The human DLBCL cell lines BJAB, U2932, OCI-Ly3 were obtained from DSMZ, SUDHL-4, and SUDHL-6 were obtained from American Type Culture Collection (ATCC; Manassas, VA, USA). All cell lines were cultured in RPMI 1640 medium supplemented with 20% FBS and 100 U/ml penicillin/streptomycin (Gibco). OCI-Ly10 was purchased from Cobioer Biosciences Co., LTD (Nanjing, China) and cultured in IMDM with 20% FBS, 100 U/ml penicillin/streptomycin (Gibco) and 50 μM β-mercaptoethanol (Sigma-Aldrich). All cell lines were cultured at 37°C in a humidified atmosphere of 5% CO₂.

z-VRPR-fmk (Enzo Life Sciences) was dissolved in ddH₂O at a concentration of 50 μM throughout all experiments. MI-2, BPTES, QNZ, CPI-613 (Selleck), and PMA/Iono (Sigma-Aldrich) were reconstituted in DMSO (final DMSO concentration 0.1%) and their final concentrations were 1 μM, 2 μM, 5 μM, 100 μM, and 50/500 ng/ml, respectively.

PBMCs from fresh blood samples of healthy adult donors were isolated using density gradient centrifugation with Ficoll-Paque (GE Healthcare). To generate Vγ9Vδ2 T lymphocytes, freshly isolated PBMCs were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated FBS, 100 U/ml penicillin/streptomycin (Gibco) and 50 μM β-mercaptoethanol at a density of 3 X 10⁶ cells/mL in a 24-well plate. At the onset of the culture, 30 μM zoledronic acid (Sigma-Aldrich) and 400 U/mL IL-2 (PeproTech, Inc.) were added. Three days later, the medium was replaced with maintenance medium containing 10 μg/mL IL-2, and the maintenance medium was changed every 2–3 days. The cells were selectively expanded after 10–14 days. This study was carried out in accordance with the recommendations of local ethics guidelines, the Ethics committee of Jinan University with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the Ethics committee of Jinan University. Experimental procedures have been carried out following the standard biosecurity and the institutional safety procedures.

The cytotoxic potential of expanded Vγ9Vδ2 T lymphocytes against DLBCL cells was assayed using propidium iodide (PI, Sangon Biotech Shanghai Co., Ltd.) staining. In brief, CFSE-labeled Vγ9Vδ2 T lymphocytes were co-incubated with DLBCL cells at different effector:target (E:T) ratios of Vγ9Vδ2 T lymphocytes to DLBCL cells (0:1, 5:1, 10:1, and 20:1) at 37°C for 6 h. After cells were washed with PBS, PI solution was added to the co-culture and incubated with the cells for 5 min. Subsequently, the proportion of PI⁺CFSE⁻ cells was quantified using flow cytometry. In some experiments, DLBCL cells were pretreated with z-VRPR-fmk, BPTES or QNZ for 12 h prior to co-culture with Vγ9Vδ2 T lymphocytes. For blocking assays, purified anti-human CD274 antibody (10 μg/ml; 29E.2A3) or isotype control antibody (10 μg/ml; MG2b-57; BioLegend) were included at the time of effector/target co-culture.

Total RNA was extracted with an RNA Simple Total RNA kit (Tiangen Biotech Beijing Co., Ltd.). Synthesis of cDNA was performed with DNA-free RNA samples (Qiagen) using reverse transcription with abPrimeScript™ RT reagent kit (Takara) according to the manufacturer's protocol. Real-time PCR was performed using SYBR Green qPCR Master Mix (Bimake) on a LC480 Lightcycler system (Roche). RNA was normalized to β-actin mRNA. The sequences of primers were as follows (5′ to 3′): (forward) TTC GGT CCA GTT GCC TTCT and (reverse) GGT GAG TGG CTG TCT GTG TG for IL-6, (forward) TGG GGG AGA ACC TGA AGA and (reverse) ATG GCT TTG TAG ATG CCT TTC for IL-10, (forward) GTG AGT CGG ATC GCA GCTT and (reverse) TCG GCT GCT GCA TTG TTC for Bal-xl, (forward) CAT GTA CGT TGC TAT CCA GGC and (reverse) CTC CTT AA TGTC ACG CAC GAT for β-actin.

To measure the proportion and differential subsets of Vγ9Vδ2 T lymphocytes, we performed surface staining with FITC-conjugated anti-CD3 (Tianjin Sungene Biotech Co., Ltd.), PerCP-conjugated anti-TCR Vδ2, Pacific Blue-conjugated anti-CD27 and APC-conjugated anti-CD45RA (BioLegend) antibodies. To evaluate the functional markers of Vγ9Vδ2 T lymphocytes, cells were restimulated with 50 ng/ml PMA and 500 ng/ml ionomycin for 6 h, permeabilized using reagents from a BD Cytofix/Cytoperm Kit and stained with PE-Cy7-conjugated anti-NKG2D, APC-conjugated anti-IFN-γ (BD Biosciences), PE-conjugated anti-Perforin and Pacific Blue-conjugated anti-Granzyme B (BioLegend) antibodies. GolgiStop (BD Biosciences) was added for the last 4 h of culture. To evaluate intracellular protein, DLBCL cells were sorted from the co-culture, fixed, permeabilized using BD Phosflow™ (BD Biosciences) and stained with Alexa Fluor® 647-conjugated anti-STAT3 Phospho (Tyr705) (Biolegend). To measure PD-L1⁺ or PD-1⁺ cells, Vγ9Vδ2 T lymphocytes were co-incubated with DLBCL cells for 6 h. We then performed surface staining with BV421-conjugated anti-PD-L1 (Biolegend) or Pacific Blue-conjugated anti-PD-1 (Biolegend) antibodies. For the cell survival assay, an Annexin V/PI apoptosis kit (Bimake) was used. The data were analyzed using FlowJo software (Tree Star, lnc).

Cell pellets were collected after washing with PBS. Total cellular protein was extracted using lysis buffer (20 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, 1 mM sodium orthovanadate, 1 μg/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride), and the protein concentration was determined with a BCA Kit (Pierce). After boiling in SDS-PAGE loading buffer, 20–40 μg of protein in total cell lysates was fractionated via SDS-PAGE and transferred to PVDF membranes (Millipore). Blots were blocked in 5% BSA (Sigma-Aldrich) and incubated with the appropriate antibodies. The primary antibodies anti-Bcl-10, anti-RelB, anti-p65/Phospho-p65, anti-AKT/Phospho-AKT, anti-ERK/Phospho-ERK, anti-STAT3/Phospho-STAT3 (Cell Signaling), anti-Bcl-xl (Proteintech), anti-GLS1, anti-VDAC1 (Sangon Biotech Shanghai Co., Ltd.), and anti-β-actin (Tianjin Sungene Biotech Co., Ltd.) were used. Goat-anti-mouse IgG/HRP and goat-anti-rabbit IgG/HRP (Cell Signaling) were the secondary antibodies used. The data were acquired using a ChemiDoc™ XRS+ System (Bio-Rad) with enhanced chemiluminescence HRP substrate (Millipore).

OCR and ECAR were assessed using a Seahorse XF-96 extracellular flux analyzer (Seahorse Bioscience). DLBCL cells were pooled, carefully counted and plated (1 × 10⁵ cells/well) in assay medium onto Seahorse cell plates coated with Cell-Tak (Corning, Inc.). For OCR, assay medium was supplied with XF Base Medium with 1 mM pyruvate, 2 mM glutamine, and 10 mM glucose (Sigma-Aldrich). Baseline OCR was defined as the OCR readings before 1 μM oligomycin (Seahorse Bioscience) injection. SRC was defined as the difference between baseline OCR and maximal OCR after carbonyl cyanide-4-(trifluoromethoxy) phenylhydrazone (FCCP, 0.5 μM, Seahorse Bioscience) injection to uncouple oxidative phosphorylation and electron transport. For ECAR, assay medium was supplied with XF Base Medium with 2 mM glutamine. Glycolysis was defined as the ECAR reached by a given cell after the addition of saturating amounts of glucose. Glycolytic reserve was defined as the difference in ECAR between the glucose and 1 μM oligomycin injections.

Treated DLBCL cell lines were washed with RPMI 1640 medium without phenol red and stained with MitoTracker Green (ThermoFisher) for 30 min at 37°C. After washing, cells were resuspended with PBS and dropped in a glass slides, cover-slipped with a fluoroshield mounting medium with DAPI (Abcam). Images were collected on a Leica TCS SP8 confocal microscopy.

U2932 cells were grown in [U-¹³C]-glucose medium and were treated with z-VRPR-fmk or not for 12 h. After collection, 1 mL of chloroform and 1 mL of methanol/H₂O solvent (v/v = 1:1) were successively added into a tube of cells. The suspension was vortexed for 30 s and incubated at 4°C for 2 h. The mixture was processed using 7 cycles of ultrasonication for 2 min and incubated at −40°C for 2 min before centrifugation at 16, 000 g and 4°C for 15 min. In total, 400 μL of supernatant was transferred into a glass vial containing 10 μL of internal standards (0.05 mg/mL 13C6-15N-L-isoleucine), and dried under a gentle nitrogen stream. The dry residues were reconstituted in 30 μL of 15 mg/mL methoxylamine hydrochloride in anhydrous pyridine. The resulting mixture was vortex-mixed vigorously for 30 s and incubated at 37°C for 90 min. Then, 30 μL of MTBSTFA (with 1% BDMCS) was added to the mixture and derivatized at 55°C for 60 min. Metabolomic instrumental analysis was performed on an Agilent 7890A gas chromatography system coupled to an Agilent 5975C inert MSD system (Agilent Technologies Inc.). An HP-5ms fused-silica capillary column (30 m × 0.25 mm × 0.25 μm; Agilent J&W Scientific, Folsom, CA) was utilized to separate the derivatives. Helium (> 99.999%) was used as a carrier gas at a constant flow rate of 1 mL/min through the column. The injection volume was 1 μL (splitless mode), and the solvent delay time was 5.5 min. The initial oven temperature was held at 100°C for 2 min, ramped to 140°C at a rate of 10°C/min, to 260°C at a rate of 5°C/min, to 320°C at a rate of 10°C/min, and finally held at 320°C for 8 min. The temperatures of the injector, transfer line, and electron impact ion source were set to 250°C, 250°C, and 230°C, respectively. The electron energy was 70 eV, and data were collected in full scan mode (m/z 50–600). At least three separate injections were measured per sample, and the percent enrichment of ¹³C-labeled metabolites in the total metabolite pool was calculated for each metabolite.

Glutamate assays were performed according to the manufacturer's protocols (Sigma-Aldrich).

Two-tailed, unpaired or paired Student's t-test was used to compare differences between treated and control groups using GraphPad Prism 6.0 software. Differences with p-values < 0.05 were considered statistically significant: ^*p < 0.05; ^**p < 0.01; ^***p < 0.001.

To investigate whether ABC-DLBCL and GCB-DLBCL cells exert different immune-evasion effects on T lymphocytes, we expanded Vγ9Vδ2 T lymphocytes by culturing human peripheral blood mononuclear cells (PBMCs) from healthy adults and co-cultured them with three ABC-DLBCL cell lines (U2932, OCI-Ly3, and OCI-Ly10) and three GCB-DLBCL cell lines (BJAB, SUDHL-4, and SUDHL-6) in vitro. Expanded Vγ9Vδ2 T lymphocytes typically accounted for more than 95% of the cultured PBMCs (Figure 1A) and were mainly effector memory (EM, CD45RA⁻CD27⁻) cells, which displayed efficient cytotoxic activity and cytokine production (Figure 1B). Flow cytometric analysis confirmed that the expanded Vγ9Vδ2 T lymphocytes expressed the innate receptor NKG2D and produced IFN-γ, the cytolytic protein perforin, and granzyme B (Figure 1C). As expected, ABC-DLBCL cells showed fewer PI⁺ populations than GCB-DLBCL cells when selected by Vγ9Vδ2 T lymphocytes at the various E/T ratios tested (Figure 1D), suggesting that ABC-DLBCL cells were more inclined to subvert the cytotoxicity of Vγ9Vδ2 T lymphocytes. To further illustrate whether MALT1 protease activity is involved in the tolerance of ABC-DLBCL cells to Vγ9Vδ2 T lymphocytes, we pretreated DLBCL cell lines with z-VRPR-fmk, a cell-permeable and irreversible MALT1 inhibitor. z-VRPR-fmk treatment efficiently blocked the proteolytic activity of MALT1, as indicated by the absence of the cleaved-BCL10 fragment (Figure 1E). Interestingly, all three z-VRPR-fmk pretreated ABC-DLBCL cell lines doubled their cell death in response to Vγ9Vδ2 T lymphocytes (Figure 1E). In stark contrast, z-VRPR-fmk did not have a significant effect on any of the three GCB-DLBCL cell lines tested (Figure 1E). Moreover, z-VRPR-fmk alone had no impact on the survival of these DLBCL cells (Figure 1F). These results indicated that MALT1 protease activity has an essential role in assisting ABC-DLBCL cell escape from the cytotoxicity of Vγ9Vδ2 T lymphocytes.

To confirm that PD-L1 plays a key role in ABC-DLBCL cell resistance to the cytotoxicity of Vγ9Vδ2 T lymphocytes, we screened a panel of DLBCL cell lines for PD-L1 expression using flow cytometry. Surprisingly, only the OCI-Ly10 cell line showed high levels of PD-L1 expression. The others, including three GCB-DLBCL cell lines (BJAB, SUDHL-4 and SUDHL-6) and two ACB-DLBCL cell lines (U2932 and OCI-Ly3), expressed little PD-L1 (Figure 2A). Considering that PD-L1 expression is immunologically active by suppressing the activation of tumor-associated T cells (9, 25), we checked the kinetics of PD-L1 expression on these DLBCL cell lines in response to Vγ9Vδ2 T lymphocytes. We observed that the proportion of PD-L1⁺ cells increased in the DLBCL cell lines, suggesting that Vγ9Vδ2 T lymphocyte stress might promote PD-L1 expression (Figure 2B). Interestingly, the proportion of PD-L1⁺ ABC-DLBCL cells was generally higher than that of PD-L1⁺ GCB-DLBCL cells. Treatment with z-VRPR-fmk significantly decreased the proportion of PD-L1⁺ ABC-DLBCL cells (Figure 2B). These results revealed that MALT1 protease activity is essential for PD-L1 expression on ABC-DLBCL cells. Importantly, blocking PD-L1 with anti-PD-L1 antibodies effectively restored the cytotoxicity of Vγ9Vδ2 T lymphocytes (Figure 2C). In addition, administration of anti-PD-L1 antibody induced no further increase in cytotoxicity when combined with z-VRPR-fmk, suggesting that z-VRPR-fmk enhances Vγ9Vδ2 T lymphocyte cytotoxicity toward ABC-DLBCL cells mainly by inhibiting PD-L1 expression (Figure 2C). On the other hand, we found that an average of ~20% of the expanded Vγ9Vδ2 T lymphocytes expressed PD-1 in the basal state (Figure 2D). The proportion remained nearly the same when Vγ9Vδ2 T lymphocytes were incubated with GCB-DLBCL cells, but markedly increased when they were mixed with ABC-DLBCL cells (Figures 2E,F). However, the PD-1 expression remained unchanged in Vγ9Vδ2 T lymphocytes incubated with GCB-DLBCL or ABC-DLBCL cells pretreated with z-VRPR-fmk (Figures 2E,F). Together, these data support the idea that z-VRPR-fmk selectively decreases the generation of PD-L1⁺ ABC-DLBCL cells in response to Vγ9Vδ2 T lymphocytes, thereby inhibiting the immunosuppressive property of ABC-DLBCL cells.

NF-κB, a common transcription factor downstream of the CBM complex, has been shown to regulate PD-L1 expression in tumor cells (24, 26). Thus, we assessed whether NF-κB is indispensable for the PD-L1 expression on ABC-DLBCL cells mediated by MALT1 protease activity. We found that NF-κB transcription activity was significantly inhibited in ABC-DLBCL cells based on the diminished expression of NF-κB target genes, including Bcl-xl protein expression (Figure 3A) and the mRNA levels of Bcl-xl, IL-6, and IL-10 (Figure 3B), in response to z-VRPR-fmk treatment. However, the abolished NF-κB transcription activity induced by z-VRPR-fmk recovered in the presence of Vγ9Vδ2 T lymphocytes (Figures 3A,B). We excluded the possibility that z-VRPR-fmk lost its NF-κB inhibition activity by including BJAB cells as a negative control; in these cells, NF-κB target genes remained constant under these conditions. Since the proportion of PD-L1⁺ ABC-DLBCL cells was profoundly decreased by z-VRPR-fmk (Figure 2B), NF-κB might not be involved in the PD-L1⁺ ABC-DLBCL cell generation under the stress of Vγ9Vδ2 T lymphocytes. To confirm the role of NF-κB in generating PD-L1⁺ ABC-DLBCL cells, we directly blocked NF-κB transcription using QNZ. Distinct from z-VRPR-fmk, QNZ completely blocked NF-κB transcription under Vγ9Vδ2 T lymphocyte stress, as shown by the extreme decreases in Bcl-xl protein expression (Figure 3C) and the Bcl-xl, IL-6, and IL-10 mRNA levels (Figure 3D). Importantly, QNZ-induced blockage of NF-κB transcription activity only slightly increased cell death in ABC-DLBCL cells (Figure 3E) and slightly decreased PD-L1⁺ABC-DLBCL cell generation (Figure 3F). Taken together, NF-κB is dispensable for the generation of PD-L1⁺ ABC-DLBCL cells mediated by MALT1 protease activity under Vγ9Vδ2 T lymphocyte stress.

A possible link was recently reported between metabolic alterations and PD-L1 expression (18, 19). Hence, we proposed that MALT1 protease activity might be involved in regulating cell metabolic reprogramming to support PD-L1⁺ ABC-DLBCL cell generation. To test our hypothesis, we examined the bioenergetic profiles of z-VRPR-fmk- or vehicle-treated DLBCL cells, after confirming z-VRPR-fmk efficiently blocked the proteolytic activity of MALT1 (Supplementary Figure 1A). We found that the basal oxygen consumption rate (OCR) was slightly lower in ABC-DLBCL cells primed with z-VRPR-fmk, while mitochondrial spare respiratory capacity (SRC) was substantially decreased, as revealed through the difference between basal OCR and maximal OCR after FCCP treatment (Figures 4A,B). This absent mitochondrial SRC suggested that mitochondria in these cells were operating close to their bioenergetics limit under stressful conditions, which was further confirmed by decreased ATP-coupled OCR and mitochondrial ATP levels (Supplementary Figure 1B). However, mitochondria biogenesis was not influenced by VRPR, which was indicated by unchanged VDAC1 protein levels and intensity of MitoTracker Green staining (Supplementary Figures 1C,D). For GCB-DLBCL cells, these were not influenced by treatment with z-VRPR-fmk (Figures 4C,D; Supplementary Figures 1B–D). Furthermore, we found that GCB-DLBCL cells (BJAB and SUDHL-4 cells) with activated MALT1 protease activity presented higher basal OCR and mitochondrial SRC than the control group or the z-VRPR-fmk combined with PMA/Iono group (Supplementary Figure 2A; Figures 4E,F). We also measured the changes in extracellular acidification rate (ECAR), a typical readout of cellular glycolytic activity, in these DLBCL cells by administering z-VRPR-fmk (Supplementary Figures 2B,C). The data were analyzed and reported as the rate of glycolysis under basal conditions and glycolytic reserve calculated as the difference between the maximal glycolytic capacity and the basal glycolysis, which revealed unchanged after z-VRPR-fmk treatment (Supplementary Figures 2D,E). All the above data showed that MALT1 protease activity endowed ABC-DLBCL cells with a latent ability for mitochondrial bioenergetics, which might be fully applied in response to Vγ9Vδ2 T lymphocytes.

To understand how MALT1 protease activity supports mitochondrial bioenergetics in ABC-DLBCL cells, we analyzed the metabolic consequences in the U2932 cell line, whose MALT1 protease activity was profoundly blocked by z-VRPR-fmk (Supplementary Figure 3A). We grew these cells in uniformly labeled [U-¹³C]-glucose medium. As the results showed, there was no change in the level of the m+3 isotopolog of lactate (¹³C₃-lactate) derived from [U-¹³C]-glucose, which indicated an unchanged ECAR of ABC-DLBCL cells primed with z-VRPR-fmk (Figure 5A). In addition, z-VRPR-fmk did not impact pyruvate derived from [U-¹³C]-glucose (m+3 forms). However, glucose-derived TCA cycle intermediates, such as the doubly ¹³C-labeled isotopolog of citrate, a-ketoglutarate, succinate, fumarate, and malate (m+2 forms, circled green), were increased (Figure 5A). Notably, a large fraction of these TCA metabolites (m+0 forms, circled red) were not derived from the [U-¹³C]-glucose (Figure 5A), suggesting that an alternative source supported mitochondrial bioenergetics via TCA cycling. Meanwhile, these TCA metabolites (m+0 forms, circled red) decreased under MALT1 protease activity inhibition conditions (Figure 5A). These data suggest that glutamate from glutaminolysis likely entered the TCA cycle, which was attenuated by MALT1 protease activity inhibition. Subsequently, we found that the levels of GLS1 protein (Figure 5B) and intracellular glutamate (Figure 5C) were down-regulated in ABC-DLBCL cells without MALT1 protease activity but up-regulated in GCB-DLBCL cells with MALT1 protease activity (Supplementary Figures 3B,C). These data indicate that MALT1 protease activity promoted glutaminolysis by enhancing GLS1 expression, producing substantial glutamate that entered the TCA cycle to support mitochondrial bioenergetics. Next, we determined whether enhanced mitochondrial SRC was attributed to glutaminolysis-mediated mitochondrial bioenergetics. We exposed ABC-DLBCL cells to the GLS1 inhibitor BPTES, which blocks glutaminolysis by inhibiting the chemical conversion of glutamine to glutamate (Supplementary Figure 3D). Treatment with either BPTES or z-VRPR-fmk impaired mitochondrial SRC in ABC-DLBCL cells (Figures 5D,E). Administration of BPTES showed no additional decline in mitochondrial SRC when combined with z-VRPR-fmk (Figures 5D,E), suggesting that the enhancement effect of MALT1 on mitochondrial SRC was dependent on glutaminolysis in ABC-DLBCL cells. Together, these data established that MALT1 protease activity up-regulated GLS1 expression, thereby promoting glutaminolysis-mediated mitochondrial bioenergetics in ABC-DLBCL cells.

To explore whether glutaminolysis enhanced the generation of PD-L1⁺ ABC-DLBCL cells, we exposed DLBCL cells to BPTES and co-cultured them with Vγ9Vδ2 T lymphocytes. After treatment with BPTES, glutaminolysis was effectively blocked, as indicated by the decreased intracellular glutamate levels (Figure 6A), while survival of these DLBCL cells was not influenced before exposure to Vγ9Vδ2 T lymphocytes (Supplementary Figure 4). Compared with that of GCB-DLBCL cells, PD-L1⁺ cell generation was significantly depressed (Figure 6B) and ABC-DLBCL cell death was promoted (Figure 6C) by BPTES exposure. BPTES administration did not further improve the cytotoxicity of Vγ9Vδ2 T lymphocytes combined with anti-PD-L1 antibodies (Figures 6D,E), suggesting that the attenuation effect of BPTES on immune-suppressive function was mediated by PD-L1 expression in ABC-DLBCL cells. To determine the extent to which the anti-cytotoxicity effects of MALT1 protease activity in ABC-DLBCL cells are glutaminolysis-dependent, we pretreated ABC-DLBCL cells with z-VRPR-fmk and BPTES separately or in combination followed by co-culture with Vγ9Vδ2 T lymphocytes. The combination of both inhibitors did not increase the cell death rate compared with the effect of either inhibitor alone (Figure 6F), which suggested that glutaminolysis mediated the immunosuppressive effects of MALT1 protease activity on ABC-DLBCL cells.

To understand how glutaminolysis promotes PD-L1⁺ ABC-DLBCL cell generation, we blocked glutaminolysis in ABC-DLBCL cells using BPTES and then co-cultured them with Vγ9Vδ2 T lymphocytes. After sorting DLBCL cells, we measured the changes in total and phosphorylated p65, AKT, ERK, and STAT3 protein levels. As shown in Figure 7A, STAT3 phosphorylation was almost completely inhibited by treatment with BPTES in ABC-DLBCL cells, indicating its activation was inhibited by blocking glutaminolysis. In contrast, p65, AKT and ERK phosphorylation was not impacted, suggesting that STAT3 is the major regulator of PD-L1. Re-supplementation of glutamate completely recovered the reduced levels of intracellular glutamate and phospho-STAT3 caused by BPTES treatment (Figures 7B,C), indicating that glutamate or glutamate-mediated mitochondrial bioenergetics plays an important role in maintaining phospho-STAT3. To further confirm that glutamate-mediated mitochondrial bioenergetics maintains phospho-STAT3 or glutamate, we detected the mitochondrial respiration profiles of ABC-DLBCL cells after glutamate re-supplementation in the presence of BPTES. ABC-DLBCL cells showed a lower mitochondrial SRC in the presence of BPTES, which recovered to levels as high as those observed in the original state when glutamate was resupplied, indicating that glutamate entered the TCA cycle to support mitochondrial bioenergetics (Figures 7D,E). Moreover, glutamate re-supplementation did not recover the phospho-STAT3 levels to those observed in the original state in cells treated with CPI-613, which functions as a TCA cycle blocker (Figure 7F). These data illustrate that glutaminolysis-mediated mitochondrial bioenergetics helped to sustain phospho-STAT3, a PD-L1 transcription factor, which finally led to generation of PD-L1⁺ ABC-DLBCL cells under Vγ9Vδ2 T lymphocyte stress.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.