Regeneration is widespread in nature but poorly understood. Regenerating animals are confronted with unique challenges that defy explanation—production of fully differentiated structures starting from an adult rather than embryonic context, integration of newly regenerated and previously existing tissues, and specification of the identity of regenerating tissues. What molecular and cellular processes do regenerating animals use to solve these problems?

We are addressing these problems by turning to an animal model that possesses almost unlimited regenerative abilities, the planarian Schmidtea mediterranea. Planarians are free-living, freshwater flatworms of the phylum Platyhelminthes that grow to between 1-30 mm and possess complex organ systems (see for example, an image of the worm's nervous system shown on this page). Remarkably, planarians are capable of regenerating any missing tissues after wounding. Regeneration in planarians requires a large pool of adult, possibly totipotent, stem cells termed neoblasts. The activity of neoblasts forms a regeneration blastema from which some missing tissues are derived, and their activity is also required for homeostatic maintenance of adult tissues in the absence of wounding. The completion of the S. mediterranea genome project, and the accomplishment of the first large-scale RNA interference (RNAi) screen in S. mediterranea in recent years have transformed this classical developmental system into an emerging model for the molecular study of regeneration and stem cells.

The specification of proper tissue pattern is a central mystery of regeneration. How do regenerating animals "know" what missing tissues to regenerate? As a paradigm for understanding how this process works, we are focusing on the dramatic decision of planarians to regenerate a head versus a tail after amputation. We have discovered that a planarian homolog of beta-catenin is required for this decision (1). Inhibition of beta-catenin-1 by RNAi causes animals to regenerate a head rather than a tail at posterior-facing wounds. Beta-catenin is the effector of the highly conserved canonical Wnt signal transduction pathway, which is used for diverse processes in development and disease and acts in many animals to pattern the anterioposterior axis running from head to tail (reviewed in (2)). A Wnt gene, wntP-1, is transcriptionally activated by wounding and is required for tail-versus-head regeneration in planarians. These observations indicate that unlike in development, regeneration requires specific wound-induced pathways for pattern control (3). Our ongoing studies seek to understand the mechanism that ensures this polarity of regeneration and how stem cell activity is controlled to fully restore a head-to-tail axis truncated by amputation.

Our ultimate goals are to understand how spatial cues signal to stem cells to replace damaged or entirely missing structures and what molecular processes allow robust pattern restoration after diverse injuries.