457 - PTPN22 and HLA-SE are not Associated with Discontinuation of TNFα Inhibitors or Methotrexate in a Large RA Cohort

Jenny E. Heller1, Sandeep K. Agarwal1, Lori B. Chibnik1, Nancy E. Maher1, Daniel H. Solomon1, Alex Parker2, Ronenn Roubenoff2, Robert M. Plenge1, Michael E. Weinblatt1, Nancy A. Shadick1. 1Brigham and Women's Hospital, Boston, MA; 2Millennium Pharmaceuticals, Cambridge, MA
Presentation Number: 457

PURPOSE: Multiple replication studies have demonstrated the PTPN22 single nucleotide polymorphism to be a predictor of rheumatoid arthritis (RA) risk and likely of RA severity. The HLA-DRB1 “shared epitope” alleles (HLA-SE) are established severity markers. It is not known, however, if these genes predict drug response. We aim to determine whether HLA-SE and PTPN22 are associated with the discontinuation of TNFα inhibitors and methotrexate (MTX).
METHODS: We enrolled RA patients in a registry that is collecting prospective genetic, demographic and functional status data. All patients were diagnosed by a rheumatologist according to ACR criteria. We assessed the HLA-DRB1 alleles by low-resolution genotyping and the PTPN22 missense SNP (rs2476601) by Sequenom genotyping. Genotypes were classified as single or double HLA-SE alleles and as PTPN22 TC or TT alleles. At enrollment and one year, we determined medication use through patient self-report and assessed the multi-dimensional health assessment questionnaire (MDHAQ).
RESULTS: At enrollment, 918 registry patients had a mean age (±SD) of 56.9 years (±14.1), a mean disease duration of 14.1 years (±12.4) and were 82% female.
A subset of 740 subjects had PTPN22 genotypes, with 25% (n=183) having the missense polymorphism. At enrollment, 36% (n=117) of those who reported current or previous TNFα inhibitor use had discontinued at least one. PTPN22 did not show any association with TNFα discontinuation in univariate or multivariate models adjusted for sex and disease duration (multivariate OR=1.15 95%CI 0.70-1.90). Further controlling for disease severity through MDHAQ did not affect the relationship.
In a subset of 758 subjects with HLA-SE genotypes, 64% (n=482) had at least one allele. Thirty-five percent (n=116) of those reporting TNFα use had discontinued at least one. The presence of at least one HLA-SE allele was not associated with TNFα inhibitor discontinuation (univariate OR=0.74, 95%CI 0.45-1.21). Adjusting for sex, disease duration, MDHAQ and MTX use in a multivariate model had no effect on the odds ratio (OR=0.74, 95%CI 0.43-1.27).
At enrollment, 641 patients reported current or previous use of MTX, and 256 patients (40%) had stopped MTX. PTPN22 and HLA-SE were not significantly associated with MTX discontinuation (PTPN22 OR=1.08 95%CI 0.70-1.64 and HLA-SE OR=1.10 95%CI 0.74-1.63; multivariate adjusted for sex, disease duration and MDHAQ).
Odds ratios remained constant with discontinuations at one year added.
CONCLUSION: In our large cohort, the disease susceptibility and severity markers PTPN22 and HLA-SE have no predictive value in determining discontinuation of TNFα inhibitors and MTX. These results indicate that a mechanism beyond the effect of increased disease severity may underlie inadequate drug response.

J.E. Heller, Millennium Pharmaceuticals, 2 Research grants.