Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Mar 6.
Published in final edited form as: J Allergy Clin Immunol. 2025 Mar 6;156(1):108–117. doi: 10.1016/j.jaci.2025.02.030

Network Analysis Reveals Protein Modules Associated with Childhood Respiratory Diseases

Nicole Prince 1,2, Sofina Begum 1,2, Kevin M Mendez 1,2,3, Lourdes G Ramirez 1,2,4, Yulu Chen 1,2, Qingwen Chen 1,2, Su H Chu 1,2, Priyadarshini Kachroo 1,2, Ofer Levy 2,5, Joann Diray-Arce 2,5, Paolo Palma 6,7, Augusto A Litonjua 8, Scott T Weiss 1,2, Rachel S Kelly 1,2,*, Jessica A Lasky-Su 1,2,*
PMCID: PMC12517469  NIHMSID: NIHMS2063644  PMID: 40057284

Abstract

Background:

The first year of life represents a dynamic immune development period that impacts the risk of developing respiratory-related diseases, including asthma, recurrent infections, and eczema. However, the role of immune-mediating proteins in childhood respiratory diseases is not well characterized in early life.

Methods:

We applied weighted gene correlation network analysis (WGCNA) to derive modules of highly correlated proteins during early life immune development using plasma samples collected from children at age 1 year (N=294) in the Vitamin D Antenatal Asthma Reduction Trial (VDAART). Using regression analysis, we evaluated relationships between protein modules at age 1 and respiratory-related diseases by age 6. We integrated protein modules with additional ‘omics and social, demographic, and environmental data for further characterization.

Results:

Our analysis identified four protein modules at age 1 year associated with incidence of childhood asthma and/or recurrent wheeze (Padj range: 0.02–0.03), respiratory infections (Padj range: 6.3×10–9–2.9×10–6), and eczema (Padj=0.01) by age 6 years; associations between modules and clinical outcomes were temporally sensitive and were not recapitulated using protein profiles at age 6 years. Age 1 modules were associated with environmental factors (Padj range: 2.8×10–10–0.03) and alteration in metabolomic pathways (Padj range: 2.8×10–6–0.04). No genome-wide SNPs were identified for any protein module.

Conclusion:

These findings suggested protein profiles as early as age 1 year predicted development of respiratory-related diseases by age 6. In the future, applying network approaches to study protein profiles may represent a new strategy to identify children susceptible to respiratory-related diseases in the first year of life.

Keywords: Asthma, respiratory infections, immune development, proteomics, metabolomics

CAPSULE SUMMARY:

Network analysis of protein profiles in the first year of life facilitated prediction of respiratory-related diseases, including identification of driving proteins that could be used to identify at-risk children.

Graphical Abstract

graphic file with name nihms-2063644-f0004.jpg

INTRODUCTION

The first year of life is critical to healthy immune and respiratory development,1 and disruptions in normal physiological processes during this period can have long-lasting health consequences, including increased risk of respiratory-related diseases, such as asthma, recurrent wheeze, eczema, and frequent infections.2,3 However, it is difficult for clinicians to identify children prone to these diseases during infancy based solely on clinical information.4 Protein profiling may represent an alternative approach to identifying vulnerable subsets of children, as it has shown success in classifying adult patients based on respiratory disease status and in the formation of disease subtypes.5–7 Protein profiles are currently understudied with respect to their role in childhood respiratory diseases, but additional investigation may reveal important information about the complex, dynamic biological processes mediated by levels of the proteins they capture, including cytokines, chemokines, and growth factors, that have established clinical relevance in respiratory diseases as biomarkers and therapeutic targets.8,9 Further study of early life protein profiles could yield novel insights into the connection between immune development during the first year of life and childhood respiratory diseases.

The position of the proteome in the central dogma of molecular biology10,11 adds complexity to its study, as the proteome is dynamic to changes in upstream and downstream ‘omics. Protein profiles represent the combined influence of genetic and environmental cues12 and can reflect inter-individual variability in disease responses.13 They are further related to downstream impacts on the metabolome14,15 that can provoke alterations to biochemical pathways, ultimately influencing disease. This presents a challenge to the traditional method of investigating the proteome with respect to individual protein features. Applying a network approach to study early life protein profiles – rather than proteins as individual components – is an emerging conceptual framework that has elucidated their role in various disease mechanisms16 and may yield insights when applied to early life protein profiles towards understanding the development of childhood respiratory diseases. Further, integration of clinically-relevant protein networks with additional ‘omics could facilitate characterization of upstream and downstream factors to improve understanding of disease antecedents and pathophysiology.

In this study, we sought to define relationships between protein profiles at age 1 year and the development of childhood respiratory-related diseases by age 6 years- including asthma, recurrent wheeze, respiratory infections, and eczema- using data from the Vitamin D Antenatal Asthma Reduction Trial (VDAART)17 and applying a network approach, weighted correlation network analysis (WGCNA).18 Our hypothesis was that network analysis would identify clinically-relevant modules of highly-correlated proteins during the first year of life; further, that these protein modules would be related to genetic, environmental, and metabolomic factors, enhancing molecular insights across multiple layers of systems biology.19

METHODS

Vitamin D Antenatal Asthma Reduction Trial (VDAART)

The Vitamin D Antenatal Asthma Reduction Trial17,20 was a clinical trial from 2009–2015 that recruited pregnant women between 10–18 weeks’ gestation (GW) and randomized them to a daily vitamin D dose of 4400 IU or 400 IU as normal pregnancy care. VDAART specifically recruited mother-child pairs based on a maternal or paternal history of asthma, eczema, or allergic rhinitis. A subset of 294 mother-child pairs from VDAART were utilized in this study, based on availability of plasma sample for proteomic profiling. Detailed information on VDAART individuals can be found in the Supplementary Methods and in Table 1. Pregnant mothers completed monthly questionnaires throughout the duration of pregnancy, including questions about diet quality,21 exposure to smoking, and demographic/social characteristics, including household income and self-reported race. At delivery, birth characteristics were collected, including birth weight, gestational age at delivery, and mode of delivery. Offspring of VDAART mothers were monitored over the first six years of life at yearly clinical visits and through quarterly questionnaires completed by parents/caregivers. At clinical visits, blood samples were collected from offspring for ‘omic profiling. Informed consent was obtained from all study participants, and ethical approval for this study was obtained from the Partners Human Research Committee at Brigham and Women’s Hospital (Protocol 2014P001109).

Table 1. Characteristics of VDAART Children at Age 1 Year.

A subset of 294 children from the VDAART study with proteomic profiling were included in this study. Clinical outcomes, demographic characteristics, and other social and environmental exposure variables are reported below compared to all VDAART individuals with available clinical data.

Individuals with Protein Data (N=294) VDAART Overall (N=880)
Childhood Clinical Outcomes
Reported Infections ages 0–6 years, Mean (SD) 28.45 (12.94) 28.54 (13.30)
Physician-Diagnosed Asthma ages 0–6 years, N (%) 79 (26.9) 198 (22.5)
Asthma and/or Wheeze Diagnosis ages 0–6 years, N (%) 135 (45.9) 359 (44.6)
Recurrent Wheeze Diagnosis ages 0–6 years, N (%) 129 (43.9) 343 (42.6)
Eczema Diagnosis ages 0–6 years, N (%) 151 (52.8) 369 (52.3)
Demographic Characteristics
Child sex, N male (%) 151 (54.1) 254 (54.2)
Reported Race Category, N (%)
  Black 134 (48.0) 227 (48.4)
  White 89 (31.9) 152 (32.4)
  Other 56 (20.1) 90 (19.2)
Study Site
  Boston, MA, N (%) 89 (31.9) 144 (30.7)
  San Diego, CA, N (%) 93 (33.3) 158 (33.7)
  St. Louis, MO, N (%) 97 (34.8) 167 (35.6)
Reported Annual Household Income, N (%)
  Less than $30,000 82 (39.8) 263 (39.3)
  $30,000-$49,999 37 (18.0) 120 (17.9)
  $50,000-$74,999 32 (15.5) 101 (15.1)
  $75,000-$99,999 25 (12.1) 82 (12.3)
  $100,000-$149,999 22 (10.7) 71 (10.6)
  Over $150,000 8 ( 3.9) 32 ( 4.8)
Reported Maternal Education Level, N (%)
  Did Not Graduate High School 36 (12.9) 108 (12.3)
  High School Graduate 67 (24.0) 225 (25.6)
  Technical School, Junior College, or Some College 82 (29.4) 256 (29.1)
  College Graduate 57 (20.4) 165 (18.8)
  Graduate School or Higher 37 (13.3) 126 (14.3)
Prenatal Exposures
Maternal Vitamin D at 10–18GW in ng/mL, Mean (SD) 23.03 (9.76) 22.80 (10.25)
Maternal Vitamin D at 32–38GW in ng/mL, Mean (SD) 33.38 (14.47) 32.90 (14.63)
Exposure to smoking during pregnancy, N (%) 76 (27.3) 234 (31.1)
Perinatal Exposures
Cord Blood Vitamin D in ng/mL, Mean (SD) 24.31 (11.82) 23.16 (11.73)
Birth weight in kg, Mean (SD) 3.31 (0.50) 3.30 (0.53)
Gestational age at delivery in weeks, Mean (SD) 39.20 (1.47) 38.91 (2.27)
C-section mode of delivery, N (%) 78 (28.0) 241 (29.6)
Birth order, Mean (SD)1 0.92 (1.04) 0.91 (1.06)
Postnatal Exposures
Child Vitamin D at age 1 year in ng/mL, Mean (SD) 29.86 (10.34) 29.65 (10.32)
Child Body Mass Index (BMI) at age 1 year, Mean (SD) 17.45 (2.26) 17.43 (2.15)
Attended daycare between ages 0–1 year, N (%) 31 (11.1) 46 ( 9.8)
Breastfeeding Duration, N (%)
  No breastfeeding 51 (18.3) 106 (22.6)
  Breastfed < 6 months 115 (41.2) 184 (39.2)
  Breastfed 6–12 months 43 (15.4) 80 (17.1)
  Breastfed > 12 months 70 (25.1) 106 (22.6)
Open in a new tab
1

Birth order is derived from the number of children previously birthed by the mother

Clinical Outcomes

A detailed description of asthma, recurrent wheeze, eczema, and infection definitions have been reported previously for VDAART, and these outcomes were collected through quarterly questionnaires.20,22 Asthma was defined as caregiver report of physician-diagnosed asthma, and asthma/wheeze was defined as asthma diagnosis or wheezing in conjunction with use of asthma medication. The asthma/wheeze definition was included as a less strict definition of asthma due to the young age of the children in the VDAART cohort and the potential challenges of diagnosing asthma at this age.20 Recurrent wheeze was defined as maternal or caregiver report of wheeze or the use of any asthma medication in 2 separate years over the first 6 years. Eczema was defined as caregiver report of physician’s diagnosis of eczema. For the purposes of this study, these 3 outcomes were considered “true” if reported at any time during the first 6 years of life. Caregiver reports of numbers of infections that occurred during the first 6 years of life included reports of: pneumonia, bronchiolitis, bronchitis, sinus infection, sore throat, croup, ear infection, and colds. The sum of all infections was used as a continuous variable in this study.

Proteomic, Genomic, and Metabolomic Profiling

Non-fasting blood samples were collected from mothers during pregnancy visits (10–18 GW; 32–38 GW) and from offspring at follow-up visits. Maternal blood samples were assayed for 25OHD levels. In plasma samples collected in offspring at ages 1 and 6 years, levels of 200 immune-mediating proteins were measured using a commercially available NULISA-Seq panel that employs sequential immunocomplex capture and release, then next generation sequencing (NGS) to provide ultra-high sensitivity multiplexing (Alamar Biosciences, Freemont, CA, USA).23 Briefly, the NULISA-Seq platform utilizes a multiplexed enzyme-linked immunosorbent assay (ELISA) approach for capture of proteins that is paired with NGS readout to determine relative protein levels. A full list of protein targets can be found in Supplementary Table 1 and included 124 cytokines, chemokines, and other proteins involved in inflammatory processes across a broad range of immune-related processes. Samples were selected for protein profiling based on existing metabolomic profiling data and availability of remaining plasma sample. Following data collection, quality control processes were applied to protein data; protein levels were normalized to remove technical variation by dividing protein counts for each sample by internal control counts, then normalized across assay plates through normalizing to target-specific medians of inter-plate control samples. Normalized protein counts were centered around zero and scaled by dividing by the standard deviation of each protein column.

Genomic and metabolomic data were utilized from existing data. Briefly, genotyping was performed in children using the Illumina Infinium HumanOmniExpressExome Bead chip (San Diego, CA, USA), as described previously.24 Metabolomic profiling of offspring plasma samples was performed by Metabolon, Inc. (NC, USA) using Metabolon’s global platform that generates data using High Performance Liquid Chromatography coupled to tandem Mass Spectrometry (HPLC-MS/MS).22,25 Additional details of proteomic, genomic, and metabolomic profiling can be found in the Supplementary Methods.

Protein Modules Using Weighted Gene Correlation Network Analysis (WGCNA)

WGCNA18 was used to derive modules of highly-correlated proteins at age 1 year using the WGCNA package in R v4.3.0.26 WGCNA generates modules based on pairwise correlations between protein features. Modules were merged using a cut height (i.e., the Euclidean distance between modules) of 0.3 and a soft power threshold of 6 based on iterative process to identify an optimal number of modules. Module parameters are displayed in Supplementary Figure S1. Following WGCNA, protein groups (i.e., all protein features within a respective module) were input into the STRING database version 12.027 to identify common biological functions, and module hubs were identified based on the greatest number of edges. Modules were summarized as eigenvectors summarizing relative levels of proteins of each module for each subject,28 this eigenvector value was utilized in subsequent statistical models to estimate associations between modules and clinical outcomes, ‘omics, and social/environmental characteristics. Age 1 modules were recapitulated using protein profiling from age 6 samples to assess the consistency of the module relationships with our outcomes if interest over time.

Statistical Analysis

Associations with asthma, asthma/wheeze, recurrent wheeze, and eczema outcomes were evaluated using logistic regression, in which outcomes were defined as “true” if any incidence was reported by caregivers at any point between birth and age 6 years. Poisson regression was utilized for the count of cumulative respiratory infections, which was represented by a continuous variable. In both cases, protein module eigenvalues were treated as predictors. Fully adjusted models included potential confounding variables: sex (1=male; 0=female), race (White used as reference group), breastfeeding duration (0=>12 months; 1=6–12 months; 2=<6 months; 3=none reported), and daycare attendance between birth and age 1 (true/false for any attendance).

Associations between protein module eigenvalues and social, environmental, and demographic variables were also evaluated, specifying module eigenvalues as outcomes. Logistic regression models estimated associations between modules and sex, smoke exposure in pregnancy via cotinine measurement in plasma, mode of delivery, and daycare attendance. Linear models were used for reported annual household income, maternal education, race, breastfeeding duration, maternal diet quality scores, birth weight, birth order, age 1 body mass index (BMI), and vitamin D levels in nanograms per milliliter (ng/mL). Correlations between these variables were estimated using the cor package in R v4.3.0.

Genome-wide associations between protein modules and single nucleotide polymorphisms (SNPs) were investigated using an expression quantitative trait loci (eQTL) approach using the Matrix eQTL package in R.29 Linear regression models estimated associations between protein module eigenvalues and individual metabolites at age 1 year using the glm package in R v4.3.0. Beta estimates and P-values from these models were used as input for metabolomic set enrichment analysis using MetaboAnalyst v6.0.30

All regression analyses employed multiple testing correction controlling for false discovery rate (FDR) using the Benjamini-Hochberg procedure.31 An overview of the study design and statistical analysis procedures can be found in Figure 1.

Figure 1. Overview of Approach.

Figure 1.

Open in a new tab

Plasma samples were collected from 294 children aged 1 and 6 years in the Vitamin D Antenatal Asthma Reduction Trial (VDAART), and targeted protein profiling of 200 immune-mediating proteins was performed using the NULISA-Seq platform from Alamar Biosciences, Inc. Using Weighted Gene Correlation Network Analysis (WGCNA), we derived protein modules based on correlations between plasma protein levels at age 1. After determining protein modules, we recapitulated protein modules at age 6 years and characterized module associations with respiratory-related outcomes, including incidence of asthma, recurrent wheeze, infections, and eczema by age 6 years. Genetic, environmental, and metabolomic factors related to outcomes were assessed for age 1 modules to provide additional biological insights.

RESULTS

VDAART Offspring Clinical Outcomes

Among individuals selected for proteomic profiling, 26.9% met the strict definition of physician-diagnosed asthma, 45.9% had asthma/wheeze, 43.9% had recurrent wheeze, and 52.8% had eczema by age 6. On average, children experienced 28.5 infections in the first 6 years, as reported by caregivers (Table 1). The subset of VDAART children included in this study was heterogeneous in demographic characteristics such as race (48.0% Black, 31.9% White and 20.1% Other), geographical location (31.9% in Boston, MA; 33.3% in San Diego, CA; 34.8% in St. Louis, MO), and maternal education status (33.7% reported bachelor’s degree or higher). A substantial proportion of VDAART participants reported low household income levels, with 39.8% of families reporting less than $30,000 per year. The subset of 294 individuals with protein profiling included in this study was representative of the overall VDAART cohort with respect to outcomes, demographic characteristics, and environmental exposures of interest (Table 1).

Protein module associations with clinical outcomes were time-sensitive

WGCNA produced seven total protein modules based on protein profiles at age 1 year, four of which were associated with clinical outcomes by age 6 years using an FDR-adjusted P-value cutoff of 0.05 (Fig. 2a-b). All P-values reported in results are P-values after FDR correction, unless otherwise specified. Higher eigenvalues for Module 1 and Module 2 at age 1 were associated with higher incidence of asthma/wheeze (Beta[CI]Module1= 5.8[1.6, 9.9], PModule1=0.03; Beta[CI]Module2= 5.2[1.1, 9.2], PModule2=0.03), recurrent wheeze (Beta[CI]Module1= 5.7[2.5, 9.8], PModule1=0.03; Beta[CI]Module2= 4.7[0.7, 8.7], PModule2=0.03), and cumulative infections (Beta[CI]Module1= 1.2[0.8, 1.7], PModule1=6.3×10–9; Beta[CI]Module2= 1.0[0.6, 1.4], PModule2=2.9×10–6), while Module 3 was the only module associated with higher incidence of eczema (Beta[CI]Module3= 6.4[2.2, 10.6], PModule3=0.01). Module 4 demonstrated the opposite directions of effect with higher eigenvalues associated with reduced incidence of asthma/wheeze (Beta[CI]Module4= −4.8[−8.9, −0.7], PModule4=0.03) and recurrent wheeze (Beta[CI]Module4= −4.9[−9.0, −0.8], PModule4=0.03) (Fig. 2a). The directions of effect with the stricter definition of asthma were consistent in direction of effect but did not meet a nominal significance threshold. Higher eigenvalues for all modules were correlated with elevated levels of individual proteins within each respective module (Supplementary Table S1). These results were unchanged after adjusting for sex, race, breastfeeding duration, and daycare attendance (Supplementary Table S2). These protein modules could not be recapitulated in protein profiles at 6 years (Fig. 2b); only the association between Module 2 at 6 years and cumulative infections retained significance (Beta[CI]Module2= 0.6[0.3, 1.01], PModule2=1.1×10−3). Protein correlations with module groupings at 6 years can be found in Supplementary Table S3.

Figure 2. WGCNA identified four time-sensitive protein modules associated with clinically-relevant outcomes.

Figure 2.

Open in a new tab

Four modules were associated with outcomes in a temporally-sensitive manner; associations between modules and clinical outcomes are shown at age 1 year (A) and age 6 years (B). Beta and 95% confidence interval are displayed for each forest plot and colored by module; triangle point shape indicates the association was significant at an adjusted P-value < 0.05. (C) Correlations between individual proteins and Module 1, Module 2, Module 3, and Module 4 plotted in volcano plots. (D) Shows a list of individual protein features sorted into each module by WGCNA in alphabetical order, and module hubs are denoted with a hash mark (#).

Shared biological functions of proteins within modules were characterized using the STRING database (Supplementary Figure S2), and correlations between individual protein levels at age 1 and module eigenvalues were assessed. Module 1 included eleven proteins (Fig. 2c); interleukin-1 receptor antagonist (IL-1Rn) demonstrated the largest correlation (r2=0.88) with Module 1, followed by matrix metallopeptidase 8 (MMP8; r2=0.88). However, IL-1β and C-X-C Motif Chemokine Ligand 8 (CXCL8; also known as IL-8) showed the largest number of edges in the STRING network (i.e., connections between individual proteins based on publicly available sources of protein–protein interaction information) and were considered module “hubs” (Supplementary Figure S2a). Module 2 included six proteins (Fig. 2c), and its eigenvalue showed the strongest correlations with oncostatin M (OSM; r2=0.92) and hepatocyte growth factor (HGF; r2=0.91); OSM, glycoprotein colony-stimulating factor 3 (CSF3), and matrix metalloproteinase-9 MMP9 showed the largest number of edges and were module hubs (Supplementary Figure S2b). Module 3 included only four chemokines with all correlations above 0.8 (Fig. 2c): chemokine ligand 1 (CXCL1; r2=0.85), CXCL2 (r2=0.94), CXCL3 (r2=0.92), and CXCL5 (r2=0.83) with the module eigenvalue; all four proteins demonstrated an equal number of edges (Supplementary Figure S2c). Module 4 included the largest number of individual proteins (Fig. 2c), with TNF Receptor Superfamily Member 8 (TNFRSF8, r2=0.84), cytotoxic T-lymphocyte associated protein 4 (CTLA4, r2=0.82), and IL-2Ra (r2=0.82) demonstrating the highest correlations with module eigenvalue; CTLA4 and TNF were module hubs (Supplementary Figure S2d).

For all modules, determination of module hubs was based on top 2 proteins with largest number of edges. Information for module hubs can be found in Supplementary Table S4.

Protein module associations with environmental variables and ‘omics

Environmental variables related to social determinants of health across the prenatal, postnatal, and demographic categories were associated with at least one protein module at a P-value threshold of 0.05 after multiple testing correction by FDR (Fig. 3a). However, the majority of associations with these variables was present only for Module 3, suggesting higher levels of Module 3 proteins in children were associated with maternal exposure to first or secondhand smoke (Beta[CI]= 9.16[4.3, 14.0]; P=9.1×10−4), non-White race (Beta[CI]= 5.7[4.1, 7.4]; P=2.8×10−10), lower maternal education (Beta[CI]= 4.1[1.7, 6.5]; P=4.4×10−3), and lower household income (Beta[CI]= 6.2[2.8, 9.7]; P=2.2×10−3). Module 3 was associated with poor maternal diet during the first (Beta[CI]= −6.3[−9.6, −2.9]; P=1.09×10−3) and third trimester (Beta[CI]= −7.2[−10.7, −3.6]; P=3.6×10−4) and lower vitamin D levels during the first trimester (Beta[CI]= −3.1[−5.1, −1.2]; P=6.9×10−3). VDAART was originally part of a vitamin D supplementation study, but only one significant association with vitamin D levels at any time period was observed (Supplementary Table S5).

Figure 3. Protein modules were associated with other ‘omics.

Figure 3.

Open in a new tab

(A) Forest plots depict associations between environmental, social, and demographic variables and age 1 modules. Beta and 95% confidence intervals are shown colored by module. Associations meeting a significance threshold of adjusted P-value < 0.05 are denoted with large, open triangle shape. Details of coding of non-continuous variables is available in the Methods section. (B) Correlations between environmental, social, and demographic variables in all VDAART children at age 1 year (N=294); correlation coefficient is shown, and significance is noted by asterisks for P<0.05*, P<0.01**, and P<0.001***. Axis labels are colored by the relevant time period: prenatal (orange), perinatal (light blue), postnatal (forest green), and demographic (pink). (C) Bubble plot displaying enriched metabolomic pathways for each module at age 1 year; size of circle corresponds to the number of metabolites in each pathway, and coloring by module is consistent with panel (A). (D) Genome-wide associations with each age 1 protein module. Genome-wide significance threshold of 5×10−8 is marked by a red line and relaxed threshold of 5×10−7 is marked by a blue line.

Breastfeeding duration and study site were the only variables associated with multiple modules. Shorter breastfeeding duration (i.e., higher continuous values of the coded breastfeeding variable) was associated with higher levels of proteins in Module 1 (Beta[CI]= 2.9[0.9, 5.0], P=0.01), Module 2 (Beta[CI]= 2.4[0.3, 4.5], P=0.03), and Module 3 (Beta[CI]= 5.6[3.6, 7.6]; P=6.1×10−7). Study site was associated with Module 2 (Beta[CI]= 3.1[1.5, 4.9], P=3.9×10−4) and Module 3 (Beta[CI]= 3.1[1.5, 4.8]; P=3.9×10−4) in a direction that suggested location at urban centers located in Boston, MA and St. Louis, MO were associated with higher levels of proteins within these 2 modules compared to San Diego, CA. Correlations between environmental variables demonstrated a number of significant correlations (Fig. 3b) that should be considered when interpreting the influence of any individual factor.

Individual metabolites were associated with Module 1 (N=255 metabolites), Module 2 (N=150 metabolites, Module 3 (N=149 metabolites), and Module 4 (N=83 metabolites), and all associations between metabolites and modules are shown in Supplementary Table S6. The results of the MetaboAnalyst enrichment was based on KEGG pathways (Fig. 3c); Module 1 demonstrated the largest number of enriched pathways, with a total of 12 at an FDR-significant threshold (P-values=2.8×10−6 to 0.04) followed by Module 2 with 5 enriched pathways (P-values=1.49×10−3 to 0.04), then Module 3 with 3 enriched pathways (P-values=0.02), and finally Module 4 with only 1 enriched pathway (P-value=1.39×10−4). All metabolomic pathway enrichment results are available in Table 2. Genetic risk factors were also evaluated, but no genome-wide significant associations between SNPs and protein modules at age 1 year were observed based on a P-value threshold of 5×10−8; only two SNPs demonstrated associations with any module at a relaxed threshold of 5×10−7: rs6465878 of FBXL13 and rs4878832 of ANKRD18A with Module 3 (Fig. 3d-g).

Table 2. Enriched Metabolomic Pathways for Protein Modules at 1 Year of Age.

Regression estimates and P-values from associations between individual metabolites and protein modules at age 1 year were used as input for MetaboAnalyst software to perform metabolite set enrichment analysis using KEGG pathway assignments. The number of metabolites enriched in each pathway for each protein module is shown alongside the FDR-corrected P-value calculated by MetaboAnalyst. Pathways with an FDR<0.05 are shown in the table.

Metabolomic Pathway Protein Module Number of Metabolites Adjusted P-value
Valine, leucine and isoleucine biosynthesis Module 1 4 0.00514
Valine, leucine and isoleucine biosynthesis Module 4 4 0.000139
Taurine and hypotaurine metabolism Module 1 3 0.0403
Taurine and hypotaurine metabolism Module 2 3 0.0286
Purine metabolism Module 1 10 0.0229
Phenylalanine, tyrosine and tryptophan biosynthesis Module 1 4 0.000172
Phenylalanine metabolism Module 1 3 0.0403
Pantothenate and CoA biosynthesis Module 1 5 0.0236
Pantothenate and CoA biosynthesis Module 2 4 0.0448
Glyoxylate and dicarboxylate metabolism Module 1 6 0.0318
Glycine, serine and threonine metabolism Module 1 8 0.0033
Glycine, serine and threonine metabolism Module 3 5 0.0178
Citrate cycle (TCA cycle) Module 1 6 0.00514
Citrate cycle (TCA cycle) Module 2 6 0.00149
Citrate cycle (TCA cycle) Module 3 4 0.0178
Arginine biosynthesis Module 1 8 3.18e-06
Arginine biosynthesis Module 2 4 0.0177
Arginine and proline metabolism Module 1 7 0.0205
Alanine, aspartate and glutamate metabolism Module 1 11 2.75e-06
Alanine, aspartate and glutamate metabolism Module 2 6 0.006
Alanine, aspartate and glutamate metabolism Module 3 5 0.0178
Open in a new tab

DISCUSSION

In this study, we developed modules of highly correlated proteins at age 1 year that were associated with incidence of asthma, asthma/wheeze, recurrent wheeze, respiratory infections, and eczema by age 6 years towards identification of novel biology contributing to the development of these diseases during childhood. The biological milieu present during this critical period of immune development can be challenging to unravel, but applying a network approach allowed our investigation to identify clinically-relevant protein modules and additional exploration of the overlap in modules associated with multiple diseases. This builds upon previous studies that sought to characterize factors leading to asthma,24 recurrent infections,32 and eczema33 during childhood and their biological overlap. Our investigation identified two protein modules associated with the overlap of asthma and recurrent infections and two additional protein modules specifically associated with asthma and with eczema, respectively. Further, protein module associations with outcomes were temporally sensitive, which highlights the particular importance of proteins in first year of life with respect to respiratory health trajectories. These findings may ultimately aid in our understanding of childhood respiratory diseases.

Our results demonstrated overlap in clinical outcomes between protein Modules 1 and 2, which were consistently associated with asthma/wheeze, recurrent wheeze, and infections by age 6 and were consistent in direction of effect with the strictest definition of asthma included in our study. Module 1 captured proteins central to acute phase immune response;34 specifically, IL-1β and CXCL8- two potent pro-inflammatory factors- were Module 1 hubs,35 suggesting alterations in acute phase response during the first year of life may contribute to development of asthma and recurrent respiratory infections. Module 1 demonstrated the highest degree of metabolomic enrichment, including 12 pathways across amino acid and energy metabolism,36 which may reflect subsequent biochemical changes in response to the actions of acute phase response proteins belonging to module. Module 2 was made up of proinflammatory and profibrotic mediators,37,38 further suggesting innate immune mechanisms in early life are critical to respiratory disease development. Module 2 hubs CSF3, MMP9, and OSM participate in IL-6-type signaling that ultimately impacts a range of mucosal immune mechanisms38 and contributes to inflammatory diseases. Metabolomic enrichment of 5 pathways was consistent across Modules 1 and 2, which may suggest similar metabolomic alterations despite differences in innate immune mechanisms represented by each protein module. Reduced breastfeeding, a recognized risk factor in asthma39 and infections,32 was also associated with higher levels of proteins in Modules 1 and 2; no other genetic or environmental variables were identified, including vitamin D levels during pregnancy or in offspring. Due to the early ages included in this analysis, it is possible the associations between these modules and asthma/wheeze and recurrent wheeze are capturing early alterations in protein profiles that impact asthma development.

In contrast, reduced levels of proteins in Module 4 at age 1 year were associated with development of asthma/wheeze and recurrent wheeze. As was observed for Modules 1 and 2, these relationships were consistent using the strictest definition of asthma, but did not meet the threshold for significance. Module 4 proteins were broadly related to T cell regulatory processes, and the hubs of this module- CTLA4 and TNF- are regulators of T cell activation.40 Our findings suggested that reduced levels of these proteins at age 1 were associated with increased risk of childhood asthma and wheeze. This suggested a contrast with previous studies that demonstrated elevated TNF predicted asthma development.41 However, this may be a time-sensitive impact that reflects a failure to establish Th1/Th2 balance postnatally, which can lead to disease.42,43 In addition to TNF as a Module 4 hub, we noted elevation of five TNF superfamily proteins known to participate in regulating the immune responses of T cells;44 with established connections to CXCL1344, further suggesting that lack of T cell regulation in early life may impact subsequent asthma risk. Reduced TNF and TNF superfamily proteins may reflect a lower extent of T cell immune regulation mechanisms at age 1 that ultimately has a negative impact on respiratory disease development. Module 4 was the only module with no associations to environmental or genetic variables. While asthma researchers have uncovered multiple genetic and environmental risk factors for disease,24,45,46 our results suggested that the resultant inflammatory profile represented by Module 4 proteins is not induced by a single genetic or environmental factor. Further, a single metabolomic pathway- valine, leucine and isoleucine biosynthesis- was enriched in this module, and the lowest number of individual metabolite associations were observed for Module 4.

Module 3 was associated with childhood eczema and was comprised of four neutrophil attractant chemokines.47 CXCL1 and CXCL5 have been implicated in neutrophil chemotaxis to the skin during inflammatory episodes prevalent in atopic diseases,48 and our findings suggested increased neutrophilic inflammation is present in early life among infants at high risk of developing eczema, even outside of specific inflammatory episodes concomitant with disease. Enrichment of glycine, alanine, and TCA cycle pathways were also observed for Module 3, which could reflect the increased demand for energy sources in children with higher levels of inflammation mediated by these proteins.49 Importantly, prenatal and demographic variables were highly influential, so the temporal relationship between environmental characteristics and neutrophilic inflammation was confounded. Social determinants of health can greatly influence immune development in early life,50 and our study observed associations between elevated neutrophil chemokines of Module 3 and poor socioeconomic status. Additionally, lower vitamin D levels in early pregnancy were associated with higher levels of Module 3 proteins. While our study does not distinguish whether these are cause-effect relationships, these may represent relevant biomarkers to identify eczema-susceptible children at a very early time period. Further, as Module 3 proteins at age 6 years were associated with increased prevalence of recurrent wheeze, it may stand to reason that these neutrophil attractants could represent relevant targets in the overlap of eczema and other respiratory diseases.

One strength of this study is the availability of longitudinal clinical, demographic, environmental and systems biology data captured in VDAART,20 enabling identification of biomarkers and modules across multiple ‘omics that may influence clinical outcomes. Additionally, the VDAART study collected dense information about asthma and related conditions, so we were able to include three definitions of asthma (physician-diagnosed, asthma and/or wheeze, and recurrent wheeze) to thoroughly investigate this problem. Our study included protein profiles at two time points, age 1 and age 6 years, which allowed us to examine relationships between protein profiles prior to and following diagnosis of respiratory diseases. However, assessing social determinants of health was limited by a lack of information on income in the subset of participants with protein profiling. Another strength is that our protein profiling platform included broad coverage of immune-related proteins and was agnostic to specific immune mechanisms, allowing discovery across multiple immune-relevant pathways. However, relative quantification limited our ability to fully evaluate the time-sensitive impacts, especially given the dynamic nature of the immune system during early life.51 Fully quantitative profiling could enhance the clinical applicability of our findings and will be pursued in future studies. Additionally, the exploratory nature of this study necessitated a large amount of data, presenting a challenge in replication, as it is difficult to identify other cohorts with existing ‘omic profiling that are consistent with respect to this immune development time period. Following the identification of particular protein groups, metabolomic pathways, and environmental exposures presented here, future studies will attempt replication in external populations to validate specific findings and explore relationships over a broader timespan.

In summary, this study demonstrated the utility of applying network analysis to assess the relationship between protein profiles in early life and subsequent development of asthma, wheeze, respiratory infections, and eczema. Our results emphasized the importance of studying protein profiles specifically in the early life period to predict childhood respiratory diseases. Our analysis found that elevated levels of pro-inflammatory proteins of acute phase immune response and mucosal immune mechanisms were observed in children that went on to develop asthma and experience more frequent respiratory infections. Further, that this relationship was specific to the immune development period and was not recapitulated in samples collected at age 6 years. The protein hubs or modules derived in this study could represent novel markers to identify children susceptible to these diseases, ultimately allowing early clinical monitoring and intervention to combat these diseases in a vulnerable population.

Supplementary Material

1
3
4
5
6
7
8
9

KEY MESSAGE:

  • Protein signatures in the first year of life were predictive of childhood asthma, eczema, and more frequent respiratory infections by age 6 years

  • Protein networks were temporally sensitive, highlighting the dynamic role of protein profiles during early life immune development

  • Protein modules were associated with altered metabolomic pathways and were related to social and environmental characteristics

ACKNOWLEDGEMENTS

We thank the participants and study staff of the original VDAART study and the team at Alamar Biosciences, Inc. for their assistance in quality control of the protein data acquired and used for this study.

Funding:

This work was supported by the National Institutes of Health [R01HL123915, R01HL141826, K01HL146980, K01HL175261, T32HL007427, and U19AI168643].

ABBREVIATIONS:

25OHD

25-hydroxy vitamin D

BMI

Body mass index

CI

Confidence interval

CSF3

Colony-stimulating factor 3

CTLA4

Cytotoxic T-lymphocyte associated protein 4

CXCL1

C-X-C motif chemokine ligand 1

CXCL2

C-X-C motif chemokine ligand 2

CXCL3

C-X-C motif chemokine ligand 3

CXCL5

C-X-C motif chemokine ligand 5

CXCL8

C-X-C motif chemokine ligand 8

ELISA

Enzyme-linked immunosorbent assay

eQTL

Expression quantitative trait loci

FDR

False discovery rate

GW

Gestational weeks

HPLC-MS/MS

High-performance liquid chromatography coupled to tandem mass spectrometry

IL-1β

Interleukin 1 beta

IL-1Rn

Interleukin-1 receptor antagonist

IL-2Ra

Interleukin-2 receptor alpha

MMP8

Matrix metallopeptidase-8

MMP9

Matrix metalloproteinase-9

NGS

Next generation sequencing

OSM

Oncostatin M

SNP

Single nucleotide polymorphism

STRING

Search Tool for the Retrieval of Interacting Genes/Proteins

TNF

Tumor necrosis factor

TNFRSF

TNF receptor superfamily

VDAART

Vitamin D Antenatal Asthma Reduction Trial

WGCNA

Weighted gene correlation network analysis

Footnotes

Conflicts of Interest: J. L-S. is a scientific advisor to Precion, Inc and a consultant to Tru Diagnostic, Inc. STW receives royalties from UpToDate and is on the Board of Histolix. OL has consulted for GlaxoSmithKline and HIllevax. All other authors have no relevant competing interests to disclose.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

REFERENCES

  • 1.Moraes-Pinto MI, Suano-Souza F, Aranda CS. Immune system: development and acquisition of immunological competence. J Pediatr (Rio J). Mar-Apr 2021;97 Suppl 1(Suppl 1):S59–S66. doi: 10.1016/j.jped.2020.10.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Nascimento JXPT, Ribeiro CCC, Batista RFL, Alves MTSSBet al. The First 1000 Days of Life Factors Associated with “Childhood Asthma Symptoms”: Brisa Cohort, Brazil. Scientific Reports. 2017/November/22 2017;7(1):16028. doi: 10.1038/s41598-017-16295-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Brodin P, Davis MM. Human immune system variation. Nat Rev Immunol. Jan 2017;17(1):21–29. doi: 10.1038/nri.2016.125 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Baena-Cagnani CE, Badellino HA. Diagnosis of allergy and asthma in childhood. Curr Allergy Asthma Rep. Feb 2011;11(1):71–7. doi: 10.1007/s11882-010-0156-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Kelly RS, Mendez KM, Huang M, Hobbs BD, Clish CB, Gerszten R, et al. Metabo-Endotypes of Asthma Reveal Differences in Lung Function: Discovery and Validation in two TOPMed Cohorts. Am J Respir Crit Care Med. Nov 12 2021;doi: 10.1164/rccm.202105-1268OC [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Zhang Y, Filiou MD, Reckow S, Gormanns P, Maccarrone G, Kessler MS, et al. Proteomic and metabolomic profiling of a trait anxiety mouse model implicate affected pathways. Research Support, Non-U.S. Gov’t. Molecular & cellular proteomics : MCP. Dec 2011;10(12):M111 008110. doi: 10.1074/mcp.M111.008110 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Zhang YH, Hoopmann MR, Castaldi PJ, Simonsen KA, Midha MK, Cho MH, et al. Lung proteomic biomarkers associated with chronic obstructive pulmonary disease. Am J Physiol Lung Cell Mol Physiol. Dec 1 2021;321(6):L1119–L1130. doi: 10.1152/ajplung.00198.2021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Arai K-i, Lee F, Miyajima A, Miyatake S, Arai N, Yokota T. CYTOKINES: COORDINATORS OF IMMUNE AND INFLAMMATORY RESPONSES. Annual Review of Biochemistry. 1990;59(1):783–836. doi: 10.1146/annurev.bi.59.070190.004031 [DOI] [PubMed] [Google Scholar]
  • 9.Lambrecht BN, Hammad H, Fahy JV. The Cytokines of Asthma. Immunity. Apr 16 2019;50(4):975–991. doi: 10.1016/j.immuni.2019.03.018 [DOI] [PubMed] [Google Scholar]
  • 10.Costa Dos Santos G Jr., Renovato-Martins M, de Brito NM. The remodel of the “central dogma”: a metabolomics interaction perspective. Metabolomics. May 9 2021;17(5):48. doi: 10.1007/s11306-021-01800-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Cob M. 60 years ago, Francis Crick changed the logic of biology. PLoS Biol. Sep 2017;15(9):e2003243. doi: 10.1371/journal.pbio.2003243 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Pantazi AC, Balasa AL, Mihai CM, Chisnoiu T, Lupu VV, Kassim MAK, et al. Development of Gut Microbiota in the First 1000 Days after Birth and Potential Interventions. Nutrients. Aug 20 2023;15(16)doi: 10.3390/nu15163647 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Maiorino E, Baek SH, Guo F, Zhou X, Kothari PH, Silverman EK, et al. Discovering the genes mediating the interactions between chronic respiratory diseases in the human interactome. Nat Commun. Feb 10 2020;11(1):811. doi: 10.1038/s41467-020-14600-w [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Fu J, Zhu F, Xu CJ, Li Y. Metabolomics meets systems immunology. EMBO Rep. Apr 5 2023;24(4):e55747. doi: 10.15252/embr.202255747 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Wishart DS. Metabolomics for Investigating Physiological and Pathophysiological Processes. Physiol Rev. Oct 1 2019;99(4):1819–1875. doi: 10.1152/physrev.00035.2018 [DOI] [PubMed] [Google Scholar]
  • 16.Kaczorowski KJ, Shekhar K, Nkulikiyimfura D, Dekker CL, Maecker H, Davis MM, et al. Continuous immunotypes describe human immune variation and predict diverse responses. Proc Natl Acad Sci U S A. Jul 25 2017;114(30):E6097–e6106. doi: 10.1073/pnas.1705065114 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Litonjua AA, Lange NE, Carey VJ, Brown S, Laranjo N, Harshfield BJ, et al. The Vitamin D Antenatal Asthma Reduction Trial (VDAART): rationale, design, and methods of a randomized, controlled trial of vitamin D supplementation in pregnancy for the primary prevention of asthma and allergies in children. Contemp Clin Trials. May 2014;38(1):37–50. doi: 10.1016/j.cct.2014.02.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. Research Support, N.I.H., Extramural. BMC bioinformatics. 2008;9:559. doi: 10.1186/1471-2105-9-559 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Babu M, Snyder M. Multi-Omics Profiling for Health. Mol Cell Proteomics. Jun 2023;22(6):100561. doi: 10.1016/j.mcpro.2023.100561 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Litonjua AA, Carey VJ, Laranjo N, Stubbs BJ, Mirzakhani H, O’Connor GT, et al. Six-Year Follow-up of a Trial of Antenatal Vitamin D for Asthma Reduction. N Engl J Med. Feb 6 2020;382(6):525–533. doi: 10.1056/NEJMoa1906137 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Savage JH, Lee-Sarwar KA, Sordillo JE, Lange NE, Zhou Y, O’Connor GT, et al. Diet during Pregnancy and Infancy and the Infant Intestinal Microbiome. J Pediatr. Dec 2018;203:47–54 e4. doi: 10.1016/j.jpeds.2018.07.066 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Prince N, Kim M, Kelly RS, Diray-Arce J, Bonnelykke K, Chawes BL, et al. Reduced Steroid Metabolites Identify Infection-Prone Children in Two Independent Pre-Birth Cohorts. Metabolites. Nov 13 2022;12(11)doi: 10.3390/metabo12111108 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Feng W, Beer J, Hao Q, Ariyapala IS, Sahajan A, Komarov A, et al. NULISA: a novel proteomic liquid biopsy platform with attomolar sensitivity and high multiplexing. bioRxiv. Jun 2 2023;doi: 10.1101/2023.04.09.536130 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Kelly RS, Chawes BL, Guo F, Zhang L, Blighe K, Litonjua AA, et al. The role of the 17q21 genotype in the prevention of early childhood asthma and recurrent wheeze by vitamin D. Eur Respir J. Oct 2019;54(4)doi: 10.1183/13993003.00761-2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Huang M, Kelly RS, Chu SH, Kachroo P, Gurdeniz G, Chawes BL, et al. Maternal Metabolome in Pregnancy and Childhood Asthma or Recurrent Wheeze in the Vitamin D Antenatal Asthma Reduction Trial. Metabolites. Jan 23 2021;11(2)doi: 10.3390/metabo11020065 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.R: A language and environment for statistical computing. 2012. http://www.R-project.org/ [Google Scholar]
  • 27.Szklarczyk D, Kirsch R, Koutrouli M, Nastou K, Mehryary F, Hachilif R, et al. The STRING database in 2023: protein–protein association networks and functional enrichment analyses for any sequenced genome of interest. Nucleic Acids Research. 2022;51(D1):D638–D646. doi: 10.1093/nar/gkac1000 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Lee-Sarwar KA, Kelly RS, Lasky-Su J, Zeiger RS, O’Connor GT, Sandel MT, et al. Integrative analysis of the intestinal metabolome of childhood asthma. J Allergy Clin Immunol. Aug 2019;144(2):442–454. doi: 10.1016/j.jaci.2019.02.032 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Shabalin AA. Matrix eQTL: ultra fast eQTL analysis via large matrix operations. Bioinformatics. 2012;28(10):1353–1358. doi: 10.1093/bioinformatics/bts163 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Lu Y, Pang Z, Xia J. Comprehensive investigation of pathway enrichment methods for functional interpretation of LC–MS global metabolomics data. Briefings in Bioinformatics. 2022;24(1)doi: 10.1093/bib/bbac553 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Benjamini Y, Hochberg Y. Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society: Series B (Methodological). 1995;57(1):289–300. [Google Scholar]
  • 32.Vissing NH, Chawes BL, Rasmussen MA, Bisgaard H. Epidemiology and Risk Factors of Infection in Early Childhood. Pediatrics. Jun 2018;141(6)doi: 10.1542/peds.2017-0933 [DOI] [PubMed] [Google Scholar]
  • 33.Novak N, Bieber T, Leung DY. Immune mechanisms leading to atopic dermatitis. J Allergy Clin Immunol. Dec 2003;112(6 Suppl):S128–39. doi: 10.1016/j.jaci.2003.09.032 [DOI] [PubMed] [Google Scholar]
  • 34.Pepys MB. Acute Phase Proteins. In: Delves PJ, ed. Encyclopedia of Immunology (Second Edition). Elsevier; 1998:18–20. [Google Scholar]
  • 35.Gruys E, Toussaint MJ, Niewold TA, Koopmans SJ. Acute phase reaction and acute phase proteins. J Zhejiang Univ Sci B. Nov 2005;6(11):1045–56. doi: 10.1631/jzus.2005.B1045 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Le Floc’h N, Melchior D, Obled C. Modifications of protein and amino acid metabolism during inflammation and immune system activation. Livestock Production Science. 2004/April/01/2004;87(1):37–45. 10.1016/j.livprodsci.2003.09.005 [DOI] [Google Scholar]
  • 37.Turner MD, Nedjai B, Hurst T, Pennington DJ. Cytokines and chemokines: At the crossroads of cell signalling and inflammatory disease. Biochim Biophys Acta. Nov 2014;1843(11):2563–2582. doi: 10.1016/j.bbamcr.2014.05.014 [DOI] [PubMed] [Google Scholar]
  • 38.Heinrich PC, Behrmann I, Muller-Newen G, Schaper F, Graeve L. Interleukin-6-type cytokine signalling through the gp130/Jak/STAT pathway. Biochem J. Sep 1 1998;334 ( Pt 2)(Pt 2):297–314. doi: 10.1042/bj3340297 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Xue M, Dehaas E, Chaudhary N, O’Byrne P, Satia I, Kurmi OP. Breastfeeding and risk of childhood asthma: a systematic review and meta-analysis. ERJ Open Res. Oct 2021;7(4)doi: 10.1183/23120541.00504-2021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Sun L, Su Y, Jiao A, Wang X, Zhang B. T cells in health and disease. Signal Transduct Target Ther. Jun 19 2023;8(1):235. doi: 10.1038/s41392-023-01471-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Halonen M, Lohman IC, Stern DA, Ellis WL, Rothers J, Wright AL. Perinatal tumor necrosis factor-alpha production, influenced by maternal pregnancy weight gain, predicts childhood asthma. Am J Respir Crit Care Med. Jul 1 2013;188(1):35–41. doi: 10.1164/rccm.201207-1265OC [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Brunne PM Early immunologic changes during the onset of atopic dermatitis. Ann Allergy Asthma Immunol. Aug 2019;123(2):152–157. doi: 10.1016/j.anai.2019.03.033 [DOI] [PubMed] [Google Scholar]
  • 43.Luo W, Hu J, Xu W, Dong J. Distinct spatial and temporal roles for Th1, Th2, and Th17 cells in asthma. Front Immunol. 2022;13:974066. doi: 10.3389/fimmu.2022.974066 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Croft M The role of TNF superfamily members in T-cell function and diseases. Nat Rev Immunol. Apr 2009;9(4):271–85. doi: 10.1038/nri2526 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Ferreira MA, McRae AF, Medland SE, Nyholt DR, Gordon SD, Wright MJ, et al. Association between ORMDL3, IL1RL1 and a deletion on chromosome 17q21 with asthma risk in Australia. Research Support, N.I.H., Extramural Research Support, Non-U.S. Gov’t. European journal of human genetics : EJHG. Apr 2011;19(4):458–64. doi: 10.1038/ejhg.2010.191 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Stern J, Pier J, Litonjua AA. Asthma epidemiology and risk factors. Semin Immunopathol. Feb 2020;42(1):5–15. doi: 10.1007/s00281-020-00785-1 [DOI] [PubMed] [Google Scholar]
  • 47.Rajarathnam K, Schnoor M, Richardson RM, Rajagopal S. How do chemokines navigate neutrophils to the target site: Dissecting the structural mechanisms and signaling pathways. Cell Signal. Feb 2019;54:69–80. doi: 10.1016/j.cellsig.2018.11.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Walsh T, McClellan JM, McCarthy SE, Addington AM, Pierce SB, Cooper GM, et al. Rare structural variants disrupt multiple genes in neurodevelopmental pathways in schizophrenia. Research Support, N.I.H., Extramural Research Support, N.I.H., Intramural Research Support, Non-U.S. Gov’t. Science. Apr 25 2008;320(5875):539–43. doi: 10.1126/science.1155174 [DOI] [PubMed] [Google Scholar]
  • 49.Chen S, Saeed A, Liu Q, Jiang Q, Xu H, Xiao GG, Rao L, et al. Macrophages in immunoregulation and therapeutics. Signal Transduct Target Ther. May 22 2023;8(1):207. doi: 10.1038/s41392-023-01452-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Georgountzou A, Papadopoulos NG. Postnatal Innate Immune Development: From Birth to Adulthood. Front Immunol. 2017;8:957. doi: 10.3389/fimmu.2017.00957 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Lee AH, Shannon CP, Amenyogbe N, Bennike TB, Diray-Arce J, Idoko OT, et al. Dynamic molecular changes during the first week of human life follow a robust developmental trajectory. Nat Commun. Mar 12 2019;10(1):1092. doi: 10.1038/s41467-019-08794-x [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS2063644-supplement-1.docx (9.8MB, docx)
3
NIHMS2063644-supplement-3.jpg (1.5MB, jpg)
4
NIHMS2063644-supplement-4.xlsx (39KB, xlsx)
5
NIHMS2063644-supplement-5.xlsx (12.9KB, xlsx)
6
NIHMS2063644-supplement-6.xlsx (39KB, xlsx)
7
NIHMS2063644-supplement-7.xlsx (11.8KB, xlsx)
8
NIHMS2063644-supplement-8.xlsx (17.9KB, xlsx)
9
NIHMS2063644-supplement-9.xlsx (308.2KB, xlsx)

RESOURCES