Characteristics and Immunogenicity of Gluten Peptides in Enzyme-Treated and -Untreated Beers for Celiac Patients

Abstract

The peptidomes from the literature of 24 prolyl-endopeptidase-treated beers during fermentation, declared gluten-free, and 13 untreated beers have been characterised and subjected to an extensive study to investigate their safety for celiac patients. The analysis contains 1996 gluten peptides, ascribed to the treated beers, and 1804 to the untreated beers. The prolyl-endopeptidase-untreated malt beers are hazardous for celiac patients. Peptides of most of these beers showed matches with complete celiac immunogenic motifs, and an additional 28% of the peptides have partial matches with complete immunogenic motifs. On the other hand, after the enzyme treatment during fermentation no celiac hazardous gluten peptides are identified in the treated beers. Due to partial matches with complete celiac immunogenic motifs, 11% potentially hazardous gluten peptides are still identified in the treated beers. Only a maximum of 17% of these peptides can be detected by ELISA analysis. A mass spectrometry analysis or the recently developed method based on G12/A1 monoclonal antibody lateral flow immunochromatographic assay seems necessary to thoroughly reveal the potential risk of the treated beers. The actual immune response of treated beer, described in the literature by the response of the serum antibodies of celiac disease (CD)-active patients and by in vitro immune response, could not be related to the presence of known (partial) CD-immunogenic motifs in the gluten peptides.

1. Introduction

Gluten is a restricted term for the complex mixture of grain seed storage proteins found in wheat, barley, rye and possibly oats [1,2,3]. Within the different grains, gluten proteins have been given different trivial names: gliadins (prolamins) and glutenins (glutelins) in wheat, hordeins in barley, secalins in rye and avenins in oat [3,4]. All gluten proteins have long repetitive amino acid sequences rich in glutamine (Q, 35 to 37%) and proline (P, 17 to 23%) [2,3].

Celiac disease (CD) is an immune-mediated inflammatory disease of the small intestine triggered by the ingestion of dietary gluten in genetically susceptible individuals [5,6]. The characteristic high levels of Q- and P-residues in gluten have a strong impact on the CD pathogenesis. Due to the high levels of P-residues, gluten peptides are partly digested by gastro-intestinal enzymes and rather large gluten peptides enter the upper small intestine. It is suggested that in CD-susceptible individuals, certain specific gluten sequences can manifest an inflammatory state of the intestinal mucosa by stimulating both the innate and adaptive arms of the immune system, which in time will result in (complete) mucosal villous atrophy [7,8]. Adaptive immunity is initiated by binding specific gluten peptides to the peptide groove of HLA-DQ2/DQ8 molecules on the surface of Antigen-Presenting Cells, allowing for their presentation to CD4+ T-cells. P-residues will enhance the binding of gluten peptides to HLA-DQ2/DQ8 molecules in two ways. Firstly, proline can alter the gluten peptide to a left-handed poly-Pro II helical conformation preferred for HLA-DQ2/DQ8 binding. Secondly, proline enhances the specificity of extracellular tissue transglutaminase (tTG2) towards glutamine. The tTG2 enzyme has deamidase activity and hydrolyses glutamine into glutamic acid (Q→E). This facilitates the binding of the gluten peptide to HLA-DQ2/DQ8 molecules [4,7].

CD is becoming epidemiologically more relevant as illustrated by an increase in CD incidence over time [6]. According to King et al. [6] the prevalence of biopsy-confirmed CD is estimated at 0.7%, while diagnosis based on serology is estimated at 1.4% of the population. However, CD is underdiagnosed due to individuals presenting an atypical or silent form of the disease with minimal or no symptoms [6]. Individuals with CD have to adhere to a lifelong gluten-free diet based on the strict avoidance of common food products (e.g., bread, pasta, beer) made from wheat, barley, rye and, in rare cases, oats [7,8].

The maximum safe level of 20 ppm gluten in gluten-free products is defined by legislative authorities [9]. This level meets the tolerable daily intake (TDI) for subchronic (repeated) exposure of 0.4 mg gluten/day and for chronic exposure of 7.0 mg gluten/day. The latter TDI values are for wheat gluten and were derived from the NOAEL and LOAEL of exposure, taking into account safety margins for inter-individual differences [10].

Gluten-free food products are divided in two different groups. The first group consists of products made from non-gluten-containing grains (e.g., rice, sorghum). The second group are products made from gluten-containing grains (i.e., wheat, barley, rye, oats and their crossbred varieties) that are specifically processed to products with a final gluten content below 20 ppm. The labelling of group two of gluten-free products is either by a ‘gluten-free’ claim in Europe or a ‘gluten-removed’ claim in US. However, on the label of prepacked food products and beverages of group two gluten-free products, the presence of gluten-containing grains must be declared [11,12].

A gluten-free diet is associated with significant burdens including elevated financial costs, inadvertent exposure to gluten and psychosocial impacts [6]. Hence, the availability of gluten-free products safe for consumption by individuals with CD is essential. The gluten-free market has witnessed an enormous growth over the years. In the brewing industry, there is also a thirst for gluten-free beers.

Beer is a well-beloved and consumed alcoholic beverage traditionally mainly made from barley malt. Therefore, its place in gluten-free diet has been questioned. Several screening studies have shown that the presence of gluten (poly)peptides in the final beers is often still too high (>20 ppm) to be safely consumed by most patients with CD [13,14,15,16,17]. In addition, through the use of mass spectrometry (MS) techniques, there was an alert that some of the remaining gluten peptides in apparently gluten-free beer have CD-immunogenic properties [18,19,20,21].

A highly sensitive and easy-to-use method based on G12/A1 monoclonal antibody lateral flow immunochromatographic assay (LFIA) was developed by Segura et al. [22] (2022) that requires no extraction and has an LOD of 0.5 ppm gluten. In addition, the method can measure the level of bioavailable gluten-immunogenic peptides in hydrolysed foods. Their results showed that, although regulations allow beer products to contain less than 20 ppm of gluten to be considered gluten-free, they still contain gluten-immunogenic peptides, which should not be consumed by patients with CD [22].

Prolyl-endopeptidase (PEP) from Aspergillus niger (AN) is a food-grade serine protease (EC 3.4.21.26; AN-PEP) which selectively hydrolyses peptide bonds at the carboxyl side of internal P-residues [23]. In the field of brewing, this proline-specific enzyme is commercially available and is used to clarify beer and also specifically to produce gluten-free malt beers.

Several studies have used R5-cELISA (RIDASCREEN^® gliadin competitive R5-cELISA R7021 (R-Biopharm)) to quantify gluten (poly)peptides in AN-PEP-treated beers and observed that AN-PEP could successfully lower the gluten concentrations of malt beers below the gluten-free threshold [24,25,26,27,28]. The gluten content of these beers was even below the LOQ of 10 ppm of the R5-cELISA. However, it cannot be guaranteed that the residual amount of gluten (peptides) below the threshold as found in the competitive ELISAs based on R5 and G12 antibodies is without hazard for individuals with CD [27]. Therefore, the actual safety of gluten-free and gluten-removed beers by AN-PEP, with gluten concentrations below the threshold, can still be questioned.

In a study by Allred et al. [29], a response of serum antibodies of CD-active patients towards a commercially available gluten-removed beer (defined by regulation as <20 ppm gluten) was observed. This suggests that there are residual gluten peptides in enzymatically treated beer that may be specifically recognised by individuals susceptible to CD. Spada et al. [30] found that an AN-PEP-treated malt beer containing 11 ppm gluten, assayed with the G12-cELISA, elicited a weak in vitro CD-immune response.

These findings support the results of the evaluation of the CD-immunostimulatory properties of peptides of AN-PEP-treated barley beers [31,32,33,34]. The peptidomes of the AN-PEP-treated barley beers were generated by MS technology.

Akeroyd et al. [31] identified gluten peptides in five lab-scale AN-PEP-treated beers and their untreated counterparts as references, and in two commercial AN-PEP-treated beers. The untreated beers contained gluten peptides with (multiple) CD-immunogenic motifs. In none of the seven AN-PEP-treated beers peptides were CD-immunogenic motifs found. Akeroyd et al. [31] noted that CD-immunogenic motifs are hydrolysed in AN-PEP-treated barley malt beers.

Colgrave et al. [32] identified gluten peptides in sixteen beers (four control, i.e., gluten-containing beers; ten AN-PEP-treated beers; one gluten-free beer produced by an undisclosed proprietary process; and one brewed from an ultralow gluten barley variety). In all four control beers, gluten peptides that carry (multiple) complete CD-immunogenic motifs were found. In four out of the ten AN-PEP-treated beers, (multiple) potentially CD-immunogenic motifs (i.e., with a homologous amino acid sequence) were found. In the other beers, no CD-immunogenic motifs were found.

Fiedler et al. [33] identified gluten peptides in AN-PEP-treated and -untreated 100% barley malt beer (both beers produced at 19 L fermentor scale). In these two beers, gluten peptides containing CD-immunogenic motifs were identified, although fewer were in the AN-PEP-treated beer.

The discrepant conclusions on the ability of AN-PEP to eliminate CD-immunogenic properties of gluten peptides, made by Akeroyd et al. [31] on the one hand and Colgrave et al. [32] and Fiedler et al. [33] on the other hand, are because of the different criteria that were used for evaluation/prediction. Akeroyd et al. [31] only screened the found peptides for complete CD-immunogenic motifs. Colgrave et al. [32] and Fiedler et al. [33] carried out a deeper search. They screened their peptides for complete CD-immunogenic motifs and for homologous motifs, and mismatches or insertions in the amino acid sequence of the CD-immunogenic motifs occurred. However, the latter can be discussed, i.e., the number, nature and position of the allowed mismatches or insertions. Pilolli et al. [35] argued to harmonise search strategies to evaluate the CD-immunostimulatory property of gluten peptides. Therefore, Pilolli et al. [35] recommended the use of guidelines for risk assessments of allergenicity provided by the EFSA (European Food Safety Authority) [36].

The latter guidelines were for the first time adapted by Watson et al. [34] to analyse the CD-immunostimulatory property of residual gluten peptides in an industrial produced AN-PEP-treated gluten-free beer and its untreated counterpart as reference. Gluten peptides with complete CD-immunogenic motif(s) were only found in the reference beer. However, in both beers a set of gluten peptides that contain one or more partial CD-immunogenic motifs were identified making the AN-PEP-treated beer also potentially hazardous for patients with CD.

Watson et al. [34] laid the foundation to study the gluten peptidome of AN-PEP-treated gluten-free beers in a comprehensive manner.

This study aims to thoroughly investigate peptides from more AN-PEP-treated gluten-free beers and AN-PEP-untreated beers on various aspects. The ultimate goal of the research is to improve the safety of beer for celiac patients, which could lead to new brewing processes. In this study, an in-depth investigation of available appropriate data and knowledge from the literature is used to generate and process novel data to achieve four objectives. These objectives are as follows: (i) a better insight in the AN-PEP activity to make gluten-free beer, (ii) a well-founded risk assessment of the CD-immunostimulatory property of gluten peptides in AN-PEP-treated gluten-free beers while resolving the discrepant conclusions between the literature studies, (iii) an evaluation of the identified (potentially) immunogenic peptides for their ability to be detected by ELISA analysis and (iv) the search for the relationship between the actual immune response of treated beer and the presence of known (partial) CD-immunogenic motifs in the gluten peptides.

2. Materials and Methods

2.1. Data Collection

For this study, available research data from four studies on gluten peptides in AN-PEP-treated beers during fermentation and AN-PEP-untreated (some counterpart references) beers were used. These studies are Akeroyd et al. [31], Colgrave et al. [32], Fiedler et al. [33] and Watson et al. [34]. In this way, the gluten peptidome was analysed from 37 different barley malt beers, of which 24 were AN-PEP-treated beers and 13 were AN-PEP-untreated beers (Table S1).

The four studies are eligible for the research in this manuscript because they fulfil the four requirements of detailed data of the peptides of the beers being available, AN-PEP-treated- and-untreated beers are analysed, the beers are barley beers and the treated beers are gluten-free beers. The 24 AN-PEP-treated gluten-free beers from the four studies have a gluten content < 20 ppm and therefore meet the legal gluten-free safe threshold of 20 ppm and as such can all be officially labelled as gluten-free beer independent of the scale and process of brewing. In total, 12 of the 24 AN-PEP-treated beers have an AN-PEP-untreated counterpart beer (Table S1).

The AN-PEP enzyme dosage and fermentation temperature, if specified, are included in Table S1. The pH during fermentation, and therefore also during enzyme treatment, drops from the average pH 6 of the wort to the average pH 4.5 of the beer. This pH drop is due to yeast activity.

For fermented or hydrolysed products such as beer, the recommended test for gluten analysis is the R5-cELISA test [1]. The test is widely approved: AACC Method 38-55.01, AOAC OMA Final Action 2015.05 and ASBC internationally approved method Beer-49. Another used method for gluten quantification in beer is the G12-cELISA. Although the US FDA issued a dossier in August 2020 (FDA-2014-N-1021-0560) stating that it does not know of “any scientifically valid analytical method that is effective in accurately detecting and quantifying the protein content of gluten in fermented or hydrolyzed foods in terms of equivalent amounts of intact gluten proteins”.

The gluten content of 14 AN-PEP-treated beers from three studies was analysed using the competitive ELISA with the R5-antibody. The treated beers from the research by Akeroyd et al. [31] and Watson et al. [34] have a gluten content < LOQ (<10 ppm) (Table S1). The six treated beers from the study by Fiedler et al. [33] have a gluten content of 12–14 ppm (Table S1). The 10 AN-PEP-treated beers from the study by Colgrave et al. [32] have a gluten content < 20 ppm as measured using ELISA by the beer manufacturer (Table S1) and can also be considered as gluten-free. No further information is available on the type of ELISA used for these 10 beers.

From all studies, the set of described peptides were collected, characterised and cleaned-up. The gluten peptides were aligned to barley gluten (hordeins), wheat gluten (gliadins and glutenins), rye gluten (secalins) and Avenin-Like A protein and the adjacent amino acids at the C- and N-terminus determined. The final peptide list for the study contained 3800 gluten peptides of which 1996 are ascribed to the 24 AN-PEP-treated beers and 1804 to the 13 AN-PEP-untreated beers (Table S2).

2.2. Data Analysis

For the data analysis of the peptides, python scripts were designed by Bjorn Criel. The studied gluten peptides in beers were searched for exact and/or partial matches with the known 28 CD-immunogenic motifs [37] according to a stepwise search strategy that is schematically depicted in Figure 1.

The strategy is based on the EFSA guidelines [36], which are outlined by Watson et al. [34]. Three search steps are performed and each step has an annotated CD hazard. First, searches were performed for an exact match with the 28 CD-immunogenic motifs (Table S3). The found peptides are thereby annotated as CD-hazardous peptides. Second, an exact match with the partial CD-immunogenic motif(s) Q-X1-P-X2 (Table S4) was sought. Third, a partial match (1–3 AA mismatches) with complete CD-immunogenic motif(s) was screened. The found peptides of the latter two steps are annotated as potentially CD-hazardous peptides. After a gluten peptide was matched in a previous step, it was omitted from the further search to avoid double counting. This is justified because gluten peptides with an exact match with complete CD-immunogenic motifs certainly have partial matches. Gluten peptides that did not show matching in one of the three steps and/or gluten peptides that are less than nine-AA-long are considered as non-CD-hazardous [34].

The gluten peptides of the 37 studied beers were also searched for the presence of the distinct antibody-recognition epitopes of both ELISA tests used nowadays to quantify gluten in beer.

The used stepwise search strategy for CD-immunogenic properties of gluten peptides in beers is also employed for the data of Spada et al. [30] (Table S8) from an AN-PEP-treated beer and four untreated beers upon in vitro simulated gastro-duodenal (GD) digestion.

3. Results and Discussion

3.1. Evaluation of CD-Immunostimulatory Properties of Gluten Peptides

The results of the evaluation of CD-immunostimulatory properties of gluten peptides characterised in AN-PEP-treated and -untreated beers are summarised in

Table 1. CD-immunogenic motifs of gluten peptides (sub-10 kDa proteinaceous fraction) in prolyl-endopeptidase from Aspergillus niger (AN-PEP)-treated and -untreated beers. The Beer_IDs followed by an identical letter (a–l) are, respectively, AN-PEP-treated and their untreated counterpart beers. Three untreated beers are counterparts of two beers. For example beer 34 ^f,i is counterpart of treated beers 18 ^f and 21 ⁱ.

		Complete Match CD-Motif	Partial Match Q-X1-P-X2	Partial Match CD-Motif	No Match + <9 AA Long
AN-PEP treated beers	n	Rel. nr.	Rel. nr.	Rel. nr.	Rel. nr.
Beer1 a	38	0	0	0	100
Beer2 b	39	0	0	5	95
Beer 3 c	44	0	0	5	95
Beer 4 d	138	0	0	9	91
Beer5 e	118	0	0	8	92
Beer6	39	0	0	0	100
Beer7	43	0	0	0	100
Beer8	28	0	0	4	96
Beer9	33	0	0	0	100
Beer 10	36	0	3	0	97
Bee r 11	31	0	3	3	94
Beer 12	44	0	5	7	89
Beer 13	62	0	3	5	92
Beer 14	59	0	5	7	88
Beer 15	30	0	0	10	90
Beer 16	22	0	5	5	91
Beer 17	19	0	0	5	95
Beer 18 f	160	0	7	9	84
Beer 19 g	176	0	5	8	87
Beer 20 h	191	0	7	9	84
Beer 21 i	173	0	3	9	87
Beer 22 j	161	0	2	9	88
Beer 23 k	157	0	5	9	86
Beer 24 l	155	0	7	5	88
Total	1996	0	4	7	89
AN-PEP untreated beers
Beer 25 a	99	3	12	11	74
Beer 26 b	66	2	29	9	61
Beer 27 c	68	4	9	7	79
Beer 28 d	220	5	12	14	70
Beer 29 e	135	8	16	12	64
Beer 30	23	0	9	4	87
Beer 31	48	2	8	13	77
Beer 32	59	0	17	7	76
Beer 33	22	0	9	0	91
Beer 34 f,i	280	5	16	14	66
Beer 35 g,j	274	4	15	14	67
Beer 36 h,k	352	3	18	14	65
Beer 37 l	158	4	18	6	72
Total	1804	4	16	12	69

AA: amino acid; complete match CD motif: exact match with complete CD-immunogenic motif(s) (9/9 AA match); partial match Q-X1-P-X2: exact match with partial CD-immunogenic motif(s), i.e., Q-X1-P-X2 (4/4 AA match); partial match CD-motif: partial match (1–3 AA mismatches) with complete CD-immunogenic motif(s) (6/9 to 8/9 AA match); no match + <9 AA long: no match with complete CD-immunogenic motif(s) (0/9 to 5/9 AA match) and gluten peptides with less than 9 AAs in the sequence; n: number of peptides found in the beer or total number of peptides found in the group of treated or untreated beers; Rel. nr.: relative number of peptides (in %) per beer or per total of the two groups of beers with CD-immunogenic motifs (complete or partial matches or no matches). For example, of the 191 peptides in beer 20 ^h, 0% have a complete match, 7% have a partial match Q-X1-P-X2, 9% have a partial match CD motif and 84% have no match.

(next page).

Complete CD-immunogenic motifs were absent in all peptides of the 24 AN-PEP-treated beers, while in the 13 AN-PEP-untreated beers, 0 to 8% (4% of the total of peptides of the untreated beers) (Table 1) of the gluten peptides comprise one to three complete CD-immunogenic motifs (Table S2). For example, in the gluten peptide collection of untreated beers, peptide F.PLQPQQPFPQQPQQPFPQPQQP.F (peptide_ID: 2304, beer_ID: 28) can be considered as the most CD-hazardous as it contains three different complete CD-immunogenic motifs: QQPQQPFPQ (motif_ID: 10), LQPQQPFPQ (motif_ID: 12) and QQPFPQQPQ (motif_ID: 13).

Only 10 out of the 28 known CD-immunogenic motifs (motif_IDs: 6, 7, 9, 10, 12, 13, 14, 18, 19 and 21) were found as complete epitopes in gluten peptides of AN-PEP-untreated beers, with QQPQQPFPQ (motif_ID: 10), LQPQQPFPQ (motif_ID: 12) and QQPFPQQPQ (motif_ID: 13) being the most common ones (14, 34 and 29 times, respectively; Table S6). Although the beers under study are barley malt-based beers, the latter three CD-immunogenic motifs are ascribed to wheat gluten. This is not unexpected as 21 of the 28 motifs are ascribed to wheat gluten (Table S3) and wheat, barley and rye gluten contain homologous amino acid sequences because these grains are closely related (Triticeae tribe within the Pooideae subfamily of grasses). In addition, historically, CD-immunogenicity studies have mainly focused on wheat gluten rather than on barley, rye and oat gluten. Therefore, CD-immunogenic motifs are mainly described from wheat, and only four from barley.

It is a promising result that no complete CD-immunogenic motifs could be found in the 24 AN-PEP-treated beers. However, it would be too simplistic to conclude from this result that the remaining gluten peptides in beer are non-CD-hazardous. It can still be that these gluten peptides contain ‘unknown CD-motifs’ of which the CD-immunostimulatory effect (HLA-DQ2 or -DQ8 binding, or CD T-cell proliferation in human samples or subjects) has not been confirmed yet [35]. In addition, in the CD pathogenesis, cross-reactivity by T-cells towards gluten peptides that contain homologous sequences to known CD-immunogenic motifs might also occur. The CD-immunostimulatory effect of such sequences is challenging to define as HLA-DQ2 or -DQ8 binding, and CD T-cell proliferation is highly dependent on the nature and position of some amino acids in the nonamer epitope. As for this, EFSA guidelines include the motif Q-X1-P-X2 (L, Q, F, S or E at position X1; and Y, F, A, V or Q at position X2) in the search strategy. Q-X1-P-X2 is a characteristic motif in most HLA-DQ2 epitopes. Because not all HLA-DQ2 or -DQ8 epitopes contain this characteristic motif, known CD-immunogenic motifs with up to three amino acid mismatches were included in the search strategy as well [34] (Figure 1).

Table 1 illustrates that among 24 gluten-free AN-PEP-treated beers, there is a wide variation in the number of exact matches of the gluten peptides to partial CD-immunogenic motif(s) Q-X1-P-X2. Similar variation applies to the number of partial matches to complete CD-immunogenic motif(s) (Table 1). In their study of gluten-free beers, Nye-Wood et al. [20] found variation between the beers in both the total gluten content and the presence/absence of immunogenic epitopes.

In AN-PEP-treated beers, 0 to 7% (4% of the total) of the gluten peptides comprise one to two exact matches with partial CD-immunogenic motif(s) Q-X1-P-X2. In addition, 0 to 10% (7% of the total) of the gluten peptides comprise one or more (median 1, but up to 16 in peptide Q.QPQPYPQQPYPQQP.Q (peptide_ID: 1442, beer_ID: 21)) partial matches with complete CD-immunogenic motif(s) (Table 1 and Table S2). Thus, although the gluten peptides in AN-PEP-treated beers do not comprise complete CD-immunogenic motifs, 11% of the total of the gluten peptides are potentially hazardous for patients with CD. In comparison, in AN-PEP-untreated beers notably more, i.e., 8 to 29% (16% of the total) of the gluten peptides comprise one to four exact matches with partial CD-immunogenic motif(s) Q-X1-P-X2, and 0 to 14% (12% of the total) of the gluten peptides comprise one or more (median 6, but up to 32 in peptide P.QPFPQQPQPFPQQPIPQQPQPYPQQPQPFPQQP.F (peptide_ID: 2016, beer_ID: 25)) partial matches with complete CD-immunogenic motif(s) (Table 1 and Table S2). Thus, for AN-PEP-untreated beers, besides the 4% of gluten peptides that contain known complete CD-immunogenic motifs, an additional 28% of the total of the gluten peptides are potentially hazardous for patients with CD.

In total, 7 out of 25 partial CD-immunogenic motif(s) Q-X1-P-X2 (partial motif_IDs: 1, 6, 7, 8, 10, 15 and 20) were found in the gluten peptides of AN-PEP-treated beers. Meanwhile 13 out of 25 (partial motif_IDs: 1, 2, 3, 5, 6, 7, 8, 9, 10, 15, 20, 21 and 25) were found in the gluten peptides of untreated beers (Table S7). For both, the two most common ones were QQPF (partial motif ID_7) and QQPQ (partial motif ID_10) (Table S7). The latter two partial motifs can be linked to 11 different complete CD-immunogenic motifs (Table S5). Partial matches with one to three amino acid mismatches to almost all 28 complete CD-immunogenic motif(s) were found in AN-PEP-treated and -untreated beers.

Several researchers have evaluated the ability of AN-PEP to eliminate the CD-immunogenicity of beer. The reported results are sometimes conflicting, mainly because of the different criteria for the evaluation of CD-immunogenic properties of gluten peptides. Cognizant of this hiatus, we re-evaluated in this study the CD-immunogenic properties of the previously described gluten peptides in AN-PEP-treated beers by a comprehensive approach according EFSA guidelines. By doing so, gluten peptides with sequence identity to partial CD-immunogenic motifs were found in 20 out of the 24 analysed AN-PEP-treated beers (Table 1). Probably the number of (potentially) hazardous gluten peptides in AN-PEP-treated beers is higher. This is because the findings are based on the sub-10 kDa proteinaceous fraction of AN-PEP-treated beers and larger peptides in AN-PEP-treated beers can be expected as mentioned by Colgrave et al. [32] and Fiedler et al. [33]. It is not excluded that larger gluten peptides still harbour a CD-immunogenic motif. Spada et al. [30] could characterise gluten peptides (sequences not available) that contain the complete CD-immunogenic motifs IQPQQPAQL, QQPFPQQPQ, PFPQPQQPF, PFSQQQQPV and FSQQQQSPF (or homologs thereof) in the sub-6 kDa proteinaceous fraction of a gluten-free (11 ppm gluten; G12-cELISA) AN-PEP-treated beer. However, after a chymotrypsin digest to split the larger polypeptides that initially could not be identified by MS/MS, several gluten peptides (sequences not available) that contain the complete CD-immunogenic motifs PFPQPQLPY and PQPQLPYPQ (or homologs thereof) were characterised as well. It is striking that the latter two motifs were not found in the peptides of the studied 13 AN-PEP-untreated beers (Table S2). This points also towards a different action of the malt peptidases and the chymotrypsin.

It is still difficult to interpret the CD-immunostimulatory potential of (AN-PEP-treated) beers based on the set of (potentially) hazardous gluten peptides they contain. Their bioaccessibility and bioactivity are of importance as well. The regulatory gluten-free safe threshold of 20 ppm was the result of analysis of intact gluten [38]. It is not well known whether hydrolysed gluten generate the same biological activity [39]. In addition, it has not been investigated how beer polypeptides change upon gastro-duodenal (GD) digestion. GD digestion of gluten peptides in AN-PEP-treated beer might increase the CD-immunogenicity by the release of epitopes previously inaccessible in longer polypeptides, or it could eliminate sensitive sequences while digestion-resistant sequences remain [30].

Bioaccessibility of beer gluten peptides can be determined experimentally by in vitro digestion with simulated gastric fluid, liposomes, pepsin and simulated intestinal fluid [30]. Bioactivity can be determined by in vivo or in vitro experiments. In vivo clinical challenge studies include small bowel histology. Valuable studies are in vitro T-cell activation tests [21,30,40] or in vitro antibody recognition tests [29] using sera of CD-active patients and a non-CD-control group. Both bioaccessibility and bioactivity of beer gluten peptides has been studied by Spada et al. [30]. They examined five beers in an in vitro simulated GD digestion, where gluten levels and the (gluten) proteome of the beer were analysed to assess the bioaccessibility of beer gluten peptides. In addition, Spada et al. [30] used the GD hydrolysed beer samples to stimulate in vitro intestinal T-cell lines from three different patients with acute CD to assess the bioactivity of beer gluten peptides. Before digestion, an AN-PEP-treated malt beer contained 11 ppm gluten, an all-barley malt beer contained 16 ppm gluten and a Weiss (wheat) beer and two einkorn beers contained >80 ppm gluten (G12-cELISA). After digestion, all their examined beers contained <20 ppm gluten (G12-cELISA). In agreement, their HPLC-MS/MS analysis showed a significant reduction in protein components, including in gluten peptides/proteins, in the digested beer samples in comparison to in the undigested ones. Nevertheless, both digested and undigested beer samples could elicit a weak in vitro immune response (i.e., IFN-γ response), although, in general, the response was noticeably reduced by GD digestion and mostly below medium or comparably medium. In order to understand the CD bioactivity of the gluten peptides that remain after GD digestion [30], these peptides were searched in this study for CD-immunogenic motifs (Table S8). In AN-PEP-treated beer containing, according to Spada et al. [30], multiple complete CD-immunogenic motifs or homologs thereof, these motifs were absent upon in vitro simulated GD digestion (Table S8). However, one gluten peptide with the QFPQ motif could still be identified, as well as two more additional gluten peptides that contain partial matches to complete CD-immunogenic motifs with three amino acid mismatches (Table S8). Surprisingly, Spada et al. [30] characterised no gluten peptides in the GD-digested sample of the all-barley malt beer. As expected, our analysis of the peptides of the GD-digested einkorn beers and the Weiss beer showed complete and partial CD-immunogenic motifs and partial matches (Table S8). The immune responses elicited by Spada et al. [30] for the GD-digested beer samples (including the AN-PEP-treated beer) could not be correlated with our number and the type of found CD-immunogenic motifs (Table S8).

Spada et al. [30] reported that the immune response to CD-immunogenic gluten peptides is variable, as the response depends on the individual immune system used. They clearly illustrated this for the GD-undigested beers whereby most responses were above medium. The responses they found those which were lower than medium were the GD-undigested AN-PEP-treated beer and the GD-undigested all-barley malt beer. These two beers showed only one of the three CD T-cell lines with a positive response [30]. This response for the GD-undigested AN-PEP-treated beer was a substantially high response and unexpected as no hazardous gluten peptides were characterised in the 24 studied AN-PEP-treated beers of this study.

From one of the CD T-cell lines used by Spada et al. [30], it is known that it recognises the CD-immunogenic motifs PFPQPQLPY, PQPQLPYPQ and PFPQPQQPF (corresponding motifs_ID 1, 3 and 14 in this study; Table S3). In this study, multiple peptides were characterised that contain these (partial) motifs in the GD-digested Weiss beer of Spada et al. [30] (Table S9), while none were characterised in the corresponding AN-PEP-treated beer. Yet, the measured immune response was higher for the GD-digested AN-PEP-treated beer than it was for the GD-digested Weiss beer [30].

Peptides of 8 beers of the 24 AN-PEP-treated beers studied contain the partial motifs QQPF (in 46 peptides of 8 beers) or QLPY (only in one peptide of beer 21) of two (motifs_ID 3 and 14) of the three motifs recognised by the CD T-cell lines.

This all makes the relationship between actual immune response and the presence of known (partial) CD-immunogenic motifs confusing and demonstrates that remaining vigilant is the golden advice for celiac patients.

3.2. Evaluation of R5- and G12 Antibody-Recognition Epitopes in Gluten Peptides

The monoclonal R5-Antibody (Ab) recognises the pentamer epitopes QQPFP, QQQFP, LQPFP and QLPFP [39]. As expected, the gluten quantitative results (mostly obtained by R5-cELISA analysis; Table S1) reveal that the number of gluten peptides that comprise R5-Ab-recognition epitopes is lower in AN-PEP-treated beers than in AN-PEP-untreated beers, i.e., 0 to 7% (2% of the total) vs. 4 to 23% (15% of the total), respectively (Tables S2 and S10).

G12-Ab recognises the hexamer epitopes QPQLPY, QPQLPF, QPQLPL, QPQQPY and QPQQPF [21,41]. Real et al. [21] identified several motifs associated with the induction of celiac disease in hordein- and gliadin-derived peptides from beer fractions. Among these motifs, they found the PQQPF sequence described by Chen et al. [42] in their phage display study as one of the gliadin toxic motifs. This also explains why in both AN-PEP-treated and -untreated beers there are fewer gluten peptides with an G12-Ab-recognition epitope than there are gluten peptides with an R5-Ab recognition epitope. The G12-Ab-hexamer-recognition epitopes very often lacked the PQQPF sequence. For the treated beers, this was the case for 25 peptides because 22 of these peptides are hexamers IPQQPF and 3 hexamers are FPQQPF. The 25 peptides were identified by the R5-Ab-pentamer-recognition epitope PQQPF through recognition of the hexamer PQQPFP (Table S2). As expected, the number of gluten peptides that comprise G12-Ab recognition epitopes is lower in AN-PEP-treated beer than in AN-PEP-untreated beers, i.e., 0 to 5% (1% of the total) vs. 3 to 15% (11% of the total), respectively (Tables S2 and S10).

Efforts have been made to link the results of gluten analysis of fermented–hydrolysed products, such as beer, to immunogenic activity in patients with CD [43,44]. Therefore, the antibody must specifically recognise CD-immunogenic gluten peptides. If this is not the case, the gluten quantitative result could underestimate or overestimate the net CD-immuno-activity of the product [44]. For this purpose, G12-based ELISA might be more suitable compared to R5-based ELISA [39]. The G12 antibody was generated against the highly immunogenic 33-mer gluten peptide LQLQPFPQPQLPYPQPQLPYPQPQLPYPQPQPF of α-gliadin. Highly immunogenic gluten peptides contain multiple (overlapping) known CD-immunogenic motifs. Several such peptides have been characterised and studies have shown that they are responsible for more than 90% of celiac disease-specific responses [44]. Previous gluten analysis of beer with G12-Ab has shown that the reactivity of G12-Ab correlates with the CD-immunogenicity of the beer [21,40]. Later, it was observed that gluten peptides in beer that are recognised by R5-Ab and not by the G12-Ab are not CD-immunogenic or only have low CD-immunogenicity [44]. R5-Ab overestimates true CD-immunogenicity and should be considered less specific than G12-Ab. The lower specificity of R5-Ab can be explained by the epitopes of the R5-Ab having a shorter recognition sequence (pentamers) than that of the hexamer epitopes of G12-Ab.

The recommendation to use the R5-cELISA method for the quantification of gluten in beer [1] can be justified by the fact that it gives a higher semi-positive result. Thus, gluten content determined by R5-cELISA may increase consumer alertness, although it is unlikely to be directly related to CD-immunogenicity.

According to Wieser et al. [45], the difference in results between different antibody-based methods is useful for ELISA analysis of hydrolysed food or beverages due to the different resistance of the corresponding epitopes observed. Wieser et al. [45] then conclude: “Thus, the concept of gluten content should ideally be changed to gluten immunogenic peptides in beer and other hydrolyzed food, as gluten proteins are actually hydrolyzed and peptides are what remain. A potential risk is that absorption is faster because it may not need digestion in the stomach and intestine to have immunogenic peptides that may be rapidly absorbed”. The latter can be interpreted as a non-negligible risk that also arises from the potentially dangerous peptides identified in the beers treated with AN-PEP.

Moreno et al. [44] stated that even quantifying the total gluten content is very difficult. The same certainly applies to the quantification of individual peptides in complex gluten peptide mixtures. Determining the relative immuno-activity of the different identified peptides is even more challenging. Moreno et al. [44] performed immunoprecipitation of in vitro simulated gastrointestinal-digested hydrolysed PWG gliadin (Prolamin Working Group) and one commercial barley beer sample with the monoclonal G12 antibody covalently coupled to agarose beads to study the immuno-activity of mixtures of peptides. They found that the peptide mixtures of their used AN-PEP-untreated commercial beer, which was tested with low gluten content, still contained hazardous peptides that were not hydrolysed by the malt peptidases. The low gluten content (<20 ppm) of the beer used is probably due to the treatment (silica gel, tannins, filtration, centrifugation) or the cold step of the beer to clarify the beer. These treatments work best on proteins and long peptides, while an additional AN-PEP treatment is usually applied in the production of gluten-free beer. The AN-PEP enzyme then eliminates the shorter peptides with hazardous complete immunogenic motifs. The AN-PEP enzyme is a prolyl endopeptidase and its enzyme activity is very different from the activity of gastrointestinal peptidases and barley malt peptidases. The AN-PEP enzyme delivers a different peptide profile of a beer than that of an untreated beer. It is therefore not possible to draw conclusions from the immuno-activity results of Moreno et al. [44] for the peptides from AN-PEP-treated beers.

In this study, we have found that in both AN-PEP-treated and -untreated beers a significant number of (potentially) hazardous gluten peptides are missed with a R5- and/or G12-cELISA as these peptides do not contain distinct antibody-recognition epitopes (Figure 2 and Table S10). Of the 201 characterised potential hazardous gluten peptides of the AN-PEP-treated beers, 166 (83%) are missed with the R5-cELISA and 191 (95%) with the G12-cELISA (Figure 2 and Table S2). Luckily, with both ELISAs, only a minor number of the 70 peptides of the untreated beers with a complete and hazardous CD-immunogenic motif are missed: 8 (11%) with the R5-cELISA and 13 (19%) with the G12-cELISA (Table S2).

The gluten quantitative results of the beers evaluated in this study were mostly obtained by R5-cELISA analysis (Table S1). For the untreated non-gluten-free beers, i.e., beers with more than 20 ppm quantifiable gluten, the R5-cELISA analysis provides a preliminary estimate of the CD-immunogenicity of the beers. For (AN-PEP-treated) gluten-free beers, i.e., beers with no more than 20 ppm quantifiable gluten, the R5-cELISA might not be thorough enough to assess the potential CD-immunogenicity of the beer. For example, in AN-PEP-treated beers 3 and 4, and surprisingly their untreated reference beers 27 and 28 as well, low gluten concentrations < 10 ppm (R5-cELISA LOQ) are measured. Nevertheless, the gluten peptidome in these counterpart beers is completely different. In the two untreated reference beers, more gluten peptides contain (multiple) P-residues and as a result 3 peptides of beer 27 and 11 peptides of beer 28 are CD-hazardous due to an exact match with CD-immunogenic motifs (Table S2).

The results of this study stress the complementary use of ELISA and MS analysis to properly evaluate the CD-immunogenicity of gluten-free beers, as also suggested by some authors [19,20,33,34]. The untargeted MS analysis of gluten-free beers can be used to verify the absence of hazardous peptides with a complete match to CD-motifs and to determine the number of potentially hazardous peptides that have a partial match to CD-immunogenic motifs. When peptides are later identified that can serve as markers for important potentially hazardous peptides, the presence or absence of these markers can also be used to better evaluate the potential CD-immunogenicity of gluten-free beers. A targeted and not yet validated quantitative MS analysis of gluten-free beers can then be avoided. The use of MS, untargeted and certainly targeted, is not at all obvious due to the need for complex instruments, expensive equipment, high operating costs and specialised personnel, as well as due to the complexity of the beer sample containing many different peptides [45]. The developed method by Segura et al. [22] based on G12/A1 LFIA offers good opportunities to measure the low gluten-immunogenic peptide level in gluten-free beers. Mass spectrometric analysis may then become redundant.

4. Conclusions

In total, 1996 peptides were found in 24 AN-PEP-treated malt beers and 1804 peptides in 13 AN-PEP-untreated malt beers. This clearly demonstrates that by AN-PEP treatment, the gluten peptidome of the beer drastically changes. AN-PEP-untreated malt beers should be considered as hazardous for celiac patients. The comprehensive approach, using EFSA guidelines for immunogenicity, reveals that no CD-hazardous gluten peptides are characterised in all the treated beers after enzyme treatment. Potentially CD-hazardous gluten peptides still remain in 20 of the 24 treated beers. The CD-immunostimulatory potential of AN-PEP-treated beers, based on the set of potentially hazardous gluten peptides, cannot be easily interpreted. The bioaccessibility and bioactivity of the peptides is also of importance. A huge number (83 or 95%) of the 201 characterised potentially hazardous peptides in the gluten-free AN-PEP-treated beers are missed with the R5-cELISA or the G12-cELISA. ELISA analysis alone of these gluten-free treated beers is therefore not sufficient and should be supplemented with MS analysis or another assay to get a realistic picture of the potential risk for celiac patients due to the peptides still present in these beers with partial matches with CD-immunogenic epitopes.