Neurotrophin-3 (NT-3) promotes enteric neuronal development in vitro; nevertheless, an enteric nervous system (ENS) is present in mice lacking NT-3 or TrkC. We thus analyzed the physiological significance of NT-3 in ENS development. Subsets of neurons developing in vitro in response to NT-3 became NT-3 dependent; NT-3 withdrawal led to apoptosis, selectively in TrkC-expressing neurons. Antibodies to NT-3, which blocked the developmental response of enteric crest-derived cells to exogenous NT-3, did not inhibit neuronal development in cultures of isolated crest-derived cells but did so in mixed cultures of crest- and non-neural crest-derived cells; therefore, the endogenous NT-3 that supports enteric neuronal development is probably obtained from noncrest-derived mesenchymal cells. In mature animals, retrograde transport of (125)I-NT-3, injected into the mucosa, labeled neurons in ganglia of the submucosal but not myenteric plexus; injections of (125)I-NT-3 into myenteric ganglia, the tertiary plexus, and muscle, labeled neurons in underlying submucosal and distant myenteric ganglia. The labeling pattern suggests that NT-3-dependent submucosal neurons may be intrinsic primary afferent and/or secretomotor, whereas NT-3-dependent myenteric neurons innervate other myenteric ganglia and/or the longitudinal muscle. Myenteric neurons were increased in number and size in transgenic mice that overexpress NT-3 directed to myenteric ganglia by the promoter for dopamine beta-hydroxylase. The numbers of neurons were regionally reduced in both plexuses in mice lacking NT-3 or TrkC. A neuropoietic cytokine (CNTF) interacted with NT-3 in vitro, and if applied sequentially, compensated for NT-3 withdrawal. These observations indicate that NT-3 is required for the normal development of the ENS.