Treatment of advanced leukemia in mice with mRNA engineered T cells

Abstract

Cytotoxic T lymphocytes (CTLs) modified with chimeric antigen receptors (CARs) for adoptive immunotherapy of hematologic malignancies are effective in preclinical models and are being tested in several clinical trials. Although CTLs bearing stably expressed CARs generated by integrating viral vectors are efficacious and have potential long-term persistence, this mechanism of CAR expression can potentially result in significant toxicity. T cells were electroporated with an optimized in vitro transcribed RNA encoding a CAR against CD19. These RNA CAR CTLs were then tested in vitro and in vivo for efficacy. We found that T cells expressing an anti-CD19 CAR introduced by electroporation with optimized mRNA were potent and specific killers of CD19 target cells. CD19 RNA CAR T cells given to immunodeficient mice bearing xenografted leukemia rapidly migrated to sites of disease and retained significant target-specific lytic activity. Unexpectedly, a single injection of CD19 RNA CAR T cells reduced disease burden within 1 day after administration, resulting in a significant prolongation of survival in an aggressive leukemia xenograft model. The surface expression of the RNA CARs may be titrated, giving T cells with potentially tunable levels of effector functions such as cytokine release and cytotoxicity. RNA CARs are a genetic engineering approach that should not be subject to genotoxicity, and they provide a platform for rapidly optimizing CAR design before proceeding to more costly and laborious stable expression systems.

FIG. 1.

Optimized mRNA electroporation procedure results in uniform high-level surface expression and specific function of CAR T cells in vitro. (A) CAR expression as measured by mean fluorescent intensity (MFI) at different time points after electroporation with CD19-BBz mRNA in anti-CD3 and CD28 stimulated peripheral blood T cells (open histograms). Nonelectroporated T cells were used as negative control (filled histogram). (B) RNA CAR⁺ T cells specifically kill CD19 targets. A flow-based CTL assay was conducted on the indicated day post electroporation with K562-CD19 as target and K562-meso as control (data not shown). (C) Peripheral blood lymphocytes (PBLs) were electroporated with ss1-BBz, CD19-BBz, or no mRNA (Mock). Four hours post electroporation, the T cells were cocultured with K562, NALM-6, or K562 expressing either CD19 (K562-CD19) or mesothelin (K562-meso) and analyzed for CD107a staining. CD3⁺ T cells were gated. Only antigen-specific CD107a expression is observed. (D) Four hours post electroporation, T cells electroporated with mRNA encoding for CD19-BBz or ss1-BBz were cocultured with K562-meso or K562-CD19 target cells for 16 hr. IL-2 production was measured in the supernatant by ELISA, with significant increases in IL-2 production in an antigen-specific manner (*p<0.01). Data are representative of at least two independent experiments.

FIG. 2.

Potentially tunable transgene expression and effector functions of RNA CAR⁺ T cells. (A) Transgene expression of RNA CAR⁺ T cells electroporated with the indicated amounts of CD19-BBz RNA plotted as a function of time. Histograms of transgene expression of electroporated 19-BBz CAR mRNA are shown. (B) Transgene expression data from (A) plotted as a line graph. Rate of decline is similar despite different MFI seen in (A). (C) Specific lysis of CD19⁺ tumor cells with CAR T cells electroporated with the indicated amounts of RNA. Lysis was measured by a flow-cytometric CTL assay using K562-CD19 as targets on day 1 (left panel) and day 3 (right panel) after electroporation. Although little difference exists on day 1, by day 3 a dose-dependent decrease in specific lysis relative to RNA dose is observed. (D) IFN-γ secretion by CAR T cells (4 hr after electroporation) with the indicated amounts of RNA cocultured overnight with serially diluted target cells (K562‐CD19) or control targets (K562-meso at 1:1) assayed by ELISA. IFN-γ secretion titrates with the amount of target, and is not statistically significantly different among CD19 CAR⁺ groups, although a trend toward lower cytokine secretion with lower RNA doses is suggested.

FIG. 3.

Expression and function of RNA CARs in vivo. (A) NOD/SCID/γc^–/– null (NSG) mice were injected with 10⁷ PBLs either intravenously (IV) or intraperitoneally (IP) 4 hr after electroporation with ss1-BBz or CD19-BBz. Mice were sacrificed after 48 hr, and human PBLs were isolated from peripheral blood, spleen, bone marrow, and intraperitoneal washing (IP) by using a T cell negative selection kit (Dynal Magnetic Beads). The purified cells were stained for human CD3 and CAR expression (via a panspecific goat anti-mouse IgG) and analyzed by flow cytometry. Significant background staining of mouse marrow precursor cells is observed in the bone-marrow compartment despite negative selection. CD3⁺CAR⁺ cells are recovered from blood, spleen, and a peritoneal washing, but rarely from the femoral bone marrow at this time point. (B) Purified T cells recovered from intraperitoneal washings 2 days after injection of mice by the IV or IP route with CD19-BBz were used in a flow-based CTL assay. CD19-BBz RNA electroporated T cells that had been cultured in vitro for 2 days (CD19 in vitro) and mock electroporated T cells (No mRNA) were used as controls. The graph shows percent lysis of the purified PBLs against K562-CD19 or K562-meso targets. Target-specific lysis observed in recovered CAR CTLs is comparable to that of in vitro cultured CAR⁺ PBLs and is significantly higher than no mRNA controls (p<0.01).

FIG. 4.

Specific trafficking and proliferation of RNA CARs in tumor-bearing mice. (A) NOD/SCID/γc^–/– (NSG) mice were injected IV with 10⁶ Nalm-6 cells followed 7 days later with 5×10⁶ T cells 4 hr after electroporation with the indicated mRNA constructs. The T cells had been stably transfected with a lentiviral construct to express firefly luciferase, and mice were imaged for bioluminescence. The graph indicates average of individual total photon flux±the standard error for each of the indicated groups (n=8). (B) CD19 RNA CARs exhibit increasing bioluminescence signal and anatomic distribution consistent with migration to sites of disease and CAR T-cell proliferation. Photon density heat maps on day 3 post injection suggest that mock T cells or T cells expressing RNA CARs with irrelevant specificity against mesothelin pool passively in the spleen (left flank on heat map) and do not increase in photon density, indicating a lack of proliferation. Note that the 5×10⁶ cells produce a p/sec/cm² flux of ∼2×10⁷, equivalent among all groups immediately after injection. Saline-treated mice represent the background autoluminescence of 5×10⁵ p/sec/cm².

FIG. 5.

Therapeutic efficacy and specificity of a single injection of RNA CAR⁺ T cells in Nalm-6 xenograft model. (A) NSG mice were injected with 10⁶ Nalm-6 transduced to stably express firefly luciferase, as in Supplementary Fig. S1, followed by a single tail-vein injection of 2.5×10⁷ T cells electroporated with CD19-BBz or meso-BBz mRNA 7 days later (arrow). Animals were imaged at the indicated time points post injection, with total photon flux±SE indicated on the y-axis; 5×10⁵ p/sec/cm²/sr represents mice with no luciferase-containing cells. *p<0.01. (B) Photon-density heat maps of firefly luciferase-positive leukemia in representative mice at day 5 (2 days before treatment) and day 8 (24 hr post CAR⁺ PBLs). Mice start with an equal burden of leukemia, but CD19-directed CAR⁺ PBLs reduce disease burden by 2 logs (but do not eliminate it) as measured by photon density. (C) Survival of those mice treated with CD19-BBz RNA CAR⁺ T cells is significantly prolonged compared with that of saline controls and meso-BBz CAR T cell groups. p<0.01 by log rank analysis. (D) Survival with RNA CAR CTLs compares favorably with that with lentiviral generated CAR CTLs in the same model, although no long-term survivors are noted with a single infusion of RNA CAR CTLs, consistent with our observation that single injection does not entirely eliminate disease (n=12, summation of two independent experiments).