T cell repertoire breadth is associated with the number of acute respiratory infections in the LoewenKIDS birth cohort

Abstract

We set out to gain insight into peripheral blood B and T cell repertoires from 120 infants of the LoewenKIDS birth cohort to investigate potential determinants of early life respiratory infections. Low antigen-dependent somatic hypermutation of B cell repertoires, as well as low T and B cell repertoire clonality, high diversity, and high richness especially in public T cell clonotypes reflected the immunological naivety at 12 months of age when high thymic and bone marrow output are associated with relatively few prior antigen encounters. Infants with inadequately low T cell repertoire diversity or high clonality showed higher numbers of acute respiratory infections over the first 4 years of life. No correlation of T or B cell repertoire metrics with other parameters such as sex, birth mode, older siblings, pets, the onset of daycare, or duration of breast feeding was noted. Together, this study supports that-regardless of T cell functionality-the breadth of the T cell repertoire is associated with the number of acute respiratory infections in the first 4 years of life. Moreover, this study provides a valuable resource of millions of T and B cell receptor sequences from infants with available metadata for researchers in the field.

Figure 1

Acute respiratory infections (ARI) in the LoewenKIDS subcohort. (A) Visual description of the study design. (B) ARI episodes of all children with > 80% symptom diary completeness (n = 67) in the first 4 years of life.

Figure 2

Blood immune repertoire metrics of LoewenKIDS subcohort sampled at 12 months compared to older control individuals sampled in their 1st to 9th decade of life. (A) T cell receptor (TCR) repertoire clonality, richness and two diversity measures are shown for the LoewenKIDS subcohort (“infant”) versus control immune repertoires from older individuals in their 1st to 9th decade of life (“decade 1–9”; dec). n(dec0) = 116, n(dec1) = 6, n(dec2) = 9, n(dec3) = 54, n(dec4) = 69, n(dec5) = 70, n(dec6) = 63, n(dec7) = 54, n(dec8) = 43, n(dec9) = 7. (B) B cell receptor (BCR) repertoire clonality, richness and two diversity measures are shown for the LoewenKIDS subcohort (“infant”) versus control immune repertoires from older individuals in their 1st to 9th decade of life (“decade 1–9”; dec). n(dec0) = 116, n(dec1) = 7, n(dec2) = 11, n(dec3) = 49, n(dec4) = 55, n(dec5) = 61, n(dec6) = 60, n(dec7) = 48, n(dec8) = 38, n(dec9) = 7. (C) Mean lengths of TCR complementarity-determining region 3 (CDR3) in LoewenKIDS subcohort (“infant”; n = 116) versus controls (“decade 1–9”; n = 377). (D) Mean lengths of BCR CDR3 in LoewenKIDS subcohort (“infant”; n = 116) versus controls (“decade 1–9”; n = 336). (E) Somatic hypermutation (SHM) of BCR in LoewenKIDS subcohort (“infant”) versus controls (“decade 1–9”). (F) Generation probability (Pgen) of TCR rearrangements in LoewenKIDS subcohort (“infant”) versus controls (“decade 1–9”). (G) Principal component analysis (PCA) of TCR V gene usage in LoewenKIDS subcohort (“infant”) versus controls in their 1st to 9th decade of life. V genes contributing most to the repertoire skewing across all age groups are shown. The dotted line indicates the contribution if variables were evenly distributed. (H) PCA of BCR V gene usage in LoewenKIDS subcohort (“infant”) versus controls in their 1st to 9th decade of life. V genes contributing most to the repertoire skewing across all age groups are shown. The dotted line indicates the contribution if variables were evenly distributed. One-way ANOVA was used for Panels (A) and (B). For Panels (C–F) an unpaired two-tailed t-test was performed. For Panels G and H, Pillai-MANOVA was used as statistical test. Analyses and data plotting were performed using RStudio (version 1.1.456) and the tcR, ade4 and tidyverse packages.

Figure 3

Correlation of blood immune cell metrics with acute respiratory infections (ARI) in LoewenKIDS subcohort. (A) T cell receptor (TCR) repertoire clonality, richness and two diversity measures are shown for the LoewenKIDS subcohort in relation to ARI in the 1st year of life. (B) TCR repertoire clonality, richness and two diversity measures are shown for the LoewenKIDS subcohort in relation to ARI in the 2nd to 4th year of life. (C) B cell receptor (BCR) repertoire clonality, richness, two diversity measures and somatic hypermutation (SHM) are shown for the LoewenKIDS subcohort in relation to ARI in the 1st year of life. (D) BCR repertoire clonality, richness, two diversity measures and SHM are shown for the LoewenKIDS subcohort in relation to ARI in the 2nd to 4th year of life. Only subjects with > 80% of days covered in the symptom diary were included in the analyses (89 subjects for the analysis of year 1 and 65 subjects for the analysis of clonality as well as 66 subjects for the analysis of the other immune metrics in year 2–4). One-way ANOVA was used as statistical test and squared Pearson correlation coefficients R² are shown. Analyses and data plotting were performed using RStudio (version 1.1.456) and the tcR, ade4 and tidyverse packages.

Figure 4

Correlation of blood T cell metrics with birth mode, breast feeding, siblings, daycare and other potential determinants in LoewenKIDS subcohort. T cell receptor (TCR) repertoire clonality, richness and two diversity measures are shown for the LoewenKIDS subcohort in relation to sex, birth mode, siblings, pets, onset of daycare and duration of breast feeding. Unpaired two-tailed t-test and one-way ANOVA were used as statistical tests. Squared Pearson correlation coefficients R² are shown. Analyses and data plotting were performed using RStudio (version 1.1.456) and the tcR, ade4 and tidyverse packages.

Figure 5

Adjusted linear effect on the cumulative number of ARI at the age of 4 years. Separate models show the association for each T cell repertoire index (Shannon diversity, Simpson diversity, richness, clonality) in comparison with the effect of older siblings, while mutually adjusting (n = 67). Estimates and 95% confidence intervals are shown. Estimates indicate number of additional infections. Analyses and data plotting were performed using RStudio (version 1.1.456).