Restricting CAR T Cell Trafficking Expands Targetable Antigen Space

Abstract

Chimeric antigen receptor (CAR) T cells are an effective treatment for some blood cancers. However, the lack of tumor-specific surface antigens limits their wider use. We identified a set of surface antigens that are limited in their expression to cancer and the central nervous system (CNS). We developed CAR T cells against one of these antigens, LINGO1, which is widely expressed in Ewing sarcoma (ES). To prevent CNS targeting, we engineered LINGO1 CAR T cells lacking integrin α₄ (A4^ko), an adhesion molecule essential for migration across the blood-brain barrier. A4^ko LINGO1 CAR T cells were efficiently excluded from the CNS but retained efficacy against ES. We show that altering adhesion behavior expands the set of surface antigens targetable by CAR T cells.

Figure 1:. Tumor-associated single-tissue surface antigens are enriched in the CNS and can be targeted using CAR T cells.

(A) Flowchart outlining process for the identification of single-tissue surface antigens using consensus normal tissue and cancer cell line RNA sequencing (RNAseq) data. (B) Frequency of single-tissue surface and non-surface antigens expressed in different normal tissues. (C) List of CNS-restricted STSAs with highest RNA expression levels in cancer. (D) RNA expression of LINGO1 in osteosarcoma (OS) and Ewing sarcoma (ES) cell lines as determined by RNAseq. Data represent mean ± S.D. from 6 cell lines per diagnosis. Statistical significance was determined by Mann-Whitney U test. (E) Expression of LINGO1 in primary human ES tumor biopsies, normal tissues, bone marrow mesenchymal stem cells (BMMSC), neural crest mesenchymal stem cells (NCMSC), and neural crest stem cells (NSCS) as determined by HuEx microarray (20). Data represent mean ± S.D. from indicated number of patients or tissues. Statistical significance was determined by one-way ANOVA. (F) LINGO1 expression in normal human and mouse tissues as determined by RNAseq. RPKM = reads per kilobase per million mapped reads. (G) LINGO1 expression in human and mouse brain regions as determined by RNAseq. (H) Schema of LINGO1 CAR and scFv constructs. (I) Surface expression of LINGO1 on ES cell lines using single-chain variable fragments derived from monoclonal LINGO1 antibody clones Li62 and Li81 as determined by flow cytometry. (J) IFN-03B3 secretion and (K) target cell killing by LINGO1 CAR T cell variants after 16-hour incubation with ES cell line A673 expressing firefly luciferase at an effector-target ratio of 1:1. Data represent mean ± S.D. from 3 technical replicates. Statistical significance was determined by one-way ANOVA. (L) Expansion of LINGO1 CAR T cells with and without overexpression of c-Jun during manufacturing. Data represent mean ± S.D. from 3 donors. Statistical significance was determined by two-sided Student’s t test. (M) Expression of transcription factors c-Jun and BATF3 in LINGO1 CAR T cells with and without overexpression of c-Jun as determined by western blot. (N) Killing of K562 cells engineered to express CD19, human LINGO1 (huLINGO1), or mouse LINGO1 (msLINGO1) following 16hour co-culture with c-Jun (CJ)-expressing T cells expressing an Li81-based LINGO1 CAR or a control CAR without a binding domain (ΔscFV). Data represent mean ± S.D. from 3 technical replicates.

Figure 2:. LINGO1 CAR T cells efficiently eradicate ES in vitro and in vivo.

(A) CAR surface and GFP reporter expression in primary human T cells at the end of gammaretroviral manufacturing. Data are representative of 4 independent CAR T cell productions. (B) Killing of 3 ES cell lines after 16-hour co-culture with T cells expressing LINGO1 CAR or ΔscFv CAR at the indicated effector-target ratios as determined by luciferase-based cytotoxicity assay. Data represent mean ± S.D. from 3 technical replicates. (C) Cytokine concentrations in co-culture supernatants after 16-hour as determined by ELISA. Data represent mean ± S.D. from 3 technical replicates. (D) Schema of in vivo experiment to determine efficacy of LINGO1 CAR T cells. (E) Tumor burden in mice bearing systemic A673 tumors and treated with LINGO1 CAR T cells or ΔscFv control CAR T cells as determined by in vivo imaging system (IVIS). (F) Survival of ES tumor bearing mice treated with LINGO1 CAR T cells. Statistical significance was determined by log-rank test.

Figure 3:. Integrin α-4 knockout in human T cells prevents adhesion to VCAM-1 and does not alter T cell expansion or phenotype.

(A) Schema of integrin α₄-mediated T cell attachment and extravasation. (B) Schema of integrins and their respective heterodimerization partners. (C) Knockout efficiency of integrin α₄ and β₁ in Jurkat cells transduced with a lentiviral construct expressing Cas9 and non-specific (WT), or integrin α₄ or integrin β₁-targeting gRNA constructs as determined by flow cytometry after cell sorting. Unstained control in grey. (D) Adhesion of Jurkat-Fluc cells to recombinant VCAM-1 following integrin α₄ or integrin β₁ knockout as determined by luminescence. Data indicate mean ± S.D. from 2 technical replicates. Statistical significance was determined by two-way ANOVA. (E) Integrin α₄ knockout efficiency in primary human T cells after electroporation with CRISPR/Cas9 RNP as determined by flow cytometry. (F) Expansion of stimulated T cells following integrin α₄ knockout. Data represent mean ± S.D. from 3 different donors. (G) T cell phenotype 10 days after integrin α₄ knockout as determined by flow cytometry.

Figure 4:. A4^ko LINGO1 CAR T cells can be manufactured efficiently and maintain anti-tumor activity against ES.

(A) Schema of A4^ko LINGO1 CAR T cell manufacturing workflow. (B) Integrin α₄ knockout and CAR transduction efficiency in 3 donors as determined by flow cytometry. (C) Expression of T cell markers in WT and A4^ko LINGO1 CAR T cells from 3 healthy donors as determined by spectral flow cytometry. Data were normalized to the mean of the WT condition. Statistical significance of differences between WT and A4^ko LINGO1 CAR T cells was determined by two-way ANOVA. (D) Expansion of WT or A4^ko LINGO1 CAR T cells as determined by automated cell counting. (E) Schematic showing location of distal (d), peripheral (p), and central (c) supramolecular activation complex (SMAC) in classical and non-classical/CAR immune synapses (IS). (F) Killing of 2 ES cell lines after 16-hour co-culture with T cells expressing LINGO1 CAR or ΔscFv CAR at the indicated effector-target ratios as determined by luminescence cytotoxicity assay. Data represent mean ± S.D. from 3 technical replicates. (G) Cytokine concentrations in supernatants of WT or A4^ko LINGO1 CAR T cells co-cultured with A673 cells after 16 hours as determined by CodePlex assay and (H) ELISA. (I) Schema of streptavidin (SA)-bead-based approach for the depletion of remaining integrin α₄^pos LINGO1 CAR T cells. (J) Quantification of integrin α−4^pos LINGO1 CAR T cell following bead-based depletion as determined by flow cytometry. (K) Killing of A673 cells with or without depletion of integrin α₄^pos LINGO1 CAR T cells as determined by luminescence-based cytotoxicity assay. Data represent mean ± S.D. from 3 technical replicates.

Figure 5:. Integrin α₄ knockout prevents trafficking of CAR T cells to the CNS but not to ES tumors.

(A) Schema of in vivo experiment to determine tissue trafficking of primary human T cells expressing firefly luciferase (Fluc). (B) Quantification of relative luminescence units (RLU) and representative example images of organs from one animal of explanted organs normalized to the mean signal of WT LINGO1 CAR T cells. Data represent mean ± S.D. from 4 animals. Statistical significance was determined by Mann-Whitney U test. (C) Example of luminescence observed in tissues from mice injected with A4^ko T cells. (D) Schema of integrin α₄β₁ activation status in Jurkat and primary T cells. (E) Schema of in vivo experiment to determine CNS trafficking of A4^ko Jurkat cells. (F) In vivo imaging of mice injected with A4^ko Jurkat cells. Dotted lines indicate head and lower body regions of interest (ROI) used for luminescence quantification. (G) Quantification of luminescence signal of indicated ROIs from whole body imaging of mice injected with A4^ko Jurkat cells. Data represent mean ± S.D. from 5 animals per group. Statistical significance was determined by two-sided Student’s t test. (H) Quantification and images of luminescence signal measured in explanted brains from mice injected with WT of A4^ko Jurkat cells. Data represent mean ± S.D. from 5 animals per group. Statistical significance was determined by Mann-Whitney U test. (I) Number of T cells isolated from brains of mice injected with WT or A4^ko Jurkat cells. Data represent mean ± S.D. from 5 animals per group. Statistical significance was determined by two-way ANOVA. (J) Schema of orthotopic in vivo experiment to determine efficacy of A4^ko LINGO1 CAR T cells against ES. (K) Tumor burden in 4 mice per group bearing intratibial A673 tumors and treated with intravenous A4^ko LINGO1 CAR T cells or ΔscFv control CAR T cells as determined by IVIS. (L) Survival of ES tumor-bearing mice treated with LINGO1 CAR T cells. Statistical significance was determined by log-rank test.