The peripheral differentiation of human natural killer T cells

Abstract

The peripheral maturation of human CD1d-restricted natural killer T (NKT) cells has not been well described. In this study, we identified four major subsets of NKT cells in adults, distinguished by the expression of CD4, CD8 and CCR5. Phenotypic analysis suggested a hierarchical pattern of differentiation, whereby immature CD4⁺ CD8^- CCR5^- cells progressed to an intermediate CD4⁺ CD8^- CCR5⁺ stage, which remained less differentiated than the CD4^- CD8^- and CD4^- CD8⁺ subsets, both of which expressed CCR5. This interpretation was supported by functional data, including clonogenic potential and cytokine secretion profiles, as well as T-cell receptor (TCR) excision circle analysis. Moreover, conventional and high-throughput sequencing of the corresponding TCR repertoires demonstrated significant clonotypic overlap within individuals, especially between the more differentiated CD4^- CD8^- and CD4^- CD8⁺ subsets. Collectively, these results mapped a linear differentiation pathway across the post-thymic landscape of human CD1d-restricted NKT cells.

Keywords: NKT cell; T-cell differentiation; TCR; TREC.

Figure 1

Identification and phenotype of human NKT cells. (a) Viable CD3⁺ cells (left) were assessed for TRAV10 (Vα24) and TRBV25 (Vβ11) expression (top right) or PBS57‐hCD1d multimer binding (bottom right). Data are representative of three independent experiments (n = 5–12 subjects). PI, propidium iodide. (b) Expression of the indicated phenotypic markers is shown for NKT (PBS57‐hCD1d⁺) and conventional T cells (PBS57‐hCD1d⁻). Plots are gated on viable CD3⁺ cells. Data are representative of five independent experiments (n = 12 subjects). Summary statistics are provided in Table 1.

Figure 2

Subsets of NKT cells. (a) Expression of CD4, CD8α and CD8β is shown for NKT (PBS57‐hCD1d⁺) and conventional T cells (PBS57‐hCD1d⁻) from a representative individual. Plots are gated on viable CD3⁺ cells. Quadrant numbers indicate percentage values. The percentages are averaged across three independent experiments (n = 7 subjects). (b) Expression of the indicated cytokines is shown for CD4⁺, DN and CD8⁺ NKT cells from a representative individual. The percentages are averaged across two independent experiments (n = 5 subjects). (c) CD4⁺, DN and CD8⁺ NKT cells (n = 10 000 per subset) were stimulated with the PBS57‐hCD1d multimer. The amount of each cytokine released into the supernatant was measured using cytokine bead array. Results are shown as the mean ± one standard deviation from two independent experiments (n = 3–5 subjects). *P < 0.05 (Kruskal–Wallis test with Steel–Dwass correction).

Figure 3

Differentiation of NKT cells. (a) Expression of the indicated phenotypic markers is shown for CD4⁺ (top) and CD4⁻ NKT cells (bottom) from a representative individual. Quadrant numbers indicate percentage values. The percentages are averaged across four independent experiments (n = 12 subjects). (b) Expression of the indicated phenotypic markers (left panels) and cytokines produced in response to stimulation with the PBS57‐hCD1d multimer (right panels) is shown for cord blood NKT (PBS57‐hCD1d⁺) and conventional T cells (PBS57‐hCD1d⁻) from a representative individual. The percentages are averaged across four independent experiments (n = 5 subjects). (c) TREC levels were quantified in sort‐purified NKT cell subsets. Horizontal bars indicate mean values. Each data point represents one individual. *P < 0.05 (Kruskal–Wallis test with uncorrected pairwise comparisons for adult PBMCs). (d) Single CD4⁺, DN and CD8⁺ NKT cells were sorted and cloned. The fraction of cells that expanded in vitro is shown as the mean ± one standard deviation from three independent experiments (n = 3 or 4 subjects). ***P < 0.005 (Kruskal–Wallis test with Steel–Dwass correction).

Figure 4

Conventional analysis of TCR use in NKT cell subsets. CD4⁺, DN and CD8⁺ NKT cells were sort‐purified from healthy subjects (n = 3). A conventional sequencing approach was used to profile the corresponding TCRα and TCRβ repertoires. The fraction of each repertoire expressing a particular TRAV (a) or TRBV gene (c) is shown together with the fraction of each repertoire expressing a specific TCRα (b) or TCRβ sequence (d). The invariant TCRα sequence TRAV10/CVVSDRGSTLGRLY/TRAJ18 is shown in black.

Figure 5

High‐throughput analysis of TCR use in NKT cell subsets. CD4⁺ CCR5⁻, CD4⁺ CCR5⁺, DN and CD8⁺ NKT cells were sort‐purified from healthy subjects (n = 3). A high‐throughput sequencing approach was used to profile the corresponding TCRβ repertoires. The experiment was repeated 6 months later. (a) The number of unique TCRβ sequences identified for each NKT cell subset is listed in parentheses. Numbers below the diagonal indicate how many of these unique TCRβ sequences were shared across NKT cell subsets from the same individual at a single time point. (b) The number of shared unique TCRβ sequences is shown for each combination of NKT cell subsets within an individual at the indicated time points.

Figure 6

The peripheral ontogeny of human NKT cells. Based on the phenotypic, functional and molecular data presented in this study, we propose a linear differentiation model that describes the peripheral ontogeny of human NKT cells. The depicted lineage relationships are consistent with previous reports, as reviewed in detail elsewhere.13