Severe food allergy reactions are associated with α-tryptase

Abstract

Background:

Increased TPSAB1 copy number encoding α-tryptase is associated with severe reactions in adults with Hymenoptera venom allergy, systemic mastocytosis, and idiopathic anaphylaxis.

Objective:

The primary objective was to assess the association between α-tryptase and food allergy severity.

Methods:

119 subjects underwent tryptase genotyping, 82 from an observational food allergy cohort at National Institute of Allergy and Infectious Diseases (NIAID) and 37 from a cohort of children who reacted to peanut oral food challenge (OFC) at Lurie Children’s. The primary predictor was presence or absence of α-tryptase. The primary outcomes for both cohorts were measures of food allergy reaction severity. Secondary outcomes included OFC symptom scores [(Bock/PRACTALL and Severity Grading Score for Acute Reactions (SGSAR)]. Correlation between total α-tryptase isoforms and OFC scores was also assessed to account for gene dosage effects.

Results:

The presence of α-tryptase was associated with increased prevalence of food-triggered anaphylaxis as compared to those with only β-tryptase (p=0.026) in the NIAID cohort. Similarly, only 1/6 (17%) subjects in the OFC cohort with only β-tryptase had a severe reaction while 20/31 (65%) of subjects with α-tryptase had a severe reaction (p=0.066). Subjects with α-tryptase also had higher total SGSAR scores as compared to subjects with no α-tryptase (p=0.003). In addition, there were also significant positive correlations between α-tryptase isoform copy numbers and both higher total SGSAR and Bock/PRACTALL OFC scores (p=0.008 and 0.003, respectively).

Conclusion:

The presence of α-tryptase correlates with increased prevalence of anaphylaxis or severe reaction to food as compared to subjects without any α-tryptase.

Capsule Summary

This is the first report investigating TPSAB1 and tryptase genotype as a potential biomarker for food allergy reaction severity.

Introduction

Food allergies affect a significant proportion of the population, and the prevalence of food allergy continues to increase^1–3. While there have been recent advances in the diagnosis and treatment of food allergy, clinicians still face challenges in predicting reaction severity. Previous studies have identified cofactors such as sleep deprivation, exercise, and febrile illness as contributors to increased severity and lower reaction threshold in patients with food allergy^4,5. While basophil activation testing (BAT) has shown promise in identifying severe food allergy when compared to traditional antigen-specific IgE (sIgE) and skin prick testing (SPT), BAT is not routinely available or used in clinical practice^6–8. Component resolved diagnostics (CRD) with measurement of sIgE to individual allergen components have also been reported to help predict reaction severity^9–11. However, CRD is not available for all foods and result interpretation varies by geographic location and patient age^11–13. As such, there is currently no reliable or readily available clinical biomarker that accurately identifies patients with food allergies who are at risk for severe life-threatening reactions.

Elevated baseline serum tryptase (BST) (defined as >11.4 ng/mL) has long been linked to increased reaction severity in patients with Hymenoptera venom allergy^14–16. Additionally, detection of elevated serum tryptase levels following an acute allergic reaction can assist in the diagnosis of anaphylaxis irrespective of trigger. However, the utility of tryptase measurement in food allergy remains unclear¹⁷. While a small study demonstrated that higher BST may be associated with increased risk for severe anaphylaxis in children with food allergies¹⁸, children with food allergy often have low BST levels overall¹⁹. Some studies have reported that comparison of baseline and peak tryptase levels can be useful in diagnosis of food-induced anaphylaxis^19,20, but this requires serial tryptase measurements during acute reactions which limits its application in clinical practice. Furthermore, despite an increase in serum tryptase levels from baseline during food allergy reactions, these changes frequently do not meet current diagnostic criteria for significant tryptase elevations during anaphylaxis²¹, though new criteria have been proposed that would be more sensitive among children with low BST levels in confirming the clinical diagnosis of anaphylaxis²². Given these limitations, alternative strategies may be needed to further elucidate the role of tryptase in food allergy.

Elevated BST in the general population results from increased α-tryptase encoding copy number at TPSAB1 in the vast majority of individuals; this is a genetic trait known as hereditary α-tryptasemia (HαT)^23,24. Patients with normal TPSAB1 copy numbers may have 0, 1, or 2 copies of α-tryptase, while patients with HαT have at least 2 copies on the same allele, with up to 10 α-tryptase replications being reported to date²⁵. It has previously been shown in adult patients with Hymenoptera venom allergy and idiopathic anaphylaxis that patients with HαT are at increased risk for severe anaphylaxis^26,27. Similarly, in patients with systemic mastocytosis (SM), the prevalence of HαT is approximately three-times greater than the general population and an independent risk factor for severe anaphylaxis^{26, 28, 29}. Mechanistically, increased α-tryptase expression is linked to increased formation of mature α/β-tryptase heterotetramers that are released during mast cell degranulation and contribute to sensitization of mast cells to vibration and vascular endothelial cell permeability in vitro^26,30. Indeed, the increased prevalence of HαT among individuals with SM has been proposed to result from the distinct properties of α/β-tryptase heterotetramers, and the prevalence of any α-tryptase containing allele has been reported to be significantly increased among individuals with SM³¹.

Tryptase gene composition as a potential biomarker for food allergy severity has not been studied. The primary objectives of this study were to assess the associations between HαT, tryptase gene composition, and food allergy reaction severity. We initially performed observational analyses in a cohort of patients with known food allergy to assess differences in prevalence of anaphylaxis followed by a small pilot study of patients who underwent peanut oral food challenge (OFC) to evaluate the severity of clinical reactions. We hypothesized that increased relative α-tryptase expression would be associated with higher prevalence of anaphylaxis to food and increasingly severe reactions, and consequently, that patients with HαT would be more likely to have severe reactions to food.

Methods

Study Cohorts

For the retrospective cohort, tryptase genotyping was performed initially on banked samples from subjects previously enrolled on a natural history of food allergy protocol between 2015 and 2020 at the National Institute of Allergy and Infectious Diseases (NIAID) in Bethesda, Maryland (NCT02504853). To be eligible, subjects had to have at least one IgE-mediated food allergy, defined as a history of a convincing type I hypersensitivity reaction to the food in addition to positive food-specific IgE (≥ 0.35 kU_A/L), or for subjects who had never ingested the implicated food, a sIgE level to the respective food ≥ 95% positive predictive value^32,33. Individuals with a history of immunodeficiency, mastocytosis, mast cell activation syndrome, or known genetic disorder were excluded from the study. Informed consent was provided by all participants or their legal guardians. Demographic information, clinical characteristics, and detailed allergy history were obtained from patient and family interview and review of medical records.

For the peanut allergy cohort, subjects who reacted during standardized in-office open peanut OFC between January 2012 and December 2021 were recruited from the Division of Allergy and Immunology at Ann and Robert H. Lurie Children’s Hospital of Chicago. Subjects who chose to participate in the study completed a short electronic survey about demographic and clinical characteristics at time of enrollment, and genomic DNA samples were tryptase genotyped in a blinded fashion. Allergy testing results, including SPT wheal size, whole peanut sIgE, and peanut component testing (PCT) sIgE were also evaluated. Information about the OFC including the approximate amount of peanut protein consumed prior to reaction and details about reaction severity was extracted from clinical notes and flowsheet documentation. All subjects had at least one documented objective symptom prior to termination of the OFC. Study data were collected and managed using REDCap electronic data capture tools hosted at Northwestern University³⁴.

Food Allergy Severity

The primary outcome in the NIAID cohort was history of at least one episode of anaphylaxis to food by patient or family report with verification of symptoms and classification by a clinician based on NIAID/FAAN criteria³⁵. Additionally, the most severe reported anaphylactic reaction was graded based on the Severity Grading Scale for Acute Reactions (SGSAR)³⁶ and used as a secondary outcome measure.

In the peanut allergy cohort, reactions to peanut challenge were designated as either a severe reaction (defined by symptoms as reported on published anaphylaxis consensus statements^35,37 and either documentation of epinephrine administration or diagnosis code of anaphylaxis in clinical notes) or mild reaction. Secondary outcome measures included calculated OFC scores to provide increased granularity of individual symptom severity, including both a modified Bock/PRACTALL score^38,39 and SGSAR³⁶ score.

Tryptase Genotyping

Tryptase genotyping was performed on extracted genomic DNA from stored biospecimens (blood or saliva) from the NIAID cohort and buccal swabs from the peanut allergy cohort by digital droplet PCR (ddPCR) as described²⁴.

Statistical Methods

Descriptive statistics summarized demographic and clinical characteristics for both cohorts, with continuous variables expressed as median and interquartile ranges (IQR) and categorical variables by counts and percentages. Fisher’s exact test or Kruskal-Wallis tests evaluated associations between α-tryptase isoforms and clinical characteristics. Fisher’s exact test assessed for associations between presence of α-tryptase (0 versus 1 or more copies) and the primary outcome variable of food allergy reaction severity. Wilcoxon tests assessed differences between presence of any α-tryptase isoforms and secondary OFC symptom scores (SGSAR and Bock/PRACTALL scores). Spearman’s rank correlation was also performed between total number of α-tryptase isoforms and OFC symptoms scores (SGSAR and Bock/PRACTALL) to assess for possible gene dosage effects. Analyses were conducted in R version 4.0.2 (R Core Team 2020) and GraphPad Prism version 8.4.2. p-values <0.05 were considered significant.

Results

Study Population

A total of 119 subjects underwent tryptase genotyping among the two groups, 82 in the observational NIAID food allergy cohort and 37 in the peanut allergy cohort. The majority of subjects in both groups were male and white. The median age in the observational food allergy cohort was higher than that of the peanut allergy cohort (9 versus 3 years, respectively). Most subjects had at least one other co-morbid allergic condition. Complete demographic and clinical characteristics of the two groups are summarized in Table 1.

Table 1.

Summary of self-reported demographic and clinical characteristics of entire study population.

	NIAID FA Cohort			Peanut OFC Cohort
	History of FA Anaphylaxis (n=61)	No History of FA Anaphylaxis (n=21)	P-value	Severe Reaction (n=21)	Mild Reaction (n=16)	p-value
Female [n, (%)]	21 (34.4)	5 (23.8)	0.529	8 (38.1)	6 (37.5)	0.999
Age in Years [median (IQR)]	9 (4,13)	7 (4,14)	0.562	3 (1,9)	2.5 (1,7)	0.632
Self-Identified Race [n, (%)]
Asian	7 (11.5)	5 (23.8)	0.418	1 (4.8)	1 (6.2)	0.486
Black	8 (13.1)	3 (14.3)		0 (0.0)	0 (0.0)
Multiracial	13 (21.3)	2 (9.5)		3 (14.2)	2 (12.5)
White	33 (54.1)	11 (52.4)		17 (81.0)	11 (68.8)
Missing/Not Disclosed	0 (0.0)	0 (0.0)		0 (0.0)	2 (12.5)
Ethnicity [n, (%)]
Hispanic or Latin-X	6 (10)	0 (0.0)		0 (0.0)	1 (6.2)	0.432
Non-Hispanic or Latin-X	53 (86.9)	19 (90.5)		21 (100.0)	15 (93.8)
Missing/Not Disclosed	2 (3.3)	2 (9.5)		0 (0.0)	0 (0.0)
Eczema History [n, (%)]	50 (82.0)	18 (85.7)	0.954	16 (76.2)	12 (75.0)	0.572
Asthma History [n, (%)]	32 (52.5)	8 (38.1)	0.377	5 (23.8)	2 (12.5)	0.223
Allergic Rhinitis History [n, (%)]	44 (72.1)	16 (76.2)	0.939	13 (61.9)	4 (25.0)	0.116
Eosinophilic GI Disease [n, (%)]	6 (9.8)	0 (0.0)	0.314	Data not available		‐‐

Among the peanut allergy cohort subjects, there were no significant differences in demographic or baseline clinical characteristics between those with severe (n=21) versus mild reactions (n=16). Overall, subjects consumed a median of 700 mg (IQR 340 mg, 1533 mg) cumulative peanut protein prior to reaction without a significant difference between severe and mild reactors (p=0.141). The median total Bock/PRACTALL OFC score (maximum score = 30) for subjects with mild reactions and severe reactions was 2 and 6, respectively (p=0.001). The median SGSAR score (possible range 0–5) for subjects with mild and severe reactions was 1 and 3, respectively (p=0.001). Of the patients with severe reactions, only three (14%) had lower respiratory symptoms and five (24%) had mild cardiovascular symptom of tachycardia. Most mild reactions consisted of limited cutaneous symptoms of urticaria, angioedema, rash, or erythema. Additional details about symptoms and treatment during OFC are shown in Table E1.

Tryptase Genotyping

Among the 82 subjects included in the NIAID cohort, three subjects had HαT (3.7%). In the peanut allergy cohort, 2/37 (5%) participants had HαT. The majority of subjects in both groups had at least one α-tryptase encoding gene copy. Of the total 119 subjects, 26% (n=31) were genotyped as ββ:ββ. This frequency was approximately 20% lower than the expected prevalence of this genotype in the general population (32%). The most common tryptase genotype in both cohorts was αβ:ββ, which is also the most common genotype in the general population.⁴⁰ Complete tryptase genotypes for the two cohorts are shown in Table 2.

Table 2.

Tryptase genotypes of observational food allergy cohort and peanut allergy cohort.

Genotype, n (%)	ααβ:αβ*	ααβ:ββ*	αβ:αβ	αβ:ββ	ββ:ββ	HαT
NIAID FA Cohort (n=82)	1 (1.2)	2 (2.4)	23 (28.0)	31 (35.4)	25 (30.5)	3 (3.7)
Peanut Allergy Cohort (n=37)	2 (5.4)	‐‐	12 (32.4)	17 (45.9)	6 (16.2)	2 (5.4)

^*Genotypes consistent with HαT (at least 2 copies of α-tryptase on the same allele)

Presence of α-tryptase is associated with severity of food allergy reactions

In the observational NIAID cohort, all three patients with HαT had a history of anaphylaxis to food. Additionally, individuals with α-tryptase were significantly more likely to have a history of anaphylaxis to food compared to subjects with only β-tryptase (p=0.026) (Figure 1A). Subjects with α-tryptase also had higher SGSAR scores that trended higher than subjects without α-tryptase (p=0.07) (Figure 1B).

Figure 1.

Increased α-tryptase is associated with increased prevalence and severity of anaphylaxis to food in subjects with known food allergy. A. Prevalence of anaphylaxis triggered by food reported by in patients with history of food allergy by relative α-tryptase encoding copy number in the NIAID observational cohort. Numbers above each bar represent the number of subjects (total n=82) in each group by tryptase genotype [(either 0, 1, 2, or >2 α-tryptase copies/HαT (≥1 α-tryptase replication)]. Analysis performed using Fisher’s exact test. B. Most severe reported anaphylactic reaction Severity Grading Scale for Acute Reactions (SGSAR) score by relative α-tryptase encoding copy number. Wilcoxon tests assessed for association between presence of any α-tryptase and SGSAR score; Spearman’s rank correlation was also performed between total number of α-tryptase isoforms and SGSAR to assess for possible gene dosage effects.

In the peanut allergy cohort, 2 of 21 (10%) of subjects with severe reactions had HαT, while none with mild reactions had HαT [0 of 16 (0%)] (p= 0.496). 1 of 6 (17%) subjects with only β-tryptase had a severe reaction during peanut OFC while 20 of 31 (65%) of subjects with α-tryptase had a severe reaction (p=0.066) (Table E2). Similarly, subjects with α-tryptase had higher total SGSAR scores during OFC as compared to subjects with no α-tryptase isoforms (ββ:ββ) (p=0.003) (Figure 2A). In addition, there were significant positive correlations between α-tryptase copy number and both higher total SGSAR and Bock/PRACTALL symptom scores (p=0.008 and 0.003, respectively) (Figure 2A&B).

Figure 2.

Increased α-tryptase isoform expression is associated with increased severity of reaction to peanut oral food challenge (OFC). A. Subjects with any α-tryptase isoforms were more likely to have higher Severity Grading Score for Acute Reactions (SGSAR) scores as compared to subjects with no α-tryptase isoforms. B. There was a statistically significant positive correlation between α-tryptase isoform copy numbers and total modified Bock/PRACTALL symptom score. Wilcoxon tests assessed for association between presence of any α-tryptase and OFC scores; Spearman’s rank correlation was also performed between total number of α-tryptase isoforms and SGSAR to assess for possible gene dosage effects.

Associations between TPSAB1 genotype, allergic co-morbidities, and allergy testing results

There were no significant differences in the prevalence of asthma, eczema, or allergic rhinitis between patients with or without HαT in either cohort. However, there was a trend for subjects with HαT to have increased rates of drug allergy and Hymenoptera venom allergy (Figure E1, A&B). Among those subjects in the NIAID observational cohort with HαT, a trend towards an increased total number of allergic reactions and reactions during OFC was also observed (Figure E1, C&D). There were no significant differences in the prevalence of multiple food allergies, prevalence of peanut allergy, history of auto-injectable epinephrine usage, history of eosinophilic GI disease (EGID), or history of urticaria (Figure E2, A–E).

Most patients in the peanut allergy cohort had at least one result from peanut SPT, whole peanut sIgE, or PCT sIgE prior to challenge. When results of testing were categorized by α-tryptase isoform number, there were no significant differences in results between the groups (Table E3).

Discussion

Our study demonstrates that the presence of germline α-tryptase encoding sequences is associated with more severe allergic reactions to foods. In both a retrospective observational cohort and a cohort of children who reacted during standardized open peanut OFC, the presence of α-tryptase was associated with an increased prevalence of anaphylaxis and severe reactions to food as compared to subjects without α-tryptase. This is the first report investigating TPSAB1 and tryptase genotype as a potential biomarker for food allergy reaction severity.

We initially explored the association of tryptase genotype with reaction severity in an observational cohort of food allergic patients. All individuals in this cohort with HαT had a history of anaphylaxis triggered by foods, and the presence of α-tryptase was associated with a trend of increasingly severe maximum single-organ severity scores. To test the strength of these associations, we then recruited a homogenous group of pediatric patients who had documented reactions during peanut OFC in order to determine whether tryptase genotype could predict reaction severity. Whereas the total number of food allergic subjects with HαT was lower than expected in the general population, the prevalence of any α-tryptase containing allele was higher and the prevalence of HαT among subjects with severe reactions to peanut OFC was approximately twice that of the general population^24,41 (10% versus 5.7%). These findings are remarkably similar to previous data of adult patients with Hymenoptera venom allergy²⁶ where HαT was not increased among all venom allergic patients when compared to the general population but was associated with a two-fold increased risk for severe anaphylaxis. Among patients with systemic mastocytosis (SM), HαT has also been associated with an increased prevalence of anaphylaxis^26,28,29. Interestingly, in the largest of these studies, individuals with both SM and HαT had a rate of food anaphylaxis that was approximately three-fold (5.4% vs. 1.9%) that of those with SM alone²⁹. Collectively, these results suggest that HαT, and furthermore, α-tryptase may be associated with increased anaphylaxis severity generally.

Previous studies demonstrate mechanisms by which α/β-tryptase heterotetramers may contribute to the pathogenesis of increased allergic reaction severity in patients with HαT³⁰. The formation of heterotetramers depends on the relative copy number of α- and β -tryptase encoding genes. Our study demonstrates that patients with only β-tryptase are less likely to have severe food allergy reactions, which is ostensibly due to the lack of α/β-tryptase heterotetramers. A previous study demonstrated that increased active β-tryptase copy number was inversely correlated with T_H2 asthma and response to omalizumab⁴² suggesting that lacking α-tryptase - which is relatively common⁴³ - may somehow limit allergic disease. However, data to support this hypothesis are currently lacking, and in this study the absence of α-tryptase did not preclude the development of anaphylaxis. Additional study of the potential mechanisms of tryptase gene composition on food allergy severity is needed.

In exploratory analyses, there were no significant differences in other allergic co-morbidities between patients with and without HαT. However, while tryptase genotype does not appear to impact the development of allergic disease, the presence of drug allergy and Hymenoptera venom allergy seemed to be more common in patients with HαT as seen in other studies^26,28,29. There was also a trend for subjects with α-tryptase or HαT to have higher total numbers of reactions to foods and to be less likely to tolerate peanut OFC. These trends were observed even though there were no significant differences in food allergy testing results (SPT, sIgE, and Ara h2) between those with or without HαT or α-tryptase isoforms. These findings suggest that tryptase genotype may be independent predictors of food allergy severity and could be useful as a clinical biomarker.

Our analysis of the NIAID cohort was limited by retrospective recall with reliance of reported symptoms by patients and their families. However, the diagnosis of anaphylaxis was adjudicated by a trained team of allergy physicians with expertise in food allergy and associated reactions, and these data demonstrated an initial signal that the presence of HαT - and α-tryptase generally - is associated with increased prevalence of anaphylaxis and severe allergic reactions to foods. The peanut allergy cohort controlled for recall and reporting biases as potential confounders of severity. Additionally, documentation of objective symptoms during standardized open OFC allowed for increased granularity to determine true reaction severity. However, we do acknowledge that very few subjects who underwent open peanut OFC had severe reactions with involvement of lower respiratory or cardiovascular system. This was expected though, as subjects with history of severe reactions or a high pretest probability of reaction were likely not offered an OFC. Additionally, there were only five total subjects in both cohorts with HαT which limits the generalizability and estimates of true effect size. Despite these limitations, the results were similar in both cohorts and suggest that HαT, as well as the presence of α-tryptase generally, may be enriched in patients with severe allergic reactions to food.

Recently, interest in genetic biomarkers for allergic diseases has increased⁴⁴. Tryptase genotyping has been useful as a biomarker for anaphylaxis severity in patients with mast cell diseases⁴¹ and systemic mastocytosis^28,29. The use of tryptase genotype instead of serum tryptase level as a biomarker ameliorates issues of obtaining serum tryptase levels before, during, and after acute allergic reactions, as well as difficulties with interpreting changes in serum tryptase that may differ by type of allergen trigger (i.e., lesser acute increases in serum tryptase in food allergy reactions as compared to Hymenoptera venom or drug allergies)^19,20. The use of tryptase genotype as a biomarker in clinical practice would be easy to implement given that sequencing of the TPSAB1 gene can be performed easily through non-invasive means, and testing is currently offered by a commercial laborator⁴⁵. Additionally, as new studies and research investigate the role of monoclonal antibodies (mAbs) against tryptase⁴², the use of tryptase genetics to risk stratify and identify potential patients who may benefit from this therapy will be helpful.

Our study suggests that there is a gene-dosage effect of relative α- to β-tryptase gene composition on the severity of food allergy reactions and that tryptase genotyping and determination of α-tryptase-encoding copy number may be useful as a genetic biomarker in food allergy. Additional data incorporating tryptase genotyping in larger studies including evaluation of its predictive value in test and validation cohorts are needed to determine the efficacy of this approach in risk stratification of patients with food allergy.

Supplementary Material

1

Figure E1. Prevalence of allergic disorders and history of allergic reactions in subjects with HαT in NIAID food allergy observational cohort. While not statistically significant, there was a trend for subjects with increased α-tryptase copies to have increased prevalence of drug allergy (A) and Hymenoptera venom allergy (B). Similarly, patients with increases in α-tryptase copy number were more likely to have higher number of allergic reactions in the past (C) and were less likely to pass OFC (D).

2

Figure E2. Prevalence of multiple food allergies (A), peanut allergy (B), history of auto-injectable epinephrine usage (C), eosinophil-associated gastrointestinal disorders (EGID (D), and history of urticaria unrelated to food allergy (E) were not significantly different between subjects in the NIAID cohort with differing α-tryptase isoform number.

*Urticaria unrelated to food exposure

Clinical Implications.

The presence of germline α-tryptase isoforms is associated with more severe allergic reactions to food. Tryptase genotyping may be useful as a genetic biomarker in predicting severity of food allergy reactions.

Abbreviations:

BST

Baseline serum tryptase

BAT

basophil activation testing

sIgE

specific IgE

SPT

skin prick testing

CRD

component resolved diagnostics

HαT

hereditary α-tryptasemia

SM

systemic mastocytosis

OFC

oral food challenge

NIAID

National Institute of Allergy and Infectious Diseases

PCT

peanut component testing

SGSAR

Severity Grading Scale for Acute Reactions

ddPCR

digital droplet PCR