Background/Purpose: In the treatment of rheumatoid arthritis (RA), molecular targeted therapies induced different changes in different immune cell phenotypes. For instance, our previous study showed that abatacept decreased the number of Tfh cells, while TNF inhibitors increased that of Tfh cells. This fact forms the basis for subgroup identification for precision medicine. In this study, we stratified RA patients based on immunophenotyping and investigated the response for targeted therapies.

Methods: Peripheral blood mononuclear cells were obtained from 224 bio-naive RA patients and 33 healthy individuals. Circulating B, T and dendritic cells were defined based on flow cytometric analysis for human immune system termed "the Human Immunology Project". Based on these results, RA patients were classified into subgroups by cluster analysis. The human ethics review committee of our university reviewed and approved this study.

Results: The proportions of T effector and activated Tfh cells, but not Th17 and Treg cells, were higher in RA than the age matched control. Likewise, the proportions of B effector cells and plasmacytoid DCs were higher. None of them correlated with disease activity. Cluster analysis (Figure1) stratified RA patients into 2 groups: patients with less immune abnormality and patients with immune abnormality. The group with immune abnormality was further divided statistically into 3 groups (with high proportions of B effector, Th17, and Tfh cells in all groups): patients with less T cell abnormality (T cell low dependent group), patients with high percentage of Tfh and Treg cells (Tfh/Treg dominant group), and patients with conspicuously high proportion of T effector cells (T effector dominant group). Among these 4 groups, simplified disease activity index (SDAI) and clinical disease activity index (CDAI) components were similar. However, the treatment responses by biologics were different. The number of patients with insufficient response at week 52 was more frequent in T effector dominant group. In addition, TNF inhibitors were numerically more effective in the group with less immune abnormality. In contrast, abatacept had similar efficacies in the group with less immune abnormality and the group with immune abnormality, and tocilizumab had strong efficacy in Tfh/Treg dominant group.

Conclusion: Our study indicates that active RA patients can be divided into four subgroups based on immunophenotype. The identification of immunophenotypic subgroups may enhance treatment effect of molecular targeted drugs and serve as a step towards precision medicine.