Quantifying requirements for mitochondrial apoptosis in CAR T killing of cancer cells

Abstract

Chimeric antigen receptor (CAR) T cell therapy is an FDA-approved treatment for several hematologic malignancies, yet not all patients respond to this treatment. While some resistance mechanisms have been identified, cell death pathways in target cancer cells remain underexplored. Impairing mitochondrial apoptosis via knockout of Bak and Bax, forced Bcl-2 and Bcl-XL expression, or caspase inhibition protected several tumor models from CAR T killing. However, impairing mitochondrial apoptosis in two liquid tumor cell lines did not protect target cells from CAR T killing. We found that whether a cell was Type I or Type II in response to death ligands explained the divergence of these results, so that mitochondrial apoptosis was dispensable for CART killing of cells that were Type I but not Type II. This suggests that the apoptotic signaling induced by CAR T cells bears important similarities to that induced by drugs. Combinations of drug and CAR T therapies will therefore require tailoring to the specific cell death pathways activated by CAR T cells in different types of cancer cells.

Fig. 1. Loss of Bak and Bax is associated with resistance to granzyme B cytotoxicity and CAR T killing.

A Annexin V / PI staining of vector and Bak/Bax null pre-B cell lines following exposure to perforin and/or granzyme B. N = 2, each point is a biological replicate. Significance corresponds to unpaired t test, **p = 0.003. B Quantification of CD19 staining intensity via flow cytometry in two lymphoma cell lines, DHL4 and BBDL. C Immunoblotting for Bak and Bax, with actin included as a loading control. D BH3 profiling of DHL4 and BBDL cell lines, N = 2, each point is a biological replicate. E Annexin V viability assay of DHL4 and BBDL cells 18 h after co-culture with CD19 CAR T cells. N = 2, each point is a biological replicate. Significance corresponds to unpaired t test, ****p < 0.0001.

Fig. 2. Loss of Bak and Bax confers resistance to CAR T killing of HeLa target cells.

A Annexin V / Hoechst viability staining of HeLa-19 and HeLa-DKO-19 cells following 48 hour exposure to etoposide, N = 3. B BH3 profiling assay comparing mitochondrial response to the Bim peptide in HeLa-19 (N = 2) and HeLa-DKO-19 (N = 1) cell lines. Figure C–F depicts target cell viability following exposure to CD19 CAR T cells assessed via annexin V/Hoechst staining at 20 h (C, D), an impedance assay (E, N = 2), and a colony formation assay (F, N = 1). C presents data from 4 biological replicates using a single donor’s CAR T cells. Curves were fitted with non-linear regression, error bars are technical replicates. All biological replicates are shown in Fig. S2. The left panel of D presents the E:T 50 value for every paired co-culture experiment performed (N = 20 for 28z, N = 12 for BBz, 9 donors total). The right panel presents the mean difference in E:T50 (DKO-HeLA, middle bar) with the 95% confidence interval. For all graphs except panel C, each point is a biological replicate. Significance corresponds to either an unpaired (A, E) or paired (D) t test, ****p < 0.0001, **p = 0.0015.

Fig. 3. Forced expression of Bcl-2 and Bcl-XL confers resistance to CAR T killing of HeLa target cells.

A Immunoblotting for Bcl-2, Bcl-XL, or actin (loading control) following transfection of HeLa-19 cells. B Annexin V / Hoechst viability staining at 20 h following etoposide treatment of each transfected population (N = 3). C Left panel: Annexin V / Hoechst viability staining of each cell population following 20 h of CAR T co-culture (N = 3). Right: estimated E:T 50 values −/+ 95% confidence interval for each curve, with significance calculated from unpaired t test, Bonferroni corrected for three comparisons. D, E Annexin V / Hoechst staining of HeLa-19 (D, N = 4) and HeLa-DKO-19 (E, N = 2) target cells following CAR T co-culture −/+ caspase inhibition with Z-VAD-FMK. Each point is a biological replicate, significance calculated from unpaired t test, ****p < 0.0001, **p = 0.0013.

Fig. 4. CAR T killing of HeLa target cells is mediated by granzymes, TNFα, and IFNγ.

A, B Annexin V/Hoechst viability staining of HeLa-19 and HeLa-DKO-19 cells following 24 h (A, HeLa N = 5, DKO N = 2) or 48 h (B, HeLa N = 2, DKO N = 1) of treatment with the indicated death ligand or cytokine. C, D Results of Fas, TRAIL, and TNFα blockade in CAR T co-culture experiments. For C N = 3, for D Control N = 3, Anti-TNFα N = 2, Anti-IFNγ N = 2, Anti-TNFα + IFNγ N = 1. The estimated E:T50 values and 95% confidence intervals are presented in E. In F, target cells were co-cultured with CD19 CAR T cells pre-treated with 3,4 dichloroisocoumarin (DCI), N = 6. Viability was assessed by Annexin V / Hoechst staining at 24 h. Each point is a biological replicate, significance corresponds to unpaired t test, **p = 0.0066, ***p = 0.0008, ****p < 0.0001.

Fig. 5. CAR T killing of HeLa target cells is mediated by soluble factors including TNFα.

A, B Conditioned media was collected as described in the text and incubated with HeLa-19 or HeLa-DKO-19 target cells for 24 h. For A, HeLa N = 6, DKO N = 3; in B HeLa N = 4, DKO N = 1. In C, conditioned media from two paired co-cultures (HeLa-19 or HeLa-DKO-19 with 28z CAR T cells) was incubated with HeLa-19 or HeLa-DKO-19 target cells for 24 h. N = 1. Figures D and E present Cytokine Array data from conditioned media with intensity normalized to the included positive and negative controls. Figure D highlights cytokines that were unique to HeLa-19 co-culture with a specific CAR by subtracting the cytokines present in HeLa-19/Δζ co-culture (N = 1). Figure E presents results from all three cytokine arrays performed, highlighting the top cytokines noted in Figure D. F Annexin V / Hoechst viability assay performed 24 h after HeLa-19 incubation with conditioned media and the indicated blocking antibodies, N = 3. Each point represents a biological replicate, significance corresponds to unpaired t test, **p = 0.0017.

Fig. 6. Loss of Bak and Bax confers resistance to CAR T killing in solid tumor target cells.

A EGFR expression quantified by flow cytometry for paired WT and DKO HCT-116 and HeLa cell lines. Nalm6 included as negative control, Iso = isotype control. Panels B–E depict target cell viability following exposure to EGFR CAR T cells assessed via annexin V/Hoechst staining at 20 h. B, D Pooled data from biological replicates using a single donor’s CAR T cells, with a curve fit using non-linear regression. In C and E, curves were fit for each biological replicate performed, and E:T 50 values are presented for each wildtype/DKO pair (total of 3 donors and 5 replicates). Each point is a biological replicate, significance corresponds to either an unpaired (B, D) or paired (C, E) t test. *p = 0.0156, ***p = 0.0008, ****p < 0.0001.

Fig. 7. Loss of Bak and Bax confers minimal resistance to CAR T killing in lymphoma target cells.

A Sensitivity of JeKo-1 variants to co-culture with CD19 CAR T cells (Vector N = 5, Bcl-2 N = 5, Bcl-XL N = 3) assessed by annexin V/Hoechst staining at 20 h. B Viability of JeKo-1 variants following 24 h of TRAIL treatment (N = 2). C Sensitivity of Nalm-6 variants to co-culture with CD19 BBz CAR T cells (N = 3, single donor) assessed by annexin V / Hoechst staining at 20 h. D Summary statistics for the biological replicates shown in C as well as for co-culture with 28z CD19 CAR T cells. E Viability of Nalm-6 variants following 24 hours of TRAIL treatment (N = 2). Each point is a biological replicate, significance corresponds to an unpaired (A, B, D, E) t test. **p = 0.0013 in B, =0.0017 in D, ***p = 0.009, ****p < 0.0001.