We analyzed the recent evolutionary history of IL genes in humans by integrating information on environmental variables with classical population genetics and association studies. Results herein suggest that microbes and parasitic worms played a relevant role as selective agents, but the pressure imposed by helminths on IL genes has been stronger than the one caused by viral and microbial agents. Helminths were present among our ancestors before the emergence of humans as a species (for review see reference [40
]). These parasites evolve at lower rates than viruses and bacteria and, in contrast to most viral/microbial agents, are able to maintain themselves in small human communities (41
). Notably, by establishing chronic infections, parasitic worms affect the susceptibility of their host to viruses, bacteria, and protozoa (for review see reference [2
]). Therefore, helminths might have represented a stable threat to human populations and their distribution, which is not associated with sudden epidemics and, as in the case of micropathogens, might have left stronger genetic signatures.
A limitation of our study is that we implicitly assumed that the number of different pathogen species/genera per country has been maintained proportionally unchanged along human evolutionary history. Although clearly an oversimplification, this might reflect reality to some degree, given that climatic variables (e.g., precipitation rates and temperature) have a primary importance in driving the spatial distribution of human micro- and macropathogens (21
). Therefore, while the fitness cost imposed by specific species/genera might have evolved rapidly, the relative number of pathogen species per country might have changed proportionally less.
Another possible caveat of our results concerns the definition of “pathogen,” in that we included any organism that can cause a disease irrespective of its virulence or pathogenicity. The reason for this choice is that the fitness of a pathogen is a direct measure of the ability of such pathogen to replicate within a given environment. Fitness is dependent on both the features of the pathogen and of the host. The features of the pathogen include its abilities to (a) replicate within the host, (b) select a proper cell target (tropism), (c) avoid the immune response, and (d) escape the obstacles posed by drugs. The features of the host include the immune response, the genetic background, and the availability of target cells. The reciprocal balance between these factors determines the virulence of a pathogen, a feature that, therefore, cannot be considered as a constant, but rather evolves dynamically over time.
Despite these limitations, we were able to identify several variants that are likely candidates for pathogen-driven selection, and the suitability of this approach is confirmed by the previous demonstration that variability at HLA
genes correlates with a similar measure of pathogen richness (10
). Our data point to the SNPs in as good candidates for experimental analyses aimed at inferring their role in IL gene function, given that a signature of natural selection necessarily implies the presence of a functional variant (either the correlated variant itself or a linked one). Also, genes subjected to a selective pressure from infectious diseases should be regarded (42
) as obvious candidates for genetic epidemiology studies (e.g., case-control studies).
It is worth noting that most members of the IL-1 signaling pathway were observed to correlate with pathogen richness. IL-1A and IL-1B are pleiotropic cytokines with central roles in immune and inflammatory responses (43
) required for the development of Th2-mediated immunity and protection against chronic infection in mice (44
). The observation that SNPs strongly correlated with micro- and macro-pathogen richness map to IL1RAPL1
is more puzzling, as this gene is not known to be associated with infection and immune response, but is involved in brain development and function (45
); nonetheless, this gene is expressed in tissues that are different from brain (46
), suggesting that it plays other roles apart from neurodevelopment.
Strong correlations with macropathogen richness were obtained for IL4
, and IL10
. These molecules are pivotal to the elicitation of Th2 response, which is the immune response central to helminth resistance (40
). The strongest correlation with macropathogens was observed for rs12044804 in IL19
(τ = −0.62). This gene is located within an IL cluster, which also comprises IL10
, and is encompassed by a low-frequency CNV. The low levels of LD across the IL cluster (Fig. S2) suggest that IL19
, and IL10
SNPs are independently associated with pathogen richness. IL19
, and IL26
all belong to the IL-20 subfamily, are all produced by different leukocyte populations, and all bind to partially shared receptors that are mainly expressed by epithelial cells and known to promote keratinocyte growth and to induce skin inflammatory responses (48
). The correlation of SNPs in most of these genes with micropathogen and helminth richness suggests that modulation of skin immunological properties might represent an adaptive response to parasite species that infect humans through the skin.
SNPs in IL2RB
, and IL15RA
also correlated with macropathogen richness; these genes converge on the same pathway as IL-2RB and IL-15RA are part of a trimeric complex that binds IL-15 (50
). IL-15 has a central role in intestinal inflammatory processes and in the pathogenesis of CD and CeD (51
), possibly participating in the immune protection of gut tissues. An interesting observation is that IL-2RB, via IL-15 binding, regulates the intestinal epithelial barrier by transducing signals that result in tight junction formation (52
). Variation of mucosal permeability after nematode infection is a host defense response and is important for efficient parasite expulsion (53
); the correlation we observed between SNPs in these genes and macropathogen richness might indicate selection for improved intestinal clearance of nematodes. Again, mouse models will prove central in addressing the role of these loci in parasite expulsion; e.g., in the case of IL4R
, which also correlates with helminth richness (), treatment with antibodies or use of il4r
-deficient mice prevented expulsion of Heligmosomoides polygyrus
, Trichuris muris
, and Nippostrongylus brasiliensis
A major locus for susceptibility to Schistosoma mansoni
) has previously been mapped to 5q31-q33 (55
). This region covers IL4
, and IL12B
, as well as other candidate loci (IRF1
. Although our data do not allow inference on which gene is responsible for increased susceptibility to schistosomiasis, it is worth noting that four SNPs in IL4
display very strong correlation with helminth richness. IL4
might therefore be regarded as a candidate locus for susceptibility to S. mansoni
infection, with dedicated association studies being required to verify this prediction. Conversely, the SM2
region (6q22-q23), where a locus controlling hepatic fibrosis in S. mansoni
infection has been mapped (56
), harbors no IL genes.
Pathogen-driven variations in allele frequencies can occur under different selection scenarios, such as directional or balancing selection; the latter itself is the result of an initial stage of positive selection that favors the spread in a population of a new allele until selection opposes its fixation and a balanced situation is established. Common causes of balancing selection include heterozygote advantage, changing environmental conditions, and frequency-dependent selection, all of these possibly applying to host–pathogen interactions. Also, given the pleiotropic roles of many IL genes, selective pressures different from pathogen richness might affect the evolutionary fate of these loci. Classical population genetics analyses indicated that five IL genes are likely targets of balancing selection. IL1F5
, and IL1F10
are recently discovered IL-1 family members (57
) that are located within the IL1
gene cluster; based on comparative analyses, IL-1F5 and IL-1F10 are predicted to act as antagonists (58
is a regulator of skin and brain inflammation (59
) and it is expressed in many different human tissues (57
). Interestingly, an excess of heterozygotes was observed for IL1F5
, suggesting that overdominance might underly the maintenance of a balanced polymorphism in the gene. Overdominance is rare in humans (36
) and is hypothesized to enhance immune response flexibility by modulating allele-specific gene expression in different cell types and in response to diverse stimuli/cytokines (62
). Whether this is the case for IL1F5
remains to be verified.
is a still relatively unknown protein mainly expressed in skin, proliferating B cells, and tonsils (63
). One of the intermediate frequency SNPs in the gene is accounted for by a missense substitution that replaces an aspartic acid residue with an alanine (Asp51Ala); the presence of a negatively charged residue at this position is conserved among mammals (Fig. S4
), possibly suggesting functional significance and awaiting experimental testing.
In analogy to IL1F5
, the other IL1 family member we identified as rejecting neutral evolution, i.e., IL1F7
, also acts as an antiinflammatory molecule (64
, and IL1F10
are not known to be involved in human diseases; in contrast, IL18RAP
play a role in triggering immune responses. The IL-7/IL-7R ligand-receptor pair is central to the proliferation and survival of B and T leukocytes. We identified one SNP in the gene as highly correlated with macropathogen richness (). This is not surprising given the role of Th2 responses in helminth infection and the involvement of IL-7R in the TSLP signaling pathway (65
), which in turn regulates Th2-mediated inflammatory responses (66
Similar to IL7R
plays a known role in human pathology. The gene encodes a component of the protein complex involved in transducing IL-18 signal, resulting in the activation of NF-κB (67
). The IL-18 receptor complex is expressed in the intestine (38
), and one SNP immediately downstream IL18RAP
(rs917997) has been associated with both CeD and IBD (38
). We found the predisposing allele of rs917997 and a linked variant (rs2272128) to correlate with pathogen richness. The location of rs2272128 in the 5′ gene region and its strong correlation with pathogens might suggest that it (rather than rs917997) represents or is in close LD with the functional allele. Moreover, the correlation of a risk allele for autoimmune diseases with pathogen-richness suggests an interesting link between adaptation and disease. Indeed, we observed that five more risk alleles for either IBD or CeD significantly correlate with micro- and macropathogen richness. Albeit preliminary, these data suggest that infectious agents have shaped the genetic variation at IL loci involved in intestinal inflammatory processes and, as a consequence, the genetic predisposition to both CeD and CD/IBD.
A north–south gradient for IBD prevalence has been described in both the US and Europe (68
). This observation, together with the increase of IBD prevalence in the last 40 yr (68
) and the hypothesis that helminths elicit Th2-mediated responses, led to the proposal that lower exposure to parasitic worms in the setting of industrialized countries results in unbalanced immune response, and eventually predisposes to IBD (68
). The so-called hygiene hypothesis, which clearly implies evolutionary considerations concerning human–pathogen interactions, has been supported by recent studies in both humans and mice (40
). Data herein seem to indicate that a portion of CeD- and IBD-predisposing alleles have been selected by micropathogen richness, pointing to an adaptive role for these variants. Although not directly supportive of the “IBD hygiene hypothesis,” these results indicate a higher disease predisposition in subjects carrying IL SNP variants that confer stronger protection against viruses/bacteria and therefore likely elicit more vigorous Th1 responses. Living conditions in industrialized countries have resulted in a reduction of both helminth and bacterial/viral infection. The effect of this environmental change on the homeostasis of immune responses might be difficult to reconcile with simple theories (71
). In this complex scenario, we consider that evolutionary studies and population genetics approaches, such as the one proposed here, provide some insight into the genetic basis of predisposition to infectious and autoimmune diseases.