The DARC side of GWAS

In this issue of Blood, Schnabel and colleagues show that genetic variants in a genomic region harboring the DARC are associated with circulating serum levels of the inflammatory protein, MCP-1.

In the enthusiasm generated with the first draft sequence of the human genome, many predicted that human genomics would take us places never before imagined. Space-age images were invoked to portray future advances in knowledge as well as overly optimistic application to risk assessment in clinical and public health venues.

One of the first outputs is a catalog of new regions associated with complex diseases and traits. The age of genome-wide association studies (GWAS) has ushered in an era of remarkable discovery, which some have suggested has been unprecedented in modern biology.^1,2  In the past 5 years, the GWAS approach has already discovered more than 650 regions that have been mapped to 125 complex diseases and traits (http://www.genome.gov/gwastudies).³  In nearly every instance, the findings have uncovered genetic markers that merit extensive follow-up analyses.^4,5  It is rare that the “smoking gun” is immediately apparent. Instead, GWAS findings point to “new candidate regions” with many harboring unfamiliar genes or no genes at all. In a minority of cases, the markers point to known genes not previously implicated in the disease or trait of interest.

Schnabel et al report a set of studies that together point toward a region of chromosome 1 harboring Duffy antigen receptor for chemokines (DARC) that accounts for a proportion of the genetic variance observed with circulating serum MCP-1 levels.⁶  They began with a large genome-wide association study of unrelated subjects, one that has combined set of scans, that some refer to as “supersized” GWAS. Follow-up in a family-based analysis (generally preferred by classical geneticists) confirmed the association, giving it even more traction. Laboratory analysis was undertaken to determine the molecular basis of the difference in the association results for serum versus plasma levels. Together, the results provide conclusive evidence of the relationship between genetic markers in the DARC and circulating serum levels of MCP-1.

The investigators were fortunate to land upon a coding variant DARC, Asp42Gly, that had been previously shown to discriminate between 2 blood group antigens, F_y^b (as-paratate) and F_b^a (glycine).⁷  The authors provide insight into why this is apparent in serum but not in plasma, illustrating the importance of fibrinogen on the measured stability of MCP-1.

However, the story is not complete. More work is needed to investigate the mechanism of how the coding variant directly or indirectly alters circulating levels of MCP-1. Moreover, they showed that clotting and heparan sulfate (unfractionated haparin) can release MCP-1 as well, but it is not clear that the genetic variant Arg42Gly alone modulates the differential cytokine binding. Mapping studies of the region clearly are needed to identify the set of common and uncommon variants that should be investigated further. Although it is easiest to interpret a change in amino acid sequence, it is not clear how such a shift influences circulating serum levels. Without the parallel genetic work to rule in or out additional markers for laboratory evaluation, the results reported in this paper remain promising.

The association of DARC variants with MCP-1 levels would not have been high on most investigators' list of likely associations, but now that it has been established, it opens a new universe for research. In this regard, the agnostic approach of GWAS is hypothesis-free, something that is uncomfortable to some investigators, but nonetheless provides strong statistical grounds for determining whether variants are conclusively associated with a phenotype (such as a disease or trait).

Now, why is it important to investigate genetic variants that influence serum levels of MCP-1, a protein that has been studied in blood disorders, cancers, and inflammatory conditions? The authors suggest that it could be of clinical value to assess serum MCP-1 levels in the context of genetic differences. Although this is theoretically appealing, it is still in another galaxy to apply the notable variants to clinical decision-making. In the short term, the discovery of a genomic region that harbors a suitable candidate gene is notable in that it uncovers not only a possible gene, but a pathway or process that could be important for targeting in therapy or prevention.¹  To better apply the current findings of Schnabel et al to risk models for clinical purposes, the optimal markers will need to be tested in well-designed studies, not those that are conveniently available.

One of the sobering lessons of GWAS is that the discovery of new variants still requires extensive follow-up. For GWAS, the bar for the heralded statistical threshold, genome-wide significance, is high to guard against false positives.⁴  Still, it is not yet clear what will be the metric for declaring an association fully explained. So, we know we will have to explore further and dig deeper into the biology of the DARC side of MCP-1.