Maintenance of T cell specification and differentiation requires recurrent notch receptor-ligand interactions

Abstract

Notch signaling has been shown to play a pivotal role in inducing T lineage commitment. However, T cell progenitors are known to retain other lineage potential long after the first point at which Notch signaling is required. Thus, additional requirements for Notch signals and the timing of these events relative to intrathymic differentiation remain unknown. Here, we address this issue by culturing subsets of CD4 CD8 double negative (DN) thymocytes on control stromal cells or stromal cells expressing Delta-like 1 (Dll1). All DN subsets were found to require Notch signals to differentiate into CD4+ CD8+ T cells. Using clonal analyses, we show that CD44+ CD25+ (DN2) cells, which appeared committed to the T cell lineage when cultured on Dll1-expressing stromal cells, nonetheless gave rise to natural killer cells with a progenitor frequency similar to that of CD44+ CD25- (DN1) thymocytes when Notch signaling was absent. These data, together with the observation that Dll1 is expressed on stromal cells throughout the thymic cortex, indicates that Notch receptor-ligand interactions are necessary for induction and maintenance of T cell lineage specification at both the DN1 and DN2 stages of T cell development, suggesting that the Notch-induced repression of the B cell fate is temporally separate from Notch-induced commitment to the T lineage.

Figure 1.

Developmental potential of early DN thymocytes cultured on OP9 bone marrow stromal cells in the presence (OP9-DL1) or absence (OP9-control) of Dll1 expression. (a) Day 14 fetal thymocytes were sorted for the following populations: DN1 (R1-gated region A, CD117⁺ CD44⁺ CD25⁻), DN2 (R1-gated region B, CD117⁺ CD44⁺ CD25⁺), and DN3 (R2-gated region C, CD117⁻ CD44⁻ CD25⁺). (b) Thymocyte populations, sorted as indicated, were cultured (1,000 cells/well) on OP9-control cells or OP9-DL1 cells for 7 d, and then analyzed for surface expression of the indicated lineage markers by flow cytometry. Total cellularity is indicated in parentheses below each label. *, samples that could not be accurately counted by hemocytometer, and for which cell counts were determined flow cytometrically. (c) Total number of DN2-derived NK cells generated from DN2 thymocytes cultured on OP9-control and OP9-DL1 stromal cell lines.

Figure 2.

Clonal analysis of T and/or NK cell progenitor potential from DN1 and DN2 thymocytes cultured on OP9 bone marrow stromal cells in the presence (OP9-DL1) or absence (OP9-control) of DL1 expression. (a) Single DN1 (CD117⁺ CD44⁺ CD25⁻) and DN2 (CD117⁺ CD44⁺ CD25⁺) thymocytes were sorted into 96-well plates containing OP9-DL1 cells, and then cultured for up to 14 d. Clonal populations were analyzed by flow cytometry to determine the developmental potential of the individually sorted cells (as shown in c). Wells that contained only CD3⁻ NK1.1⁺ NK cells were scored as NK restricted, wells that contained only CD3⁺ T cells were scored as T cell restricted, and wells that contained both CD3⁻ NK1.1⁺ NK cells and CD3⁺ T cells were scored as T/NK cell bipotent. 192 wells, each containing a single DN1 or DN2 cell, were analyzed for each population, and the total number of wells scoring positive for the presence of T, NK, or both T and NK cells are illustrated (left). The frequency of each progenitor type is also presented as a percentage of the total number of wells that successfully reconstituted (right). (b) A comparison of the NK cell progenitor frequency observed for DN1 and DN2 cells when cultured on either OP9-control (left) or OP9-DL1 cells (middle), and the T cell progenitor frequency observed for DN1 and DN2 cells when cultured on OP9-DL1 cells (right). The data shown were obtained from the limiting dilution analysis presented in Tables I and II. (c) Representative flow cytometric analysis of individual wells seeded with a single cell with NK cell (left), T cell (middle), and T/NK bipotent (right) potential.

Figure 3.

PCR analysis of DN2-derived NK cells for TCR-β gene rearrangements. DN2 (CD117⁺ CD44⁺ CD25⁺) thymocytes were cultured on OP9-control or OP9-DL1 cells for 14 d. DNA was purified from total DN2-derived T cells (from OP9-DL1 cultures) and from DN2-derived NK cells (from OP9-control cultures). The rearrangement status of the TCR-β gene locus was determined by PCR using Dβ2-Jβ2 specific primers. PCR reactions containing DNA from day 14 fetal thymus and fibroblast DNA were included as TCR-rearranged and germline controls, respectively.

Figure 4.

Inhibition of Notch signaling blocks T cell development and promotes NK cell development. CD117⁺ Sca-1^hi day 14 fetal liver HPCs were cultured on OP9-DL1 cells for 7 d with DMSO alone, or in the presence of Inhibitor X, a presenilin 1/2 inhibitor, at the indicated concentrations. (a) Cells from these cultures were analyzed by flow cytometry for the presence of CD19⁺ NK1.1⁻ B cells and CD19⁻ NK1.1⁺ NK cells. (b) CD19⁻ cells from these cultures were electronically gated and analyzed for the expression of CD44 and CD25. (c) Total number of NK cells generated from 1.5 × 10³ HPCs cultured on OP9-DL1 stromal cell lines for 7 d in the absence (DMSO alone) or presence of Inhibitor X at the indicated concentrations. The data are representative of at least three independent trials.

Figure 5.

Analysis of Dll1 expression by thymic epithelial cells. Immunofluorescent staining of thymus sections reveals that Dll1 (top left, red) is expressed in a reticular pattern in the thymic cortex (in all panels, the capsule is in the extreme top left corner and the medulla begins in the bottom right corner, with the cortico-medullary junction indicated by the yellow dotted line). This pattern of expression is mirrored by cortical stromal cells that express cytokeratin (bottom left, green), as can be seen when the Dll1 and cytokeratin images are merged (top right, counterstained with DAPI, blue). Expression of Dll1 generally appears to be strongest in the deep cortex, but is present throughout the cortex on most cytokeratin⁺ stromal cells. The bottom right panel shows a serial section of the same thymus stained with a nonspecific control antibody of the same isotype and concentration as the Dll1 antibody.