Recombinant human bone morphogenetic protein (rhBMP) is a promising therapeutic cytokine for the induction of bone formation, but a weak response in humans remains a major hurdle in its therapeutic application. We have previously reported an rhBMP-2-induced increase in the bone mass of mice receiving systemic rolipram, a specific inhibitor of phosphodiesterase-4. To overcome the side effects of systemic administration of rolipram, we examined the effects of its local release. Polyethylene glycol discs were used as a delivery system. The discs were impregnated with rhBMP-2 and rolipram and implanted into the dorsal muscle pouches in mice. Bone formation was assessed by measuring the bone mineral content (BMC) of the formed bone. First, to determine the optimal dose of rolipram, we added 0-5000 nmol rolipram and 5 microg rhBMP-2 to the pellets and found that 500 nmol rolipram was the most effective concentration for inducing bone formation after 4 weeks. Second, to examine the time course of bone formation, we implanted 5 microg rhBMP-2 with 0 or 500 nmol rolipram and killed mice 5, 7, 10, 14, or 21 days after implantation. Bone formation was accelerated in the rolipram group. Finally, to determine the rolipram-induced increase in the effect of BMP, BMC obtained after treatment with 5 microg rhBMP-2 and 500 nmol rolipram was compared with that obtained after treatment with 5-9 microg rhBMP-2 without rolipram, 4 weeks after implantation. The results indicated that 500 nmol rolipram enhanced the effect of rhBMP-2 by almost 1.5-fold. In summary, locally released rolipram enhanced the capacity of rhBMP-2 to induce bone formation, an effect previously reported with systemic administration. These findings may decrease the cost and increase the efficacy of rhBMP-2 treatment.