αβ-T cell receptor transduction gives superior mitochondrial function to γδ-T cells with promising persistence

Summary Adoptive cell therapy using allogeneic γδ-T cells is a promising option for off-the-shelf T cell products with a low risk of graft-versus-host disease (GVHD). Long-term persistence may boost the clinical development of γδ-T cell products. In this study, we found that genetically modified Vγ9+Vδ2+ T cells expressing a tumor antigen-specific αβ-TCR and CD8 coreceptor (GMC) showed target-specific killing and excellent persistence. To determine the mechanisms underlying these promising effects, we investigated metabolic characteristics. Cytokine secretion by γδ-TCR-stimulated nongene-modified γδ-T cells (NGMCs) and αβ-TCR-stimulated GMCs was equally suppressed by a glycolysis inhibitor, although the cytokine secretion of αβ-TCR-stimulated GMCs was more strongly inhibited by ATP synthase inhibitors than that of γδ-TCR-stimulated NGMCs. Metabolomic and transcriptomic analyses, flow cytometry analysis using mitochondria-labeling dyes and extracellular flux analysis consistently suggest that αβ-TCR-transduced γδ-T cells acquire superior mitochondrial function. In conclusion, αβ-TCR-transduced γδ-T cells acquire superior mitochondrial function with promising persistence.


Highlights
Long-term persistence may boost the clinical development of gd-T cell products ab-TCR transduction into gd-T cells showed promising persistence ab-TCR transduced gd-T cells could utilize more mitochondrial energy metabolism

INTRODUCTION
A new era of immunotherapy for cancer has arrived with the application of immune checkpoint inhibitors and adoptive cell therapy with genetically modified T cells directed against cancer antigens.Chimeric antigen receptor-engineered T cells (CAR-T cells) targeting CD19 have shown clinical benefits in clinical trials [1][2][3] and are an available standard therapy for B cell malignancies.T cell receptor (TCR)-engineered T cells (TCR-T cells) are less developed than CAR-T cells, but no conclusion has been reached as to whether CAR-T cells or TCR-T cells are superior.One of the characteristics of TCR-T cells is their ability to target intracellularly expressed antigens.5][6] NY-ESO-1 is the most studied antigen in TCR-T cell clinical trials. 7Our group conducted clinical trials of TCR-T cells targeting NY-ESO-1-expressing tumors 8,9 using the G50A + A51E TCR, an NY-ESO-1-specific TCR recognizing the NY-ESO-1 p157-165 peptide (NE1 p157 ) and HLA-A*02:01 complex.This TCR had two amino acid replacements in the CDR2 region of the TCRb chain to produce high affinity, 10 which differs from the one reported by Robbins et al. [11][12][13] G50A + A51E TCR-T cells showed tolerable toxicity and promising efficacy. 8,9espite the successes described previously, there are two main issues with adoptive T cell therapy using ab-T cells: mispairing of endogenous and engineered ab-TCRs 14 and difficulties in using allogeneic ab-T cells for rapid and chemotherapy-damage-free cell product preparation due to graft-versus-host disease (GVHD).To address these issues, we focused on gd-T cells.6][17] gd-T cells recognize tumors in an MHC-independent manner 18,19 and have no risk of mispairing endogenous and engineered ab-TCRs.Tumor infiltration by gd-T cells is a good prognostic marker for many cancers. 20Although adoptive cell transfer of nongene-modified gd-T cells (NGMCs) has been studied, [21][22][23][24][25] these cells have not achieved a remarkable improvement in survival.Redirection of gd-T cells by ab-TCR gene transduction may yield antitumor activities against NGMC-resistant tumors and resolve the two aforementioned issues with ab-T cell products.In this study, we evaluated genetically modified Vg9 + Vd2 + T cells expressing NY-ESO-1-specific ab-TCR and the CD8 coreceptor (NE1-GMCs), with a specific focus on their metabolic preferences in association with their effector function and persistence.

Effector function and metabolism dependency of IFN-g secretion in NE1-GMCs
In cultures without cytokine supplementation, cell growth of PTA-treated gd-T cells was suppressed (Figures S4A and S4B).PTA-treated NGMCs functioned as antigen-presenting cells, and the possibility of mutual cytotoxicity could not be ruled out (Figure S4C).In contrast, compared to NE1-GMCs, TCRgd-specific agonistic mAb (clone IMMU510)-stimulated NGMCs remained at a similar frequency (Figure S4A) and showed similar cell proliferation (Figure S4D).For the aforementioned reasons, IMMU510 was used to stimulate NGMCs in subsequent experiments.

Metabolomic analysis of NGMCs and NE1-GMCs
Metabolites from unstimulated and stimulated NGMCs and NE1-GMCs were analyzed using a combination of capillary electrophoresis and Fourier transform mass spectrometry (CE-FTMS), and 357 metabolites (cationic:195, anionic:162) were detected (Figure 3A).Metabolites involved in the TCA cycle and OXPHOS were strongly correlated with PC1, which separated unstimulated NE1-GMCs from unstimulated NGMCs (Figure 3B and Table S2).NE1-GMCs contained more metabolites of the tricarboxylic acid (TCA) cycle and oxidative phosphorylation (OXPHOS) in both unstimulated and stimulated states (Figure 3C).The frequencies of Vd2 + cells in NGMCs and CD8 + cells in NE1-GMCs used in the study were similar (Figures S6A and S6B).

Transcriptomic analysis
RNA microarray analysis was performed using cells from the four groups as in the metabolomic analysis.Cells were newly induced on different days from those used for metabolomic analysis.Gene set enrichment analysis (GSEA) showed a trend toward increased expression of genes related to mitochondria in NE1-GMCs compared to NGMCs.Gene expression associated with the ''electron transport chain: oxphos system in mitochondria'' of WikiPathways was higher in NE1-GMCs than in NGMCs under both unstimulated and stimulated conditions (Figure 4A and Table 1).Gene expression associated with the ''oxidative phosphorylation'' of WikiPathways was higher in NE1-GMCs than in NGMCs in the stimulated state (Figures 4B and Table 1).

Mitochondrial phenotype assessed by flow cytometry
The mitochondrial phenotype of NGMCs and NE1-GMCs was assessed using MitoTracker Green and tetramethylrhodamine methyl ester (TMEM).NE1-GMCs had higher mitochondrial mass and mitochondrial membrane potential in the unstimulated state (Figure 5A).Another ab-TCR-transduced GMCs, p40Tax-GMCs, also showed a similar mitochondrial phenotype (Figure S7A).

Metabolic status analyzed by extracellular flux
Cellular metabolic status was assessed by extracellular flux analysis, measuring the oxygen consumption rate (OCR) and extracellular acidification rate (ECAR), with the results representing the levels of mitochondrial respiration and glycolysis, respectively.In the absence of TCR stimulation, the median basal OCR/basal ECAR ratio of NGMCs was similar to or higher than that of NE1-GMCs (NGMCs: 0.95 vs. NE1-GMCs: 0.65) (Figures 5B and S7B).With TCR stimulation, the median basal OCR/basal ECAR ratio increased only in NE1-GMCs, indicating enhanced mitochondrial metabolism in NE1-GMCs upon ab-TCR stimulation (NGMCs: 0.92 vs. NE1-GMCs: 2.33) (Figure 5C).A similar change was observed in p40Tax-GMCs (Figure S7C).

DISCUSSION
In a study on TCR-T cells, an autoreactive TCR consisting of endogenous TCR and engineered TCR caused lethal toxicity in a mouse model. 14onsidering the use of allogeneic gd-T cells for ab-TCR-transduced T cell therapy, T cell products would be safe because genetically engineered T cells are at least dual receptor-like, with monoclonal ab-TCR and gd-TCR.Acute rejection due to HLA incompatibility can be controlled by prior lymphodepleting chemotherapy.Therefore, if ab-TCR-transduced gd-T cells persist, the clinical development of offthe-shelf ab-TCR-transduced allogeneic gd-T cell products will advance greatly.The in vivo persistence of ab-T cells is thought to be related  to antitumor efficacy, and mainly in the context of CAR-T cell development, incorporation of 4-1BB (CD137) signaling [28][29][30] and cytokines such as IL-7 and IL-15 31 into T cells has been studied.Vg9 + Vd2 + T cells are an innate-like T cell population. 32Clinical trials on these cells have not shown a remarkable survival benefit.The reason for this lack of efficacy is unclear 33 ; however, their short persistence after activation may be a cause.In our study, ab-TCR-transduced gd-T cells showed promising persistence in vitro and in vivo.Lope et al. reported that IFN-g secretion by gd-T cells was glycolysis-dependent, 34 which is consistent with our data.Mitochondrial function is useful for assessing the persistence of ab-T cells 35,36 and may be relevant for gd-T cells.Here, we found that ab-TCR-transduced gd-T cells could utilize more mitochondrial energy metabolism.
In our study, ab-TCR transduction increased the mitochondrial mass and mitochondrial membrane potential of gd-T cells.ab-TCR-transduced gd-T cells contained more metabolites in the TCA cycle and OXPHOS and had higher gene expression of the mitochondrial electron transport chain in both unstimulated and stimulated states.In the stimulated state, IFN-g production was more dependent on ATP synthesis, which was supported by the higher OXPHOS gene expression observed in ab-TCR-transduced gd-T cells.Although not high on the list in our gene analysis, our examinations using metabolic inhibitors suggest that energy through glutamine and fatty acid metabolism might also be important for the characteristics of ab-TCR-transduced gd-T cells.The metabolism of glutamine and fatty acids in ab-TCR-transduced gd-T cells requires further study.
Evaluation of spare respiratory capacity (SRC) is a useful method for assessing mitochondrial function in T cells. 37,38However, it was difficult to identify differences in SRC between TCR-stimulated NGMCs and NE1-GMCs.In contrast to SRC, the basal OCR/basal ECAR ratio can compensate for individual differences. 39An increased basal OCR/basal ECAR ratio was observed in NE1-pMHC-stimulated NE1-GMCs, which might reflect an inclination toward mitochondrial energy metabolism.Similar results were obtained for p40Tax-pMHC-stimulated p40Tax-GMCs.In a mouse model, a relationship between gd-T cell subsets classified by cytokine secretion (IFN-g or IL-17) and mitochondrial function was reported, 34 but there were no obvious differences in cytokine profiles between NGMCs and NE1-GMCs induced from human donors in our study.The small sample size is one of the limitations of our study and requires further validation.However, our experiments consistently suggest that ab-TCR-transduced gd-T cells acquire superior mitochondrial function.
Recently, Stenger et al. reported the importance of the presence of ab-TCR for ab-T cell persistence by showing that knock out of endogenous TCR in CD19 CAR-T cells resulted in a lack of persistence, although the mechanism remains unclear. 40Mitochondria play an important role in the sustained killing of cytotoxic T cells. 41In our study, ab-TCR and CD8 transduction to gd-T cells yielded promising T cell persistence with higher utilization of mitochondrial function.CD8 is necessary to stabilize ab-TCR recognition of the peptide epitope present on MHC molecules. 16he recruitment of CD8-associated Lck is essential for the augmentation of TCR signal transduction 42 and in vivo persistence of ab-T cells. 43These results support the enhanced antitumor activity of ab-TCR and CD8 gene-cotransduced gd-T cells.5][46][47] Appropriate Lck activation seems to be important for CD8 + T cell effector function and proliferation.
In conclusion, ab-TCR transduction into gd-T cells showed promising in vitro and in vivo persistence.As a mechanism, gd-T cells could utilize more mitochondrial energy metabolism due to the acquisition of ab-TCR and CD8.

Limitations of the study
The study has several limitations: small sample size and reporting from a single group.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:    TCR (NE1 p157 /HLA-A*02:01-specific TCR-a and TCR-b chains with G50A + A51E amino acid replacements) and human CD8 (Figure 1A) were used.An NY-ESO-1-specific TCR has been reported by Schmid et al. 10 On day 8, gd-T cells were harvested and frozen with CELLBANKER1 (Takara Bio) until use, except for metabolomic and transcriptome analyses.Each experiment was performed within two months of gd-T-cell freezing.Thawed gd-T cells were cultured in RPMI 1640 medium with 10% FCS for 4-6 h and then used in each experiment.
All procedures performed in this study involving human participants were conducted in accordance with the Japanese law of Ethical Guidelines for Medical and Biological Research Involving Human Subjects.

Metabolomic analysis
Comprehensive analysis of hydrophilic and ionic metabolites using a combination of capillary electrophoresis and Fourier transform mass spectrometry (CE-FTMS) was performed by Human Metabolome Technologies, Inc. (Yamagata, Japan).NE1-GMCs were purified using CD8 Microbeads (Miltenyi Biotec) on day seven.Metabolic extracts were prepared from 2.0-3.5 3 10 6 cells per sample, with three samples from three different healthy donors included in each group, and methanol containing an internal standard solution after washing with mannitol-containing solution on day 8. Prior to extraction, T cells were cultured for 6 h in RPMI 1640-10% FCS medium with or without TCR stimulation.TCRgd-specific agonistic mAb-and NE1-pMHC-coated plates were used for NGMCs and NE1-GMCs, respectively.

Transcriptomic analyses
RNA extracts were prepared from 3.1 3 10 5 cells per sample from a healthy donor using a QIAamp RNA Blood Mini kit (QIAGEN K.K., Tokyo, Japan).Prior to RNA extraction, NGMCs and NE1-GMCs purified with CD8 microbeads were cultured with or without TCR stimulation, as in metabolomic analysis.RNA microarray analysis was performed by Agilent Technologies, Inc. Briefly, the RNA integrity of the samples was determined using an Agilent 2100 Bioanalyzer and Agilent RNA 6000 Pico Kit.All samples showed an RNA integrity number (RIN) higher than eight and qualified for further processing.Total RNA was reverse-transcribed to cDNA and then synthesized into Cy3-labeled cRNA by in vitro transcription.Labeled samples were hybridized to a SurePrint G3 Human GE Microarray 8 3 60 K Ver3.0.Scanning was performed using a SureScan Microarray scanner.Gene set enrichment analysis (GSEA) 48 was performed using GSEA software version 4.3.2.(https://www.gsea-msigdb.org/gsea/index.jsp).The log2-transformed, normalized and validated gene data by Agilent Technologies, Inc. was used.The genes were ranked based on the gene set 'WikiPathways subset of Canonical Pathways'.

Measurement of the oxygen consumption rate (OCR) and extracellular acidification rate (ECAR)
OCR and ECAR were measured with a Seahorse XF HS Mini Analyzer using a Seahorse XF Cell Mito Stress Test Kit (Agilent Technologies Inc., Santa Clara, CA, USA).Briefly, after thawing, gd-T cells were cultured in RPMI 1640-10% FCS medium.The TCR stimulation time was 72 h.After washing with assay medium, gd-T cells were plated at 2.0-3.0 3 10 5 cells/well, with the same number of cells included in each experiment, in a cell culture plate with assay medium and cultured for a total of 50-60 min at 37 C in a CO 2 -free incubator until analysis.The final concentrations of oligomycin, FCCP, and rotenone/antimycin A were 1.5, 1.0 mM and 0.5 mM, respectively.The basal OCR/basal ECAR ratio was calculated using the equation: (pmol/min)/(mpH/min). 39

QUANTIFICATION AND STATISTICAL ANALYSIS
For cytokine inhibition studies, a t test was used if Levene's test showed a value above the significance level; otherwise, Welch's t test was used.For metabolic status analysis using Seahorse, the Mann-Whitney U test was used.Calculations were performed using SPSS Statistics version 26 (IBM Japan, Ltd., Tokyo, Japan).For the metabolomic analysis, we described the results of Welch's test reported by Human Metabolome Technologies, Inc. p values less than 0.05 were considered statistically significant (*p < 0.05, **p < 0.01).

Figure 2 .
Figure 2. NY-ESO-1-specific response and persistence of NE1-GMCs and comparison of the effects of metabolic inhibition on IFN-g secretion between NGMCs and NE1-GMCs (A) Frequencies of IFN-g-secreting cells in NGMC Vd2 + cells and NE1-GMC Vd2 low CD8 high cells in media with different 2-DG concentrations.The means of 3 experiments using gd-T cells induced from 2 healthy donors are shown in the bar graph.(B) Frequencies of IFN-g-secreting cell ratios determined by comparing IMMU510-stimulated NGMC Vd2 + cells (6 experiments using gd-T cells induced from 3 healthy donors) and NE1-pMHC-stimulated NE1-GMC Vd2 low CD8 high cells (6 experiments using gd-T cells induced from 3 healthy donors).The concentrations of metabolic inhibitors were 2-DG (10 mM), etomoxir (100 mM), CB-839 (2 mM), or oligomycin (2 mg/mL).The error bar shows the standard deviation.

Figure 3 .
Figure 3. Metabolomic analysis of NGMCs and NE1-GMCs Three samples from three different healthy donors per group were assessed.The Roman numerals (I, II, and III) at the top of each sample name indicate that the T cells were derived from the same healthy donor and induced on the same day.(A) Heatmap of 357 metabolites (cationic:195, anionic:162) detected by CE-FTMS compared among unstimulated NGMCs (blue; I-1, II-1, and III-1), unstimulated NE1-GMCs (green; I-2, II-2, and III-2), IMMU510-stimulated NGMCs (red; I-3, II-3, and III-3), and NE1-pMHC-stimulated NE1-GMCs (orange; I-4, II-4, and III-4).The bars on the right show the range of metabolites involved in glycolysis (lower panel) and the TCA cycle and OXPHOS (upper panel).(B) Principal component analysis of 357 detected metabolites.Each circle indicates unstimulated NGMCs (blue), unstimulated NE1-GMCs (green), IMMU510stimulated NGMCs (red), and NE1-pMHC-stimulated NE1-GMCs (orange).(C) Bar graph of the metabolites involved in glycolysis, the TCA cycle, and OXPHOS in unstimulated and stimulated states.Each bar represents the mean relative areas of the indicated metabolites.Error bars indicate the standard deviation.

Figure 4 .
Figure 4. Continued (A) Heatmap of rank metric scores and plots of the enrichment scores for the '' WP electron transport chain: oxphos system in mitochondria'' gene set in the unstimulated and TCR-stimulated states are shown.(B) Heatmap of rank metric scores and plots of the enrichment scores for the ''WP oxidative phosphorylation'' gene set in the TCR-stimulated state are shown.The ''unstimulated'' rank metric score shows unstimulated NE1-GMCs versus unstimulated NGMCs, and the ''stimulated'' rank metric score shows NE1-pMHCstimulated NE1-GMCs versus IMMU510-stimulated NGMCs.Rank metric score >0 (red) means higher in NE1-GMCs, <0 (blue) means lower in NE1-GMCs, compared with NGMCs.One sample per group was analyzed.

Figure 5 .
Figure 5. Mitochondrial phenotype assessment by flow cytometry using mitochondria-labeling dyes and extracellular flux analysis of NGMCs and NE1-GMCs (A) The mitochondrial phenotype of NGMCs and NE1-GMCs was assessed by flow cytometry using mitochondria-labeling dyes.The bar graph shows the average of five measurements of the mean fluorescence intensity (MFI) of MitoTracker Green and TMRM.(B)Boxplots (median with the 25th and 75th percentiles) of the basal OCR/basal ECAR ratio comparing unstimulated NGMCs (total 18 measurements from 2 experiments) and unstimulated CD8 + cells selected from NE1-GMCs (total 18 measurements from 3 experiments).The median basal OCR/basal ECAR ratios were 0.95 and 0.65, respectively.(C) Boxplots (median with the 25th and 75th percentiles) of the basal OCR/basal ECAR ratio of IMMU510-stimulated NGMCs (light gray, total 12 measurements from 2 experiments) and NE1-pMHC-stimulated CD8 + cells purified from NE1-GMCs (dark gray, total 15 measurements from 2 experiments).The median basal OCR/basal ECAR ratios were 0.92 and 2.33, respectively.

TABLE
d RESOURCE AVAILABILITY B Lead contact B Materials availability B Data and code availability d EXPERIMENTAL MODELS AND STUDY PARTICIPANTS B Animals B Cell lines d METHOD DETAILS B Vector construction and preparation of ab-TCR-engineered gd-T cells B In vitro functional assay based on flow cytometry B Metabolomic analysis