Targeting BCL11B in CAR-engineered lymphoid progenitors drives NK-like cell development with prolonged anti-leukemic activity

Abstract

Chimeric antigen receptor (CAR)-induced suppression of the transcription factor B cell CLL/lymphoma 11B (BCL11B) propagates CAR-induced killer (CARiK) cell development from lymphoid progenitors. Here, we show that CRISPR-Cas9-mediated Bcl11b knockout in human and murine early lymphoid progenitors distinctively modulates this process either alone or in combination with a CAR. Upon adoptive transfer into hematopoietic stem cell recipients, Bcl11b-edited progenitors mediated innate-like antigen-independent anti-leukemic immune responses. With CAR expression allowing for additional antigen-specific responses, the progeny of double-edited lymphoid progenitors acquired prolonged anti-leukemic activity in vivo. These findings give important insights into how Bcl11b targeting can be used to tailor anti-leukemia functionality of CAR-engineered lymphoid progenitor cells.

Keywords: BCL11B; NK cell differentiation; acute myeloid leukemia; adoptive immunotherapy; gene editing; hematopoietic stem cell transplantation; lymphoid progenitors.

Graphical abstract

Figure 1

Isolated BCL11B knockdown in cord-blood-derived CD34⁺ HSPCs inhibits T cell differentiation in vitro less profoundly than early hCD123-CAR expression (A) Representation of the lentiviral vectors: shLuc (luciferase-targeted shRNA used as control), shBcl11b (Bcl11b-targeted shRNA), h12328bbz1 (human CD123 CAR), and h12328bbz1_shBcl11b (human CD123 CAR + Bcl11b-targeted shRNA). All constructs are equipped with a GFP reporter gene. (B) Responses of h12328bbz-transduced PBMCs upon stimulation with either naive or hCD123-transduced Jurkat or 293T cells. Stimulation was quantified by measuring CD107a degranulation. Representative data from one of three independent experiments are shown. (C–G) Human CD34⁺ UCB-derived HSPCs were engineered with the respective constructs and consecutively differentiated on OP9-DL1 stromal cells. Expression and flow cytometry analysis were performed within the GFP⁺ gate on day 14 of co-culture. Representative results from one of two independent experiments are shown. (C) qPCR analysis of Bcl11b expression in transduced lymphoid progenitors. (D and E) Representative flow cytometry analysis of engineered human hematopoietic progenitors showing different stages of T cell development as determined by CD5, CD7, and CD34 expression. (F) Histograms represent NOTCH1 expression on respectively engineered lymphoid progenitors on day 14 of culture. (G) Lymphoid progenitors derived from modified HSPCs were comparatively analyzed for NKG2D, NKp46, CD16, and CD161 expression on day 14 of culture. Data from one experiment are shown. Statistics were performed using Student’s t test (two tailed) and one-way ANOVA with Tukey’s post hoc test. Each data point represents an individual sample. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns, not significant.

Figure 2

Generation of a CAR against murine CD123 (mCD123 CAR, im12328z1) (A) Antigen binders against murine CD123 were identified using HAL9/10 antibody phage display libraries. (B) His-tagged scFv binding of cell-bound mCD123 was validated by flow cytometry. Representative flow cytometry panels depict the three most promising binders, namely C2, A11, and A8. They demonstrated a binding pattern similar to the positive control that was generated with a commercially available mCD123 antibody (APC; catalog 4330936; clone 5B11). Experiments were performed once. (C) The murine AML cell line C1498 was transduced to overexpress murine CD123 and was used for im12328z1 stimulation. A representative histogram from one of three independent experiments is shown. (D) The selected three scFv binders (SH1836b-A11, -A8, and -C2) were cloned into a second-generation CAR backbone. The lentiviral CD123 CAR construct im12328z1 (inducible murine CD123 CAR, CD28 co-stimulation, one functional ITAM containing CD3ζ domain) was linked to an IRES dTomato cassette for easy detection. SIN LTR, self-inactivating long terminal repeats; T11, Dox-inducible promoter; scFv, single-chain variable fragment; TM, transmembrane domain; IRES, internal ribosome entry site; PRE, woodchuck hepatitis virus posttranscriptional regulatory element. (E–H) Comparative functionality testing of the three im12328z1 CAR constructs made with the selected three binders. Representative data from one of three independent experiments are shown. (E) im12328z1 expression (carrying the SH1836b-A11, -A8, or -C2 scFv) was assessed by flow cytometry on transduced NFAT cells using protein L staining. (F) Responses of im12328z1-transduced primary murine T cells to stimulation with either C1498-mCD123 or C1498 control cells were quantified via CD107a degranulation (left) and IFN-γ production (right). (G) Intracellular signal strength (GFP expression) of im12328z1-transduced NFAT cells upon stimulation with either C1498-mCD123 or C1498 control cells was assessed by flow cytometry. (H) Specific cytotoxic activity of im12328z1 (A11)-engineered primary murine T cells against either C1498-mCD123 or C1498 control cells using the JAM assay. MFI, mean fluorescence intensity. Statistics were performed using Student’s t test (two tailed) and one-way ANOVA with Tukey’s post hoc test. Each data point represents an individual sample. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns, not significant.

Figure 3

im12328z1-engineered HSPCs differentiate into CARiK cells in vitro and provide strong anti-leukemia effects in HCT recipients (A) Representation of the lentiviral vectors: inducible dTomato reporter gene only (iTom) and im12328z1. (B–E) Murine LSK cells were transduced with either im12328z1 or the iTom control vector and consecutively differentiated on OP9-DL1 stromal cells. Flow cytometry analysis was performed within the Tom⁺ gate on day 20 of co-culture. Representative results from one of three independent experiments are shown. (B) Representative flow cytometry plot for CD44 and CD25 expression of im12328z1-engineered lymphoid progenitors (blue) and respective iTom controls (red, left) and frequencies of NK1.1⁺ cells (right). DN, double negative. (C) Responses of engineered lymphoid progenitors upon stimulation with C1498/C1498_mCD123 target cells were quantified by degranulation. (D) CD123 expression on engineered lymphoid progenitors and (E) AnnexinV/PI positivity were measured at indicated time points. (F and G) Irradiated female B6 mice were reconstituted with 3 × 10⁶ B6 TCD-BM cells (n = 10/group) and co-transplanted with 8 × 10⁶ B6 im12328z1-engineered lymphoid progenitors. Studies were performed under permanent administration of Dox-containing water or food for transgene expression. Recipients were intravenously challenged with 1 × 10⁶ C1498-mCD123 cells on day 21 after transplantation. Survival was monitored. Results from two independent transplantations were pooled. (H) Survivors were re-challenged with 1 × 10⁶ C1498-mCD123 cells on day 100 and assessed for survival. TCD-BM-only recipients served as controls. (I) Myeloid and lymphoid recovery in irradiated female B6 recipients of 3 × 10⁶ TCD-BM cells (n = 4/group) with or without 8 × 10⁶ syngeneic iTom/im12328z1-engineered lymphoid progenitors was graphed. At indicated time points, blood counts were measured and peripheral blood was analyzed for CD45 (total white blood cells), Gr.1 (granulocytes), CD19 (B cells), and CD3 (T cells) by flow cytometry. Experiments were performed once. (J) Primary murine BM, C1498, and C1498-mCD123 cells were comparatively assessed for CD123 expression strength by flow cytometry. Respective results from one of three independent experiments are shown. (K) Frequencies of Tom⁺ progeny in the spleen of im12328z1 transplant recipients over time by flow cytometry analysis (n = 3 mice/time point). Experiments were performed once. (L) Splenocytes of recipients of either im12328z1-engineered lymphoid progenitors or respective iTom controls were harvested on day 12 and re-cultured ex vivo under CARiK-optimized conditions (n = 12). Data are shown as mean ± SEM. Student’s t test (two tailed) and one-way ANOVA with Tukey’s post hoc test were used for analysis. Each data point represents an individual sample. Survival curves were compared using Mantel-Cox (log rank) test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Figure 4

Bcl11b knockout in CAR-engineered lymphoid progenitors alters the cytokine release profile and cytotoxic responsiveness after specific stimulation in vitro (A) The lentiviral vectors iTom, luciferase-targeted sgRNA for control (sgLuc), Bcl11b-targeted sgRNA (sgBcl11b), im12328z1, im12328z1_sgBcl11b (im12328z1 + sgBcl11b). All constructs are equipped with a dTomato reporter gene. (B–I) Murine LSK cells were transduced with lentiviral vectors and consecutively differentiated on OP9-DL1 stromal cells. Doxycycline was added during the entire culture period from day 1 onward. Cells were sorted for Tom⁺ on day 10 of culture. Flow cytometry analysis was performed within the Tom⁺ gate on day 20 of co-culture. Representative results from one of two independent experiments are shown. (B) Western blot analysis of BCL11B in lysates from bulk cultures of transduced lymphoid progenitor cells. (C) Representative flow cytometry plots of CD44 and D25 expression on in vitro-generated engineered lymphoid progenitors. DN, double negative. (D) qPCR analysis of Bcl11b expression in transduced lymphoid progenitors. (E) Engineered lymphoid progenitors were assessed for the expression of NK1.1, NKp46, and CCR9 by flow cytometry. (F) Recombination of D and J regions of the TCR-β locus in engineered lymphoid progenitors. Genomic DNA of engineered progenitors was isolated on day 20 of culture and rearrangements were detected by PCR. Splenocytes and thymocytes from WT B6 mice as well as non-transduced lymphoid progenitors were used as controls. GL, germ line. (G) Antigen specificity and functionality of engineered lymphoid progenitors upon stimulation were quantified via granzyme B and CD107a degranulation. (H) Specific cytotoxic activity of engineered lymphoid progenitor cells against either C1498-mCD123 or C1498 control cells using a flow cytometry-based assay. (I) Multiplex cytokine/chemokine analysis of supernatant from engineered lymphoid progenitors that had been co-cultured with target cells for 24 h. A set of 24 different cytokines and chemokines were randomly measured. Data from one experiment are shown as mean ± SEM. Statistics were performed using one-way ANOVA with Tukey’s post hoc test. Each data point represents an individual sample. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns, not significant.

Figure 5

Potential CAR gene-driven mutagenesis of CAR-modified lymphoid progenitors (A) Murine LSK cells were transduced with sgBcl11b, im12328z1, im12328z1_sgBcl11b, or the iTom control vector and differentiated into lymphoid progenitors using the OP9-DL1 co-culture system. ISA was performed during in vitro differentiation on days 17, 20, and immediately before co-transplantation on day 24 (n = 3 independent in vitro cultures). Irradiated female B6 recipients were reconstituted with 3 × 10⁶ B6 TCD-BM and co-transplanted with 8 × 10⁶ engineered lymphoid progenitors (n = 3–5 mice/group). Studies were performed under permanent administration of Dox-containing water or food for transgene expression. Progeny cells were retrieved from the recipients and assessed for clonal evolution 12 days after transfer. (B) The top 10 integration sites at the different time points. Data of three independent cultures are shown. Each colored bar represents a unique insertion site from the top 10 most abundant sequences. The gray bars represent all other background insertions. All shown insertion sites correspond to a unique chromosomal position. (C) PCA of the four gene-edited lymphoid progenitor groups on day 20 of culture. Data of 4–7 independent cultures are shown. They were compared to sorted DN2 cells either non-transduced (mock) or transduced with genotoxic vectors. The support vector machine (SVM) predictions are represented with different shapes (circle, non-transforming; triangle, transforming) and the vector designs with color codes (MOCK, gray; SIN-LV.EFS, green; RSF91, red; SIN_LV.LMO2, dark gray). (D) NES values obtained with SAGA-XL-GSEA using the DN2 features. The four constructs were compared to metadata of DN2-sorted cells. (E) Lmo2 and Mef2 mRNA levels measured by ddPCR in gene-edited lymphoid progenitors and respective controls on day 20 of culture. All values were normalized to the expression of the housekeeping gene β-actin. Each data point represents an individual sample. Data are shown as mean ± SEM.

Figure 6

Combining CAR engineering with CRISPR-Cas9-induced Bcl11b knockout enhances NK cell-like properties and prolongs anti-leukemic efficacy in vivo (A) Irradiated female B6 recipients were reconstituted with 3 × 10⁶ B6 TCD-BM and co-transplanted with 8 × 10⁶ engineered lymphoid progenitors (n = 4 mice/group). Studies were performed under permanent administration of Dox-containing water or food for transgene expression. (B) Thymic sections were imaged for Tom⁺ cells. Original magnification, ×20. (C) Cells from harvested thymi were analyzed by flow cytometry for Tom⁺ progeny of co-transplanted lymphoid progenitors. (D) The progeny of lymphoid progenitors were retrieved from thymus on day 12 after transfer. Representative flow cytometry plots show CD4/CD8 expression (left) and frequencies of TCR-β⁺ and CD4⁺CD8⁺ progeny within the Tom⁺ gate (right). Respective analysis of the spleen (E) and BM (F) are depicted as absolute numbers of Tom⁺, NK1.1⁺, and NKp46⁺ cells. Experiments were performed once. (G and H) Irradiated female B6 recipients of 3 × 10⁶ TCD-BM and 8 × 10⁶ syngeneic B6 engineered lymphoid progenitors (n = 13–15 mice/group) were challenged with 1 × 10⁶ C1498-mCD123 leukemia cells on day 21 after transplantation and consecutively monitored for survival. Results from three independent transplantations were pooled. (I) Survivors were re-challenged with 1 × 10⁶ C1498-mCD123 cells on day 100. Data from two independent experiments were pooled (n = 10–13 mice/group). Student’s t test (two tailed) and one-way ANOVA with Tukey’s post hoc test were used for analysis. Survival curves were compared using Mantel-Cox (log rank) test. Data are shown as mean ± SEM. Each data point represents an individual sample. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns, not significant.

Figure 7

The in vivo progeny of im12328z_sgBcl11b lymphoid progenitors shows a shift toward transcriptional activation associated with NK cell identity and cytotoxicity (A) Irradiated B6 recipients were reconstituted with 3 × 10⁶ B6 TCD-BM cells and co-transplanted with 8 × 10⁶ engineered lymphoid progenitors (n = 5 mice/group). Studies were performed under permanent administration of Dox-containing water or food for continuous induction of transgene expression. scRNA-seq analysis was accomplished on spleen-derived progeny of gene-edited lymphoid progenitors 12 days after further in vivo differentiation. (B) UMAP visualization of scRNA-seq data displaying cell types identified from isolated lymphocytes across all samples. Integrated datasets (n = 7,715 cells) reveal T cells (green circle), NK cells (red circle), and monocytes (light blue circle) as major cell types. Each major cell type is further subdivided into smaller subtypes, classified based on a combination of cell type and gene expression clusters (UMAP clusters Cl). (C) Dimensional reduction plot for each sample, focusing on T and NK cells. The integrated dataset was split into individual samples (im12328z1, im12328z1_sgBcl11b, iTom, sgBcl11b) to visualize the cell type plasticity introduced by cell engineering. (D) Differently engineered cells exhibit a distinct cell type profile. Bar plots for each cell subtype display normalized cell frequencies for each sample. (E) Clustered heatmap visualizing distinct expression profiles of differently engineered samples. Genes selected for the heatmap show differential expression compared to the iTom control dataset (absolute log 2 fold change [log2FC] >1, FDR < 0.05). The heatmap displays Z-scored average gene expression of 263 selected genes, with hierarchical clustering isolating six clusters (hierarchical clusters) with similar expression profiles. (F) Expression of selected marker genes. The dot plot shows the expression levels of specific marker genes across all cells within each sample. T cell markers: Cd3g, Cd3d, Cd3e. NK cell markers: Ncr1, Klrb1c, Klrk1. Cytotoxicity markers: Prf1, Ifng, Gzma, Gzmb. Activation/homing markers: Sell, Ccr2. Each row represents a marker gene, while each column corresponds to a distinct sample. Dot size indicates the proportion of cells in each sample expressing the respective marker gene, and color intensity reflects the average expression level of that gene. This figure highlights the differential expression of key marker genes across all samples, providing insights into the impact of different cellular modifications. (G) Gene signatures for regulation of cell killing and NK T cell differentiation are enriched in double-edited samples (im12328z1_sgBcl11b). GSEA was conducted on clusters identified by hierarchical clustering using the Gene Ontology (GO) category "Biological Process." The top five significant categories for each cluster are plotted, with dot size representing the gene ratio (number of genes detected in each cluster divided by the total number of genes in each category). The p.adjust values indicate the significance of enrichment.