Food-specific immunoglobulin A does not correlate with natural tolerance to peanut or egg allergens

Abstract

ImmunoglobulinA (IgA) is the predominant antibody isotype in the gut, where it regulates commensal flora and neutralizes toxins and pathogens. The function of food-specific IgA in the gut is unknown but is presumed to protect from food allergy. Specifically, it has been hypothesized that food-specific IgA binds ingested allergens and promotes tolerance by immune exclusion; however, the evidence to support this hypothesis is indirect and mixed. Although it is known that healthy adults have peanut-specific IgA in the gut, it is unclear whether children also have gut peanut-specific IgA. We found in a cohort of non-food-allergic infants (n = 112) that there is detectable stool peanut-specific IgA that is similar to adult quantities of gut peanut-specific IgA. To investigate whether this peanut-specific IgA is associated with peanut tolerance, we examined a separate cohort of atopic children (n = 441) and found that gut peanut-specific IgA does not predict protection from development of future peanut allergy in infants nor does it correlate with concurrent oral tolerance of peanut in older children. We observed higher plasma peanut-specific IgA in those with peanut allergy. Similarly, egg white-specific IgA was detectable in infant stools and did not predict egg tolerance or outgrowth of egg allergy. Bead-based epitope assay analysis of gut peanut-specific IgA revealed similar epitope specificity between children with peanut allergy and those without; however, gut peanut-specific IgA and plasma peanut-specific IgE had different epitope specificities. These findings call into question the presumed protective role of food-specific IgA in food allergy.

Fig. 1.. Stool peanut-specific IgA is detectable in non–food-allergic infants.

Samples from the GMAP study were used in this figure. (A) Peanut-specific IgA was measured in infant stool samples (n = 109) from 51 nonallergic individuals aged 0.1 to 19.1 months. Three samples were excluded because of poor dilution on ELISA. Dashed lines denote previously determined healthy adult at 10th and 90th percentiles for stool peanut-specific IgA. Best-fit line was determined with simple linear regression. (B) Stool peanut-specific IgA specificity was determined with a competitive ELISA using blocking with peanut, soy, ovalbumin, ovomucoid, or wheat protein (200 μg/ml; n = 8, age 0.3 to 12.6 months). Samples are color-coded by individual. (C) Stool peanut-specific IgA was measured in infants aged 0.1 to 6.5 months who were breastfeeding (n = 41) or not breastfeeding (n = 7). (D) Stool peanut-specific IgA was measured in infants with no eczema (n = 69) and eczema (n = 40). Dotted lines in (B) to (D) indicate the limit of detection. Data were shown with means and SEM. Means were compared using one-way repeated-measures ANOVA with Dunnett’s multiple comparisons test in (B) or with unpaired t tests in (C) and (D), and P values are labeled. *P < 0.05; ****P < 0.0001.

Fig. 2.. Atopic infants make peanut-specific IgA.

(A) Schematic of samples obtained from atopic infants from the CoFAR2 cohort. Baseline CoFAR2 samples were analyzed in this figure. (B) Stool total IgA in infants (n = 439) by age is shown. (C) Stool peanut-specific IgA in infants (n = 432) by age is shown. (D) Stool peanut-specific IgA was measured in infants who were breastfeeding (n = 86) and not breastfeeding (n = 53). (E) Stool peanut-specific IgA was measured in infants aged 9.2 to 15.6 months who were consuming (n = 17) and avoiding (n = 262) peanut. (F) Stool peanut-specific IgA was measured with a competitive ELISA using blocking with of peanut, soy, ovalbumin, ovomucoid, or wheat protein (200 μg/ml; n = 8, age 3.1 to 6.0 months). Samples are color-coded by individual. (G) Plasma peanut-specific IgA in infants (n = 330) by age is shown. Dotted lines indicate the limit of detection. Dashed lines show previously determined healthy adult at 10th and 90th percentile values. Best-fit lines were obtained with second-order polynomial or simple linear regression. Data were shown with means and SEM, with range displayed in (E). Means were compared using unpaired t tests in (E) or one-way repeated-measures ANOVA with Dunnett’s multiple comparisons test in (F). P values are labeled. ****P < 0.0001.

Fig. 3.. Peanut-specific IgA does not predict protection from future peanut sensitization or allergy in an unsensitized cohort.

CoFAR2 patients with a baseline plasma peanut-specific IgE of ≤0.1 kU_A/liter and at least 2 years of follow-up were analyzed in this figure. Baseline stool peanut-specific IgA concentrations are shown, stratified by follow-up (A) peanut-specific IgE (n = 55 ≤ 0.1 kU_A/ml and n = 39 > 0.1 kU_A/ml), (B) peanut SPT score (peanut wheal minus saline wheal) (n = 68 ≤ 3 mm and n = 25 > 3 mm), and (C) clinical peanut allergy (n = 92 not allergic, n = 17 peanut-allergic, and n = 5 equivocal peanut allergy samples excluded). Baseline plasma peanut-specific IgA concentrations are shown, stratified by (D) peanut-specific IgE (n = 42 ≤ 0.1 kU_A/ml and n = 33 > 0.1 kU_A/ml), (E) peanut SPT score (n = 54 ≤ 3 mm and n = 21 > 3 mm), and (F) clinical peanut allergy (n = 77 not allergic, n = 13 peanut-allergic, and n = 6 equivocal peanut allergy samples excluded). Data were shown with means and SEM. Means were compared using unpaired t tests, with P values as labeled.

Fig. 4.. Gut peanut-specific IgA does not correlate with concurrent peanut sensitization or allergy.

Follow-up samples from CoFAR2 were analyzed. Follow-up stool peanut-specific IgA concentrations are shown, stratified by concurrent (A) peanut-specific IgE (n = 18 ≤ 0.1 kU_A/ml and n = 102 > 0.1 kU_A/ml), (B) peanut SPT score (n = 31 ≤ 3 mm and n = 89 > 3 mm), and (C) clinical peanut allergy (n = 50 not allergic, n = 55 peanut-allergic, and n = 19 equivocal peanut allergy samples were excluded). Follow-up plasma peanut-specific IgA concentrations are shown, stratified by concurrent (D) peanut-specific IgE (n = 17 ≤ 0.1 kU_A/ml and n = 67 > 0.1 kU_A/ml), (E) peanut SPT score (n = 27 ≤ 3 mm and n = 57 > 3 mm), and (F) clinical peanut allergy (n = 40 not allergic, n = 27 peanut-allergic, and n = 17 equivocal peanut-allergic samples excluded). Correlations of follow-up plasma peanut-specific IgE and (G) follow-up stool peanut-specific IgA (n = 120 pairs) and (H) follow-up plasma peanut-specific IgA (n = 84 pairs) are shown. (I) Shown is a summary of univariate and multivariate regression of effect of plasma peanut-specific IgA ± plasma peanut-specific IgE on allergy outcome. Dotted lines indicate the limit of detection. Data were shown with means and SEM. Means were compared using unpaired t tests. Correlations were assessed using two-tailed Pearson tests, with Pearson’s r and 95% confidence intervals (CI) denoted. P values are as labeled. **P < 0.01.

Fig. 5.. Gut peanut-specific IgA targets different epitopes than plasma peanut-specific IgE.

Matched stool and plasma samples (n = 9 pairs) from the CoFAR2 follow-up time point were analyzed in this figure. Stool peanut-specific IgA and plasma peanut-specific IgE BBEA against (A) Ara h 1, (B) Ara h 2, and (C) Ara h 3 sequential epitopes are shown, expressed as heatmaps of scaled MFI.

Fig. 6.. Peanut-specific IgA is not related to tolerance to oral peanut challenge.

The subset of children in CoFAR2 who had an OFC to peanut was analyzed for (A) to (D). (A) Baseline stool peanut-specific IgA concentrations are shown, stratified by OFC outcome (n = 79 negative and n = 28 positive). (B) Follow-up stool peanut-specific IgA concentrations are shown, stratified by OFC outcome (n = 21 negative and n = 19 positive). (C) Baseline plasma peanut-specific IgA concentrations are shown, stratified by OFC outcome (n = 62 negative and n = 23 positive). (D) Follow-up plasma peanut-specific IgA concentrations are shown, stratified by OFC outcome (n = 17 negative and n = 10 positive). (E) Demographics of healthy individuals from the Stanford Twin Registry and matched peanut-allergic POISED trial patients are presented. (F) Stool peanut-specific IgA concentrations were measured in healthy (n = 9) and peanut-allergic (n = 9) individuals from the Twins Registry and POISED trial, respectively. Dashed lines denote healthy adult at 10th and 90th percentiles for stool peanut-specific IgA. Dotted lines indicate the limit of detection. Data were shown with means and SEM. Means were compared using unpaired t tests, with P values as labeled.

Fig. 7.. Peanut-specific IgA subtype and epitope specificity patterns do not correlate with protection from allergy.

Samples from the follow-up time point of CoFAR2 were used for (A) to (G); and GMAP, POISED, and Twins samples were used in (F) and (G). (A) IgA1 and (B) IgA2 optical density (OD) dilution curves are shown for matched plasma and stool samples (n = 14). (C) Plasma peanut-specific IgA1, (D) stool peanut, and (E) plasma peanut-specific IgA2 concentrations are shown, separated by concurrent allergy status [stool: n = 45 not allergic, n = 46 peanut-allergic, and n = 33 excluded samples (18 equivocal peanut allergy, one poor dilution, and 14 completely depleted by other assays); plasma: n = 40 not allergic, n = 27 peanut-allergic, and n = 17 samples excluded for equivocal peanut allergy]. Stool peanut-specific IgA BBEA results against (F) Ara h 1, (G) Ara h 2, and (H) Ara h 3 sequential epitopes are expressed as heatmaps of scaled MFI for stool samples from n = 20 peanut-allergic, n = 20 not allergic, and n = 7 peanut-specific IgA-deficient individuals. Dotted lines indicate the limit of detection. Data were shown with means and SEM. Means compared using unpaired t tests (C and E) or a Mann-Whitney test (D), with P values as labeled. **P < 0.01

Fig. 8.. Gut IgA to egg white is not related to tolerance and does not predict outgrowth of egg allergy.

CoFAR2 samples were used. Baseline stool egg white–specific IgA concentrations are shown in infants (A) by age in months (n = 438), (B) who were breastfeeding (n = 87) and not breastfeeding (n = 57), (C) who were eating (n = 24) and avoiding egg (n = 379), (D) who had outgrown (n = 163) and persistent egg allergy (n = 88), and (E) who were clinically not allergic (n = 60) or egg allergic (n = 274); n = 103 equivocal egg allergy samples excluded. (F) Follow-up egg white–specific stool IgA concentrations are shown for children who were not allergic (n = 79) and egg allergic (n = 31); n = 14 equivocal egg allergy samples excluded. (G) Baseline stool egg white–specific IgA concentrations are shown for infants with negative (n = 102) and positive (n = 16) egg OFC. (H) Follow-up stool egg white–specific IgA concentrations are shown for children with negative (n = 35) and positive (n = 10) egg OFC. Dotted lines indicate the limit of detection. Best-fit line in (A) was determined with simple linear regression. Data were shown with median and IQR and compared using Mann-Whitney tests, with P values as labeled. *P < 0.05.