Shared pathways to infectious disease susceptibility?

The recent advent of genomic approaches for association testing is starting to enable a more comprehensive understanding of the role of human immune response in determining infectious disease outcomes. Progressing from traditional linkage approaches using microsatellite markers to high-resolution genome-wide association scans, these new approaches are leading to the robust discovery of a large number of disease susceptibility genes and the beginnings of an appreciation of their connections. In this commentary, we discuss how this technology development has led to increasingly complex and common infectious diseases being unraveled, and how this is starting to dissect pathogen-specific human responses. Intriguingly, these still preliminary findings suggest that pathogen innate detection mechanisms may not be as shared among diseases as immune response mechanisms.


Severe combined immune deficiency and partial immune deficiency
The first studies of immune response defects and infec tious disease outcomes strongly suggested that infectious diseases would have shared pathways to disease. Severe combined immune deficiency (SCID) a genetic disorder that affects the adaptive immune system demonstrates the importance of the immune system in protection against infection, as patients with SCID quickly succumb to lifethreatening sepsis from a very wide range of infectious agents. The gene responsible for the majority of SCID cases, CD132 (the common γchain), was identified using the traditional linkage approach [13]. Because the common γchain is shared by many cytokine receptors (receptors for IL2, IL4, IL7, IL9, IL15, and IL21), mutations that result in a nonfunctional common γchain cause widespread defects in IL signaling and failure of B and T cells to develop and mature. This is a rather profound example, and knowledge of the respon sible gene has opened avenues for therapeutic inter ven tion in cases of Xlinked SCID, where gene transfer of the normal CD132 appears to be a promising, even curative, treatment modality [4,5]. Similar observations and clinical conclusions have been made with another form of SCID caused by defects in the variable, diversity, joining recombination process [6].
Partial immune deficiency, referring to the inactivation of a single immune pathway due to genetic mutations in one or more members of the network, was initially suspected to act in a similar way to the SCID mutations, with patients having a predisposition to a wide variety of infections. As these mutations are highly penetrant, traditional genetic linkage studies have been able to isolate the responsible genes and pathways for some of

Abstract
The recent advent of genomic approaches for association testing is starting to enable a more comprehensive understanding of the role of human immune response in determining infectious disease outcomes. Progressing from traditional linkage approaches using microsatellite markers to highresolution genome-wide association scans, these new approaches are leading to the robust discovery of a large number of disease susceptibility genes and the beginnings of an appreciation of their connections. In this commentary, we discuss how this technology development has led to increasingly complex and common infectious diseases being unraveled, and how this is starting to dissect pathogen-specific human responses. Intriguingly, these still preliminary findings suggest that pathogen innate detection mechanisms may not be as shared among diseases as immune response mechanisms. these conditions, many of which have severe or even fatal clinical presentations. Notable examples include the IFNγ receptor (IFNGR1) in cases of severe mycobacterial infections [7], the lysosomal trafficking regulator (LYST) for ChediakHigashi syndrome [8], and Bruton's tyrosine kinase for Xlinked agammaglobulinemia [9,10]. While the immune system has multiple pathways, each is thought to be active against a wide range of micro organisms. Thus, it is perhaps surprising that the spectrum of infection susceptibility for these three genes appears to be fairly constrained. For example, IFNGR1 mutations mainly predispose to mycobacterial infections, and LYST and Bruton's tyrosine kinase mutations mainly predispose to predominantly pyogenic bacterial infec tions (for example, Streptococcus pyogenes, pneumo coccus, and Haemophilus influenzae). This surprising finding echoes another example: that of complement deficiency and its observed confinement to increased susceptibility towards encapsulated bacteria, particularly Neisseria spp., despite consistent documentation on the importance of the complement pathway in common blood pathogen clearance [11,12].

Subtle changes to immune genes as a result of genetic polymorphism
In the last 5 years, advances in genomic research have enabled the use of the GWAS design, permitting the robust identification of host susceptibility genes for infectious pathogens in increased detail. Although the advent of GWAS in infectious diseases has been slow compared with GWAS in many other complex disease phenotypes [13,14], this approach has nonetheless allowed unprecedented identification (or confirmation) of genes associated with susceptibility to the infectious diseases studied so far [1518].
Infectious disease genes identified via the GWAS approach have thus far been relatively 'pathogen specific' , with crosspathogen sharing being the exception rather than the rule.
The greatest surprise has related to observations within the human leukocyte antigen (HLA) locus, where broad based sharing between pathogen groups could be ex pected, given the central role played by HLA in immune recognition of pathogens and the initiation of immune responses. However, recent GWAS have found that while different HLA alleles appear to be shared between various autoimmune phenotypes (such as psoriasis [19], vitiligo [20], or systemic lupus erythematosus [21]), there appears to be little evidence of substantial susceptibility sharing for HLA between different infectious diseases. This disparity is striking even among supposedly 'similar' pathogen groups: a very strong contribution of the HLA locus has been observed for leprosy (Mycobacterium leprae) [15,22], but not for tuberculosis (Mycobacterium tuberculosis) susceptibility [23,24], highlighting the vast complexity of the immune response to diverse pathogens. This complexity is further shown in two genetic studies of vaccine response: while the HLA is crucial for success (or otherwise) for hepatitis B virus vaccination [25], it is not the case for H. influenzae type b vaccination [26].
Although findings thus far have suggested that the main pathways underlying susceptibility to pathogenic infection are pathogen specific, it is nonetheless relevant to understand similarities in the infection and invasion process for the very diverse pathogens that the human body encounters daily. For example, knowledge gleaned from the rare primary immune deficiencies has shown that defects in extremely central components of the immune system result in broad, general susceptibility to multiple infectious diseases. Elucidation of the respon sible genes and pathways often yields considerable insight into disease mechanisms and pathogenesis, and can inform medical treatment (a good example being the case of Xlinked agammaglobulinemia where common anti biotic treatments do not suffice, and immunoglobulin replacement is necessary). If some of these therapeutic findings could be extended to the commoner forms of infection that clinicians encounter on a daily basis, this could have an immediate public health impact.
Genomic approaches have played a supporting role in the identification of potential therapeutic targets. One of the virulent capabilities of Neisseria meningitidis is its ability to recruit human factor H (encoded by CFH) via molecular mimicry [27], thus avoiding complement mediated killing of the bacterium. A recently completed GWAS on host susceptibility to meningococcal disease has shown that genetic polymorphisms within CFH as well as CFHrelated genes are very strongly associated with susceptibility to meningococcal disease [18]. As the bacterial factor Hbinding protein is now a strong meningo coccal vaccine candidate [28], it is clear that genomic approaches can be used independently to discover or to confirm therapeutic targets identified by functional approaches.
Leprosy, tuberculosis, Crohn's disease, and ulcerative colitis have also been the subjects of recent investigations. Indeed, genomic approaches have underlined the impor tance of the nucleotidebinding oligomerization domain containing (NOD)2 signaling cascade in inflammatory and infectious diseases. Different members of the path way (NOD2, TNFSF15, and RIPK2) have been strongly implicated (P < 10 10 ) in susceptibility to leprosy [15,29], as well as Crohn's disease, which is an inflammatory bowel disease with substantial links to infectious agents [30,31]. However, observations by us and others [24] do not implicate NOD2 pathway members as strongly in tuberculosis susceptibility. The same has been observed for CARD9 (a member of the extended NOD family of genes), which has been strongly associated with ulcer ative colitis susceptibility [32] and functionally implicated in tuberculosis [33], but not with leprosy [15].
Another gene observed to be a strong (P < 10 10 ) suscep tibility factor for leprosy is TNFSF15 [15]. The exact same SNP (rs6478108) has also been found to be strongly associated (P < 10 10 ) with Crohn's disease [34]. However, the minor allele has been found to be associated with increased risk of leprosy, but with decreased risk of Crohn's disease. This suggests that differential regulation of the immune response appears to be critical to reach a balance between protection from invading pathogens and susceptibility to inflammatory disorders, thus highlight ing TNFSF15 as a potential therapeutic target in manipu lation of the strength of the immune response.
Indeed, the disparity in susceptibility genes for Crohn's disease and ulcerative colitis (different members of the pathogensensing CARD family are involved: NOD2 for Crohn's, and CARD9 for ulcerative colitis) is likely to reflect different sites of pathology, different bacterial spectra, and therefore different mechanisms for microbial recognition. Strikingly, there is evidence suggesting that the downstream, cytokinemediated pathogenesis path ways (for example, those involving IL23) could be common across both diseases [32].

Missing 'shared susceptibility'?
The majority of evidence collected so far from genomic approaches appears to suggest that, barring specific exceptions, common host susceptibility to different invad ing pathogens appears to be controlled by discrete genes and pathways. As genomic approaches are still limited in resolution, there is a possibility that truly shared mechanisms governing susceptibility across mul tiple pathogen groups have yet to be found. This seems to be likely in view of the finite number of pathogen recognizing receptors available at the first lines of defence, coupled with their convergent, downstream signaling pathways. An example for 'shared susceptibility' would be TIRAP, which encodes a common adaptor for two major pathogensensing receptors (TLR2 and TLR4); an inactivating mutation (S180L) within TIRAP appears to associate with susceptibility to invasive pneumococcus and H. influenzae, as well as Trypanosoma cruzi infec tions [26,35,36], with a yet unconfirmed association with tuberculosis [37,38].
Considering the evidence now available, there is reason to believe that host susceptibility to infection could be specific and 'mechanism dependent' in terms of pathogen recognition and the site of infection, with shared common signaling pathways playing an important role down stream of pathogen contact points.
A further benefit in the largescale deployment of hypothesisfree genomic approaches is that they could help in establishing the etiology of certain autoimmune diseases where infectious diseases have long been suspected but not confirmed to be a predisposing factor. For example, in the case of type 1 diabetes, GWAS have implicated an extraordinary number of immune related genes, some of which overlap with infection pathways, thus lending credence to the above hypothesis.
Genomic approaches have resulted in improved treatment strategies for patients with severe and partial immune deficiencies by identifying the deficient protein component(s), but for many of the common infections, due to their complex nature, genomic approaches have not yet been as fruitful in informing medical treatment. Emerging antibiotic resistance is putting additional pressure on research efforts in host and pathogen genomics for further improvement of therapeutic strategy beyond antimicrobial treatment, but there is hope that this research will lead to novel human proteinbased therapies.

Competing interests
The authors declare that they have no competing interests.

Authors' contributions
Both authors have contributed to the conceptualization and preparation of this manuscript and approved the final version.