Individual variability in patterns and dynamics of fecal gluten immunogenic peptides excretion after low gluten intake

Celiac disease (CD) is a chronic, multiorgan autoimmune disease that occurs in genetically predisposed individuals with well-known genetic components such as human leukocyte antigen (HLA-DQ2 and HLA-DQ8), an auto-antigen (tissue trans-glutaminase), and gluten [1]. Additional environmental contributors have been suggested to be involved in its development, such as infections, imbalanced intestinal microbiota, and increased intestinal permeability [1‐3]. The pathogenesis of CD involves the passage of gluten immunogenic peptides (GIP) through the intestinal barrier into the lamina propria via the trans- or paracellular route, with consequent activation of both adaptive and innate immune responses [4‐7]. Reportedly, the α-gliadin 33-mer peptide is one of the most dominant immunogenic peptides, which contains three–six different potential T-cell epitopes of CD [8]. Three other gluten peptides are shown to produce most of the immunogenic responses observed in patients with CD [9].

Gluten is a water-insoluble polymorphic mixture of storage proteins, which are mainly categorized as alcohol-soluble prolamins and the alcohol-insoluble glutelins. Cereal grains such as wheat, rye, and barley contain high quantities of gluten, whereas oats show much lower content [10]. Prolamins impart specific functional properties such as viscoelasticity to food product; therefore, these cereals are used in a broad range of foodstuffs [11]. Prolamins are structurally characterized by unique repetitive amino acid sequences, rich in glutamine and proline; therefore, they are not easily digested by gastric and pancreatic enzymes [10].

Dietary gluten proteins are partially hydrolyzed in the stomach over a few minutes to 2–4 h, depending on the diet. Proteolytic degradation primarily occurs in the small intestine as a result of pancreatic enzyme activity, which cleaves polypeptides into small peptides and amino acids that are absorbed by transport systems [12, 13]. Previously, it was assumed that only di- or tripeptides are absorbed thorough the intestines; however, studies have shown that longer-chain gluten peptides resistant to digestion can enter the portal circulation, undergo filtration by the kidneys, and get excreted in urine [14]. Undigested gluten proteins and large peptides that remain intact in the small intestine may serve as substrate for local microbiota able to hydrolyze gluten [15]. The residual undigested peptides finally enter the large intestine, which contains a high density of living bacteria, and are further hydrolyzed over at least 10 h or even several days. The length and the activity of the hydrolytic process are largely dependent on the gut microbiota and the nature of the protein and the food matrix. Food components, including gluten proteins, that are not digested by enzymes and intestinal microbiota undergo fecal elimination [12, 13, 16, 17].

Strict lifelong compliance with a gluten-free diet (GFD) is the only treatment currently available for patients with CD, which implies avoidance of all gluten-containing foods and close attention to cross-contamination [18]. However, complete exclusion of dietary gluten is difficult in real-world practice because of the ubiquitous nature of gluten, social aspects and socioeconomic factors [19‐21]. Consequently, GFD compliance in patients with CD was reported to be between 12 and 90% in adults [22‐24] and between 23 and 98% in children [25], being asymptomatic patients more susceptible to regular gluten ingestion [22, 23]. A recent study reported that frequent gluten consumption showed by excreted GIP was associated with histological lesions, which may lead to future complications as a result of their condition [23].

Currently, a safe threshold of daily gluten intake among those with CD is unavailable, and the immune response to gluten significantly varies among this patient population [26]. Catassi et al. [27] concluded that the daily ingestion of contaminating gluten should be < 50 mg, based on histopathological findings observed in patients who received this dose. Additionally, other authors reported a gluten intake below 30 mg to avoid intestinal mucosal abnormalities [28].

Estimation of fecal and urinary GIP is the most direct method to monitor gluten ingestion [14, 17, 29]. In addition to control the adherence to a GFD, this method may also enable direct and quantitative assessment of gluten exposure, which refers to unintended low gluten intake. Information about the specific time, amount of ingested gluten and expected sensitivity for occasional gluten exposure may be important to design either the frequency of testing or the protocols to assess GFD compliance in patients with CD. Despite other authors have reported information about the process of GIP elimination in feces [17, 30‐32], no data are available to date regarding the dynamic of GIP excretion, the influence of individual variability and the sensitivity of the method after punctual gluten exposure maintaining identical daily diet. In this study, we determined fecal GIP excretion collecting samples from all the depositions of participants under a GFD, with two separated ingestions of 50 mg and 2 g of gluten with minimal dietary variations.

In this article, we describe the dynamics of human fecal GIP excretion in conditions simulating occasional gluten exposure under controlled dietary conditions to study individual variability. This is the first study collecting samples of all the following defecations after gluten intake and controlling diet variables to determine the range of individual variability in the excretion of fecal GIP.

Our results agreed to those reported by other studies. Comino et al. [17] first reported that the ELISA method based on anti-33mer monoclonal antibodies could detect gluten-derived peptides in the feces of patients with CD and healthy volunteers, following ingestion of processed bread containing between 50 mg and 30 g of gluten. These authors observed that the time required for gluten-derived peptide excretion was from 2–4 days in six volunteers. Silvester et al. observed similar findings in patients with CD in whom they confirmed an association between definite gluten exposure and GIP positivity in stools 2–4 days after gluten ingestion [32]. However, it was not fully feasible to accurately establish the expected period of detection owing to the high interindividual variability observed in these studies. Our results obtained in a larger number of participants (n = 20) showed GIP detection between the 2nd and 7th days of a gluten challenge.

A study performed in a small group of healthy participants observed the kinetics of fecal GIP elimination over a week after the start of a GFD, and the authors found high interindividual variability; detection of GIP occurred over > 3 days in some and even until the end of the week in other participants [30]. Although the specific moment and quantity of gluten ingestion were not reported, the results of our study were consistent with those of the aforementioned study with regard to the diversity in GIP elimination patterns. In fact, all the examples described in this publication were represented by the participants in our study: however, our results revealed that the individual GIP excretion pattern may be independent of the amount of gluten ingested, considering the initial time of detection and peak of maximum excreted GIP. The amount of GIP is expected to decrease over time for a single gluten dose; however, interestingly, in both studies, we observed intermittent GIP-positive and GIP-negative samples. Furthermore, in two participants a significantly higher fecal GIP content was observed after several days without any fecal GIP detection. The eventual ingestion of gluten was unlikely because parallel urine test controls were negative. These findings may be attributable to the discrete gluten particles that could remain temporarily encapsulated by food matrixes or occasionally retained in the intestine. Although this phenomenon only occurred after the ingestion of 2 g of gluten, it cannot be ruled out that this pattern was also present in the 50 mg dose, but perhaps GIP were undetectable because the amount of gluten was 40-fold lower.

Roca et al. [31] investigated the dynamics of GIP clearance in stools over 6 days in 18 recently diagnosed pediatric patients with CD, who were provided a GFD, and observed that GIP decreased over time in a non-linear manner; however, GIP recovery decreased in the first 48 h, with minimal detection in most samples between 48 and 72 h. Considering the clearance time for quantities of gluten ranging from 0.7 g to 11.6 g/day used in this study, we observed discrepancies in our results; we could expect at least minimal GIP excretion at 72 h. However, this discordance could be attributed to possible inaccuracies in the food-recall questionnaire used for gluten intake estimation.

With regard to the fecal GIP concentration, previous studies have described a weak association between the amount of gluten ingested and fecal GIP excretion [17, 31, 36, 37]. Our results confirmed this association; we observed significant differences in fecal GIP content between volunteers who were administered 50 mg and 2 g gluten (P = 0.013) with a significantly greater variation when results were determined in the 12–84 h range (P < 0.001). However, as expected, we observed high interindividual variability, with median GIP ranging from 0.08 µg/g–0.22 µg/g for the 50 mg dose and 0.09 µg/g–0.42 µg/g for the 2 g dose. The aforementioned studies reported fecal GIP quantities of 0.4 µg GIP/g feces–7 µg GIP/g feces with daily gluten ingestion ranging from 50 mg to 1 g in patients with CD and 0.2 µg GIP/g feces–29 µg GIP/g feces with a normal GCD in healthy subjects [17]. Roca et al. [37] also reported a mean of 13 μg GIP/g feces (range 0.56 μg/g–47 μg/g) following gluten intake ranging from 0.5 g/day to 10.5 g/day. The disparity between our results and those of previous studies could be attributed to the matrix containing gluten used in the study (capsule with partially purified gluten in this study vs. normal food in previous studies), the methodology used for the estimation of gluten ingestion, and the frequency of gluten ingestion.

In this study, we confirmed the sensitivity of the ELISA method for frequent detection of gluten ingestion of 50 mg, which concurs with results of a previous study [17]. Although the ELISA method showed higher sensitivity than the LFIA test, both techniques showed concordance. Similar results were reported by other studies [17, 36, 37], in which diagnostic sensitivity of the ELISA method ranged from 98.5 to 100% with diagnostic specificity of 100%, and the diagnostic sensitivity and specificity of the LFIA method were 75% and 100%, respectively, showing moderate concordance [17, 37]. False-positive results were discarded in our study because the extended experience in the specificity of the G12 immunomethods either in food or in human samples. According to this assumption, GIP detection was observed after gluten ingestion and subsequently to consecutive GIP-negative samples that confirmed the compliance of the volunteers performing the GFD. Samples from those participants with moderate levels of GIP in feces after the wash-out week were excluded as we could not assure that GIP detection was from the 50 mg gluten dose. Previous studies with volunteers undergoing a GFD showed absence or very low rate of positive excreted GIP [17, 37]. Moreover, we have a rigorous control of the gluten content of the food consumed by the participants (see above in “Materials and methods”).

Discordant results could be explained by differences in the configuration of immunomethods (A1/G12 antibodies for the LFIA test and G12/G12 for the ELISA test) and the different efficiency of GIP extraction methods. Usually, samples with GIP-negative result on ELISA testing also show negative results with the LFIA test; but some samples that show a GIP-positive result by ELISA could have a low GIP content that is below to the limit of detection of the LFIA test [22, 31, 37]. However, some discrepancies may occur; for example, we observed six GIP-positive results using LFIA, although these samples tested negative using ELISA testing. This difference may be associated with the heterogeneous distribution of GIP in feces, or the antibody pair used in each method, which may show differing affinity for some peptides, as mentioned earlier. The large heterogeneity of feces composition, even in the same sample of the same individual was already described by other authors [38‐40]. In this study, we aimed to reduce such heterogeneity by the homogenization of each fecal sample prior to analysis and by the determination of GIP from different aliquots taken from the same sample from each participant with each method. Nonetheless, as LFIA and ELISA tests have different extraction protocols for feces samples, each aliquot was analyzed independently with both tests.

Our results can serve to design the recommendations for the sampling of the immunotechniques to increase the sensitivity and estimate the source of GFD transgressions. Reportedly, patients with CD tend to show frequent dietary transgressions despite efforts to strictly follow a GFD [17, 22, 24, 32]; however, inadvertent gluten intake is difficult to determine. Gluten intake in individuals who are prescribed a GFD, could be secondary to regular dietary non-compliance or inadvertent transgressions. Detection of fecal GIP could likely indicate gluten intake in the preceding 12–72 h. An increase in the frequency of stool tests appears to be a convenient approach to avoid false-negative results in those showing non-compliance with GFD resultant from occasional gluten intake [22, 31]. Stefanolo et al. [22] reported that 62% of patients showed at least one GIP-positive result in weekly stool samples obtained over a month, with a median of three positive results during this period. Considering that minimal daily gluten ingestion (50 mg) can cause mucosal injury in most patients with CD and being it useful to adopt a more realistic approach with regard to a GFD, it may be convenient the use of several fecal GIP tests for a week and separated 3–4 days to include weekdays and weekends (for example, Tuesday/Wednesday and Friday/Saturday). The increase of frequency of stool sample collection could solve any clinical need to improve sensitivity of the method for detection of low-single gluten intakes. For instance, the collection of three stool samples to cover gluten exposure on both weekdays and weekends may increase the diagnostic sensitivity of the ELISA to 61.5% for 50 mg of gluten intake and 96.6% for 2 g gluten ingestion.

Overall, our study highlighted high interindividual variability in excretion time and GIP concentrations, which concurs with the results of previous research in this domain. The heterogeneity observed could be attributed to multiple contributors including gastrointestinal transit time, intestinal permeability, hydrolytic capacity of human enzymes, and intestinal microbiota activity. Several studies have reported that both patients with CD and healthy subjects have gluten-degrading bacteria that can hydrolyze immunogenic peptides and fecal glutenasic activity in individuals is inversely associated with the amount of gluten excreted [15‐17, 41, 42]. These factors may explain the negative results in all the stool samples collected after 50 mg and 2 g gluten ingestions despite positive results for urinary GIP after the second dose in one participant in this study [33]. However, the possibility of a missing sample in this participant might not be discarded. In our study, we controlled the gluten dosage and dietary composition to minimize the variables analyzed. The significant differences among participants highlight the realistic scenario that should be considered when establishing protocols for stool collection for GFD monitoring.

In the present study, the variability found between participants in the detection of GIP in urine [33] and feces makes difficult to establish a general pattern between the individual excretion of GIP in both types of samples. Apparently, some of the factors previously mentioned might affect GIP excretion in urine and feces in like manner, but we observed that elements such as fluid intake may significantly modify the sensitivity of the GIP test in urine samples [33], independently of the individual. Despite the lack of strict associations between both type of samples, it was possible to observe detectable amounts of GIP in almost all individuals after the two gluten intakes in either urine or feces (only 4 participants have negative results in both samples after the 50 mg dosage). Besides, we perceived some coincidences in the detection of GIP in urine and feces in several participants. For instance, considering the initial time of GIP detection after 2 g gluten intake, 4 participants were the first ones to excrete detectable amounts of GIP in urine (3–5 h) and feces (22–25 h), and 2 participants had a delayed detection of GIP (8 h and 51–52 h, respectively). In addition, 3 participants showed low median concentrations of GIP in both types of samples after the 2 g ingestion and, on the contrary, the urine and feces samples from one participant were within the highest GIP concentration medians.

Immunoassays for fecal GIP detection are useful in patients diagnosed with CD and gluten-related disorders. A limitation of this study was that only healthy volunteers were included, which was mainly due to ethical issues. There is no sufficient evidences that the metabolism of gluten proteins is different between patients with CD and healthy individuals; however, previous research has reported that patients with CD could show digestive abnormalities in the digestion process [43]. It has also been reported that intestinal microbiota involved in gluten metabolism could be altered in patients with CD [41]. Future studies in these patient populations may confirm the equivalences and any potential deviations in the dynamics of gluten excretion compared to healthy population. Keeping in mind this concern, this report may be valuable to define the protocols for the application of fecal GIP estimation to assess gluten exposure during follow-up of gluten-induced disorders in real-world clinical practice.