Longcan Li, Xin Chen, Shengpeng Wang, Zhengle Li, Xingwen Xie
Rheumatoid arthritis is sustained by a synovial cytokine network that acquires the capacity to maintain inflammation independently of its initiating stimulus. The biologic and targeted synthetic antirheumatic drugs developed over the past two decades each interrupt this network at a defined molecular node, and their differential effects in patients therefore reveal how the network is organized. This review is structured around that principle, treating each agent as a probe of the axis it targets rather than as a therapeutic option. Following an account of the synovial signaling nodes on which diverse cytokines converge, four axes are examined: TNF-NF-κB, where receptor-level bifurcation between inflammatory and regulatory signaling defines the class; IL-6 and JAK-STAT, where classic and trans-signaling separate pathogenic from homeostatic function; costimulatory and lymphocyte-directed signaling, targeted by CTLA4-Ig and B-cell depletion; and the effector axes of tissue destruction driven by RANKL and GM-CSF. A recurring theme is that persistent synovitis reflects failure of the circuits that normally terminate cytokine signaling, and that fibroblast programs, once epigenetically imprinted, become progressively independent of the cytokines that established them. Synovial pathotypes correspond to differences in which axis is transcriptionally dominant, providing a biological rationale for stratified therapy; the same framework identifies a fibroblast-dominant state refractory to every current agent. Whether axis dominance is fixed or shifts under treatment remains the central unresolved question determining whether mechanism-guided therapy is clinically achievable.