Blocking the binding of TNF-alpha to TNF receptor subtype-1 (TNFR1) is an important strategy for the treatment of rheumatoid arthritis (RA). We recently succeeded in developing a TNFR1-selective antagonistic TNF mutant, R1antTNF. Here, we report the anti-inflammatory effects of R1antTNF in a murine collagen-induced arthritis model. To improve the in vivo stability of R1antTNF, we first engineered PEG (polyethylene glycol)-modified R1antTNF (PEG-R1antTNF). In prophylactic protocols, PEG-R1antTNF clearly improved the incidence, and the clinical score of arthritis due to its long plasma half-life. Although, the effect of PEG-R1antTNF on the incidence and production of IL1-beta was less than that of the existing TNF-blocking drug Etanercept, its effect on severity was almost as marked as Etanercept. Interestingly, in therapeutic protocols, PEG-R1antTNF showed greater therapeutic effect than Etanercept. These data suggest that the anti-inflammatory effects of PEG-R1antTNF depend on the stage of arthritis. Recently, there has been much concern over the reactivation of viral infection caused by TNF blockade. Unlike Etanercept, PEG-R1antTNF did not reactivate viral infection. Together, these results indicate that selective inhibition of TNF/TNFR1 could be effective in treating RA and that PEG-R1antTNF could serve as a promising anti-inflammatory drug for this purpose.