Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2025 Feb 27;14(5):344.
doi: 10.3390/cells14050344.

Predicting Tolerance to Cow's Milk Allergy in Children Using IgE and IgG4 Peptide Binding Profiles

Carlos Fernández-Lozano  1   2 , Sergio Olmos-Piñero  2 , Laura Sánchez-Ruano  1   2 , Soledad Terrados  2 , Mª Del Carmen Diéguez  2 , Montserrat Fernández-Rivas  3   4   5 , Cristina Vlaicu  2 , Inmaculada Cerecedo  6 , Alejandro Gonzalo-Fernandez  3   7 , Belén de la Hoz  2   5 , Javier Martínez-Botas  1   2
Affiliations

Predicting Tolerance to Cow's Milk Allergy in Children Using IgE and IgG4 Peptide Binding Profiles

Carlos Fernández-Lozano et al. Cells. .

Abstract

Cow's milk allergy (CMA) is the most common food allergy in infants. This study aimed to identify peptide biomarkers predictive of tolerance in a Spanish population of children with CMA. We investigated specific IgE and IgG4 binding to sequential epitopes of the five major CM allergens (α-s1-, α-s2-, β-, and κ-caseins as well as β-lactoglobulin) using a microarray-based immunoassay. Microarray analysis was performed in 118 patients at baseline and after 6, 18, 30, 42, and 54 months. Most patients tolerated CM at 6 months (40.7%) and 18 months (35.4%). We found significant differences in IgE and IgG4 binding intensity and diversity between allergic and tolerant patients. No differences were observed at baseline. Combining baseline IgE and IgG4 serology variables and peptide microarray analysis results, a predictive model was developed using the XGBoost algorithm to classify tolerance status at different time points. The generated models showed high predictive value at 6 and 30 months with AUCs of 0.883 and 0.833, respectively. Therefore, using IgE and IgG4 antibody-binding peptides at baseline, we generated two models predicting tolerance in children with cow's milk allergy at 6 and 30 months.

Keywords: IgE; IgG4; biomarker; cow’s milk allergy; follow up; microarray; natural history; peptide; tolerance.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflicts of interest.

Figures

Figure 1
Figure 1
Study design and analysis strategy. Strategy 1 for the search for informative epitopes: the informative peptides were defined as those recognized by more than 75% of allergic patients at each time point. Strategy 2 for a predictive model of tolerance: the XGBoost algorithm was combined with decision trees to develop the predictive model of tolerance at visit 2 (T6), visit 3 (T18), and visit 4 (T30). Each model combined the serological variables and epitope binding at baseline (visit 1-T0) with the tolerance status (allergic or tolerant) at each time point. Patients were defined as tolerant if tolerated at or before the specified time (accumulative tolerance). The analyses of visits 5 (T42) and 6 (T42) were not performed due to the reduced number of remaining allergic patients. Lost patients voluntarily withdrew from the study.
Figure 2
Figure 2
(A) IgE recognition pattern of peptides of αs1-cas, αs2-cas, β-cas, κ-cas, and β-lac proteins after grouping by hierarchical clustering at visits 1 (0 months), 2 (6 months), 3 (18 months), 4 (30 months), 5 (42 months), and 6 (54 months). The intensity of IgE binding is represented by weighted average Z-scores on a grading scale according to the scale bar at the top of the figure. (B) Percentage of IgE recognition of peptides of αs1-cas, αs2-cas, β-cas, κ-cas, and β-lac proteins. The y-axis shows the percentage of patients within each visit showing positive binding (Z-score < 3) to each peptide. The x-axis shows the number of peptides for each protein. The TileMap on top of each protein represents the informative epitopes that were recognized by more than 75% of allergic patients within each visit according to the color code of each visit.
Figure 3
Figure 3
(A) IgG4 recognition pattern of peptides of αs1-cas, αs2-cas, β-cas, κ-cas, and β-lac proteins after grouping by hierarchical clustering at visits 1 (0 months), 2 (6 months), 3 (18 months), 4 (30 months), 5 (42 months), and 6 (54 months). The intensity of IgG4 binding is represented by weighted average Z-scores on a grading scale according to the scale bar at the top of the figure. (B) Percentage of IgG4 recognition of peptides of αs1-cas, αs2-cas, β-cas, κ-cas, and β-lac proteins. The y-axis shows the percentage of patients within each visit showing positive binding (Z-score < 3) to each peptide. The x-axis shows the number of peptides for each protein. The TileMap on top of each protein represents the informative epitopes that were recognized by more than 75% of allergic patients within each visit according to the color code of each visit.
Figure 4
Figure 4
Metrics of the model to predict tolerance at visit 2 (T6). (A) ROC curve of the training and the testing group. (B) Radar plot of the metrics of the training and the testing group. (C) Contribution of the top 10 variables to the predicting model.
Figure 5
Figure 5
Metrics of the model to predict tolerance at visit 3 (T18). (A) ROC curve of the training and the testing group. (B) Radar plot of the metrics of the training and the testing group. (C) Contribution of the top 10 variables to the predicting model.
Figure 6
Figure 6
Metrics of the model to predict tolerance at visit 4 (T30). (A) ROC curve of the training and the testing group. (B) Radar plot of the metrics of the training and the testing group. (C) Contribution of the top 10 variables to the predicting model.
See this image and copyright information in PMC

References

    1. Switkowski K.M., Oken E., Rifas-Shiman S.L., Camargo C.A., Jr., Gold D.R., Sordillo J.E., Lightdale J.R. Timing of Cow’s Milk Protein Introduction and Childhood Adverse Reactions to Cow’s Milk. J. Allergy Clin. Immunol. 2022;10:2713–2721.e2. doi: 10.1016/j.jaip.2022.06.022. - DOI - PMC - PubMed
    1. Fiocchi A., Schunemann H.J., Brozek J., Restani P., Beyer K., Troncone R., Martelli A., Terracciano L., Bahna S.L., Rance F., et al. Diagnosis and Rationale for Action Against Cow’s Milk Allergy (DRACMA): A summary report. J. Allergy Clin. Immunol. 2010;126:1119–1128.e12. doi: 10.1016/j.jaci.2010.10.011. - DOI - PubMed
    1. Wal J.M., Bernard H., Creminon C., Hamberger C., David B., Peltre G. Cow’s milk allergy: The humoral immune response to eight purified allergens. Adv. Exp. Med. Biol. 1995;371B:879–881. - PubMed
    1. Cerecedo I., Zamora J., Shreffler W.G., Lin J., Bardina L., Dieguez M.C., Wang J., Muriel A., de la Hoz B., Sampson H.A. Mapping of the IgE and IgG4 sequential epitopes of milk allergens with a peptide microarray-based immunoassay. J. Allergy Clin. Immunol. 2008;122:589–594. doi: 10.1016/j.jaci.2008.06.040. - DOI - PubMed
    1. Lin J., Bruni F.M., Fu Z., Maloney J., Bardina L., Boner A.L., Gimenez G., Sampson H.A. A bioinformatics approach to identify patients with symptomatic peanut allergy using peptide microarray immunoassay. J. Allergy Clin. Immunol. 2012;129:1321–1328.e5. doi: 10.1016/j.jaci.2012.02.012. - DOI - PMC - PubMed

Publication types

LinkOut - more resources