Eradicating Drug-tolerant Persister Cells in EGFR-Mutated Non-Small Cell Lung Cancer by Targeting TROP2 with CAR-T Cellular Therapy

Abstract

EGFR tyrosine kinase inhibitors have dramatically improved outcomes for patients with EGFR-mutated non-small cell lung cancer (NSCLC), but relapse frequently occurs because of drug-tolerant persister (DTP) cells that can evolve and develop diverse mechanisms of drug resistance. In samples from patients with EGFR-mutated NSCLC treated with EGFR tyrosine kinase inhibitors in the neoadjuvant setting, we observed enriched expression of the cell surface protein TROP2, a target of clinically active antibody-drug conjugates (ADC). We confirmed these findings across multiple EGFR-mutated NSCLC cell line and patient-derived xenograft models treated with osimertinib in vivo. Treatment with the TROP2 ADC sacituzumab govitecan at the time of osimertinib-induced minimal residual disease only modestly delayed tumor recurrence in vivo, whereas a single infusion of sacituzumab-based TROP2-directed chimeric antigen receptor (CAR) T cells significantly prolonged relapse-free survival, with evidence of cure. These data highlight the potential of engineering TROP2 CAR T-cell therapy to eliminate EGFR DTPs in patients.

Significance: We provide a rationale for targeting TROP2 in EGFR-mutated NSCLC DTPs. In contrast to TROP2 ADC therapy, targeting of TROP2 with CAR-T cells can eliminate osimertinib-induced DTPs in vivo, revealing the promise of developing novel TROP2-based CAR-T cells to promote durable response and prevent disease relapse in patients.

Figure 1.. Interrogation of ADC target mRNA after osimertinib treatment reveals TROP2 as a top hit in EGFR-mutant NSCLC DTPs.

A. Schematic of the clinical scenario in patients with locally-advanced EGFR-mutant NSCLC, receiving neo-adjuvant EGFR-TKI treatment. The surgical samples underwent single-cell RNA sequencing (scRNA-Seq) and immunohistochemistry (IHC) for evaluating RNA and protein expression of ADC targets, respectively. Radiological response to neo-adjuvant osimertinib in Patient #1 shown in bottom portion of figure with CT chest and PET CT representative images shown. B. The scRNA-Seq of Patient #1 surgical sample allowed the identification and grouping of cell type clusters. This approach revealed TACSTD2 mRNA overexpression in the tumor cell population (corresponding to DTP) compared to non-malignant cells. ERBB3 RNA is expressed at low levels in cancer cells. CD70 RNA was not found overexpressed in tumor cells, but clustered with B cells. C. tSNE projection of single cells from EGFR-mutant NSCLC patients reported in Maynard et al.(25) colored by treatment timepoint (TKI naive (TN): red and residual disease (RD): blue). Middle: Violin plot comparing TACSTD2 transcriptional expression levels between TN and RD samples. Right: Violin plot of ERBB3 transcriptional expression levels between TN and RD samples. D. Growth curves of PC9 (n=9 mice in vehicle treated group and 6 in osimertinib (10mg/kg qd po) treated group) and DFCI-282 (n=6 mice in vehicle treated group and n=10 in osimertinib (10mg/kg qd po) treated group) xenografts harvested for scRNA-Sseq (3 mice per group) and immunofluorescence (3 mice per group). E. Differential transcriptional expression measured by scRNA-Seq in osimertinib induced DTPs and parental cell lines in vivo. F. scRNA-Seq expression of TACSTD2 in osimertinib (osi) vs vehicle (veh) treated mice from (1D). G. RT-qPCR analysis of TACSTD2 mRNA expression in EGFR-mutated NSCLC cell lines treated as indicated in vitro, (n=3 independent experiments). For (1D) graphs, data are mean ± SEM.

Figure 2.. TROP2 protein is enriched in DTPs induced by osimertinib in EGFR-mutated NSCLC cell lines and PDX.

A. TROP2 expression is maintained at the protein level assessed by IHC (represented as H-score, percentage of tumor cells for intensity evaluated in a 0–3 range) in the tumor cells of Patient #1 sample, and in three additional surgical samples from patients treated with pre-operative osimertinib for advanced disease. Representative images of each case are shown. Scale bars, 50 μm. B. Immunofluorescence staining for TROP2 (green) in xenograft tumors from the indicated cell lines following vehicle or osimertinib (osi) treatment (tumor samples from in vivo experiment represented in (1D) and Supplementary Figure S5A and S5B). Below, fluorescence intensity quantification of TROP2 membranous staining. C. Western blot analysis of TROP2 total expression in EGFR-mutated NSCLC cell lines treated as indicated in vitro, (n=3 independent experiments). D. Schematic of in vivo experiment with EGFR-mutant NSCLC PDX DFCI-243 treated either with vehicle, or with osimertinib for 42 days or with osimertinib for 42 days and sacituzumab govitecan (SG). Spider plots demonstrating tumor volume (mm³) of DFCI-243 PDX vs time for vehicle group (n=8 mice), osimertinib (10 mg/kg qd po) for 42 days group (n=10 mice) and osimertinib (10 mg/kg qd po) for 42 days and sacituzumab govitecan (0.5 mg/mouse q2w iv) from day 22 to 42 group (n=10 mice). The green dotted lines indicate sacituzumab govitecan injection and the beige area indicates the osimertinib or vehicle treatment phase. E. Right panel: relapse-free survival (RFS) curves. Relapse was defined as when tumors reached 400 mm³. In (2B) Mean ± SD is shown and Mann-Whitney test was used for statistical analyses. In (2E), median relapsed free survival and statistical analysis were determined via Log Rank Mantel Cox analysis. **p<0.01, ****p < 0.0001.

Figure 3.. TROP2 can be targeted by CAR-T cells in osimertinib induced DTPs.

A. Schematic of second generation lentiviral CAR-T vector with single-chain variable fragment (scFv) derived from sacituzumab with CD28 transmembrane domain (CD28TM), 41BB costimulatory domain and CD3z signaling domain. B. Cell surface TROP2 expression assessed by live-cell flow cytometry in PC9, PC9 CTL sgRNA clone, and PC9 TROP2 sgRNA clone. C. In vitro cytotoxicity of TROP2 CAR-T cells against PC9, PC9 CTL sgRNA clone, and PC9 TROP2 sgRNA clone, stably expressing luciferase after lentiviral transduction, (n=3 independent experiments). Cells were exposed to CAR-T at indicated E:T ratio for 24 h. Viability plotted as relative luciferase unit signals compared to BCMA control at a given E:T ratio. D. In vitro cytotoxicity against DTPs derived from PC9, PC9 CTL sgRNA clone, and PC9 TROP2 sgRNA clone stably expressing luciferase with TROP2 CAR-T cells combined with osimertinib 100nM. DTPs were derived from the indicated cellular model after treatment for 10 days with 100 nM osimertinib followed by co-incubation with CAR-T at indicated E:T ratio for 24h. Viability plotted as relative luciferase unit signals compared to BCMA control at a given E:T ratio. E. Schematic and results of in vitro apoptosis assay in osimertinib DTPs derived from PC9, PC9 CTL sgRNA clone, and PC9 TROP2 sgRNA clone stably expressing Histone H2B fused with mCherry after lentiviral transduction. DTPs were derived from the indicated cellular model after treatment for 10 days with 100 nM osimertinib followed by co-incubation with CAR-T at indicated E:T ratio for 96h, (figure representative of 3 independent experiments). Apoptosis was measured via incucyte live cell imaging analysis of Caspase 3/7 measured with green dye. F. Representative images of (E). Mean ± SD are shown in all plots except (E) where mean ± SEM are shown. One-way ANOVA was used in (C) and (D) and two-way ANOVA was used for statistical analyses in (E). **p<0.01, ***p<0.001, ****p < 0.0001; ns: non significant.

Figure 4.. TROP2 CAR-T generate durable responses in EGFR-mutant PDX model of DTP state.

A. Schematic of in vivo model with EGFR-mutant NSCLC xenograft with osimertinib alone group, osimertinib and sacituzumab govitecan (SG) group, osimertinib and TROP2 CAR-T group, vehicle and SG group, vehicle and TROP2 CAR-T group. B. Mean tumor volume vs time for HCC4006 xenograft treated with indicated regimen. Mice received either vehicle or osimertinib via oral gavage at 10 mg/kg daily until day 42 and depending on treatment group, mice could also receive at day 22 CAR-T at a dose of 0.6 10⁶ CAR-T cells IV or started on SG at 0.5 mg iv biweekly x 6 doses. Standard error of the mean shown. C. Spider plots demonstrating tumor volume of DFCI-243 PDX vs time for osimertinib alone group, osimertinib and SG group and osimertinib and TROP2 CAR-T group. Groups correspond to similar figure key in (4B). D. Relapse free survival of groups shown in (4C) as well as median relapsed free survival. BCMA CAR-T control groups is also shown. Groups correspond to similar figure key in (4B). E. Hematoxylin and eosin staining and IHC of TROP2 of relapsed mice from the indicated treatment group of experiment shown in (4C). F. Tumor volumes of TROP2 relapsed DFCI-243 PDX after TROP2 CAR-T treatment with either sacituzumab govitecan or TROP2 CAR-T. Top graph with circle indicates the initial relapsed tumor from Figure 4B which was isolated and reimplanted into NSG mice (n=3 per group) and when tumors reached ~200 mm³ mice were treated with sacituzumab govitecan 0.5 mg iv biweekly x 6 doses or 10⁶ TROP2 CAR-T iv at day 1. Two-way ANOVA was used in (4B) and Log Rank Mantel Cox analysis was used for statistical analysis in (4C). **p<0.01, ****p < 0.0001.