Theme leaders: Frits Koning & Peter van Veelen

Background

Autoimmune diseases are caused by a lack or loss of tolerance to autoantigens. Tolerance is normally established and maintained by three major mechanisms: deletion of autoreactive T cells in the thymus, the induction of anergy in the peripheral lymphoid organs and through the activity of regulatory circuits. The frequent occurrence of autoimmunity is a clear indication that these processes are only partially capable of maintaining tolerance. Many autoimmune diseases are HLA-associated, i.e. individuals expressing particular HLA-alleles have an increased risk of developing disease. The most likely explanation is that such HLA-molecules preferentially bind and present peptides of (modified) autoantigens which leads to the induction of inflammatory T cell responses and disease. Although the disease initiating events are still largely unknown, evidence that PTM of proteins generates neo-epitopes that facilitates the breaking of tolerance now comes from two independent lines of research.

RA is a common, chronic inflammatory disorder affecting joints and can have severe disabilitating consequences (for reviews see ref. 17 and 23). In the past decade pivotal pathophysiological insight has been obtained by the identification of antibodies that are directed against citrullinated proteins and hence are called Anti-Citrullinated Protein Antibodies (ACPA). ACPA are highly specific for RA and directed against antigens that are also expressed in the inflamed joint. ACPA can sometimes be detected up to ten years before disease development and their presence in undifferentiated arthritis patients predicts progression to RA as well as severe clinical outcome. The relevance of these recent insights was further boosted by the demonstration that infusion of ACPA exacerbates arthritis in animals. An uncontrolled immune response to posttranslational modified proteins thus plays a pivotal role in disease development.

Similarly, in CD, a common food induced disorder of the upper gastrointestinal tract,  PTM of gluten is required for the development of full blown disease (for review see ref. 1). The gluten proteins present in wheat are partially degraded by enzymes in the gastrointestinal tract. Subsequently, these peptides are a substrate for the enzyme tissue transglutaminase (tTG) present in the intestine which converts glutamine residues into glutamic acid. As the result of the introduction of this negative charge, these peptides can bind with high affinity to the disease predisposing HLA-DQ2/8 molecules and trigger inflammatory T cell responses leading to disease. Moreover, in analogy with RA, CD patients uniquely generate antibodies reactive with these modified gluten peptides. Strikingly, patients also generate antibodies to the enzyme tTG itself, pointing to an autoimmune component in the disease process. Moreover, it is well established the CD patients have an increased risk of developing several other autoimmune diseases, raising the possibility that the elevated tTG activity in CD patients generates PTM autoantigens that trigger disease causing immune responses.

Type I diabetes (T1D) is another common HLA-associated disease where autoreactive T cell are thought to be responsible for destruction of the insulin-producing beta-cells in the pancreas (for a recent commentary see ref. 28). Similar to RA and CD, responses to PTM-proteins may play a crucial role in the development of autoimmunity in this disease as well.

Together, these results suggest that PTM of protein antigens is a crucial step towards the development of HLA-associated autoimmune diseases as the formation of neo-epitopes facilitates the generation of T cell responses to such modified proteins, leading to inflammation and the induction of autoantibodies that further perpetuate the disease process.

The aim of the current project is to obtain a comprehensive knowledge on the extent to which posttranslational modification of protein antigens plays a role in the initiation and perpetuation of autoimmune diseases, with a focus of RA, CD and T1D.

Approach

Rheumatoid arthritis. Protein citrullination is known to play a crucial role in the RA-specific immune response but hardly anything is known on the identity of the antigens initially responsible for the induction of the ACPA and the identity of the citrullinated proteins in inflamed joints. To identify these proteins we will take two approaches. In the first, citrullinated proteins and peptides will be isolated from synovial tissue samples from RA and control patients using monoclonal antibodies recognizing a broad spectrum of citrullinated proteins. In the second approach we intend to  identify the citrullinated antigens primarily recognized by ACPA using pre-disease serum from (still) healthy ACPA-positive subjects from native American communities in Canada (in which the incidence of ACPA-positivity and RA is very high). We will use purified ACPA to isolate citrullinated proteins from relevant tissue homogenates, including RA-synovium. In both approaches the identity of these proteins/peptides and the positions of modified residues (citrulline as well as other modified amino acids) will be determined by mass spectrometry.

The well established association between HLA-DR4 and RA indicates that an autoantigen specific T cell response likely has preceded the ACPA response. This is corroborated by the observation that IgG ACPA is the predominant isotype and T cell help is necessary for isotype switching. It is thus conceivable that T cells recognizing citrullinated epitopes exist that provide help to citrulline-specific B cells. Therefore, we will test T cell responses against citrullinated proteins and peptides. Moreover, as it is now possible to obtain a comprehensive inventory of the peptide ligandome bound by HLA-molecules through mass spectrometry, we will characterize the peptide repertoire present in HLA-DR4 molecules isolated from patient derived material and compare this with a similar repertoire isolated from control material in order to determine if PTM-peptides are naturally bound by these molecules in RA.

Type 1 diabetes. Although there is at present no conclusive evidence that PTM plays a role in the development of T1D, it is conceivable that PTM could lead to a breach of tolerance in this disease as well. Indeed, proof of principle exists that T-cells in T1D recognize an insulin epitope after formation of a disulfide bond between adjacent cysteine residues at 6 and 7 of the insulin A-chain (Mannering, J Exp Med 2005). Therefore, we will determine if PTM-peptides can be found in association with disease predisposing HLA-molecules. As it is essentially impossible to obtain patient-derived material for this approach we will study this in the well established NOD mouse model for T1D as these mice express a mouse MHC-class II molecule, I-Ag7, that has peptide-binding properties that closely resemble that of HLA-DQ8, the HLA-molecule that predisposes to T1D in humans. For this purpose we will purify I-Ag7 molecules from mice at various stages of disease development and determine the nature of the peptide repertoire bound.

Celiac Disease. While it is well known that tTG deamidates gluten peptides involved in CD, it is unknown if the increased tTG activity in the small intestine of patients also leads to modification of autoantigens which could lead to autoimmune responses. tTG mediates two types of modifications: deamidation and crosslinking. The former can be detected by mass spectrometry, the latter by antibodies specific for the crosslink after which the proteins can be identified by mass spectrometry. We will therefore determine if such PTM’s can be detected in biopsy material of CD patients and controls.

Technology development/mass spectrometry. The known modifications in CD (deamidation of glutamine residues leading to glutamic acid, Δm=+1), and in RA (deimination of arginine residues leading to citrulline, Δm=+1), can in principle be analyzed in a straightforward manner. A complicating factor is the interference of the isotope patterns of the modified and unmodified peptides. Both deamidation and deimination lead to a change in charge. Therefore, chromatographic and electrophoretic methods for separation of the modified and unmodified peptides can be optimally used, including peptide IEF of modified proteins and HLA-ligands. In case of incomplete separation, high resolution measurements (LTQ-FT ultra available) can handle these differences. Targeted labeling, including an affinity tag, of citrullinated HLA-ligands will be developed. In CD the action of tTG can, next to deamidation, produce neoepitopes via an isopeptide bond at susceptible glutamine residues. E.g. a residual lysine residue can be screened for, both in the MS2 spectra and, in a very sensitive manner, by parent ion scanning (feasibility tested successfully in Utrecht/NPC). A collaboration with the NPC for this triple quadrupole application would be of great value. Moreover, monoclonal antibodies specific for the crosslink are available.

In RA, the formed citrulline residue represents a unique amino acid residue mass (157 Da) and this mass difference can be screened for in database searches. In addition, a tryptic cleavage site is lost, which leads to a different digest pattern, which can be picked up. As yet unknown modifications will be screened for by inclusion of these in the database search routines.

In addition, we will generate novel reagents to determine PTM. To generate new tools for the isolation of post-translationally modified proteins from crude samples, two approaches will be taken. First, RNA aptamers will be selected from randomized libraries using synthetic peptides containing a specific modified amino acid (e.g. citrulline, methylated arginine, acetylated lysine). To obtain aptamers that are not, or only to a limited degree dependent on the flanking amino acids, the reiterated screening will be performed with glycine-rich peptides or with a combination of different peptides sharing the specific modified amino acid. The selected aptamers will be optimized by directed mutagenesis to increase their affinity for the target (if required) and to increase their stability (5’ and 3’ end stabilization; replacement of phosphates in the backbone). The applicability of these PTM-specific aptamers will be tested by using them as affinity-resins for the isolation of modified proteins from autoimmune patient tissue samples. Secondly, for certain modifications donor analogues will be generated that can be used by the protein modifying enzymes to attach (part of) the analogue to the substrate protein. These analogues will contain a moiety that allows their attachment to detection/purification modules via click chemistry. For example, a clickable lysine analogue (e.g. azido-cadaverine) will be used for the detection and isolation of substrates for transglutaminase. Similarly, S-adenosylmethionine analogues will be generated to detect and isolate protein methyltransferase targets. The development of such analogues will provide new tools to localize and identify targets of modifying enzymes.

Deliverables

• Inventory of proteins and their post-translational modification state in diseased tissues of RA patients.
• Identity of citrullinated autoantigens present in inflamed RA patient tissues.
• Identification of the HLA-DR4 bound peptide repertoire in RA patients and PTM’s in the se peptides.
• Identification of PTM’s induced by tissue transglutaminase.
• Identification of the I-Ag7 bound peptide repertoire in NOD mice at various stages of disease development and PTM’s in these peptides.
• Aptamers specific for PTMs in a peptide context.
• Clickable donor molecules for protein modifying enzymes.