Performance assessment of a new G12/A1 antibody-based rapid ELISA using commercially available and gluten-spiked food samples

Objective Food products with <20 mg/kg gluten can be labeled ‘gluten-free’ according to international regulations. Several antibodies-based ELISAs have been develop to track gluten traces in food products. Among them, R5 and G12 antibody-based ELISAs are the frequently used methods. However, these antibodies have certain limitations. We evaluated the accuracy of G12/A1 antibody-based ‘Glutentox ELISA Rapid G12’ and compared the results with the current reference method i.e., R5 antibody-based ‘Ridascreen R5 ELISA’. Methods In the first step, the performance of Glutentox ELISA Rapid G12 kit was inspected by determination of the threshold value i.e., > or <20 mg/kg gluten in different food products. In the second step, quantification accuracy was assessed by quantification of gluten in gluten-free food products spiked with gliadin reference material. Results In total 47 food products (naturally and labeled gluten-free, and food with traces of gluten) were included. Of them, 29 products were quantified with <20 mg/kg, and 18 with a low level of gluten by both the kits. Six out of 29 gluten-free products were used for the recovery test at different spike levels. Gluten concentration and mean recovery rates of individual kits showed consistency. Conclusion GlutenTox Rapid G12 ELISA could be an appropriate choice for detecting gluten in food products but needs more in-house validation and collaborative tests.


Introduction
Celiac disease (CD) is an autoimmune condition that occurs in genetically predisposed individuals who develop an intense T-cell-mediated immune response on exposure to the gluten protein found in wheat, barley, rye, and related grains [1].Approximately 1% of the world's population is affected with CD [2].A life-long complete elimination of gluten from the diet is the only accepted treatment for CD so far [3].According to the European Commission, Food and Drug Administration (FDA), and Codex Alimentarius, food products containing <20 mg/ kg of gluten can be labeled as 'gluten-free' and are safe for CD patients [4][5][6].However, CD patients, similar to healthy individuals, are strongly dependent on commercially available food, and this is a problem for them since gluten contamination may occur along the food chain [7][8][9].Therefore, an accurate gluten detection tool is needed to quantify the amount of gluten in gluten-free food products [10].To track the minute amount of gluten in food, the ELISA method is the most used and considered a 'method of choice' [8,11].Different antibodies such as R5, G12, α-20, MIoBS, DQ2.5-glia-α3, and Skerritt have been developed to detect gluten traces in food products [8,12].However, they have certain limitations.There are no official reference materials approved, and there is no universal unit of measurement of gluten [10,[13][14][15].
Among above-mentioned antibodies, the R5 antibodybased gluten ELISA test kit i.e.RIDASCREEN ELISA is the most prevalent antibody used for the gluten analysis [8,11].
A1/G12 antibody-based sandwich ELISA is another frequently used ELISA method to detect gluten in the foodstuff [16,17].The G12 antibody efficiently measures the immunotoxic fraction of gluten in native and partially hydrolyzed cereals [18,19].
In the present study, we assessed the efficacy of an improved and rapid version of the traditional wellestablished A1/G12 antibody-based GlutenTox ELISA, that is, AOAC PTM approved G12/A1 antibody-based 'Glutentox ELISA rapid G12' developed by Hygiena Diagnóstica España S.L., Sevilla, Spain.Though, a limited number of studies compared both the ELISA methods (A1/G12 and R5) [15,20,21].In our study, we assessed the performance of the Glutentox ELISA rapid G12 kit using a range of commercially available gluten-free, and partially gluten-free products along with spiked samples.R5 ELISA was used as a reference kit and the gluten quantification as well as the gliadin recovery results obtained from the Glutentox ELISA rapid G12 were matched with the R5 ELISA.

Collection of food products
Food products were purchased randomly from different supermarkets in Ancona, Italy.Commercially available food products 'containing gluten', 'naturally gluten-free', 'labeled Gluten-free', and food products with the label 'may contain traces of gluten' were included in the range of purchased food products.

Gluten estimation ELISA kits
Gluten quantification in food products was assessed through newly developed G12/A1 antibody-based 'GlutenTox ELISA rapid G12' (KIT3075, Hygiena Diagnóstica España S.L., Sevilla, Spain), and the results were compared with Ridascreen Gliadin (R5) ELISA, R-7001 (R-Biopharm, Darmstadt, Germany) ELISA kits at the Celiac Disease Research Laboratory, Department of Pediatrics, Polytechnic University of Marche, Ancona, Italy.During each ELISA run, standard and samples were added in duplicates into pre-defined ELISA wells.During the ELISA process, the recommended manufacturer's guidelines were strictly followed for each ELISA method.For both methods, food samples quantified with gluten <20 mg/kg were considered gluten-free.The absorption value of ELISA standards was assured with the quality assurance certificate provided with the respective ELISA kits.Food products with a gluten level higher than 20 mg/ kg were re-extracted and re-analyzed second time in duplicates before confirming the results.

Gliadin reference material
For spiking the food samples, the PWG-gliadin standard (protein content 90.7%, gliadin content 87.6%) was purchased from AGF e.V. (Detmold, Germany).Gliadin stock solution was prepared as per the manufacturer's instructions.

ELISA kits evaluation strategy
The G12/A1 antibody-based GlutenTox rapid G12 ELISA kit was evaluated in two steps.In the first step, the performance of the ELISA kit was inspected by the determination of the threshold value of the gluten in the different food products (gluten-containing as well as gluten-free) i.e. <20 mg/kg (above or below).In the second step (spike and recovery), the accuracy of the quantification was assessed by the quantification of gluten in gluten-free food spiked with PWG-gliadin reference material at different gliadin spike levels.

Statistical analysis
Quantitative variables were summarized as the mean ± SD and 95% confidence interval (CI).Paired t-test was applied to compare the means between the study ELISA kits for both the gluten concentration and PWGgliadin recovery (%).Cohen's kappa test was run to determine the agreement between the gluten detection ELISA methods.The Pearson correlation analysis was performed to analyze the consistency between both the gluten detection ELISA methods.ELISA results <LOQ were excluded from the statistical tests.Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) IBM version 23, Chicago (IL), USA.
Determination of gluten content in gluten-free food samples by G12/A1 antibody-based GlutenTox ELISA rapid G12

Sample extraction
Briefly, 0.5 grams of homogenized food samples were measured in sterile tubes, and 5 mL of universal gluten extraction solution (UGES) was added.A vigorous vortex was provided.In the case of liquid samples, a 0.5 mL food sample was measured and 4.5 mL of UGES was added.Further, tubes were heated in a water bath at 50 °C for 40 min in the case of non-heat-processed/simple matrix samples/liquid samples without solids, the samples were incubated at room temperature (20-26 °C) for 40 min with mild agitation.At the final step (for all types of samples), tubes were centrifuged at 2500g for 10 min, and transferred the supernatant in a fresh sterile tube.

Gluten quantification
Extracted food samples were treated with ready-to-use sample diluent.In each ELISA run, 100 μL of standard, positive control, negative control, internal control, and sample were loaded into the pre-defined well, and the ELISA plate was incubated for 30 min at room temperature.In subsequent steps 100 μL of the conjugate antibody (A1-HRP antibody), 100 μL of the substrate solution were added with an incubation of 30 min.Between these steps, an ELISA wash was provided to the ELISA plate.At the final step, the reaction was stopped by adding 100 μL stop solution and ELISA plates were read in ELISA reader.

Determination of gluten content in gluten-free food samples by R5 antibody-based Ridascreen Gliadin R5 ELISA Extraction and preparation of samples
Five grams of each food sample were homogenized, 0.25 grams of homogenized samples were measured and 2.5 mL of recommended cocktail solution (R-7016) was added in each tube and heated in a water bath at 50 o C for 40 min followed by a treatment of 7.5 mL with freshly prepared 80% ethanol.Tubes were centrifuged (at least 2500 g for 10 min), and the supernatant was collected.

Gluten quantification
Extracted samples were diluted at 1:12.5 ratio in provided sample dilution buffer, standard and samples were added in duplicates and the conjugate was added to each well followed by a wash of ELISA plate and incubated for 30 min at room temperature.Substrate and chromogen were added, and the ELISA reaction was terminated using stop solution.At the final step, the reaction was stopped by adding 100 μL stop solution and ELISA plates were read in ELISA reader.

Spike and recovery test
The PWG-gliadin standard stock solution was prepared by dissolving 1 mg/ml gliadin standard in 60% (v/v) ethanol corresponding to 876 µg/ml gliadin.This stock solution was diluted with the sample dilutions of the respective ELISA kits to prepare different concentrations of gliadin reference.material (0,5,10,20,50 and 80 mg/kg) for the spiking experiments.The respective amount of gliadin reference solution was added in the food products during the gluten extraction step.Afterward, the spiked sample extraction and the respective ELISAs were performed as described in the ELISA methodology.
The gluten recovery (R) was calculated from the average ELISA concentration (M) and spiked level (S) using the equation R = (M/S) × 100 [22].Recovery of gliadin standards between the range of 80-120% was considered ideal, and recovery between 50-150% was considered relevant as per the already developed spike and recovery protocol by Abbott et al., 2010 [23] (Table 1).

Comparison of the ELISA kits for the testing of gluten traces in food products
From December 2022 to May 2023, 47 commercially available food products ('containing gluten', 'naturally glutenfree', 'labeled Gluten-free', and food products with the label 'may contain traces of gluten') were included in the study and clustered into five different food categories, that is, dairy products (n = 2), legumes (n = 7), meat products (n = 6), cereals (n = 19), and bakery products (n = 13).Of the total food products, 23 were naturally gluten-free, 6 were with 'gluten-free' label, 11 were with the label 'may contain gluten traces' and 7 had gluten as one of the main ingredients.Out of total food samples (n = 47), 29 products were quantified within the accepted gluten threshold (<20 mg/kg) and hence considered gluten-free.Of these 29 gluten-free foods, 23 products were detected with less than the limit of quantification (LOQ) by both ELISA kits.However, one product was quantified with <LOQ by G12/A1 ELISA but R5 ELISA detected it within the acceptable level of gluten (5.9 mg/ kg) i.e. 5-20 mg/kg.Other 5 food products were quantified within the acceptable level of gluten from both the Kits, that is, <200 mg/kg gluten (Table 2).The remaining 18 food products ['containing gluten' (7/7), 'naturally glutenfree' (1/23), 'labeled Gluten-free' (0/6), and food products with the label 'may contain traces of gluten' (10/11)] were quantified with a low level of gluten (20-100 mg/kg) by both the kits (Table 2).A good agreement was observed between both the ELISA methods.Moreover, the paired t-test showed no significant difference in the mean gluten concentration between the ELISA kits (P = 0.32) (Table 2).

Accuracy of the quantification of the G12/A1 antibodybased GlutenTox rapid G12 ELISA and its comparison with R5 antibody-based Ridascreen R5 ELISA: spike and recovery test
To evaluate the matrix effect of the ELISA kits, a spike and recovery test was performed using PWG-gliadin reference material.For this purpose, six food products found gluten-free by both ELISA kits were chosen from the original samples list (Table 2).These products were chicken (Sample ID 438), chickpeas (sample ID 341), lentil (Sample ID 420), corn pasta (sample ID 397), gluten-free crackers (sample ID 446), and rice flour (sample ID 305).They were chosen among items characterized by a difficult gluten extraction process due to high-fat content (cracker), extruded food products (pasta), and legumes [24].
Recovery results showed an overall mean recovery for GlutenTox rapid G12, and R5 ELISA of 97.4 ± 19.4 (95% CI: 91-104; range: 74-141 mg/kg), and 99.4 ± 18.2 (95% CL: 92-105; range: 66-141 mg/kg), respectively.Figure 1 shows a correlation in the gluten concentration and recovery obtained from GlutenTox G12 rapid ELISA and Ridascreen R5 ELISA.The recovery rate of G12/A1 GlutenTox rapid G12 kit and Ridascreen R5 ELISA were similar, and the average mean of recovery for the PWGgliadin standard was in the range of 50-150% for both the ELISA methods (Fig. 2).A similar pattern was achieved on the spiked food matrices mean recovery of PWG-gliadin standard (Table 3).There was a consistency found in the mean gluten concentration mean gluten recovery in the individual kits.However, paired t-test between both the ELISA kits showed no significant differences in gluten concentration and recovery means (P = 0.48).Additionally, Cohen's kappa test showed good strength of agreement between the two methods in terms of recovery, k = 0.66 (95% CI, 0.39 to 0.93), P < 0.001 (individual categories were <50%, 50-100%, >100%) (Table 3) (Figs. 1 and 2).

Discussion
The G12/A1 antibody-based GlutenTox rapid G12 ELISA Kit has AOAC PTM approval with PTM certificate No. #042301.In this kit, the G12 mAb (capture antibody) is coated on the ELISA well, and treatment of A1-HRP antibody (detection antibody) is provided as conjugate, which is precisely the reverse of the traditional GlutenTox G12 test.The G12 antibody is restricted to the detection of the QPQLPY sequence of 33-mer toxic peptide found in gluten fragment while A1 has a broader range of epitope recognition, and apart from the QLPYPQP sequence, it also recognizes for two more sequences (QQPFPQP and QLPFPQP) [18].Therefore, the reversal of the location of antibodies (G12 and A1) is an improved fit with a polynomial function.
As for the ability to detect the threshold quantity of the gluten (>20 mg/kg), we found no significant difference in the mean gluten concentration between the tested ELISA kits (P = 0.32) (Table 2).This result is similar to the study conducted by Hochegger R. et al.,2015 where the authors compared R5 antibody-based ELISA with A1/G12 antibodybased conventional GlutenTox G12 ELISA kits in various food products (flours, bakery, and soy products) and found similar results using either R5 or conventional G12 method [15].
The PWG-gliadin recovery results showed no significant difference in the recovery percentage mean (P = 0.48) and found a strong agreement between both the ELISA kits as analyzed by the kappa test (K = 0.66).Recovery of gliadin standard constantly remained in the acceptable range (50-150%) and yielded consistent results in both the kits (Fig. 2) [23].Moreover, raw data obtained by G12/A1 antibody-based 'GlutenTox rapid G12' ELISA was accurate against the spike levels (Table 3).
Both the ELISA kits use different gluten extraction solutions and calibration standards.However, both the gluten extraction solutions are patented.Though 'GlutenTox rapid G12' detects secalin (gluten found in rye), our random gluten-containing food products selection lacked food products containing rye.Our study included a limited  number of food products for the gluten quantification (n = 47) as well as for the spike and recovery test (n = 6).Nevertheless, we included a wide range of food matrices in the study and selected 6 out of 47 food products for the recovery test that were detected gluten-free by both GlutenTox rapid G12 and Ridascreen R5 gluten ELISA.Moreover, most food matrices, included in the study, make the recovery process complicated.We used PWG-gliadin standard reference material for the spike and recovery test at 6 different spike levels (i.e.0, 5, 10, 20, 50, and 80 mg/kg) in duplicates.These increase the reliability of the results.However, we agree that an extended number of food samples may have provided a more significant result.

Conclusion
Our results suggest that G12/A1 antibody-based GlutenTox rapid G12 test is an appropriate choice for the detection of gluten in food products and performs accurately as the current reference method, that is, R5 ELISA.It is a faster gluten detection method and could be a fair alternative gluten ELISA test to the Ridacreen R5 ELISA.However, needs more in-house validation and collaborative tests.Copyright © 2024 Wolters Kluwer Health, Inc. reproduction of this article is prohibited.

Table 2 .
Quantification of gluten (mg/kg) in different food products by G12/A1 antibody-based Glutentox ELISA rapid G12 and R5 antibodybased Ridascreen Gliadin R5 ELISA a Food products also used for spike and recovery test.
a Indicates mean ± SD.