespie et al. 2004). This changed distribution of HLA

haplotypes could indicate that environmental factors are important for development of type 1

diabetes, not only in children at later ages, but also in young children as in our study. Three

percent (1/32) of the diabetic children in our study did not have any HLA-risk alleles, in

accordance with a study by Gillespie et al. where 5-8% of children diagnosed before 5 years

of age lacked risk-associated alleles (Gillespie et al. 2004).

By studying samples from the time point preceding the onset of type 1 diabetes we found that

64% of the children were positive for GADA and/or IA-2A. This is lower than what have

been reported by other studies (Bonifacio et al. 1995; Sabbah et al. 1999; Strebelow et al.

1999; Winter et al. 2002). The lower frequency of autoantibody positivity in our study might

be due to a longer interval between collections of samples, and in many individuals the last

sample before onset was taken a long time before onset (range 0-2.5 years). Thus, it cannot be

excluded that autoantibodies may have developed short after the last sampling. Another factor

that could contribute to the lower percentage of autoantibody positivity is that IAA was not

Type 1 diabetes Permanent aab´s Transient aab´s Negative aab´s

Risk

91%

Protective

0%

Neutral

3%

Risk and

protective

6%

Risk

50%

Protective

25%

Risk and

protective

25%

Neutral

0%Risk

28%

Protective

40%

Risk and

protective

19%

Neutral

13% Risk

27%

Protective

38%

Neutral

14%

Risk and

protective

21%


66

measured in our study, because of the small volumes of EDTA blood that was collected from

the children at 1 and 2.5 years of age.

In conclusion, the strong relation between HLA risk alleles and type 1 diabetes observed in

our study confirms that HLA risk genotypes play an important role for development of type 1

diabetes. In contrast, HLA genotypes seem not to explain induction of autoantibodies,

especially transient autoantibodies, in the general population. Our results indicate that other

factors than HLA are involved in the initiation of autoimmunity.

DEVELOPMENT OF AN IMMUNOASSAY BASED ON SPR (PAPER II)

Despite a lot of research in the field of type 1 diabetes, there is still no prevention for the

disease. However, a number of ongoing and planned trials raise the hope for successful

prevention strategies (Schloot et al. 2007; Pescovitz et al. 2009; Rother et al. 2009; Tian et al.

2009; Ludvigsson 2011). If any treatment demonstrates to be beneficial for prevention, the

need to identify risk individuals is large. Although HLA-risk genotypes define some of the

risk individuals (Ilonen et al. 2002), the best way to identify risk individuals is through

autoantibodies (Verge et al. 1996; Bingley et al. 1997). Thus, large scale screening of

autoantibodies in children from the general population will be necessary to identify high-risk

individuals suitable for prevention treatment. At present, screening for autoantibodies is

commonly performed by the well established RIA methods. However, there are some

disadvantages with these methods, including the use of radioactively labelled material,

variations among the “in-house” batches of the GAD and IA-2 proteins, low sensitivity for the

IAA assay and long assay times for a sample. Thus, there is a need for better methods for

detection of autoantibodies.

The surface plasmon resonance (SPR) technology is an interesting candidate for detection of

autoantibodies, mainly due to the label-free detection, the short analysis time for each sample

and the possibility to simultaneously measure multiple antibodies. We therefore decided to

develop a method based on this technology, by using a Biacore instrument, for detection of

three of the autoantibodies commonly related to type 1 diabetes; GADA, IA-2A and IAA. We

aimed to decrease the analysis time per sample, and if possible, to increase the sensitivity of

the autoantibody measurement compared to RIA. We decided to start with GADA, since this

autoantibody is one of the most commonly observed in pre-diabetic children. In addition, a


67

better method for GADA was of special interest for us since the GAD-alum trial was about to

start. This treatment might be a good candidate to be used in prevention trials with GADA

positive individuals.

Attempts to develop an SPR-based immunoassay for GADA

We immobilized GAD65 at pH 5.0 via amine coupling, resulting in moderate levels of

immobilized protein at the dextran surface. However, the injections of GAD65 mAbs and a

GADA positive serum, to the surface immobilized with GAD65, induced low and inconsistent

binding responses. A possible explanation for this could be that the conformation of the

GAD65 protein might have changed upon binding to the surface. In addition, after

regeneration with NaOH, the binding responses were different from the ones observed before

regeneration. We tried to overcome these problems by attaching the GAD65 protein to the

surface using another coupling chemistry. Instead of immobilizing GAD65 directly to the

sensor surface, we immobilized streptavidin to the surface and then added biotinylated

GAD65, which binds to streptavidin. However, this setup did not perform better than GAD65

immobilized directly to the surface.

A number of different regeneration solutions and cocktails, such as NaOH (pH ranging from

10.7 to 12.7), glycine pH 3.0 and different ionic solutions, were tested without finding any

solution that regenerated the surface sufficiently. David Myszka, in cooperation with Chris

Hampe, has previously observed difficulties to regenerate the surface by using acidic or basic

regeneration solutions after interaction with GAD65 (unpublished report, partially published

in (Raju et al. 2005)). The problems with the low binding of antibodies to GAD65 and the

altered binding after regeneration, suggest that GAD65 is a pH-sensitive and unstable protein.

Consequently, since the GAD65 protein seemed to be problematic in this kind of assay, we

did not continue the development of the assay.

We did not develop an SPR-based method for detection of IA-2A, mainly due to difficulties

to obtain the IA-2 protein and antibodies towards IA-2 to test the assay.


68

Development of the SPR-based immunoassay for IAA

Pre-concentration of human recombinant insulin to the dextran matrix was optimized. Then,

insulin (100 μg/ml) at pH 5.0 was immobilized via amine coupling to a CM5 chip. This

procedure was started by activation of the carboxyl groups at the sensor chip with 0.2 M N-

ethyl-N-(3-diethylaminopropyl) carbodiimide (EDC) and 0.05 M N-hydroxysuccinimide

(NHS) during 7 minutes (Figure 15). Thereafter, insulin was injected during 7 minutes for

immobilization, followed by blocking of excess activated carboxyl groups with ethanolamine

hydrochloride (1 M, pH 8.5) for 7 min. The surface was finally washed twice with glycine-

HCl (10 mM, pH 1.8) for 1 min. The level of immobilized insulin varied to some extent

between different sensor chips, but stayed at 1300 ± 200 RU. This was the highest level

possible to immobilize, and could not be increased by injection of higher insulin

concentrations or by increasing the time that insulin was allowed to pass the surface.

Figure 15. Immobilization of insulin followed by regeneration and binding of a monoclonal anti-

human insulin antibody. Injection of EDC/NHS (1), insulin (2), ethanolamine (3), glycine-HCl (4) and

finally a monoclonal anti-human insulin antibody (5) are shown in the sensorgram.

To assure that binding to the immobilized insulin worked properly, a standard curve, with a

monoclonal anti-human insulin antibody (Art. No. I7660-14, Biosite, recognizing an epitope

including amino acid B30) diluted in running buffer, was run once (Figure 16A). In following

experiments, new surfaces immobilized with insulin were checked for activity with the same

insulin antibody at a concentration of 20 µg/ml.

1

2

5

3

4

Time

s


69

Figure 16. The anti-human insulin antibody concentrations injected were 10000, 5000, 2500, 1250,

625, 313, 156 and 78 ng/ml, resulting in responses of 1352, 630, 290, 137, 66, 31, 14 and 6 RU,

respectively, as shown in the standard curve (A). In panel B, the anti-human insulin antibody at

concentrations of 10000, 5000, 2500, 1250, 625 and 313 ng/ml were diluted in an IAA-negative serum

instead of buffer. The responses were 948, 361, 163, 102, 41 and 4 RU, after subtraction of the

response for the serum alone (169 RU), and were presented in panel B (black squares) together with

the standard curve from panel A (black circles).

Efforts to reduce non-specific binding of serum proteins to the sensor surface

The next step in the assay development was to introduce serum to the immobilized surface.

Non-specific binding of serum components to the surface of the sensor chip often gives a

large SPR signal, which masks the analyte response. This is especially evident when

measuring molecules of low concentration in serum, which is the case for IAA. We observed

high non-specific binding of serum proteins during initial experiments. It has been reported

that when the concentration of a molecule is high enough to be detected after diluting serum

1/20, the problem with non-specific binding of serum matrix proteins can be overcome

(Karlsson et al. 1993). However, when detecting IAA in human serum, the generally very low

concentrations do not allow the use of diluted samples in the assays.

Other ways to overcome the problem with non-specific binding of serum proteins are to use a

proper design of the surface chemistry on the sensor chip and/or to introduce additives to the

analyte solution. Different non-sticky surface coatings, based on dextran (Löfås et al. 1990) or

PEG (Herrwerth et al. 2003; Gobi et al. 2007; Larsson et al. 2007; Shankaran et al. 2007),

have been developed. For the development of our SPR-based immunoassay we used the CM5

chip (Figure 11). A CM4 chip, with low carboxylation, is sometimes used to reduce non-

A.

050

0

1000

1500

2000

2500

0

50

100

150

200

250

2600

1000

0

3001400

Insulin ab (ng/ml)

Resp

on

se (

RU

)B.

0

2000

4000

6000

8000

1000

00

500

1000

1500

Insulin ab (ng/ml)

Resp

on

se (

RU

)


70

specific binding (Raju et al. 2005) (www.biacore.com). However, since assays for

determination of the concentration of an analyte generally require a high surface

concentration of the immobilized molecule, this chip with fewer immobilization sites was not

a suitable choice for our assay. Addition of CM dextran to a serum sample has shown to

reduce the non-specific binding to CM-coated surfaces with up to 75% via a competition

effect (Karlsson et al. 1993). We therefore chose to use CM dextran and NaCl for reduction of

non-specific binding. When the concentrations of these additives were optimized (1 mg/ml

CM dextran and 0.35 M NaCl) an evident reduction of the non-specific binding signal was

observed.

After reduction of the non-specific binding it was possible to detect the insulin antibodies in

serum. A standard curve, with different concentrations of the monoclonal anti-human insulin

antibody diluted in an IAA-negative serum sample, was run. The shape of the curve was

similar, but with lower responses, than the standard curve obtained with different

concentrations of insulin antibody diluted in buffer (Figure 16B).

Detection of IAA in serum using the 3-day and 5-min methods

Since the non-specific binding fluctuated among sera from different individuals, it was

difficult to measure IAA directly in serum. To overcome this problem, an indirect competitive

immunoassay was developed. In this method, each serum sample was divided in two equal

aliquots; one was mixed with buffer and the other with an excess amount of insulin, similarly

to the procedure for detection of IAA by RIA. In RIA, the following step is incubation of the

serum solutions for 72 hours. Since one of our objectives was to develop a method which was

not as time consuming as RIA, we aimed to reduce the time required for analysis of one

sample. Thus, the serum mixtures were incubated under agitation either for 72 hours at 4ºC

(according to the 3-day method) or for 5 minutes at room temperature (according to the 5-min

method). Thereafter, each serum mixture was diluted with CM dextran and NaCl for reduction

of non-specific binding, and then injected over the sensor surface immobilized with insulin.

When the portion of serum mixed with buffer was introduced to the surface, IAA in the

sample was free to bind the immobilised insulin on the surface (Figure 17, lower left). In

contrast, when serum mixed with insulin was injected, competition for binding of IAA

between insulin in solution and insulin immobilised on the surface took place. However, due

to the large excess of insulin molecules in solution, very little binding of IAA to the surface-


71

bound insulin would occur. The resonance signal obtained from this sample is instead a

measure of non-specific binding to the immobilized insulin and to the dextran matrix.

The levels of IAA interacting with insulin were too low to be directly quantified. To enhance

the response signal, a secondary polyclonal anti-human IgG antibody (50 µg/ml in HBS-EP)

was injected immediately after the serum injection was completed. After a short regeneration-

pulse with glycine-HCl at pH 1.8 the surface was ready to be used for a new serum injection.

Regeneration of the surface removed both specifically and non-specifically bound serum

components from the surface. The regeneration was very reliable, providing a stable baseline,

and a chip could be used for numerous (at least 50) repeated cycles without any visible

decrease in response.

To calculate the level of IAA in a serum sample (R), the secondary antibody response from

serum mixed with buffer (Ba-Bb) was subtracted from the response obtained when the serum

was mixed with excess insulin (Ia-Ib) (Figure 17), according to the formula:

)()( baba IIBBR ,

where Ba = response for sample incubated with buffer after secondary antibody, Bb = before

secondary antibody, Ia = response for sample incubated with excess insulin after secondary

antibody and Ib = before secondary antibody.

To assess if we were able to detect different levels of IAA in serum, we run a standard curve,

using both the 3-day and the 5-min method. A pooled high IAA-positive serum was diluted to

different concentrations of IAA (0-40 U/ml) in pooled serum from IAA-negative non-diabetic

individuals. This was done to yield a range of positive samples with almost identical serum

matrix composition. The SPR-run showed that the responses increased with increasing IAA

concentrations, both for the 3-day (Figure 18A) and the 5-min method (Figure 18B).


72

Figure 17. The principle of IAA detection with the indirect competitive immunoassay based on SPR

(lower part of the figure) as well as a full response cycle for a serum sample mixed with buffer (upper

left) and with insulin (upper right) is shown. In the figure, the different steps during the cycle are

illustrated as follow: 1 = start of injection of serum mixed with buffer, 2 = end of serum injection, 3 =

start of secondary antibody injection, 4 = end of secondary antibody injection, 5 = start of

regeneration, 6 = end of regeneration, 7 = start of injection of serum mixed with insulin, 8 = end of

serum injection, 9 = start of secondary antibody injection, 10 = end of secondary antibody injection,

11 = start of regeneration and 12 = end of regeneration. The difference in secondary antibody response

from a serum sample incubated in buffer (Ba-Bb) and insulin (Ia-Ib), respectively, gives a measure of

the level of IAA in the sample.

100 200


73

Figure 18. Standard curves from (A) the 3-day and (B) the 5-min methods. A pooled high IAA

positive serum was diluted to concentrations of 0, 4, 6, 10, 20 and 40 U/ml for the 3-day method,

resulting in SPR responses of -4, 8, 15, 17, 20 and 29 RU. For the 5-min method, the concentrations of

the pooled high IAA positive serum were 0, 4, 6, 7.5, 10, 20 and 40 U/ml, giving SPR responses of 45,

44, 47, 54, 61, 67 and 80 RU. For both methods, the IAA positive serum was diluted in pooled serum

from IAA negative non-diabetic individuals. U/ml is an arbitrary unit from the reference RIA method.

Since the non-specific binding of serum proteins to the surface is very different among

different individuals, we wanted to evaluate if a similar standard curve could be obtained

when using sera from 7 non-diabetic individuals instead of one pooled serum as background

for the standard curve. Therefore, another set of standard curves was run, with both the 3-day

and the 5-min method. Each IAA-containing sample was diluted in a single IAA-negative

serum, and different single IAA-negative serum samples were used for the different

concentrations of the IAA-positive sample. Again, the SPR responses increased with

increasing IAA concentrations (Figure 19 and 20), confirming that the SPR-based indirect

competitive immunoassay was suitable for the measurement of IAA in serum from different

individuals. To determine variations in the method, we run the standard curve with the 5-min

method three times (Figure 20). Reproducibility was found to be excellent, with low variation

between the runs (Table 4).

A

5 10 15 20 25 30 35 40 45

-10

-5

0

5

10

15

20

25

30

35

40

U/ml

RU

B

0 5 10 15 20 25 30 35 40 4540

45

50

55

60

65

70

75

80

85

U/ml

RU


74

Figure 19. Standard curve from the 3-day method. Different concentrations of a pooled high IAA-

positive sample (0, 4, 6.6, 10, 20 and 40 U/ml) were diluted in different single IAA negative serum

samples from non-diabetic individuals, resulting in SPR responses of -8, 4.9, 15.9, 23.2, 37.6 and 45.2

RU.

Figure 20. Standard curve from the 5-min method, based on three repeated runs. A high IAA-positive

sample was diluted to different concentrations in different single serum samples from IAA-negative

non-diabetic individuals. The values of the mean responses from the SPR run for these samples are

shown below in Table 4. Sample A, B, C and D are non-pooled sera from type 1 diabetic patients (A =

32.5 U/ml, B = 24 U/ml, C = 11.3 U/ml and D = 5.2 U/ml).

5 10 15 20 25 30 35 40 45

-20

-10

0

10

20

30

40

50

60

U/ml

RU


75

Table 4. Test of reproducibility of a standard curve from the 5-min method, with different

concentrations of an IAA-positive sample diluted in different single serum samples from IAA-negative

non-diabetic individuals. The corresponding SPR responses for the three runs are shown, as well as

mean and CV (coefficient of variation).

High-affinity IAA

It has been shown that IAA of high affinity is related to progression to type 1 diabetes

(Achenbach et al. 2004a; Schlosser et al. 2005b). Thus, methods for detection of IAA should

be designed to measure high-affinity antibodies. We were therefore interested to find out if

the IAA measured with the SPR-based immunoassay were of high affinity. The sites for

antigen immobilization in an immunoassay might affect the nature of the antibodies being

captured. In our experiments, possible coupling sites on the insulin molecule to the dextran

matrix are the N-termini of the A and B chains and the lysine residue at position B29, since

these positions contain amino groups. Thus, the A8-A13 region, which is important for

binding of high-affinity IAA (Achenbach et al. 2004a), should be available for binding after

immobilization of the insulin molecule. This speaks in favor of that the observed response in

the SPR-based assays can be originated from high-affinity IAA. Other important arguments

supporting this idea are that the same conditions as in the incubation step in RIA were used in

our 3-day method, and that all the IAA-positive serum samples measured in the SPR-based

system were from newly diagnosed type 1 diabetic patients, and consequently should be

disease-specific and thereby IAA of high affinity.

The 5-min method compared to RIA

We evaluated differences between the responses detected by the 3-day and the 5-min

methods. The methods showed to perform similarly, suggesting that the 5-min method

probably also detect high-affinity IAA. Thus, we believe the 5-min method is suitable for

SPR-based detection of IAA in serum. In RIA, the mixture of IAA-containing serum and

IAA (U/ml) Mean CV (%)

0 46.7 46.5 44.7 46.0 2.4

2 48.9 35.3 37.6 40.6 17.9

4 41.1 39.9 38.4 39.8 3.4

6 64.7 61.7 61.4 62.6 2.9

8 84.7 85.8 83.5 84.7 1.4

10 86.0 88.8 90.0 88.3 2.3

20 92.0 87.8 91.8 90.5 2.6

40 136.0 132.6 135.3 134.6 1.3

Response (RU)


76

labelled insulin is incubated for 3 days, capturing the high-affinity antibodies (Wilkin 1990).

In the Biacore system, the continuous flow enabling repeated associations and dissociations

between antigens and antibodies, might explain the functionality of the 5-min method,

although the incubation time is shortened to only 5 minutes.

By using a calculated cut-off value for the 5-min method, the serum samples were classified

as IAA-positive or -negative in accordance with the results obtained from the measurement

with RIA (Figure 20). To determine the cut-off level for the 5-min method, a number of IAA-

negative samples from non-diabetic individuals as well as low IAA-positive samples from

newly diagnosed type 1 diabetic patients were measured with the SPR method. The cut-off for

positivity was set to 6 U/ml, by calculating the mean value of the low IAA positive samples

subtracted by 3 SD. None of the IAA-negative samples exceeded this value, indicating that

this cut-off value was suitable for our assay. Since the RU value for the same sample might

differ to some extent between different runs, depending on the amount of insulin immobilized

on the surface, the cut-off value cannot be expressed as RU. In RIA, the cut-off for positivity

for IAA was 6.45 U/ml, corresponding to the 98th

percentile of 114 healthy Swedish children

(Holmberg et al. 2006).

The similar cut-off levels, i.e. 6 and 6.45 U/ml, for the SPR-based method and RIA

respectively, indicate that the sensitivity for detection of IAA with both methods was similar.

However, the IAA levels differed to some extent between the 5-min indirect competitive

immunoassay and the RIA. A possible explanation might be that with the SPR-based assay

we probably detected IAA of the IgG3 subclass, which cannot be detected in RIA using

protein A sepharose for precipitation. In a study by Hoppu et al, IAA of IgG1 and IgG3

subclasses were predominant in genetically susceptible young children who rapidly

progressed to clinical type 1 diabetes, while a weak or absent IgG3 response was associated

with relative protection from disease (Hoppu et al. 2004c).

The SPR-based immunoassay for detection of IAA was time-saving compared to RIA. The

RIA procedure takes 4 days to perform, and 15 samples can be run simultaneously in one

plate. The same number of samples in the SPR-based immunoassay takes only approximately

13 hours to run, since the time to run one sample in this system is about 50 minutes and the

samples are run after each other. One disadvantage with the SPR-based immunoassay

compared to RIA is the larger volume of serum needed for the assay; 100 µl compared to only


77

30 µl in RIA. However, the advantages of the 5-min method, with no need for labelling of

interactants and that it is time-saving compared to RIA, speak in favor for the use of the SPR-

based assay when sample volume is not a limiting issue.

Type 1 diabetes-related analytes measured with SPR: Results from other

studies

SPR has been used in several studies to measure biomolecules in diluted serum (Vikinge et al.

1998; Gomara et al. 2000; Jongerius-Gortemaker et al. 2002) and some have detected

biomolecules in undiluted serum (Masson et al. 2007; Phillips et al. 2007). A number of

studies have measured type 1 diabetes-related analytes with SPR. For example, Ayela et al

detected a polyclonal antibody against a synthetic peptide from the juxtamembrane region of

IA-2, in diluted serum (1/100), using mixed self-assembled monolayers (SAM) on a gold

surface as the sensor surface (Ayela et al. 2007). Lee et al measured monoclonal anti-GAD65

antibodies diluted in buffer, using an alkanethiol SAM sensor surface, by immobilizing

streptavidin on the surface followed by injection of biotin-GAD (Lee et al. 2005). Another

study focused on measurement of insulin, both in buffer and in diluted serum (1/4 and 1/10),

using a PEG-containing SAM sensor surface (Gobi et al. 2007). They used an indirect

competitive immunoreaction principle, in which a monoclonal anti-insulin antibody was

mixed with different concentrations of insulin prior to injection into the SPR instrument.

Thereafter, the insulin concentration was indirectly measured by detecting the binding of the

anti-insulin antibody to the sensor surface immobilized with insulin. In another study, the

level and affinity of insulin antibodies (IA) in diabetic patients after receiving exogenous

insulin were measured with SPR, using a CM5 chip (Kure et al. 2005). Since the levels of IA

in diabetic patients after receiving insulin are generally much higher than the levels of IAA

commonly observed in newly diagnosed type 1 diabetic patients and non-diabetic individuals,

their assay did not detect low concentrations of antibodies, as we aimed to do. Also, due to the

relatively high concentrations of IA, they were able to overcome the problems with non-

specific binding of serum proteins to the dextran matrix by diluting the serum. Our SPR-based

indirect competitive immunoassay for IAA is, to our knowledge, the only published report on

IAA quantification in serum samples using SPR.


78

Further and future experiments with the 5-min method

After publication of the methodological paper with the SPR-based indirect competitive

immunoassay for detection of IAA (Paper II), further experiments were performed. To

evaluate the capability of the method for large scale screening of serum samples, 86 serum

samples were analysed. The method detected the majority of the IAA-positive samples,

previously classified as positive by RIA, and also most of the IAA-negative samples were

correctly classified. However, we observed that some serum samples showed aberrant high

responses, higher than what is commonly observed in positive samples with high IAA levels.

To date, we have no explanation for this result, but it was interesting to observe that some of

the sera were collected from diabetic patients while others came from non-diabetic

individuals. Further experiments are needed to clarify the reason for the aberrant high

responses.

In conclusion, we have developed an SPR-based indirect competitive immunoassay for

detection of IAA in serum. The main advantages with this assay are that there is no need for

labelling of interactants and it is time-saving compared to RIA, but further development of the

method is needed.

EFFECTS OF THE GAD-ALUM TREATMENT ON THE HUMORAL

IMMUNE RESPONSE (PAPER III AND IV)

Increased levels of GADA

In the phase II clinical trial performed in children and adolescents with type 1 diabetes,

treatment with GAD-alum resulted in preservation of residual insulin secretion in children

with short duration of the disease, shown as a declined decrease in C-peptide levels. A

number of effects on the immune system have been observed after this treatment (Axelsson et

al. 2010; Hjorth et al. 2010), including increased levels of GADA (Ludvigsson et al. 2008)

(Figure 21). The GADA levels started to increase already after the first injection of GAD-

alum, and peaked at 3 months, after the booster dose. Further, still 30 months after the first

injection the levels were higher compared to the placebo group.


79

Figure 21. Median levels of GADA for the GAD-alum group (black circles) and the placebo group

(white circles) at baseline and at 1, 3, 9, 15, 21 and 30 months after initial injection. GADA titers

increased after injections of GAD-alum, and at month 3, the titers were significantly higher in the

GAD-alum compared to the placebo group. This difference persisted throughout the 30 months study

period.

In a GAD-alum dose-finding study in LADA patients, GADA levels did not increase after

injection of 20 μg of GAD-alum, the dose that resulted in clinical effect (Agardh et al. 2005)

and was used in our study. However, in the same study, increased GADA levels were

observed in patients treated with 500 μg of GAD-alum, and were accompanied by a

temporary increase in GADA-binding to one epitope specificity often related to type 1

diabetes (Bekris et al. 2007). We were therefore interested to study whether the epitope

binding pattern of GADA was affected by the GAD-alum treatment in type 1 diabetic

patients.

Temporary increase in binding to one rFab-defined epitope

To investigate if the epitope specificity of GADA was affected by the GAD-alum treatment,

we used six rFab (b96.11, DPA, DPD, MICA3, b78 and N-GAD65 mAb) recognizing

different epitope regions on GAD65 (Paper III). The b96.11-defined epitope was generally

0

1000

2000

3000

4000

P=0.001

N.S

N.S

P=0.01

P=0.04P=0.01

P=0.03

0 1 3 9 15 21 30

Month

GA

DA

(U

/ml)

↑ ↑Injections


80

predominant and observed in the majority of the tested sera of both GAD-alum and placebo

treated children (Figure 22). This was expected, since this epitope specificity has previously

been shown to be predominant in type 1 diabetic patients (Padoa et al. 2003). Further, binding

to the DPD- and MICA3-defined epitopes was also observed in our study, previously found in

other studies with type 1 diabetic individuals (Richter et al. 1993; Padoa et al. 2003).

( )


81

Figure 22. Binding to GAD65 in the presence of rFab b96.11 (A), DPD (B), MICA-3 (C), DPA (D),

b78 (E) and N-GAD65 mAb (F) at baseline, 3 months and 15 months for GAD-alum and placebo

groups, presented as ratio of competed / non-competed. A higher binding to GAD65 in the presence of

an rFab indicates a lower proportion of GADA binding to the respective epitope. Ratios above 1.4 are

placed within brackets outside the range of the y axis.

Overall, there were no longitudinal changes in binding to any of the epitopes, including

b96.11, over the 15 months study period (Table 5). Binding levels to the DPA-, N-GAD65

mAb- and b78-defined epitopes were very low throughout the study (Figure 22) and were

therefore excluded from the statistical analyses.

Table 5. The variability over time (baseline to 15 months) was investigated by calculation of the sum

of absolute differences between consecutive measurements, and compared for the GAD-alum and

placebo groups.

Epitope Sum of absolute differences (min, max) p-value

GAD-alum Placebo

b96.11 40.0 (3.6, 108.2) 30.1 (3.9, 78.9) 0.07

DPD 37.3 (17.3, 76.6) 46.3 (13.9, 92.8) 0.48

MICA3 42.1 (16.3, 108.7) 40.8 (11.1, 90.4) 0.75

Next, as GADA titers peaked at 3 months, we were interested to study whether this increase

in GADA titer was accompanied by any change in GADA epitope recognition between two

visits. For this analysis, binding of GADA to rFab-defined epitopes was compared between

samples collected at two time points (baseline to 1 month, baseline to 3 months, and 3 months

( ) ( ) ( ) ( )


82

to 15 months). Our results showed that binding to the b96.11-defined epitope changed

temporarily, with increased binding between baseline and 3 months in the GAD-alum treated

group compared to the placebo group (Table 6). This effect was especially evident in the

group of GAD-alum treated patients who experienced an increase of more than 100% in their

GADA titer from baseline to 3 months. The increase in binding to the b96.11-defined epitope

was followed by decreased binding between 3 and 15 months, resulting in similar levels at 15

months as the ones observed at baseline. Similar results were observed in the GAD-alum

dose-finding study on LADA patients, in which treatment with the highest dose of GAD-alum

(500 µg) was accompanied by increased GADA levels (Agardh et al. 2005), and temporarily

increased binding to the b96.11-defined epitope (Bekris et al. 2007). The changes in b96.11

epitope recognition after treatment with 20 µg GAD-alum found in our study, was not seen in

the group of LADA patients receiving the same dose of GAD-alum resulting in clinical effect.

This might be due to the small number of individuals in the LADA study, or alternatively,

changes in b96.11 epitope recognition might be explained by the increased GADA levels in

the type 1 diabetic patients, which was not observed in LADA patients after the same dose of

GAD-alum.

Higher GADA titers have previously been associated with broader epitope reactivity (Gilliam

et al. 2004; Ronkainen et al. 2004). In our study, we found no correlation between GADA

titer and epitope binding at 3 and 15 months in the GAD-alum group, but in the placebo group

binding to the b96.11- and DPD-defined epitopes correlated with GADA titer (Table 7). We

cannot explain the lack of correlation in the GAD-alum group after treatment. The treatment

induced very high GADA titers, often much higher than the ones detected in the majority of

GADA positive diabetic individuals. We could speculate that the binding to the b96.11-

defined epitope did not increase in the same proportion as the induced GADA. It could also be

possible that some of the induced GADA bind to other epitopes on GAD65 than those tested

in our study. The temporary increased binding to the b96.11-defined epitope could be due to

the activation and further upregulation of the already existing memory B cell responses.

Although the increase in binding to the b96.11-defined epitope was only temporary, the

GADA levels remained higher in the GAD-alum group compared to placebo even 30 months

after treatment. It cannot be excluded that as a part of the immunomodulatory effect of the

treatment, GAD-alum might have induced memory B cells producing GADA with other

epitope specificities than those tested in our study. This is an interesting hypothesis to be

confirmed in future studies.


83

Table 6. Differences in epitope binding between GAD-alum and placebo patients, from baseline to 1

month, baseline to 3 months, 3 months to 15 months and baseline to 15 months. Changes in median

binding to GAD65 in the presence of rFab b96.11, DPD and MICA3 between two time points were

calculated and the differences in median response were compared between the GAD-alum and placebo

groups. This was also done for individuals in the GAD-alum group who experienced an increase in

GADA titer of more than 100% between baseline and 3 months (high GADA increasers, HGI)

compared to individuals in the same group without this increase (non-HGI).

Epitope Months compared Difference in median response % (min, max) p-value

GAD-alum Placebo

0-1 -2.8 (-54.4, 52.6) 0.5 (-28.4, 28.2) 0.45

b96.11 0-3 -8.1 (-72.4, 39.6) 1.5 (-28.3, 28.6) 0.02 *

3-15 8.3 (-17.1, 36.7) -2.4 (-32.8, 30.1) <0.05 *

0-15 -5.6 (-62.5, 62.4) 1.6 (-25.6, 35.7) 0.24

0-1 0.9 (-24.1, 23.3) 3.2 (-15.3, 32.5) 0.59

DPD 0-3 -1.8 (-43.6, 41.5) 2.4 (-28.6, 32.7) 0.87

3-15 3.2 (-15.9, 24.3) 2.6 (-37.4, 26.7) 0.66

0-15 -2.5 (-42.3, 42.1) -0.1 (-49.8, 37.3) 0.85

0-1 5.6 (-21.8, 51.7) 1.2 (-9.5, 49.4) 0.30

MICA3 0-3 3.1 (-31.1, 56.8) 0.1 (-36.6, 25.4) 0.18

3-15 5.3 (-34.3, 20.7) -0.9 (-42.1, 36.1) 0.20

0-15 0.1 (-33.9, 39.2) -0.1 (-36.2, 25.3) 0.89

GAD-alum HGI GAD-alum non-HGI

b96.11 0-3 -10.8 (-72.4, 30.5) 3.1 (-7.9, 39.6) 0.04 *


84

Table 7. Correlations between GADA titer and binding of GADA to GAD65 in the presence of rFab

b96.11 or DPD for all individuals, and for the GAD-alum and placebo groups, at baseline, 3 months

and 15 months after the first injection.

Epitope Time point Individuals Correlation with GADA titer

p-value r

b96.11 baseline all individuals 0.001 ** -0.423

GAD-alum 0.021 * -0.400

placebo 0.073 -0.350

3 months all individuals 0.018 * -0.305

GAD-alum 0.688 -0.073

placebo 0.005 ** -0.524


GAD-alum 0.283 -0.193

placebo 0.019 * -0.448

DPD baseline all individuals 0.003 ** -0.377

GAD-alum 0.011 * -0.436

placebo 0.088 -0.335

3 months all individuals 0.075 -0.232

GAD-alum 0.840 0.036

placebo 0.004 ** -0.538


GAD-alum 0.254 -0.205

placebo 0.048 * -0.385

Changes in IgG subclasses of GADA

Since the GAD-alum treatment affected GADA levels and temporarily increased binding to

the b96.11 defined epitope, we were interested to investigate if there were any further effects

of the GAD-alum treatment on GADA. In humans, Th1 cells and their related cytokines seem

to stimulate the production of IgG1 while Th2 cells stimulate the production of IgG4 and IgE

(Gascan et al. 1991; Couper et al. 1998; Bonifacio et al. 1999; Hawa et al. 2000). Based on

results from animal studies, a Th2 profile has been suggested to be protective against the

aggressive damage inflicted on the insulin-producing cells (Cameron et al. 1997; Petrovsky et

al. 2003). The GAD-alum treatment in type 1 diabetic children induced a Th2 response, with

increased levels of IL-5 and IL-13 already 1 month after the first injection (Axelsson et al.

2010). A broad range of cytokines were induced at 3 and 9 months, but only the two Th2

associated cytokines IL-5 and IL-13 together with FOXP3 increased continuously over time.


85

Therefore, we were interested to study the IgG subclass distribution of GADA. Samples from

patients in the GAD-alum and placebo groups were investigated at baseline and at 3 and 15

months after initial treatment (Paper IV). Frequencies of the different IgG subclasses were

calculated with respect to total IgG in each sample at every time point. IgG1 was the

predominant IgG subclass of GADA, both in GAD-alum and placebo individuals (Table 8).

This is in accordance with previous studies showing a predominance of IgG1 both in

prediabetic individuals and in newly diagnosed type 1 diabetic individuals (Bonifacio et al.

1999; Hoppu et al. 2004a; Ronkainen et al. 2006).

At baseline, the percentages of IgG1, 2, 3 and 4 were similar in both groups. Three months

after initial injection of GAD-alum a reduced percentage of IgG1 and increased IgG3 and

IgG4 were detected in GAD-alum treated individuals (Figure 23). Emergence of the IgG4

subclass of GADA has been observed in genetically susceptible children who remained non-

diabetic, while children who developed type 1 diabetes often had a broad initial response of

IgG subclasses and rarely developed IgG4 before onset of the disease (Ronkainen et al. 2006).

GADA of the IgG4 subclass has been found to appear more often in LADA patients than in

type 1 diabetic patients (Hillman et al. 2009), indicating that IgG4 might be associated with

slower progression to clinical diabetes, perhaps due to a more balanced immune response in

the pancreatic tissue (Couper et al. 1998; Hillman et al. 2004). Thus, an increased percentage

of IgG4 and decreased IgG1 after GAD-alum treatment might further support the suggestion

that a predominant protective Th2 response was induced as a result of the treatment. While

IgG1 has been associated with Th1 and IgG4 with Th2 responses (Gascan et al. 1991; Couper

et al. 1998; Bonifacio et al. 1999; Hawa et al. 2000; Hillman et al. 2004), the relation to

specific T cell subsets of cytokines for IgG2 and IgG3 in humans is not well-defined. There

are some studies supporting that IgG3 might be associated with Th1 responses (Hussain et al.

1995; Widhe et al. 1998).


86

Ig

G1

Ig

G2

Ig

G3

Ig

G4

G

AD

-alu

m

Pla

ceb

o

GA

D-a

lum

P

lace

bo

G

AD

-alu

m

Pla

ceb

o

GA

D-a

lum

P

lace

bo

Ba

seli

ne

33

94

(1

10

-12

199

)

35

/35

37

54

(9

7-1

15

15)

34

/34

50

(4

-240

)

25

/35

35

(4

-291

)

24

/34

80

(1

-735

)

31

/35

90

(3

-271

4)

29

/34

47

(3

-688

)

22

/35

65

(1

-109

9)

24

/34

3 m

on

ths

74

20

(7

60

-12

569

)

35

/35

30

24

(1

19

-10

811

)

34

/34

96

(18

-93

7)

31

/35

47

(3

-252

)

24

/34

42

5 (

15

-40

86

)

32

/35

87

(1

-209

4)

31

/34

23

1 (

5-3

01

7)

29

/35

67

(6

-668

)

27

/34

15

mo

nth

s 6

06

3 (

265

-11

311

)

35

/35

34

68

(6

5-9

79

5)

34

/34

61

(5

-164

6)

29

/35

40

(5

-696

)

21

/34

23

7 (

13

-21

86

)

30

/35

10

9 (

7-1

01

6)

29

/34

10

3 (

6-2

76

2)

28

/35

57

(7

-542

)

27

/34

Ta

ble

8.

GA

DA

su

bcl

ass

dis

trib

uti

on

. T

he

med

ian

lev

els

and

ran

ge

of

IgG

are

rep

rese

nte

d in

cp

m.

Th

e n

um

ber

s o

f in

div

idu

als

wit

h d

etec

tab

le le

vel

s ar

e sh

ow

n f

or ea

ch t

ime

po

int.

To

det

erm

ine

po

siti

vit

y f

or

the

sam

ple

s, t

he

mea

n c

pm

val

ue

for

a G

AD

A

neg

ativ

e co

ntr

ol

and

3x

SD

of

the

sam

e co

ntr

ol

was

su

btr

acte

d f

rom

eac

h s

amp

le.

All v

alu

es a

bo

ve

0 a

fter

th

is c

alcu

lati

on

wer

e

reg

ard

ed a

s p

osi

tiv

e.


87

At 15 months, the distribution of IgG subclasses in the samples from GAD-alum individuals

had returned to that observed at baseline, and no differences in distribution between the GAD-

alum and placebo groups were observed. It will be interesting to see if this transient effect of

GAD-alum on IgG subclass distribution is prolonged by treatment with more than two GAD-

alum injections in the ongoing phase III GAD-alum trial.

A.

B.

% o

f to

tal

IgG

GA

DA

Baseline 3 m 15 m

IgG 1

80

85

90

95

100

p= 0.005 n.sn.s

IgG 2

0

1

2

3

4

n.sn.s

n.s

IgG 3

0

2

4

6 p= 0.01

n.sn.s

IgG 4

0

1

2

3

4

p= 0.048

n.s

n.s

Baseline 3 m 15 m

Baseline 3 m 15 m

Baseline 3 m 15 m

% o

f to

tal

IgG

GA

DA

% o

f to

tal

IgG

GA

DA

% o

f to

tal

IgG

GA

DA

GAD-alum

IgG 1 IgG 2 IgG 3 IgG 40

2

4

6

8

1080

90

100

Baseline

3 month

15 month

ns

p<0.001 p=0.001p<0.001 p=0.001

ns

ns p=0.039

ns

p=0.002 p=0.003

ns

p=0.017 p=0.04

% o

f to

tal Ig

G


88

C.

Figure 23. The effect of GAD-alum in the GADA subclass distribution. Differences in IgG1, 2, 3 and

4 subclasses between GAD-alum (black circles) and placebo (white circles) (panel A), and within the

groups (panels B and C). The results are presented as median percentage of total IgG GADA. Total

IgG was calculated with respect to the combined sum of all four subclasses in each sample.

No changes in SPS-related epitope binding or inhibition of GAD65

enzymatic activity

There has been concern that induction of high GADA levels by GAD-alum treatment might

cause adverse events, especially neurological symptoms. However, no treatment-related

adverse events have been observed during the trial (Ludvigsson et al. 2008). High GADA

levels leading to neurological symptoms have been associated with a decreased production of

GABA in the nervous system of SPS patients (Murinson 2004; Murinson et al. 2004; Raju et

al. 2008). GADA in SPS patients inhibits the enzymatic activity of GAD65, which is not the

case for GADA in type 1 diabetic patients (Raju et al. 2005). We therefore investigated the

capacity of serum from patients in the GAD-alum trial to inhibit the enzyme activity of

GAD65 in vitro (Paper IV). The results showed that serum from GAD-alum treated

individuals did not affect the enzymatic activity of GAD65, at baseline, 3 or 15 months after

treatment (Figure 24).

Placebo

IgG 1 IgG 2 IgG 3 IgG 40

2

4

6

8

1080

90

100 Baseline

3 month

15 month

ns

ns ns

ns

ns ns

ns

ns nsns

ns ns

% o

f to

tal Ig

G


89

Figure 24. GAD65 enzymatic activity assay for serum from patients treated with GAD-alum (black

circles) and placebo (white circles). GAD65 enzyme activity is reported as radioactivity (cpm) from

14CO2 released from 14C-glutamic acid measured at baseline and after 3 and 15 months. A control

sample from one SPS patient was included in each run (triangles).

Further, GADA in SPS patients recognizes mainly linear epitopes while GADA in type 1

diabetic patients generally recognizes conformational epitopes (Kim et al. 1994; Li et al.

1994; Daw et al. 1996; Raju et al. 2005). As a part of the epitope binding study (Paper III), we

therefore investigated if the treatment affected the epitope specificity of GADA regarding two

SPS associated epitopes (by using rFab b78 and N-GAD65 mAb). Sustained low binding to

the b78- and N-GAD65 mAb-defined epitopes throughout the study indicates that these

antibody specificities did not contribute to the increased GADA levels after GAD-alum

treatment. Thus, the treatment did not affect either the GAD65 enzyme activity or the binding

of GADA to SPS associated epitopes, further strengthening the safety of the treatment.

GAD-alum

Bas

elin

e

3 m

onths

15 m

onths

SPS

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

5500

6000

6500

7000

7500

8000

8500

9000

9500

10000

ns ns

ns

p<0.001

p<0.001

p<0.001

cp

m

Placebo

Bas

elin

e

3 m

onths

15 m

onths

SPS

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

5500

6000

6500

7000

7500

8000

8500

9000

9500

10000ns ns

ns

p<0.001

p<0.001

p<0.001

cp

m


90

Antigen-specific effect

Another result that speaks in favor of the safety of the GAD-alum treatment was that the

immunomodulatory effect of GAD-alum was antigen-specific (Paper IV). The levels of IA-

2A, tetanus toxoid and IgE antibodies did not change after treatment, neither in the GAD-

alum nor in the placebo group (Figure 25), indicating a GAD65 specific effect on the immune

system. At all time points the levels of IA-2A were higher in the placebo group than in the

GAD-alum group, but there were no changes over time. In addition, levels of IA-2A did not

correlate to C-peptide levels, GADA levels or responsiveness of the treatment.

A.

B.

Baseline Baseline 3 months 3 months 9 months 9 months0

500

1000

1500

20002000

4000

6000

8000

GAD-alum

Placebo

IA-2

A t

ite

r (U

/ml)

p< o.o5

p< o.o5

p< o.o5

Baseline Baseline 3 months 3 months 9 months 9 months0.0

0.5

1.0

1.5

2.0GAD-alum

Placebo

Te

tan

us

tit

er

(IU

/ml)

n.s


91

C.

Figure 25. Levels of IA-2A (A), tetanus toxoid antibodies (B) and IgE antibodies (C) before and after

treatment with GAD-alum. IA-2A and tetanus toxoid antibodies were analyzed at baseline, 3 and 9

months. IgE was analyzed at baseline and at 21 months. Antibody levels for each patient in the GAD-

alum (black circles) and placebo (white circles) groups are shown as dot plots.

The role of GADA in the effect of GAD-alum

The role of positivity to autoantibodies in the effect of autoantigen immunomodulation in type

1 diabetes is unclear. In our study, it was interesting to observe that baseline GADA levels

were higher in patients with better C-peptide preservation 15 months after the first injection of

GAD-alum (Paper IV). Our results are in line with those from the DPT-1 study, where a

beneficial effect of treatment with oral insulin was observed in individuals with high levels of

IAA pre-treatment, although this study failed to prevent type 1 diabetes (Skyler et al. 2005). A

possible explanation for the association between the response and GADA titres might be that

higher levels of GADA in circulation resulted in abundant antigen-antibody complex

formation. Complexes circulate effectively, resulting in a more efficient antigen processing

and presentation. In an experimental model of autoimmune disease, injection of specific

immune complexes was able to powerfully redirect immune responses (Barabas et al. 2007).

Thus, it is possible that induction of a more potent immune response by GAD65-antibody

complexes might be responsible for an increased response to GAD65. It cannot be excluded

that GADA levels simply reflect the current immunological status, and patients with higher

autoantibody levels may have more beta-cells left. However, this seems less likely, as it has

n.s


92

been shown that after the diagnosis of type 1 diabetes the presence of GADA is independent

of whether or not patients still have measurable C-peptide levels (Jensen et al. 2007). Also,

another study showed that presence of GADA at onset of type 1 diabetes was not correlated

with rate of beta-cell destruction after clinical onset (Batstra et al. 1997). We have to be

cautious for drawing further conclusions from our results regarding better C-peptide

preservation in individuals with higher GADA levels, based on a very small group of patients.

However, this is an interesting observation that should be further addressed in ongoing trials

including larger number of patients.

In conclusion, we found that the increase in GADA levels after GAD-alum treatment was

accompanied by a temporarily increased GADA-binding to the b96.11-defined epitope. In

addition, a decreased percentage of GADA of the IgG1 subclass and increased IgG4 were

observed, which might suggest a temporary switch towards a protective Th2 response. The

safety of the treatment was further confirmed by the antigen-specific effect together with the

lack of changes in SPS-related epitope binding and GAD65 enzymatic activity.

93

Summary and conclusions

Autoantibodies related to type 1 diabetes are commonly measured in healthy children, often

with increased genetic risk, to predict the disease, and in new onset diabetic individuals for

classification of type 1 diabetes. In this thesis, we investigated the prevalence and persistence

of autoantibodies against GAD and IA-2, and studied the relation between these

autoantibodies and HLA genotypes, both in healthy children and in newly diagnosed type 1

diabetic children from the ABIS study. Our results confirmed the importance of HLA risk

genotypes for the development of type 1 diabetes, and indicated that HLA risk genotypes play

an important role for persistence of autoantibodies. On the other hand, the induction of

autoantibodies in the general population, especially transient autoantibodies, seems to be more

influenced by environmental factors. It is possible that other genetic factors than HLA might

also play a role in the induction of autoantibodies.

Radioimmunoassays (RIAs) are commonly used for detection of autoantibodies and are

regularly validated and compared between research laboratories around the world. However,

there are some disadvantages with these methods, such as the use of radioactively labelled

material, low sensitivity for the IAA assay and that they are time consuming. Due to the need

for better methods we developed an SPR-based assay for detection of IAA. The main

advantages of this method are that there is no need for labelling, and thus no work with

radioactivity, and it is time-saving compared to RIA. We were able to measure IAA with a

similar sensitivity as RIA, but further optimization of the SPR-based method is required. We

also made attempts to develop a method for detection of GADA, but unfortunately without

successful results. I believe that this is an area where further efforts should be done in the

future. It is important to have fast and well-functioning methods for detection of

autoantibodies, especially in the view of the increased possibilities for prevention, when there

will be a need to screen the general child population for identification of high-risk individuals.

Besides measuring the prevalence and levels of autoantibodies in type 1 diabetes-related

studies, it is also of interest to investigate their properties, such as subclass distribution,

affinity and epitope reactivity. During the time from emergence of autoantibodies to the

clinical manifestation of diabetes, the properties of the autoantibodies changes gradually,

generally with increasing affinity, changed distribution of subclasses and epitope spreading. A

Summary and conclusions ’

94

new field of interest is to study autoantibodies induced by immunomodulatory treatments with

autoantigens. The fact that GADA levels were increased after two injections of 20 µg GAD-

alum in type 1 diabetic children, raised the question whether the antibodies were different

after treatment than before. Our results showed a temporary increased binding to the b96.11-

defined epitope in the GAD-alum treated type 1 diabetic children, which was also the most

prevalent epitope for GADA binding before treatment. No other changes in epitope

specificities were observed, including the two SPS-related epitopes, which together with the

antigen specific effect of the treatment and the lack of changes in GAD65 enzymatic activity,

further speak in favor of the safety of GAD-alum.

Since the GAD-alum treatment induced an early Th2-directed cytokine response, we wanted

to clarify if the distribution of IgG subclasses was changed upon treatment. We investigated

the effect of GAD-alum on the distribution of the subclasses IgG1, 2, 3 and 4. A decreased

proportion of the predominantly occurring IgG1 subclass and increased IgG4, associated in

humans with Th1 and Th2 responses respectively, suggested a switch towards a protective

Th2 response. It was interesting to observe that the change in subclass distribution was only

temporary, and occurred when GADA titers were highest.

Concluding remarks and future perspectives

This thesis mainly comprises measurements of autoantibodies related to type 1 diabetes. We

observed that HLA risk genotypes play an important role for development of type 1 diabetes,

while induction of transient autoantibodies seems to be more influenced by environmental

factors. New improved methods for detection of autoantibodies are needed, where our SPR-

based immunoassay seems to be a good candidate, but further development is required. The

GAD-alum treatment induced changes in the humoral immune response, with increased levels

of GADA and temporary changes in epitope binding and IgG subclass distribution. GAD-

alum seems to be a promising treatment for type 1 diabetes, and hopefully many of the

research questions originated from our results will be answered during the ongoing phase III

trials.

95

Acknowledgements

I would like to end this thesis by expressing my sincere gratitude and thanks to all who

contributed and helped me to accomplish this. In particular, I would like to thank:

My supervisors: Johnny Ludvigsson for giving me the opportunity to do research within the

field of diabetes and to visit and do laboratory work at the University of Washington in

Seattle for 6 weeks during 2006.Thank you for your enthusiasm and helpfulness. Then I

would like to thank Rosaura Casas, my assistant supervisor during the second half of my

research studies. Your enthusiasm and support have helped me a lot and you were always

willing to discuss small and big matters whenever I needed. Outi Vaarala, my supervisor

during the first half of my research studies. Thank you for your ideas and discussions and for

sharing your knowledge in the field of immunology.

All children and parents participating in this study.

Special thanks to Jenny Carlsson, my twinning partner within the Strategic Area for

Prevention of Diabetes and its Complications (a joint strategy area for the County Council of

Östergötland and Linköping University), working at the Division of Molecular Physics at the

Department of Physics, Chemistry and Biology. Thank you for all collaboration and our

discussions about trouble-shooting and new ideas during all the hours in front of the Biacore

instrument, and for your friendship.

I would like to thank Chris Hampe, at the Univeristy of Washington in Seattle, for the time I

stayed in Seattle, for learning me the laboratory methods and for taking care of me. Thank

you for your kindness, helpfulness and our nice chats, and also for all e-mail communication

and discussions after the visit. Åke Lernmark, for letting me stay at your laboratory.

All friends and present and former colleagues at the Pediatric Laboratory in Linköping and

at the Clinical and Experimental Research for interesting discussions and nice chats in the

coffee room.

Ingela Johansson for your friendship, for scientific and non-scientific discussions and for all

refreshing chats and nice Friday “fika”.

My former and present roommates Maria Hjorth, Anna Rydén, Cecilia Bivik, Petra

Wäster and Christina Bendrik for all nice chats and interesting discussions, both regarding

research and more personal matters.

For excellent technical skills I would like to thank Gosia Smolinska-Konefal, Ingela

Johansson and Lena Berglert.

Mikael Chéramy for our discussions and chats about matters concerning the GAD-alum trial.

Maria Jenmalm for reading and giving constructive comments on the frame work of this

thesis.

Acknowledgements ’

96

My collaborators in different projects, Jeanette Wahlberg, Jorma Ilonen, Gunilla

Westermark, Karin Enander and Bo Liedberg for your invaluable knowledge you shared

with me and for your comments on manuscripts.

Ann-Christine Gilmore-Ellis for always being helpful and for your great service with

administrative issues, and Madeleine Örlin for your great service with administrative issues

regarding the dissertation.

Forum Scientum, the graduate school at Linköping University, for providing a

multidisciplinary research environment within the Diabetes Research Centre.

This thesis was supported by grants from the Swedish Child Diabetes Foundation

(Barndiabetesfonden), Swedish Research Council (Vetenskapsrådet), Swedish Diabetes

Association, Medical Research Fund of the County of Östergötland, the National Institutes of

Health, a basic research award from the American Diabetes Association, the Strategic Area

for Prevention of Diabetes and its Complications, and an unrestricted grant from Diamyd

Medical AB.

Till sist och viktigast av alla:

Mina föräldrar Inger och Lars. Tack för allt stöd och uppmuntran genom åren och att ni alltid

tror på mig och är stolta över mig.

Min storebror Magnus, för alla trevliga stunder tillsammans och samtal om allt från politik

till fritid och barn.

Min kusin Helene med familj, och Mariana för att ni är min lilla men naggande goda släkt.

Anna-Lisa och Göran samt Ulrika med familj, för att jag alltid känner mig välkommen hos

er.

Gamla och nya vänner, i och utanför Linköping, för allt roligt vi har hittat på under åren!

Min man Johan, tack för ditt stöd och uppmuntran. Jag älskar dig!

Våra barn Anton och Amanda. Ni är två underbara solstrålar!

97

References

Achenbach, P., Bonifacio, E., Koczwara, K. and Ziegler, A. G. (2005). Natural history of type 1

diabetes. Diabetes; 54 Suppl 2: S25-31.

Achenbach, P., Koczwara, K., Knopff, A., Naserke, H., Ziegler, A. G. and Bonifacio, E. (2004a).

Mature high-affinity immune responses to (pro)insulin anticipate the autoimmune cascade that leads to

type 1 diabetes. J Clin Invest; 114(4): 589-597.

Achenbach, P., Lampasona, V., Landherr, U., Koczwara, K., Krause, S., Grallert, H., Winkler, C.,

Pfluger, M., Illig, T., Bonifacio, E. and Ziegler, A. G. (2009). Autoantibodies to zinc transporter 8 and

SLC30A8 genotype stratify type 1 diabetes risk. Diabetologia; 52(9): 1881-1888.

Achenbach, P., Schlosser, M., Williams, A. J., Yu, L., Mueller, P. W., Bingley, P. J. and Bonifacio, E.

(2007). Combined testing of antibody titer and affinity improves insulin autoantibody measurement:

Diabetes Antibody Standardization Program. Clin Immunol; 122(1): 85-90.

Achenbach, P., Warncke, K., Reiter, J., Naserke, H. E., Williams, A. J., Bingley, P. J., Bonifacio, E.

and Ziegler, A. G. (2004b). Stratification of type 1 diabetes risk on the basis of islet autoantibody

characteristics. Diabetes; 53(2): 384-392.

Achenbach, P., Warncke, K., Reiter, J., Williams, A. J., Ziegler, A. G., Bingley, P. J. and Bonifacio, E.

(2006). Type 1 diabetes risk assessment: improvement by follow-up measurements in young islet

autoantibody-positive relatives. Diabetologia; 49(12): 2969-2976.

Agardh, C. D., Cilio, C. M., Lethagen, A., Lynch, K., Leslie, R. D., Palmer, M., Harris, R. A.,

Robertson, J. A. and Lernmark, A. (2005). Clinical evidence for the safety of GAD65

immunomodulation in adult-onset autoimmune diabetes. J Diabetes Complications; 19(4): 238-246.

Akerblom, H. K., Vaarala, O., Hyoty, H., Ilonen, J. and Knip, M. (2002). Environmental factors in the

etiology of type 1 diabetes. Am J Med Genet; 115(1): 18-29.

Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K. and Watson, J. D. (2002). Molecular biology of

the cell. New York, Garland Science.

Aly, T. A., Ide, A., Jahromi, M. M., Barker, J. M., Fernando, M. S., Babu, S. R., Yu, L., Miao, D.,

Erlich, H. A., Fain, P. R., Barriga, K. J., Norris, J. M., Rewers, M. J. and Eisenbarth, G. S. (2006).

Extreme genetic risk for type 1A diabetes. Proc Natl Acad Sci U S A; 103(38): 14074-14079.

American Diabetes Association (2008). Diagnosis and classification of diabetes mellitus. Diabetes

Care; 31 Suppl 1: S55-60.

Ankelo, M., Westerlund-Karlsson, A., Ilonen, J., Knip, M., Savola, K., Kankaanpaa, P., Merio, L.,

Siitari, H. and Hinkkanen, A. (2003). Time-resolved fluorometric assay for detection of autoantibodies

to glutamic acid decarboxylase (GAD65). Clin Chem; 49(6 Pt 1): 908-915.

Atkinson, M. A., Bowman, M. A., Campbell, L., Darrow, B. L., Kaufman, D. L. and Maclaren, N. K.

(1994). Cellular immunity to a determinant common to glutamate decarboxylase and coxsackie virus

in insulin-dependent diabetes. J Clin Invest; 94(5): 2125-2129.

Atkinson, M. A. and Eisenbarth, G. S. (2001). Type 1 diabetes: new perspectives on disease

pathogenesis and treatment. Lancet; 358(9277): 221-229.

References ’

98

Axelsson, S., Hjorth, M., Akerman, L., Ludvigsson, J. and Casas, R. (2010). Early induction of

GAD(65)-reactive Th2 response in type 1 diabetic children treated with alum-formulated GAD(65).

Diabetes Metab Res Rev; 26(7): 559-568.

Ayela, C., Roquet, F., Valera, L., Granier, C., Nicu, L. and Pugniere, M. (2007). Antibody-antigenic

peptide interactions monitored by SPR and QCM-D. A model for SPR detection of IA-2

autoantibodies in human serum. Biosens Bioelectron; 22(12): 3113-3119.

Bach, J. F. (1994). Insulin-dependent diabetes mellitus as an autoimmune disease. Endocr Rev; 15(4):

516-542.

Bach, J. F. and Chatenoud, L. (2001). Tolerance to islet autoantigens in type 1 diabetes. Annu Rev

Immunol; 19: 131-161.

Baekkeskov, S., Aanstoot, H. J., Christgau, S., Reetz, A., Solimena, M., Cascalho, M., Folli, F.,

Richter-Olesen, H. and De Camilli, P. (1990). Identification of the 64K autoantigen in insulin-

dependent diabetes as the GABA-synthesizing enzyme glutamic acid decarboxylase. Nature;

347(6289): 151-156.

Baekkeskov, S., Nielsen, J. H., Marner, B., Bilde, T., Ludvigsson, J. and Lernmark, A. (1982).

Autoantibodies in newly diagnosed diabetic children immunoprecipitate human pancreatic islet cell

proteins. Nature; 298(5870): 167-169.

Baisch, J. M., Weeks, T., Giles, R., Hoover, M., Stastny, P. and Capra, J. D. (1990). Analysis of HLA-

DQ genotypes and susceptibility in insulin-dependent diabetes mellitus. N Engl J Med; 322(26): 1836-

1841.

Bakhtadze, E., Borg, H., Stenstrom, G., Fernlund, P., Arnqvist, H. J., Ekbom-Schnell, A., Bolinder, J.,

Eriksson, J. W., Gudbjornsdottir, S., Nystrom, L., Groop, L. C. and Sundkvist, G. (2006). HLA-DQB1

genotypes, islet antibodies and beta cell function in the classification of recent-onset diabetes among

young adults in the nationwide Diabetes Incidence Study in Sweden. Diabetologia; 49(8): 1785-1794.

Barabas, A. Z., Cole, C. D., Kovacs, Z. B. and Lafreniere, R. (2007). Elevated antibody response by

antigen presentation in immune complexes. Med Sci Monit; 13(5): BR119-124.

Barker, J. M., Barriga, K. J., Yu, L., Miao, D., Erlich, H. A., Norris, J. M., Eisenbarth, G. S. and

Rewers, M. (2004). Prediction of autoantibody positivity and progression to type 1 diabetes: Diabetes

Autoimmunity Study in the Young (DAISY). J Clin Endocrinol Metab; 89(8): 3896-3902.

Batstra, M. R., Pina, M., Quan, J., Mulder, P., de Beaufort, C. E., Bruining, G. J. and Aanstoot, H. J.

(1997). Fluctuations in GAD65 antibodies after clinical diagnosis of IDDM in young children.

Diabetes Care; 20(4): 642-644.

Bekris, L. M., Jensen, R. A., Lagerquist, E., Hall, T. R., Agardh, C. D., Cilio, C. M., Lethagen, A. L.,

Lernmark, A., Robertson, J. A. and Hampe, C. S. (2007). GAD65 autoantibody epitopes in adult

patients with latent autoimmune diabetes following GAD65 vaccination. Diabet Med; 24(5): 521-526.

Bennett, S. T., Lucassen, A. M., Gough, S. C., Powell, E. E., Undlien, D. E., Pritchard, L. E.,

Merriman, M. E., Kawaguchi, Y., Dronsfield, M. J., Pociot, F. and et al. (1995). Susceptibility to

human type 1 diabetes at IDDM2 is determined by tandem repeat variation at the insulin gene

minisatellite locus. Nat Genet; 9(3): 284-292.

Bingley, P. J., Bonifacio, E. and Mueller, P. W. (2003). Diabetes Antibody Standardization Program:

first assay proficiency evaluation. Diabetes; 52(5): 1128-1136.

References

99

Bingley, P. J., Bonifacio, E., Williams, A. J., Genovese, S., Bottazzo, G. F. and Gale, E. A. (1997).

Prediction of IDDM in the general population: strategies based on combinations of autoantibody

markers. Diabetes; 46(11): 1701-1710.

Bingley, P. J., Christie, M. R., Bonifacio, E., Bonfanti, R., Shattock, M., Fonte, M. T., Bottazzo, G. F.

and Gale, E. A. (1994). Combined analysis of autoantibodies improves prediction of IDDM in islet

cell antibody-positive relatives. Diabetes; 43(11): 1304-1310.

Bonifacio, E., Genovese, S., Braghi, S., Bazzigaluppi, E., Lampasona, V., Bingley, P. J., Rogge, L.,

Pastore, M. R., Bognetti, E., Bottazzo, G. F. and et al. (1995). Islet autoantibody markers in IDDM:

risk assessment strategies yielding high sensitivity. Diabetologia; 38(7): 816-822.

Bonifacio, E., Lampasona, V., Bernasconi, L. and Ziegler, A. G. (2000). Maturation of the humoral

autoimmune response to epitopes of GAD in preclinical childhood type 1 diabetes. Diabetes; 49(2):

202-208.

Bonifacio, E., Scirpoli, M., Kredel, K., Fuchtenbusch, M. and Ziegler, A. G. (1999). Early

autoantibody responses in prediabetes are IgG1 dominated and suggest antigen-specific regulation. J

Immunol; 163(1): 525-532.

Borch-Johnsen, K., Joner, G., Mandrup-Poulsen, T., Christy, M., Zachau-Christiansen, B., Kastrup, K.

and Nerup, J. (1984). Relation between breast-feeding and incidence rates of insulin-dependent

diabetes mellitus. A hypothesis. Lancet; 2(8411): 1083-1086.

Bottazzo, G. F., Florin-Christensen, A. and Doniach, D. (1974). Islet-cell antibodies in diabetes

mellitus with autoimmune polyendocrine deficiencies. Lancet; 2(7892): 1279-1283.

Brekke, O. H., Michaelsen, T. E. and Sandlie, I. (1995). The structural requirements for complement

activation by IgG: does it hinge on the hinge? Immunol Today; 16(2): 85-90.

Brooking, H., Ananieva-Jordanova, R., Arnold, C., Amoroso, M., Powell, M., Betterle, C., Zanchetta,

R., Furmaniak, J. and Smith, B. R. (2003). A sensitive non-isotopic assay for GAD65 autoantibodies.

Clin Chim Acta; 331(1-2): 55-59.

Bu, D. F., Erlander, M. G., Hitz, B. C., Tillakaratne, N. J., Kaufman, D. L., Wagner-McPherson, C. B.,

Evans, G. A. and Tobin, A. J. (1992). Two human glutamate decarboxylases, 65-kDa GAD and 67-

kDa GAD, are each encoded by a single gene. Proc Natl Acad Sci U S A; 89(6): 2115-2119.

Cameron, M. J., Arreaza, G. A., Zucker, P., Chensue, S. W., Strieter, R. M., Chakrabarti, S. and

Delovitch, T. L. (1997). IL-4 prevents insulitis and insulin-dependent diabetes mellitus in nonobese

diabetic mice by potentiation of regulatory T helper-2 cell function. J Immunol; 159(10): 4686-4692.

Christgau, S., Aanstoot, H. J., Schierbeck, H., Begley, K., Tullin, S., Hejnaes, K. and Baekkeskov, S.

(1992). Membrane anchoring of the autoantigen GAD65 to microvesicles in pancreatic beta-cells by

palmitoylation in the NH2-terminal domain. J Cell Biol; 118(2): 309-320.

Christie, M. R., Genovese, S., Cassidy, D., Bosi, E., Brown, T. J., Lai, M., Bonifacio, E. and Bottazzo,

G. F. (1994). Antibodies to islet 37k antigen, but not to glutamate decarboxylase, discriminate rapid

progression to IDDM in endocrine autoimmunity. Diabetes; 43(10): 1254-1259.

Cohen, S. (1963). Antibodies. Mod Trends Immunol; 55: 25-52.

Couper, J. J., Harrison, L. C., Aldis, J. J., Colman, P. G., Honeyman, M. C. and Ferrante, A. (1998).

IgG subclass antibodies to glutamic acid decarboxylase and risk for progression to clinical insulin-

dependent diabetes. Hum Immunol; 59(8): 493-499.

References ’

100

Dahlquist, G. and Mustonen, L. (1994). Childhood onset diabetes--time trends and climatological

factors. Int J Epidemiol; 23(6): 1234-1241.

Dahlquist, G. G., Ivarsson, S., Lindberg, B. and Forsgren, M. (1995). Maternal enteroviral infection

during pregnancy as a risk factor for childhood IDDM. A population-based case-control study.

Diabetes; 44(4): 408-413.

Daw, K., Ujihara, N., Atkinson, M. and Powers, A. C. (1996). Glutamic acid decarboxylase

autoantibodies in stiff-man syndrome and insulin-dependent diabetes mellitus exhibit similarities and

differences in epitope recognition. J Immunol; 156(2): 818-825.

Donner, H., Rau, H., Walfish, P. G., Braun, J., Siegmund, T., Finke, R., Herwig, J., Usadel, K. H. and

Badenhoop, K. (1997). CTLA4 alanine-17 confers genetic susceptibility to Graves' disease and to type

1 diabetes mellitus. J Clin Endocrinol Metab; 82(1): 143-146.

DPT-1, D. P. T.-T. D. S. G. (2002). Effects of insulin in relatives of patients with type 1 diabetes

mellitus. N Engl J Med; 346(22): 1685-1691.

Dromey, J. A., Weenink, S. M., Peters, G. H., Endl, J., Tighe, P. J., Todd, I. and Christie, M. R.

(2004). Mapping of epitopes for autoantibodies to the type 1 diabetes autoantigen IA-2 by peptide

phage display and molecular modeling: overlap of antibody and T cell determinants. J Immunol;

172(7): 4084-4090.

Eisenbarth, G. S. (2008). The Pancreatic Beta-cell. In G.S. Eisenbarth (Ed), Type 1 diabetes:

Cellular, molecular and clinical immunology. Online Edition Version 3.0. Denver, Colorado,

USA: Barbara Davis Center for Childhood Diabetes.

Eising, S., Nilsson, A., Carstensen, B., Hougaard, D. M., Norgaard-Petersen, B., Nerup, J., Lernmark,

A. and Pociot, F. (2010). Danish children born 1981-2002 with GAD65 and IA-2 autoantibodies at

birth had an increased risk to develop type 1 diabetes before 2004. Eur J Endocrinol.

Falorni, A., Ackefors, M., Carlberg, C., Daniels, T., Persson, B., Robertson, J. and Lernmark, A.

(1996). Diagnostic sensitivity of immunodominant epitopes of glutamic acid decarboxylase (GAD65)

autoantibodies in childhood IDDM. Diabetologia; 39(9): 1091-1098.

Fenalti, G., Hampe, C. S., Arafat, Y., Law, R. H., Banga, J. P., Mackay, I. R., Whisstock, J. C.,

Buckle, A. M. and Rowley, M. J. (2008). COOH-terminal clustering of autoantibody and T-cell

determinants on the structure of GAD65 provide insights into the molecular basis of autoreactivity.

Diabetes; 57(5): 1293-1301.

Fontenot, J. D., Gavin, M. A. and Rudensky, A. Y. (2003). Foxp3 programs the development and

function of CD4(+)CD25(+) regulatory T cells. Nature Immunology; 4(4): 330-336.

Gale, E. A. M., Bingley, P. J., Knip, M., Emmett, C. L., Swankie, H., Fewell, S., Kearsey, P., Schober,

E., Gorus, F. K., Dupre, J., Mahon, J. L., Profozic, V., Reimers, J. I., Mandrup-Poulsen, T., Levy-

Marchal, C., et al. (2004). European Nicotinamide Diabetes Intervention Trial (ENDIT): a randomised

controlled trial of intervention before the onset of type 1 diabetes. Lancet; 363(9413): 925-931.

Gascan, H., Gauchat, J. F., Roncarolo, M. G., Yssel, H., Spits, H. and de Vries, J. E. (1991). Human B

cell clones can be induced to proliferate and to switch to IgE and IgG4 synthesis by interleukin 4 and a

signal provided by activated CD4+ T cell clones. J Exp Med; 173(3): 747-750.

Geha, R. S. (1985). Idiotypic-antiidiotypic interactions in man. Ric Clin Lab; 15(1): 1-8.

References

101

Genovese, S., Bonfanti, R., Bazzigaluppi, E., Lampasona, V., Benazzi, E., Bosi, E., Chiumello, G. and

Bonifacio, E. (1996). Association of IA-2 autoantibodies with HLA DR4 phenotypes in IDDM.

Diabetologia; 39(10): 1223-1226.

Gillespie, K. M., Bain, S. C., Barnett, A. H., Bingley, P. J., Christie, M. R., Gill, G. V. and Gale, E. A.

(2004). The rising incidence of childhood type 1 diabetes and reduced contribution of high-risk HLA

haplotypes. Lancet; 364(9446): 1699-1700.

Gilliam, L. K., Binder, K. A., Banga, J. P., Madec, A. M., Ortqvist, E., Kockum, I., Luo, D. and

Hampe, C. S. (2004). Multiplicity of the antibody response to GAD65 in Type I diabetes. Clin Exp

Immunol; 138(2): 337-341.

Glastras, S. J., Mohsin, F. and Donaghue, K. C. (2005). Complications of diabetes mellitus in

childhood. Pediatr Clin North Am; 52(6): 1735-1753.

Gobi, K. V., Iwasaka, H. and Miura, N. (2007). Self-assembled PEG monolayer based SPR

immunosensor for label-free detection of insulin. Biosens Bioelectron; 22(7): 1382-1389.

Gomara, M. J., Ercilla, G., Alsina, M. A. and Haro, I. (2000). Assessment of synthetic peptides for

hepatitis A diagnosis using biosensor technology. J Immunol Methods; 246(1-2): 13-24.

Gorus, F. K., Goubert, P., Semakula, C., Vandewalle, C. L., De Schepper, J., Scheen, A., Christie, M.

R. and Pipeleers, D. G. (1997). IA-2-autoantibodies complement GAD65-autoantibodies in new-onset

IDDM patients and help predict impending diabetes in their siblings. The Belgian Diabetes Registry.

Diabetologia; 40(1): 95-99.

Green, A. and Patterson, C. C. (2001). Trends in the incidence of childhood-onset diabetes in Europe

1989-1998. Diabetologia; 44 Suppl 3: B3-8.

Greenbaum, C. J., Palmer, J. P., Kuglin, B. and Kolb, H. (1992). Insulin autoantibodies measured by

radioimmunoassay methodology are more related to insulin-dependent diabetes mellitus than those

measured by enzyme-linked immunosorbent assay: results of the Fourth International Workshop on

the Standardization of Insulin Autoantibody Measurement. J Clin Endocrinol Metab; 74(5): 1040-

1044.

Grubin, C. E., Daniels, T., Toivola, B., Landin-Olsson, M., Hagopian, W. A., Li, L., Karlsen, A. E.,

Boel, E., Michelsen, B. and Lernmark, A. (1994). A novel radioligand binding assay to determine

diagnostic accuracy of isoform-specific glutamic acid decarboxylase antibodies in childhood IDDM.


Hagopian, W. A., Lernmark, A., Rewers, M. J., Simell, O. G., She, J. X., Ziegler, A. G., Krischer, J. P.

and Akolkar, B. (2006). TEDDY--The Environmental Determinants of Diabetes in the Young: an

observational clinical trial. Ann N Y Acad Sci; 1079: 320-326.

Hagopian, W. A., Michelsen, B., Karlsen, A. E., Larsen, F., Moody, A., Grubin, C. E., Rowe, R.,

Petersen, J., McEvoy, R. and Lernmark, A. (1993). Autoantibodies in IDDM primarily recognize the

65,000-M(r) rather than the 67,000-M(r) isoform of glutamic acid decarboxylase. Diabetes; 42(4):

631-636.

Hagopian, W. A., Sanjeevi, C. B., Kockum, I., Landin-Olsson, M., Karlsen, A. E., Sundkvist, G.,

Dahlquist, G., Palmer, J. and Lernmark, A. (1995). Glutamate decarboxylase-, insulin-, and islet cell-

antibodies and HLA typing to detect diabetes in a general population-based study of Swedish children.

J Clin Invest; 95(4): 1505-1511.

References ’

102

Hamilton, R. G. (1987). Human IgG subclass measurements in the clinical laboratory. Clin Chem;

33(10): 1707-1725.

Hampe, C. S., Hammerle, L. P., Bekris, L., Ortqvist, E., Persson, B. and Lernmark, A. (2002). Stable

GAD65 autoantibody epitope patterns in type 1 diabetes children five years after onset. J Autoimmun;

18(1): 49-53.

Harjutsalo, V., Sjoberg, L. and Tuomilehto, J. (2008). Time trends in the incidence of type 1 diabetes

in Finnish children: a cohort study. Lancet; 371(9626): 1777-1782.

Hawa, M. I., Beyan, H., Buckley, L. R. and Leslie, R. D. (2002). Impact of genetic and non-genetic

factors in type 1 diabetes. Am J Med Genet; 115(1): 8-17.

Hawa, M. I., Fava, D., Medici, F., Deng, Y. J., Notkins, A. L., De Mattia, G. and Leslie, R. D. (2000).

Antibodies to IA-2 and GAD65 in type 1 and type 2 diabetes: isotype restriction and polyclonality.

Diabetes Care; 23(2): 228-233.

Hermann, R., Knip, M., Veijola, R., Simell, O., Laine, A. P., Akerblom, H. K., Groop, P. H.,

Forsblom, C., Pettersson-Fernholm, K. and Ilonen, J. (2003). Temporal changes in the frequencies of

HLA genotypes in patients with Type 1 diabetes--indication of an increased environmental pressure?


Hermann, R., Turpeinen, H., Laine, A. P., Veijola, R., Knip, M., Simell, O., Sipila, I., Akerblom, H.

K. and Ilonen, J. (2003). HLA DR-DQ-encoded genetic determinants of childhood-onset type 1

diabetes in Finland: an analysis of 622 nuclear families. Tissue Antigens; 62(2): 162-169.

Herold, K. C., Gitelman, S., Greenbaum, C., Puck, J., Hagopian, W., Gottlieb, P., Sayre, P., Bianchine,

P., Wong, E., Seyfert-Margolis, V., Bourcier, K. and Bluestone, J. A. (2009). Treatment of patients

with new onset Type 1 diabetes with a single course of anti-CD3 mAb Teplizumab preserves insulin

production for up to 5 years. Clin Immunol; 132(2): 166-173.

Herold, K. C., Gitelman, S. E., Masharani, U., Hagopian, W., Bisikirska, B., Donaldson, D., Rother,

K., Diamond, B., Harlan, D. M. and Bluestone, J. A. (2005). A single course of anti-CD3 monoclonal

antibody hOKT3gamma1(Ala-Ala) results in improvement in C-peptide responses and clinical

parameters for at least 2 years after onset of type 1 diabetes. Diabetes; 54(6): 1763-1769.

Herold, K. C., Hagopian, W., Auger, J. A., Poumian-Ruiz, E., Taylor, L., Donaldson, D., Gitelman, S.

E., Harlan, D. M., Xu, D., Zivin, R. A. and Bluestone, J. A. (2002). Anti-CD3 monoclonal antibody in

new-onset type 1 diabetes mellitus. N Engl J Med; 346(22): 1692-1698.

Herrwerth, S., Rosendahl, T., Feng, C., Fick, W., Eck, M., Himmelhaus, R., Dahint, R. and Grunze,

M. (2003). Covalent coupling of antibodies to self-assembled monolayers of carboxy-functionalized

poly(ethylene glycol): Protein resistance and specific binding of biomolecules. Langmuir; 19: 1880-

1887.

Hillman, M., Torn, C. and Landin-Olsson, M. (2009). The glutamic acid decarboxylase 65

immunoglobulin G subclass profile differs between adult-onset type 1 diabetes and latent autoimmune

diabetes in adults (LADA) up to 3 years after clinical onset. Clin Exp Immunol; 157(2): 255-260.

Hillman, M., Torn, C., Thorgeirsson, H. and Landin-Olsson, M. (2004). IgG4-subclass of glutamic

acid decarboxylase antibody is more frequent in latent autoimmune diabetes in adults than in type 1

diabetes. Diabetologia; 47(11): 1984-1989.

References

103

Hjorth, M., Axelsson, S., Ryden, A., Faresjo, M., Ludvigsson, J. and Casas, R. (2010). GAD-alum

treatment induces GAD(65)-specific CD4(+)CD25(high)FOXP3(+) cells in type 1 diabetic patients.

Clin Immunol.

Holmberg, H., Vaarala, O., Sadauskaite-Kuehne, V., Ilonen, J., Padaiga, Z. and Ludvigsson, J. (2006).

Higher prevalence of autoantibodies to insulin and GAD65 in Swedish compared to Lithuanian

children with type 1 diabetes. Diabetes Res Clin Pract; 72(3): 308-314.

Holmberg, H., Wahlberg, J., Vaarala, O. and Ludvigsson, J. (2007). Short duration of breast-feeding

as a risk-factor for beta-cell autoantibodies in 5-year-old children from the general population. Br J

Nutr; 97(1): 111-116.

Hoppu, S., Harkonen, T., Ronkainen, M. S., Akerblom, H. K. and Knip, M. (2004b). IA-2 antibody

epitopes and isotypes during the prediabetic process in siblings of children with type 1 diabetes. J

Autoimmun; 23(4): 361-370.

Hoppu, S., Ronkainen, M. S., Kimpimaki, T., Simell, S., Korhonen, S., Ilonen, J., Simell, O. and Knip,

M. (2004c). Insulin autoantibody isotypes during the prediabetic process in young children with

increased genetic risk of type 1 diabetes. Pediatr Res; 55(2): 236-242.

Hoppu, S., Ronkainen, M. S., Kulmala, P., Akerblom, H. K. and Knip, M. (2004a). GAD65 antibody

isotypes and epitope recognition during the prediabetic process in siblings of children with type I

diabetes. Clin Exp Immunol; 136(1): 120-128.

Hori, S., Nomura, T. and Sakaguchi, S. (2003). Control of regulatory T cell development by the

transcription factor Foxp3. Science; 299(5609): 1057-1061.

http://clinicaltrials.gov (2010a). The Protégé Study - Clinical Trial of MGA031 in Children and Adults

With Recent-Onset Type 1 Diabetes Mellitus. From

http://clinicaltrials.gov/ct2/show/NCT00385697?term=MGA031-01&rank=2.

http://clinicaltrials.gov (2010b). A Phase III Study to Investigate the Impact of Diamyd in Patients

Newly Diagnosed With Type 1 Diabetes (EU). From

http://clinicaltrials.gov/ct2/show/NCT00723411?term=diamyd&rank=1.

Hummel, M., Bonifacio, E., Schmid, S., Walter, M., Knopff, A. and Ziegler, A. G. (2004). Brief

communication: early appearance of islet autoantibodies predicts childhood type 1 diabetes in

offspring of diabetic parents. Ann Intern Med; 140(11): 882-886.

Hussain, R., Kifayet, A. and Chiang, T. J. (1995). Immunoglobulin G1 (IgG1) and IgG3 antibodies are

markers of progressive disease in leprosy. Infect Immun; 63(2): 410-415.

Hyoty, H., Hiltunen, M., Knip, M., Laakkonen, M., Vahasalo, P., Karjalainen, J., Koskela, P.,

Roivainen, M., Leinikki, P., Hovi, T. and et al. (1995). A prospective study of the role of coxsackie B

and other enterovirus infections in the pathogenesis of IDDM. Childhood Diabetes in Finland (DiMe)

Study Group. Diabetes; 44(6): 652-657.

IDF (2009). International Diabetes Federation Diabetes Atlas online version: www.eatlas.idf.org.

Ilonen, J., Sjoroos, M., Knip, M., Veijola, R., Simell, O., Akerblom, H. K., Paschou, P., Bozas, E.,

Havarani, B., Malamitsi-Puchner, A., Thymelli, J., Vazeou, A. and Bartsocas, C. S. (2002). Estimation

of genetic risk for type 1 diabetes. Am J Med Genet; 115(1): 30-36.

Imagawa, A., Hanafusa, T., Itoh, N., Waguri, M., Yamamoto, K., Miyagawa, J., Moriwaki, M.,

Yamagata, K., Iwahashi, H., Sada, M., Tsuji, T., Tamura, S., Kawata, S., Kuwajima, M., Nakajima,

References ’

104

H., et al. (1999). Immunological abnormalities in islets at diagnosis paralleled further deterioration of

glycaemic control in patients with recent-onset Type I (insulin-dependent) diabetes mellitus.


Janeway, C., Travers, P., Walport, M. and Shlomchik, M. (2005). Immunobiology, Garland Science

Publishing.

Jensen, R. A., Gilliam, L. K., Torn, C., Landin-Olsson, M., Karlsson, F. A., Palmer, J. P., Kockum, I.,

Akesson, K., Lernmark, B., Lynch, K., Breslow, N. and Lernmark, A. (2007). Multiple factors affect

the loss of measurable C-peptide over 6 years in newly diagnosed 15- to 35-year-old diabetic subjects.

J Diabetes Complications; 21(4): 205-213.

Johansson, C., Samuelsson, U. and Ludvigsson, J. (1994). A high weight gain early in life is

associated with an increased risk of type 1 (insulin-dependent) diabetes mellitus. Diabetologia; 37(1):

91-94.

Jongerius-Gortemaker, B. G., Goverde, R. L., van Knapen, F. and Bergwerff, A. A. (2002). Surface

plasmon resonance (BIACORE) detection of serum antibodies against Salmonella enteritidis and

Salmonella typhimurium. J Immunol Methods; 266(1-2): 33-44.

Jun, H. S., Chung, Y. H., Han, J., Kim, A., Yoo, S. S., Sherwin, R. S. and Yoon, J. W. (2002).

Prevention of autoimmune diabetes by immunogene therapy using recombinant vaccinia virus

expressing glutamic acid decarboxylase. Diabetologia; 45(5): 668-676.

Karlsson, R., Fagerstam, L., Nilshans, H. and Persson, B. (1993). Analysis of active antibody

concentration. Separation of affinity and concentration parameters. J Immunol Methods; 166(1): 75-

84.

Karvonen, M., Pitkaniemi, J. and Tuomilehto, J. (1999). The onset age of type 1 diabetes in Finnish

children has become younger. The Finnish Childhood Diabetes Registry Group. Diabetes Care; 22(7):

1066-1070.

Kim, J., Namchuk, M., Bugawan, T., Fu, Q., Jaffe, M., Shi, Y., Aanstoot, H. J., Turck, C. W., Erlich,

H., Lennon, V. and Baekkeskov, S. (1994). Higher autoantibody levels and recognition of a linear

NH2-terminal epitope in the autoantigen GAD65, distinguish stiff-man syndrome from insulin-

dependent diabetes mellitus. J Exp Med; 180(2): 595-606.

Kimpimaki, T., Kupila, A., Hamalainen, A. M., Kukko, M., Kulmala, P., Savola, K., Simell, T.,

Keskinen, P., Ilonen, J., Simell, O. and Knip, M. (2001). The first signs of beta-cell autoimmunity

appear in infancy in genetically susceptible children from the general population: the Finnish Type 1

Diabetes Prediction and Prevention Study. J Clin Endocrinol Metab; 86(10): 4782-4788.

Knip, M. (1997). Disease-associated autoimmunity and prevention of insulin-dependent diabetes

mellitus. Ann Med; 29(5): 447-451.

Knip, M., Korhonen, S., Kulmala, P., Veijola, R., Reunanen, A., Raitakari, O. T., Viikari, J. and

Akerblom, H. K. (2010). Prediction of type 1 diabetes in the general population. Diabetes Care; 33(6):

1206-1212.

Knip, M. and Siljander, H. (2008). Autoimmune mechanisms in type 1 diabetes. Autoimmun Rev; 7(7):

550-557.

Knip, M., Virtanen, S. M., Seppa, K., Ilonen, J., Savilahti, E., Vaarala, O., Reunanen, A., Teramo, K.,

Hamalainen, A. M., Paronen, J., Dosch, H. M., Hakulinen, T. and Akerblom, H. K. (2010). Dietary

intervention in infancy and later signs of beta-cell autoimmunity. N Engl J Med; 363(20): 1900-1908.

References

105

Koeleman, B. P., Lie, B. A., Undlien, D. E., Dudbridge, F., Thorsby, E., de Vries, R. R., Cucca, F.,

Roep, B. O., Giphart, M. J. and Todd, J. A. (2004). Genotype effects and epistasis in type 1 diabetes

and HLA-DQ trans dimer associations with disease. Genes Immun; 5(5): 381-388.

Korn, T., Bettelli, E., Oukka, M. and Kuchroo, V. K. (2009). IL-17 and Th17 Cells. Annu Rev

Immunol; 27: 485-517.

Krischer, J. P., Cuthbertson, D. D., Yu, L., Orban, T., Maclaren, N., Jackson, R., Winter, W. E.,

Schatz, D. A., Palmer, J. P. and Eisenbarth, G. S. (2003). Screening strategies for the identification of

multiple antibody-positive relatives of individuals with type 1 diabetes. J Clin Endocrinol Metab;

88(1): 103-108.

Kukko, M., Kimpimaki, T., Korhonen, S., Kupila, A., Simell, S., Veijola, R., Simell, T., Ilonen, J.,

Simell, O. and Knip, M. (2005). Dynamics of diabetes-associated autoantibodies in young children

with human leukocyte antigen-conferred risk of type 1 diabetes recruited from the general population.

J Clin Endocrinol Metab; 90(5): 2712-2717.

Kulmala, P., Rahko, J., Savola, K., Vahasalo, P., Sjoroos, M., Reunanen, A., Ilonen, J. and Knip, M.

(2001). Beta-cell autoimmunity, genetic susceptibility, and progression to type 1 diabetes in

unaffected schoolchildren. Diabetes Care; 24(1): 171-173.

Kulmala, P., Savola, K., Petersen, J. S., Vahasalo, P., Karjalainen, J., Lopponen, T., Dyrberg, T.,

Akerblom, H. K. and Knip, M. (1998). Prediction of insulin-dependent diabetes mellitus in siblings of

children with diabetes. A population-based study. The Childhood Diabetes in Finland Study Group. J

Clin Invest; 101(2): 327-336.

Kulmala, P., Savola, K., Reijonen, H., Veijola, R., Vahasalo, P., Karjalainen, J., Tuomilehto-Wolf, E.,

Ilonen, J., Tuomilehto, J., Akerblom, H. K. and Knip, M. (2000). Genetic markers, humoral

autoimmunity, and prediction of type 1 diabetes in siblings of affected children. Childhood Diabetes in

Finland Study Group. Diabetes; 49(1): 48-58.

Kupila, A., Muona, P., Simell, T., Arvilommi, P., Savolainen, H., Hamalainen, A. M., Korhonen, S.,

Kimpimaki, T., Sjoroos, M., Ilonen, J., Knip, M. and Simell, O. (2001). Feasibility of genetic and

immunological prediction of type I diabetes in a population-based birth cohort. Diabetologia; 44(3):

290-297.

Kure, M., Katsura, Y., Kosano, H., Noritake, M., Watanabe, T., Iwaki, Y., Nishigori, H. and

Matsuoka, T. (2005). A trial to assess the amount of insulin antibodies in diabetic patients by surface

plasmon resonance. Intern Med; 44(2): 100-106.

Laaksonen, M., Pastinen, T., Sjoroos, M., Kuokkanen, S., Ruutiainen, J., Sumelahti, M. L., Reijonen,

H., Salonen, R., Wikstrom, J., Panelius, M., Partanen, J., Tienari, P. J. and Ilonen, J. (2002). HLA

class II associated risk and protection against multiple sclerosis-a Finnish family study. J

Neuroimmunol; 122(1-2): 140-145.

Ladner, M. B., Bottini, N., Valdes, A. M. and Noble, J. A. (2005). Association of the single nucleotide

polymorphism C1858T of the PTPN22 gene with type 1 diabetes. Hum Immunol; 66(1): 60-64.

LaGasse, J. M., Brantley, M. S., Leech, N. J., Rowe, R. E., Monks, S., Palmer, J. P., Nepom, G. T.,

McCulloch, D. K. and Hagopian, W. A. (2002). Successful prospective prediction of type 1 diabetes in

schoolchildren through multiple defined autoantibodies: an 8-year follow-up of the Washington State

Diabetes Prediction Study. Diabetes Care; 25(3): 505-511.

References ’

106

Lambert, A. P., Gillespie, K. M., Thomson, G., Cordell, H. J., Todd, J. A., Gale, E. A. and Bingley, P.

J. (2004). Absolute risk of childhood-onset type 1 diabetes defined by human leukocyte antigen class

II genotype: a population-based study in the United Kingdom. J Clin Endocrinol Metab; 89(8): 4037-

4043.

Lan, M. S., Lu, J., Goto, Y. and Notkins, A. L. (1994). Molecular cloning and identification of a

receptor-type protein tyrosine phosphatase, IA-2, from human insulinoma. DNA Cell Biol; 13(5): 505-

514.

Lan, M. S., Wasserfall, C., Maclaren, N. K. and Notkins, A. L. (1996). IA-2, a transmembrane protein

of the protein tyrosine phosphatase family, is a major autoantigen in insulin-dependent diabetes

mellitus. Proc Natl Acad Sci U S A; 93(13): 6367-6370.

Larsson, A., Ekblad, T., Andersson, O. and Liedberg, B. (2007). Photografted poly(ethylene glycol)

matrix for affinity interaction studies. Biomacromolecules; 8(1): 287-295.

Lee, J. W., Sim, S. J., Cho, S. M. and Lee, J. (2005). Characterization of a self-assembled monolayer

of thiol on a gold surface and the fabrication of a biosensor chip based on surface plasmon resonance

for detecting anti-GAD antibody. Biosens Bioelectron; 20(7): 1422-1427.

Leslie, R. D., Atkinson, M. A. and Notkins, A. L. (1999). Autoantigens IA-2 and GAD in Type I

(insulin-dependent) diabetes. Diabetologia; 42(1): 3-14.

Levy, L. M., Dalakas, M. C. and Floeter, M. K. (1999). The stiff-person syndrome: an autoimmune

disorder affecting neurotransmission of gamma-aminobutyric acid. Ann Intern Med; 131(7): 522-530.

Li, L., Hagopian, W. A., Brashear, H. R., Daniels, T. and Lernmark, A. (1994). Identification of

autoantibody epitopes of glutamic acid decarboxylase in stiff-man syndrome patients. J Immunol;

152(2): 930-934.

Liedberg, B., Nylander, C. and Lundstrom, I. (1995). Biosensing with surface plasmon resonance--

how it all started. Biosens Bioelectron; 10(8): i-ix.

Lin, R. Y., Schwartz, R. A. and Vafaie, J. (2010). Immunoglobulin G deficiency. From

http://emedicine.medscape.com/article/136897-overview.

Liu, E. and Eisenbarth, G. S. (2007). Accepting clocks that tell time poorly: fluid-phase versus

standard ELISA autoantibody assays. Clin Immunol; 125(2): 120-126.

Lu, J., Li, Q., Xie, H., Chen, Z. J., Borovitskaya, A. E., Maclaren, N. K., Notkins, A. L. and Lan, M. S.

(1996). Identification of a second transmembrane protein tyrosine phosphatase, IA-2beta, as an

autoantigen in insulin-dependent diabetes mellitus: precursor of the 37-kDa tryptic fragment. Proc

Natl Acad Sci U S A; 93(6): 2307-2311.

Ludvigsson, J. (2006). Why diabetes incidence increases--a unifying theory. Ann N Y Acad Sci; 1079:

374-382.

Ludvigsson, J. (2010a). Immune intervention in children with type 1 diabetes. Curr Diab Rep; 10(5):

370-379.

Ludvigsson, J. (2011). Diabetes vaccines: review of the clinical evidence. Clin. Invest.; E-pub ahead of

print.

Ludvigsson, J., Faresjo, M., Hjorth, M., Axelsson, S., Cheramy, M., Pihl, M., Vaarala, O., Forsander,

G., Ivarsson, S., Johansson, C., Lindh, A., Nilsson, N. O., Aman, J., Ortqvist, E., Zerhouni, P., et al.

References

107

(2008). GAD treatment and insulin secretion in recent-onset type 1 diabetes. N Engl J Med; 359(18):

1909-1920.

Ludvigsson, J., Hjorth, M., Cheramy, M., Axelsson, S., Pihl, M., Forsander, G., Nilsson, N. O.,

Samuelsson, B. O., Wood, T., Aman, J., Ortqvist, E. and Casas, R. (2010b). Extended evaluation of

the safety and efficacy of GAD treatment of children and adolescents with recent-onset type 1

diabetes: a randomised controlled trial. Diabetologia.

Ludvigsson, J., Ludvigsson, M. and Sepa, A. (2001). Screening for prediabetes in the general child

population: maternal attitude to participation. Pediatr Diabetes; 2(4): 170-174.

Lumley, T., Diehr, P., Emerson, S. and Chen, L. (2002). The importance of the normality assumption

in large public health data sets. Annu Rev Public Health; 23: 151-169.

Lundgren, M., Persson, U., Larsson, P., Magnusson, C., Smith, C. I., Hammarstrom, L. and

Severinson, E. (1989). Interleukin 4 induces synthesis of IgE and IgG4 in human B cells. Eur J

Immunol; 19(7): 1311-1315.

Luo, D., Rogers, C. N., Steed, J. T., Gilliam, L. K., Hampe, C. S. and Lernmark, A. (2004).

Development of glutamic acid decarboxylase 65 (GAD65) autoantibody assay using biotin-GAD65

fusion protein. J Biotechnol; 111(1): 97-104.

Löfås, S. and Johnsson, B. (1990). A novel hydrogel matrix on gold surfaces in surface plasmon

resonance sensors for fast and efficient covalent immobilization of ligands. J Chem Soc, Chem

Commun; 21: 1526-1528.

MacCuish, A. C., Irvine, W. J., Barnes, E. W. and Duncan, L. J. (1974). Antibodies to pancreatic islet

cells in insulin-dependent diabetics with coexistent autoimmune disease. Lancet; 2(7896): 1529-1531.

MacLennan, I. C. M., GulbransonJudge, A., Toellner, K. M., CasamayorPalleja, M., Chan, E., Sze, D.

M. Y., Luther, S. A. and Orbea, H. A. (1997). The changing preference of T and B cells for partners as

T-dependent antibody responses develop. Immunological Reviews; 156: 53-66.

Masson, J. F., Battaglia, T. M., Khairallah, P., Beaudoin, S. and Booksh, K. S. (2007). Quantitative

measurement of cardiac markers in undiluted serum. Anal Chem; 79(2): 612-619.

Mayr, A., Schlosser, M., Grober, N., Kenk, H., Ziegler, A. G., Bonifacio, E. and Achenbach, P.

(2007). GAD autoantibody affinity and epitope specificity identify distinct immunization profiles in

children at risk for type 1 diabetes. Diabetes; 56(6): 1527-1533.

McKee, A. S., MacLeod, M., White, J., Crawford, F., Kappler, J. W. and Marrack, P. (2008). Gr1+IL-

4-producing innate cells are induced in response to Th2 stimuli and suppress Th1-dependent antibody

responses. Int Immunol; 20(5): 659-669.

Melendez-Ramirez, L. Y., Richards, R. J. and Cefalu, W. T. (2010). Complications of type 1 diabetes.

Endocrinol Metab Clin North Am; 39(3): 625-640.

Menser, M. A., Forrest, J. M. and Bransby, R. D. (1978). Rubella infection and diabetes mellitus.

Lancet; 1(8055): 57-60.

Milgrom, F. and Witebsky, E. (1963). Antoantibody Formation as a Failure of "Self-Recognition".

Proc Can Cancer Conf; 5: 319-336.

Mire-Sluis, A. R., Gaines Das, R. and Lernmark, A. (2000). The World Health Organization

International Collaborative Study for islet cell antibodies. Diabetologia; 43(10): 1282-1292.

References ’

108

Moriwaki, M., Itoh, N., Miyagawa, J., Yamamoto, K., Imagawa, A., Yamagata, K., Iwahashi, H.,

Nakajima, H., Namba, M., Nagata, S., Hanafusa, T. and Matsuzawa, Y. (1999). Fas and Fas ligand

expression in inflamed islets in pancreas sections of patients with recent-onset Type I diabetes

mellitus. Diabetologia; 42(11): 1332-1340.

Mosmann, T. R. and Sad, S. (1996). The expanding universe of T-cell subsets: Th1, Th2 and more.

Immunol Today; 17(3): 138-146.

Mowen, K. A. and Glimcher, L. H. (2004). Signaling pathways in Th2 development. Immunological

Reviews; 202: 203-222.

Moya-Suri, V., Schlosser, M., Zimmermann, K., Rjasanowski, I., Gurtler, L. and Mentel, R. (2005).

Enterovirus RNA sequences in sera of schoolchildren in the general population and their association

with type 1-diabetes-associated autoantibodies. J Med Microbiol; 54(Pt 9): 879-883.

Mueller, P. (2007). Workshop report: Autoantibody markers for Type 1 diabetes. Diabetes

Autoantibodies Standardization Program 2007. Discussion and future direction. 9th International

Congress of Immunology of Diabetes Society, Nov 14-18, 2007, Florida. Acta Diabetologica; 44

((Suppl 1) Nov).

Murinson, B. B. (2004). Stiff-person syndrome. Neurologist; 10(3): 131-137.

Murinson, B. B., Butler, M., Marfurt, K., Gleason, S., De Camilli, P. and Solimena, M. (2004).

Markedly elevated GAD antibodies in SPS: effects of age and illness duration. Neurology; 63(11):

2146-2148.

Mölne, J. and Wold, A. (2007). Inflammation, Liber.

Nanto-Salonen, K., Kupila, A., Simell, S., Siljander, H., Salonsaari, T., Hekkala, A., Korhonen, S.,

Erkkola, R., Sipila, J. I., Haavisto, L., Siltala, M., Tuominen, J., Hakalax, J., Hyoty, H., Ilonen, J., et

al. (2008). Nasal insulin to prevent type 1 diabetes in children with HLA genotypes and autoantibodies

conferring increased risk of disease: a double-blind, randomised controlled trial. Lancet; 372(9651):

1746-1755.

Nejentsev, S., Sjoroos, M., Soukka, T., Knip, M., Simell, O., Lovgren, T. and Ilonen, J. (1999).

Population-based genetic screening for the estimation of Type 1 diabetes mellitus risk in Finland:

selective genotyping of markers in the HLA-DQB1, HLA-DQA1 and HLA-DRB1 loci. Diabet Med;

16(12): 985-992.

Nistico, L., Buzzetti, R., Pritchard, L. E., Van der Auwera, B., Giovannini, C., Bosi, E., Larrad, M. T.,

Rios, M. S., Chow, C. C., Cockram, C. S., Jacobs, K., Mijovic, C., Bain, S. C., Barnett, A. H.,

Vandewalle, C. L., et al. (1996). The CTLA-4 gene region of chromosome 2q33 is linked to, and

associated with, type 1 diabetes. Belgian Diabetes Registry. Hum Mol Genet; 5(7): 1075-1080.

Norris, J. M., Beaty, B., Klingensmith, G., Yu, L. P., Hoffman, M., Chase, H. P., Erlich, H. A.,

Hamman, R. F., Eisenbarth, G. S. and Rewers, M. (1996). Lack of association between early exposure

to cow's milk protein and beta-cell autoimmunity - Diabetes autoimmunity study in the young

(DAISY). Jama-Journal of the American Medical Association; 276(8): 609-614.

Norris, J. M., Yin, X., Barriga, K., Eisenbarth, G. S. and Rewers, M. (2007). Timing of cereal

introduction in the infant diet is associated with earlier onset but not later onset islet autoimmunity:

The diabetes autoimmunity study in the young (DAISY). Diabetes; 56: A262-A262.

References

109

Notkins, A. L. and Lernmark, A. (2001). Autoimmune type 1 diabetes: resolved and unresolved issues.

J Clin Invest; 108(9): 1247-1252.

Notkins, A. L., Lu, J., Li, Q., VanderVegt, F. P., Wasserfall, C., Maclaren, N. K. and Lan, M. S.

(1996). IA-2 and IA-2 beta are major autoantigens in IDDM and the precursors of the 40 kDa and 37

kDa tryptic fragments. J Autoimmun; 9(5): 677-682.

Oak, S., Gilliam, L. K., Landin-Olsson, M., Torn, C., Kockum, I., Pennington, C. R., Rowley, M. J.,

Christie, M. R., Banga, J. P. and Hampe, C. S. (2008). The lack of anti-idiotypic antibodies, not the

presence of the corresponding autoantibodies to glutamate decarboxylase, defines type 1 diabetes.

Proceedings of the National Academy of Sciences of the United States of America; 105(14): 5471-

5476.

Onengut-Gumuscu, S., Ewens, K. G., Spielman, R. S. and Concannon, P. (2004). A functional

polymorphism (1858C/T) in the PTPN22 gene is linked and associated with type I diabetes in

multiplex families. Genes Immun; 5(8): 678-680.

Onkamo, P., Vaananen, S., Karvonen, M. and Tuomilehto, J. (1999). Worldwide increase in incidence

of Type I diabetes--the analysis of the data on published incidence trends. Diabetologia; 42(12): 1395-

1403.

Padoa, C. J., Banga, J. P., Madec, A. M., Ziegler, M., Schlosser, M., Ortqvist, E., Kockum, I., Palmer,

J., Rolandsson, O., Binder, K. A., Foote, J., Luo, D. and Hampe, C. S. (2003). Recombinant Fabs of

human monoclonal antibodies specific to the middle epitope of GAD65 inhibit type 1 diabetes-specific

GAD65Abs. Diabetes; 52(11): 2689-2695.

Padoa, C. J., Crowther, N. J., Thomas, J. W., Hall, T. R., Bekris, L. M., Torn, C., Landin-Olsson, M.,

Ortqvist, E., Palmer, J. P., Lernmark, A. and Hampe, C. S. (2005). Epitope analysis of insulin

autoantibodies using recombinant Fab. Clin Exp Immunol; 140(3): 564-571.

Palmer, J. P., Asplin, C. M., Clemons, P., Lyen, K., Tatpati, O., Raghu, P. K. and Paquette, T. L.

(1983). Insulin antibodies in insulin-dependent diabetics before insulin treatment. Science; 222(4630):

1337-1339.

Patterson, C. C., Dahlquist, G. G., Gyurus, E., Green, A. and Soltesz, G. (2009). Incidence trends for

childhood type 1 diabetes in Europe during 1989-2003 and predicted new cases 2005-20: a multicentre

prospective registration study. Lancet; 373(9680): 2027-2033.

Peng, H. and Hagopian, W. (2006). Environmental factors in the development of Type 1 diabetes. Rev

Endocr Metab Disord; 7(3): 149-162.

Pescovitz, M. D., Greenbaum, C. J., Krause-Steinrauf, H., Becker, D. J., Gitelman, S. E., Goland, R.,

Gottlieb, P. A., Marks, J. B., McGee, P. F., Moran, A. M., Raskin, P., Rodriguez, H., Schatz, D. A.,

Wherrett, D., Wilson, D. M., et al. (2009). Rituximab, B-lymphocyte depletion, and preservation of

beta-cell function. N Engl J Med; 361(22): 2143-2152.

Petrovsky, N., Silva, D. and Schatz, D. A. (2003). Vaccine therapies for the prevention of type 1

diabetes mellitus. Paediatr Drugs; 5(9): 575-582.

Phillips, K. S., Han, J. H. and Cheng, Q. (2007). Development of a "membrane cloaking" method for

amperometric enzyme immunoassay and surface plasmon resonance analysis of proteins in serum

samples. Anal Chem; 79(3): 899-907.

Potter, K. N. and Wilkin, T. J. (2000). The molecular specificity of insulin autoantibodies. Diabetes

Metab Res Rev; 16(5): 338-353.

References ’

110

Press release Macrogenics and Lilly (2010). MacroGenics and Lilly Announce Pivotal Clinical Trial

of Teplizumab Did Not Meet Primary Efficacy Endpoint. From

https://investor.lilly.com/releasedetail2.cfm?ReleaseID=521014.

Pugliese, A., Gianani, R., Moromisato, R., Awdeh, Z. L., Alper, C. A., Erlich, H. A., Jackson, R. A.

and Eisenbarth, G. S. (1995). HLA-DQB1*0602 is associated with dominant protection from diabetes

even among islet cell antibody-positive first-degree relatives of patients with IDDM. Diabetes; 44(6):

608-613.

Purkerson, J. and Isakson, P. (1992). A two-signal model for regulation of immunoglobulin isotype

switching. FASEB J; 6(14): 3245-3252.

Raju, R., Foote, J., Banga, J. P., Hall, T. R., Padoa, C. J., Dalakas, M. C., Ortqvist, E. and Hampe, C.

S. (2005). Analysis of GAD65 autoantibodies in Stiff-Person syndrome patients. J Immunol; 175(11):

7755-7762.

Raju, R. and Hampe, C. S. (2008). Immunobiology of stiff-person syndrome. Int Rev Immunol; 27(1-

2): 79-92.

Rakocevic, G., Raju, R. and Dalakas, M. C. (2004). Anti-glutamic acid decarboxylase antibodies in

the serum and cerebrospinal fluid of patients with stiff-person syndrome: correlation with clinical

severity. Arch Neurol; 61(6): 902-904.

Rautajoki, K. J., Kylaniemi, M. K., Raghav, S. K., Rao, K. and Lahesmaa, R. (2008). An insight into

molecular mechanisms of human T helper cell differentiation. Ann Med; 40(5): 322-335.

Redondo, M. J. and Eisenbarth, G. S. (2002). Genetic control of autoimmunity in Type I diabetes and

associated disorders. Diabetologia; 45(5): 605-622.

Redondo, M. J., Rewers, M., Yu, L., Garg, S., Pilcher, C. C., Elliott, R. B. and Eisenbarth, G. S.

(1999). Genetic determination of islet cell autoimmunity in monozygotic twin, dizygotic twin, and

non-twin siblings of patients with type 1 diabetes: prospective twin study. BMJ; 318(7185): 698-702.

Revoltella, R. P., Laricchia Robbio, L. and Liedberg, B. (1998). Comparison of conventional

immunoassays (RIA, ELISA) with surface plasmon resonance for pesticide detection and monitoring.

Biotherapy; 11(2-3): 135-145.

Rich, S. S., Akolkar, B., Concannon, P., Erlich, H., Hilner, J. E., Julier, C., Morahan, G., Nerup, J.,

Nierras, C., Pociot, F. and Todd, J. A. (2009). Current status and the future for the genetics of type I

diabetes. Genes Immun; 10 Suppl 1: S128-131.

Richter, W., Shi, Y. and Baekkeskov, S. (1993). Autoreactive epitopes defined by diabetes-associated

human monoclonal antibodies are localized in the middle and C-terminal domains of the smaller form

of glutamate decarboxylase. Proc Natl Acad Sci U S A; 90(7): 2832-2836.

Ronkainen, M. S., Hoppu, S., Korhonen, S., Simell, S., Veijola, R., Ilonen, J., Simell, O. and Knip, M.

(2006). Early epitope- and isotype-specific humoral immune responses to GAD65 in young children

with genetic susceptibility to type 1 diabetes. Eur J Endocrinol; 155(4): 633-642.

Ronkainen, M. S., Savola, K. and Knip, M. (2004). Antibodies to GAD65 epitopes at diagnosis and

over the first 10 years of clinical type 1 diabetes mellitus. Scand J Immunol; 59(3): 334-340.

Rose, N. R. and Bona, C. (1993). Defining criteria for autoimmune diseases (Witebsky's postulates

revisited). Immunol Today; 14(9): 426-430.

References

111

Rother, K. I., Brown, R. J., Morales, M. M., Wright, E., Duan, Z., Campbell, C., Hardin, D. S.,

Popovic, J., McEvoy, R. C., Harlan, D. M., Orlander, P. R. and Brod, S. A. (2009). Effect of ingested

interferon-alpha on beta-cell function in children with new-onset type 1 diabetes. Diabetes Care;

32(7): 1250-1255.

Roux, K. H., Strelets, L. and Michaelsen, T. E. (1997). Flexibility of human IgG subclasses. J

Immunol; 159(7): 3372-3382.

Sabbah, E., Savola, K., Kulmala, P., Veijola, R., Vahasalo, P., Karjalainen, J., Akerblom, H. K. and

Knip, M. (1999). Diabetes-associated autoantibodies in relation to clinical characteristics and natural

course in children with newly diagnosed type 1 diabetes. The Childhood Diabetes In Finland Study

Group. J Clin Endocrinol Metab; 84(5): 1534-1539.

Sakaguchi, S. (2004). Naturally arising CD4(+) regulatory T cells for immunologic self-tolerance and

negative control of immune responses. Annual Review of Immunology; 22: 531-562.

Salminen, K., Sadeharju, K., Lonnrot, M., Vahasalo, P., Kupila, A., Korhonen, S., Ilonen, J., Simell,

O., Knip, M. and Hyoty, H. (2003). Enterovirus infections are associated with the induction of beta-

cell autoimmunity in a prospective birth cohort study. J Med Virol; 69(1): 91-98.

Sanjeevi, C. B., Landin-Olsson, M., Kockum, I., Dahlquist, G. and Lernmark, A. (1995). Effects of the

second HLA-DQ haplotype on the association with childhood insulin-dependent diabetes mellitus.

Tissue Antigens; 45(2): 148-152.

Savola, K., Bonifacio, E., Sabbah, E., Kulmala, P., Vahasalo, P., Karjalainen, J., Tuomilehto-Wolf, E.,

Merilainen, J., Akerblom, H. K. and Knip, M. (1998). IA-2 antibodies--a sensitive marker of IDDM

with clinical onset in childhood and adolescence. Childhood Diabetes in Finland Study Group.


Savola, K., Laara, E., Vahasalo, P., Kulmala, P., Akerblom, H. K. and Knip, M. (2001). Dynamic

pattern of disease-associated autoantibodies in siblings of children with type 1 diabetes: a population-

based study. Diabetes; 50(11): 2625-2632.

Savola, K., Sabbah, E., Kulmala, P., Vahasalo, P., Ilonen, J. and Knip, M. (1998). Autoantibodies

associated with Type I diabetes mellitus persist after diagnosis in children. Diabetologia; 41(11):

1293-1297.

Schloot, N. C., Meierhoff, G., Lengyel, C., Vandorfi, G., Takacs, J., Panczel, P., Barkai, L., Madacsy,

L., Oroszlan, T., Kovacs, P., Suto, G., Battelino, T., Hosszufalusi, N. and Jermendy, G. (2007). Effect

of heat shock protein peptide DiaPep277 on beta-cell function in paediatric and adult patients with

recent-onset diabetes mellitus type 1: two prospective, randomized, double-blind phase II trials.

Diabetes Metab Res Rev; 23(4): 276-285.

Schlosser, M., Banga, J. P., Madec, A. M., Binder, K. A., Strebelow, M., Rjasanowski, I., Wassmuth,

R., Gilliam, L. K., Luo, D. and Hampe, C. S. (2005a). Dynamic changes of GAD65 autoantibody

epitope specificities in individuals at risk of developing type 1 diabetes. Diabetologia; 48(5): 922-930.

Schlosser, M., Koczwara, K., Kenk, H., Strebelow, M., Rjasanowski, I., Wassmuth, R., Achenbach, P.,

Ziegler, A. G. and Bonifacio, E. (2005b). In insulin-autoantibody-positive children from the general

population, antibody affinity identifies those at high and low risk. Diabetologia; 48(9): 1830-1832.

Schlosser, M., Mueller, P. W., Torn, C., Bonifacio, E. and Bingley, P. J. (2010). Diabetes Antibody

Standardization Program: evaluation of assays for insulin autoantibodies. Diabetologia; 53(12): 2611-

2620.

References ’

112

Schmidli, R. S., Colman, P. G. and Bonifacio, E. (1995). Disease sensitivity and specificity of 52

assays for glutamic acid decarboxylase antibodies. The Second International GADAB Workshop.

Diabetes; 44(6): 636-640.

Schmidli, R. S., Colman, P. G., Bonifacio, E., Bottazzo, G. F. and Harrison, L. C. (1994). High level

of concordance between assays for glutamic acid decarboxylase antibodies. The First International

Glutamic Acid Decarboxylase Antibody Workshop. Diabetes; 43(8): 1005-1009.

Schroeder, H. W., Jr. and Cavacini, L. (2010). Structure and function of immunoglobulins. J Allergy

Clin Immunol; 125(2 Suppl 2): S41-52.

Seissler, J., Eikamp, K., Schott, M. and Scherbaum, W. A. (2002). IA-2 autoantibodies restricted to

the IgG4 subclass are associated with protection from type 1 diabetes. Horm Metab Res; 34(4): 186-

191.

Sepa, A., Frodi, A. and Ludvigsson, J. (2005). Mothers' experiences of serious life events increase the

risk of diabetes-related autoimmunity in their children. Diabetes Care; 28(10): 2394-2399.

Sepa, A., Wahlberg, J., Vaarala, O., Frodi, A. and Ludvigsson, J. (2005). Psychological stress may

induce diabetes-related autoimmunity in infancy. Diabetes Care; 28(2): 290-295.

Shankaran, D. R. and Miura, N. (2007). Trends in interfacial design for surface plasmon resonance

based immunoassays. J Phys D Appl Phys; 40: 7187-7200.

She, J. X. (1996). Susceptibility to type I diabetes: HLA-DQ and DR revisited. Immunol Today; 17(7):

323-329.

Sjoroos, M., Ilonen, J., Reijonen, H. and Lovgren, T. (1998). Time-resolved fluorometry based

sandwich hybridisation assay for HLA-DQA1 typing. Dis Markers; 14(1): 9-19.

Skyler, J. S. (2010). Glycemia and cardiovascular diseases in type 2 diabetes. J Intern Med; 268(5):

468-470.

Skyler, J. S., Krischer, J. P., Wolfsdorf, J., Cowie, C., Palmer, J. P., Greenbaum, C., Cuthbertson, D.,

Rafkin-Mervis, L. E., Chase, H. P. and Leschek, E. (2005). Effects of oral insulin in relatives of

patients with type 1 diabetes: The Diabetes Prevention Trial--Type 1. Diabetes Care; 28(5): 1068-

1076.

Snapper, C. M. and Mond, J. J. (1993). Towards a comprehensive view of immunoglobulin class

switching. Immunol Today; 14(1): 15-17.

Sodoyez-Goffaux, F., Koch, M., Dozio, N., Brandenburg, D. and Sodoyez, J. C. (1988). Advantages

and pitfalls of radioimmune and enzyme linked immunosorbent assays of insulin antibodies.


Solimena, M., Dirkx, R., Jr., Hermel, J. M., Pleasic-Williams, S., Shapiro, J. A., Caron, L. and Rabin,

D. U. (1996). ICA 512, an autoantigen of type I diabetes, is an intrinsic membrane protein of

neurosecretory granules. EMBO J; 15(9): 2102-2114.

Soltesz, G., Patterson, C. C. and Dahlquist, G. (2007). Worldwide childhood type 1 diabetes

incidence--what can we learn from epidemiology? Pediatr Diabetes; 8 Suppl 6: 6-14.

References

113

Stead, J. D. and Jeffreys, A. J. (2002). Structural analysis of insulin minisatellite alleles reveals

unusually large differences in diversity between Africans and non-Africans. Am J Hum Genet; 71(6):

1273-1284.

Steck, A. K., Zhang, W., Bugawan, T. L., Barriga, K. J., Blair, A., Erlich, H. A., Eisenbarth, G. S.,

Norris, J. M. and Rewers, M. J. (2009). Do non-HLA genes influence development of persistent islet

autoimmunity and type 1 diabetes in children with high-risk HLA-DR,DQ genotypes? Diabetes;

58(4): 1028-1033.

Stolt, U. G., Helgesson, G., Liss, P. E., Svensson, T. and Ludvigsson, J. (2005). Information and

informed consent in a longitudinal screening involving children: a questionnaire survey. Eur J Hum

Genet; 13(3): 376-383.

Stolt, U. G., Liss, P. E., Svensson, T. and Ludvigsson, J. (2002). Attitudes to bioethical issues: a case

study of a screening project. Soc Sci Med; 54(9): 1333-1344.

Strebelow, M., Schlosser, M., Ziegler, B., Rjasanowski, I. and Ziegler, M. (1999). Karlsburg Type I

diabetes risk study of a general population: frequencies and interactions of the four major Type I

diabetes-associated autoantibodies studied in 9419 schoolchildren. Diabetologia; 42(6): 661-670.

Stride, A. and Hattersley, A. T. (2002). Different genes, different diabetes: lessons from maturity-

onset diabetes of the young. Ann Med; 34(3): 207-216.

SWEDIABKIDS (2009). The Swedish national pediatric diabetes registry. Yearly Report 2009. From

https://www.ndr.nu/ndr2/.

Szabo, S. J., Sullivan, B. M., Peng, S. L. and Glimcher, L. H. (2003). Molecular mechanisms

regulating Th1 immune responses. Annu Rev Immunol; 21: 713-758.

The DIAMOND Project Group (2006). Incidence and trends of childhood Type 1 diabetes worldwide

1990-1999. Diabet Med; 23(8): 857-866.

The TEDDY Study Group (2007). The Environmental Determinants of Diabetes in the Young

(TEDDY) study: study design. Pediatr Diabetes; 8(5): 286-298.

The TEDDY Study Group (2008). The Environmental Determinants of Diabetes in the Young

(TEDDY) Study. Ann N Y Acad Sci; 1150: 1-13.

Tian, J., Atkinson, M. A., Clare-Salzler, M., Herschenfeld, A., Forsthuber, T., Lehmann, P. V. and

Kaufman, D. L. (1996). Nasal administration of glutamate decarboxylase (GAD65) peptides induces

Th2 responses and prevents murine insulin-dependent diabetes. J Exp Med; 183(4): 1561-1567.

Tian, J., Clare-Salzler, M., Herschenfeld, A., Middleton, B., Newman, D., Mueller, R., Arita, S.,

Evans, C., Atkinson, M. A., Mullen, Y., Sarvetnick, N., Tobin, A. J., Lehmann, P. V. and Kaufman, D.

L. (1996). Modulating autoimmune responses to GAD inhibits disease progression and prolongs islet

graft survival in diabetes-prone mice. Nat Med; 2(12): 1348-1353.

Tian, J. and Kaufman, D. L. (2009). Antigen-based therapy for the treatment of type 1 diabetes.

Diabetes; 58(9): 1939-1946.

Tisch, R., Wang, B. and Serreze, D. V. (1999). Induction of glutamic acid decarboxylase 65-specific

Th2 cells and suppression of autoimmune diabetes at late stages of disease is epitope dependent. J

Immunol; 163(3): 1178-1187.

References ’

114

Tisch, R., Yang, X. D., Singer, S. M., Liblau, R. S., Fugger, L. and McDevitt, H. O. (1993). Immune

response to glutamic acid decarboxylase correlates with insulitis in non-obese diabetic mice. Nature;

366(6450): 72-75.

Torn, C., Mueller, P. W., Schlosser, M., Bonifacio, E. and Bingley, P. J. (2008). Diabetes Antibody

Standardization Program: evaluation of assays for autoantibodies to glutamic acid decarboxylase and

islet antigen-2. Diabetologia; 51(5): 846-852.

TRIGR. (2010). Trial to reduce type 1 diabetes in the genetically at risk, from

http://trigr.epi.usf.edu/index.html.

Undlien, D. E., Lie, B. A. and Thorsby, E. (2001). HLA complex genes in type 1 diabetes and other

autoimmune diseases. Which genes are involved? Trends Genet; 17(2): 93-100.

Vaarala, O. (1999b). Gut and the induction of immune tolerance in type 1 diabetes. Diabetes Metab

Res Rev; 15(5): 353-361.

Vaarala, O., Knip, M., Paronen, J., Hamalainen, A. M., Muona, P., Vaatainen, M., Ilonen, J., Simell,

O. and Akerblom, H. K. (1999a). Cow's milk formula feeding induces primary immunization to

insulin in infants at genetic risk for type 1 diabetes. Diabetes; 48(7): 1389-1394.

Wahlberg, J., Fredriksson, J., Nikolic, E., Vaarala, O. and Ludvigsson, J. (2005). Environmental

factors related to the induction of beta-cell autoantibodies in 1-yr-old healthy children. Pediatr

Diabetes; 6(4): 199-205.

Wahlberg, J., Vaarala, O. and Ludvigsson, J. (2006). Dietary risk factors for the emergence of type 1

diabetes-related autoantibodies in 21/2 year-old Swedish children. Br J Nutr; 95(3): 603-608.

van der Werf, N., Kroese, F. G., Rozing, J. and Hillebrands, J. L. (2007). Viral infections as potential

triggers of type 1 diabetes. Diabetes Metab Res Rev; 23(3): 169-183.

Vardi, P., Ziegler, A. G., Mathews, J. H., Dib, S., Keller, R. J., Ricker, A. T., Wolfsdorf, J. I.,

Herskowitz, R. D., Rabizadeh, A., Eisenbarth, G. S. and et al. (1988). Concentration of insulin

autoantibodies at onset of type I diabetes. Inverse log-linear correlation with age. Diabetes Care;

11(9): 736-739.

Wenzlau, J. M., Juhl, K., Yu, L., Moua, O., Sarkar, S. A., Gottlieb, P., Rewers, M., Eisenbarth, G. S.,

Jensen, J., Davidson, H. W. and Hutton, J. C. (2007). The cation efflux transporter ZnT8 (Slc30A8) is

a major autoantigen in human type 1 diabetes. Proc Natl Acad Sci U S A; 104(43): 17040-17045.

Verge, C. F., Gianani, R., Kawasaki, E., Yu, L., Pietropaolo, M., Jackson, R. A., Chase, H. P. and

Eisenbarth, G. S. (1996). Prediction of type I diabetes in first-degree relatives using a combination of

insulin, GAD, and ICA512bdc/IA-2 autoantibodies. Diabetes; 45(7): 926-933.

Verge, C. F., Stenger, D., Bonifacio, E., Colman, P. G., Pilcher, C., Bingley, P. J. and Eisenbarth, G.

S. (1998). Combined use of autoantibodies (IA-2 autoantibody, GAD autoantibody, insulin

autoantibody, cytoplasmic islet cell antibodies) in type 1 diabetes: Combinatorial Islet Autoantibody

Workshop. Diabetes; 47(12): 1857-1866.

WHO (1999). Definition, Diagnosis and Classification of Diabetes Mellitus and its Complications:

Report of a WHO consultation Part 1: Diagnosis and Classifications of Diabetes Mellitus. Geneva,

World Health Organisation.

References

115

Widhe, M., Ekerfelt, C., Forsberg, P., Bergstrom, S. and Ernerudh, J. (1998). IgG subclasses in Lyme

borreliosis: a study of specific IgG subclass distribution in an interferon-gamma-predominated disease.

Scand J Immunol; 47(6): 575-581.

Vikinge, T. P., Askendal, A., Liedberg, B., Lindahl, T. and Tengvall, P. (1998). Immobilized chicken

antibodies improve the detection of serum antigens with surface plasmon resonance (SPR). Biosens

Bioelectron; 13(12): 1257-1262.

Wilkin, T. J. (1990). Insulin autoantibodies as markers for type I diabetes. Endocr Rev; 11(1): 92-104.

Wilkin, T. J. (2001). The accelerator hypothesis: weight gain as the missing link between Type I and

Type II diabetes. Diabetologia; 44(7): 914-922.

Williams, A. J., Bingley, P. J., Bonifacio, E., Palmer, J. P. and Gale, E. A. (1997). A novel micro-

assay for insulin autoantibodies. J Autoimmun; 10(5): 473-478.

Williams, A. J., Norcross, A. J., Dix, R. J., Gillespie, K. M., Gale, E. A. and Bingley, P. J. (2003). The

prevalence of insulin autoantibodies at the onset of Type 1 diabetes is higher in males than females

during adolescence. Diabetologia; 46(10): 1354-1356.

Winter, W. E., Harris, N. and Schatz, D. (2002). Type 1 diabetes islet autoantibody markers. Diabetes

Technol Ther; 4(6): 817-839.

Virtanen, S. M. and Knip, M. (2003). Nutritional risk predictors of beta cell autoimmunity and type 1

diabetes at a young age. Am J Clin Nutr; 78(6): 1053-1067.

Virtanen, S. M., Rasanen, L., Ylonen, K., Aro, A., Clayton, D., Langholz, B., Pitkaniemi, J., Savilahti,

E., Lounamaa, R., Tuomilehto, J. and et al. (1993). Early introduction of dairy products associated

with increased risk of IDDM in Finnish children. The Childhood in Diabetes in Finland Study Group.

Diabetes; 42(12): 1786-1790.

Viskari, H., Ludvigsson, J., Uibo, R., Salur, L., Marciulionyte, D., Hermann, R., Soltesz, G.,

Fuchtenbusch, M., Ziegler, A. G., Kondrashova, A., Romanov, A., Kaplan, B., Laron, Z., Koskela, P.,

Vesikari, T., et al. (2005). Relationship between the incidence of type 1 diabetes and maternal

enterovirus antibodies: time trends and geographical variation. Diabetologia; 48(7): 1280-1287.

Viskari, H., Ludvigsson, J., Uibo, R., Salur, L., Marciulionyte, D., Hermann, R., Soltesz, G.,

Fuchtenbusch, M., Ziegler, A. G., Kondrashova, A., Romanov, A., Knip, M. and Hyoty, H. (2004).

Relationship between the incidence of type 1 diabetes and enterovirus infections in different European

populations: results from the EPIVIR project. J Med Virol; 72(4): 610-617.

Viskari, H. R., Koskela, P., Lonnrot, M., Luonuansuu, S., Reunanen, A., Baer, M. and Hyoty, H.

(2000). Can enterovirus infections explain the increasing incidence of type 1 diabetes? Diabetes Care;

23(3): 414-416.

Xie, H., Zhang, B., Matsumoto, Y., Li, Q., Notkins, A. L. and Lan, M. S. (1997). Autoantibodies to

IA-2 and IA-2 beta in insulin-dependent diabetes mellitus recognize conformational epitopes: location

of the 37- and 40-kDa fragments determined. J Immunol; 159(7): 3662-3667.

Yoon, J. W. and Jun, H. S. (2005). Autoimmune destruction of pancreatic beta cells. Am J Ther; 12(6):

580-591.

Yu, J., Yu, L., Bugawan, T. L., Erlich, H. A., Barriga, K., Hoffman, M., Rewers, M. and Eisenbarth,

G. S. (2000). Transient antiislet autoantibodies: infrequent occurrence and lack of association with

"genetic" risk factors. J Clin Endocrinol Metab; 85(7): 2421-2428.

References ’

116

Yu, L., Rewers, M., Gianani, R., Kawasaki, E., Zhang, Y., Verge, C., Chase, P., Klingensmith, G.,

Erlich, H., Norris, J. and Eisenbarth, G. S. (1996). Antiislet autoantibodies usually develop

sequentially rather than simultaneously. J Clin Endocrinol Metab; 81(12): 4264-4267.

Zhang, B., Lan, M. S. and Notkins, A. L. (1997). Autoantibodies to IA-2 in IDDM: location of major

antigenic determinants. Diabetes; 46(1): 40-43.

Ziegler, A. G., Hummel, M., Schenker, M. and Bonifacio, E. (1999). Autoantibody appearance and

risk for development of childhood diabetes in offspring of parents with type 1 diabetes: the 2-year

analysis of the German BABYDIAB Study. Diabetes; 48(3): 460-468.

Ziegler, A. G. and Nepom, G. T. (2010). Prediction and pathogenesis in type 1 diabetes. Immunity;

32(4): 468-478.

Ziegler, A. G., Schmid, S., Huber, D., Hummel, M. and Bonifacio, E. (2003). Early infant feeding and

risk of developing type 1 diabetes-associated autoantibodies. JAMA; 290(13): 1721-1728.

Documents

Autoantibodies related to type 1 diabetes in childrenliu.diva-portal.org/smash/get/diva2:392919/FULLTEXT01.pdfIn conclusion, measurement of autoantibodies related to type 1 diabetes