Impact of STAT6 Variants on the Response to Proton Pump Inhibitors and Comorbidities in Patients with Eosinophilic Esophagitis

Proton pump inhibitors (PPIs) are the first-line drug for eosinophilic esophagitis (EoE), although it is estimated that there is a lack of histological remission in 50% of patients. This research aimed to identify pharmacogenetic biomarkers predictive of PPI effectiveness and to study their association with disease features. Peak eosinophil count (PEC) and the endoscopic reference score (EREFS) were determined before and after an eight-week PPI course in 28 EoE patients. The impact of the signal transducer and activator of transcription 6 (STAT6), CYP2C19, CYP3A4, CYP3A5, and ABCB1 genetic variations on baseline PEC and EREFS, their reduction and histological response, and on EoE symptoms and comorbidities was analyzed. PEC reduction was higher in omeprazole-treated patients (92.5%) compared to other PPIs (57.9%, p = 0.003). STAT6 rs12368672 (g.18453G>C) G/G genotype showed higher baseline PEC values compared to G/C and C/C genotypes (83.2 vs. 52.9, p = 0.027). EREFS reduction in STAT6 rs12368672 G/G and G/C genotypes was higher than in the C/C genotype (36.7% vs. −75.0% p = 0.011). However, significance was lost after Bonferroni correction. Heartburn incidence was higher in STAT6 rs167769 (g.27148G>A) G/G patients compared to G/A (54.55% vs. 11.77%, p = 0.030). STAT6 rs12368672G>C and rs167769G>A variants might have a relevant impact on EoE status and PPI response. Further research is warranted to clarify the clinical relevance of these variants.


Introduction
Eosinophilic esophagitis (EoE) is an immune-mediated disease characterized by an eosinophil-predominant inflammation restricted to the esophagus and by the appearance of symptoms of esophageal dysfunction [1].It is considered a type 2 inflammatory response mainly triggered by food antigens [1,2].In EoE, dietary allergens activate the esophageal epithelium and trigger the production of T-helper (Th)-2 cytokines that stimulate different immune cells.Interleukin (IL)-4 and IL-13 locally produced activate the signal transducer and activator of transcription 6 (STAT6), which induces eotaxin-3 (also called CCL26) expression.Eotaxin-3 promotes the recruitment of eosinophils from blood to esophageal tissues, leading to eosinophilic inflammation [3].Additionally, lymphocytes in the esophagus produce other Th-2 cytokines, such as IL-5 and IL-15, that contribute to the perpetuation of inflammation, altering the barrier function, increasing esophageal permeability to dietary antigens, and promoting tissue remodeling [2], which in the long term results in fibrosis that is manifested as rings and strictures in endoscopy.These alterations determine a variety of symptoms that include dysphagia, food impaction, heartburn, regurgitation, vomiting, nausea, and abdominal pain [1,4].
Fist-line treatment options for EoE include dietary modifications, proton pump inhibitors (PPIs), and swallowed topical corticosteroids [4].PPIs are widely accessible and convenient drugs, therefore representing the most commonly used therapy at all ages and in most settings [5][6][7].The primary mechanism of action of PPIs is the blockage of the gastric H, K-ATPase, thereby inhibiting gastric acid secretion [8].The CYP2C19 genotype-informed metabolic phenotype was shown to impact PPI pharmacokinetics, safety, and efficacy in several diseases, including peptic ulcer, gastroesophageal reflux disease, and Helicobacter pylori eradication.For this reason, the Clinical Pharmacogenetics Implementation Consortium (CPIC) has published a pharmacogenetic guideline recommending omeprazole, pantoprazole, and lansoprazole dose adjustments based on the CYP2C19 phenotype [9].To further optimize and personalize EoE treatment with PPIs, other biomarkers have been proposed [10,11], albeit their clinical relevance has not yet been proven.One of these biomarkers is STAT6, which has been identified as a PPI target in EoE patients, as it is involved in drug pharmacodynamics since PPIs block STAT6 binding to the CCL26 promoter [10].Therefore, the aim of this research was to analyze the impact of CYP2C19 and STAT6 genetic variation (and variation in other relevant pharmacogenes such as CYP3A4, CYP3A5, or ABCB1) on PPI response, and of STAT6 variants on EoE baseline status (i.e., peak eosinophils count, endoscopic phenotype, and symptoms) and comorbidities.It should also be noted that with STAT6, as with other pharmacogenes, there is a fine line between human genetics and pharmacogenetics; some biomarkers can be both diagnostic and pharmacogenetic.Thus, this research also aimed to understand the role of STAT6 genetic variants in the development and progression of EoE.This research is part of the La Princesa Multidisciplinary Initiative for the Implementation of Pharmacogenetics (PriME-PGx) [12].

Impact of STAT6 Genetic Variation in Baseline Scores
A higher baseline peak eosinophil count (PEC) was observed in patients with the STAT6 rs12368672 G/G genotype compared to those with the G/C + C/C genotypes (p = 0.027); likewise, a trend towards higher baseline exudates, rings, edema, furrows, and stricture (EREFS) scale (or endoscopic reference scale) was observed in patients with the STAT6 rs12368672 G/G + G/C genotypes compared to those with the C/C genotype (p = 0.086) (Table 2).However, significance was not reached after Bonferroni correction for multiple comparisons (p-value established at p < 0.004).No further differences in baseline scores and disease duration according to STAT6 genotypes were observed (Table 2).

Impact of STAT6 Genetic Variation on Symptom Onset and Comorbidity
Heartburn incidence was higher in patients with the STAT6 rs167769 G/G genotype compared to G/A (54.55% vs. 11.77%,p = 0.030) (Table 3).Also, a tendency towards higher heartburn incidence in the STAT6 rs324011 G/G genotype compared to G/A + A/A genotypes (50.00% vs. 16.67%,p = 0.091), and in the STAT6 rs12368672 G/G genotype compared to the G/C + C/C genotypes (46.15% and 13.33%, p = 0.096), was observed (Table 3).A tendency towards lower food allergy incidence in patients with the G/G + G/C genotypes for STAT6 rs12368672 compared to those with the C/C genotype (p = 0.074), and a tendency towards higher asthma incidence in carriers of the STAT6 rs324015 A/G genotype compared to those with the G/G genotype (p = 0.097), were observed (Table 3).

Impact of Genetic Variation on Effectiveness Variables and Histological Response
Patients with the STAT6 rs12368672 C/C genotype showed a lower reduction in EREFS score compared to patients with G/C + G/G genotypes (p = 0.011) (Table 5).Furthermore, a higher EREFS score reduction was observed in individuals with ABCB1 rs2032582 T/T+T/G genotypes compared to those with A/A+G/A+G/G genotypes (p = 0.045) (Table 5); none of these differences reached the threshold for statistical significance after the Bonferroni correction for multiple comparisons (p-value established at p < 0.004).Regarding CYP2C19, no significant differences in PEC or EREFS score reduction were observed, although an approximately 35% lower PEC reduction was observed in rapid metabolizers (RM) compared to normal, intermediate, and poor metabolizers (NM, IM and PM, respectively) (p = 0.359).No significant differences were observed for PEC or EREFS score reduction according to the remaining genotypes (Table 5).No associations between genotypes or phenotypes and histological response were found (Table S1).

Impact of Genetic Variation on Symptom Variations
All patients with the CYP3A4 *1/*1 genotype suffered dysphagia (n = 26), which disappeared in seven of them after treatment (26.92%), compared to one out of the two patients with the CYP3A4 *1/*22 genotype, in whom it also disappeared (p = 0.021).Fifteen patients with the STAT6 rs1059513 A/A genotype (71.42%) suffered from food impaction; in fourteen of them it improved after treatment (93.33%) compared to the seven patients with the STAT6 rs1059513 A/G genotype who suffered food impaction (100%); in three of them (42.85%), the symptom was relieved (p = 0.011).No further association in symptom changes and genotypes or phenotypes were found.

Discussion
PPIs are the most widely used first-line treatment for EoE due to their safety profile, easy administration, and low cost [5,6].However, it is estimated that only 50% of patients under PPI treatment reach histological remission [13], and personalizing pharmacological therapy might be a key tool to increase this rate of response.Thus, the objective of this research was to find useful predictors of PPI response in patients with EoE, which might be useful not only to guide drug selection but also dose optimization, which is common and highly relevant in the treatment of EoE with PPIs.
Eighty-nine percent of patients included in this research were men (8:1 ratio), which is concordant with the well-defined higher incidence of EoE among males [14][15][16].As in other series [17], omeprazole was the most commonly used PPI (53.6%), and it led to higher PEC reduction compared to other active ingredients.These results are partially consistent with previous research, in which a trend towards a higher omeprazole and esomeprazole effectiveness was observed [17].In Spain, the cost for an omeprazole 20 mg capsule is EUR 0.09, compared to EUR 0.45, 0.57, and 0.62 for esomeprazole 20 mg, lansoprazole 30 mg, and pantoprazole 40 mg, respectively [18].This five-to seven-fold price difference might justify omeprazole predominance.The better response observed in patients treated with omeprazole may be explained by their lower disease duration, as longer disease duration is associated with a transition to a fibrotic state, resulting in a lower response to pharmacological therapy [19].
Several studies have shown the role of STAT6 in the eosinophilic inflammation of EoE by inducing CCL26 expression [3].However, its role as a pharmacogenetic biomarker in this disease and the association with its symptoms and comorbidities have not been widely studied.In this work, the g.18453G>C far upstream variant in STAT6 (rs12368672) was related to lower baseline PEC and to lower EREFS score reduction.In a previous article, an association between this variant and higher eosinophil/hpf levels previous to PPI treatment was found [11].Although these results appear to be contradictory, it should be taken into account that STAT6 is located in the reverse strand of DNA, the probe used in both studies (C_31186828_10, Applied Biosystems, Thermofisher, Waltham, MA, USA) provides the genotype in the forward strand, and this variant entails a change between complementary nucleotides, which might act as a cofounding factor [20].Thus, a clear description of the nomenclature and reference sequence is needed to enable comparison and conclusion drawing.Additionally, differences in the study populations (adult versus pediatric) should also be considered.In addition, three articles have supported the relevance of this variant in food allergy, which is consistent with the trend observed in our work towards a higher prevalence of food allergy in patients with the STAT6 g.18453G>C (rs12368672) C/C genotype compared to the G/G and G/C genotypes [21][22][23].The lack of a clear description of the reference sequence used, as mentioned above, prevents us from jointly weighing the direction of the association.Nevertheless, the fact that different and independent articles have reported an effect of the same STAT6 variant suggests its potential relevance, not only in EoE but in its comorbidities and other allergic or atopic diseases.Additionally, patients with the STAT6 g.27148G>A (rs167769) G/G genotype suffered from heartburn with higher frequency than patients with the G/A genotype, and a similar trend was observed for the g.28741G>A (rs324011) and g.18453G>C (rs12368672) G/G genotypes, probably due to its high, although not complete, linkage disequilibrium with rs167769 [11].Lastly, patients with the STAT6 g.41214A>G (rs1059513) A/A genotype had a higher frequency of food impaction improvement after treatment compared to those with the A/G genotype.To our knowledge, this is the first work to find such associations.Although these results should be considered cautiously, they suggest the relevance of STAT6 genetic variation in EoE baseline status, symptoms, and comorbidities, shedding some light onto its impact on EoE mechanism of action.Due to the fine line between human genetics and pharmacogenetics, these results may open a way to predicting the risk of EoE development, thus facilitating early diagnosis, which would possibly lower the progression to a fibrotic phenotype, and the discovery of new targets for the treatment of this illness.
In our research, PEC reduction was approximately 35% lower in CYP2C19 RMs compared NMs, IMs, and PMs, although this difference was not statistically significant, likely due to the reduced sample size.However, CPIC only considers dose adjustments for ultrarapid metabolizers (UMs) [9], a phenotype absent in our study and which shows even higher enzymatic activity than that of RMs; therefore, greater differences with respect to NMs, IMs, and PMs could be expected.Further research with increased sample sizes and including UMs is needed to assess whether CYP2C19 RMs may also benefit from a PPI dose increase in EoE treatment, as shown in a different study [24].
Lastly, PPIs are proposed to be substrates and inhibitors of the ABCB1-coded transporter, P-glycoprotein (P-gp) [25].In this research, a nominal association between the ABCB1 g.186947T>G/A (rs2032582) G/G, G/A, and A/A genotypes and lower EREFS reduction was observed compared to the T/T and T/G genotypes.Concordantly, these ABCB1 genetic variants were found to alter PPI pharmacokinetics or pharmacodynamics in previous studies [26,27].However, ABCB1 structural and functional characterization is required prior to concluding the clinical relevance of variant-phenotype associations.
This study is intended as an exploratory and descriptive study, where statistical significance does not imply clinical relevance, especially in light of the limited sample size.Therefore, the main limitation of this study was the small sample size available, especially for the variants with a low prevalence within the population, which reduces the statistical power.This, along with the low incidence of ADRs associated with PPI treatment at standard doses [28,29], led to the lack of meaningful conclusions in the analysis of drug tolerability.Nevertheless, further research is needed on the safety of long-term treatment with high-dose PPIs [30,31] and on the ability of STAT6 genetic variants to predict this response.In addition, a better analysis of the linkage disequilibrium between STAT6 variants should be performed, which might lead to allele and posterior phenotype definition, and a clear description of the reference sequence used is also needed to allow for comparison between results [32,33].Furthermore, a functional characterization of STAT6 variants (i.e., their impact on STAT6 expression and/or function) would also be of interest to better predict their clinical consequences.Thus, further research is warranted to clarify the impact and clinical relevance of these associations, not only in adult patients but also in pediatric EoE patients.However, this study also has some strengths.It is a prospective study analyzing the main candidate genes for PPI response and EoE development in a population representative of the real EoE population, and in which variability was reduced by standardizing PPI treatment, duration, and dose.Additionally, this research also sheds some light on the impact of STAT6 genetic variation on the EoE mechanism of action, which might open an avenue to its study as a diagnostic and pharmacogenetic biomarker.

Study Population and Procedures
This was an observational prospective study on 28 patients with newly diagnosed EoE according to current criteria [14] at two Spanish hospitals: Hospital Universitario de La Princesa (Madrid, Spain) and Hospital General de Tomelloso (Ciudad Real, Spain).They routinely attended the gastroenterology departments of either hospital as part of routine clinical practice between February 2018 and November 2020.They all gave informed consent to participate in the present study.The inclusion criteria were as follows: to be an adult patient newly diagnosed with EoE, and therefore naïve to EoE treatment, and to have been prescribed but not yet have started omeprazole, esomeprazole, lansoprazole, or pantoprazole treatment at least at double dose for eight weeks.The only exclusion criteria were pregnancy or lactation.Institutional review boards at both sites approved the study protocol.
At the baseline endoscopy, three esophageal biopsies were obtained from each proximal and distal esophagus for histopathological evaluation after hematoxylin and eosin staining.Esophageal eosinophilia was defined as an eosinophil count of ≥15 cells per highpower field (hpf) (corresponding to an area of 0.24 mm 2 ) in one or more biopsy specimens at any esophageal level.The exclusion of other potential causes of the esophageal eosinophilia and the absence of eosinophilic infiltration in gastric and duodenal mucosa biopsies led to EoE diagnosis when symptoms of esophageal dysfunction were present [14,34].In addition, three additional biopsies were collected at the mid esophageal third for investigational purposes.Patients underwent an eight-week period of PPI therapy with omeprazole, esomeprazole, lansoprazole, or pantoprazole.Drug selection was performed based on a physician-patient decision, and the dose administered was at a least double dose in every patient: omeprazole 40 or 80 mg daily, esomeprazole 40 or 80 mg daily, lansoprazole 60 mg daily, or pantoprazole 80 mg daily (Table S2).Endoscopy was repeated after an eight-week treatment.Endoscopic features were assessed by the EREFSscale [35], and PEC was counted in all esophageal biopsies.Percentage reductions from baseline PEC and EREFS to after an eighth-week PPI course were analyzed as effectiveness variables.Those patients that achieved histological response (i.e., less than 15 eosinophils per hpf in the biopsy) after PPI treatment were classified as responders, and those who did not (i.e., 15 or more eosinophils per hpf) were considered non-responders.Additionally, clinical data including demographics, symptoms, disease duration (defined from symptoms onset to baseline endoscopy), and atopic background were collected from all patients' clinical records [36].
This study was approved by the Research Ethics Committee of the Hospital Universitario de La Princesa (PI17/0008, registration number 3107, 8 June 2017) and all subjects gave informed consent to participate.During the research, the Declaration of Helsinki, the Good Clinical Practice guidelines, and the Patient Autonomy Law (41/2002) were followed [37,38].

Esophageal Biopsies Processing, Genotyping, and Phenotyping
The three endoscopic biopsies obtained for research were collected and frozen under liquid N 2 conditions and then disrupted with a mortar and pestle, grinding them to a fine powder used for DNA extraction with a NZYtech tissue genomic DNA isolation kit (BM13502) (NZYtech, Lisbon, Portugal).
Bonferroni correction for multiple comparisons was performed to control for type I error.The correlation between effectiveness variables, baseline scores, and disease duration was calculated with Pearson's correlation coefficient.The Pearson coefficient (r) is shown for significant associations (p < 0.05).Additionally, a Chi 2 or Fisher's exact test were performed to search for associations between STAT6 genetic variation and histological response (responders vs. non-responders) and symptoms and atopic disease incidence.

Conclusions
STAT6 g.27148G>A (rs167769), g.18453G>C (rs12368672), and g.41214A>G (rs1059513) may have potential relevance as biomarkers that are predictive of EoE development and PPI response.However, their exact role on the disease and how it can be used to guide treatment require further investigation.

Institutional Review Board Statement:
This study was conducted according to the guidelines of the Declaration of Helsinki, and was approved by the Ethics Committee of Hospital Universitario de La Princesa (PI17/0008, registration number 3107, 8 June 2017).
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Table 1 .
Demographic characteristics and symptoms and atopic disease incidence.

Table 2 .
Baseline scores and disease duration according to sex, treatment, and STAT6 variants.

Table 3 .
Differences in symptom onset and comorbidity incidence according to STAT6 variants.

Table 4 .
Percentage of PEC reduction and EREFS score reduction from baseline to after eight-week treatment according to histological response, sex, and treatment.
Data shown as median and interquartile range (IQR).PEC: peak eosinophils count; EREFS: endoscopic reference score.*: p < 0.05 compared to responders.$ : p < 0.004 compared to responders and to esomeprazole, pantoprazole, and lansoprazole treatment (threshold for significance adjusted after multiple comparisons).

Table 5 .
Percentage of PEC reduction and EREFS score reduction from baseline to after eight-week treatment according to genotypes and phenotypes.
+ G/G genotypes or with ABCB1 rs2032582 T/T + T/G genotypes (nominally significant); no association reached the significance threshold after Bonferroni correction for multiple comparisons (p < 0.004).