Normal wounds can heal by secondary intention (epidermal migration to cover a denuded surface) or by approximation of the wound edges (e.g., suturing). In healing by secondary intention, epidermis-derived MMPs are important. Keratinocyte migration begins within 3-6 hr post injury, as basal cells detach from underlying basal lamina and encounter a dermal substratum rich in type I collagen. Cell contact with type I collagen in vitro stimulates collagenase-1 expression, which is mediated by the alpha 2 beta 1 integrin, the major keratinocyte collagen-binding receptor. Collagenase-1 activity alone is necessary and sufficient for keratinocyte migration over a collagen subsurface. Stromelysins-1 and -2 are also found in the epidermis of normal acute wounds. Stromelysin-2 co-localizes with collagenase-1 and may facilitate cell migration over non-collagenous matrices of the dermis. In contrast, stromelysin-1 is expressed by keratinocytes behind the migrating front and which remain on basal lamina, i.e., the proliferative cell population. Studies with stromelysin-1-deficient mice that suggest this MMP plays a role in keratinocyte detachment from underlying basement membrane to initiate cell migration. In chronic ulcers, MMP levels are markedly elevated, in contrast to their precise temporal and spatial expression in acute wounds. Both collagenase-1 and stromelysin-1 are found in fibroblasts underlying the nonhealing epithelium, and stromelysin-1 expression is especially prominent. Two key questions underlie the use of MMP inhibitors and wound healing: (1) will these agents impair normal reepithelialization in wounds that heal by secondary intention; and (2) can MMP inhibitors be effective therapy for chronic ulcers? The answer to neither is known. Batimastat and marimastat appear not to interfere with normal wound healing, but only in sutured surgical wounds, a situation in which MMP expression has practically no role. We also show the first example of an in vivo immune response, contact hypersensitivity, which is dependent upon MMP activity. Using gene-deficient mice, we demonstrate that stromylysin-1 (MMP-3) is required for sensitization, whereas gelatinase B (MMP-9) is required for timely resolution of the reaction to antigenic challenge.