Sustained Successful Peanut Oral Immunotherapy Associated with Low Basophil Activation and Peanut-Specific IgE

Abstract

Background

Oral immunotherapy (OIT) can successfully desensitize many peanut allergic subjects, but clinical tolerance diminishes over time upon discontinuation, or low dose maintenance, of peanut. Therefore, in order to improve the efficacy and sustainability of such therapy, we sought to identify biomarkers and clinical tools that can predict therapeutic outcomes and monitor treatment responses.

Objective

We evaluated whether basophil activation in whole blood, and plasma levels of peanut-specific immunoglobulins, are useful biomarkers for peanut OIT.

Methods

We longitudinally measured, before, during and after OIT, basophil activation in whole blood ex vivo in response to peanut stimulation, and peanut-specific IgE and IgG4, in a large, single-site, double-blind, randomized, placebo-controlled, phase 2 peanut OIT study. We compared basophil responsiveness and peanut specific immunoglobulins between those who were clinically reactive vs. tolerant to peanut oral challenges.

Results

Peanut OIT significantly decreased basophil activation, peanut-specific, Ara h 1, Ara h 2 and Ara h 3 IgEs, and sIgE/total IgE, but increased sIgG4/sIgE. Participants who became reactive to 4 g of peanut 13 weeks off active OIT exhibited higher peanut-induced basophil activation ex vivo and higher peanut-specific IgEs and sIgE/total IgE, but lower sIgG4/sIgE. Notably, participants entering the study with low basophil responsiveness were more likely to achieve treatment success. Substantial suppression of basophil activation was required to maintain long-term clinical tolerance after peanut OIT.

Conclusion

Assessments of peanut-specific basophil activation and peanut-specific immunoglobulins can help to predict treatment outcomes, and to differentiate transient desensitization vs. sustained unresponsiveness after OIT.

Capsule summary

Assessment of basophil activation in whole blood and specific immunoglobulins in plasma before, during and after peanut OIT can help to differentiate transient desensitization vs. sustained unresponsiveness.

INTRODUCTION

There has been an increasing prevalence of food allergy, in both children and adults.^{1, 2}. However, there is currently no cure, and standard care, e.g., dietary avoidance and symptomatic treatments, are suboptimal. Peanut allergy is particularly problematic. Only ~20% of peanut allergic subjects are expected to outgrow this allergy³. Moreover, it is associated with high rates of accidental exposure, which carries a dangerous risk of systematic anaphylaxis⁴. Finally, diagnosing peanut allergy without oral peanut challenge is clinically challenging⁵.

Experimental therapies using oral immunotherapy (OIT), sublingual immunotherapy (SLIT), or epicutaneous immunotherapy (EPIT) have evaluated the safety, efficacy, and immunological changes associated with peanut immunotherapy⁶. Such studies have demonstrated that immunotherapy can successfully desensitize patients to peanut^7–16. However, it isn’t clear whether long-term tolerance is sustainable without continuous peanut dosing after active therapy.

We recently completed a single-site, double-blind, randomized, placebo-controlled, phase 2 study (POISED) to evaluate the long-term effects of peanut OIT¹⁷. POISED participants underwent peanut OIT build-up and maintenance followed by either peanut withdrawal or low dose (300 mg) peanut daily dosing. More than 80% of participants tolerated up to 4 g of peanut protein after 2 years of active OIT therapy, but the tolerance declined substantially with or without low dose daily peanut intake¹⁷. Our findings in POISED¹⁷, and others’ studies^{12, 18–20}, indicate that maintenance dosing and treatment regimens need to be improved further to achieve long-term protection after peanut immunotherapy.

Biomarkers for improving accuracy in diagnosis, and/or differentiating short-term desensitization vs. long-term unresponsiveness or other outcomes, would be useful in improving treatment protocols that could ultimately be based on an individual’s response to therapy. While several immunological parameters are being evaluated for such purposes, the basophil activation test (BAT) has emerged as one of the most promising tools for food allergy testing^{21, 22}. However, routine applications of BATs in clinical and research settings is not yet feasible due to the skills required to handle blood basophils, the lack of standardization of BAT protocols, and validation of BAT results with clinical outcomes.

To facilitate the utility of BATs, we recently developed a simple BAT protocol that produces consistent and reproducible results using whole blood samples collected in heparin, stored at 4°C, and processed within 24 hou rs after blood collection²³. Using this protocol, basophil activation can be quantified in small volumes of whole blood by flow cytometric measurements of activation markers, such as CD63 and CD203c²³. In the present study, we performed a comprehensive longitudinal analysis of basophil activation in the whole blood of participants enrolled in the POISED phase 2 peanut OIT study, and examined the association between basophil responsiveness and treatment outcomes after 2 years of peanut OIT followed by 13 weeks of peanut withdrawal (Peanut 0) or 300 mg low dose peanut maintenance (Peanut 300). We reported recently that lower peanut IgE and higher sIgG4/sIgE at baseline are associated with treatment success in our POISED study¹⁷. In this study, we examined whether OIT-induced changes in peanut specific IgEs and IgG4 are associated with basophil activation and are predictive of sustained outcomes.

METHODS

Study design

Whole blood for BATs was collected from peanut allergic participants (adults and children) enrolled in the POISED trial (ClinicalTrials.gov ) at the Sean N. Parker Center for Allergy and Asthma Research at Stanford University. The clinical research protocol was approved by the Division of Allergy, Immunology, and Transplantation (DAIT)/National Institute of Allergy and Infectious Diseases (NIAID) Allergy and Asthma Data Safety Management Board, the DAIT/NIAID Clinical Review Committee, the Stanford Institutional Review Board, and the Food and Drug Administration (FDA). For complete study design, baseline characteristics of the participants, and treatment outcomes, see Chinthrajah et al¹⁷. Briefly, 120 participants who reacted to a cumulative ingested dose of < 500 mg peanut by DBPCFCs at enrollment were randomized into three arms: Peanut 0 (60 participants), Peanut 300 (35 participants) and Placebo (25 participants). Peanut 0 and Peanut 300 participants received OIT build up to, and maintenance of, 4 g of peanut protein until week 104, when participants in the Peanut 0 arm discontinued peanut and those in the Peanut 300 arm maintained daily dosing of 300 mg peanut protein. Placebo received oat flour throughout the study. DBPCFCs were performed at baseline (week 0), week 104, week 117, and every 13 weeks for those who passed the previous DBPCFCs until the last DBPCFC at week 156. Intention-to- treat (ITT) populations included all 120 participants randomized at enrollment. Treatment “success” is defined as passing DBPCFCs to 4 g of peanut protein. Participants exiting the study before week 104 were considered “dropouts”. There were 9 dropouts in Peanut 0, 5 dropouts in Peanut 300, and 3 dropouts in Placebo, including 2 Placebo participants who exited the clinical study before week 104 and another Placebo participant whose basophil data were not available at weeks 104 and 117. All dropouts were excluded from longitudinal analyses. One participant was randomized to Peanut 300, failed DBPCFC at week 104, and was not on 300 mg peanut from weeks 104–117, nor received food challenge at week 117. Data from this participant were included up to week 104.

Blood collection

Blood was collected from each participant before, during, and after OIT at weeks 0, 12, 52, 104 and 117.

Basophil activation test

BATs were performed using whole blood preserved in heparin as described in Mukai et al²³. Briefly, basophil activation was assessed after stimulation for 30 min at 37°C with IL-3 (PeproTech, 2 ng/mL), polyclonal rabbit anti-human IgE (Bethyl Laboratoies, 1 μg/mL), or peanut protein extracted from the same peanut flour used in OIT at 0, 0.1, 1, 10, 100, 1000 (ng/mL). All stimuli were prepared in RPMI. Basophils were gated as CD123+HLA−DR− cells and %CD63^high basophils and CD203c expression levels were quantified by flow cytometry. To assess changes in basophil reactivity and sensitivity in responses to OIT, up-regulation of the basophil activation marker CD63 induced by peanut ex vivo stimulation was calculated as area under curve (AUC) of %CD63^high basophils as a function of the peanut dose²⁴. %CD63^high PE AUC was calculated for each participant at each time point. Expression of CD203c was measured by MFI. ΔCD203c is defined as CD203c MFI after anti-IgE, IL-3, or peanut stimulation minus CD203c MFI incubated with RPMI.

Immunoglobulin measurements

Total IgE, peanut-specific IgE and IgG₄ levels and component IgEs (Ara h 1, 2, 3) were measured at baseline, week 104 and week 117 using standardized methods in CLIA-approved laboratories (Johns Hopkins University for peanut-specific IgE, peanut-specific IgG_4, and component IgE testing, and Stanford Clinical laboratories for total IgE measurements). Both laboratories (at Stanford and Johns Hopkins) measured immunoglobulin levels using ImmunoCAP on the Phadia 250 Immunoassay Analyzers (Thermo Fisher Scientific).

Statistical analysis

Statistical comparisons were performed using GraphPad Prism 8 or R version 3.6.0. Mixed-effects analysis with the Geisser-Greenhouse correction was used for the longitudinal analysis of basophil CD63 and CD203c. Mann-Whitney tests were performed for comparisons between different treatment arms and between success and failure outcomes. Wilcoxon matched-pairs signed rank test was used to assess changes in immunoglobulin levels in the same group. Spearman correlations were used to assess associations between basophil activation and immunoglobulins levels. Fisher’s exact test was used to test for differences in treatment outcomes between basophil responders.

RESULTS

Peanut OIT suppresses basophil activation and reduces peanut specific IgE

Before OIT, there were no differences in the induction of %CD63^high basophils by ex vivo stimulation with peanut or anti-IgE between Peanut 0, Peanut 300 and Placebo (Fig E1. A, B). Levels of CD203c expression in basophils after anti-IgE, IL-3 or peanut stimulation were also similar between the 3 treatment arms at baseline (Fig E1. C). Consistent with previous OIT studies of peanut allergy^{7, 18, 25–27}, our OIT protocol significantly decreased peanut-induced %CD63^high basophils as early as week 12 of the OIT build up phase (Fig 1, A). Up-regulation of CD63 in basophils was suppressed during the OIT maintenance phase from weeks 52–104. Basophils of Peanut 0 and Peanut 300 participants continued to exhibit reduced levels of peanut induced activation until week 117 (Fig 1, A).

FIG 1. Peanut OIT reduces basophil responsiveness to peanut and anti-IgE over time.

A. Peanut-induced %CD63^high basophils represented by area under the curve (AUC) as a function of the peanut extract (PE) protein dose (0–1000 ng/mL) in Peanut 0, Peanut 300, and Placebo arms at weeks 0, 12, 52, 104 and 117. B. Anti-IgE (1 μg/mL) induced %CD63^high basophils in Peanut 0, Peanut 300 and Placebo arms at weeks 0, 12, 52, 104 and 117. Data from dropout participants (i.e., those who exited the study before week 104) were excluded. Whiskers represent min to max, boxes extend from the 25^th to 75^th percentiles. The lines in the middle of the boxes are medians. Individual values are shown as circles. P values were determined by mixed-effects analysis with the Geisser-Greenhouse correction for longitudinal changes, or by Wilcoxon matched-pairs signed rank test for comparisons between adjacent columns. Source of Wk 0 data in A were from Fig 4 of Chinthrajah et al (Ref. 17) with permission from the Publisher (ELSEVIER).

Peanut OIT also significantly reduced %CD63^high basophils in response to anti-IgE (1 μg/mL) stimulation in Peanut 0 and Peanut 300 participants as early as week 12 (Fig 1, B). Interestingly, we detected a small but significant increase in peanut-, but not anti-IgE-, induced %CD63^high basophils in Peanut 0 participants between week 104 and week 117. This increase is likely due to peanut avoidance by the Peanut 0 participants during this period. On the other hand, there was no significant difference in %CD63^high PE AUC between weeks 104 and 117 in Peanut 300, during which these participants continued taking 300 mg peanut after active OIT (Fig 1).

Similarly, peanut OIT significantly inhibited basophil CD203c expression induced by anti-IgE, IL-3 and peanut stimulation over time (Fig E2, A, B). No significant change in CD63 and substantially lesser changes in CD203c were observed in basophils of Placebo participants (Fig 1 and Fig E2, C). Basophil responses were strongly suppressed by peanut OIT at weeks 104 and 117 compared to Placebo, however, there was no difference in peanut-induced basophil activation between Peanut 0 and Peanut 300 participants (Fig E3) at either time, suggesting 300 mg peanut maintenance did not influence basophil responses at this timepoint.

Our peanut OIT also significantly decreased plasma levels of peanut, Ara h 1, Ara h 2 and Ara h 3 IgEs, and the ratio of peanut specific IgE (sIgE) vs. total IgE (Fig 2). By contrast, peanut specific IgG4 (sIgG4), and ratios of sIgG4 vs. sIgE were elevated by peanut OIT (Fig 2). There were no changes in sIgE, Ara h 1, Ara h 2, or Ara h 3 IgE, sIgE/total IgE or sIgG4/sIgE between week 0 and 117 in Placebo participants (Table E1). However, we detected some small but significant reductions in IgEs specific for peanut, Ara h 1 and Ara h 2, and sIgE/total IgE, in Placebo participants at week 104. Importantly, IgEs specific for peanut, Ara h 1, Ara h 2 and Ara h 3, and sIgG4, sIgE/total IgE, and sIgG4/sIgE, between Peanut 0 and Peanut 300 participants were not different at week 117, indicating that 300 mg of peanut dosing for an additional 13 weeks after active OIT did not influence plasma levels of peanut specific immunoglobulins.

FIG 2. Changes in peanut-specific immunoglobulins associated with peanut OIT.

Plasma immunoglobulins were quantified at weeks 0, 104 and 117. A. Peanut 0. B. Peanut 300. Values <0.1 (equivalent to zero) in Ara h 1, h 2 or h 3 IgEs are displayed as 0.01. P values were determined by the Wilcoxon matched-pairs signed rank test. * P<0.05, **P<0.01, ***P<0.001. N.S. P>0.05. Horizontal bar = median. Data from dropout participants (i.e. those who exited the study before week 104) were excluded.

Correlations of basophil activation and peanut specific immunoglobulins

Induction of allergen-specific IgG4 is a common outcome of immunotherapy and thought to be the primary mechanism that inhibits IgE-mediated effector cell activation^{28, 29}. We recently reported lower peanut-induced basophil activation and higher peanut sIgG4/sIgE before treatment is associated with OIT-induced sustained effects (i.e. passing DBPCFCs to 4 g peanut) after 13 weeks of peanut avoidance¹⁷. Furthermore, lower Ara h 2 IgE and peanut specific IgE are significantly associated with treatment success at primary end point at week 117 in both Peanut 0 and Peanut 300 participants, and baseline sIgE/total IgE is associated with a higher % of dose-related adverse events¹⁷.

We examined whether basophil activation was correlated with the plasma levels of sIgE or sIgG4 before and after active OIT treatment. At baseline before OIT, there were weak, but significant correlations between peanut-induced basophil activation (%CD63^high PE AUC) vs. IgEs specific for peanut, Ara h 1, Ara h 2 or Ara h 3, or sIgE/total IgE, in 120 intention-to-treat (ITT) participants (Fig 3, A). While there was no correlation between basophil responsiveness and sIgG4, we detected a weak, but significant, negative correlation between %CD63^high PE AUC and sIgG4/sIgE at baseline (Fig 3, A).

FIG 3. Correlations between basophil responses (%CD63^high PE AUC) and plasma immunoglobulins before and after OIT.

A. Baseline at week 0. All participants in ITT were included in the analysis. B. 13 weeks after active OIT at week 117. Data from dropout participants (i.e., those who exited the study before week 104) were excluded. Values <0.1 in Ara h 1, h 2 and h 3 IgEs are displayed as 0.01. Spearman correlation coefficient (r) and P values are provided on each graph. In B, dots in red circles are data from 2 placebo participants and one basophil high responder in the Peanut 0 arm who failed week 117 DBPCFC.

We analyzed the primary end-point at week 117 (i.e., 13 weeks after active OIT treatment). We found significant weak to moderate correlations between peanut-induced basophil activation and IgE for peanut, Ara h 1, Ara h 2 or Ara h 3, and sIgE/total IgE or sIgG4/sIgE, but no correlation between sIgG4 and peanut-induced basophil activation (Fig 3, B). These data suggest that peanut-induced basophil activation is regulated not just by specific IgG4 alone, but by the interactions of basophils with peanut specific IgEs, peanut specific IgG4, and total IgE.

Peanut-induced basophil activation and specific IgEs differentiate transient desensitization vs. sustained unresponsiveness

The main goals of our peanut OIT study were to: 1) assess sustained unresponsiveness to a cumulative dose of 4 g peanut protein after active peanut OIT treatment followed by 13 weeks of peanut avoidance, and 2) test the sustained effect of 300 mg daily dosing following active therapy. We compared longitudinally, between those passing vs. failing food challenges at week 117, peanut-induced %CD63^high basophils at weeks 0, 12, 52, 104 and 117.

Peanut 0 and Peanut 300 participants who failed 4 g peanut DBPCFCs at week 117 exhibited higher %CD63^high basophils upon peanut stimulation at every time (Fig 4, A, B), while there was no significant difference in %CD63^high basophils to anti-IgE stimulation (1 μg/mL) between those who passed vs. failed week 117 peanut challenges (Fig 4, C, D). Peanut 0 participants who failed week 117 oral food challenge also exhibited higher ΔCD203c in response to peanut stimulation, but not to anti-IgE or IL-3 (Fig E4, A, B). In Peanut 300, lower ΔCD203c in basophils of those who passed 4 g peanut DBPCFCs at week 117 were detected with high dose peanut stimulation at week 117 (Fig E4, C, D). In addition, participants who failed week 117 DBPCFCs exhibited significantly higher IgEs specific for peanut, Ara h 1, Ara h 2 and Ara h 3, and sIgE/total IgE, at week 117 (Fig 5). Peanut 0 participants who failed week 117 DBPCFCs had (unexpectedly) significantly higher sIgG4, but lower sIgG4/sIgE, at week 117 (Fig 5, A). sIgG4 and sIgG4/sIgE were not different between Peanut 300 participants who failed vs. passed week 117 DBPCFCs (Fig 5, B).

FIG 4. Peanut- and anti-IgE-induced %CD63^high basophils over time by treatment outcome at week 117.

A. Peanut-induced %CD63^high basophils are represented by area under the curve (AUC) as a function of the peanut extract (PE) protein dose (0–1000 ng/mL) in Success vs. Failure in Peanut 0. B. Peanut-induced %CD63^high basophils are represented by area under the curve (AUC) as a function of the peanut extract (PE) protein dose (0–1000 ng/mL) in Success vs. Failure in Peanut 300. C. Anti-IgE (1 μg/mL) induced %CD63^high basophils in in Success vs. Failure in Peanut 0. D. Anti-IgE (1 μg/mL) induced %CD63^high basophils in Success vs. Failure in Peanut 300. Whiskers represent min to max, boxes extend from the 25^th to 75^th percentiles. The lines in the middle of the boxes are medians. Individual values are shown as circles. P values were determined by Mann-Whitney test. “Success” is tolerant to DBPCFCs to 4 g peanut protein at week 117.

FIG 5. Plasma immunoglobulin levels by treatment outcome at week 117.

Plasma immunoglobulins were quantified at week 117. A. Peanut 0. B. Peanut 300. Horizon bar = median. Values <0.1 in Ara h 1, h 2 and h 3 IgEs are displayed as 0.01. P values were determined by Mann-Whitney test. “Success” is tolerant to DBPCFCs to 4 g peanut protein at week 117. Source of the data in this figure is from Figs S5 and S6 of Chinthrajah et al (Ref. 17) with permission from the Publisher (ELSEVIER).

We did not find any significant differences between Peanut 0 and Peanut 300 participants who failed or between Peanut 0 and Peanut 300 participants who passed week 117 DBPCFCs in %CD63^high PE AUC at weeks 0, 12, 52, 104 and 117, or in ΔCD203c and peanut specific IgEs and IgG4 at week 0 and week 117.

Basophil non/low responders to peanut exhibit lower peanut-specific IgE and better treatment outcome

Basophils of the 120 participants enrolled in this study expressed a wide range of peanut-induced CD63 upregulation before treatment. Based on the sensitivity and extent of their basophil responses to ex vivo peanut stimulation at enrollment, participants could be classified as basophil “non/low”, “intermediate”, or “high’ responders. We found that the induction of %CD63^high basophils (calculated by AUC as a function of the peanut dose response curves [0–1000 ng/mL]) ranged from 0 to 192.6 with a mean of 54.73 (Fig 6, A). We defined the “non/low responders” (LR) as those with peanut-induced %CD63^high dose response AUC<12.09 (= mean (54.73)-1x SD (42.64) of %CD63^high PE AUC of 120 participants) and the “high responders” (HR) as those with peanut induced %CD63^high dose response AUC>97.37 (= mean+1xSD of %CD63^high PE AUC of 120 participants). The “intermediate responders” (IR) had peanut-induced %CD63^high dose response AUC >12.09 and <97.37 (Fig 6, B).

FIG 6. Three subgroups of peanut allergy participants based on basophil responses to peanut before OIT.

A. Peanut induced %CD63^high basophils (%CD63^high PE AUC) in the 120 participants at baseline. Scatter dot plot with mean ± SD. B. Peanut-induced %CD63^high basophils of non/low responders (PE AUC<12.09), intermediate responders (PE AUC >12.09 and <97.37), and high responders (PE AUC>97.37). Individual values are shown as circles. C. Distribution of three subgroups of basophil responders in intention-to-treat (ITT) population, Peanut 0, Peanut 300 and Placebo. D. Anti-IgE (1 μg/mL) induced %CD63^high basophils in the three subgroups of basophil responders. Whiskers represent min to max, boxes extend from the 25^th to 75^th percentiles. The lines in the middle of the boxes are medians. Individual values are shown as circles. E. Cumulative tolerated peanut challenge dose (CTD) at week 0 among the three subgroups of basophil responders. Data are mean+ SEM. LR, non/low responder; IR, intermediate responder, HR, high responder. P values were determined by Mann-Whitney test.

There were 18 (15%) LRs, 83 (69%) IRs and 19 (16%) HRs in our ITT population (Fig 6, C). Moreover, there was a similar distribution of LRs, IRs and HRs among the three treatment arms (Figure 6, C). Participants were classified based on their basophil responses to peanut, but we found that the basophil LRs to peanut also exhibited lower anti-IgE induced CD63 up-regulation and vice versa for the HRs (Fig 6, D). On the other hand, we found that basophil LRs tolerated 2–3 times more peanut protein at enrollment than the IRs and HRs (LR vs. IR vs HR: [150.6 ± 35.51] mg vs. [66.51± 10.92] mg vs. [49.21± 15.96] mg) (Fig 6, E). Furthermore, before treatment, basophil LRs also had lower peanut-specific, Ara h 1, Ara h 2 and Ara h 3 IgE, and sIgE/total IgE, compared to IRs and/or HRs (Fig 7, A). By contrast, basophil HRs had higher levels of peanut-specific, Ara h 1, Ara h 2 and Ara h 3 IgE, and sIgE/total IgE. Interestingly, no difference was detected in levels of peanut specific IgG4 among LRs, IRs or HRs, but the ratio of sIgG4/sIgE was significantly higher in LRs compared to IRs or HRs (Fig 7, A).

FIG. 7. Plasma immunoglobulin levels in the three groups of basophil responders before and after OIT.

A. Before OIT at week 0. Data from ITT participants were included in the analysis. B. After OIT at week 117. Data from Peanut 0 and Peanut 300 participants were included; data from dropouts were excluded. Horizon bar = median. LR, non/low responder; IR, intermediate responder; HR, high responder. Values <0.1 in Ara h 1, h 2 and h 3 IgEs were displayed as 0.01. P values were determined by Mann-Whitney test.

We also compared plasma IgE and IgG4 levels between the 3 groups of basophil responders at week 117 after active OIT treatment. Basophil LRs exhibited the lowest levels of IgE specific for peanut, Ara h 1, Ara h 2 and Ara h 3 compared to IRs and/or HRs, whereas the levels of these IgEs were not different between IRs and HRs (Fig 7, B). Notably, the levels of peanut-specific IgG4 were also lower in the basophil LRs compared to either IRs or HRs after OIT, whereas peanut sIgG4/sIgE was not significantly different between 3 groups after OIT (Fig 7, B).

LRs not only exhibited lower basophil activation following peanut or anti-IgE stimul\ation, and lower peanut specific IgEs, they also responded to OIT with better outcomes. 91% and 55% of LRs passed oral food challenge up to 4 g of peanut protein at week 117 and week 156, respectively (Table 1). By contrast, only 17% of basophil high responders passed oral food challenges at both weeks 117 and 156. 48% and 22% of the IR recorded treatment successes at weeks 117 and 156, respectively.

Table 1.

Low basophil activation at week 0 is associated with sustained effectiveness of peanut OIT.

Basophil status (week 0)	Peanut 0+Peanut 300						Peanut 0						Peanut 300
	Wk 117 DBPCFC			Wk 156 DBPCFC*			Wk 117 DBPCFC			Wk 156 DBPCFC*			Wk 117 DBPCFC			Wk 156 DBPCFC*
	Pass	Fail	P value	Pass	Fail	P value	Pass	Fail	P value	Pass	Fail	P value	Pass	Fail	P value	Pass	Fail	P value
Non/low responder	10 (91%)	1 (9%)	0.001	6 (55%)	5 (45%)	0.08	5 (83%)	1 (27%)	0.02	2 (33%)	4 (67%)	0.48	5 (100%)	0 (0%)	0.13	4 (80%)	1 (20%)	0.23
Intermediate responder	28 (48%)	30 (52%)		13 (22%)	45 (78%)		15 (42%)	21 (58%)		5 (14%)	31 (86%)		13 (59%)	9 (41%)		8 (36%)	14 (64%)
High responder	2 (17%)	10 (83%)		2 (17%)	10 (83%)		1 (11%)	8 (89%)		1 (11%)	8 (89%)		1 (33%)	2 (67%)		1 (33%)	2 (67%)

P values were determined by Fisher’s exact test.

^*Those who failed at week 117 were not retested (for DBPCFC) at week 156. However, in the table, these participants are included as “failures”.

Notably, levels of peanut-induced %CD63^high basophils of IRs and HRs in the Peanut 0 arm who passed week 117 DBPCFCs were reduced to 20 ± 4% of their baseline levels by OIT, whereas those who failed still had 42 ± 7% of their baseline levels (P = 0.01). Similarly, peanut-induced %CD63^high basophils in IRs and HRs in the Peanut 300 arm who passed week 117 DBPCFCs were reduced to 11 ± 2% of their baseline levels, compared to 45 ± 18% in those who failed (P < 0.0001).

These data indicate that achieving sustained unresponsiveness 13 weeks off active OIT is observed primarily in two groups of subjects, those with lower basophil responses at study entrance and those that undergo a substantial reduction (80–90%) of their peanut-induced basophil activation after OIT.

Basophil responsiveness in LR participants who reached treatment success remained low after OIT (%CD63^high PE AUC=3.78 ± 1.21 & 5.67 ± 2.45 at week 0 & week 117, respectively. P = 0.30). Peanut OIT also did not alter specific IgE in these LRs (peanut specific IgE (kU/L) =25.70 ± 10.92 & 13.39 ± 5.26 at week 0 & week 117, respectively. P = 0.13), but increased specific IgG4 (ng/mL) levels from 0.82 ± 0.30 at week 0 to 8.93 ± 4.53 at week 117 (P = 0.03). By contrast, sustained tolerance in IR and HR subgroups was associated with lower %CD63^high PE AUC and lower IgEs specific for peanut, Ara h 1, Ara h 2 and Ara h 3, but not with higher sIgG4, before and/or after OIT treatment (Table E2).

In summary, in our POISED study, assessment of basophil activation and measurement of peanut specific IgEs are useful biomarkers in differentiating transient desensitization vs. sustained tolerance.

DISCUSSION

Since its first application in food allergy two decades ago³⁰, the BAT has been actively evaluated in this disorder^{21, 22}. Enhanced spontaneous and food-induced basophil activation was shown to be associated with clinical reactivity to milk^{31, 32}, peanut^{33, 34}, egg^{33, 35}, red meat³⁶ and other foods^{21, 22}. Some studies have suggested that BATs are more reliable than skin prick tests and specific IgE levels in differentiating food allergy from tolerance in children^{32, 34}. While suppression of basophil activation by immunotherapy has been demonstrated in food allergy to egg³⁷, milk³⁸, and peanut^{7, 8, 18, 25–27, 39, 40}, only a few studies have examined the association of basophil activation with treatment outcomes^{18, 26, 27, 37, 40, 41}. Our analysis of BAT results obtained for more than 2 years in the POISED study shows that basophil activation remains suppressed even after 13 weeks off active OIT, and that sustained protection is associated with lower levels of basophil activation by peanut before, during, and after OIT.

In addition to inhibiting peanut-induced basophil activation, our peanut OIT also decreased upregulation of CD63 and CD203c in responses to anti-IgE and/or IL-3. Antigen-nonspecific inhibition of basophil activation also was observed in other peanut OIT studies^{18, 25}. This “bystander effect” is likely due to reduction of FcεRI on basophils as a consequence of IgE reduction by OIT^{42, 43}.

The mechanisms by which OIT suppress basophil activation are still being investigated. We think that understanding the suppression of basophils by immunotherapy can help in efforts to devise better treatment strategies or preventive approaches for allergies to peanuts and other foods. Depletion of IgG, specifically IgG4, has been shown to eliminate the inhibitory effect of OIT plasma in basophil activation^{28, 40, 44}. We did not find direct associations between peanut induced %CD63^high basophils and levels of peanut specific IgG4 alone, however, we detected significant weak correlations between peanut-induced %CD63^high basophils with sIgG4/sIgE before and after OIT, as well as a moderate correlation between peanut-induced %CD63^high basophils and sIgE/total IgE after OIT. These findings suggest that basophil activation is regulated by the summation of activating (such as sIgE) and inhibitory (e.g., IgG1, IgG4) immunoglobulins. Indeed, our findings indicate not only that the levels of sensitization at study entrance, but also immune responses shaped by OIT, can have significant impacts on treatment outcome.

Furthermore, other immunoglobulin isotypes induced by OIT can also potentially inhibit basophil activation^{45, 46}. It is conceivable that peanut immunotherapy induces immunological modifications (such as reduction of peanut specific IgE and increase of peanut specific IgG, IgA, and IgD) which in turn alter basophil phenotype (such as reduction of FcεR, Syk⁴⁷ and other intracellular signaling machinery) and thus reduces basophil responsiveness to food allergens, and clinical reactivity. Therefore, we think that BATs ideally should be performed in whole blood, as we did in this study, to evaluate OIT-induced immune modifications external to basophils, as well as intrinsic changes in basophils.

We found that participants with or without 300 mg peanut daily dosing after OIT exhibited similar basophil responsiveness and peanut-specific immunoglobulins tested 13 weeks after active OIT. These findings are in line with clinical observations that Peanut 0 and Peanut 300 participants achieved similar treatment success rates 13 weeks after active OIT¹⁷. However, Peanut 300 participants were 3 times more likely to tolerate 4 g peanut DBPCFCs than Peanut 0 participants one year after active OIT. One limitation of our study is that we were not able to associate BAT results with clinical outcomes beyond 3 months after stopping active OIT, because most of the participants who failed week 117 DBPCFCs did not return to the study and their basophils were not tested further.

In summary, we have demonstrated that BATs and blood biomarkers (including peanut and peanut component-specific IgEs, sIgG4/sIgE, and sIgE/total IgE) measured before, during, and after immunotherapy, can differentiate transient desensitization vs. sustained tolerance after peanut OIT. Our analysis of basophil responsiveness to ex vivo peanut stimulation reveals phenotypic heterogeneity among basophils in peanut allergy, and identifies a peanut allergy subpopulation (i.e., basophil non/low responders) who are more amenable to OIT treatment. Our results support the hypothesis that BATs, in conjunction with other blood biomarkers, can identify allergic subjects who can most benefit from OIT. They also raise the possibility that prolonged administration of antigen, or antigen updosing, may improve the efficacy and durability of OIT in difficult-to-treat populations

Key messages.

• Peanut OIT suppresses basophil activation in responses to peanut and anti-IgE, reduces peanut-specific, Ara h 1, Ara h 2 and Ara h 3 IgEs, and increases peanut-specific IgG4.

• Measurements of basophil activation by peanut and peanut-specific IgEs can help to differentiate those who will achieve transient desensitization vs. sustained tolerance after peanut OIT.

• Basophil non/low responders exhibit higher treatment success to peanut OIT and, among other subjects, 80–90% suppression of peanut-specific basophil activation is required to maintain long-term clinical tolerance after peanut OIT.

Abbreviations

POISED

Peanut oral immunotherapy study: safety, efficacy and discovery

BAT

Basophil activation test

OIT

Oral immunotherapy

DBPCFC

Double-blind, placebo-controlled, oral food challenge

SU

Sustained unresponsiveness

ITT

Intention to treat

AUC

Area under the curve

LR

Basophil non/low responder

IR

Basophil intermediate responder

HR

Basophil high responder

MFI

Mean florescent intensity