ARTHRITIS & RHEUMATISM Vol. 43, No. 11, November 2000, pp 2391–2396 © 2000, American College of Rheumatology SYSTEMIC ANTI–TUMOR NECROSIS FACTOR THERAPY IN RHEUMATOID ARTHRITIS DOWN-REGULATES SYNOVIAL TUMOR NECROSIS FACTOR SYNTHESIS ANN-KRISTIN ULFGREN, ULF ANDERSSON, MARIANNE ENGSTRO ¨ M, LARS KLARESKOG, RAVINDER N. MAINI, and PETER C. TAYLOR Objective. To investigate the hypothesis that tu- mor necrosis factor (TNF) blockade in rheumatoid arthritis (RA) diminishes synovial synthesis of TNF, interleukin-1(IL-1), and IL-1. Methods. Patients with active RA received a single 10 mg/kg infusion of infliximab. Multiple synovial bi- opsy specimens were obtained from a knee the day before infusion and 14 days later. A modified immuno- histochemical method detecting cytokine-producing rather than cytokine-binding cells was applied to deter- mine synthesis of TNF, IL-1, and IL-1in fixed, cryopreserved sections. Computerized image analysis using two different methodologies was performed by independent observers blinded to the identity of samples. Results. All 8 patients met the American College of Rheumatology 20% improvement response criteria (ACR 20) at 2 weeks, and half of these patients met the ACR 50. With a few exceptions, there was concordance between both image analysis methodologies regarding the direction of change in immunopositive area fraction for all cytokines analyzed. TNFsynthesis was signifi- cantly reduced after treatment (P 0.05 at the Karo- linska Institute, Stockholm, Sweden; P 0.008 at the Kennedy Institute, London, UK). Patients meeting the ACR 50 were those with the highest baseline levels of TNFsynthesis. There was a significant correlation between baseline levels of TNFexpression and change in TNFlevels in response to therapy. Both IL-1and IL-1synthesis were reduced in 3 patients; IL-1 synthesis alone was reduced in 2 patients and IL-1 synthesis alone was reduced in 2 patients. In 1 patient, neither IL-1nor IL-1synthesis was reduced. Conclusion. Analysis of synovial tissue by means of immunomorphology and image analysis in a clinical trial setting may allow the drawing of biologically mean- ingful conclusions. Synovial TNFsynthesis was re- duced 2 weeks after infliximab treatment. Reductions in IL-1and IL-1synthesis were demonstrated in a subgroup of patients. High levels of synovial TNF production prior to treatment may predict responsive- ness to therapy. Clinical trials of therapy for rheumatoid arthritis (RA) based on tumor necrosis factor (TNF) neutral- ization using TNF-specific antibodies or soluble recep- tor fusion proteins demonstrate marked dose-dependent improvement in both clinical and laboratory measures of inflammation (1). Evidence to date indicates that there are at least two major mechanisms of action. First, there is diminished influx of inflammatory cells to the joints (2,3). The reduced cell margination (4) and migration are secondary to a down-regulation of both synovial endothelial adhesiveness (2) and chemotactic gradient (3). Second, there is deactivation of the proinflamma- tory cytokine cascade (5), as demonstrated by reductions in serum C-reactive protein and interleukin-6 (IL-6) concentrations. By means of serum studies, however, it has not been possible to test unequivocally the predic- tion, based on ex vivo studies (6), that TNFblockade will inhibit IL-1 and TNFproduction in vivo. IL-1 is either undetectable or present in very low concentrations in the peripheral blood of most patients. The Kennedy Institute is supported by a core grant from the Arthritis Research Campaign, UK, and holds a grant from Centocor Inc. for the investigation of mechanisms of action of anti-TNF therapy. Ann-Kristin Ulfgren, PhD, Ulf Andersson, MD, Marianne Engstro ¨m, BSc, Lars Klareskog, MD: Karolinska Hospital, Stockholm, Sweden; Ravinder N. Maini, FRCP, Peter C. Taylor, MA, PhD, FRCP: Kennedy Institute of Rheumatology, London, UK. Address reprint requests to Ann-Kristin Ulfgren, PhD, Rheu- matology Research Laboratory, CMM L8-004, Karolinska Hospital, S-171 76 Stockholm, Sweden. Submitted for publication December 31, 1999; accepted in revised form June 27, 2000. 2391