Large-scale genetic and immunological profiling reveals key environmental and genetic drivers of immunological diversity within the healthy human population.
Diversity is an essential characteristic of the immune system, a critical bulwark against pathogen specialization. The idiosyncrasy of each person’s immune system not only slows the spread of pathogens within a population but also contributes to that person’s risk of developing immune system–mediated disease. In a study published in this issue of Nature Immunology, the Milieu Intérieur Consortium identifies critical genetic and environmental factors that modify the immune system of healthy people1.
Variation in the immune system has been delineated by several systems-immunology studies showing that 50% of such variation is driven by environmental effects2 and 30–40% of the variation is driven by genetic variation3,4. The Milieu Intérieur Consortium has produced a large-scale analysis of 166 immune-system phenotypes of 1,000 healthy people that identifies previously unknown environmental and genetic factors that drive distinct immunological changes (Fig. 1). Participants in the study were surveyed extensively for multiple demographic variables, including age, sex, past infections, vaccination and health-related habits. Strong influences of age, sex and cytomegalovirus (CMV) infection on immunological traits are observed, as has been reported previously2,3,5,6. The strongest effect of age and CMV status is skewing within T cell populations from naive subsets to inflammatory subsets, a common reoccurring theme of studies of aging of the immune system. The effect of sex on the immune system replicates previously identified T cell phenotypes but also finds a surprising association of sex with the number of mucosa-associated T cells. Such differences might reflect hormone-dependent changes, with a potential effect on the differential susceptibility of males and females to immune system–related diseases. Alternatively, the differences attributed to biological sex could instead reflect gender, as the different environments that men and women are typically exposed to are known to drive profound alterations in the peripheral blood, independently of biological sex7. In addition to those known relationships, Albert and colleagues here add the influence of smoking. Smoking influences 36% of all traits, with active smoking being associated with the largest effects on the composition of the T cell and B cell pools. In particular, strong skewing toward memory-type B cell subsets and activated regulatory T cell subsets is observed, indicative of the inflammatory processes associated with smoking. Other physiological factors (metabolic score and body mass index), economic factors (income, education and housing) and infections (rubella, mumps, measles and chicken pox) have no measurable effect on the immune system in this cohort.
On the other side of the ‘nurture-versus-nature’ dichotomy, Albert and colleagues use a genome-wide association study approach1. Three pioneering immunology genome-wide association studies have identified 32 associations corresponding to 26 loci associated with at least one immunological trait3,6,8. Albert and colleagues report 15 loci associated with 42 immunological traits, of which 12 loci are previously unknown1. This brings the total to 47 genome-wide significant associations, which correspond to 38 distinct loci. Only 9 of the 38 loci are reported in more than one study, with immunological genome-wide association studies still in the novel discovery phase. Notably, of the 42 immunological traits with significant genetic association, 86% relate to the expression of a single marker, and 78% of these associations are within the vicinity of the corresponding gene (i.e., local protein quantitative trait loci). Natural killer cells are particularly rich in genetic influences, as more than half of the identified genetic interactions are based on measurements of natural killer cells. The findings are in contrast with those of an earlier study in which overall heritability was lower for surface expression of proteins than for cellular traits4; however, direct comparisons between studies are problematic.
A key feature of this study1 is the striking differences between the innate immune system and adaptive immune system that are observed. While high levels of variation are observed in both, the global analysis highlights the role of the environment in shaping the adaptive immune system and the role of genetics in molding the innate immune system. Around 50% less of the variation in the innate immune system could be explained through the environment, while, conversely, ~70% more of the variation in the innate immune system could be explained by genetic background, relative to such explanations for variations in adaptive immunity1. To a degree, these findings could be driven by the selection of immunological phenotypes measured, as the inverse result was found in another published study4. While roughly equal numbers of innate phenotypes and adaptive phenotypes are tested in the Milieu Intérieur cohort, the adaptive phenotypes are heavily skewed toward cellular subsets, while innate phenotypes are based largely on the expression of functional markers (which allows the detection of protein quantitative trait loci). However, the observed dichotomy of genes versus environment cannot be explained simply by cellular subsets versus functional markers, which indicates that a fundamental difference might exist in the origin of variation in the innate immune system versus that in the adaptive immune system. There are many potential sources for this division. Lymphoid cells in general live longer than myeloid cells, which gives the former a greater chance for the accumulation of environmental influences. Lymphoid cells also rely on clonal expansion and cellular differentiation, with decades-long immunological memory built into their cellular response. Indeed, the genetic associations observed in the adaptive immune system are identified largely in the naive compartment, while strong environmental effects such as CMV drive changes in the antigen-experienced compartment. Alternatively, the division might reflect an ancient evolutionary separation in mechanisms to achieve diversity. Natural killer cells have a large number of highly polymorphic functional receptors, which diminishes capacity of viruses to evade the immune system9. In contrast, in the adaptive immune system, the obligate diversity can be encoded either somatically (for T cell and B cell antigen receptors) or in a single hyper-variable locus (HLA). Ultimately, larger and deeper systems-immunology studies will be needed to find the answer to this intriguing dilemma.
A final ‘take-away’ from this study1 is just how much remains to be discovered about the drivers of variation in the human immune system. Known intrinsic factors (age and sex), environmental factors (CMV infection and smoking) and genetic factors (with a P value of <5 × 10−8) explain at most 15–20% of the variance in the number of cells in leukocyte subsets (with the exception of a few T cell subsets for which the combined effect of age and CMV infection is up to 40%)1. Likewise, for immunological markers, known factors control only ~5–10% of the expression of most such markers, and even the exceptions (smoking and expression of CD38 on memory B cells; protein quantitative trait loci for natural killer cell markers) explain less than 50% of the observed variance1. On both the genetic front and environmental front, the clear majority of variation in the immune system of healthy people remains unaccounted for. Key environmental influencers, the effect of rare genetic variants and, most problematically, gene–environment interactions must be elucidated for full understanding of the diversity of the human immune system.