Tetracyclines exert, independently of their antimicrobial activity, anti-collagenolytic effects by inhibiting activities of human interstitial collagenases and by preventing the oxidative activation of latent pro-collagenases. We tested the clinical response to a 3-month doxycycline in concert with collagenase activity in 12 rheumatoid arthritis (RA) patients. Patients received 150 mg/day of doxycycline for 3 months. Clinical assessments at zero, six and 12 weeks comprised classification of the functional class, joint score index, Hb, CRP, ESR, health assessment questionnaire, visual analogue scale (VAS) of pain, pain disability index, comprehensible psychopathological rating scale (CPRS), SDS-PAGE laser densitometric collagenase activity measurements and Western blots. Significant reductions were seen in joint score index (P < 0.01), pain VAS (P < 0.05) and some CPRS parameters. Furthermore, collagenase activities measured from saliva by quantitative SDS-PAGE electrophoresis were significantly reduced during the 12-week intervention (P < 0.01). Western blots demonstrated intact 75-80 kDa enzyme protein (classic neutrophil collagenase), but also a newly discovered mesenchymal, less glycosylated 40-55 kDa MMP-8 subtype of fibroblast/chondrocytic origin. These results indicate that the documented favourable clinical response may in part be due to in vivo inhibition of classic neutrophil and mesenchymal collagenase/MMP-8 activities produced by doxycycline. This anti-collagenolytic doxycycline effects is mediated through inhibition of the enzyme activity and not through degradation of the enzyme, which may have contributed to the reportedly reduced tissue destruction, as has been seen in clinical studies concerning RA as well as reactive arthritis.