Answers are grounded in this paper and cite their sources.
Patients with celiac disease (CD) have a higher risk of malignancies, in particular, non-Hodgkin lymphoma. In CD the small-intestinal intraepithelial T-cell population is increased, and T-cell lymphomas arising from these T cells form the most common primary T-cell lymphoma in the intestine. Strikingly, patients who adhere to a strict gluten-free diet have no higher risk of malignancies. Therefore, it is likely that chronic stimulation of T cells in the small intestine from the continuous presence of gluten in the diet can eventually result in malignancy. It is well established that the large majority of CD patients express the HLA-DQ2 class II molecules. The function of the HLA class II molecules is to present peptides to T cells, which initiates specific immune responses against the pathogens from which the peptides were derived. Therefore, it is assumed that the HLA-DQ2 molecules can present particular gluten-derived peptides to gut-associated T cells, leading to disease. The identification of gluten-derived peptides that can bind to HLA-DQ2, but not to other HLA class II molecules, and that stimulate T cells may thus reveal the immunological basis for the development of CD and associated malignancies. For this purpose, we have purified HLA-DQ2 from professional antigen-presenting cells by affinity chromatography and characterized the bound peptides. This is accomplished through separation of the mixture of peptides by reverse-phase high-performance liquid chromatography and through sequencing of the individual peptide peaks by Edman degradation for the determination of the amino acid sequences of the peptides. The results indicate that particular amino acids in these peptides are responsible for binding to HLA-DQ2. This so-called peptide binding motif is specific for HLA-DQ2, suggesting that DQ2 can bind a well-defined set of peptides distinct from those bound by other HLA class II molecules. These results may be useful for the identification of gluten-derived peptides that can bind to HLA-DQ2. DISCUSSION Peña: I noticed that some of the binding peptides to HLA-DQ2 are HLA class I. What does this mean? Koning: It has been observed that HLA class I-derived peptides can be found in HLA class II molecules. The explanation is that HLA class I molecules are present on the cell surface, some of which is taken up by the cell during endocytosis. The class I molecules end up in the endosomes, where the class II molecules have access to the peptides. Thus, the peptides from class I molecules will actually bind to class II molecules. In some cases ≤50% of the peptides that are bound to class II molecules may be derived from class I. Peña: The HLA-DQ heterodimer (α1 *0501, β1 *0201) associated with CD is closely linked to other genes in the same region. What is the biological significance of this phenomenon, for example, in relation to the tumor necrosis factor genes? Koning: There are different opinions about this. For example, John Trowsdale thinks that these genes are kept together because it is relevant to their function. On the other hand, we should not forget that in humans, β2-microglobulin, which is associated with class I molecules and necessary for their cell-surface expression, is not located on chromosome 6. I find it difficult to assume that these genes are kept together because of their function. Probably they would function just as well if they were scattered throughout a chromosome. Of course, if you have polymorphisms, for example, in proteases or in transporters, a linkage disequilibrium among these genes could affect the peptide pool generated, and this influences the class I and class II peptide complexes formed. Feighery: If I understood it correctly, there are some difficulties in matching up the gliadin peptides with the peptides naturally bound by HLA-DQ2. What do you think about other approaches to the same problem, for example, looking at HLA-DQ2 binding to gliadin molecules in competitive assays or, as has been reported, studying the cell binding to gliadin molecules using fluorescence-activated cell sorter (FACS) analysis? In addition, what is the possibility of studying in vivo the incorporation of gliadin molecules by cells bearing HLA-DQ2? Koning: Yvonne van de Wal, from our group, is actually trying to do some of the experiments you are referring to. For example, we have cultured about thirty billion HLA-DQ2-bearing B cells together with a gliadin preparation and looked at gliadin-derived peaks in the peptide preparation from the B-cell surfaces. But she could not identify any new sequence in the HLA profile of the B-cell surfaces. One possible explanation is that we are using the wrong antigen-presenting cells, and, as a consequence, the B cells do not take up the gliadin efficiently enough. Perhaps B cells with specific surface IgA antigliadin are better antigen-presenting cells. Marsh: In my opinion, there already is sufficient information about which gliadin peptides apparently are active within the framework of gluten sensitivity. This information has mostly been gained from in vivo experiments, and it has been published. How do those results square with your own? Koning: Are you referring to gliadin peptides 31 to 43? Marsh: One of the best in vivo works is the one published by Sturgess and colleagues in Lancet last year. They reported quite a big gliadin peptide, which was said not to be active. My group has shown that gliadin peptides 31 to 43, 44 to 55, and 78 to 88 are active in vivo, while peptide 56 to 68 is inactive. All the active gliadin peptides have one thing in common: they have a PQQP motif. I just wonder why you have not, in fact, taken on board what is already known in the literature about possible active gliadin peptides. Koning: I totally agree with you, but the part of the gliadin molecule that is associated with in vivo activity does not necessarily have to be the same one that binds to the class II DQ2 molecule. For example, the part of the gliadin molecule associated with in vivo activity may stick out from the class II-binding place but nevertheless be capable of stimulating a T cell. But a binding motif to HLA-DQ2 is always necessary. Let us take gliadin peptide 31 to 43. If we want to come to a binding motif with HLA-DQ2, we have to make the assumption that the prolines make the peptide space one amino acid smaller. In such a case, this gliadin peptide with a large aromatic residue, a glutamine, near the end terminus could bind to HLA-DQ2. But this is an assumption, and we still have to test whether this peptide actually binds to HLA-DQ2 with any affinity. Kasarda: My question is close to the one of Mike Marsh. Do you think there is any difference in the in vivo uptake and processing of small peptides, like 31 to 43, compared with these larger peptides or intact gliadin protein? Koning: To use whole gliadin in vitro, it is necessary to dissolve it in 70% ethanol. Thus, smaller fragments have to be tested anyway. We have used trypsin-digested gliadin in our assays, as well as 18-mer synthetic peptides, but we did not find different results. Perhaps our synthetic peptides are not of the right size. There is a serious problem with the synthesis of these peptides because of all the glutamines and prolines. If you make a synthetic peptide, you expect to find one peak on reverse-phase high-performance liquid chromatography, but what you actually get is a whole series of peaks, depending on the size and part of the molecule. Kasarda: We have synthesized 26-residue gliadin peptides. Koning: Did you check them on high-performance liquid chromatography and mass spectrometry? Kasarda: Yes. The chemist who did the synthesis did get double peaks on high-performance liquid chromatography because of proline isomerigation, but this is in the nature of these peptides. But there were single peaks by mass spectrometry. Peña: Many years ago Falchuk and Strober, using organ cultures, showed that HLA class I -A1 was important in the response to gliadin in vitro. If class II molecules contain some parts from class I, could it be possible that a part of class I -A1 helps to present gliadin peptides? Koning: There is one article published in the Proceedings of the National Academy of Sciences in which, using fluorescent techniques, presentation of two peptides to one class II molecule was shown. That probably is a very particular case and not what people believe to be common. Perhaps some peptides have an affinity for each other, just like proteins have an affinity for each other.
No takes yet. Share an insight, caveat, or question.
Frits Koning (1997) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: