Wheat is one of the most important cereals in the world. Although starch is the major component of wheat grains (60–75%), the proteins of the grain (9–18%) are essential for bread-making quality. According to their functionality, wheat grain proteins are divided into two types: gluten and non-gluten proteins. About twenty per cent of the total grain protein corresponds to non-gluten proteins, comprising albumins and globulins, which have metabolic and structural functions with a minor role in wheat quality. In contrast, gluten proteins represent about 80% of the total grain proteins and they are mainly responsible for the rheological properties of the dough. These gluten proteins, also called prolamins given their high content of the amino acids proline and glutamine ^[1][2][1,2], include gliadins (α/β-, ω-, and γ-gliadins) and glutenins, comprising high molecular weight glutenin subunits (HMW-GS) and low molecular weight glutenin subunits (LMW-GS) (Figure 1). Most of the gliadins are monomeric proteins and they form intramolecular disulphide bonds; however, glutenins are polymeric complexes linked by inter- and intramolecular disulphide bonds to glutenins and gliadins. Both gliadins and glutenins form a viscoelastic network that traps the CO₂ released during fermentation, providing the typical texture characteristics of the wheat bread. In this process, gliadins are responsible for the viscosity and extensibility of the dough. Nevertheless, glutenins provide elasticity and strength to the dough, contributing, especially the HMW-GS, to the formation of long polymers [3]. Gluten proteins are coded by multiple genes at complex loci present on chromosomes 1 and 6. In particular, α-gliadins are coded by genes at the Gli-2 loci present on the short arm of chromosome group 6 [4], ω-, and γ-gliadins are genetically linked and are coded by genes at the Gli-1 loci on the short arm of chromosome group 1, typical LMW-GS are coded by genes at Glu-3 loci, genetically linked to Gli-1 loci, and finally, HMW-GS are coded by genes at Glu-1 loci present on the long arm of chromosome group 1 (Figure 2). The presence of gliadins and glutenins, and the balance between these two types of proteins is essential for the quality of the final product [5].

Wheat is associated with pathologies such as coeliac disease (CD), which affects about 1% of the population worldwide [6], non-coeliac wheat sensitivity (NCWS) [7] and allergies; baker’s asthma [8] and wheat-dependent exercise-induced anaphylaxis (WDEIA) [9].

Baker’s asthma is a respiratory allergy triggered by a wide range of wheat proteins that react with immunoglobulin E (IgE). Wheat proteins responsible for baker´s asthma comprise gliadins, glutenins, serine proteinase inhibitors (serpins), thioredoxin, agglutinin and several enzymes [10]. The α-amylase inhibitors are included among these enzymes, which are proteins soluble in chloroform: methanol (CM-like proteins) [8]. These α-amylase inhibitors have been described as the major group of proteins responsible for this allergy. Wheat is also responsible for WDEIA, which is an allergic reaction caused by combining the ingestion of wheat food and subsequent physical exercise. The major allergens associated with WDEIA are the ω-5 gliadins ^[9][11][9,11]. Palosuo et al. [12] suggested that the activation of transglutaminase in the intestinal mucosa could be provoked by the development of large allergen complexes responsible for triggering anaphylactic reactions during physical exercise in patients with WDEIA.

NCWS is a widespread and heterogeneous pathology. The disease has been described as a reaction to gluten proteins, in which allergic and autoimmune mechanisms have been excluded. In fact, other proteins such as metabolic proteins called α-amylase/trypsin inhibitors (ATI) [13], and FODMAPS (fermentable oligo-saccharides, disaccharides, monosaccharides and polyols) seem likely candidates to cause this pathology. Removing gluten from the diet is the only way to normalise the small-bowel mucosa as well as improving the symptoms. On the other hand, there are conflicting results about the existence of a wheat/gluten-induced inflammation in the majority of patients, as the mucosa from patients with gluten/wheat sensitivity does not express markers of inflammation, and their basophils are not activated by gliadin [14].

Coeliac disease is the most studied of these pathologies. It is an autoimmune disorder that occurs in genetically predisposed individuals triggered by gluten proteins from wheat (gliadins and glutenins), rye (secalins), barley (hordeins), oats (avenins), and also, all hybrids in which any of the toxic cereals are involved. CD has a strong environmental component, gluten, but also a genetic component, concerning the human leukocyte antigen (HLA)-DQ2 and HLA-DQ8 [15].

Gluten proteins are characterised by a high content of proline and glutamine residues making their complete digestion difficult. This gluten composition produces large peptides with immunostimulatory activity in the intestinal lumen [16]. These immunogenic peptides cross the intestinal epithelium and are deamidated by the tissue transglutaminase 2 (tTG2) in the lamina propria [17], providing a negative charge to gliadin peptides and hence enhancing their affinity to bind HLA-DQ2/8. It causes the activation of CD4⁺, triggering intestinal damage and malabsorption symptoms. For immunological reasons, 95% of coeliac patients present the HLA-DQ2 antigen and 5% contain the HLA-DQ8 [18]. The α-gliadin protein family is highly stimulatory because the 33-mer, the main immunodominant peptide for coeliac patients, is located in the repetitive region of these proteins [19]. The 33-mer peptide is resistant to gastric and pancreatic digestion, playing an essential role in the development of coeliac disease ^[16][20][16,20]. α-Gliadins also contain an additional DQ2-restricted epitope which partially overlaps with the 33-mer peptide [21]. However, it has been seen that not only the α-gliadins fraction is responsible for the stimulation of T-cells in people with CD. In fact, for the purpose of investigating the recognition profile of gluten immunogenic peptides in HLA-DQ2 coeliac patients, Camarca et al. [22] analysed the ability of stimulatory peptides to stimulate T-cell clones from CD patients. They concluded that there is indeed a substantial heterogenicity in the response of T-cells to gluten, also highlighting the stimulatory relevance of ω-, and γ- peptides.

The development of wheat varieties with reduced immunogenic gluten proteins could be noteworthy for CD and NCWS people, not only to improve their quality of life but also to reduce the incidence of those pathologies, given that a relationship between the amount and exposition to gluten and these diseases has been suggested ^[23][29]. The appeal of these wheat varieties could undoubtedly also apply to the general population, in particular for those who want to reduce the intake of gluten. The search for wheat varieties, which naturally are devoid of coeliac epitopes in the gliadin sequences encoded by the A, B, and D genomes of wheat, has been discussed for a long time. Hence, van den Broeck el al. ^[24][30], using monoclonal antibodies, analysed different wheat lines from Chinese Spring with partial deletions in the short and long arms of the group 6 chromosomes. They obtained a reduced T-cell stimulatory response in those lines that carried the deletion of α-gliadins in the D genome. These findings supported the study of van Herpen et al. ^[25][31] who showed that epitopes are not randomly distributed in the wheat genome and the three genomes contribute differently to the total epitope content and immunogenicity. Also, Spaenij-Dekking et al. ^[26][32] reported the existence of wheat genotypes with a low content of T-cell stimulatory sequences, suggesting that selection of less toxic lines through classical breeding could be possible. On the other hand, Juhász et al. ^[27][33] combined the use of both the genome sequence of the cultivar Chinese Spring and wheat protein/peptide databases. They carried out a comprehensive analysis of wheat genes that encode proteins related to digestive disorders and allergies, and their chromosomal locations. This reference mapping of immunostimulatory wheat proteins provides a new tool to select programs targeting traits, such as producing low-gliadin lines. With the exception of this last study, the rest of them described above were based on α-gliadin fractions. Thus, it cannot be concluded that they are exempt from other sequences that can stimulate the autoimmune response in coeliac individuals since it is known that many epitopes are present in other gliadin fractions ^[28][34], especially the ω-gliadins that could also play an important role in CD [19]. Moreover, the sequences of the individual genes within the same family of gliadins are very akin, and they might contain multiple and different T-cell stimulatory epitopes. This high complexity coupled with the fact that gliadin genes are inherited in blocks makes obtaining wheat lines with reduced toxic epitopes by using conventional breeding really difficult.

One promising approach would be to reduce the amount of CD immunogenic gluten proteins by the application of the latest developments in genetic engineering to toxic cereals. In this case, the silencing of gliadin genes, specifically the immunodominant α-gliadin fractions ^[29][35] could be achieved.

Given that gliadins are formed by groups of proteins encoded by multigene families that contain epitopes related to CD, the objective of several researchers has been the silencing of groups of gliadin genes rather than single genes ^[30][31][36,37]. In a first attempt, Gil-Humanes et al. ^[32][38] carried out the silencing of the γ-gliadins fraction using a D-Hordein promoter ^[33][39]. These transgenic lines were obtained from the T. aestivum cv Bobwhite (denoted as BW208, and here used as a control). In this work, they demonstrated that using RNAi it was possible to silence a complex gene family. Furthermore, the results obtained from A-PAGE gels and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) revealed an increment in the other gliadin fractions, suggesting a compensatory effect as a result of silencing of γ-gliadins. As result, total gluten content, detected by monoclonal antibodies did not decrease but increased for some lines, indicating that a compensation mechanism with other gliadin fractions is operating, and the silencing of a single gliadin family does not provide wheat lines devoid of toxicity ^[34][40].

Grain protein profile and specific reduction of the different gliadin fractions can be achieved by using different silencing fragments for targeting prolamins in bread wheat ^[35][45]. Lines were analysed by proteomic and genetic techniques, and specifically, the gluten immunogenicity was determined by using competitive anti-gliadin 33-mer moAb. The results indicated a decrease in the gluten content in all silencing-fragment combinations except for lines with only γ-gliadins silenced (pghpg8.1 construction). Furthermore, the CD epitope analysis by mass spectrometry of pepsin and trypsin digested protein extracts, provided the identification of the E82 line as containing the lowest amount of CD immunogenic epitopes.

The stability of gliadin silencing is a key factor for the usefulness of the low-gliadin lines, either for its introgression to commercial cultivars or as raw material for food processing. In order to quantify the robustness of this silencing, we have collected the data from several years of experiments with 15 different RNAi lines, derived from two wild type parental lines, carrying the hairpin ω/α fragment ^[35][45], and also from three nitrogen fertilization levels. Therefore, the data provide valuable information on the performance of the construct in different environments (Figure 3 and Figure 4). These data comprise values for several variables concerning the grain protein content as quantified by RP-HPLC: gliadin content and its fractions (omega-, alpha-, and gamma-gliadins); glutenin content and its fractions (high molecular weight- and low molecular weight-glutenin); the content of total gliadin and glutenin; and the gliadin to glutenin ratio. Values for the gluten content (parts per million; ppm) as determined by monoclonal antibodies are also available in most of the works. All data have been transformed into proportions to their respective wild type so that they are comparable. The objective is to eliminate or, at least, reduce the content in this toxic fraction of the gluten.

Overall, in the set of six trials, reductions of more than 75% in the gliadin content have been obtained with respect to the wild type, and in some cases, around 90% (Figure 3), of the value of the wild type line from which they are derived (see wild-type values in Table 1). However, the decrease is not of the same magnitude for all the gliadin fractions, producing a greater reduction, especially, in γ-gliadins (Figure 3). Of all the silenced lines, three of them were highly promising and therefore they were present in most of the published trials (D783, D793 and E82) ^{[34][35][36][37][38]}[40,45,46,47,48]. Of these, lines D793 and E82 show a lower content of α- and γ-gliadins, which results in a reduction of 80% in the content of total gliadins (Figure 4). The E82 line presents, in addition, a significant reduction in the content of LMW glutenins ^[34][40] and, consequently, in the content of total glutenins (Figure 4). Comparing the variation in the prolamin content of these genotypes, it can be concluded that line E82 is of great interest in obtaining lines with steadily reduced toxicity.