Innate and adaptive immune traits are differentially affected by genetic and environmental factors

Abstract

The diversity and activity of leukocytes is controlled by genetic and environmental influences to maintain balanced immune responses. However, the relative contribution of environmental compared with genetic factors that affect variations in immune traits is unknown. Here we analyse 23,394 immune phenotypes in 497 adult female twins. 76% of these traits show a predominantly heritable influence, whereas 24% are mostly influenced by environment. These data highlight the importance of shared childhood environmental influences such as diet, infections or microbes in shaping immune homeostasis for monocytes, B1 cells, γδ T cells and NKT cells, whereas dendritic cells, B2 cells, CD4⁺ T and CD8⁺ T cells are more influenced by genetics. Although leukocyte subsets are influenced by genetics and environment, adaptive immune traits are more affected by genetics, whereas innate immune traits are more affected by environment.

Figure 1. Estimating heritability of immune traits.

(a) Trait values from MZ twins, DZ twins or longitudinal specimens were correlated; the Pearson''s correlation coefficients for all traits are shown as histograms. (b) The distribution of heritability of all traits as estimated by Falconer's formula (that is, twice the difference of the Pearson's r for MZ and DZ twins). (c) SeqM was performed on all robust traits. The traits were grouped according to cell subsets as shown under the axis. The distribution of genetic associations for traits in each category is shown as a bar chart: thick bars indicate the interquartile range (with a break at the median); thin lines indicate the 10th–90th percentiles, with dots for outliers. It is noteworthy that traits with heritabilities <60% are not shown here.

Figure 2. Proportion of trait variation explained by heritable or environmental influence.

Based on SeqM, the fraction of variation in a trait is calculated for heritable, shared environmental or unique environmental factors. This is shown in circle plots for CSF traits (a) and SPEL traits (b). Traits are arranged in order by the major lineage to which they belong (as indicated inside each circle); for SPEL traits, they were further ordered according to which cell surface protein was quantified (some of which are indicated outside the circle).

Figure 3. Trait variation determination in T cell subsets.

The subset of T cell traits shown in Fig. 2a are expanded. (a,b) T-cell traits for non-CD4 T cells (a), including CD8 T, γδ T and NKT cells (defined as CD1d-multimer binding), and for CD4 T cells (b). Traits were arranged according to the subsets from which they were defined: ‘Differentiation'—subsets defined by the markers CD27, CD28, CD31, CD45RA, CD57, CD95, CD127 and CD244 (Staining panel 1; see Supplementary Fig. 2 in Roederer et al.8). ‘Activation'—subsets defined by the markers CD25, CD38, HLA-DR and CD279/PD-1 (Staining panel 2). ‘Polarization'—subsets defined by the markers CCR4, CCR6, CCR10, CXCR3 and CXCR5 (Staining panel 3). ‘Treg'—subsets defined by the markers CD25, CD127, CD39, CD45RO and CD73 (Staining panel 2). (c) Traits for only Treg cells (expressing CD39 and/or CD73). Of the 20 traits showing shared environmental influence, 18 arise from subsets co-expressing CD25 and CD73 with variable expression of other markers.

Figure 4. Trait variation determination in non-T-cell subsets.

(a) B-cell traits were arranged according to lineage (B1 or conventional B2) and among B2, by surface isotype expression (here, cells that co-express IgM and IgD are labelled as IgD; those that express only IgM are labelled as IgM). (b) DC traits are arranged by subset phenotype. ‘pDC', plasmacytoid DC (CD11c⁻CD123⁺). ‘xDC', DC co-expressing CD11c and CD123. All others are myeloid DC (mDC, CD11c⁺CD123⁻), including mDC-1 (CD1c⁻CD16⁻), antigen-presenting cells for CD4 T cells (‘CD4 APC', CD1c⁺), antigen-presenting cells for CD8 T cells (‘CD8 APC', CD141⁺) and inflammatory mDC (‘imDC', CD1c⁻CD16⁺). (c) Monocyte traits are divided by the coordinate expression of CD16 and CD274.

Figure 5. Influences on leukocyte subset representation.

SEqM quantifies the degree of variation in traits that can be ascribed to heritable versus shared environmental influences. Here are illustrated divergent trends in major subsets. For each comparison, the illustration indicates a relative increase in either heritable or shared environmental influences. For example, subsets of CD4 T cells defined by polarization markers (chemokine receptor expression) show increased shared environmental influences compared to C4 subsets defined by activation markers (CD38, HLA-DR).