Rejuvenated iPSC-derived GD2-directed CART Cells Harbor Robust Cytotoxicity Against Small Cell Lung Cancer

Abstract Small cell lung cancer (SCLC) is exceptionally aggressive, with limited treatment options. Disialoganglioside (GD2) is highly expressed on SCLC and is considered a good target for chimeric antigen receptor (CAR) T cells (CART). Although GD2-directed CARTs (GD2-CART) exhibit cytotoxicity against various GD2-expressing tumors, they lack significant cytotoxicity against SCLC. To enhance cytotoxicity of GD2-CARTs against SCLC, we introduced GD2-CAR into induced pluripotent stem cells (iPSC)-derived rejuvenated cytotoxic T lymphocytes (GD2-CARrejT). GD2-CARrejTs acted much more strongly against SCLC cells than did GD2-CARTs both in vitro and in vivo. Single-cell RNA sequencing elucidated that levels of expression of TIGIT were significantly lower and levels of expression of genes associated with cytotoxicity were significantly higher in GD2-CARrejTs than those in GD2-CARTs. Dual blockade of TIGIT and programmed death-1 (PD-1) increased the cytotoxicity of GD2-CARTs to some extent, suggesting that low TIGIT and PD-1 expression by GD2-CARrejTs is a major factor required for robust cytotoxicity against SCLC. Not only for robust cytotoxicity but also for availability as “off-the-shelf” T-cell therapy, iPSC-derived GD2-CARrejTs are a promising novel treatment for SCLC. Significance: This research introduces iPSC-derived rejuvenated GD2-CARTs (GD2-CARrejT) as a novel approach to combat SCLC. Compared with conventional GD2-CARTs, GD2-CARrejTs with reduced TIGIT and PD-1 expression demonstrate robust cytotoxicity against SCLC and would be a promising therapy for SCLC.

The experimental protocols were approved by the Research Ethics Committee for the Faculty of Medicine, Juntendo University (Bunkyo-ku, Tokyo, Japan), and were in accordance with the Declaration of Helsinki.Written informed consent was obtained from healthy donors and patients.

Flow Cytometry
Flow cytometry was performed on BD FACSAria II (RRID:SCR_018934) or BD LSRFortessa equipment (RRID: SCR_023638, BD Biosciences).The acquired data were analyzed using FlowJo software 10.5.3 (RRID:SCR_008520; Tree Star).Propidium iodide or 7-aminoactinomycin D (catalog no.P1304MP) was used to gate in live cells in all analyses.A fluorescence-minus-one technique was used to interpret flow cytometry data in all antibody combinations.
We also evaluated the antitumor efficacy of GD2-2840z-CARTs with anti-PD-1 antibody (catalog no.BE0146, InVivoMAb anti-human PD-1, BioXCell) and anti-TIGIT antibody (Tiragolumab, Seleck Biotech) against SCLC using RTCA.The samples were divided into four groups: a target only group, single blockade by anti-PD-1 antibody group, single blockade by anti-TIGIT antibody group, and dual blockade by anti-PD-1 and TIGIT antibodies group.H446 cells (3 × 10 4 cells per well) were seeded and cultured for 18.5 hours in microelectrodecoated 96-well plates to permit target cell adherence.Anti-PD-1 antibody was added at a final concentration of 20 μg/mL on coculture days −4, −1, and 0, while anti-TIGIT antibody was added at a final concentration of 20 μg/mL on coculture days −1 and 0. Effector T cells were then added (E:T ratio, 1:2).Electrical impedance changes were recorded automatically and continuously as CI by RTCA every 15 minutes for up to 100 hours.

Antitumor Activity, In Vivo Model
All in vivo studies were approved by the Animal Research Committees of Juntendo University School of Medicine (Tokyo, Japan).To verify the tumorsuppressive effect of GD2-CARrejTs against SCLC, the H446 SCLC cell line was transduced with a γ-retroviral vector encoding a fusion protein composed of GFP and FFluc.GFP/FFluc H446 cells (1 × 10 6 cells/mouse) were injected intravenously into 6-week-old female nonobese diabetic/Shi-severe combined immunodeficiency, IL2Rγ knockout Jic (NOG) mice (In-Vivo Science) and tumor growth was monitored using the Caliper in vivo imaging system (Caliper Life Sciences).Firefly d-luciferin substrate (OZ Biosciences) was intraperitoneally injected into mice 15 minutes before imaging.Mice were divided into three groups: no-treatment group, GD2-2840z-CART therapy treated group, and GD2-CARrejT treated group.Four days after tumor inoculation, mice were intravenously treated with GD2-2840z-CARTs or GD2-CARrejTs (1.0 × 10 7 cells).Living Image software version 4.7.2 (RRID:SCR_014247; PerkinElmer) was used for luminescence analyses.The intensity of signal was measured as total photon/s/cm 2 /steradian (p/s/cm 2 /sr).

qRT-PCR
Genomic DNA was isolated from samples of mouse whole blood and from cultured GD2-CARTs using a QIAamp DNA blood mini kit (catalog no.51183, QIAGEN).GD2-CAR vector copies were quantified using the StepOnePlus real-time PCR system (RRID:SCR_015805; Applied Biosystems).Primers and probes for the GD2-CAR transgene and GAPDH were custom-ordered (Applied Biosystems) as described previously (34).Individual PCR reaction results were normalized against GAPDH levels.Copies of transgene/μg of DNA were quantified as copies calculated from GD2 standard curve/input DNA (ng) correction factor (ng detected/ng input) 1,000 ng, as described previously (35).
Table 1 provides primer and probe sequences.

IHC Staining and Image Capture
Mouse tissue samples were fixed in phosphate-buffered aqueous 4% paraformaldehyde solution (catalog no.161-20141, Wako Pure Chemicals).

Primer and probe Sequence
Forward 5 -GCTGCACCAACTGTATCCATCTT-3 Reverse 5 -GGTCCAGACTGCTGAAGCT-3 Probe 5 -CACCCGACCCACCACC-3 Sections of paraffin-embedded tissues, picked up on glass slides, were stained with hematoxylin and eosin for histopathologic examination.Anti-human CD3 rabbit mAb (SP7; 1:50 dilution; ab16669, Abcam, RRID:AB_443425) was used for immunostaining (29,(36)(37)(38).Tissue sections were deparaffinized with xylene and ethanol and then rehydrated with H 2 O.To recover antigen, slides were heated in a microwave oven at 95°C for 20 minutes in ethylenediaminetetraacetic acid aqueous buffer solution (pH 8.0).After cooling and washing with buffer solution, the slides were treated using a Dako autostainer (Dakocytomation).Endogenous peroxidase activity was blocked by 5 minutes' incubation in 3% H 2 O 2 .The slides were incubated with antibody at room temperature in a humidified chamber for 30 minutes and then incubated with the Dako REAL EnVision Detection System horseradish peroxidase-conjugated anti-rabbit secondary antibody (K4003; Dakocytomation) at room temperature for 30 minutes.Reaction products were visualized by treatment with diaminobenzidine chromogen (Dakocytomation) for 5 minutes.Immunohistopathologic images were obtained using a BX53 microscope (Olympus).

Library Preparation for Single-cell RNA Sequencing and Cellular Indexing of Transcriptomes and Epitopes by Sequencing
Approximately 1 × 10 6 sample cells (in vitro cultured GD2-2840z-CARTs and GD2-CARrejTs, obtained 14 days after stimulation with 5 μg/mL PHA-L (Sigma-Aldrich) mixed with irradiated PBMCs in NS-A2 CTL medium (Nissui) supplemented with 1% PSG in the presence of 10 ng/mL IL7 and 10 ng/mL IL15 (both Miltenyi Biotec), were resuspended with 50 μL of chilled PBS + 0.04% BSA (Thermo Fisher Scientific).CAR transgene expression in GD2-CARTs uniformly was approximately 80%, as CAR-positive T cells were sorted before using the assay.Samples were labeled with TotalSeq-B Human Universal Cocktail, V1.0 (RRID: AB_2892472, BioLegend), an antibody cocktail conjugated with a unique identifier in the form of a barcode and used for tracking expression profiles.
After staining, cells were washed four times to remove residual antibodies.The labeled cells were then loaded onto the 10x Chromium system with the Chromium Next Gem Single Cell 3 Reagent Kits v 3.1 (Dual Index; catalog no.PN-1000128, 10x Genomics).Libraries generated from cell surface protein labeling with TotalSeq-B antibodies were sequenced (Macrogen).The Shirokane supercomputer system and Cell Ranger (RRID:SCR_017344, version 7.0.0,10x Genomics) were used for demultiplexing and alignment (human reference genome GRCh38).Downstream analyses used Seurat (v4.0; ref. 39).Doublet Finder was used to remove doublets (40) and samples were filtered for cells with < 15% mitochondrial gene expression in which >200 genes were detected per cell.Samples were normalized using Normalize-Data and ScaleData followed by integration of samples with clustering and Uniform Manifold Approximation and Projection (UMAP) analysis using

Statistical Analysis
All data are presented as mean ± SD or SEM.Results were analyzed by unpaired Student t test (two-tailed) or ANOVA as stated, with a P value <0.05 indicating statistical significance.Survival curves were compared using Kaplan-Meier analysis with log-rank testing.Software for all statistical analyses was Excel (Microsoft) and Prism 9.0 (RRID:SCR_002798; GraphPad Software).

Data Availability
The single-cell RNA sequencing (scRNA-seq) data of the GD2-2840z-CARTs and GD2-CARrejTs are deposited in the Gene Expression Omnibus database with accession number GSE245955.
These results established that GD2-2840z-CARTs showed killing activity against GD2-expressing tumor cells without SCLC cells and that cytotoxicities of GD2-2840z-CARTs against these six cell lines did not directly correlate with GD2 expression levels.
In addition, expression of genes encoding transcription factors known to be cytotoxic effectors (TBX and RUNX) and to mark tissue resident memory T cells (ZNF and ZBTB) were high in GD2-CARrejTs (Fig. 4H).
We also compared expression levels of these genes in GD2-2840z-CARTs, GD2-CARrejTs, and healthy donor-derived CD3 + T cells as control T cells.Control T cells expressed genes associated with exhaustion markers (such as TIGIT and LAG3) and activation marker CD226 at low levels (Supplementary Fig. S3A).
These scRNA-seq analyses indicated that GD2-CARrejTs more strongly expressed genes associated with cytotoxicity, the immunologic synapse, and transcription factors than did GD2-2840z-CARTs.Furthermore, expression levels of genes associated with T-cell exhaustion, especially TIGIT and LAG, were clearly lower in GD2-CARrejTs than in GD2-2840z-CARTs.

Discussion
Constitutive phosphorylation of the GD2-CAR CD3-ζ domain due to CAR clustering in an antigen-independent manner reportedly leads to early exhaustion of CART cells, an obstacle to their clinical deployment (43).In this study, GD2-CARTs expressed exhaustion markers (LAG-3, PD-1, and TIM-3) only 12 days after GD2-CAR transduction (Fig. 1D).Furthermore, scRNA-seq revealed that GD2-2840z-CARTs expressed genes associated with exhaustion at levels significantly higher than those of GD2-CARrejTs (Fig. 4G and H).
Because expression of TIGIT was clearly lower on scRNA-seq in GD2-CARrejTs than in GD2-2840z-CARTs, we focused on TIGIT as a potential contributor to inefficacy of GD2-2840z-CARTs against SCLC.TIGIT is an immunoglobulin superfamily member expressed on T cells and natural killer cells.TIGIT on CD8 T cells regulates costimulatory receptor CD226 through sharing with it the ligand CD155.TIGIT binds more strongly to CD155 than does CD226.(54).Therefore, GD2-CARrejTs that express both TIGIT and PD-1 at low levels can sustain cytotoxic activity with diminished susceptibility to exhaustion, a feature that holds therapeutic promise against SCLC.
Furthermore, GD2-CARrejTs with higher proportion of stem cell memory T cells (Fig. 2E), produced perforin, IFNγ, TNF, IL2, IL4, and IL10 at higher levels than did GD2-CARTs, leading to an advantage in GD2-CARrejTs against SCLC.This high cytotoxic potential of GD2-CARrejTs could be explained by differences between original cells; GD2-CARrejTs were derived from LMP2-specific CTLs, but GD2-2840z-CARTs were derived from bulk activated T cells.
scRNA-seq analysis revealed that GD2-CARrejTs expressed genes associated with cytotoxicity (PRF, GZMB, IFNG, NKG, LTB, LTA, GNLY, and LAMP; Fig. 4D and H) and genes associated with the immunologic synapse at levels higher than those of GD2-2840z-CARTs (Fig. 4E and H).LFA-1 (CD11a/CD18) is activated by CD226 binding to CD155 and activated LFA-1 allows binding to ICAM-1 with high affinity, enhancing cytotoxicity by forming an immunological synapse with ICAM-1 (55).Therefore, examination of tumor infiltration markers such as LFA-1 and ICAM-1 could be important.We infer that potent cytotoxic potential and robust immunologic synapse formation in GD2-CARrejTs may contribute to effective and elimination of SCLC.
The greatest advantage of iPSC-derived T-cell therapy is the unlimited supply of therapeutic T cells feasible for use "off-the-shelf".We have developed clinical grade HLA class I-edited iPSCs that were established from CTLs to evade rejection by recipient immune cells.GD2-CAR vector transduction into these gene-edited iPSCs will simultaneously make allogeneic GD2-CARrejT therapy available for many patients with SCLC.Although SCLC is extremely difficult to treat, GD2-CARrejTs can be prepared from iPSCs in virtually unlimited quantities, allowing multiple administrations and conferring enhanced efficacy.
In conclusion, GD2-CARrejTs demonstrated robust efficacy against SCLC both in vitro and in vivo.These findings may lead to promising therapy using iPSC technology in patients with SCLC.

FIGURE 2
FIGURE 2 Generation of GD2-CARrejTs.A, Schema, generation of GD2-CARrejTs.T-iPSCs established from LMP2-specific CTL clones were transduced with lentiviral GD2-CAR and GD2-CAR-iPSCs were differentiated into GD2-CARrejTs.B, Flow cytometric analysis of GD2-2840z-CARTs and GD2-CARrejTs to evaluate CAR transgene expression and LMP2 antigen specificity.Histogram in blue, fluorescence minus one (FMO) control.Histogram in red, GD2-2840z-CARTs and GD2-CARrejTs.The plots represent three independent experiments.C, Intracellular perforin and granzyme B marking, GD2-2840z-CARTs and GD2-CARrejTs (red).Histogram in blue, FMO control.The plots represent three independent experiments.D, CBA for measuring cytokines (IFNγ, TNF, and IL2) produced by effector T cells (GD2-2840z-CARTs and GD2-CARrejTs) after 24 hours of coculture with SCLC-J1.Error bars represent ± SEM.E, Representative flow cytometrogram of memory phenotype (CD45RA and CD62 L population) among GD2-2840-CARTs and GD2-CARrejTs.F, In vitro 51 Cr assays of GD2-2840z-CARTs, GD2-CARrejTs, and control T cells against SCLC cell lines.Error bars represent ± SD.Data represent three independent triplicate experiments.G, RTCA continuous graphical output of tumor proliferation indices up to 40 hours for H69 cells cocultured with control T cells, GD2-2840z-CARTs, and GD2-CARrejTs and for tumor only; E:T ratio uniformly 1:1.Data were plotted and are shown as mean ±SD.Data represent three independent triplicate experiments.H, RTCA continuous graphical output of tumor proliferation indices up to 100 hours for H446 cells cocultured with control T cells, GD2-2840z-CARTs, and GD2-CARrejTs and for tumor only; E:T ratio uniformly 1:1.Data were plotted and are shown as mean ±SD.Data represent three independent triplicate experiments.
Outcompeting CD155 away from CD226 attenuates effector T-cell function (Fig. 5A and F; ref. 51).In addition, PD-1-PD-L1 interaction inhibits not only the CD28 signaling pathway but also the CD226 signaling pathway (Fig. 5A; refs.52, 53).Although single blockade of either TIGIT or PD-1 slightly improved GD2-2840z-CART cytotoxicity against SCLC, dual blockade of TIGIT and PD-1 significantly enhanced it.These results indicated that blockade of both TIGIT and PD-1 was important in restoring effector function of GD2directed T-cell therapy against SCLC cells.Dual downregulation of TIGIT and PD-1 enhances short-term effector function through PD-1 suppression, while TIGIT downregulation is responsible for maintaining a less exhausted state AACRJournals.orgCancer Res Commun; 4(3) March 2024 733

FIGURE 5
FIGURE 5 Blockade of TIGIT and PD-1 augments cytotoxicity of GD2 2840z-CARTs.A, Schema of immune-regulatory mechanism of T cells against SCLC.B, Flow cytometric analysis of TIGIT and CD226 expression on GD2-2840z-CARTs and GD2-CARrejTs.The plots represent independent triplicate experiments.C, Flow cytometric analysis of CD155 expression on H69, H446, and SCLC-J1 (red).Histogram in blue, fluorescence minus one (FMO) control.The histograms represent independent triplicate experiments.D, Flow cytometric analysis of PD-L1 expression on GD2-expressing tumor cell lines (H69, H446, and SCLC-J1; red).Histogram in blue, FMO control.The histograms represent independent triplicate experiments.E, RTCA continuous graphical output of tumor proliferation indices up to 80 hours alone and in coculture with GD2-2840z-CARTs with/without anti-TIGIT (aTIGIT) and/or anti-PD-1 antibodies (aPD-1).Data were plotted and are shown as mean ± SD.Data represent three independent triplicate experiments.F, Schema of immune-regulatory mechanism of GD2-2840z-CARTs and GD2-CARrejTs against SCLC.