The development of functionally responsive T cells. 
The work reviewed in this article separates T cell development into four phases. First is an expansion phase prior to TCR rearrangement, which appears to be correlated with programming of at least some response genes for inducibility. This phase can occur to some extent outside of the thymus. However, the profound T cell deficit of nude mice indicates that the thymus is by far the most potent site for inducing the expansion per se, even if other sites can induce some response acquisition. Second is a controlled phase of TCR gene rearrangement. The details of the regulatory mechanism that selects particular loci for rearrangement are still not known. It seems that the rearrangement of the TCR gamma loci in the gamma delta lineage may not always take place at a developmental stage strictly equivalent to the rearrangement of TCR beta in the alpha beta lineage, and it is not clear just how early the two lineages diverge. In the TCR alpha beta lineage, however, the final gene rearrangement events are accompanied by rapid proliferation and an interruption in cellular response gene inducibility. The loss of conventional responsiveness is probably caused by alterations at the level of signaling, and may be a manifestation of the physiological state that is a precondition for selection. Third is the complex process of selection. Whereas peripheral T cells can undergo forms of positive selection (by antigen-driven clonal expansion) and negative selection (by abortive stimulation leading to anergy or death), neither is exactly the same phenomenon that occurs in the thymic cortex. Negative selection in the cortex appears to be a suicidal inversion of antigen responsiveness: instead of turning on IL-2 expression, the activated cell destroys its own chromatin. The genes that need to be induced for this response are not yet identified, but it is unquestionably a form of activation. It is interesting that in humans and rats, cortical thymocytes undergoing negative selection can still induce IL-2R alpha expression and even be rescued in vitro, if exogenous IL-2 is provided. Perhaps murine thymocytes are denied this form of rescue because they shut off IL-2R beta chain expression at an earlier stage or because they may be uncommonly Bcl-2 deficient (cf. Sentman et al., 1991; Strasser et al., 1991). Even so, medullary thymocytes remain at least partially susceptible to negative selection even as they continue to mature .