Adult naive T cells, which are at rest in normal conditions, proliferate strongly when transferred to lymphopenic hosts. In neonates, the first mature thymocytes to migrate to the periphery reach a compartment devoid of preexisting T cells. We have extensively analyzed the proliferation rate and phenotype of peripheral T cells from normal C57BL͞6 and T cell antigen receptor transgenic mice as a function of age. We show that, like adult naive T cells transferred to lymphopenic mice, neonatal naive T cells proliferate strongly. By using bone-marrow transfer and thymicgraft models, we demonstrate that the proliferation of the first thymic emigrants reaching the periphery requires T cell antigen receptor-self-peptide͞self-MHC interactions and is regulated by the size of the peripheral T cell pool. P eripheral T cells expand strongly after transfer to lymphopenic hosts (1, 2). This proliferation has been considered as a homeostatic mechanism that fills the peripheral T cell pool. It was initially suggested that homeostatic T cell proliferation was restricted to activated͞memory T cells. Indeed, most authors inferred that naive and memory T cell pools would be regulated independently. The memory T cell pool would be renewed and filled by proliferation of preexisting activated͞memory cells, whereas only continuous output of naive T cells by the thymus could generate a full compartment of truly naive T cells (3). This theory is supported by the observation that a significant proportion of memory T cells proliferate in normal mice, whereas naive T lymphocytes do not cycle in normal conditions (4-8).Nevertheless, using monoclonal naive T cells from T cell antigen receptor (TCR) transgenic mouse strains and purified polyclonal naive T cells from normal mice, it has recently been shown that naive T cells proliferate when transferred to lymphopenic hosts (8-21). Such peripheral T cell expansion would be depend on self-peptide-specific, low-affinity interactions similar to those required for positive selection in the thymus (10-12), although other reports do not support this view (13). By following the long-term fate of naive T cells transferred to lymphopenic hosts (rag-deficient, CD3 -deficient, and irradiated normal mice), we and others have observed that the transferred cells fail to fill the peripheral naive T cell pool. Indeed, absolute numbers of recovered T cells are far below those found in the full peripheral naive T cell pool of a normal mouse. Moreover, injected naive T cells acquire a memorylike phenotype that remains stable with time, despite the absence of foreign antigenic stimulation, and their functional capacities are modified, enhanced, or abolished (18-21). Recent studies have demonstrated a role of IL-7, IL-12, and p56 lck in the proliferation of naive T cells in lymphopenic mice (22)(23)(24)(25), but the mechanisms underlying this process are not fully understood.Furthermore, it is crucial to show the physiological relevance of this mechanism. Indeed, injection of purified naive T cells to irradiated or...

Daily lymphocyte production in both central and peripheral lymphoid organs was evaluated by associating in vivo incorporation of bromodeoxyuridine (BrdUrd) with cell surface labeling and multi-parameter flow analysis. At least 10% of mature T and B lymphocytes are generated every 24 h. The kinetic behavior of these cell populations differs, however, in that mature B cells are generated predominantly in the precursor compartments of the bone marrow, while most mature T cell generation occurs at the periphery. Therefore, peripheral expansion is the major mechanism of mature T cell production in the adult mouse. By following the accumulation of BrdUrd-labeled cells in peripheral lymphoid organs we found that the progeny of the daily lymphocyte production was sufficient to renew 30%-40% of all peripheral T and B cells every 48 h, demonstrating a high turnover rate of mature lymphocytes. We also examined the conditions of BrdUrd labeling of cycling cells in vivo. We found that while greater than 90% of bone marrow and thymus cells in S phase were labeled with a single injection of BrdUrd, in peripheral lymphoid compartments 70% of T and B cells in S failed to incorporate BrdUrd. Particular schedules of BrdUrd administration were required to overcome the low labeling efficiency of mature cells in vivo. Prolonged BrdUrd administration, however, had toxic effects on resident cells. The low labeling efficiency of BrdUrd incorporation by mature cells, as well as its potential toxicity during prolonged administration, may explain controversial results obtained by the different strategies used to study lymphocyte population dynamics.