BCL11B depletion induces the development of highly cytotoxic innate T cells out of IL-15 stimulated peripheral blood αβ CD8+ T cells

Abstract

BCL11B, an essential transcription factor for thymopoiesis, regulates also vital processes in post-thymic lymphocytes. Increased expression of BCL11B was recently correlated with the maturation of NK cells, whereas reduced BCL11B levels were observed in native and induced T cell subsets displaying NK cell features. We show that BCL11B-depleted CD8+ T cells stimulated with IL-15 acquired remarkable innate characteristics. These induced innate CD8+ (iiT8) cells expressed multiple innate receptors like NKp30, CD161, and CD16 as well as factors regulating migration and tissue homing while maintaining their T cell phenotype. The iiT8 cells effectively killed leukemic cells spontaneously and neuroblastoma spheroids in the presence of a tumor-specific monoclonal antibody mediated by CD16 receptor activation. These iiT8 cells integrate the innate natural killer cell activity with adaptive T cell longevity, promising an interesting therapeutic potential. Our study demonstrates that innate T cells, albeit of limited clinical applicability given their low frequency, can be efficiently generated from peripheral blood and applied for adoptive transfer, CAR therapy, or combined with therapeutic antibodies.

Keywords: ADCC; AICC; BCL11B; CRISPR/Cas9; IL-15; alpha-beta CD8+ T cells; human; innateness; knock-out.

Figure 1.

Efficacy of BCL11B knock-out in CD8+ T cells enriched from PBMC. (a) BCL11B locus with magnified exon 4 including primer (arrows) and cRNA binding sites. Fragments are labeled with the predicted sizes of PCR products. (b) Flow cytometry histogram of fluorescence-labeled tracrRNA signal 24 h after transfection. (c) Representative agarose gel showing PCR products of genomic DNA templates with CRISPR/Cas9 target loci-spanning primers as indicated in (a). (d) Representative Western blot showing BCL11B levels in relation to HDCA1 in the nuclear fraction. (e) Normalized protein level to nt of mock condition and BCL11B knock-out, mean ±SEM. Statistical significance was calculated by unpaired t test. nt: not transfected control, trans: transfected-only control (no gRNA), mock: mock gRNA transfected, BCL11B KO: BCL11B knock-out, TSS: transcription start site. ****P ≤ 0.0001.

Figure 2.

BCL11B deletion in CD8+ T cells results in altered immunophenotype (a) First row: tSNE plots showing clustering of the complete concatenated sample as an overlay of all conditions or separately for each experimental condition. Second till last row: tSNE plots showing the surface abundance of the indicated markers in a complete concatenated sample (overlay) or separately for each experimental condition. Last column: flow cytometry histograms show staining of the indicated markers separately for each experimental condition in the concatenated sample. All plots contain cumulative results of 16 different individuals. (b) Percentage of CD16+ cells in each experimental condition measured by flow cytometry, mean ± SEM. (c) Top: flow cytometric staining for CD56 and NKp30 in one representative individual. Bottom: MFI of CD56+ and NKp30+ cells in each experimental condition, mean ± SEM, n = 15. (d) Percentage of indicated populations over time measured by flow cytometry, mean ± SEM, n = 18–31. (e) Top: frequency of CD45RO and CD45RA cells in each experimental condition determined by flow cytometry in one representative individual. Bottom: MFI of CD45RA and CD45RO in each experimental condition, mean ± SEM, n = 15. (f) Percentage of CD57+ cells in each experimental condition measured by flow cytometry, mean ± SEM, n = 25. nt: not transfected control, mock: mock gRNA transfected control, BCL11B KO: BCL11B knock-out. MFI: mean fluorescence intensity. Statistical significance of not normally distributed data sets (b, c [MFI NKp30+], e [MFI CD45RA], f) was calculated by the Kruskal–Wallis test with Dunn’s multiple comparisons test correction and of normally distributed data sets (c [MFI CD56+], e [MFI CD45RO]) was calculated by the One-way ANOVA with Tukey’s multiple comparisons test correction. Statistical significance of time-dependent kinetics (d) was calculated by the Two-way ANOVA with Tukey’s multiple comparisons test correction. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

Figure 3.

Global transcriptome analysis of BCL11B knock-out CD8+ T cells exhibits a unique phenotype characterized by upregulation of various genes related to innate lymphocyte populations. (a) Heatmap of detected genes in BCL11B knock-out (KO) cells, not transfected (nt) control, and mock control cells. Hierarchical clustering of columns was performed using the Euclidean distance method. (b) Volcano plot showing differential gene expression in BCL11B knock-out compared to mock control cells. Genes significantly upregulated and downregulated are shown in magenta and teal, respectively. Dashed lines indicate cutoff at absolute log₂ ratio ≥ 0.5 and −log₁₀ false discovery rate (FDR) q-value < 0.05. (c) A combined bar and dot plot shows enriched KEGG pathways in BCL11B knock-out cells compared to mock control cells. The normalized enrichment score (NES) is shown as black bars. Positive NES values indicate enrichment in the BCL11B knock-out condition, while negative values result from enrichment in the mock condition. FDR q-values are plotted as magenta dots. (d, e, f) Heatmaps of NK, T, and overlapping cell identity genes which are curated by a public database and were detected in this dataset (g) Gene expression of ZNF683 in BCL11B knock-out cells and mock control relative to ZNF683 expression in nt control cells, mean ± SEM, n = 10 (dashed line set to 1). Statistical significance was calculated by the Mann–Whitney test. nt: not transfected control, mock: mock gRNA transfected control, BCL11B KO: BCL11B knock-out. **P ≤ 0.01.

Figure 4.

Proteins with altered abundances after BCL11B knock-out in CD8+ T cells displayed enrichment in pathways related to signal transduction and actin cytoskeleton regulation. (a) A combined bar and dot plot shows KEGG pathways with enrichment of proteins displaying altered protein abundances in BCL11B knock-out cells compared to mock control cells. The normalized enrichment score (NES) from GSEA is shown as black bars. Positive NES values indicate increased abundances in the BCL11B knock-out condition, while negative values result from higher protein abundances in the mock condition. FDR q-values are plotted as magenta dots. (b) Selected proteins from KEGG pathways shown in Figure 4 A. Compared to mock control, the depicted proteins exhibited significantly changed abundances in the global proteome analysis (absolute fold change ≥ 1.5, P ≤ 0.05) in BCL11B knock-out CD8+ T cells. (c) Top: representative flow cytometry plots showing ITGA1 and ITGAX signals in each experimental condition. Bottom: percentage of ITGA+, ITAGX+, and ITGA1+ ITAGX+ cells detected by flow cytometry, mean ± SEM, n = 3. Statistical significance was calculated by the One-way ANOVA with Tukey’s multiple comparisons test correction. nt: not transfected control, mock: mock gRNA transfected control, BCL11B KO: BCL11B knock-out. *P ≤ 0.05, **P ≤ 0.01.

Figure 5.

Single-cell transcriptome profiles of BCL11B knock-out in comparison to control T cells. Cells lacking BCL11B exhibited a partial loss of naivety and gain of innate-like phenotype. Dot plots show changes in (a) naivety score or (b) innateness score between the three experimental conditions. Scores were calculated by average expression of lineage-specific genes listed in Supplementary Table 6. (c) Heatmap generated from microarray results of global transcriptome analysis showing differential expression of PLCG1 and PLCG2 as well as LYN and FCER1G, n = 3. (d) Representative Western blot shows differential protein abundance of PLZF, ID2, and HOPX in relation to HDAC1 used as a nucleus-specific loading control. nt: not transfected control, mock: mock gRNA transfected control, BCL11B KO: BCL11B knock-out.

Figure 6.

Gene expression and surface marker profile of costimulatory and inhibitory receptors upon BCL11B knock-out. (a) Gene expression heatmap of curated costimulatory and inhibitory receptor genes, n = 3. (b) Top: representative flow cytometry histograms showing surface expression of indicated markers in nt control (gray filled), mock control (magenta line), and BCL11B knock-out (teal line) cells. Bottom: MFI or percentage (as indicated) of CD63, CD40LG, CD161, and TNFRSF18. Mean ± SEM, n = 3. nt: not transfected control, mock: mock gRNA transfected control, BCL11B KO: BCL11B knock-out. Statistical significance was calculated by Kruskal–Wallis test with Dunn’s multiple comparisons test (CD63, CD161) or by One-way ANOVA with Tukey’s multiple comparisons test (CD40LG, TNFRSF18). **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

Figure 7.

Degranulation and IFNγ production by BCL11B knock-out CD8+ T cells compared to controls. (a–c) Representative flow cytometry plots (left) showing degranulation as CD107a induction and IFNγ production and percentage (right) of CD107a+ or IFNγ+ cells after stimulation with beads presenting (a) a single antibody, (b) a combination of two antibodies, and (c) anti-CD3 alone or in combination with anti-NKp30. Mean ± SEM, n = 5. Statistical significance was calculated by two-way ANOVA with Tukey’s multiple comparisons test. nt: not transfected control, mock: mock gRNA transfected control, BCL11B KO: BCL11B knock-out. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

Figure 8.

Cytotoxic activity of BC11B knock-out CD8+ T cells against leukemic or neuroblastoma tumor cell lines. (a, b) Spontaneous cell-mediated cytotoxicity of BCL11B knock-out cells and controls against (a) K562 infrared fluorescent protein (iRFP) expressing or (b) K562 B7H6 knock-out (KO) iRFP target cells. (c) Antibody-dependent cell-mediated cytotoxicity of BCL11B knock-out cells and controls in presence of anti-GD2 against CHLA-136 iRFP target cells. (d) Cell-mediated cytotoxicity of BCL11B knock-out cells in the absence or presence of anti-GD2 or an isotype control antibody (ITC). Mean ± SEM, n = 5–6. nt: not transfected control, mock: mock gRNA transfected control, BCL11B KO: BCL11B knock-out. Statistical significance was calculated by Mann–Whitney test. **P ≤ 0.01.