Tissue Transglutaminase: Master Regulator of Celiac Disease?

In 1997, Dieterich and colleagues (1) described that tissue transglutaminase (tTG) is the target of the endomysium specific autoantibodies that are characteristic for celiac disease (CD). As the result of that seminal observation, a series of studies (2–6) have revealed that screening for the presence of tTG-specific antibodies is a very specific indicator for disease and an article in the current issue of the JPGN strengthens that view (7). In addition, there is now overwhelming evidence that tTG is not only linked to the humoral response but is also involved in generating gluten peptides that stimulate T cells present in the small intestine of CD patients (8–13). Due to the activity of tTG glutamine residues in gluten-derived peptides are converted into glutamic acid. This introduction of negative charges results in peptides that bind with relatively high affinity to the disease-associated HLA-DQ2 or HLA-DQ8 molecules and trigger inflammatory T-cell responses (8–15). This provides a satisfactory explanation for the well-known fact that CD is almost exclusively found in HLA-DQ2 or HLA-DQ8 positive individuals (16). The importance of the modification of gluten by tTG has been underscored by the observation that the enzyme specificity correlates with the toxicity of cereals for CD patients (13). We found that the enzyme specificity could be combined with the requirements for peptide binding to HLA-DQ2 and that this resulted in search patterns that identified T-cell stimulatory peptides in wheat, barley, and rye but not in oats, the latter of which is considered safe for celiac disease patients (17). Truly, therefore, tTG can be considered a master regulator of CD, involved in both the humoral and cellular immune response and tightly linked to cereal toxicity. Together, these observations have raised multiple questions: why do CD patients make tTG-specific antibodies? Do we still require a small intestinal biopsy for diagnosis since tTG-specific antibodies are such a good indicator of disease? Can we design protocols for mass screening for detecting silent disease? Can we use this knowledge to develop an alternative therapy or better disease prevention? As shown by Agardh et al. in this issue of the JPGN (7), an intestinal biopsy remains the gold standard diagnostic procedure for CD, as a number of patients have false positive or negative results of their “celiac” serum antibodies. On the other hand, the determination of IgA-tTG antibodies in serum is an accurate instrument for screening of CD, especially when preformed in for CD genetic-predisposed individuals carrying the HLA-DQ haplotypes associated with the disease (18). However, to decide whether mass screening programs for CD are worthwhile we need more information, among others, on the natural history of CD and on the health effects and costs associated with such screening programs. It is likely, for example, that large-scale screening will identify many individuals with “silent” CD but it is unknown what the long-term benefits of this will be. Why do CD patients make tTG specific antibodies? The most likely explanation was put forward by Sollid and collegues (19). tTG can crosslink itself to gluten and these gluten-tTG complexes will be taken up by B-cells that express tTG-specific immunoglobulin on their membrane. As the result of this uptake, the gluten-tTG complex will be degraded intracellularly and gluten peptides will bind to HLA-DQ and be expressed on the cell surface. In CD patients, gluten-specific T cells will recognize this HLA-DQ-gluten peptide complex and this results in T cell “help” for the production of tTG-specific antibodies by the B cells. Since only CD-patients have measurable numbers of gluten-specific T cells, this explains why only CD patients make tTG-specific antibodies and the antibody titers drop after gluten withdrawal. Can we use our knowledge to develop an alternative therapy or for better disease prevention? It is safe to assume that the HLA-DQ2/8 restricted, gluten-specific T cell responses are the main cause of the disease. Hypothetically, we could try to intervene in this process at three levels. First, we could block the binding of gluten peptides to HLA-DQ with specific inhibitors. This may be less farfetched than it seems. Much is known about the peptide-binding properties of HLA-DQ2 molecules (14,15) which facilitates the development of a specific blocking compound that would leave the HLA-DR molecules intact. Moreover, the intestine is readily accessible for drugs. Second, we could block the activity of tTG and thus the generation of the majority of T-cell stimulatory gluten peptides. Again, the knowledge of the substrate specificity of tTG (13) will facilitate the design of specific blockers but we have to take into account that long-term blocking of tTG might have unwanted side effects. Third, we could try to block or eliminate the gluten-specific T cells. This is complicated by the fact that gluten-specific T cells are quite abundant in the intestine of patients and directed to a variety of gluten peptides (8–13,20). It is difficult, therefore, to envisage compounds that would specifically block or eliminate the gluten-specific T cells but not other T cells. Disease prevention, however, may also prove a valuable option. Several regions on chromosome 5 and 11 have been identified that show linkage to celiac disease. It is reasonable to assume that in the coming 5 to 10 years we will learn much more about the additional genes and gene products in those regions that predispose to disease. This may offer novel possibilities to identify individuals at risk before disease development and the development of alternative therapies. The development of safer food products for patients can now be envisaged since we have detailed information about the toxic sequences in gluten. This allows the rational development of tests to detect these toxic peptides in food products and we are currently developing such test systems. These can also be used to identify wheat varieties with less-toxic properties. Moreover, our results indicate that the replacement of particular key amino acids in gluten by other amino acids would yield less harmful gluten molecules. It remains to be determined, however, if such molecules still possess the typical and desired characteristics of gluten molecules. We can conclude that there has been major progress in the area of CD research in recent years. We now understand the interaction between gluten, tissue transglutaminase, HLA-DQ, and T cells and how this leads to disease and the development of tTG-specific antibodies in CD patients. A critical question that must be investigated is why HLA-DQ-restricted, gluten-specific T cells are commonly found in CD patients but not in healthy, HLA-DQ2/8 positive individuals. Furthermore, we must use our knowledge to improve the quality of life of patients and to prevent and treat disease.