WASHINGTON, DC—Genomics researchers are steadily completing the puzzle of how genes contribute to the risk for rheumatoid arthritis (RA). Peter K. Gregersen, MD, from the Feinstein Institute, North Shore Long Island Jewish Health System, in Manhasset, New York, discussed recent breakthroughs in a special translational research session titled "Translating Risk of Disease from Genetic Elements" at the American College of Rheumatology 2006 Annual Meeting.¹

"We are mostly looking for genes associated with a relative risk (RR) of 1.5 to 2.0. Some have asked whether it is worth looking for genes that convey such a modest risk, but they can help in identification of disease pathways. In addition, gene-gene and gene-environment interactions can be the basis for stratifying patients in research trials, which may make it easier to identify environmental factors in rheumatoid diseases," Dr. Gregersen said.

SNP studies find 5 RA-associated gene polymorphisms

The current paradigm for how RA develops in susceptible individuals is that there is a susceptibility stage, followed by a preclinical stage, and finally clinically apparent disease; genetic elements may come into play at any of these stages. The well-recognized aggregation of autoimmune (AI) diseases in families, for example, means an increased RR of 15 to 20 for individuals in such families.

"Understanding genetics depends on the definition of [disease] phenotype," Dr. Gregersen said.  For example, RA in a patient who is seropositive for anti-CCP autoantibodies is different from anti-CCP-negative RA. The influence of the HLA shared epitope (SE) affects only anti-CCP-positive disease, according to Dr. Gregersen.

Using random forest analysis (which improves the prediction accuracy of classification trees by building a collection of trees), Dr. Gregersen and colleagues analyzed whole-genome screen data including 381 microsatellite markers for 583 Caucasian, RA-affected sibling pairs from the North American Rheumatoid Arthritis Consortium (NARAC). They looked for non-major histocompatability complex (non-MHC) loci associated with RA. They found several non-MHC regions that were predictive of age of onset, CCP status, and RF status, as well as gene-gene interactions and regions predictive of SE status. They concluded that multiple non-MHC loci "may be influencing clinical and HLA-associated susceptibility for RA, and with the exception of CCP status, these relationships are independent of gender." ²

High-density single nucleotide polymorphism (SNP) analysis of NARAC data also identified three subregions within chromosome 18q in which there were "clusters of two or more SNPs with significant allele frequency differences in cases compared with controls (P <.005)." ³  Further analysis revealed five SNPs that were significantly associated with RA. This suggests that this region contains one or more variants of novel genes that influence susceptibility to RA, according to Dr. Gregersen. RA-related SNPs have also been identified on chromosome 11.⁴

Data from the NARAC study also showed an association between PTPN22 and various AI diseases, and the PTPN22 R620W allele appears to be a "gain of function" variant. "This might be associated with a higher threshold for thymic deletion of T cells, leading to release of self-reactive T cells into the periphery," Dr. Gregersen said.

A number of common genes are associated with multiple AI phenotypes, and Dr. Gregersen listed eight genes with confirmed associations with RA. "On the same chromosome, some genes predispose to RA but one reduces anti-CCP levels," Dr. Gregersen said.

However, "HLA-DR4 is only part of the story," he warned. "You cannot attribute the entire HLA effect to DR4."

Important to control for ancestry in genetic studies

More recently Dr. Gregersen's group has been using "tagging" SNPs to detect entire haplotype blocks, using Illumina Whole-Genome SNP arrays. He said that 1 year ago he was excited to be receiving 16 million genotypes in 6 months, but this year he is receiving 100 million genotypes per week.

This type of analysis also identifies ancestral chromosomes with common segments resulting from residual relatedness in modern populations. Dr. Gregersen noted that the Plenge study of RA-linked genes in subjects of European ancestry highlighted the importance of matching cases and controls when seeking genetic factors in AI diseases. "Populations in some of the Nordic countries are relatively homogenous, but in New York, people are from everywhere," he noted. Controlling for this is critical, since some genetic associations are affected by northern European vs southern European ancestry. He recommended doing marker studies first, then stratifying the study population in order to match cases and controls. "Ancestry informative markers are an important tool, especially for US studies," he said.

References

1. Gregersen PK. Translating risk of disease from genetic elements. Presented at: American College of Rheumatology Meeting; November 11, 2006; Washington, DC.
2. Barcellos LF, Ransay PP, Madden E, et al. Random Forest analysis: a novel approach for exploring the complex genetic component in rheumatoid arthritis. Presented at: American College of Rheumatology Meeting; November 11, 2006; Washington, DC. Abstract 189.
3. Remmers EF. Le JM, Li W, et al. High-density SNP-typing identifies two regions on chromosome 18q that are associated with susceptibility to rheumatoid arthritis. Presented at: American College of Rheumatology Meeting, November 11, 2006; Washington, DC. Abstract 895.
4. Plenge RM, Lee AT, Li W, et al. Dense SNP genotyping under 5 linkage peaks in rheumatoid arthritis. Presented at: American College of Rheumatology Meeting, November 11, 2006; Washington, DC. Abstract 917.