Cell therapy for a rare disease- hairy cell leukemia variant

Abstract

Hairy cell leukemia variant (HCL-v) is a rare malignancy of clonal mature B-cells that follows a chronic disease course. HCL-v patients are often resistant to purine nucleoside analogs, which are the first-line therapy. To address the shortcomings of current therapy for HCL-v, we investigated the activity of a BAFF ligand-based CAR-T cell which binds to all three BAFF receptors, BAFF-receptor, TACI, and BCMA. Here, we demonstrate that HCLv patient-derived cells highly express all three BAFF receptors and that BAFF CAR-T cells induce significant cytotoxicity in vitro against both cell lines and HCL-v patient cells. This cytotoxicity corresponds with significant CAR-T cell activation, degranulation, and release of pro-inflammatory cytokines after co-incubation with HCLv cells. Furthermore, we successfully generated BAFF CAR-T cells directly from an HCLv patient and observed direct autologous killing against patient tumor cells in vitro. These HCLv patient-derived CAR-T cells were also effective in killing the Hair-M cell line and tumor cells derived from a different HCLv patient. Lastly, we also developed two mouse xenograft models for HCL, a subcutaneous Bonna-12 model and intravenous Hair-M xenograft model. We observed decreases in tumor burden and prolonged overall survival without significant toxicity. In conclusion, here we show that BAFF CAR-T cells exert anti-tumor effects in vitro and in vivo against multiple cell lines and patient-derived HCL-v samples and may be a successful therapeutic strategy for HCLv patients.

Keywords: BAFF; CAR-T cells; HCL; cytotoxicity; leukemia.

Figure 1.

HCL cell lines express all three BAFF receptors and are sensitive to BAFF CAR-T cells. HCL cells were stained with CD19 (FITC), BCMA (PE), TACI (PE/CY7), BAFF-R (APC) antibodies. The mantle cell lymphoma line JeKo-1 is included as a positive control and primary natural killer cells (NK) as an antigen negative control for receptor expression. Expression level of each receptor was measured via flow cytometry. Histograms were derived from live cell population from FSC/SSC gating (a). Median fluorescent intensity (MFI) of each receptor is derived from the respective histograms and the unstained control and stained samples are shown (b). Fluorescently labeled HCL cell lines were co-cultured with BAFF CAR-T cells at various effector: target ratios (2.5:1, 5:1, 10:1). After 16-24 h of incubation, cells were stained with propidium iodide and percentage of pi-positive cancer cells was quantified via flow cytometry (C shows gating strategy and D represents graphed percentage of PI+ (dead) cancer cells). ****p < 0.0001. Mean ± SD, n = 3 biologically independent co-cultures, two-way ANOVA with Holm- šídák’s multiple comparison test.

Figure 2.

Hcl-v patient samples express all the BAFF receptors and are sensitive to BAFF CAR-T cells. hcl-c/v patient splenocytes were stained as in figure 1, with CD19 (FITC) BCMA (PE), TACI (PE/CY7), BAFF-R(APC) antibodies and overlayed flow histograms comparing the samples and controls are represented. Primary NK cells (NK) are used as an antigen negative control and the mantle cell lymphoma cell line JeKo-1, known to overexpress the receptors, is included as a positive control. The receptor expression is quantified from the histograms as MFI values and the unstained control and stained sample is shown (b). Multiple HCL patient-derived samples were labeled with eFluoro670 dye and co-incubated with BAFF CAR-T cells at various ratios (2.5:1, 5:1, 10:1, 20:1) for 16-24 h. The co-culture was then stained with propidium iodide. The flow gating strategy was to gate on APC positive cells (population P2) to distinguish the target cells and measure percent PI positivity in that population (shown as V2R) (c). The percent PI positive cancer cells were reported in bar graphs. ****p < 0.0001. Mean ± SD, n = 3 biologically independent co-cultures, two-way ANOVA with Holm- šídák’s multiple comparison test. (D).

Figure 3.

BAFF CAR-T cells are activated and degranulated when co-cultured with HCL cells. BAFF CAR-T cells were co-cultured with hcl-c/v cell lines or patient samples at a 5:1 effector: target ratio, and PC7:CD107a was immediately added. Monensin Golgistop was added after an hour, and cells were stained with PE:CD3 after 7 total hours of co-incubation. The percent positive CD3+ cells expressing CD107a was quantified via flow cytometry analysis. ****p < 0.0001. Mean ± SD, n = 3 biologically independent experiments. ****p < 0.0001. Mean ± SD, two-way ANOVA with Holm- šídák’s multiple comparison test. (A). hcl-c/v cell lines or patient samples were co-cultured at a 5:1 effector: target ratio with BAFF CAR T-cells for 16 h, after which cells were stained with hCD3 (PE) and CD69 (PerCP5.5). Percentage of CD3+ cells expressing CD69 was quantified. ****p < 0.0001. Mean ± SD, n = 3 biologically independent experiments. ****p < 0.0001. Mean ± SD, two-way ANOVA with Holm- šídák’s multiple comparison test. (B). At least three T-cell donors were used for CAR-T generation.

Figure 4.

Patient-derived BAFF CAR-T cells induce autologous cancer cell killing and BAFF CAR-T cells release proinflammatory cytokines when co-incubated with HCL cells. Total patient-derived PBMCs were stained with a panel of CD19 (FITC), PE(BCMA), TACI (PCY7), BAFF-R (APC) or CD22 (APC) flow antibodies and receptor expression was measured on flow cytometry and compared to a healthy donor control (a). The expression of BAFF receptors on cancer cells is analyzed by gating on the CD19+ population. T-cells were isolated from the remaining PBMCs, activated, and transduced with the pHR- BAFF CAR vector. The GFP reporter shows percent positive CAR-T cells following transduction and expansion (b). PBMCs from the same patient were co-incubated with control T cells or patient derived BAFF CAR-T cells for 16-24 h. Cells were stained with CD22 (APC) and live CD22+ cells were measured by FSC/SSC gating and live cell percentages are reported. ****p < 0.0001, *** = 0.002 for control T versus CAR-T in 10:1 experiment. Mean ± SD, one-way ANOVA with Tukey’s multiple comparison test. (c). Patient-derived BAFF CAR-T cells were co-incubated with fluorescently labeled hair-M cells (right) or cancer cells from a different patient (left), and live cancer cell percentage was measured by FSC/SSC gating via flow cytometry. ****p < 0.0001. Mean ± SD, one-way ANOVA with Tukey’s multiple comparison test (C). The supernatant from T-cells after co-culture with bonna-12 and hair-M cells for cytotoxicity experiments (figure 1D) was collected after 16 h of incubation. Various pro-inflammatory cytokines (d) and cytotoxic lytic enzymes (e) were measured with a multiplex cytokine release assay. ****p < 0.0001. Mean ± SD, two-way ANOVA with Holm- šídák’s multiple comparison test.

Figure 5.

CAR-T cells reduce HCL tumor burden and prolong overall survival in vivo without toxicity. NSG mice were inoculated subcutaneously with 10 million bonna-12 cells (a). At formation of a palpable tumor of 300 mm³ at day 40, mice were treated intra-tumorally with PBS or 5 million BAFF CAR T-cells. Mice were retreated at day 47, when tumors volumes were approximately 500 mm.³ Mean ± SD, n = 5 mice per group. Two-way ANOVA with Holm-šídák correction for multiple comparisons was applied to tumor volume analysis over time, ****p-value <0.0001. Log rank (mantel-cox) test was applied for survival analysis, p = 0.0011. 10 million bonna-12 cells were injected subcutaneously in the right flank (B). Mice were treated intratumorally with 5 million untransduced control T-cells or BAFF CAR-T cells on days 21 and 39. 2-way ANOVA with Holm-šídák correction for multiple comparisons was applied to tumor volume analysis over time. ****p-value <0.0001, p = 0.049 (day 33), p = 0.0024 (day 36). Log rank (mantel-cox) test was applied for survival analysis, p = 0.0080,.