Generation of PLZF+ CD4+ T cells via MHC class II-dependent thymocyte-thymocyte interaction is a physiological process in humans

Abstract

Human thymocytes, unlike mouse thymocytes, express major histocompatibility complex (MHC) class II molecules on their surface, especially during the fetal and perinatal stages. Based on this observation, we previously identified a novel developmental pathway for the generation of CD4+ T cells via interactions between MHC class II-expressing thymocytes (thymocyte-thymocyte [T-T] interactions) with a transgenic mouse system. However, the developmental dissection of this T-T interaction in humans has not been possible because of the lack of known cellular molecules specific for T-T CD4+ T cells. We show that promyelocytic leukemia zinc finger protein (PLZF) is a useful marker for the identification of T-T CD4+ T cells. With this analysis, we determined that a substantial number of fetal thymocytes and splenocytes express PLZF and acquire innate characteristics during their development in humans. Although these characteristics are quite similar to invariant NKT (iNKT) cells, they clearly differ from iNKT cells in that they have a diverse T cell receptor repertoire and are restricted by MHC class II molecules. These findings define a novel human CD4+ T cell subset that develops via an MHC class II-dependent T-T interaction.

Figure 1.

PLZF expression and acquisition of innate properties in mouse CD4⁺ T cells is dependent on MHC class II–dependent T–T interactions. (A) Flow cytometry of thymocytes and splenocytes from WT, CIITA^tg, CIITA^tgpIV^−/−, CIITA^tgCD1d^−/−, and CD1d^−/− mice assessing PLZF expression in permeabilized CD1d/αGalCer tetramer–positive (Tetr⁺) and –negative (Tetr⁻) populations. The numbers indicate the percentage of each PLZF⁺ subset (Tetr⁻PLZF^hi, Tetr⁻PLZF^lo, Tetr⁺PLZF^hi, and Tetr⁺PLZF^lo) among total cells. Representative data from two independent experiments are shown. (B) A representative profile of CD24, CD44, CD62L, NK1.1, and CD122 expression versus PLZF expression in gated CD4 SP thymocytes and splenic CD4⁺ T cells from CIITA^tgpIV^−/− mice. PLZF^hi and PLZF^lo populations could be identified. The numbers indicate the percentage of cells in each quadrant. Representative data from six independent experiments are shown. (C) Intracellular flow cytometry for IL-4 and IFN-γ in WT and CIITA^tgpIV^−/− mice. Splenic CD4⁺ T cells from each mouse were activated with PMA and ionomycin for 5 h and assessed for their cytokine secretion. The numbers in the dot plots indicate the percentage of cells in each quadrant. Representative data from three independent experiments are shown. (D) Flow cytometry of thymocytes to assess PLZF expression in OT-II TCR transgenic thymocytes after CD45.1/OT-II and CD45.2/CIITA^tg BM cells were mixed and transferred into lethally irradiated pIV^−/− hosts. The dot plot shows PLZF expression in CD4 SP thymocytes 7 wk after transfer. The numbers indicate the percentage of PLZF⁺ cells among the CD45.1⁺ (OT-II) and CD45.1⁻ (WT) CD4 SP thymocytes. Spl, spleen; Thy, thymus.

Figure 2.

Rat and human thymocytes drive MHC class II–dependent T–T interaction and PLZF expression in a chimeric mouse system. (A) Generation of PLZF⁺ CD4⁺ T cells in a rat→mouse BM chimera. Rat BM stem cells were transplanted into irradiated RAG-1^−/−γc^−/−MHCII^−/− mice with (middle) or without (top) fetal rat thymus graft. 8 wk after transplant, the recipient mouse thymus, the grafted rat thymus, and the recipient spleen were harvested, and PLZF expression in each subset of rat T cells was compared with that of WT rats. CD4 SP thymocytes were subdivided into two stages based on CD8 expression, as indicated. The numbers indicate the percentage of cells in each quadrant. Representative FACS data of two mice from two independent experiments are shown. The forward scatter (FSC) value is displayed as a linear scale. (B) T cell development and PLZF expression in CD34⁺ cord blood–reconstituted NOG mice with (bottom) or without a human fetal thymus graft (top). 14 wk after the transfer of cord blood cells, PLZF expression in human T cells was analyzed in the recipient mouse thymus and spleen as well as the grafted human thymus. CD4 SP thymocytes were subdivided into three populations based on CD8 expression, as indicated. The numbers indicate the percentage of cells in each quadrant. Representative FACS data from three independent experiments are shown.

Figure 3.

PLZF is expressed in a subset of fetal human CD4⁺ T cells that have properties similar to mouse T–T CD4⁺ T cells during the second trimester of gestation. (A) Representative expression profile of PLZF in the fetal thymus (GA = 16 wk) and spleen (GA = 23 wk), as well as the neonatal (day 10) thymus. The numbers in each quadrant indicate the percentage of cells present. The forward scatter (FSC) value is displayed as a linear scale. (B) Summary of PLZF⁺ cell frequency in the fetal human thymus and spleen. The PLZF⁺ populations in CD4 SP thymocytes and splenic CD4⁺ T cells were enumerated by flow cytometry at a GA of 16–26 wk. (C) CD45RO and CD161 expression on PLZF⁺ and PLZF⁻ CD4 SP thymocytes or splenic CD4⁺ T cells from human fetuses at a GA of 19 wk. (D) Intracellular flow cytometry of IFN-γ and PLZF in human fetal CD4⁺ T cells. MACS-enriched CD4⁺ T cells from human fetal splenocytes were activated with PMA and ionomycin for 5 h. The representative data of two independent experiments. (E) Flow cytometric assessment of the PLZF⁺ non-iNKT population in human fetal thymocytes (CD4 SP and DP gated) and CD4⁺ splenocytes. The iNKT cells were identified using CD1d/αGalCer tetramers after comparison with cells stained with unloaded CD1d (uCD1d) tetramer. The numbers indicate the percentage of cells in each quadrant. DN, double negative.

Figure 4.

PLZF⁺ T–T CD4⁺ T cells in mice and humans have diverse TCR Vβ usage. The bar graph shows the Vβ chain distribution of PLZF⁺ and PLZF⁻ CD4 SP thymocytes and splenic CD4⁺ T cells from CIITA^tgpIV^−/− mice (A) and a human fetus (GA = 18 wk; B). This is compared with conventional CD4⁺ T cells from a WT mouse (A) and human cord blood (B). The data are mean values ± SEM of four animals in each group of mice, with one human fetus and four cord blood samples.