DURING UNTREATED HIV infection, naïve CD4+ T cells, and eventually CD8+ cells, are depleted more rapidly than memory cells. An impaired supply of T cells from a functional thymus has been proposed as a possible mechanism for this depletion. Recent thymic emigrants (RTE) carry a traceable molecular marker, a T cell receptor excision circle (TREC). TREC content in a cell population is affected both by changes in thymic output and by the division history of the cells1. This is because TRECs do not replicate during mitosis. Indeed, memory T-cell populations contain 10 to 100-fold fewer TRECs than naïve populations. This has been taken into account in evaluating TREC data1. However, it has been believed that proliferative dilution of TRECs among naïve cells was slow and most importantly, constant. In this issue, Hazenberg et al2. present data to suggest that this assumption may be wrong. Hazenberg et al. analyzed the relative impact of changes in thymic output and in naïve T-cell division rate on TREC content2. From equations describing how numbers of naïve T cells (N), and of TRECs (T) change with time, the steady-state TREC content, T/N, was derived. It is written as c/(1+ D), where c is T/N in pure RTE and D a dilution factor. D is essentially Nα/σ where α is the per-cell division rate and σ is the immigration rate of RTE. Thus, dilution depends on the ratio between the production rate of non-TREC containing cells through cell division and the thymic output of TREC containing cells. This analysis has important implications for naïve T-cell homeostasis. In the absence of cell division, TREC content does not change even when σ does. Because naïve T-cell TRECs decline as people age, division must contribute to maintenance of the population. Furthermore, N is not just passively adjusted to a gradually declining thymic output, because then N would remain proportional to σ and T/N would not change. This argues for active homeostasis, whereby division rate/cell increases and/or death rate diminishes as N decreases. The authors’ simulations favor increased division rate, in agreement with recent experiments showing accelerated naïve T-cell division in lymphopenic mice3. TREC content of blood cells is reduced within a few months after infection1. Hazenberg et al. show that such change cannot result from a sharp decline in thymic output, even if such decline did occur, because of the low death rate of naïve cells2. By contrast, a several-fold increase in division rate of naïve cells can lead to a rapid fall in TREC content. If such an increase occurs, the reductions in TREC content during HIV infection cannot be taken as evidence for diminished thymic function.

Naïve T cells express both CD27 and the RA isoform of CD45, while most memory cells are believed to express the RO isoform. Hazenberg et al. measured TREC content of CD45RA+CD4+ and CD45RA+CD8+ cells in HIV-infected individuals and controls and also the frequency of naïve (CD45ROCD27+) cells that expressed Ki67, a cellcycle marker2. They report reduced TREC