The success of checkpoint inhibitors in cancer treatment is associated with the infiltration of tissue-resident memory T (Trm) cells. In this study, we found that about 30% of tumor-infiltrating lymphocytes (TIL) in the tumor microenvironment of gastric adenocarcinoma were CD69+CD103+ Trm cells. Trm cells were low in patients with metastasis, and the presence of Trm cells was associated with better prognosis in patients with gastric adenocarcinoma. Trm cells expressed high PD-1, TIGIT, and CD39 and represented tumor-reactive TILs. Instead of utilizing glucose, Trm cells relied on fatty acid oxidation for cell survival. Deprivation of fatty acid resulted in Trm cell death. In a tumor cell–T-cell coculture system, gastric adenocarcinoma cells outcompeted Trm cells for lipid uptake and induced Trm cell death. Targeting PD-L1 decreased fatty acid binding protein (Fabp) 4 and Fabp5 expression in tumor cells of gastric adenocarcinoma. In contrast, the blockade of PD-L1 increased Fabp4/5 expression in Trm cells, promoting lipid uptake by Trm cells and resulting in better survival of Trm cells in vitro and in vivo. PD-L1 blockade unleashed Trm cells specifically in the patient-derived xenograft (PDX) mice. PDX mice that did not respond to PD-L1 blockade had less Trm cells than responders. Together, these data demonstrated that Trm cells represent a subset of TILs in the antitumor immune response and that metabolic reprogramming could be a promising way to prolong the longevity of Trm cells and enhance antitumor immunity in gastric adenocarcinoma.

Gastric cancer is the second most common cancer worldwide, and gastric adenocarcinoma accounts for 95% of the gastric cancer (1). Despite significant achievements in the management of gastric adenocarcinoma, the prognosis remains dismal with a 5-year survival rate of about 20% (2). Surgical resection remains the first choice for patients with resectable tumors, but recurrence occurs in 20% to 50% of the patients following gastric resection (2). More than 50% of patients present with locally advanced or metastatic gastric adenocarcinoma at diagnosis, leaving chemotherapy as the main therapeutic option for these patients (3). Thus, the development of novel therapeutic agents and strategies for gastric adenocarcinoma is eagerly awaited due to its high morbidity and mortality.

The emergence of immunotherapy has changed the landscape of cancer treatments (4). Immune checkpoint blockade of programmed death 1 (PD-1) is successful in the treatment of lung cancer, melanoma, kidney cancer, and various other cancers (5–7). Anti–PD-1 treatment is effective for gastric cancer (8, 9). However, only 11.2% and 22% of the treated patients responded to such treatments in these two clinical trials, respectively. It is urgent to identify new therapeutic targets to approach better outcomes for patients with gastric adenocarcinoma.

Tissue-resident memory T (Trm) cells are a T-cell subset that resides in the tissue and does not circulate back to the blood or secondary lymphoid organs (10, 11). Trm cells produce higher amounts of cytokines than their circulating counterparts and provide enhanced local immunity in response to infection (12, 13). CD8+ Trm cells are associated with antitumor immune responses (14, 15). The presence of Trm cells correlates with improved prognosis in patients with cancer (16, 17). CD8+ Trm cells promote melanoma-immune equilibrium in skin, whereas mice deficient in Trm cells' formation are more susceptible to tumor development (18). These properties give Trm cells great potential in the treatment of cancer. However, the roles of Trm cells in GCA have not yet been reported, and the detailed maintenance mechanism of CD8+ Trm cells in the tumor microenvironment (TME) remains to be addressed.

In the current study, we found that Trm cells were present in the TME of gastric adenocarcinoma, which indicated better prognosis. Trm cells relied on lipid uptake and metabolism for cell survival. Cancer cells of gastric adenocarcinoma outcompeted Trm cells for lipid uptake and induced apoptosis of Trm cells, which could be reversed by blocking PD-L1 on cancer cells. Anti–PD-L1 treatment unleashed Trm cells specifically in the patient-derived xenograft (PDX) mice.

Patient samples

Primary tumor tissues were collected from patients with gastric adenocarcinoma with surgically resectable tumors in The First Affiliated Hospital (cohort 1, n = 180), Cancer Center, and The Eighth Affiliated Hospital (cohort 2, n = 152) of Sun Yat-sen University. For cell isolation and mouse experiments, tumor tissues and blood samples were collected freshly from The First Affiliated Hospital, Sun Yat-sen University (cohort 3, n = 43). Patients' clinical manifestations and laboratory results were evaluated carefully, and patients with infection or autoimmune diseases were excluded from this study. Demographics of included patients (cohort 1, cohort 2, and cohort 3) are shown in Supplementary Tables S1–S3. This study was approved by the Ethics Committee of The First Affiliated Hospital, Sun Yat-sen University. Consent was informed, and consent forms were obtained from all patients. The studies were conducted in accordance with recognized ethical guidelines of the Declaration of Helsinki.

Cell isolation

Blood samples were collected from patients with gastric adenocarcinoma. Peripheral blood mononuclear cells (PBMC) were isolated by Ficoll–Hypaque through density gradient centrifugation. Freshly collected blood samples were diluted with PBS in a ratio of 1:1. The diluted blood was then placed on top of Ficoll–Hypaque separation medium carefully, centrifuging at 1,500 rpm for 30 minutes. Buffy coat layer was collected and washed with PBS twice. To prepare single-cell suspensions, fresh tumor tissues were minced and digested with type I collagenase (2 mg/mL, Sigma) and DNAse (50 U/mL) in RPMI 1640 medium in 37°C. Digestion buffer was replaced with fresh buffer every 30 minutes. Cells were then filtered through a cell strainer (70 μm) and washed with PBS. CD8+CD103Hi, CD8+CD103Med, or CD8+CD103Neg cells were sorted from tumor-infiltrating lymphocytes (TIL) using a BD FACS Influx. CD8+ T cells were purified from PBMCs by negative selection using the EasySep human total CD8+ T Cell enrichment kit according to the manufacturer's instructions (STEMCELL Technologies Inc.; Catalog#19053). Cell purity was checked by fluorescence-activated cell sorting meter (FACS; >96%, BD FACS Influx).

Autologous tumor cells were prepared as described previously (19). Resected gastric adenocarcinoma tumor samples were immediately processed into single-cell suspensions by mechanical dissociation and enzymatic digestion. Tumor tissues were minced and incubated with collagenase I (2 mg/mL) and DNAse type I (50 U/mL). Tumor cells digested from tissue were filtered through a cell strainer. CD45 population (nonhematopoietic) was further sorted by FACS from the tumor digest and used as autologous tumor cells. The staining details for cell sorting were described below in the Flow cytometry section.

Circulating tumor cell enrichment and analysis

Circulating tumor cells (CTC) were separated by the NanoVelcro system (Cytolumina) as previously described (20). A total amount of 5 ml blood specimens was collected from each of the patients with gastric adenocarcinoma, and the samples were processed within 24 hours. Red blood cells were removed by incubating cells with red blood cell lysing buffer (BioLegend) at room temperature for 5 minutes. Cells were then washed with PBS twice. IHC was applied to visualize the captured cells on SiNW substrate. The captured cells were stained with 4′, 6-diamidino-2-phenylindole (DAPI, nuclear marker), TRITC-conjugated anti-CD45 antibody (WBC marker; Abcam; 1:400), and FITC-conjugated anti-CK antibody (cancer cell marker; Abcam; 1:100). Characteristic phenotypes and morphology of CTCs were scrutinized by an experienced pathologist. CTCs were identified by combination of the following criteria: positive staining of anti-CK and DAPI, and negative staining of anti-CD45.

Cell culture

The human gastric cancer cell line SGC7901 and normal gastric epithelial cell line GES-1 were obtained from the Type Culture Collection of Chinese Academy of Sciences in 2015 (Shanghai, China; ref. 21). These cells were authenticated and certified by cell viability analysis, short tandem repeat profiling, and isoenzyme analysis, and were also screened for mycoplasma contamination by Type Culture Collection of Chinese Academy of Sciences. Cells were not reauthenticated. Cells were grown in complete culture medium RPMI 1640 supplemented with 10% FBS, 50 U/mL penicillin, and 50 mg/mL streptomycin in a humidified atmosphere at 37°C with 5% CO2. These cells were cultured for a maximum of 15 passages after thawing from our stocks that were frozen at passage 3.

Flow cytometry

Cells were isolated from tumor tissue of patients with gastric adenocarcinoma as described above and were stained with the following antibodies: FITC-conjugated anti-CD45, phycoerythrin-conjugated anti-CD3, APC-conjugated anti-CD8, BV421-conjugated CD69, PE-CY7–conjugated CD103, BV605-conjugated PD-1, PE-conjugated TIGIT, FITC-conjugated CD39, PerCP5.5-conjugated CD26 (all were 1:100 and purchased from BioLegend). Cells were incubated with antibodies at 4°C for 30 minutes and then washed with PBS twice. SGC7901 and GES-1 were stained with primary antibodies against fatty acid binding protein (Fabp) 4 (1:200) and Fabp5 (1:200; both from Abcam) at 4°C overnight. Cells were then washed with PBS twice and incubated with Alexa Fluor 488–conjugated anti-mouse IgG (1:1,000) or Alexa Fluor 555–conjugated goat anti-rabbit IgG (1:1,000; Life Technologies) at 4°C for 30 minutes. Cells were then washed with PBS twice and analyzed by flow cytometry.

To measure cytokine production in T cells, cells were stimulated with 500 ng/mL phorbol myristate acetate, 1 μg/mL ionomycin, and 5 μg/mL Brefeldin A (Sigma-Aldrich) for 5 hours at 37°C with 5% CO2. Cells were collected, permeabilized, and fixed using the Fixation/Permeabilization Solution (BD Bioscience) on ice for 30 minutes. Cells were then washed with Perm/Wash Buffer and stained with V450-conjugated anti-IFNγ (BD Biosciences) and FITC-conjugated anti-TNFα antibodies (BioLegend) in Perm/Wash Buffer at 4°C for 30 minutes. Cells were washed twice with Perm/Wash Buffer and analyzed by flow cytometry.

To measure T-cell apoptosis, cells were stained with Pacific Blue–conjugated Annexin V and 7-AAD (BioLegend). Samples were analyzed using a BD FACS ARIA (BD Bioscience). FlowJo (Tree Star) software was used for data analysis.

Transfection

One day before the transfection, SGC7901 cells were seeded on 24-well culture plates. Fabp4, Fabp5, PD-L1 siRNA, and scramble siRNA were purchased from Santa Cruz Biotechnology (Catalog#, sc-43592, sc-41237, and sc-39699). siRNA oligomers were diluted in Opti-MEM I Reduced Serum Medium without serum to a final concentration of 50 nmol/L. Lipofectamine 2000 (1 μL) was diluted with Opti-MEM I Reduced Serum Medium (50 μL) and incubated for 5 minutes at room temperature. The diluted oligomer was then mixed with the diluted Lipofectamine 2000 and added to each well. Cells were incubated at 37°C with 5% CO2. The siRNA–Lipofectamine complex was removed after a 6-hour incubation, and the cells were cultured at 37°C with 5% CO2 overnight in RPMI 1640 medium supplemented with 10% FBS.

Coculture

Sorted CD8+CD103Hi, CD8+CD103Med, or CD8+CD103Neg T cells were cocultured with autologous tumor cells or SGC7901 cells in 48-well plates at a ratio of 5:1 in RPMI 1640 medium supplemented with 10% FBS at 37°C with 5% CO2. T cells were stimulated with anti-CD3/CD28 antibodies (BioLegend) according to the manufacturer's instructions. PD-L1–blocking antibody (Bio X Cell, 1 μg/mL) was included in some experiments. T cells were collected to determine apoptosis and lipid uptake by flow cytometry.

Cytotoxicity assay

In vitro cytotoxicity assays were performed by coculturing T cells and cancer cells as described previously (22, 23). Sorted CD8+ T cells were cocultured with autologous cancer cells for 16 hours at a ratio of 4:1. After coculture, adherent and nonadherent cells were collected, stained with 7-AAD, and analyzed by flow cytometry to determine the number of dead tumor cells. The killing percentage was calculated by the percentage of 7-AAD+ cancer cells.

Western blot

Cells were lysed with RIPA Buffer including 1% v/v Halt Protease/Phosphatase Inhibitor (Thermo Fisher Scientific). The BCA protein assay (Thermo Fisher Scientific) was used for quantitation of total protein. A total of 30 μg protein was loaded on SDS–polyacrylamide gels (4% and 15%) and electrotransferred onto polyvinylidene difluoride membranes after separation. Membranes were blocked with 10% BSA in Tris-Buffered Saline Tween-20 buffer and incubated with anti-CD36 (1:1,000; Abcam) and anti–β-actin (1:2,000; Cell Signaling Technology) primary antibodies at 4°C overnight. The membranes were then washed with Tris-Buffered Saline Tween-20 and incubated with horseradish peroxidase–conjugated anti-rabbit IgG (1:2,500; Cell Signaling Technology) at room temperature for 60 minutes. Signals were detected by enhanced chemiluminescence (Thermo Fisher Scientific) with the Odyssey Fc imaging system (LI-COR Bioscience).

Immunometabolism

To measure lipid uptake or glucose uptake, cells were incubated at 2 × 105 cells in complete medium containing either 1 μmol/L Bodipy 500 (Thermo Fisher) or 20 μmol/L 2-NBDG (Thermo Fisher), for 30 minutes at 37°C in a cell incubator. To measure lipid content in the cells, Bodipy 493 (1 μg/mL, Thermo Fisher) was added and incubated for 1 hour at 4°C. Cells were analyzed by flow cytometry. To further detail the metabolic alterations in Trm cells, a Seahorse assay was performed to measure glycolysis and mitochondria oxygen consumption. T cells were incubated in a CO2-free incubator in RPMI 1640 medium supplemented with glucose (20 mmol/L) and sodium pyruvate (1 mmol/L) for 30 minutes. Measurements were performed using an XF96 extracellular analyzer (Seahorse Bioscience). The Seahorse XF Cell Mito Stress Test Kit was used for the measurement of mitochondrial oxygen consumption rate. During the measurements, cells were treated with oligomycin (1 μmol/L), carbonyl cyanide p-trifluoromethoxyphenylhydrazone (1.5 μmol/L), and rotenone/antimycin A (1 μmol/L), respectively, as indicated.

ELISA

To measure cytokines released by T cells, culture supernatant was collected for the measurement of IFNγ and TNFα using commercial ELISA kits (Catalog# DIF50 and DTA00D) according to the manufacturer's instructions (R&D Systems).

Immunofluorescence and immunohistochemistry

Paraffin-embedded tissues were cut into 4 μm sections. The slides were deparaffinized with Xylene and rehydrated in 100% ethanol, 95% ethanol, 70% ethanol, and 50% ethanol sequentially. Antigen retrieval was performed in a pressure cooker using Antigen Retrieval Buffer (Citrate Buffer pH 6.0). Slides were blocked with 5% BSA in PBS and then incubated with primary antibody against CD8 (1:1,000) and CD103 (1:200; Abcam) at 4°C overnight. Slides were washed with PBS twice and incubated with Alexa Fluor 555–conjugated goat anti-rabbit IgG (1:500) and Alexa Fluor 488–conjugated anti-mouse IgG (1:500; Life Technologies) at room temperature for 30 minutes. Sections were counterstained with DAPI in mounting medium. For IHC staining, slides were stained with primary antibody against CD103 (1:200) at 4°C overnight. The sections were then incubated with an horseradish peroxidase–conjugated secondary antibody (1:500) for 1 hour at room temperature. Peroxidase was visualized with 3,3′ diaminobenzidine, and the slides were counterstained with hematoxylin. Slides were examined using fluorescence microscopy (Zeisse), and the number of CD8+CD103+ T cells was counted from 5 different high-power areas.

Humanized PDX tumor model

Six- to 8-week-old NOD.Cg-PrkdcscidIl2rgtm1Sug/JicCrl (NOG) mice (Vital River) were used to establish the PDX mouse model as described previously (24). Tumor specimens of gastric adenocarcinoma were collected and snap frozen in optimal cutting temperature. Tissue blocks were sectioned (4 μm) and fixed in precold acetone at 4°C for 10 minutes. Tissue slides were then washed with PBS twice and stained with Mayer hematoxylin solution for 5 minutes. Slides were washed in warm running tap water for 10 minutes and counterstained in eosin for 30 seconds. Slides were mounted with xylene-based mounting medium. Immune infiltrates in the TME were evaluated under light microscopy (Supplementary Fig. S1; Zeiss). Upon arrival, necrotic and supporting tissues were carefully removed using a surgical blade. Approximately 20 to 30 mg tissue fragments with immune infiltrates were implanted subcutaneously into the flank region of NOG mice. Successful established mouse models were monitored, and engrafted tumors were collected for analysis. Established PDX mice were treated with an anti–PD-L1 antibody (5 mg/kg) or isotype control (Bio X Cell). The mice were monitored 3 times per week for evidence of morbidity and mortality associated with tumor growth and metastasis. All animal procedures were conducted in accordance with, and with the approval of, the Institute Animal Care and Use Committee of Sun Yat-sen University.

Statistical analysis

Data were presented as mean ± SEM. Statistical analysis was performed using SPSS (version 13.0 for windows; SPSS). Comparisons were assessed using either the Student t test, paired Student t test, or one-way ANOVA with or without repeated measurements followed by Bonferroni multiple comparison posttest, as appropriate. The Cox univariate and multivariate analyses were used to explore the influences of different prognostic factors on overall survival. P values < 0.05 were considered as statistically significant.

Trm predicted better survival in gastric adenocarcinoma

Trm cells are associated with better survival in lung cancer and ovarian cancer (17, 25). In this study, we investigated the role and the predictive potential of Trm cells in gastric adenocarcinoma. Trm cells were identified firstly in the TME of gastric adenocarcinoma by dual staining of CD8 and CD103. It revealed that CD8+CD103+ Trm cells were detectable in the TME of gastric adenocarcinoma (Fig. 1A), and CD103+ Trm cells were more abundant in primary tumor compared with metastatic tumor as quantified by IHC (Fig. 1B). To further define and assess Trm cells in gastric adenocarcinoma, single-cell suspensions from tumor tissues were analyzed by flow cytometry. Trm cells were identified as CD69-positive or CD103-positive, or both. Because the vast majority of CD103+ T cells were simultaneously CD69+ (Fig. 1C; Supplementary Fig. S2), Trm cells were defined as CD103+ or CD69+CD103+ in this study. Enhanced frequency of Trm cells was confirmed in tumor without metastasis as measured by flow cytometry (Fig. 1C and D).

Figure 1.

Trm cells in gastric adenocarcinoma. A, Immunofluorescence staining of CD8 (green) and CD103 (red) in tissue sections from gastric adenocarcinoma. Slides were counterstained with DAPI. Representative images are shown. B, Tissue sections from primary or local metastasized gastric adenocarcinoma were stained with anti-CD103 antibody. Slides were counterstained with hematoxylin. Representative images are shown (n = 180). C, Fresh tumor tissues from primary or local metastasized gastric adenocarcinoma were collected, and single-cell suspension was prepared. Cells were stained with antibodies against CD8, CD69, and CD103 and analyzed by flow cytometry. Representative counter plots gated on CD8+ cells are shown. FMO, fluorescence minus one. D, Percentages of CD8+CD69+CD103+ T cells in primary (n = 24) or local metastasized (n = 8) gastric adenocarcinoma (mean ± SEM, t test). E, Fresh collected tumor tissues from gastric adenocarcinoma were engrafted into NOG mice to establish a PDX model. F, Mouse blood and tumors were collected 3 weeks after engraftment. Single cells were prepared and analyzed for CD69 and CD103 expression on CD45+CD8+ T cells by flow cytometry. Representative counter plots of eight experiments are shown. G, Kaplan–Meier plots representing the probability of overall survival (OS) in cohort 1 and cohort 2. ***, P < 0.001.

Figure 1.

Trm cells in gastric adenocarcinoma. A, Immunofluorescence staining of CD8 (green) and CD103 (red) in tissue sections from gastric adenocarcinoma. Slides were counterstained with DAPI. Representative images are shown. B, Tissue sections from primary or local metastasized gastric adenocarcinoma were stained with anti-CD103 antibody. Slides were counterstained with hematoxylin. Representative images are shown (n = 180). C, Fresh tumor tissues from primary or local metastasized gastric adenocarcinoma were collected, and single-cell suspension was prepared. Cells were stained with antibodies against CD8, CD69, and CD103 and analyzed by flow cytometry. Representative counter plots gated on CD8+ cells are shown. FMO, fluorescence minus one. D, Percentages of CD8+CD69+CD103+ T cells in primary (n = 24) or local metastasized (n = 8) gastric adenocarcinoma (mean ± SEM, t test). E, Fresh collected tumor tissues from gastric adenocarcinoma were engrafted into NOG mice to establish a PDX model. F, Mouse blood and tumors were collected 3 weeks after engraftment. Single cells were prepared and analyzed for CD69 and CD103 expression on CD45+CD8+ T cells by flow cytometry. Representative counter plots of eight experiments are shown. G, Kaplan–Meier plots representing the probability of overall survival (OS) in cohort 1 and cohort 2. ***, P < 0.001.

Close modal

To further investigate the tissue residency of Trm in the TME, a PDX tumor-bearing mouse model was established by engrafting tumor tissues from patients with gastric adenocarcinoma into the immunodeficient mice. By gating on human CD45+CD8+ population, we found that the CD69+CD103+ Trm cells resided in the TME and did not recirculate back to the blood. The population of CD8+ T cells that recirculated back to the blood is CD69CD103 (Fig. 1E and F; Supplementary Fig. S2). We next assessed the relationship between Trm cells and patient survival. CD103+ Trm cells were quantified (>5 CD103+ TILs per 0.6 mm core) to stratify the cohort into high-density versus low-density subsets as previously described (25). Using this strategy, we found that the presence of high-density CD103+ Trm cells positively correlated with survival in two cohorts of patients with gastric adenocarcinoma (Fig. 1G). The number of Trm cells was negatively correlated with the number of CTCs (Supplementary Fig. S3), indicating the potential function of Trm cells in suppressing tumor metastasis. Both univariate analysis and multivariate analysis performed using the COX proportional hazard regression model showed that the number of CD103+ Trm cells was an independent prognosis-related marker for gastric adenocarcinoma (Supplementary Tables S4 and S5).

Tumor reactivity of Trm cells in gastric adenocarcinoma

TILs highly express high immune inhibitory molecules (26). To understand the expression patterns of the immune inhibitory molecules in Trm cells, we evaluated the expression of PD-1, TIGIT, CD39, and CD26 in Trm cells from gastric adenocarcinoma specimens by flow cytometry (Fig. 2A and B). We found that the expression of PD-1, TIGIT, and CD39 correlated with that of CD103. CD103Hi subpopulation expressed the highest amount of inhibitory molecules, whereas CD26 expression did not correlate to that of CD103 (Fig. 2BF). High PD-1/PD-L1 in the TME predicts better immune response and outcomes in patients with cancer to checkpoint blockade therapy (27). To investigate the antitumor response by Trm cells, we sorted the CD103Hi, CD103Med, and CD103Neg populations and cultured the cells with or without autologous tumor cells. We found that cytokine production of TILs from patients with gastric adenocarcinoma correlated with CD103 expression. CD103Hi cells produced lower amounts IFNγ and TNFα when cultured alone as measured by ELISA. On the contrary, CD103Hi cells produced higher amounts IFNγ and TNFα in the supernatant when cultured with autologous tumor cells (Fig. 2G). These results were further confirmed by intracellular cytokine production measured by flow cytometry (Fig. 2H and I). These data suggested that Trm cells are tumor-reactive TILs.

Figure 2.

Tumor reactivity of Trm cells in gastric adenocarcinoma. A, Single-cell suspensions from gastric adenocarcinoma were analyzed by flow cytometry. Gating strategy of CD8+CD103+ cells according to CD103 expression for subpopulation analysis. Representative histogram plots of gastric adenocarcinoma specimens are shown. B, Coexpression of PD-1, TIGIT, CD39, and CD26 on CD103 subpopulations was analyzed by flow cytometry. Representative histograms are shown. C–F, PD-1, TIGIT, CD39, and CD26 expression on CD103 subpopulations was summarized from 12 samples. CD103 subpopulations were FACS sorted. Cells were cultured with or without autologous tumor cells. Tumor reactivity of CD103 subpopulations was evaluated by measuring IFNγ and TNFα in supernatant by ELISA (G) or intracellular cytokine measuring by flow cytometry (H and I). Data are mean ± SEM. *, P < 0.05; **, P < 0.01; and ***, P < 0.001 by one-way ANOVA. MFI, mean fluorescence intensity; ns, not significant.

Figure 2.

Tumor reactivity of Trm cells in gastric adenocarcinoma. A, Single-cell suspensions from gastric adenocarcinoma were analyzed by flow cytometry. Gating strategy of CD8+CD103+ cells according to CD103 expression for subpopulation analysis. Representative histogram plots of gastric adenocarcinoma specimens are shown. B, Coexpression of PD-1, TIGIT, CD39, and CD26 on CD103 subpopulations was analyzed by flow cytometry. Representative histograms are shown. C–F, PD-1, TIGIT, CD39, and CD26 expression on CD103 subpopulations was summarized from 12 samples. CD103 subpopulations were FACS sorted. Cells were cultured with or without autologous tumor cells. Tumor reactivity of CD103 subpopulations was evaluated by measuring IFNγ and TNFα in supernatant by ELISA (G) or intracellular cytokine measuring by flow cytometry (H and I). Data are mean ± SEM. *, P < 0.05; **, P < 0.01; and ***, P < 0.001 by one-way ANOVA. MFI, mean fluorescence intensity; ns, not significant.

Close modal

Increased lipid uptake and metabolism in Trm cells

Metabolic regulation is critical for T-cell activation and effector functions (28). Specifically, lipid metabolism enhances CD8+ T-cell memory (29). To understand the metabolic status of Trm cells, we first measured the expression of CD36, which imports lipids into the cells (30), in CD103Hi, CD103med, and CD103Neg populations by Western blot. CD103Hi Trm cells had the highest expression of CD36 when compared with CD103Med and CD103Neg Trm cells (Fig. 3A). Along with increased CD36 expression, lipid content (measured by Bodipy 493) was dramatically increased in CD103Hi Trm cells (Fig. 3B). In addition, by loading the cells with Bodipy 500 to measure lipid uptake, CD103Hi cells showed a clear increase in lipid uptake (Fig. 3B, D, and E). However, glucose uptake (measured by 2-NBDG) of CD103Hi Trm cells decreased when compared with CD103med or CD103Neg Trm cells (Fig. 3B and C). To further detail the metabolic alterations in Trm cells, a Seahorse assay was performed to measure glycolysis and mitochondria oxygen consumption. CD103Hi Trm cells showed decreased extracellular acidification rate (ECAR; Fig. 3F) and increased oxygen consumption rate (OCR; Fig. 3G). CD103Hi Trm cells had significantly higher basal metabolic activity, ATP-coupled OCR, stronger spare mitochondrial capacity, and maximum respiratory (Fig. 3HK). These data implied that Trm cells underwent a metabolic reprogram and switched to mitochondria fatty acid oxidation to meet their energy requirements.

Figure 3.

Glucose and lipid metabolism in Trm cells from gastric adenocarcinoma. A, Single-cell suspensions were prepared from gastric adenocarcinoma–derived tumor specimens. CD103 subpopulations were FACS sorted. CD36 expression on the CD103+ subpopulations was measured by Western blot. Representative bands are shown. B, Sorted CD103+ subpopulations were incubated with 2-NBDG (glucose uptake), Bodipy 493 (lipid content), and Bodipy 500 (lipid uptake). Cells were analyzed by flow cytometry, and representative histograms are shown. C–E, Data are summarized from 12 samples. F and G, Sorted CD103 subpopulations were analyzed with a Seahorse Bioscience XF96 analyzer. Glycolytic activity was measured by ECAR (F), and mitochondrial activities were measured by OCR (G). Summarized baseline respiration (H), respiration coupled to ATP production (I), respiratory spare capacity (J), and maximal (Max.) respiration (K). Data are mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; and ****, P < 0.0001 by one-way ANOVA. FCCP, carbonyl cyanide p-trifluoromethoxyphenylhydrazone; MFI, mean fluorescence intensity; min, minute(s); ns, not significant.

Figure 3.

Glucose and lipid metabolism in Trm cells from gastric adenocarcinoma. A, Single-cell suspensions were prepared from gastric adenocarcinoma–derived tumor specimens. CD103 subpopulations were FACS sorted. CD36 expression on the CD103+ subpopulations was measured by Western blot. Representative bands are shown. B, Sorted CD103+ subpopulations were incubated with 2-NBDG (glucose uptake), Bodipy 493 (lipid content), and Bodipy 500 (lipid uptake). Cells were analyzed by flow cytometry, and representative histograms are shown. C–E, Data are summarized from 12 samples. F and G, Sorted CD103 subpopulations were analyzed with a Seahorse Bioscience XF96 analyzer. Glycolytic activity was measured by ECAR (F), and mitochondrial activities were measured by OCR (G). Summarized baseline respiration (H), respiration coupled to ATP production (I), respiratory spare capacity (J), and maximal (Max.) respiration (K). Data are mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; and ****, P < 0.0001 by one-way ANOVA. FCCP, carbonyl cyanide p-trifluoromethoxyphenylhydrazone; MFI, mean fluorescence intensity; min, minute(s); ns, not significant.

Close modal

Lipid metabolism was required for the survival of Trm from gastric adenocarcinoma

CD8+ Trm cells tend to utilize mitochondrial fatty acid oxidation to support both their longevity and protective function (31). In this work, the role of lipid metabolism in the maintenance of Trm cells of gastric adenocarcinoma was explored. CD103Hi Trm cells showed increased apoptotic rates in vitro (Supplementary Fig. S4), which can be reversed by supplementing free fatty acids (FFA). However, FFAs did not change the apoptosis status of CD103Neg cells (Fig. 4A and B). FFAs also enhanced cytokine production by CD103Hi Trm cells (Fig. 4C and D). Seahorse assay showed FFAs did not affect ECAR by CD103Hi Trm cells. OCR by CD103Hi Trm cells was significantly increased when FFA was supplemented (Fig. 4E and F). These data suggested an important role of lipid metabolism in Trm cell survival by improving mitochondria activities in CD103Hi Trm cells. To further confirm the contribution of lipid metabolism to the survival of Trm cells in vivo, we established a PDX mouse model by engrafting tumor specimens from patients with gastric adenocarcinoma. PDX mice were treated with etomoxir, which inhibits fatty acid oxidation (Fig. 4G). Consistent with in vitro data, Etomoxir decreased the percentage of Trm cells in the TME (Fig. 4H and I). The total number of Trm cells was reduced dramatically by etomoxir (Fig. 4J), whereas the number of non-Trm cells in the TME was not affected by etomoxir treatment (Fig. 4K). These data demonstrated that FFAs improved mitochondria function of CD103Hi Trm cells, which was important for the survival of Trm cells.

Figure 4.

Lipid metabolism regulated survival of Trm cells from gastric adenocarcinoma. A and B, CD103Hi and CD103Neg cells were cultured in the presence or absence of palmitate (Palm, fatty acid) or vehicle for 24 hours. The apoptotic rate was measured by flow cytometry. Data are from six experiments. C and D, CD103Hi cells were cultured in the presence or absence of Palm. IFNγ and TNFα production was measured by flow cytometry. Representative count plots are shown. E and F, ECAR and mitochondrial activities were measured by OCR by Seahorse Bioscience XF96 analyzer. G, PDX was established as in Fig. 1. Mice were treated with etomoxir (CPT1 inhibitor) for 1 week. CD8+CD69+CD103+ cells in TILs were measured by flow cytometry. H, Representative counter plots of CD8+CD69+CD103+ cells are shown. Percentages (I) and number (J) of CD8+CD69+CD103+ cells in the tumor tissue. K, Number of CD103 CD8+ T cells in the tumor tissue. Data are mean ± SEM. *, P < 0.05; **, P < 0.01; and ***, P < 0.001 by t test. FCCP, carbonyl cyanide p-trifluoromethoxyphenylhydrazone; min, minute(s); ns, not significant.

Figure 4.

Lipid metabolism regulated survival of Trm cells from gastric adenocarcinoma. A and B, CD103Hi and CD103Neg cells were cultured in the presence or absence of palmitate (Palm, fatty acid) or vehicle for 24 hours. The apoptotic rate was measured by flow cytometry. Data are from six experiments. C and D, CD103Hi cells were cultured in the presence or absence of Palm. IFNγ and TNFα production was measured by flow cytometry. Representative count plots are shown. E and F, ECAR and mitochondrial activities were measured by OCR by Seahorse Bioscience XF96 analyzer. G, PDX was established as in Fig. 1. Mice were treated with etomoxir (CPT1 inhibitor) for 1 week. CD8+CD69+CD103+ cells in TILs were measured by flow cytometry. H, Representative counter plots of CD8+CD69+CD103+ cells are shown. Percentages (I) and number (J) of CD8+CD69+CD103+ cells in the tumor tissue. K, Number of CD103 CD8+ T cells in the tumor tissue. Data are mean ± SEM. *, P < 0.05; **, P < 0.01; and ***, P < 0.001 by t test. FCCP, carbonyl cyanide p-trifluoromethoxyphenylhydrazone; min, minute(s); ns, not significant.

Close modal

Cancer cells deprive Trm cells of lipid uptake, inducing their apoptosis

Metabolic shifting such as glucose deprivation in the TME impairs the antitumor effects of CD8+ T cells (32, 33). Thus, we next explored the metabolic alterations in the TME and the metabolic reprogramming in Trm cells from TME of gastric adenocarcinoma. By coculturing CD103Hi and CD103Neg T cells with gastric adenocarcinoma cells, we found that gastric adenocarcinoma cells increased the apoptosis of CD103Hi Trm cells, whereas the CD103Neg T cells were unaffected (Fig. 5A and B). Because lipid metabolism was important for Trm cells' survival (Fig. 4), we hypothesized that gastric adenocarcinoma cells deprive lipid uptake of Trm cells and induce Trm cell apoptosis. We thus measured the expression of Fabp4 and Fabp5 in gastric adenocarcinoma specimens by Western blot. We found that Fabp4 and Fabp5 expression significantly increased in gastric adenocarcinoma tumors when compared with para-cancer normal gastric tissue (Fig. 5C). Fabp4 and Fabp5 expression was also higher in gastric adenocarcinoma cells (SGC-7901) compared with healthy gastric epithelial cells (GES-1) as measured by flow cytometry (Fig. 5D and E). In contrast to GES-1, lipid uptake increased in SGC-7901 (Fig. 5F and G). Our previous study also shows gastric tumor cells outcompete CD8+ T cells for glucose consumption (34). Next, we investigated whether lipid metabolism in cancer cells would affect lipid uptake and apoptosis of Trm cells. The results revealed that lipid uptake by CD103Hi Trm cells was reduced when cocultured with cancer cells (Fig. 5H and I). In the tumor cell–T-cell coculture system, lipid uptake in Trm cells obviously increased with knockdown of both Fabp4 and Fabp5 in cancer cells when compared with knockdown of either Fabp4 or Fabp5 (Fig. 5J and K). Knockdown of Fabp4 or Fabp5 in cancer cells decreased the apoptosis of Trm cells. And knockdown of Fabp4 and Fabp5 together further decreased the apoptosis of Trm cells (Fig. 5L and M). These data demonstrated that gastric adenocarcinoma cells could induce Trm cells' apoptosis by depriving lipid uptake by Trm cells, which further confirmed that FFAs alone did not affect the Trm cells in the TME (Supplementary Fig. S5).

Figure 5.

Tumor cells deprived Trm cells of lipid uptake and induced their apoptosis. A, CD103Hi and CD103Neg subpopulations were sorted as in Fig. 2. CD103Hi and CD103Neg were cocultured with gastric cancer cells (SGC7901) at a ratio of 5:1 for 24 hours. Cells were stained with Annexin V and 7-AAD. Apoptosis was measured by flow cytometry. B, Percentages of apoptosis summarized from 6 samples. C, Expressions of Fabp4 and Fabp5 in gastric adenocarcinoma and normal gastric tissue were measured by Western blot. Representative bands are shown. D and E, Fabp4 and Fabp5 expression in SGC7901 and normal gastric epithelial cells (GES-1) was measured by flow cytometry. Representative histograms and data are from three experiments. F and G, SGC7901 and GES-1 cells were incubated with Bodipy 500, and the lipid uptake was measured by flow cytometry. Representative histograms and data are from three experiments. H and I, CD103Hi cells were cultured with or without SGC7901 and with Bodipy 500. Lipid uptake was measured by flow cytometry. Representative histograms and data are from three experiments. J and K, CD103Hi cells were cocultured with SGC7901 that were Fabp4, Fabp5, or dual knockdown. Lipid uptake was measured by incubating cells with Bodipy 500 and analyzed by flow cytometry. Representative histograms and data were from three experiments. L and M, CD103Hi cells were cocultured with SGC7901 that were Fabp4, Fabp5, or dual knockdown for 24 hours. Cells were stained with Annexin V and 7-AAD for apoptosis analysis by flow cytometry. Representative histograms and data are from three experiments. Data are mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; and ****, P < 0.0001 by t test in B, E, G, and I and one-way ANOVA in K and M. KD, knockdown; MFI, mean fluorescence intensity; ns, not significant.

Figure 5.

Tumor cells deprived Trm cells of lipid uptake and induced their apoptosis. A, CD103Hi and CD103Neg subpopulations were sorted as in Fig. 2. CD103Hi and CD103Neg were cocultured with gastric cancer cells (SGC7901) at a ratio of 5:1 for 24 hours. Cells were stained with Annexin V and 7-AAD. Apoptosis was measured by flow cytometry. B, Percentages of apoptosis summarized from 6 samples. C, Expressions of Fabp4 and Fabp5 in gastric adenocarcinoma and normal gastric tissue were measured by Western blot. Representative bands are shown. D and E, Fabp4 and Fabp5 expression in SGC7901 and normal gastric epithelial cells (GES-1) was measured by flow cytometry. Representative histograms and data are from three experiments. F and G, SGC7901 and GES-1 cells were incubated with Bodipy 500, and the lipid uptake was measured by flow cytometry. Representative histograms and data are from three experiments. H and I, CD103Hi cells were cultured with or without SGC7901 and with Bodipy 500. Lipid uptake was measured by flow cytometry. Representative histograms and data are from three experiments. J and K, CD103Hi cells were cocultured with SGC7901 that were Fabp4, Fabp5, or dual knockdown. Lipid uptake was measured by incubating cells with Bodipy 500 and analyzed by flow cytometry. Representative histograms and data were from three experiments. L and M, CD103Hi cells were cocultured with SGC7901 that were Fabp4, Fabp5, or dual knockdown for 24 hours. Cells were stained with Annexin V and 7-AAD for apoptosis analysis by flow cytometry. Representative histograms and data are from three experiments. Data are mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; and ****, P < 0.0001 by t test in B, E, G, and I and one-way ANOVA in K and M. KD, knockdown; MFI, mean fluorescence intensity; ns, not significant.

Close modal

PD-L1 regulated lipid competition in the TME

PD-1 altered T-cell metabolic reprogramming by inhibiting glucose glycolysis and increasing mitochondria fatty acid oxidation (35). We investigated if PD-L1, the ligand of PD-1, affects lipid uptake and metabolism in cancer cells and Trm cells. In the tumor cell–T-cell coculture system, we found that PD-L1 blockade decreased Fabp4/5 expression in tumor cells, whereas Fabp4/5 expression in T cells was increased when PD-L1 was blocked (Fig. 6A and B). Lipid uptake by cancer cells was reduced by blocking PD-L1 (Fig. 6C and D). Importantly, lipid uptake in CD103Hi Trm cells was increased when PD-L1 was blocked (Fig. 6E and F). We confirmed these data by knocking down PD-L1 expression in tumor cells. Knockdown of PD-L1 expression in tumor cells decreased lipid uptake by tumor cells and increased lipid uptake in CD103Hi Trm cells (Supplementary Fig. S6). Lipid uptake by CD103Hi Trm cells was decreased when PD-1 was knocked down (Supplementary Fig. S7). Meanwhile, the apoptotic rate of CD103Hi Trm cells was decreased when PD-L1 was blocked (Fig. 6G and H). These data indicated that the inhibition of PD-L1 in the TME could improve the survival of Trm cells and enhance antitumor immune response. To further confirm this result, the PDX model was established as described in Fig. 1, and the tumor-bearing mice were treated with PD-L1–blocking antibody (Fig. 6I). After the treatment, Trm cells in the TME were analyzed by flow cytometry. We found that the percentage of Trm cells was increased after PD-L1 blockade. The absolute number of Trm cells was also increased by blocking PD-L1 in the PDX mice (Fig. 6J and K).

Figure 6.

PD-L1 blockade promoted Trm cell survival. Gastric cancer cells (SGC7901) were treated with IFNγ (10 ng/mL) for 24 hours. CD103Hi CD8+ T cells were then sorted by flow cytometry and cocultured with SGC7901 in the presence of anti–PD-L1 antibody or isotype control for 24 hours. A and B, Fabp4 and Fabp5 expression in tumor cells and T cells was measured by flow cytometry. Representative histograms are shown, and data are from 5 patient samples. C–F, Cells were incubated with Bodipy 500, and lipid uptake by tumor cells and T cells was assessed by flow cytometry. Representative histograms and data are from five experiments. G and H, Annexin V expression in T cells was measured by flow cytometry. Representative histograms gated on CD8+ cells and data are from 5 patient samples. I, PDX model was established as in Fig. 1. PDX-bearing mice were then treated with anti–PD-L1 or isotype for 2 weeks. J and K, Single-cell suspension was prepared from tumor xenografts, and CD8+CD69+CD103+ T cells in TILs were analyzed by flow cytometry. Percentages (J) and cell numbers (K) are shown (n = 12). Data are mean ± SEM. *, P < 0.05 and **, P < 0.01 by t test in B, D, F, and H and paired t test in J and K. MFI, mean fluorescence intensity.

Figure 6.

PD-L1 blockade promoted Trm cell survival. Gastric cancer cells (SGC7901) were treated with IFNγ (10 ng/mL) for 24 hours. CD103Hi CD8+ T cells were then sorted by flow cytometry and cocultured with SGC7901 in the presence of anti–PD-L1 antibody or isotype control for 24 hours. A and B, Fabp4 and Fabp5 expression in tumor cells and T cells was measured by flow cytometry. Representative histograms are shown, and data are from 5 patient samples. C–F, Cells were incubated with Bodipy 500, and lipid uptake by tumor cells and T cells was assessed by flow cytometry. Representative histograms and data are from five experiments. G and H, Annexin V expression in T cells was measured by flow cytometry. Representative histograms gated on CD8+ cells and data are from 5 patient samples. I, PDX model was established as in Fig. 1. PDX-bearing mice were then treated with anti–PD-L1 or isotype for 2 weeks. J and K, Single-cell suspension was prepared from tumor xenografts, and CD8+CD69+CD103+ T cells in TILs were analyzed by flow cytometry. Percentages (J) and cell numbers (K) are shown (n = 12). Data are mean ± SEM. *, P < 0.05 and **, P < 0.01 by t test in B, D, F, and H and paired t test in J and K. MFI, mean fluorescence intensity.

Close modal

PD-L1 blockade unleashed Trm cells and their antitumor effects

The success of anti–PD-1/PD-L1 in treatment of cancers has led to a paradigm shift in the oncology field (36). We investigated how Trm cells were involved in antitumor immune response. We first tested cytotoxicity of Trm cells to tumor cells. We found that CD103Hi Trm cells showed more effectively induced tumor cell death when compared with CD103Neg cells (Fig. 7A and B). In addition, PD-L1 blockade enhanced cytotoxic function of CD103Hi Trm cells mostly (Fig. 7C and D). To further study CD103Hi Trm cells in antitumor immune response in vivo, a PDX model was established as in Fig. 1. The tumor-bearing mice were treated with anti–PD-L1 antibody or isotype control (Fig. 7E). IFNγ production in TILs was measured by flow cytometry after the treatment. We found that PD-L1 blockade mainly affected the CD103Hi Trm cells through stimulating IFNγ production, whereas there seemed to be no evident effects on CD103Neg (Fig. 7F and G). We monitored tumor growth by measuring the tumor volume in mice treated with anti–PD-L1 antibody. Seven PDX mice responded to the anti–PD-L1 treatment, whereas tumor growth in 11 of the treated mice was not affected by anti–PD-L1 treatment (Fig. 7H). The percentage of Trm cells was significantly higher in the treatment-responsive mice compared with the nonresponsive mice (Fig. 7I and J). These data indicated that CD8+ CD103Hi Trm cells contributed significantly to the antitumor immune response to gastric adenocarcinoma.

Figure 7.

Cytotoxicity of Trm cells correlated with antitumor response to PD-L1 blockade. A and B, Autologous tumor cells were cocultured with or without CD103Hi and CD103Neg CD8+ T cells for 16 hours. Cells were collected, and cell death was measured by staining cells with 7-AAD. Representative FACS plots were gated on CD45 tumor cells. Data are from 5 samples. C and D, Autologous tumor cells were cocultured with CD103Hi and CD103Neg CD8+ T cells in the presence of anti–PD-L1 antibody or isotype control for 16 hours. Cell death was measured by staining cells with 7-AAD. Representative FACS plots were gated on CD45 tumor cells. Data are from 5 samples. E, A PDX model was established and treated as in Fig. 6. F and G, TILs from anti–PD-L1– or isotype-treated mice were measured for IFNγ expression by flow cytometry. Percentages of IFNγ-producing T cells and representative counter plots gated on CD103Hi and CD103Neg CD8+ T cells (n = 8). H, Tumor volume of anti–PD-L1-treated mice. I and J, CD8+CD69+CD103+ cells in TILs from anti–PD-L1 treatment-responsive (n = 7) and nonresponsive (n = 11) mice were analyzed by flow cytometry. Representative counter plots. Data are mean ± SEM. *, P < 0.05; ***, P < 0.001; and ****, P < 0.0001 by one-way ANOVA in B, paired t test in D, and t test in G and J. ns, not significant.

Figure 7.

Cytotoxicity of Trm cells correlated with antitumor response to PD-L1 blockade. A and B, Autologous tumor cells were cocultured with or without CD103Hi and CD103Neg CD8+ T cells for 16 hours. Cells were collected, and cell death was measured by staining cells with 7-AAD. Representative FACS plots were gated on CD45 tumor cells. Data are from 5 samples. C and D, Autologous tumor cells were cocultured with CD103Hi and CD103Neg CD8+ T cells in the presence of anti–PD-L1 antibody or isotype control for 16 hours. Cell death was measured by staining cells with 7-AAD. Representative FACS plots were gated on CD45 tumor cells. Data are from 5 samples. E, A PDX model was established and treated as in Fig. 6. F and G, TILs from anti–PD-L1– or isotype-treated mice were measured for IFNγ expression by flow cytometry. Percentages of IFNγ-producing T cells and representative counter plots gated on CD103Hi and CD103Neg CD8+ T cells (n = 8). H, Tumor volume of anti–PD-L1-treated mice. I and J, CD8+CD69+CD103+ cells in TILs from anti–PD-L1 treatment-responsive (n = 7) and nonresponsive (n = 11) mice were analyzed by flow cytometry. Representative counter plots. Data are mean ± SEM. *, P < 0.05; ***, P < 0.001; and ****, P < 0.0001 by one-way ANOVA in B, paired t test in D, and t test in G and J. ns, not significant.

Close modal

Gastric cancer is one of the leading causes of cancer-related death. There has been a paradigm shift in cancer therapy in the past 5 to 10 years due to the successes of immunotherapy (37). Anti–CTLA-4 and anti–PD-1/PD-L1 treatments represent the most successful therapeutics targeting immune system to cure cancer (38). Data from clinical trials show the promising results using anti–PD-1/PD-L1 antibody to treat gastric cancer (8, 9). However, due to the low response rate, there is an urgent need to develop novel strategies to improve the therapeutic effect. In this study, we investigated the roles and the metabolic regulation of CD8+ Trm cells in gastric adenocarcinoma, focusing on the roles of fatty acid oxidation Trm cell death in gastric adenocarcinoma. We found that Trm cells were present in the TME of gastric adenocarcinoma and predicted better survival. Fatty acid oxidation was required for the survival of Trm cells. Gastric adenocarcinoma cells outcompeted Trm cells for lipid uptake, and PD-L1 blockade reversed this competition, unleashing Trm cells, leading to tumor regression.

Immune checkpoint blockade unleashes CD8+ T cells and enhances their antitumor response in the TME (39). Trm cells are T cells that reside in the tissue and provide stronger and faster immune response during antigen rechallenge (40). Trm cells can be harnessed to enhance the efficacy of cancer vaccines (41). CD103+ Trm cells are present in the TME and predict better outcomes in breast cancer, lung cancer, and ovarian cancer (16, 17, 25). In the current study, we first confirmed the presence of CD103+ Trm cells in the TME of gastric adenocarcinoma by detecting the coexpression pattern of CD8 and CD103. Meanwhile, these cells were also CD69+, which was in consistent with previous reports (17). Results from the PDX model proved that the CD69+CD103+ Trm cells were TILs that do not participate in lymphocyte recirculation. The percentage of CD8+ Trm cells was lower in patients with metastasis. Clinical data analysis showed that higher density of Trm cell, as defined as >5 CD103+ TILs per 0.6 mm core, was associated with better overall survival. We observed a negative correlation between the qualities of Trm cells and CTC, confirming the suppressive effect of Trm cells against metastasis. Together with the strong expression of the granzymes, perforin, and IFNγ in Trm cells (17), these results demonstrate the critical role of Trm cells in suppressing tumor growth and metastasis in gastric adenocarcinoma.

TILs highly express immune inhibitory molecules (42). To explore the special status of Trm cells, we measured the expression of PD-1, TIGIT, and CD39 on TILs from patients with gastric adenocarcinoma. CD103Hi cells more highly expressed PD-1, TIGIT, and CD39. PD-1 and TIGIT are highly expressed in TILs from gastric cancer (34). Coexpression of CD39 and CD103 identifies tumor-reactive T cells in human malignancies (43). CD39+ CD8+ T cells in human tumor infiltrates are antigen specific, and CD8+ TILs without CD39 expression could be defined as bystander (not responsive to tumors; ref. 44). The high CD39 expression on Trm cells indicates that Trm cells are antigen-specific T cells involved in the antitumor immune response. In this work, our results showed that CD103Hi Trm cells responded to autologous tumor cells by producing IFNγ and TNFα, whereas CD103Neg T cells barely responded to autologous tumor cells. From the PDX mouse experiments, we learned that the T cells that recirculated back to the blood were CD69CD103. Together, CD103Neg T cells could be identified as the so-called bystander in the TME. These data support the notion that Trm cells are the major effector T cells in response to cancer (43).

Energy demand is dramatically increased during T-cell activation, proliferation, and differentiation (28, 45). TILs display profound metabolic reprograming (19), whereas glucose uptake by cancer cells outcompetes TILs, which leads to impaired effector functions of CD8+ T cells (33). Glucose deprivation suppresses effector functions of TILs, and metabolic reprogramming of TILs enhances their antitumor response (46). In the current study, we found that CD103Hi Trm cells express the fatty acid translocase CD36. We also observed increased lipid uptake and mitochondrial activities synchronous with decreased glucose uptake and glycolytic activity in Trm cells. CD103Hi Trm cells from TME of gastric adenocarcinoma displayed higher amounts of apoptosis in in vitro culture, and FFAs rescued the Trm cells from apoptosis. By inhibiting fatty acid oxidation, the proportions and quantities of Trm cells decreased in the TME of a PDX model for gastric adenocarcinoma. These findings are consistent with a previous report showing that Trm cells rely on lipid metabolism for long-term survival (31). In the harsh TME, the supply of FFAs may be required for Trm cell maintenance and long-term survival.

Due to the high proliferation rate, cancer cells rely on glucose to generate energy in a fast but inefficient way through glycolysis (47). There is also an increased demand of lipids for the fast proliferating cancer cells for energy production and new membrane formation (48). Our study demonstrated that Fabp4/5 expression was increased in gastric adenocarcinoma cells and displayed increased lipid uptake compared with healthy gastric epithelial cells. In a T-cell–tumor cell coculture system, cancer cells suppressed the lipid uptake of Trm cells and induced Trm apoptosis. These effects could be reversed by inhibiting cancer cells from lipid uptake. These results unearthed the metabolic competition between cancer cells and Trm cells for lipid consumption, leading to the suppression of Trm cells.

PD-L1 is involved in the glucose metabolism of cancer cells (33), and PD-L1 is expressed in the TME of all stages of gastric adenocarcinoma (49), suggesting an important role of PD-L1 in the control of metabolism of gastric adenocarcinoma cells. However, the effects of PD-L1 in the regulation of lipid metabolism remain unknown. We found that PD-L1 inhibition decreased Fabp4/5 expression in gastric adenocarcinoma cells, whereas Fabp4/5 expression was increased in the T cells in the coculture system. The blockade of PD-L1 not only suppressed lipid uptake of cancer cells, but also rescued Trm cells from apoptosis in the TME. These findings were verified in the PDX models treated with anti–PD-L1 where an enrichment of Trm cells was seen. Taken together, cancer cells outcompete Trm cells for lipid uptake through PD-L1, which leads to the apoptosis of Trm cells in the TEM, dampening the antitumor immune response.

PD-1/PD-L1 blockade unleashes TILs and enhances antitumor responses to suppress tumor from growing (50). We found that CD103Hi Trm cells exhibited strong cytotoxic function in response to autologous tumor cells and anti–PD-L1 enhanced the cytotoxic function of CD103Hi Trm cells. In the PDX mice treated with anti–PD-L1 antibody, the mice responding to the treatment showed dramatically higher percentage of Trm cells in the TME. In contrast, the mice with progressing tumors presented low frequencies of Trm. Checkpoint blockade targets tumor-specific T cells (51), and CD103+ TILs represent the tumor-reactive T cells (43). Trm cells contribute to antitumor immune response by targeting PD-1/PD-L1.

Taken together, our data suggested a distinct role for tumor-specific Trm cells in mediating antitumor immunity and predicting treatment response to checkpoint blockade. Reprogramming lipid metabolism of Trm cells could be a promising therapeutic approach for gastric adenocarcinoma. Further investigation to explore the relevant mechanisms and potential clinical application targeting lipid metabolism reprogramming of Trm cells is warranted.

No potential conflicts of interest were disclosed.

Conception and design: S. Cai, Z. Ke, W. He

Development of methodology: W. He

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R. Lin, H. Zhang, Y. Yuan, Q. He, J. Zhou, S. Li, Y. Sun, H.-B. Qiu, W. Wang, Z. Zhuang, B. Chen, Y. Huang, C. Liu, Y. Wang, Z. Ke

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R. Lin, H. Zhang, Y. Yuan, D.Y. Li, W. He

Writing, review, and/or revision of the manuscript: R. Lin, H. Zhang, Y. Yuan, D.Y. Li, Z. Ke, W. He

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): W. He

Study supervision: S. Cai, Z. Ke, W. He

This work was supported by grants from the National Natural Science Foundation of China (30900650, 81372501, 81572260, 81871994, and 81701834), Guangdong Natural Science Foundation (2011B031800025, S2012010008378, 2015A030313036, 2017A010101030, and 2018A030310285), and the Guangzhou Science and Technology Planning Program (2014J4100132, 2015A020214010, 2012B031800115, 2013B02180021, 2016A020215055, 201904010398, and 201902020018).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Ajani
JA
,
Lee
J
,
Sano
T
,
Janjigian
YY
,
Fan
D
,
Song
S
. 
Gastric adenocarcinoma
.
Nat Rev Dis Primers
2017
;
3
:
17036
.
2.
Kim
Y
,
Spolverato
G
,
Ejaz
A
,
Squires
MH
,
Poultsides
G
,
Fields
RC
, et al
A nomogram to predict overall survival and disease-free survival after curative resection of gastric adenocarcinoma
.
Ann Surg Oncol
2015
;
22
:
1828
35
.
3.
Sitarz
R
,
Skierucha
M
,
Mielko
J
,
Offerhaus
GJA
,
Maciejewski
R
,
Polkowski
WP
. 
Gastric cancer: epidemiology, prevention, classification, and treatment
.
Cancer Manag Res
2018
;
10
:
239
48
.
4.
Drake
CG
,
Lipson
EJ
,
Brahmer
JR
. 
Breathing new life into immunotherapy: review of melanoma, lung and kidney cancer
.
Nat Rev Clin Oncol
2014
;
11
:
24
37
.
5.
Topalian
SL
,
Hodi
FS
,
Brahmer
JR
,
Gettinger
SN
,
Smith
DC
,
McDermott
DF
, et al
Safety, activity, and immune correlates of anti-PD-1 antibody in cancer
.
N Engl J Med
2012
;
366
:
2443
54
.
6.
Barlesi
F
,
Vansteenkiste
J
,
Spigel
D
,
Ishii
H
,
Garassino
M
,
de Marinis
F
, et al
Avelumab versus docetaxel in patients with platinum-treated advanced non-small-cell lung cancer (JAVELIN Lung 200): an open-label, randomised, phase 3 study
.
Lancet Oncol
2018
;
19
:
1468
79
.
7.
Hamid
O
,
Robert
C
,
Daud
A
,
Hodi
FS
,
Hwu
WJ
,
Kefford
R
, et al
Safety and tumor responses with lambrolizumab (anti-PD-1) in melanoma
.
N Engl J Med
2013
;
369
:
134
44
.
8.
Kang
YK
,
Boku
N
,
Satoh
T
,
Ryu
MH
,
Chao
Y
,
Kato
K
, et al
Nivolumab in patients with advanced gastric or gastro-oesophageal junction cancer refractory to, or intolerant of, at least two previous chemotherapy regimens (ONO-4538–12, ATTRACTION-2): a randomised, double-blind, placebo-controlled, phase 3 trial
.
Lancet
2017
;
390
:
2461
71
.
9.
Muro
K
,
Chung
HC
,
Shankaran
V
,
Geva
R
,
Catenacci
D
,
Gupta
S
, et al
Pembrolizumab for patients with PD-L1-positive advanced gastric cancer (KEYNOTE-012): a multicentre, open-label, phase 1b trial
.
Lancet Oncol
2016
;
17
:
717
26
.
10.
Mueller
SN
,
Mackay
LK
. 
Tissue-resident memory T cells: local specialists in immune defence
.
Nat Rev Immunol
2016
;
16
:
79
89
.
11.
Schenkel
JM
,
Masopust
D
. 
Tissue-resident memory T cells
.
Immunity
2014
;
41
:
886
97
.
12.
Watanabe
R
,
Gehad
A
,
Yang
C
,
Scott
LL
,
Teague
JE
,
Schlapbach
C
, et al
Human skin is protected by four functionally and phenotypically discrete populations of resident and recirculating memory T cells
.
Sci Transl Med
2015
;
7
:
279ra39
.
13.
Gebhardt
T
,
Wakim
LM
,
Eidsmo
L
,
Reading
PC
,
Heath
WR
,
Carbone
FR
. 
Memory T cells in nonlymphoid tissue that provide enhanced local immunity during infection with herpes simplex virus
.
Nat Immunol
2009
;
10
:
524
30
.
14.
Molodtsov
A
,
Turk
MJ
. 
Tissue resident CD8 memory T cell responses in cancer and autoimmunity
.
Front Immunol
2018
;
9
:
2810
.
15.
Corgnac
S
,
Boutet
M
,
Kfoury
M
,
Naltet
C
,
Mami-Chouaib
F
. 
The emerging role of CD8(+) tissue resident memory T (TRM) cells in antitumor immunity: a unique functional contribution of the CD103 integrin
.
Front Immunol
2018
;
9
:
1904
.
16.
Savas
P
,
Virassamy
B
,
Ye
C
,
Salim
A
,
Mintoff
CP
,
Caramia
F
, et al
Single-cell profiling of breast cancer T cells reveals a tissue-resident memory subset associated with improved prognosis
.
Nat Med
2018
;
24
:
986
93
.
17.
Ganesan
AP
,
Clarke
J
,
Wood
O
,
Garrido-Martin
EM
,
Chee
SJ
,
Mellows
T
, et al
Tissue-resident memory features are linked to the magnitude of cytotoxic T cell responses in human lung cancer
.
Nat Immunol
2017
;
18
:
940
50
.
18.
Park
SL
,
Buzzai
A
,
Rautela
J
,
Hor
JL
,
Hochheiser
K
,
Effern
M
, et al
Tissue-resident memory CD8(+) T cells promote melanoma-immune equilibrium in skin
.
Nature
2019
;
565
:
366
71
.
19.
Thommen
DS
,
Koelzer
VH
,
Herzig
P
,
Roller
A
,
Trefny
M
,
Dimeloe
S
, et al
A transcriptionally and functionally distinct PD-1(+) CD8(+) T cell pool with predictive potential in non-small-cell lung cancer treated with PD-1 blockade
.
Nat Med
2018
;
24
:
994
1004
.
20.
He
W
,
Hou
M
,
Zhang
H
,
Zeng
C
,
He
S
,
Chen
X
, et al
Clinical significance of circulating tumor cells in predicting disease progression and chemotherapy resistance in patients with gestational choriocarcinoma
.
Int J Cancer
2019
;
144
:
1421
31
.
21.
Li
G
,
Wang
Z
,
Ye
J
,
Zhang
X
,
Wu
H
,
Peng
J
, et al
Uncontrolled inflammation induced by AEG-1 promotes gastric cancer and poor prognosis
.
Cancer Res
2014
;
74
:
5541
52
.
22.
LaFleur
MW
,
Nguyen
TH
,
Coxe
MA
,
Miller
BC
,
Yates
KB
,
Gillis
JE
, et al
PTPN2 regulates the generation of exhausted CD8(+) T cell subpopulations and restrains tumor immunity
.
Nat Immunol
2019
;
20
:
1335
47
.
23.
Wang
W
,
Green
M
,
Choi
JE
,
Gijon
M
,
Kennedy
PD
,
Johnson
JK
, et al
CD8(+) T cells regulate tumour ferroptosis during cancer immunotherapy
.
Nature
2019
;
569
:
270
4
.
24.
Gao
H
,
Korn
JM
,
Ferretti
S
,
Monahan
JE
,
Wang
Y
,
Singh
M
, et al
High-throughput screening using patient-derived tumor xenografts to predict clinical trial drug response
.
Nat Med
2015
;
21
:
1318
25
.
25.
Webb
JR
,
Milne
K
,
Watson
P
,
Deleeuw
RJ
,
Nelson
BH
. 
Tumor-infiltrating lymphocytes expressing the tissue resident memory marker CD103 are associated with increased survival in high-grade serous ovarian cancer
.
Clin Cancer Res
2014
;
20
:
434
44
.
26.
Pardoll
DM
. 
The blockade of immune checkpoints in cancer immunotherapy
.
Nat Rev Cancer
2012
;
12
:
252
64
.
27.
Nishino
M
,
Ramaiya
NH
,
Hatabu
H
,
Hodi
FS
. 
Monitoring immune-checkpoint blockade: response evaluation and biomarker development
.
Nat Rev Clin Oncol
2017
;
14
:
655
68
.
28.
Fox
CJ
,
Hammerman
PS
,
Thompson
CB
. 
Fuel feeds function: energy metabolism and the T-cell response
.
Nat Rev Immunol
2005
;
5
:
844
52
.
29.
Pearce
EL
,
Walsh
MC
,
Cejas
PJ
,
Harms
GM
,
Shen
H
,
Wang
LS
, et al
Enhancing CD8 T-cell memory by modulating fatty acid metabolism
.
Nature
2009
;
460
:
103
7
.
30.
Silverstein
RL
,
Febbraio
M
. 
CD36, a scavenger receptor involved in immunity, metabolism, angiogenesis, and behavior
.
Sci Signal
2009
;
2
:
re3
.
31.
Pan
Y
,
Tian
T
,
Park
CO
,
Lofftus
SY
,
Mei
S
,
Liu
X
, et al
Survival of tissue-resident memory T cells requires exogenous lipid uptake and metabolism
.
Nature
2017
;
543
:
252
6
.
32.
Scharping
NE
,
Menk
AV
,
Moreci
RS
,
Whetstone
RD
,
Dadey
RE
,
Watkins
SC
, et al
The tumor microenvironment represses T cell mitochondrial biogenesis to drive intratumoral T cell metabolic insufficiency and dysfunction
.
Immunity
2016
;
45
:
374
88
.
33.
Chang
CH
,
Qiu
J
,
O'Sullivan
D
,
Buck
MD
,
Noguchi
T
,
Curtis
JD
, et al
Metabolic competition in the tumor microenvironment is a driver of cancer progression
.
Cell
2015
;
162
:
1229
41
.
34.
Chevrier
S
,
Levine
JH
,
Zanotelli
VRT
,
Silina
K
,
Schulz
D
,
Bacac
M
, et al
An immune atlas of clear cell renal cell carcinoma
.
Cell
2017
;
169
:
736
49.e18
.
35.
Patsoukis
N
,
Bardhan
K
,
Chatterjee
P
,
Sari
D
,
Liu
B
,
Bell
LN
, et al
PD-1 alters T-cell metabolic reprogramming by inhibiting glycolysis and promoting lipolysis and fatty acid oxidation
.
Nat Commun
2015
;
6
:
6692
.
36.
Chen
L
,
Han
X
. 
Anti-PD-1/PD-L1 therapy of human cancer: past, present, and future
.
J Clin Invest
2015
;
125
:
3384
91
.
37.
Ribas
A
. 
Releasing the brakes on cancer immunotherapy
.
N Engl J Med
2015
;
373
:
1490
2
.
38.
Boutros
C
,
Tarhini
A
,
Routier
E
,
Lambotte
O
,
Ladurie
FL
,
Carbonnel
F
, et al
Safety profiles of anti-CTLA-4 and anti-PD-1 antibodies alone and in combination
.
Nat Rev Clin Oncol
2016
;
13
:
473
86
.
39.
Tumeh
PC
,
Harview
CL
,
Yearley
JH
,
Shintaku
IP
,
Taylor
EJ
,
Robert
L
, et al
PD-1 blockade induces responses by inhibiting adaptive immune resistance
.
Nature
2014
;
515
:
568
71
.
40.
Park
CO
,
Kupper
TS
. 
The emerging role of resident memory T cells in protective immunity and inflammatory disease
.
Nat Med
2015
;
21
:
688
97
.
41.
Nizard
M
,
Roussel
H
,
Diniz
MO
,
Karaki
S
,
Tran
T
,
Voron
T
, et al
Induction of resident memory T cells enhances the efficacy of cancer vaccine
.
Nat Commun
2017
;
8
:
15221
.
42.
Baumeister
SH
,
Freeman
GJ
,
Dranoff
G
,
Sharpe
AH
. 
Coinhibitory pathways in immunotherapy for cancer
.
Annu Rev Immunol
2016
;
34
:
539
73
.
43.
Duhen
T
,
Duhen
R
,
Montler
R
,
Moses
J
,
Moudgil
T
,
de Miranda
NF
, et al
Co-expression of CD39 and CD103 identifies tumor-reactive CD8 T cells in human solid tumors
.
Nat Commun
2018
;
9
:
2724
.
44.
Simoni
Y
,
Becht
E
,
Fehlings
M
,
Loh
CY
,
Koo
SL
,
Teng
KWW
, et al
Bystander CD8(+) T cells are abundant and phenotypically distinct in human tumour infiltrates
.
Nature
2018
;
557
:
575
9
.
45.
MacIver
NJ
,
Michalek
RD
,
Rathmell
JC
. 
Metabolic regulation of T lymphocytes
.
Annu Rev Immunol
2013
;
31
:
259
83
.
46.
Ho
PC
,
Bihuniak
JD
,
Macintyre
AN
,
Staron
M
,
Liu
X
,
Amezquita
R
, et al
Phosphoenolpyruvate is a metabolic checkpoint of anti-tumor T cell responses
.
Cell
2015
;
162
:
1217
28
.
47.
Hsu
PP
,
Sabatini
DM
. 
Cancer cell metabolism: Warburg and beyond
.
Cell
2008
;
134
:
703
7
.
48.
Rohrig
F
,
Schulze
A
. 
The multifaceted roles of fatty acid synthesis in cancer
.
Nat Rev Cancer
2016
;
16
:
732
49
.
49.
Thompson
ED
,
Zahurak
M
,
Murphy
A
,
Cornish
T
,
Cuka
N
,
Abdelfatah
E
, et al
Patterns of PD-L1 expression and CD8 T cell infiltration in gastric adenocarcinomas and associated immune stroma
.
Gut
2017
;
66
:
794
801
.
50.
Taube
JM
. 
Unleashing the immune system: PD-1 and PD-Ls in the pre-treatment tumor microenvironment and correlation with response to PD-1/PD-L1 blockade
.
Oncoimmunology
2014
;
3
:
e963413
.
51.
Gubin
MM
,
Zhang
X
,
Schuster
H
,
Caron
E
,
Ward
JP
,
Noguchi
T
, et al
Checkpoint blockade cancer immunotherapy targets tumour-specific mutant antigens
.
Nature
2014
;
515
:
577
81
.