Soluble IgE‐binding factors in the serum of food‐allergic patients: Possible pathophysiological role of soluble FcεRI as protective factor

Abstract Background IgE‐mediated food allergy is the result of an aberrant immune response involving the interaction of a food allergen with its specific IgE bound to FcɛRI, the high affinity IgE receptor, on mast cells. Allergen‐specific IgE also binds to soluble binding factors, but, their expression and role in food allergy is not well characterized. Here, we assess the prevalence and relevance of soluble IgE binding factors in food allergy and tolerance. Methods We measured serum levels of four IgE binding factors, that is, galectin‐3, galectin‐9, soluble FcɛRI (sFcεRI) and soluble CD23 (sCD23) in 67 adults sensitized to peanut or hazelnut and sFcɛRI in 29 children sensitized to hen's egg. Adults without food allergen sensitization (n = 17) served as healthy controls. We compared serum levels of patients and controls and assessed them, in the former, for links to clinical features including allergy and tolerance. Results Serum levels of sFcɛRI and sCD23, but not galectin‐3 and galectin‐9, significantly differ in food‐sensitized patients as compared to healthy controls. A subgroup (28%) of peanut and hazelnut allergic patients had elevated sFcεRI levels, that were associated with higher total and specific IgE levels. Furthermore, sFcεRI levels were significantly higher in tolerant subjects compared to allergics. Among hazelnut allergic patients, those with high sFcεRI levels tolerated the highest protein amounts in the oral food challenge. Conclusion sFcɛRI but not sCD23, galectin‐3 and galectin‐9 might play a role in the pathophysiology of food allergy. Its functional role or use as biomarker should be assessed in further studies.

hazelnut allergic patients, those with high sFcεRI levels tolerated the highest protein amounts in the oral food challenge.
Conclusion: sFcɛRI but not sCD23, galectin-3 and galectin-9 might play a role in the pathophysiology of food allergy. Its functional role or use as biomarker should be assessed in further studies.

| INTRODUCTION
Food allergy is a frequent, serious and potential fatal allergic disease, with increasing rates over the last decades. [1][2][3] It often starts in early childhood and can be transient, but also persistent into adulthood.
Peanut and tree nut allergies typically persist and bear the potential to cause severe anaphylactic reactions. [4][5][6] Hen's egg allergy can also lead to severe reactions. In most cases, this food allergy is transient, affecting mostly young children, with a high rate of spontaneous remission after a few years. 7,8 Food allergy is caused by specific immunoglobulin E (sIgE) bound to the high affinity receptor for IgE (FcεRI) on mast cells and basophils. Its crosslinking by food allergens leads to degranulation and the release of histamine and other proinflammatory mediators, which cause the clinical symptoms of food allergy. Allergen-specific IgE also binds to its low affinity receptor, CD23, expressed by follicular B cells, T-cells, monocytes, Langerhans cells, eosinophils, and macrophages. 9 Additionally to the membrane bound receptors, soluble IgEbinding factors exist: soluble FcεRI (sFcεRI), soluble CD23 (sCD23) as well as galectin-3 and galectin-9, but their role and relevance in food allergy are not well characterized. 10,11 sFcεRI is a truncated receptor consisting of the alpha chain, the IgE-binding domain of the membrane bound FcεRI complex. In vitro, sFcεRI can block IgE binding to FcεRI on cell surfaces and inhibits basophil activation as well as anaphylaxis in mouse models. 12,13 sFcεRI has been proposed as a biomarker for atopic diseases like atopic dermatitis and allergic asthma. It has also been proposed as a biomarker to identify patients who tolerate drug desensitization protocols. 15 Despite these extensive previous data, little is known about sFcεRI in food allergy. 14,15 The same holds true for sCD23, which results from a proteolytic cleavage of the transmembrane form, for example, this cleavage can be accomplished by an enzymatic activity of the house dust mite allergen Der p 1. 10,16,17 All described isoforms of sCD23 possess the globular lectin head domain and therefore the IgE binding site. However, their different sizes seem to have different functional implications, for example, different potencies to inhibit or enhance IgE production. 18 In addition, sCD23 has been suggested as a biomarker for atopy. [19][20][21][22][23][24][25] Galectin-3 and galectin-9 are carbohydrate-binding proteins and have a large number of intracellular and extracellular binding partners; for example, galectin-3 not only interacts with IgE but also with FcεRI and therefore can activate mast cells in an IgE-dependent and -independent manner. 26,27 Galectin-3 plays a role in a variety of inflammatory processes, including neutrophil adhesion, chemoattraction of monocytes and macrophages as well as migration of dendritic cells. 28 Due to its various interaction partners, elevated galectin-3 serum levels can be found in many disorders including autoimmune diseases and cancer. [29][30][31][32][33] Galectin-9, which is also elevated in various diseases, binds to IgE stronger than galectin-3.
Galectin-9 was shown to inhibit mast cell degranulation by depriving IgE of binding to its antigen in a concentration-dependent manner. 34,35 Although in vitro soluble IgE binding factors interfere with IgE dependent mast cell activation their clinical implication in food allergy remains to be characterized in detail.

The humanized monoclonal anti-IgE antibody omalizumab binds
IgE molecules in the same domain as FcεRI and therefore, binds free IgE circulating in serum without triggering mast cell degranulation.
Clinical trials using this monoclonal antibody in patients with food allergy have shown to be efficacious as these patients developed transient tolerance to higher amounts of allergen consumption. 36,37 Like omalizumab, the soluble IgE receptors have the ability to bind IgE and thus may function as natural protectors from allergic reactions, including food allergy, by reducing the amount of available IgE molecules to bind to membrane receptors. We, therefore, hypothesized that (1) sFcεRI, sCD23, galectin-9, and galectin-3 levels are increased in patients with food allergen-specific IgE, (2)  All patients or their parents gave written informed consent, before they were enrolled in the studies. We included peanut and hazelnut sensitized adults (n = 67) as well as hen's egg sensitized children (n = 29) to reflect food allergy heterogeneity. The control group consisted of 17 serum samples from non-sensitized adults with no history of food allergy, negative skin prick test (SPT) and/or no sIgE toward peanut and hazelnut (non-allergic). Sensitized patients, with detectable elevated sIgE levels to food allergens and/or positive SPT, were divided into two groups according the comprehensive standardized diagnostic workup: sensitized as well as allergic (hereafter allergic, n = 58), sensitized but tolerant patients (hereafter tolerant, n = 38).
Clinical characteristics and demographics of all patients and healthy controls are presented in Supplemental Tables S1 and S2.
There was no significant difference in sex and age between our adult subgroups.

| Routine laboratory assessments
IgE levels were measured using the ImmunoCAP System ® (Phadia Laboratory Systems, Thermo Fisher Scientific Inc, Uppsala, Sweden) at a central laboratory (Labor Berlin GmbH, Berlin, Germany). Serum was analyzed for total IgE and IgE against the following food allergens: peanut, Ara h 1, Ara h 2, Ara h 3, Ara h 8, Ara h 9, hazelnut, Cor a 1, Cor a 8, Cor a 9, Cor a 14 and hen's egg. Total IgE levels >100 kU/l were considered as elevated, specific IgE levels of

| Soluble IgE binding factors
The soluble IgE receptor levels were measured using the following commercially available ELISA kits according to manufacturer's pro- and IgE bound levels were measured in 33 patients. As levels did not differ and were tightly correlated, for remaining measurements only total levels were determined (see Supplement 3). sFcεRI, galectin-3 and galectin-9 levels were given in ng/ml and sCD23 levels were given in U/ml. Patients' serum samples were measured in duplicates.
The optical density was read at 450 nm using the Victor V reader (Perkin Elmer, Waltham, MA, USA).

| Statistical analyses
Statistical analyses were performed using GraphPad Prism Version

| Levels of sFcεRI and sCD23, but not galectin-3 and galectin-9 differ between food allergen-sensitized subjects and healthy controls
Adult patients sensitized (allergic or tolerant) to peanut and/or hazelnut had similar serum levels of galectin-3 and galectin-9 as compared to non-sensitized healthy controls (Figure 1; for patient and healthy control demographics see Supplemental  Figure 1A. Peanut and hazelnut sensitized patients exhibited sCD23 levels ranging from undetectable to 847.4 U/ml ( Figure 1A). Since several healthy controls were also having undetectable sCD23 levels, we used the 25% percentile from healthy controls as cut-off value STEINERT ET AL. Nineteen out of 29 (=66%) children sensitized to hen's egg showed sFcεRI levels greater than 3.26 ng/ml.
In contrast to sFcεRI, low sCD23 levels were not associated with any of the assessed features (Supplement 4).

| Soluble FcεRI are linked to total IgE levels as well as allergen testing in food allergy
Fourteen of the 19 (= 74%) peanut and hazelnut sensitized adults with elevated sFcεRI levels were allergic toward their culprit allergen. Therefore, we also investigated the link between total IgE levels as well as allergen sensitization in peanut and hazelnut allergic patients with normal and elevated sFcεRI levels ( Table 1).
Allergic patients with elevated sFcεRI levels had significantly higher total IgE levels compared to patients with low levels (511  Figure 2C).
Such differences were not seen in regard to hazelnut. Furthermore, nut allergic patients with elevated sFcεRI levels had more frequently documented diagnosed secondary pollen-associated food allergies (no sensitization against storage proteins: Ara h 1, Ara h 2, Ara h 3, Ara h 6 or Cor a 9, Cor a 14).
Six out of eight (=75%) children with hen's egg allergy showed elevated sFcεRI levels. However, due to low numbers no link between total IgE, sIgE and SPT was investigated.   Figure 3A). However, as shown in Figure 3B, sIgE levels were better predictors for food allergy in our small cohort than sFcεRI levels.
In peanut tolerant patients sFcεRI serum titers were slightly, but significantly higher than in allergic ones (2.62   Measurement of sCD23, galectin-3 or galalectin-9 did not result in any differences among peanut and hazelnut tolerant or allergic patients in our nut allergic patient cohort (Supplement 5).  As third pathophysiological possibility, also, both mechanisms could be present in food allergic patients depending on the atopic background and the status of the allergy. Given this circumstance, the use of sFcεRI as a stand-alone biomarker might be limited, but could, maybe together with other factors, give a clearer picture of the allergic status of a patient and might be of use as a follow up marker during tolerance induction, for example, with the newly developed oral immunotherapy for peanut sensitized children. 47 This project was set out as an exploratory study with low patient numbers and a retrospective cross-sectional single center study design.

| DISCUSSION
For ethical reasons no age-matched, non-allergic controls exist for the hen's egg sensitized children and sCD23, galectin-3 and galectin-9 were not measured in this cohort, due to limited amount of serum sample. However, to our knowledge this is the first study indicating a potential relevance of soluble IgE binding factors, especially sFcεRI in food allergy.
As the above mentioned limitations could potentially have influenced the study outcome a prospective long-term follow up study with a larger cohort is needed to further clarify the relevance of sFcεRI in nut and other food allergies and its precise role in the pathophysiology.