from the Faculty of Medicine 999
_____________________________ _____________________________
Pteridine Dependent Hydroxylases as Autoantigens in Autoimmune
Polyendocrine Syndrome Type 1
BY
OLOV EKWALL
ACTA UNIVERSITATIS UPSALIENSIS
ABSTRACT
Ekwall, O. 2001. Pteridine dependent hydroxylases as autoantigens in autoimmune polyendocrine syndrome type I. Acta Universitatis Upsaliensis. Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 999. 81pp. Uppsala. ISBN 91-554-4941-7.
Autoimmune polyendocrine syndrome type I (APS) is a monogenous, recessively inherited disease characterised by endocrine and non-endocrine autoimmune manifestations. One fifth of APS I patients suffer from periodic intestinal dysfunction with varying degrees of malabsorbtion, steatorrhea and constipation. Alopecia areata is found in one third of APS I patients. By immunoscreening human cDNA libraries derived from normal human duodenum and scalp with APS I sera, we identified tryptophan hydroxylase (TPH) as an intestinal autoantigen and tyrosine hydroxylase (TH) as a dermal autoantigen. Forty-eight percent (38/80) of the APS I patients had TPH antibodies (Ab) and 44% (41/94) showed TH immunoreactivity. No reactivity against TPH or TH was seen in healthy controls. TPH-Abs showed a statistically significant correlation with gastrointestinal dysfunction (p<0.0001) and TH-Abs were significantly correlated to alopecia (p=0.02). TPH-Ab positive APS I sera specifically immunostained TPH containing enterochromaffin cells in normal duodenal mucosa. In affected mucosa a depletion of the TPH containing EC cells was seen. In enzyme inhibition experiments TPH and TH activity in vitro was reduced by adding APS I sera. TPH and TH together with phenylalanine hydroxylase (PAH) constitute the group of pteridine dependent hydroxylases. These are highly homologous enzymes involved in the biosynthesis of neurotransmitters. Immunoprecipitation of PAH expressed in vitro showed that 27% (25/94) of APS I patients had antibodies reacting with PAH, but no associations with clinical manifestations was observed. An immunocompetition assay showed that the PAH reactivity reflects a cross-reactivity with TPH.
In conclusion, we have identified TPH and TH as intestinal and dermal autoantigens in APS I, coupled to gastrointestinal dysfunction and alopecia. We have also demonstrated immunoreactivity against PAH in APS I patient sera reflecting a cross-reactivity with TPH.
Key words: APS I, alopecia, autoantigen, cDNA, malabsorbtion, phenylalanine hydroxylase, pteridine, tryptophan hydroxylase, tyrosine hydroxylase.
Olov Ekwall, Department of Medical Sciences, University Hospital, SE-751 85 Uppsala, Sweden, olov.ekwall@medsci.uu.se
Olov Ekwall 2001 ISSN 0282-7476 ISBN 91-554-4941-7
This thesis is based on the following papers, which will be referred to in the text by their roman numerals:
I. Ekwall O, Hedstrand H, Grimelius L, Haavik J, Perheentupa J, Gustafsson J, Husebye E, Kämpe O and Rorsman F. (1998) Identification of tryptophan hydroxylase as an intestinal autoantigen. Lancet, 1998; 352(9124): 279-283.
II. Ekwall O, Sjöberg K, Mirakian R, Rorsman F and Kämpe O. (1999)
Tryptophan hydroxylase autoantibodies and intestinal disease in autoimmune polyendocrine syndrome type 1. Lancet 1999; 354(9178): 568.
III. Hedstrand H, Ekwall O, Haavik J, Landgren E, Betterle C, Perheentupa J, Gustafsson J, Husebye E, Rorsman F and Kämpe O. (2000) Identification of Tyrosine Hydroxylase as an Autoantigen in Autoimmune Polyendocrine Syndrome Type I. Biochem Biophys Res Commun 2000; 267(1): 456-461.
IV. Ekwall O, Hedstrand H, Haavik J, Perheentupa J, Betterle C, Gustafsson J, Husebye E, Rorsman F and Kämpe O. (2000) Pteridine dependent
hydroxylases as autoantigens in autoimmune polyendocrine syndrome type I.
J Clin Endocrinol Metab 2000; 85(8): 2944-2950.
Reprints were made with the permission of the publishers.
ABBREVIATIONS 7 INTRODUCTION
General immunology 9
The innate immune system 9
The adaptive immune system 10
B-lymphocytes 12
Antibodies 13
T-lymphocytes 14
The T-cell receptor 15
The major histocompability complex 18
MHC class I 19
MHC class II 21
Tolerance 22
The “Danger” hypothesis 23
Autoimmunity 24
Genetic factors 25
Molecular mimicry 26
Aberrant expression of antigen 27
Defect immunoregulation through cytokines 28
Target organ defects 28
Superantigens 28
The TH1/TH2 paradigm 29
Apoptosis 29
Autoantigens, autoantibodies and autoreactive T-cells 30
Autoimmune polyendocrine syndrome type I 32
Addison’s disease 35
Hypogonadism 36
Alopecia 36
Intestinal dysfunction 37
Vitiligo 37
Autoimmune hepatitis 38
Pernicious anaemia 38
Insulin-dependent diabetes mellitus 39
Minor components 39
Genetics 40
Autoimmune polyendocrine syndrome type II 42
Pteridine dependent hydroxylases 44
Structure 44
Function and tissue distribution 45
TPH, PAH and TH in disease 45
CURRENT INVESTIGATION
Results 47
The identification of TPH as an autoantigen in APS I (I) 47 TPH antibodies in other autoimmune intestinal diseases (II) 49 The identification of TH as an autoantigen in APS I (III) 49 Pteridine dependent hydroxylases as autoantigens in APS I (IV) 50
Discussion 52
SUMMARY 60
FUTURE PERSPECTIVES 61
AADC Aromatic L-amino acid decarboxylase Ab Antibody
AChR Acetylcholine receptor
AIRE Autoimmune regulator (human) Aire Autoimmune regulator (mouse) APC Antigen presenting cell
APECED Autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy APS I Autoimmune polyendocrine syndrome type I
BH4 Tetrahydrobiopterin CAH Chronic active hepatitis CD Cluster of differentiation
cDNA Complementary deoxyribonucleic acid CKK Cholecystokinin
CNS Central nervous system
EAE Experimental autoimmune encephalomyelitis EC Enterochromaffin
ER Endoplasmatic reticulum GAD Glutamic acid decarboxylase HLA Human leucocyte antigen IBD Inflammatory bowel disease IDDM Insulin dependent diabetes mellitus IFNγ Interferon γ
Ig Immunoglobulin
ITT In vitro transcription and translation MG Myasthenia gravis
MHC Major histocompability complex PAH Phenylalanine hydroxylase PCR Polymerase chain reaction PKU Phenylketonurea
RT-PCR Reverse transcriptase - polymerase chain reaction SCC Side chain cleavage enzyme
SLE Systemic lupus erythematosus TCR T cell receptor
TH Tyrosine hydroxylase TPH Tryptophan hydroxylase
I N T R O D U C T I O N
General immunology
The human immune system has evolved to ensure a dynamic defence against a wide range of invading organisms. The first obstacle an invading pathogen must overcome is a surface barrier e.g. keratinized skin or an enzyme coated mucosa. Pathogens able to pass this first barrier are then met by two principally different but co-working systems: the innate and adaptive immune systems. The innate immune system is characterised by a similar response to re-invasion by the same type of invader, irrespective of the number of previous encounters. On the other hand, the adaptive immune system is characterised by its ability to strengthen the defence towards re- invasion by the same type of invader.
The innate immune system
The components of the innate immune system are immunologically active cells, complement, acute phase proteins and cytokines. The cells involved are phagocytic cells including neutrophils, monocytes and macrophages; cells that release inflammatory mediators including basophils, mast cells and eosinophils; and natural killer cells. The innate immune response is rapid and does not require cell proliferation before the intruder is attacked. The main limitation of the innate immune system is the lack of specificity. The recognising receptors used are coded by germ line genes, and the repertoire is limited to the hundreds, in contrast to 1014 – 1018 somatically recombined receptors used in the adaptive immune system (110). The structures recognised by the innate immune system are called “pathogen-associated molecular
general features of the pathogen-associated patterns are that they are specific for microbial pathogens, often crucial for the survival of the pathogen, and shared by whole classes of pathogens. Examples are lipopolysaccarides, peptidoglycans and bacteria-specific DNA. The receptors are either secreted, endocytic or signalling.
Secreted circulating receptors bind to the surface of pathogens and mark them for phagocytosis or destruction by the complement system. Endocytic receptors are expressed on phagocytes and direct the recognised pathogen to lysosomes where it is degraded. Signalling receptors induce the production of inflammatory mediators such as inflammatory cytokines, when they encounter their counterpart. In contrast to the adaptive immune system, the innate immune system cannot recognise self-structures.
This feature of the innate system is used as a control mechanism in the initiation of an adaptive immune response. The initial activation of a T-cell requires two signals: the recognition by the T-cell receptor of a peptide-MHC complex, and a co-stimulatory signal, such as B7.1 or B7.2, controlled by the innate immune system. In this way the system ensures that a peptide can activate a T-cell response only if it is derived from a pathogen recognised by the innate immune system (111).
The adaptive immune system
The basis for the adaptive immune system is clonal proliferation of T-, and B-cells, bearing receptors specific for the triggering antigens. A refined system of somatic gene rearrangements allows the adaptive system to generate approximately 1015 different receptors, each one specific to one unique epitope. T-cell receptors are expressed on T- cells and recognise epitopes presented in complex with MHC. B-cells produce antibodies when B-cell receptors are activated, and antibodies are the soluble forms of B-cell receptors. Upon recognition of the specific epitope, the mature T-, or B-cell, undergoes clonal proliferation leading to direct cell-mediated destruction,
Figure 1. Overview of lymphocyte activation. T-cell receptors recognise processed peptides presented by MHC class I or class II. A cytotoxic T lymphocyte (CTL) express CD8, which by binding to a constant region of MHC class I restricts the CTL to interact with MHC class I expressing cells, mainly presenting intracellular proteins of viral, or endogenous, origin. An activated CTL kills the infected cell by inducing lysis.
A T-helper (TH) cell is in the same way restricted, by CD4 expression, to activation by preferrably extracellular antigens presented by MHC class II. In addition to the antigen presentation, the TH cell also requires a co-stimulatory signal, i.e. binding of B7 to CD28, to be activated. The TH cell can be a TH1 or a TH2 cell. Activated TH1 cells produce IFNγ and IL-2, which activate CTLs or macrophages. Activated TH2 cells secrete IL-4, 5 and 6 which activate B-cells. B-cells recognise antigens directly through membrane bound receptors and also need a second signal from an TH cell to be activated and proliferate into antibody producing plasma cells or memory B cells
complement mediated lysis, neutralisation or phagocytosis through opsonisation by antibodies (Figure 1) (42, 43).
B-lymphocytes
B-cells are developed throughout human life from stem cells, initially in the fetal liver, and later in the bone marrow. During differentiation the cells go through specific developmental stages in which they, through somatic recombination, rearrange the variable regions of the heavy and light chain of the B-cell receptor or antibody (VDJ rearrangement). The VDJ rearrangement is mediated by recombinases coded by the recombination-activating genes (RAG 1 and RAG 2) and is estimated to result in the generation of about 1015 unique variable regions (2, 169). These are checkpoints that ensure that genetic rearrangements are no longer possible after heavy and light chain rearrangements are completed. The end products of this development are immature B- cells leaving the bone marrow, each one expressing a unique IgM on its surface.
Immature B-cells that bind antigens to their receptors undergo clonal deletion to prevent the occurrence of mature and activated B-cells able of producing antibodies against self-structures (39). The immature B-cells enter the circulation and migrate to the secondary lymphoid tissues – mainly the spleen and the lymph nodes, where they complete their maturation. Mature cells recirculate in the periphery and the fate of each cell is dependent upon encountering its specific antigen. When an antigen binds to a membrane bound IgM, the binding of the antigen sends a direct signal to the interior of the B-cell, the antigen is then internalised and presented in a complex with MHC class II to a CD4 positive T-cell (80). The direct signal from the receptor, and signalling from the T-cell through cytokines, are often both required to activate the B- cell (133). The B-cell then proliferates to become antibody-producing plasma cells, that are predominantly located in the bone marrow or on mucosal surfaces, or as long
lived memory B-cells that are mainly found in the spleen and lymph nodes. In the proliferative stage, activated B-cells further diversify the antibody specificity by somatic hypermutation in that cells producing antibodies with the highest affinity to the antigen are positively selected (80).
Antibodies
Antibodies are the soluble form of the B-cell receptors. An antibody consists of two heavy chains and two light chains held together by disulphide bonds (46). Both heavy and light chains have hypervariable regions in their amino termini and constant regions in their carboxy termini. The hypervariable regions from one light chain and one heavy chain together form an antigen recognition site, thus there are two antigen recognition sites on each antibody. All antibodies or B-cell receptors produced by a given B-cell bear the same specificity through allelic and isotypic exclusion. There are two types of light chains, namely λ and κ. There are five different classes of heavy chains with different heavy-chain constant (C) domains, and the constant domains categorise the 5 major isotypes of antibodies. The C domain can be µ, δ, γ, ε or α, giving rise to IgM, IgD, IgG, IgE and IgA, respectively. The IgGs can be further divided into four subtypes: IgG1, IgG2, IgG3 and IgG4, and IgAs can be either IgA1 or IgA2. The isotype switching is accomplished through alternative splicing and DNA rearrangements (158). General characteristics of the different isotypes are summarised in table 1. IgM is mainly produced before somatic hypermutation has occurred and therefore has a relatively low affinity. It can rapidly be secreted into the blood as a pentamer and acts as an effective activator of the complement system. IgD has an obscure function in the immune system. It is present on the surface of B-cells and may have a role in the development and/or activation of the B-cell. Very low levels of IgD
predominant antibody isotype found in the circulation. It is an efficient opsonisator and activator of complement. It can also pass the placental barrier and transmit immunological properties to the foetus before it has started its own antibody production. IgE is primarily found beneath the skin and mucosa where it acts as an activator of mast cells which induce a fast and powerful immune response, mainly through the release of histamine. IgA has the ability to associate with a J chain and form dimers that can pass through epithelial surfaces in the gut, bronchi, mammary glands etc. The secreted IgAs can inhibit the attachment of infectious agents to the epithelium and form a line of defence against pathogens from the outside (80).
Function IgM IgD IgG1 IgG2 IgG3 IgG4 IgA IgE
Neutralisation + - ++ ++ ++ ++ ++ -
Opsonisation - - +++ ++/- ++ + + -
Activation of NK cells - - ++ - ++ - - -
Activation of mast cells - - + - + - - +++
Complement activation +++ - ++ + +++ - + -
Distribution
Transport across epithelium + - - - +++ -
Transport across placenta - - +++ + ++ +/- - -
Diff. into extravascular sites +/- - +++ +++ +++ +++ ++ + Mean serum level (mg/ml) 1.5 0.04 9 3 1 0.5 2.1 3x10-5 Table 1. Characteristics of antibody isotypes and IgG subtypes.
T-lymphocytes
Immature T-cells migrate from the bone marrow to the thymus where they step-by- step differentiate into mature cells. These mature cells are efficient in recognising and reacting against non-self peptides, but also avoid attacking self structures. In this
intriguing process, referred to as positive and negative selection of T-cells, the small fraction of cells that survive are saved from apoptosis twice. First they are saved by their T-cell receptor’s ability to recognise self-MHC, and then by the same receptor’s inability to recognise self-peptides presented in a complex with MHC (Figure 2).
The T-cell receptor
Membrane bound receptors are expressed on the surface of T-cells to ensure the specificity of the T-cell response. The T-cell receptor belongs to the immunoglobulin superfamily of receptor molecules and shares many structural characteristics with B- cell receptors, examples include subunits divided into variable and constant regions, and genetic rearrangements responsible for antigenic variability (57). There are, however, two major differences. Immunoglobulins recognise native antigens in extracellular spaces while T-cell receptors only recognise processed peptides presented in complex with MHC on cell surfaces. T-cell receptors only exist in a membrane bound form while immunoglobulins can either be secreted antibodies or membrane bound B-cell receptors.
The predominant T-cell receptor is an heterodimer consisting of two transmembrane glycoprotein chains, α and β, linked by a disulphide bond. Although there are reports of T-cells that, through incomplete allelic exclusion, express T-cell receptors with more than one specificity (129), in principle all T-cell receptors on a given cell recognise the same peptide-MHC complex. An alternative γ/δ-T-cell receptor is expressed on a minority of T-cells found mainly in epithelial tissue in the epidermis and small intestine. The γ/δ-receptor differs from the α/β-receptor in that it seems to be able to recognise an antigen directly without the presence of a MHC molecule. The physiological role of these γ/δ-T-cells is still unclear (24).
In the same manner as B-cell receptors, the α-, and β-chains of the T-cell receptor are coded by sets of genes which, during T-cell differentiation in the thymus, are somatically recombined to form functional genes. The variable region of α-chain is generated by ~70 Vα-segments rearranged to ~60 Jα-segments. The β-chain gene is a result of the rearrangement of ~50 Vβ-, 2 Dβ-, and 13 Jβ-segments (42). The variability is highest in the CDR3-region of the T-cell receptor, forming the centre of the antigen binding groove, responsible for peptide recognition, and is lower in the flanking MHC recognising parts (56). In contrast to immunoglobulins, T-cell receptors do not undergo somatic hypermutation. This lowers the risk that T-cells, having passed negative selection, mutate into self-reacting cells. This may also be functional, in the sense that T-cell receptors must retain their ability to recognise MHC to be able to stimulate an immune response.
The function of a T-cell is not only determined by the nature of the T-cell receptor expressed, but also by the expression of CD4 or CD8 co-receptor molecules. A mature T-cell is either expressing CD4 or CD8 associated with the T-cell receptor on the cell surface. CD4 and CD8 exclusively bind to invariable parts of MHC class II and I, respectively. The expression of CD4 or CD8 thus restricts the T-cell to interact with peptides presented in a complex with either MHC class I or class II. Whether a T-cell should express CD4 or CD8 is determined at the end of the T-cell differentiation in the thymus (184).
Figure 2. Positive and negative selection in the thymus. Immature, CD4+CD8+
T cells are predestined to apoptosis, and saved if they recognise MHC presented by cortical epithelial cells. During this “positive selection”, 95% of lymphocytes are eliminated. The surviving cells are challenged, in the medulla, by the presentation of self peptides by dendritic cells, or macrophages. T cells that bind self peptides with too high affinity are eliminated in this “negative selection”. The remaining fraction of cells are exported to the periphery as CD4 or CD8 positive T cells (From: Delves and Riott, NEJM, 2000; 343(1); 37-49).
Progenitor T-cells leave the bone marrow and enter the thymus at the edge of the cortex as “double-negative” cells, lacking both CD 4 and CD 8 (CD 4- 8-) and a rearranged T-cell receptor on the cell surface. The differentiation is initiated by the rearrangement of the β-chain gene. When the β-chain is rearranged and expressed on
the surface, the cells becomes “double positive” (TCRβ CD4+ CD 8+), and the rearrangement of the α-chain is started. When the complete T-cell receptor is expressed on “double positive” cells (TCRαβ CD4+ CD 8+), the cells are destined to apoptosis if they are not saved in the process of positive selection, by the binding of the T-cell receptor to MHC expressed on cortical epithelial cells (173). Ninety-five percent of pre T-cells die in the thymus at this stage. Dendritic cells and macrophages in the medulla then challenge the remaining cells for self-antigens bound to MHC. The cells with T-cell receptors with high affinity to these self-antigens are directed towards apoptosis (124). The remaining small fraction of cells then, depending on the preference for MHC class I or II, cease to express either CD4 or CD8. The final result, after approximately three weeks of development in the thymus, is the export to the periphery, of single positive CD 4 or CD 8 expressing cells. These cells have a rearranged T-cell receptor able to recognise MHC class I or II, but unable to interact with self peptides in complex with MHC (Figure 2).
The major histocompability complex
All nucleated cells in the body express the major histocompability complex (MHC) antigens on their surface. MHC antigens present processed pathogen-derived, or endogenous, peptides for T-cells (87, 88). There are two classes of MHC: class I and class II. MHC class I present peptides for CD 8 positive T-cells and MHC class II present peptides for CD 4 positive T-cells. MHC class I consists of two subunits, one membrane bound α-chain with three domains, α1, α2 and α3, and β2-microglobulin.
MHC class II molecules have one α, and one β chain, each with two domains called α1/α2 and β1/β2, respectively. Human MHC antigens are also called human leukocyte antigens (HLA). All human MHC, with the exception of β2-microglobulin, are coded
by genes clustered on chromosome 6. A number of other proteins engaged in processing and presentation of peptides are also encoded by genes in the MHC region.
HLA is polygenic and polymorhpic in that there are multiple genes for each HLA locus, and there are up to as many as 200 possible alleles for each locus. Three genes encode MHC class I (A, B and C), and four sets of genes encode MHC class II (DP, DQ, DR1 and DR2). In addition, most individuals are heterozygous at HLA loci, because the alleles on both chromosomes are seldom identical. This results in a total of six different class I molecules and eight different class II molecules expressed in one individual. The diversity is even higher if class II α and β chains from different chromosomes are taken into account. The combination of HLA alleles expressed is called the HLA haplotype. The HLA haplotype determines the range of peptides that can be presented in complex with MHC and this is probably the basis for the association between HLA haplotype and increased risk for different autoimmune diseases (142).
MHC class I
MHC class I is expressed on all nucleated cells, and presents peptides generated within the cell (Figure 3) (87). Endogenous or viral proteins are degraded in proteasomes and are loaded into MHC class I in the endoplasmatic reticulum by two proteins:
“Transporters associated with Antigen Processing” -1 and –2 (TAP-1 and TAP-2). The MHC class I-peptide complex is then transported through the Golgi apparatus to the cell surface and presented for CD 8 positive T-cells (60). The α1 and α2 domain of the α-chain form a peptide-binding groove where the peptide is presented. The groove is closed at the ends restricting the length of the presented peptide to 8 – 10 amino acids (106). Different allelic variants of MHC class I have different preferences to the
Figure 3. Antigen processing. The upper part of the figure illustrates the processing of antigens presented by MHC class I. Viral or endogenous proteins are degraded in proteasomes into peptides. Peptides are transported to ER where they are loaded into MHC class I, and the complexes are transported through the Golgi apparatus to the surface of the cell. The lower part shows the processing of antigens presented by MHC class II. Extracellular proteins enter the cell through endocytosis, or phagocytosis, into early endosomes (a). MHC class II are synthesised in ER and transported through Golgi into primary lysosomes (b), which fuse with endosomes and MHC class II compartments (c) are formed. In MHC class II compartments the proteins are degraded, peptides are loaded into MHC class II, and the complexes are exported to the surface (From: Klein and Sato, NEJM, 2000; 343(10); 702-709).
for the binding. These so called anchor residues are similar in all peptides bound to a given MHC class I molecule. The MHC class I-peptide complex is finally presented to CD 8 positive cytotoxic T-cells resulting in cell death through the release of cytotoxic effector proteins e.g. pore forming perforin and apoptosis inducing gramzymes from
the cytotoxic T-cell. The mechanism is primarily designed to serve as a defence against viruses and intracellular bacteria.
MHC class II
MHC class II is expressed on a subgroup of immune cells including B-cells, activated T-cells, macrophages, dendritic cells and thymic epithelial cells. Other cells have also been found to present MHC class II in the presence of interferon-γ (87). While class I molecules present peptides derived from cytosolic proteins, MHC class II present peptides from extracellular pathogens, degraded in endocytic vesicles (Figure 3).
MHC class II chains are synthesised separately in the endoplasmatic reticulum where they are brought together and stabilised by the invariant chain protein, which binds the peptide binding groove and thereby blocks premature peptide binding. The MHC class II-invariant chain complex is transported in vesicles that eventually fuse with endocytic vesicles, containing degraded extracellular proteins, where the invariant chain is released and a peptide is loaded into the groove (26). The peptide binding cleft of MHC class II shares many features with the corresponding cleft on MHC class I, but one important difference is that the cleft is open in its ends, allowing longer peptides to be presented. In addition, the MHC class II groove has pockets where anchor residues determine whether a peptide is allowed to be presented by a specific allelic variant of MHC class II (28). The result of the presentation of the MHC class I- peptide complex for CD 4 positive T-cells is dependent on whether the T-cell is an TH1 or TH2 cell (117). Mainly through the secretion of interferon-γ, IL-2, TNF-α and TNF-β, TH 1 cells activate macrophages to destroy intracellular micro-organisms, initiate T-cell mediated cytotoxicity, and activate B-cells to produce opsonizing antibodies. The TH1 response is classically demonstrated by the delayed type
hypersensitivity reaction. Activated TH2 produce IL-4, IL-5, IL-10 and IL-13, which in turn, activate B-cells to produce immunoglobulins of different isotypes (145). TH2 cells also activate mast cells and eosinophils. TH1 and TH2 cells are thought to arise from a common precursor. The course of differentiation is determined by the genetic background, the physical form of the immunogen, the type of adjuvant, the dose of antigen and the co-stimulatory factors present at the time of antigen presentation (154).
Tolerance
The combination of a potent immune system with an ingenious capability to generate an enormous range of B- and T-cell reactivity along with thousands of self peptides displayed requires the immune system to have a number of mechanisms to prevent reaction with self structures. Central tolerance develops in the fetus and is based on the negative selection of developing T-, or B-cells in the thymus or bone marrow, where cells reacting with self antigens undergo apoptosis (113). There is a need for a second line of defence against self reactivity. This is called the peripheral tolerance and it is needed as cells can escape central tolerance mechanisms. These cells are reactive to self antigens that are of low affinity, or expressed at low concentration, or not at all in the thymus or bone marrow. Peripheral tolerance is developed postnatally and the general principles are the need for dual signals to activate T- and B-cells, and low or absense of exposure to the antigens (172). B-cells require co-stimulation by T-helper cells to become activated and thus rely on T-cell tolerance, or a “hole” in T-cell reactivity, for their own tolerance. An autoreactive B-cell that does not receive a co- stimulatory signal from a TH-cell becomes anergic. There are also Fas ligand expressing T-cells with specificity for self peptides that induce apoptosis in self reactive T-, or B-cells (172). Ignorance of self antigens by autoreactive T-cells can occur if:
1 . The antigen is anatomically sequestered behind an endothelial barrier e.g. the blood-brain barrier.
2. The antigen is presented in amounts too low to be detected.
3. The antigen is presented on cells with few or no MHC molecules.
T-cells can also become anergic if the antigen is presented by a non-professional APC lacking the ability to produce a co-stimulatory signal e.g. B7.1 or B7.2. The central and peripheral mechanisms of tolerance are believed to be our main protection against autoimmune disease (103).
The “Danger” hypothesis
An alternative model of viewing immune regulation and initiation of immune responses is called the danger hypothesis (7, 108). This model recognises the deletion of self reactive lymphocytes in the thymus, but does not see the self/non-self discrimination as the on/off-button. Instead it stresses the importance of APCs in the periphery as gatekeepers in the immune system. Autoreactive T-cells need a second signal from an activated APC to be activated. The APC, according to this model, becomes activated by endogenous alarm signals if the tissue is in danger. These alarm signals can be of two principal types: “pre-packaged” as the extracellular exposition of intracellular cell components e.g. DNA, RNA, mitochondria, or inducible as e.g. heat shock proteins and IFNα. The type of alarm signal determines which immunological effector response a given stimuli will provoke.
Autoimmunity
In a normal physiological state a number of autoreactive T-, and B-cells circulate in the body without causing disease. Transient autoimmune reactions are also common especially in connection with infectious disease or trauma. Autoimmune diseases arise when the immune system turns itself against self structures in a sustained and uncontrolled way, resulting in tissue damage or direct pathogenic action by autoantibodies. The classical definition of an autoimmune disease, formulated in 1957 (181), includes four criteria:
1. The existence of an autoantibody or cell mediated immunity.
2. The identification of the corresponding antigen.
3. The induction of disease in an experimental animal by immunisation with the antigen.
4. The transfer of disease to a healthy individual by transfer of T-cells, B-cells or autoantibodies.
Revised criteria have been formulated, but in principle these original criteria are still valid (149). Autoimmune diseases are common, affecting around 3 percent of the population, women are more frequently affected (3:1 sex ratio) (79). The autoimmune diseases can be divided into three classes: systemic disease e.g. systemic lupus erythematosus (SLE), organ specific destructive disease e.g. insulin-dependent diabetes mellitus (IDDM) and organ specific non-destructive disease e.g. myasthenia gravis. Systemic autoimmune diseases are characterised by autoreactivity against cell constituents present in most cells in the body, e.g. antinuclear antibodies, and a heterogeneous clinical presentation often mediated through immune complexes. Organ
specific destructive diseases often display tissue specific intracellular enzymes as autoantigens, e.g. glutamic acid decarboxylase (GAD) or 21-hydroxylase. The disease leads to a destruction of target cells which is probably cell mediated. Organ specific non-destructive diseases show autoantibodies against tissue specific cell surface receptors, e.g. acetylcholine receptor antibodies, directly driving the pathophysiological process, and the target tissue is not destroyed. The disease discussed in this thesis, APS I, is a typical organ specific destructive disease, with a number of target organs. The mechanisms underlying autoimmune diseases are largely unknown, but some possible mechanisms and risk factors will be discussed below.
Genetic factors
Many autoimmune diseases, especially organ specific diseases show a familial clustering. Any given individual affected by one autoimmune disease has an increased risk of being affected by a second autoimmune disease. Different HLA alleles can be protective or increase the susceptibility to different autoimmune diseases. For instance the DRB1*0401 allele increases the risk of getting rheumatoid arthritis sevenfold (96), and the DQB1*0302 allele increases the risk of getting IDDM. With the exception of the strong association between HLA-B27 and ankylosing spondylitis, most disease associations have been found with different class II loci (162). It is difficult to determine exactly which gene is responsible for a particular disease association as many genes in the MHC region are tightly linked and combinations of alleles are needed to increase susceptibility (121). The HLA associations are, in principle, understandable as different HLA alleles have varying preferences when it comes to
“choosing” what peptides to present for T-cells, both in the development of central tolerance in the thymus and in the presentation for effector T-cells in the periphery.
Substitution of one single amino acid in the peptide binding groove has been shown to
have major influence on peptide binding (95). Other genes encoded in the MHC complex that have been shown to increase the risk of autoimmunity are genes encoding complement components, especially C1, C2 and C4, in SLE (51, 120). Genes outside the MHC complex are also thought to contribute to the genetic component in autoimmune disease. Association studies in human IDDM have resulted in a number of candidate genes, among them the gene encoding insulin (82). The genetic components in most autoimmune diseases seem, however, to be very complex and gene polymorphisms coupled to disease are to be seen as permissive rather than causing the disease. One exception is APS I, which is a monogenous autosomal recessive disease with total penetrance and as such, can be of value to study mechanisms underlying autoimmune disease in general.
Molecular mimicry
The basis for molecular mimicry in autoimmune disease is that a foreign infectious particle, or food constituent, has a linear or conformational epitope that is homologous with a self antigen. The foreign particle must be structurally similar enough to share determinants, but dissimilar enough to provoke an immune response. Several mechanisms for the autoimmune reaction are possible (6, 103):
1. A viral antigen is presented in complex with MHC class I to a cytotoxic T-cell.
The cytotoxic T-cell is then activated, clonally expanded and can attack other cells presenting a self-peptide that is homologous to the viral epitope.
2. An ignorant or anergic TH cell with self specificity is activated by a cross reactive foreign peptide presented in complex with class II and co-stimulatory factors by a professional APC.
3. The foreign antigen activates a B-cell which proliferates and cross reacts with a homologous self antigen.
4 . The foreign antigen activates a B-cell which proliferates, undergoes somatic hypermutation leading to epitope spreading and cross reacts with a homologous self antigen through epitope spreading.
Once the tolerance is broken by molecular mimicry, the autoimmune process can continue without the presence of the foreign particle, as tissue damage leads to a non- physiological exposition of self peptides. A number of homologies between viral or bacterial epitopes and human autoantigens have been described (126). For example:
the homology between the AChR α subunit (the predominant autoantigen in myasthenia gravis) and herpes simplex virus glycoprotein D (152), and the shared determinants of GAD and the P2-C protein of coxackie B virus (11).
Aberrant expression of antigen
Intracellular self-antigens, or antigens normally presented in immunologically privileged compartments e.g. brain or testis, can under special circumstances be accessible to the immune system. Infections, trauma or inflammation can break down vascular or cellular barriers. In sympathetic ophthalmia, trauma against one eye releases intraocular proteins and an autoimmune reaction against both eyes follows (159). Another example is that mice infected with coxsackie virus with a tropism for pancreatic islets develop a chronic autoimmune inflammation against islet cells which is not related to molecular mimicry (72). Aberrant MHC class II expression in target cells normally only expressing MHC class I, e.g. due to stimulation by inflammatory cytokines, may lead to the presentation of intracellular self peptides for CD 4 positive TH cells.
Defect immunoregulation through cytokines
The peripheral tolerance relies on a number immunoregulatory mechanisms leading to ignorance of self epitopes. Potentially self reactive TH-cells are kept ignorant by TS- cells, IL-10, TGF-β and prostaglandins. Cytokine imbalance has also been shown to promote autoimmunity. An experimentally deleted or mutated IFNγ gene in pancreatic β-cells decrease the autoimmune destruction of the β-cells (175), and in IDDM patients an increased expression of IFNγ has been demonstrated (73). SLE patients, as well as their healthy relatives, have a defect IL-10 regulation of TS-cells (64).
Target organ defects
Pre-morbid abnormal target organ function or aberrant MHC expression are considered to be related to an increased risk of developing an autoimmune disorder.
Animal strains with high susceptibility for autoimmune disease, e.g. EAE receptive animals, have been shown to upregulate MHC class II expression in target organs (156). Obese strain chickens, prone to autoimmune thyroiditis, have a defective iodine uptake in their thyroid gland. This has been experimentally shown to be a permissive trait to develop thyroiditis (160).
Superantigens
Superantigens are proteins produced by pathogens that have the ability to activate CD 4 positive T-cells by binding both the outer faces of MHC class II and the Vβ domain of the TCR. A superantigen has a relative specificity for certain Vβ chains, and is capable of polyclonally activating 2 – 20 % of all T-cells. It has been proposed that bacterial or viral superantigens may have a role in the development of rheumatoid arthritis by activating anergic self reactive T-cells (130).
The TH1/TH2 paradigm
Studies, mainly those involving different animal models, have shown that organ specific destructive autoimmune diseases have a predominant TH1 cytokine profile, while systemic diseases show a bias towards TH2 (146). Manipulation of TH differentiation has resulted in a changed cause of the disease, e.g. in EAE where addition of an anti-B7.1 antibody reduced the incidence, while anti-B7.2 increased the severity of the disease (94). It has been proposed that an imbalance in the TH1/TH2 response may be a prerequisite for autoimmune diseases to evolve. This can open the way for new therapeutic strategies by modulating the TH1/TH2 balance (146).
Apoptosis
Apoptosis, or programmed cell death, is a mechanism for controlled cell destruction which, in contrast to necrosis, occurs without inflammatory responses. Apoptosis is mediated via Fas and its’ ligand, Fas L, in the immune system and has an important role in the modulation of the immune response. Defect regulation of apoptosis could have a possible role in autoimmunity by several mechanisms (33), (30):
1. Defect elimination of autoreactive lymphocytes.
2. Aberrant cleavage of intracellular autoantigens can result in the presentation of new, not tolerated, epitopes.
3. In contrast to apoptosis, defective apoptosis can lead to a bias towards perforin mediated cell destruction which promotes autoimmunity through an uncontrolled exposition of intracellular material.
The existence of Fas mutations in patients with autoimmune lymphoproliferative syndrome (52) supports the role of apoptosis in autoimmune disease. Experimental support also exists, e.g. the very low incidence of diabetes in perforin-negative NOD mice (83).
Autoantigens, autoantibodies and autoreactive T-cells
Self structures identified through reactivity with autoantibodies or self reactive T-cells, associated with autoimmune diseases, are called autoantigens. A large number of B- cell autoantigens have been identified through immunohistochemistry, immunoblotting and screening of cDNA libraries. T-cell autoantigens are harder to identify and it is difficult to verify the association with a disorder in a large patient sample, mainly due to the MHC restriction of T-cell reactivity. The effector mechanisms in autoimmune diseases are mediated both through autoantibodies and through T-cells. They can be classified in analogy with type II, III and IV hypersensitivity reactions (38). Type II reactions are antibody mediated with cell surface antigens resulting in phagocytosis or complement mediated lysis, as in autoimmune haemolytic anaemia (45), receptor mediated stimulation as in Graves’
disease (14) or receptor blockade as in myasthenia gravis (101). Type III reactions are immune complex mediated with extracellular antigens, matrix derived or soluble, and can be exemplified by SLE (92). Type IV reactions are T-cell mediated, organ specific destructive diseases and the antigens proposed are often intracellular as GAD in IDDM (12) and 21-OH in Addison’s disease (180).
The role of autoantigens in the aetiology of autoimmune diseases, and the value of their identification, is not obvious. B-cell activation and subsequent autoantibody production requires activated TH-cells. The specificity of these TH-cells and the
specificity of the autoantibody do not necessarily have to be the same, even if spatial and time factors suggest a shared antigen. In myasthenia gravis, both autoantibody specificity and T-cell epitopes are predominantly found within the same region of the α subunit of AChR (140). Insulin and GAD65 are both reported as B-, and T-cell autoantigens in IDDM (13, 119, 131, 132).
The value of a number of antibodies as disease-specific diagnostic markers is undisputed. The occurrence of autoantibodies often precedes the clinical onset of the disease (23), and can thus be used to screen persons at risk of developing disease. The titres of some autoantibodies, e.g. anti-dsDNA antibodies in SLE (167), are correlated to disease activity, while others are of less value for monitoring the disease.
Autoimmune polyendocrine syndrome type I
Autoimmune polyendocrine syndrome type I (APS I; OMIM acc no. 240300), also known as autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED), is an organ-specific autoimmune disease. The disease affects multiple organs (Table 2) and is characterised by circulating tissue-specific autoantibodies (Table 3).
Neufeld 1980
Brun 1982
Ahonen 1990
Zlotogora 1992
Betterle 1998
Cases 106* 166* 68 23 41
Origin Mixed Mixed Finnish Iranian Jewish Italian
Sex ratio f/m 1.4 1.7 0.8 1.1 2.4
Chronic candidiasis 78% 75% 100% 17% 83%
Hypoparathyroidism 82% 89% 79% 96% 93%
Adrenal insufficiency 67% 60% 72% 22% 73%
Hypogonadism 12% 45% 36% 36% 43%
Alopecia 26% 20% 29% 13% 37%
Intestinal dysfunction 24% 25% 18% NR 15%
Vitiligo 9% 4% 13% NR 15%
Autoimmune hepatitis 11% 9% 12% NR 20%
Pernicious anaemia 15% 16% 13% 9% 15%
Thyroid disease 10% 12% 4% 4% 10%
IDDM 2% 1% 12% 4% 2%
*a majority from a literature search; NR: not reported
Table 2. Clinical characteristics of different populations of APS I patients.
The disease is rare and can be traced back in the literature as sporadic case reports for almost 100 years (27, 37, 40, 41, 98, 136, 166). The disease was recognised as a familial syndrome in the mid-forties (161), and reported under different names during the following decades (44, 75, 116, 163, 177, 182). A more precise definition of the
syndrome, based on a large clinical review, was made in 1980 (122). APS I was then defined as the occurrence of two out of three cardinal symptoms: chronic mucocutaneous candidiasis, hypoparathyroidism and autoimmune Addison’s disease, or the occurrence of one of these three symptoms in siblings to APS I patients. APS I was separated from other polyglandular syndromes such as APS II and APS III. The wide clinical variation in APS I referring to the number of, type of and age at onset for different components, was thoroughly described in a large Finnish study of 68 patients in 1990 (5).
As will be discussed later, APS I is enriched in some genetic isolates e.g. Finland, Sardinia and among Iranian Jews. The clinical components vary in frequency between different populations (5, 17, 185). Although the clinical course of the disease is highly variable, the three main components usually develop in a specific order with candidiasis occurring before the age of 5, hypoparathyroidism before the age of 10 and finally Addison’s disease usually before the age of 15 (5, 17). Note that all patients do not develop all three main components, and that only two out of three are required for the diagnosis. In general, the earlier the age at onset of the first symptom indicates a predisposition to developing a higher number of components, and conversely, patients who have a late first manifestation of the disease are more likely to have fewer components (5).
In addition to the three main components, APS I patients develop other endocrine and non-endocrine symptoms in varying frequencies (Table 2). The different components of APS I will be discussed in detail below.
Related autoantigens Reference
Chronic candidiasis -
Hypoparathyroidism Calcium sensing receptor 100, 64 Adrenal insufficiency 21-OH (17-OH, SCC) 180 (93, 179)
Hypogonadism SCC 178
Alopecia TH 67
Intestinal dysfunction TPH 49
Vitiligo SOX9, SOX10 68
Autoimmune hepatitis CYP1A2, CYP2A6, AADC 35, 36, 58
Pernicious anaemia H+,K+-ATPase 84
Thyroid disease TPO, TG 17, 165
IDDM GAD65, ICA, insulin, IA-2 65, 89, 170
Table 3. Manifestations in APS I and related autoantigens
Mucocutaneous candidiasis
Chronic mucocutaneous candidiasis is often the first clinical manifestation of APS I and it can appear very early in childhood. The diagnosis is based on microscopic examination and culturing for Candida albicans. The candidiasis affects the nails, the dermis and the oral, vaginal and oesophageal membranes. In most patients the candidiasis responds to therapy and is restricted to a limited part of the body surface.
In rare cases the candidiasis can cause serious clinical problems. General candidiasis after immunosuppressive treatment has been reported to be the cause of death in one APS I patient (17). The occurrence of candidiasis in APS I has been seen as an evidence for a T-cell deficiency (34, 104) and experimental studies have shown defective T-cell suppressor functions (10). Other studies have shown that APS I patients lack “anti-candidial factors” in their serum (102), but that they are capable of producing anti-candidial antibodies indicating a correct B-cell response (138).
Hypoparathyroidism
Hypoparathyroidism is often the first endocrine component of APS I and usually appears before the age of 10 (mean 7.5 yr.) (5, 17). Patients with subnormal plasma calcium levels, supranormal plasma phosphate levels and normal renal function are diagnosed as having hypoparathyroidism (5). Parathyroid autoantibodies by immunofluorescence (22), and autoantibodies reacting with the extracellular domain of the calcium sensing receptor (100) have been reported. These results have not been confirmed in other studies. On the contrary, a recent study indicates that we are still lacking a good serologic test for hypoparathyroidism (164).
Addison’s disease
There are a number of causes to Addison’s disease, or adrenal insufficiency.
Historically, and world-wide, tuberculosis is probably the major cause of Addison’s disease. Other causes are metastatic malignancies, haemorrhage, sarcoidosis or rare hereditary diseases such as adrenoleucodysdrophy. In the western world, adrenal insufficiency is most often caused by an autoimmune destruction of the suprarenal cortex. The term “Addison’s disease” refers to the autoimmune variant of the disease in this paper. Addison’s disease most often occurs as an isolated disease, but can also be one component of APS I, or other syndromes. The laboratory diagnosis of Addison’s disease is defined as: low plasma cortisol, and/or low urinary cortisol, and/or high plasma ACTH concentrations, and/or defect response to acute and/or prolonged ACTH stimulation. Autoantibodies against the adrenal cortex in idiopathic Addison’s disease have been described since the 1950s (8, 21). 21-OH has been identified as the major autoantigen in Addison’s disease (180). 17α-OH and side-chain cleavage enzyme (SCC) are presented as an adrenal autoantigen in APS I patients with
different autoantigens in APS I points out 21-OH and SCC as associated with Addison’s disease in APS I (164). In APS I, the disease has an earlier age at onset (peak 10-12 yr.) compared to isolated Addison’s disease or Addison’s disease in APS II (peak 20-30 yr.) (122, 123). Historically, Addison’s disease has been a major cause of death in APS I, but with modern substitution therapy the clinical management is satisfactory. As in isolated Addison’s disease, future management will probably include substitution with androgens to women with Addison’s disease, in addition to the usual treatment with glucocorticoid and mineralocorticoid supplementation (9, 59).
Hypogonadism
Hypogonadism in APS I shows a female predominance and is mainly diagnosed at puberty, but can also be detected as secondary amenorrhea before the age of 40. The diagnosis of ovarian failure is based on amenorrhea, high levels of FSH or LH, and slow, absent or regressive pubertal development. Testicular failure is diagnosed through high base-line, or post-GRH stimulatory, levels of FSH or LH, and low testosterone levels (5). Both 17α-OH and SCC have been proposed as autoantigens associated with gonadal failure (18, 171, 178). A recent multivariate analysis of ten different autoantibodies in a large sample of APS I patients shows that SCC alone is statistically associated with hypogonadism (164).
Alopecia
Alopecia is present in APS I both as alopecia areata and as the more severe form alopecia totalis. Alopecia areata is defined as a transient patchy hair loss on the scalp.
When the hair loss covers the whole scalp and becomes permanent the condition has developed into an alopecia totalis. Alopecia totalis is twice as common as alopecia
areata in APS I (69). Autoantibodies against hair follicles have been shown, by indirect immunofluorescence, to be associated with alopecia totalis in APS I (69).
Intestinal dysfunction
Different descriptions and names such as malabsorbtion (5, 17), celiac syndrome (37, 44), constipation (29), steatorrhea and diarrhoea (41, 78, 116) have been used in the literature to characterise the intestinal dysfunction seen in APS I. This illustrates the great variations of the clinical manifestations of the gastrointestinal dysfunction seen in APS I. We define the intestinal dysfunction as periodic steatorrhea, diarrhoea or severe constipation. The dysfunction is often therapy resistant and results in marked weight loss and/or growth failure. APS I patients with intestinal dysfunction as their first manifestation of the disease generally have a high number of disease components (5). Although the malabsorbtion often is severe no specific shortages are seen. Faecal fat is often elevated. Exocrine pancreas insufficiency (176) and low levels of cholecystokinin (71, 77) have been reported. Some reports have described a covariation between intestinal symptoms and hypocalcemia (5, 71, 135). The intestinal symptoms have been explained in the literature as being secondary to hypocalcemia (71), candidial overgrowth (153), exocrine pancreas defect, or intestinal lymphangiectasia (16). Substitution with pancreatic enzymes, normalisation of serum calcium (71) and immunosuppressive treatment with cyclosporine (176) or methylprednisolone (128), has been proposed for treatment.
Vitiligo
Vitiligo is characterised by loss of pigment formation by melanocytes in the skin.
Different patterns of distribution are acro-facial, focal, segmental and generalised vitiligo. In APS I vitiligo often occurs as symmetrical depigmented patches on the
face, neck, extensor surfaces of hands, wrists and legs and in the axillae. Complement- fixing autoantibodies have been described as associated with vitiligo in APS I (137).
Recently, the transcription factors SOX 9 and SOX 10, expressed in melanocytes, have been identified as autoantigens associated with vitiligo in APS I (68).
Autoimmune hepatitis
Autoimmune chronic active hepatitis is probably the most feared component of APS I.
The clinical course is highly variable, ranging from asymptomatic changes of liver enzymes to fulminant hepatitis with a fatal outcome (5, 112). The histopathologic picture resembles that of idiopathic autoimmune chronic active hepatitis and markers of viral hepatitis are not present. Autoantibodies against cytochromes P4501A2, P4502A6 and aromatic L-amino acid decarboxylase (AADC) have been associated with autoimmune hepatitis in APS I (35, 36, 58). Treatment is based upon immunosuppressive therapy in combination with corticosteroids (112). The determination of liver enzymes should be done with regularity to promote early detection and intervention.
Pernicious anaemia
Pernicious anaemia is secondary to vitamin B12 malabsorbtion due to a shortage of intrinsic factor caused by a disturbed parietal cell function, coupled to autoimmune gastritis. Clinical features are megaloblastic anaemia and achlorhydria. Lymphocytic infiltrates and autoantibodies directed towards parietal cells have been found (116, 122). H+K+-ATPase and intrinsic factor have been described as autoantigens associated with autoimmune gastritis (84, 150).
Insulin-dependent diabetes mellitus
Insulin-dependent diabetes mellitus (IDDM) shows a great variation in occurrence between different populations of APS I patients, with frequencies ranging from 1 to 12 percent (5, 17, 122, 143). Autoantibodies against glutamic acid decarboxylase (GAD65), tyrosine phosphatase IA-2 (IA-2), insulin and/or islet cell antibodies (ICA) are present in APS I patients with IDDM (65, 89, 170). In a multivariate analysis of the clinical associations of a panel of autoantigens in APS I, only IA-2 was shown to have a predictive value (164). AADC which has been cloned from a β-cell cDNA library as an autoantigen in APS I does not seem to correlate to IDDM (76, 147), but surprisingly, it is statistically associated with autoimmune hepatitis (58).
Minor components
Ectodermal dystrophy with hypoplastic enamel of the teeth, pitted nail dystrophy and tympanic membrane dystrophy affects 77, 52 and 33 percent of APS I patients, respectively (5). Keratoconjunctivitis is seen in 8 - 41% of APS I patients as corneal opacities, bulbar conjunctival injection and corneal neovascularisation (5, 17, 109).
Autoimmune thyroid disease, mainly Hashimoto’s thyroiditis, has been reported in 2 – 13% of APS I patients (5, 17, 85). Thyroid autoantibodies are present in a high proportion of APS I patients with and without thyroid disease (17, 165). The prevalence of acquired asplenia, due to a progressive destruction of the spleen, has not been systematically examined, but has been described in a high proportion in small sample of APS I patients (55, 134). Cholelithiasis is also reported as over-represented in APS I patients (55).
identified by positional cloning and named autoimmune regulator (AIRE) (118, 168).
The AIRE gene encodes a 545 aa, 57.5 kDa protein with a conserved nuclear localisation signal, two plant homeodomain (PHD)-type zinc fingers, four LXXLL motifs, one SAND and one HSR domain (Figure 4). These features are all markers for nuclear, DNA-binding proteins suggesting a role for AIRE in transcriptional control.
AIRE has also been shown in vitro to interact with transcriptional coactivators and activate transcription (19, 139).
AIRE is expressed mainly in thymic epithelial cells and to a lesser degree also in the
spleen, lymph nodes, fetal liver and mononuclear leukocytes (20, 118, 186). Recent studies in mice also demonstrate Aire expression in a variety of tissues outside the immune system e.g. kidney, testis, adrenal glands, liver and ovary (66). The sub- cellular localisation of wild type AIRE has been described mainly as restricted to the nucleus, but a cytoplasmatic expression has also been described (20, 70, 144). In the nucleus AIRE is found in dots resembling, but not co-localising with, the nuclear bodies formed by e.g. promyelotic leukaemia protein (PML), Sp 100 or Sp140.
The function of AIRE is not yet understood. The expression pattern in man and experimental studies using murine Aire, show that AIRE is likely to be involved in the thymic selection process and so indicates a role for AIRE in the induction of self- tolerance (186).
To date, 29 different disease causing mutations of AIRE have been described which are clustered in four mutational hotspots. Although 5 – 10 % of patients who fulfil the clinical criteria for APS I lack mutations in AIRE exons, or only have one allele mutated in the exonic parts, the general agreement is that both AIRE alleles must be
mutated to cause the disease (19). A possible explanation for these APS I patients with no detectable mutations is probably that mutations do exist in promoter regions or intronic parts of the AIRE gene which have not been examined by sequencing. In Finland, Sardinia and among Iranian Jews marked founder effects are seen. 89% of Finnish APS I chromosomes have the major Finnish mutation, R257X, 92% of Sardinian APS I chromosomes have the common Sardinian mutation, R139X, and 100% of APS I patients among Iranian Jews are homozygous for a single amino acid change, Y85C (19, 148). Most mutations are, mainly through frame shifts, resulting in truncated forms of AIRE, lacking DNA-binding domains or nuclear receptor binding domains.
In experimental studies of sub-cellular localisation of mutated AIRE variants, an altered intracellular distribution pattern of the protein was found (19, 144). Almost all mutations examined, except the common Iranian missense mutation, resulted in a cytoplasmatic location of the mutated protein. In transcription activation assays truncated mutated AIRE constructs showed no, or markedly lowered, activation of transcription. The transcriptional activation ability of the AIRE construct with the common Iranian mutation was the same as wild type AIRE. The fact that the experimental behaviour of the Iranian mutation does not differ much from wild type AIRE may explain the milder clinical phenotype that can be noticed among Iranian Jewish APS I patients.
Autoimmune polyendocrine syndrome type II
The diagnosis of autoimmune polyendocrine syndrome type II (APS II) requires the presence of Addison’s disease plus autoimmune thyroid disease and/or insulin-
dependent diabetes mellitus (IDDM) (122). Other endocrinopathies and associated autoimmune diseases are also present, at low frequencies. APS II shows a female predominance (123, 165), is inherited in a dominant fashion (47), and has been shown to be associated with HLA-DR3 and -DR4 (105, 125). The age at onset for Addison’s disease is higher in APS II (peak at age 20 – 30) compared to APS I (peak at age 10 – 12) (123).
Pteridine dependent hydroxylases
Tryptophan hydroxylase (TPH; EC 1.14.16.4), tyrosine hydroxylase (TH; EC 1.14.16.2) and phenylalanine hydroxylase (PAH; 1.14.16.4) together constitute the enzyme family of tetrahydrobiopterin dependent hydroxylases. The three enzymes are closely related, share functional and structional characteristics and are thought to have evolved from a common progenitor gene (63).
Structure
The functional structures of the enzymes are homologous with an N-terminal regulatory domain, a central catalytic domain and a C-terminal tetramerization domain. The regulatory domain is specific for each enzyme with a 20% linear homology between the three enzymes. It determines substrate specificity and thus reflects the unique properties of each enzyme. The central catalytic domain, containing the active site, is highly conserved and has an overall homology of around 65%. TPH is a 230 kDa tetramer consisting of identical subunits, each with a molecular mass of 58 kDa. TH consists of four 55-59 kDa subunits, as four isoforms (TH 1 - 4) due to alternative splicing, forming tetramers with molecular masses of 204-217 kDa. PAH is found in an equilibrium between tetrameric and a dimeric forms composed of 50 kDa subunits (74). The crystal structures for the catalytic domain of TH and complete PAH have been determined (50, 91). The structures are essentially identical, and the TPH structure can be predicted, based on the structures of TH and PAH (174). The crystal structure determinations are not only useful in understanding the regulation of enzymatic activity, but may also be of great value in determining the possible
conformational epitopes involved in autoimmunity directed against these enzymes (62).
Function and tissue distribution
TPH, TH and PAH are monooxygenases, incorporating one atom of oxygen into the substrate and reducing the other atom to water, and thereby catalyze the hydroxylation of different amino acids. Tetrahydrobiopterin (BH4) supplies the two electrons required. Molecular oxygen and ferrous iron are also needed in the reaction (Figure 7).
The regulatory properties differ in detail between the enzymes, but in general they are regulated by their substrates, BH4 and phosphorylation of serines (53). TPH and TH have central roles in the biosynthesis of the neurotransmitters serotonin and dopamine (Figure 7). TPH catalyzes the hydroxylation of tryptophan into 5-OH tryptophan and is the rate-limiting enzyme in the synthesis of serotonin. It is expressed in serotonergic cells in the central nervous system and the intestine (183). TH is the rate limiting enzyme in the biosynthesis of catecholamines where it converts tyrosine into L-Dopa, and is mainly found in the adrenal medulla and catecholaminergic neurons throughout the body. PAH is primarily produced in the liver where it catalyses the conversion of phenylalanine to tyrosine. This serves two purposes: it provides an endogenous supply of tyrosine, making tyrosine a non-essential dietary component, and it is also rate limiting in the catabolism of phenylalanine (74).
TPH, PAH and TH in disease
Mutations in PAH are the most common cause of phenylketonurea (PKU). In PKU, serum levels of phenylalanine are elevated to toxic levels causing brain damage leading to mental retardation. A second result of PAH dysfunction is that tyrosine
phenylalanine and supplemented with tyrosine (48). Defects in the production, or recycling, of BH4 also leads to hyperphenylalaninaemia in combination with TPH and TH deficiencies. These conditions have a more severe clinical presentation than isolated PAH dysfunction. TH defects and polymorphisms have been coupled to bipolar affective disorders, schizophrenia and parkinsonism (61). Hitherto, no disease couplings have been made with TPH dysfunction, although theoretically TPH defects could very well be involved in affective disorders, schizophrenia, migraine, drug abuse and intestinal movement disorders (115).
C U R R E N T I N V E S T I G A T I O N
Results
The identification of TPH as an autoantigen in APS I (paper I)
Periodic gastrointestinal dysfunction represents a significant management problem in APS I, affecting 25-30% of the patients. Before this study, the pathogenesis of the gastrointestinal dysfunction was largely unknown, but explained in the literature as being secondary to hypocalcemia (71), candidial overgrowth (153), exocrine pancreas defect or intestinal lymphangiectasia (16). We undertook this study with a belief that the intestinal dysfunction, as well as the majority of the other manifestations of APS I, was of autoimmune origin. The aim of the study was to identify a potential intestinal autoantigen associated with APS I and possibly other autoimmune intestinal diseases.
A commercially available, human duodenal cDNA library was immunoscreened with sera from seven APS I patients. A positive clone was identified as tryptophan hydroxylase (TPH, EC 1.14.16.4) and was used for in vitro transcription/translation (ITT) followed by immunoprecipitation with sera from 80 APS I patients. Forty-eight percent (38/80) of the APS I patients had TPH antibodies (Ab), whereas in a large number of sera from patients with other autoimmune diseases (n=372) and healthy blood donors (n=70) no reactivity against TPH was detected. A correlation with clinical symptoms showed that 89% (17/19) of APS I patients with gastrointestinal dysfunction were TPH-AB positive, compared to 34% (21/61) without (p < 0.0001).
Immunostainings of normal human small intestine with patient sera and specific antibodies were performed. These showed that TPH-Ab positive APS I sera
specifically immunostain TPH containing enterochromaffin cells in normal duodenal mucosa. We also observed that 3/8 APS I patient sera stained goblet cells and that 3/8 patient sera reacted with Paneth cells (Figure 5). When immunostaining duodenal biopsies from APS I patients with gastrointestinal dysfunction and TPH-Abs, a total absence of enterochromaffin cells was seen. Finally, enzyme inhibition assays demonstrated that TPH-Ab positive APS I sera almost completely inhibited TPH activity at a dilution of 1:100.
Figure 5. Immunostainings of normal human duodenum with APS I showing immunoreactivity with goblet cells (left), EC cells and Paneth cells (right).
In this study we identified TPH as an endogenous intestinal autoantigen in APS I and established a statistical association between TPH-Ab and gastrointestinal dysfunction.
We also confirmed the reactivity against TPH with enzyme inhibition assays and immunohistochemical stainings of normal and affected tissues.
TPH antibodies in other autoimmune intestinal diseases (paper II)
We reported the identification of TPH as an intestinal autoantigen in APS I in paper I.
The gastrointestinal symptoms in APS I are in some respects related to both the clinical presentation and the suggested autoimmune background of more common conditions such as inflammatory bowel disease, irritable bowel syndrome and celiac disease. Autoimmune enteropathy in children and paraneoplastic pseudoobstruction of the gut are two rare disaeses with clinical or pathogenic features related to the intestinal disease in APS I. In paper II, we wanted to find out if TPH-antibodies are present in these other intestinal diseases with a possible autoimmune pathogenesis.
Sera from 22 patients with ulcerative colitis, 36 patients with Crohn’s disease, 47 patients with irritable bowel syndrome, 25 patients with celiac disease, 3 children with autoimmune enteropathy and two patients with ANNA-1 seropositive paraneoplastic pseudoobstruction of the gut were screened for the presence of TPH-antibodies using ITT and immunoprecipitation of TPH (97, 114, 155). Immunoreactivity against TPH could not be detected in sera from these patient groups. This illustrates the high specificity of TPH-antibodies as a marker for gastrointestinal disease in APS I.
The identification of TH as an autoantigen in APS I (paper III)
Ectodermal manifestations in APS I such as alopecia, vitiligo, nail and enamel dystrophy are frequent. Alopecia areata, characterized by sudden patchy hair-loss, has been reported in frequencies varying from 13 to 37% in different populations of APS I patients (5, 17, 185). In an earlier study we showed that APS I patients with alopecia have autoantibodies directed against hair follicles (69). In this investigation, we set out
