W e have determined previously that IGF-I is dependent on the presence of IGF binding protein-1 (IGFBP-1) to act as a wound healing agent. We sought to determine the mechanism whereby IGFBP-1 is able to enhance IGF-I bioactivity. As IGFBP-1 binds both the alpha5beta1 integrin as well as IGF-I in vitro, we asked which of the following interactions were important: (a) the ability of IGFBP-1 to interact with an integrin receptor, and/or (b) the binding of IGF-I by IGFBP-1. We used an IGF-1 analogue (des(1-3)IGF-I) with a > 100-fold reduction in affinity for IGFBP-1 as well as an IGFBP-1 mutant (WGD-IGFBP-1) which does not associate with the alpha5beta1 integrin to selectively abrogate each of these interactions. We also tested the ability of IGFBP-2, a related binding protein which has an arginine-glycine-aspartate sequence but does not associate with integrin family members, to enhance IGF-I bioactivity. Full-thickness dermal wounds were created on rabbit ears; various combinations of native IGF-I, native IGFBP-1, native IGFBP-2, and their respective analogues/mutants were applied to each wound. Wounds were harvested 7 d later for analysis. Only native IGF-I in combination with native IGFBP-1 was effective as a wound healing agent, enhancing reepithelialization and granulation tissue deposition by 64+/-5 and 83+/-12% over controls (P = 0.008 and 0.016, respectively). The same doses of IGF-I/WGD-IGFBP-1, des(1-3)IGF-I/IGFBP-1, and IGF-I/IGFBP-2 were ineffective. We propose that IGF-I physically interacts with IGFBP-1 and that IGFBP-1 also binds to an integrin receptor, most likely the alpha5beta1 integrin. This interaction is unique to IGFBP-1 as the closely related IGFBP-2 had no effect, a finding consistent with its inability to bind to integrin receptors. Our results suggest that activation of both the IGF-I receptor and the alpha5beta1 integrin is required for IGF-I to stimulate wound healing.