CD1d-restricted invariant natural killer T (iNKT) cells actively patrol the liver and possess valuable antitumor potential. However, clinical trials evaluating administration of iNKT cell–specific agonist α-galactosylceramide (α-GalCer) have failed to achieve obvious tumor regression. Improving the efficacy of iNKT cell–based immunotherapy requires a better understanding of the factors restraining the clinical benefits. In the context of hepatitis B virus (HBV)-related hepatocellular carcinoma (HCC), we found circulating and hepatic iNKT cells were hyperactivated but demonstrated imbalances in ratio and defective α-GalCer responsiveness. Exogenous IL2 helped to expand residual α-GalCer–responsive clones with reduced T-cell receptor diversity. However, transcriptome-wide analysis revealed activation of the senescence-associated secretory phenotype and dampened cytotoxicity in iNKT cells, weakening their immune surveillance capacity. The senescent status of iNKT cells from the patients was further illustrated by cell-cycle arrest, impaired telomere maintenance, perturbed calcium transport-related biological processes, and altered metabolism. Lipidomic profiling revealed the accumulation of long-chain acylcarnitines (LCAC) and aberrant lipid metabolism in HCC tissue. Exogenous LCACs, especially palmitoyl-carnitine and stearoyl-carnitine, inhibited iNKT cell expansion and promoted senescence. Collectively, our results provide deeper insights into iNKT cell dysregulation and identify a cell senescence–associated challenge for iNKT cell–based immunotherapy in HBV-related HCC. The mechanistic links between iNKT cell senescence and accumulated LCACs suggest new targets for anti-HCC immunotherapies.

Significance:

Patients with HBV-related HCC exhibit a cell senescence–associated dysregulation of invariant natural killer cells that is related to altered lipid metabolism and accumulated LCACs in tumor tissue.

Hepatocellular carcinoma (HCC) is a major primary liver malignancy developed from a setting of chronic liver diseases. Chronic hepatitis B virus (HBV) infection accounts for approximately 50% of total cases and virtually all of childhood HCC (1). Although 90% of HCCs are attributed to underlying chronic liver inflammation, insufficient T-cell response contributes to consequent impaired antitumor immune surveillance during persistent HBV infection (2). Invariant natural killer T (iNKT) cells are an innate-like T-cell subset recognizing lipid antigens presented by monomorphic CD1d molecules. They actively patrol the liver, directly kill target cells and act as potent immunologic modulators (3). Because placed at the interface between innate and adaptive immunity, iNKT cells have great potential to shape the immune response and are well known for their antitumor function in both humans and mice. Although human iNKT cells only range 0.001% to 0.1% in blood lymphocytes and 0.01% to 5% in hepatic lymphocytes, the frequencies are still much higher than that of naïve peptide-specific T cells in the order of magnitude (4–6). With the discovery of α-galactosylceramide (α-GalCer), a specific agonist ligand for iNKT cells, it becomes possible to specifically stimulate and expand iNKT cells (7). Therefore, the abundance and functional advantages of iNKT cells facilitate them as good candidates for antitumor immunotherapies (8).

Many preclinical models have been developed to investigate the strategies for α-GalCer and iNKT cell–based cancer immunotherapies (9). Over 30 clinical trials with α-GalCer or α-GalCer-pulsed dendritic cells (DC) have also been performed, and some show an increased level of iNKT cells and enhanced IFNγ production via targeting iNKT cells (10). Nonetheless, there are few reports of obvious tumor regression after α-GalCer and iNKT cell–based immunotherapy. Some possible mechanisms have been proposed for the suboptimal efficiency of α-GalCer in antitumor clinical trials, such as induction of iNKT cell anergy and immune suppression by tumor microenvironment. Many novel iNKT cell–based immunotherapies, including adoptive transfer of autologous iNKT, engineered iNKT-TCR–modified hematopoietic stem cell, and chimeric antigen receptor (CAR)-modified iNKT cells, are explored to circumvent the dysfunctional ones and achieve application promise (11–13). Notably, the function of iNKT cells can be readily modulated by microenvironment factors. In the context of HBV-related HCC, chronic HBV infection and carcinogenesis progress exert dual effects on the local immune microenvironment, making the development of effective iNKT cell–based immunotherapies more challenging (14). Deciphering profiles of residual iNKT cells and factors related to iNKT cell dysfunction would be helpful to design the optimal iNKT cell–based immunotherapy in HBV-related HCC.

iNKT cells are lipid-reactive T cells, and lipid components are involved in their structural function in membrane composition and cell signaling functions (15, 16). When the tumor develops, the liver always undergoes metabolic dysfunction, which shapes a special tumor microenvironment with mutative lipids. Recent research has reported a link between abnormal lipids composition in the tumor microenvironment and immune cell dysfunction (17, 18). Although iNKT cells have semi-invariant T-cell receptor (TCR) usages, their complementarity-determining region3β (CDR3β) loop is diverse, which determines differential lipid-antigen recognition and functionally heterogeneous clones (19). Exhaustion and senescence are two major dysfunctional states of T cells that coexist in chronic infections and cancers (20), which have been known to hinder effective antitumor immunity and sustain the suppressive tumor microenvironment (21). In this study, we found that patients with HBV-related HCC suffered severe iNKT cell defects in ratio and α-GalCer responsiveness. Transcriptome-wide analysis indicated that α-GalCer/IL2-rescued iNKT cells from the patients displayed reduced TCR diversity and senescent phenotype associated with aberrant calcium transportation and disturbed metabolic profiling. Widely targeted lipidomic profiling revealed the long-chain acylcarnitines (LCAC) accumulated in the tumor tissue, which contributed to iNKT cell premature senescence.

Patients and populations

Blood samples of 55 patients with chronic hepatitis B (CHB), 11 patients with HBV-negative HCC, and 32 patients with HBV-positive HCC were collected from Tongji Hospital, Wuhan, P.R. China. Blood samples from 41 healthy donors were served as controls. Nontumor liver and tumor specimens were collected from 13 patients with HBV-negative hepatic carcinoid, 8 patients with HBV-negative HCC, and 23 patients with HBV-positive HCC during hepatic resection in Tongji Hospital. Individuals with other concurrent types of viral hepatitis, human immunodeficiency virus infection, autoimmune liver disease, or alcoholic liver disease were excluded. Patient characteristics including age, gender, and correlated clinical indexes are listed in Table 1.

Table 1.

Clinical characteristics of patients enrolled in this study.

Liver and tumor samplesBlood samples
HDHBV+ HCCHBV− HCCHBV− Hepatic carcinoidHBV+ HCCHBV− HCCChronic Hepatitis B
No 41 23 13 32 11 55 
Gender (M/F) 32/9 19/4 7/1 5/8 26/6 9/2 47/8 
Age (y)a 41(18–69) 48 (25–69) 62 (33–75) 50 (21–66) 51.5 (22–69) 53 (41–75) 39 (18–63) 
HBV-DNAa  4.95 ND ND <2.7 ND <2.7 
(log10 copies/mL) <2 (<2–7.25), 3 ND ND ND (<2.7–6.04), 9 ND ND (<2.7–8.69), 2 ND 
HBeAg (+/−) ND 6/16, 1 ND 0/7, 1 ND 0/13 4/28, 4 ND 0/7, 4 ND 16/37, 2 ND 
HBeAb (+/−) ND 15/7, 1 ND 2/5, 1 ND 3/10 22/10, 4 ND 4/3, 4 ND 23/30, 2 ND 
ALTa ND 31 (10–145) 15 (8–51), 1 ND 15 (7–63) 31.5 (10–500) 31.5 (14–441) 40 (11–602) 
ASTa ND 38 (14–110) 26 (17–53), 1 ND 20 (14–38) 44 (16–1,285) 40 (18–468) 37 (20–256) 
Liver and tumor samplesBlood samples
HDHBV+ HCCHBV− HCCHBV− Hepatic carcinoidHBV+ HCCHBV− HCCChronic Hepatitis B
No 41 23 13 32 11 55 
Gender (M/F) 32/9 19/4 7/1 5/8 26/6 9/2 47/8 
Age (y)a 41(18–69) 48 (25–69) 62 (33–75) 50 (21–66) 51.5 (22–69) 53 (41–75) 39 (18–63) 
HBV-DNAa  4.95 ND ND <2.7 ND <2.7 
(log10 copies/mL) <2 (<2–7.25), 3 ND ND ND (<2.7–6.04), 9 ND ND (<2.7–8.69), 2 ND 
HBeAg (+/−) ND 6/16, 1 ND 0/7, 1 ND 0/13 4/28, 4 ND 0/7, 4 ND 16/37, 2 ND 
HBeAb (+/−) ND 15/7, 1 ND 2/5, 1 ND 3/10 22/10, 4 ND 4/3, 4 ND 23/30, 2 ND 
ALTa ND 31 (10–145) 15 (8–51), 1 ND 15 (7–63) 31.5 (10–500) 31.5 (14–441) 40 (11–602) 
ASTa ND 38 (14–110) 26 (17–53), 1 ND 20 (14–38) 44 (16–1,285) 40 (18–468) 37 (20–256) 

Abreviation: ND, not determined.

aData are shown as median (range).

Reagents and antibodies

α-Galcer (KRN7000, Avanti) was dissolved in DMSO at 1 mg/mL and diluted in RPMI1640 complete medium by sonicating at 37°C for 30–60 minutes. Lauroyl-carnitine (LC; 39953, Sigma), myristoyl-carnitine (MC; 61367, Sigma), palmityl-carnitine (PC; 61251, Sigma), and stearoyl-carnitine (SC; 61229, Sigma) were dissolved in DMSO at 100 mmol/L and diluted in RPMI1640 complete medium at 60°C for 2 hours. The FITC-labeled PC (FITC-PC) was prepared by conjugating 5-FITC-C2-amine (75453-82-6, CONFLUORE) with PC and dissolved in DMSO and diluted in RPMI1640 complete medium.

Cytometric Beads Array Flex Sets (BD Biosciences) were used for cytokine detection. CellTrace Violet (Thermo Fisher Scientific), and FITC-AnnexinV Apoptosis Detection Kit (BD Biosciences) were used in apoptosis detection. APC or BV421 labeled PBS57/hCD1d tetramer was provided by the NIH facility. Fluorescence-conjugated mAbs were purchased from BioLegend: CD3 (UCHT1), CD69 (FN50), CD95 (DX2), CD95 L (NOK-1), and CD1d (51.1). Cells were stained with corresponding antibodies and/or detection reagents according to the product introduction. Data were collected using FACSVerse (BD Biosciences) and analyzed by FlowJo software (Tree Star).

Cell preparation and functional assay

Intrahepatic mononuclear cells (MNC) were isolated by modified enzymatic dispersal protocol and Percoll density purification (6). Peripheral blood mononuclear cells (PBMC) were isolated by Ficoll density gradient centrifugation. iNKT cells were expanded from PBMCs with α-GalCer (200 ng/mL) and IL2 (50 U/mL) for 7 days, positively enriched by APC-labeled PBS57/hCD1d tetramer and magnetic-activated cell sorting with anti-APC microbeads (Miltenyi Biotec) according to the product instruction. Sulfate Latex microbeads 8% (wt/vol), 3.5 μm (Thermo Fisher Scientific) were coated with purified anti-human CD28 (BD Pharmingen) and APC-labeled PBS57/hCD1d tetramer to construct microbeads coated with PBS57/hCD1d tetramer.

PBMCs or hepatic MNCs were stimulated with α-GalCer (200 ng/mL) or microbeads coated with PBS57/hCD1d tetramer in the presence of recombinant IL2 (50 U/mL), IL15 (2.5 ng/mL, R&D Systems), PC, LC, MC and/or SC with or without mouse anti-human CD1d (5 μg/mL; BioLegend). Supernatants were collected on third day for cytokine detection and cells were harvested on the seventh day for detection of iNKT cell ratio. HepG2-tmCD1d is a CD1d transfectant of HepG2 (ATCC, HB_8065; ref. 22), which was stained with Celltrace Violet and loaded with α-GalCer (200 ng/mL) for 6 hours as stimulator or target cell. The enriched iNKT cells and target cells were cocultured at a ratio of 2:1 for 24 hours. The supernatants were collected for cytokine detection, and the killing activity was detected by propidium iodide staining of the target cells.

Senescence-associated β-gal activity and real-time PCR for cell senescence detection

α-GalCer/IL2-expanded iNKT cells were treated with vehicle, LC, MC, PC, or SC for 3 days. The senescence-associated β-gal (SA-β-Gal) activity was detected by Cellular Senescence Detection Kit- SPiDER-βGal (SG03, DojinDO), according to the manufacturer's introduction. Total RNA was extracted from enriched iNKT cells and reversed transcribed to synthesize cDNA using the Evo M-MLV Mix Kit (AG11728, Hunan Accurate Biotechnology Co., Ltd.), followed by a real-time PCR reaction on a CFX96 Real-Time PCR Detection System (Bio-Rad) using SYBR Green Premix Pro Taq HS qPCR Kit (AG11701, Hunan Accurate Biotechnology Co., Ltd.) and specific primers (p16INK4a also known as CDKN2A, forward: TTCGCTAAGTGCTCGGAGTTAATAG, reverse: ACCCTGTCCCTCAAATCCTCTG; p53 also known as TP53, forward: CAGCACATGACGGAGGTTGT, reverse: TCATCCAAATACTCCACACGC). The relative expression levels of target genes were normalized by the expression level of the housekeeping gene GAPDH.

Confocal microscopy and image processing

α-GalCer/IL2-expanded iNKT cells were purified and treated with FITC-PC (50 μmol/L) for 16 hours. The cellular membrane and mitochondria were stained with AF700-human-antiCD45 (HI30, BioLegend) and mitochondrial deep red (M22426, Thermo Fisher Scientific), respectively, according to the product introduction. The confocal images were obtained using Leica TCS SP8 MP confocal microscope, and processed in LAS X software.

RNA sequencing and analysis

RNA from nontumor liver, tumor specimens or enriched iNKT cells was used for stranded RNA sequencing (RNA-seq) library preparation using KC-Digital Stranded mRNA Library Prep Kit for Illumina (catalog no. DR08502, Seqhealth Technology Co., Ltd.). The kit eliminates duplication bias in PCR and sequencing steps, by using a unique molecular identifier of eight random bases to label the preamplified cDNA molecules. Qualified reads were mapped to the reference genome using STAR software (version 2.5.3a) and counted by featureCounts (Subread-1.5.1; Bioconductor). Differential expressed genes (DEG) were analyzed by the DEseq package on R (version 4.1.2). The genes were selected for further gene ontology (GO)/Kyoto Encyclopedia of Genes and Genomes (KEGG)/gene set enrichment analyses (GSEA) enrichment analysis and visualization with normailized enrichment score (NES). TCR repertoire of samples was analyzed by MiXCR software (available at http://mixcr.milaboratory.com/ and https://github.com/milaboratory/mixcr/).

Widely targeted lipidomic profiling

Nontumor liver and tumor specimens were placed in liquid nitrogen, then thawed on ice. A total of 20 mg sample was homogenized with a mixture (1 mL) of methyl tertiary butyl ether, methanol, and internal standard mixture, resuspended with water (500 μL) and centrifuged at 12,000 revolutions per minute at 4°C for 10 minutes. The supernatant was dried with nitrogen. The powder was dissolved with mobile phase B (100 μL) for LC/MS-MS analysis through an LC-ESI-MS/MS system (UPLC, Shim-pack UFLC SHIMADZU CBM A system; MS, QTRAP System) by Wuhan Metware Biotechnology Co., Ltd. Peak areas of lipids were obtained using the Analyst software (version 1.6.3). Relative levels of lipid metabolites were selected for orthogonal partial least squares discriminant analysis through R Studio (version 14.1.0, RStudio Inc.). Significantly changed lipids were selected to produce heatmap in R according to variable importance in projection >0.9 criteria.

Statistical analysis

Graphs were generated and analyzed by GraphPad Prism 7.0 (GraphPad Software). Data were analyzed by one-way ANOVA or Student t test for comparisons of groups with normal distribution and equal variance. For biased distributed variables, Mann–Whitney U test was performed. A P value lower than 0.05 was considered statistically significant.

Ethics

Written informed consents were obtained from all participating patients. The studies were conducted in accordance with recognized ethical guideline (Declaration of Helsinki), and were approved by the ethics committee of Tongji Medicine College, Huazhong University of Science and Technology.

Data availability statement

The transcriptome data generated in this study are publicly available in Gene Expression Omnibus at GSE208036 and GSE208038. Other data generated in this study are available within the article and its Supplementary Data.

iNKT cells from patients with HBV-related HCC display severe defects in ratio and α-GalCer responsiveness

A previous study has shown the deficiency of iNKT cells in HBV-related patients, which is associated with the degree of liver injury (6). Here, we found the patients with HBV-related HCC had even lower circulating iNKT cell ratios as compared with patients with CHB (Fig. 1A). This difference was not likely due to the enhanced liver injury, because the enrolled patients in CHB and HCC groups had similar levels of serum alanine transaminase (ALT), aspartate transaminase (AST), and total bilirubin (Supplementary Fig. S1A). Furthermore, a reduced ratio of hepatic iNKT cells was revealed in the nontumor liver and tumor tissue from patients with HBV-related HCC (Fig. 1B). In accordance with the reported activation-induced cell death for iNKT cell reduction (6), CD69, CD95 (Fas), and CD95 L (FasL) level increased in iNKT cells from patients with HBV-related HCC (Fig. 1C). In addition, patients with HBV-related HCC tended to have fewer iNKT cells with higher CD69 level as compared with patients with HBV-negative HCC, albeit HBV-negative ones had lower iNKT cell frequency than healthy donors (Supplementary Fig. S1B and S1C). These results suggested the severe reduction in iNKT cells was related to overactivation of the cells in patients with HBV-related HCC.

Figure 1.

iNKT cells from patients with HBV-related HCC show severe defects in ratio and α-GalCer responsiveness. A, Representative dot plots and summary scatter graph depict the ratio of circulating iNKT cells in CD3+ T cells from HDs (n = 41), patients with CHB (n = 55), and patients with HBV-related HCC (n = 32). B, Representative dot plots and summary scatter graph depict the ratio of hepatic iNKT cells in CD3+ T cells from the nontumor liver and tumor tissue from HBV-related patients (n = 23). Liver tissue from HBV-negative individuals (HBV liver, n = 13) served as control. Data are shown as median with 95% confidence interval with statistical significance determined by Mann–Whitney U test. C, Bar graphs with scatter plots depict CD69, CD95, and CD95 L levels on iNKT cells from indicated groups. D, Representative dot plots and bar graphs with scatter plots show the fold changes in ratio and number of iNKT cells in PBMCs from indicated groups stimulated with α-GalCer (αGC) for 7 days in the presence and/or absence of anti-CD1d antibody (HD, n = 36; CHB, n = 45; HCC, n = 7). E, The bar graph with scatter plots shows IFNγ concentration from 3 days’ αGC-stimulated PBMCs in indicated groups (HD, n = 16; CHB, n = 30; HCC, n = 6). Data are shown as mean ± SEM and the statistical significance among groups was determined by one-way ANOVA with Fisher LSD post hoc test. F, Scatter graphs show IFNγ concentration in αGC-stimulated intrahepatic MNCs from indicated tissue (n = 8). The cells stimulated with vehicle (Veh) were served as control. Data are shown as mean ± SEM with statistical significance determined by an unpaired two-tailed Student t test. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, nonsignificant.

Figure 1.

iNKT cells from patients with HBV-related HCC show severe defects in ratio and α-GalCer responsiveness. A, Representative dot plots and summary scatter graph depict the ratio of circulating iNKT cells in CD3+ T cells from HDs (n = 41), patients with CHB (n = 55), and patients with HBV-related HCC (n = 32). B, Representative dot plots and summary scatter graph depict the ratio of hepatic iNKT cells in CD3+ T cells from the nontumor liver and tumor tissue from HBV-related patients (n = 23). Liver tissue from HBV-negative individuals (HBV liver, n = 13) served as control. Data are shown as median with 95% confidence interval with statistical significance determined by Mann–Whitney U test. C, Bar graphs with scatter plots depict CD69, CD95, and CD95 L levels on iNKT cells from indicated groups. D, Representative dot plots and bar graphs with scatter plots show the fold changes in ratio and number of iNKT cells in PBMCs from indicated groups stimulated with α-GalCer (αGC) for 7 days in the presence and/or absence of anti-CD1d antibody (HD, n = 36; CHB, n = 45; HCC, n = 7). E, The bar graph with scatter plots shows IFNγ concentration from 3 days’ αGC-stimulated PBMCs in indicated groups (HD, n = 16; CHB, n = 30; HCC, n = 6). Data are shown as mean ± SEM and the statistical significance among groups was determined by one-way ANOVA with Fisher LSD post hoc test. F, Scatter graphs show IFNγ concentration in αGC-stimulated intrahepatic MNCs from indicated tissue (n = 8). The cells stimulated with vehicle (Veh) were served as control. Data are shown as mean ± SEM with statistical significance determined by an unpaired two-tailed Student t test. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, nonsignificant.

Close modal

Despite low frequency in PBMCs, iNKT cells from healthy donors were expanded efficiently via α-GalCer stimulation and enriched up to over 95% purity by PBS57/hCD1d tetramer-positive sorting (Supplementary Fig. S1D). In response to α-GalCer stimulation, PBMCs from patients with CHB and HBV-related HCC showed reduced iNKT cell expansion compared with healthy donors, and the lowest expansion rate of iNKT cells was observed in PBMCs from patients with HCC (Fig. 1D; Supplementary Fig. S1E). The enriched human iNKT cells from healthy donors exhibited Th1-biased cytokines production and displayed significant cytotoxicity against hepatoma cells in a CD1d-dependent manner, which was enhanced by replacing the presented endogenous lipids with α-GalCer (Supplementary Fig. S2). However, α-GalCer–induced CD1d-dependent IFNγ production was reduced in PBMCs from patients with CHB and lost in PBMCs and intrahepatic MNCs from patients with HBV-related HCC (Fig. 1E and F). Notably, in patients with CHB and HCC, there was no decrease in proportions and CD1d levels of monocytes and B cells, two major antigen-presenting cell population in PBMCs (Supplementary Fig. S3). This suggested in vitro hyporesponsiveness of iNKT cells was less likely attributed to impaired α-GalCer presentation. Although elevated CD1d expression in nontumor liver and tumor tissue from patients with HBV-related HCC has been reported and expected to facilitate antitumor effects of iNKT cells (6), the severe defects in iNKT cell ratio and responsiveness suggested dampened iNKT cell–mediated antitumor effects in the patients, which could be hardly restarted by the administration of α-GalCer.

IL2-rescued α-GalCer–responsive iNKT cells from patients with HBV-related HCC exhibit reduced TCR CDR3β diversity as well as exhausted and senescent phenotype

We have previously shown that IL2 and IL15 significantly increase α-GalCer–induced expansion and IFNγ production of iNKT cells from chronic HBV–infected patients (6). Here, exogenous IL2 but not IL15 increased α-GalCer–induced expansion of iNKT cells from the patients with HBV-related HCC (Fig. 2A and B). However, the IL2 effect was limited because of much fewer expanded iNKT cells from the patients as compared with those from healthy donors (Fig. 2C). This suggested reduced responsive iNKT cell clones in the patients with HBV-related HCC. IL2 also rescued the CD1d-dependent IFNγ production of iNKT cells from the patients (Fig. 2D), which could be due to the increased cell number and/or enhanced IFNγ producing capacity of residual iNKT cells by IL2 stimulation. These results indicated that IL2 helped to partially rescue the α-GalCer responsiveness of residual iNKT clones in the patients with HBV-related HCC.

Figure 2.

IL2 partially rescues α-GalCer–responsive iNKT cells from patients with HBV-related HCC with reduced TCR CDR3β diversity and changed TCR repertoire. AD, PBMCs from the HDs and patients with HBV-related HCC were stimulated with α-GalCer in the presence of IL2, IL15, and/or anti-CD1d antibody (n = 7/group). A, Representative iNKT cell ratio in indicated groups. B, Fold changes in the ratio (top) and number (bottom) of iNKT cells in indicated groups. C, The absolute ratio and cell number of iNKT cells in indicated groups. D, IFNγ concentration in indicated groups. Data are shown as mean ± SEM with statistical significance determined by paired two-tailed Student t test. EH, RNA-seq data of α-GalCer/IL2-expanded iNKT cells from three HD (HD iNKT), three patients with CHB (CHB iNKT), and three patients with HBV-related HCC (HCC iNKT) were analyzed for TCR usages by MiXCR software. E, Bar graph with scatter plots shows the distribution of CDR3β aa length in indicated groups (n = 3/group). F, Seqlogo plots for amino acid composition at each position of 15aa CDR3β sequences in indicated groups (n = 3/group), with red circles highlighting distinct amino acids in indicated populations. G, Table shows the numbers of total and distinct CDR3β sequence clones of indicated iNKT cells. Bar graph with scatter plots depicts the ratio of distinct TCR CDR3β sequence, which was calculated through dividing numbers of total detected CDR3β sequence clones by numbers of distinct TCR CDR3β sequence clones. Data are shown as mean ± SEM and the statistical significance was determined by one-way ANOVA with Fisher LSD post hoc test. H, Chord diagrams show the TRBV25-1 and TRBJ pairs in HD iNKT, CHB iNKT and HCC iNKT (n = 3/group). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 2.

IL2 partially rescues α-GalCer–responsive iNKT cells from patients with HBV-related HCC with reduced TCR CDR3β diversity and changed TCR repertoire. AD, PBMCs from the HDs and patients with HBV-related HCC were stimulated with α-GalCer in the presence of IL2, IL15, and/or anti-CD1d antibody (n = 7/group). A, Representative iNKT cell ratio in indicated groups. B, Fold changes in the ratio (top) and number (bottom) of iNKT cells in indicated groups. C, The absolute ratio and cell number of iNKT cells in indicated groups. D, IFNγ concentration in indicated groups. Data are shown as mean ± SEM with statistical significance determined by paired two-tailed Student t test. EH, RNA-seq data of α-GalCer/IL2-expanded iNKT cells from three HD (HD iNKT), three patients with CHB (CHB iNKT), and three patients with HBV-related HCC (HCC iNKT) were analyzed for TCR usages by MiXCR software. E, Bar graph with scatter plots shows the distribution of CDR3β aa length in indicated groups (n = 3/group). F, Seqlogo plots for amino acid composition at each position of 15aa CDR3β sequences in indicated groups (n = 3/group), with red circles highlighting distinct amino acids in indicated populations. G, Table shows the numbers of total and distinct CDR3β sequence clones of indicated iNKT cells. Bar graph with scatter plots depicts the ratio of distinct TCR CDR3β sequence, which was calculated through dividing numbers of total detected CDR3β sequence clones by numbers of distinct TCR CDR3β sequence clones. Data are shown as mean ± SEM and the statistical significance was determined by one-way ANOVA with Fisher LSD post hoc test. H, Chord diagrams show the TRBV25-1 and TRBJ pairs in HD iNKT, CHB iNKT and HCC iNKT (n = 3/group). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Close modal

To profile the residual α-GalCer–responsive iNKT cells, we enriched α-GalCer/IL2-expanded iNKT cells from patients with HBV-related HCC (HCC iNKT) for RNA-seq as compared with those from healthy donors (HD iNKT) or patients with CHB (CHB iNKT; Supplementary Fig. S1E). By taking advantage of MiXCR software, V, D, and J segments, as well as CDR3β of TCR were identified from RNA-seq data. Although most iNKT cells from different individuals typically expressed an invariant Vα24Jα18 segment (TRAV10-TRAJ18) paired with the Vβ11 chain (TRBV25-1) as expected (Supplementary Fig. S4A and B), their CDR3β loops were diverse with frequent CDR3β length from 14 to 17 amino acid (aa; Fig. 2EG). Enriched amino acids in CDR3β middle region were found to be different in HD iNKT, CHB iNKT, and HCC iNKT (Fig. 2F; Supplementary Fig. S4C). In particular, Pro118β and Pro119β in 15aa length CD3β loop were found to be more prevalent in CHB iNKT and HCC iNKT than HD iNKT, while Ser119 was specially enriched in HCC iNKT (Fig. 2F). Notably, the diversity of CDR3β was significantly reduced in CHB iNKT and HCC iNKT, reflected by the reduced ratio of distinct CDR3β sequence clones in total detected CDR3β sequence clones (Fig. 2G). As compared with HD iNKT, Jβ2-7 was overrepresented while Jβ1-5 decreased among CHB iNKT and HCC iNKT (Fig. 2H). Notably, Jβ2-1 usage decreased significantly in CHB iNKT but slightly increased in HCC iNKT. These results showed the common and distinguished changes in TCR repertoire between CHB iNKT and HCC iNKT, indicating both HBV infection and cancer progression contributed to biased TCR usage of HCC iNKT.

Further analyzing the transcriptome data, we found overall reduced cytotoxicity mediators but increased activation/exhaustion/inhibition markers in HCC iNKT (Fig. 3A). The levels of CD4 and chemokine receptor 7 guiding tissue exit of the cells were also found to be elevated in HCC iNKT (Fig. 3A). This suggested the increased CD4+ iNKT subset and enhanced recirculation of residual iNKT cells in the patients. Notably, HCC iNKT showed predominantly upregulated transcripts in “Cytokine-cytokine receptor interaction” pathway (Supplementary Fig. S5), among which, upregulation of IL6, chemokine (C-X-C motif) ligand 8 (CXCL8), IL10, IL18, IL1β, IL1α, CCL2, and CXCL1/2/3/5 are typical senescence-associated secretory phenotype (SASP; Fig. 3B; refs. 23, 24). Downregulated perforin and granzymes were observed in HCC iNKT, further supporting the SASP profile of the population (Fig. 3B). Moreover, HCC iNKT cells had predominantly increased genes in the “Transcriptional misregulation in cancer” pathway (Fig. 3C; Supplementary Fig. S6). After analyzing the functional protein–protein interaction network, we found an IL6/CXCL8/IL1A/IL1B/IL10/MMP9 centered interaction network between “Cytokine-cytokine receptor interaction” and “Transcriptional misregulation in cancer” pathways (Fig. 3D), suggesting tumor-associated secretome of HCC iNKT. Together, our results showed SASP and dampened cytotoxicity of HCC iNKT cells, which was supposed to increase their protumorigenic potential and weaken their immune surveillance capacity.

Figure 3.

Transcriptome-wide analysis reveals the dysregulated HCC iNKT with tumor-associated secretome and enriched senescence-associated pathways. Transcripts from RNA-seq data of α-GalCer/IL2-expanded HD iNKT (HD1–3, n = 3), CHB iNKT (CHB1–3, n = 3), and HCC iNKT (HCC1–4, n = 4) were analyzed in phenotype clusters or KEGG pathways. A, The radar map shows the relative transcript levels of indicated genes in CHB iNKT versus HD iNKT and in HCC iNKT versus HD iNKT. B, Heatmap shows relative mRNA levels of indicated genes for the SASP in different populations. C, The log2-fold changes of genes enriched in the “Transcriptional misregulation in cancer” pathway from CHB iNKT versus HD iNKT (left) and HCC iNKT versus CHB iNKT (right). D, Protein–protein interaction network shows the interaction of upregulated genes in the “Cytokine-cytokine receptor interaction” and “Transcriptional misregulation in cancer” pathways, respectively. E, The bubble chart shows enriched pathways of HCC iNKT versus HD iNKT, with top 10 significantly changed pathways labeled according to the values of −log10 (P).

Figure 3.

Transcriptome-wide analysis reveals the dysregulated HCC iNKT with tumor-associated secretome and enriched senescence-associated pathways. Transcripts from RNA-seq data of α-GalCer/IL2-expanded HD iNKT (HD1–3, n = 3), CHB iNKT (CHB1–3, n = 3), and HCC iNKT (HCC1–4, n = 4) were analyzed in phenotype clusters or KEGG pathways. A, The radar map shows the relative transcript levels of indicated genes in CHB iNKT versus HD iNKT and in HCC iNKT versus HD iNKT. B, Heatmap shows relative mRNA levels of indicated genes for the SASP in different populations. C, The log2-fold changes of genes enriched in the “Transcriptional misregulation in cancer” pathway from CHB iNKT versus HD iNKT (left) and HCC iNKT versus CHB iNKT (right). D, Protein–protein interaction network shows the interaction of upregulated genes in the “Cytokine-cytokine receptor interaction” and “Transcriptional misregulation in cancer” pathways, respectively. E, The bubble chart shows enriched pathways of HCC iNKT versus HD iNKT, with top 10 significantly changed pathways labeled according to the values of −log10 (P).

Close modal

iNKT cells from patients with HBV-related HCC display cell-cycle arrest and impaired telomere maintenance

The DEGs of HCC iNKT versus HD iNKT were further analyzed for pathway enrichment. KEGG pathway enrichment analysis identified several aging-related pathways, including “Cell senescence” and “Cell cycle” pathways in HCC iNKT (Fig. 3E). Furthermore, GSEA based on GO database (GSEA-GO) revealed the SASP-related IL6, IL1, CXCL8, IL10, and IL18 production processes were enriched in HCC iNKT cells with positive normalized enriched scores (NES > 1, q value < 0.05). Upregulated aging-related biological processes (NES > 1, q value < 0.05) and downregulated cell cycle–related biological processes (NES ≤ 1, q value < 0.05) were also found in HCC iNKT (Fig. 4A). Moreover, downregulated telomere maintenance-related biological processes indicated impaired telomere maintenance in HCC iNKT (Fig. 4A), showing another evidence for their senescent status. GSEA based on KEGG database (GSEA-KEGG) showed downregulated transcriptional signatures in “DNA replication” and “Cell cycle” pathways in HCC iNKT cells (Fig. 4B). Genes encoding cyclin (CCNA2/CCNB1/CCNB2/CCND3/CCNE2/CCNH) and cyclin-dependent kinases (CDK1/2/4/7) that control the progression of the cell cycle were downregulated, while cyclin-dependent kinase inhibitors (CDKN1A/CDKN2B) were upregulated (Fig. 4C; refs. 25–27). In addition, genes encoding cell cycle–related phosphatases CDC25A and B were downregulated in HCC iNKT cells (Fig. 4C; ref. 28). These changes were proposed to promote cell-cycle arrest at senescence. GLB1, the gene encoding lysosomal beta-D-galactosidase and responsible for the activity of SA-β-Gal (29), was upregulated in HCC iNKT cells (Fig. 4D). Consistently, we found higher SA-β-Gal levels in iNKT cells from the tumor tissue as compared with those from nontumor liver (Fig. 4E), suggesting elevated senescent profile of tumor-infiltrated iNKT cells.

Figure 4.

iNKT cells from patients with HBV-related HCC display an increased senescent phenotype compared with those from the healthy donors. Transcripts from HCC iNKT (HCC1–4, n = 4) were compared with HD iNKT (HD1–3, n = 3) and enriched for GO, KEGG, and GSEA. A, The histogram graph shows NES of senescence-related biological processes from GSEA-GO analysis in HCC iNKT versus HD iNKT. B, GSEA plots for transcriptional signatures of “DNA replication” and “Cell cycle” KEGG pathways in HCC iNKT versus HD iNKT. C, Heatmap shows the relative mRNA levels of cell-cycle pathway–related genes in indicated cells. D, Bar graph with scatter plots shows related mRNA levels of GLB1 in the indicated groups. E, Histogram and summary scatter graph show SA-β-Gal level of iNKT cells from nontumor tissue (NT) and tumor tissue (T) of patients with HBV-related HCC (n = 5/group). *, P < 0.05, determined by a paired two-tailed Student t test. F, GSEA plot for the transcriptional signatures of “Glycolysis” KEGG pathway in HCC iNKT versus HD iNKT. G, Heatmap shows the relative mRNA levels of glycolysis pathway–related genes in indicated cells. H, GSEA plot for transcriptional signatures of “Oxidation phosphorylation” KEGG pathway in HCC iNKT versus HD iNKT. I, The NES of telomere maintenance–, calcium transport–, and mitochondrial function–related biological processes from GSEA-GO analysis in HCC iNKT versus HD iNKT. J, GSEA plots for transcriptional signatures of “Calcium signaling pathway” and “Endocrine and other factor regulated calcium reabsorption” KEGG pathways in HCC iNKT versus HD iNKT. K, mRNA levels of ITPR1, ITPR2, and MCU in the indicated groups. Data are shown as mean ± SEM. *, P < 0.05, according to the result of DESeq on R.

Figure 4.

iNKT cells from patients with HBV-related HCC display an increased senescent phenotype compared with those from the healthy donors. Transcripts from HCC iNKT (HCC1–4, n = 4) were compared with HD iNKT (HD1–3, n = 3) and enriched for GO, KEGG, and GSEA. A, The histogram graph shows NES of senescence-related biological processes from GSEA-GO analysis in HCC iNKT versus HD iNKT. B, GSEA plots for transcriptional signatures of “DNA replication” and “Cell cycle” KEGG pathways in HCC iNKT versus HD iNKT. C, Heatmap shows the relative mRNA levels of cell-cycle pathway–related genes in indicated cells. D, Bar graph with scatter plots shows related mRNA levels of GLB1 in the indicated groups. E, Histogram and summary scatter graph show SA-β-Gal level of iNKT cells from nontumor tissue (NT) and tumor tissue (T) of patients with HBV-related HCC (n = 5/group). *, P < 0.05, determined by a paired two-tailed Student t test. F, GSEA plot for the transcriptional signatures of “Glycolysis” KEGG pathway in HCC iNKT versus HD iNKT. G, Heatmap shows the relative mRNA levels of glycolysis pathway–related genes in indicated cells. H, GSEA plot for transcriptional signatures of “Oxidation phosphorylation” KEGG pathway in HCC iNKT versus HD iNKT. I, The NES of telomere maintenance–, calcium transport–, and mitochondrial function–related biological processes from GSEA-GO analysis in HCC iNKT versus HD iNKT. J, GSEA plots for transcriptional signatures of “Calcium signaling pathway” and “Endocrine and other factor regulated calcium reabsorption” KEGG pathways in HCC iNKT versus HD iNKT. K, mRNA levels of ITPR1, ITPR2, and MCU in the indicated groups. Data are shown as mean ± SEM. *, P < 0.05, according to the result of DESeq on R.

Close modal

After comparing the DEGs of HCC iNKT with CHB iNKT, we found HCC iNKT also had downregulated cell cycle–related biological processes, upregulated SASP-related processes, and downregulated telomere maintenance (Supplementary Fig. S7A–S7C), suggesting the promotion of iNKT cell-cycle arrest and senescence by tumor factors. Together, our results showed that iNKT cells from patients with HBV-related HCC had elevated senescence profiles.

iNKT cells from patients with HBV-related HCC exhibit aberrant calcium transport-related biological processes and disturbed metabolic profiling

Because T-cell aging closely correlates to mitochondrial dysfunction and calcium signaling (30–32), we further focused on cell metabolism– and calcium transportation–related pathways. As compared with HD iNKT and CHB iNKT, HCC iNKT exhibited an upward trend in the “Glycolysis” pathway and a downward trend in the “Oxidation phosphorylation” pathway (Fig. 4FH; Supplementary Fig. S7C). We also found that downregulated telomere maintenance-related biological processes (NES ≤ 1, q value < 0.05) coincided with upregulated calcium transport-related biological processes (NES > 1, q value < 0.05) in HCC iNKT (Fig. 4I; Supplementary Fig. S7D). GSEA-KEGG further revealed upregulation of transcriptional signatures in the “Calcium signaling pathway” (Fig. 4J; Supplementary Fig. S7E). Endoplasmic recticulum (ER) stress is characterized by increased calcium release to the other organelle, such as mitochondria and lysosomes (33). Here, we found upregulation in transcriptional levels of 1,4,5-trisphosphate receptors (ITPR1 and ITPR2) and mitochondrial calcium uniporter (MCU; Fig. 4K; Supplementary Fig. S7F). This was supposed to promote ER calcium release and mitochondrial calcium accumulation followed by increased reactive oxygen species, thus inducing premature senescence. Consistently, transcriptional upregulation of “reactive oxygen species biosynthetic process” and “superoxide anion generation” as well as reduced mitochondria-related biological processes suggested elevated oxidative stress and mitochondria dysfunction in HCC iNKT cells (Fig. 4I; Supplementary Fig. S7D). In accordance with that iNKT cells are sensitive to mitochondrial perturbation (34), our results suggested they were susceptible to calcium transportation-related mitochondrial dysfunction and cell senescence.

Accumulation LCACs in the tumor tissue from patients with HBV-related HCC coincides with disturbed lipid metabolism of iNKT cells

On the basis of widely targeted lipidomic profiling, we found the top10 upregulated lipid components in tumor tissue from patients with HBV-related HCC as compared with nontumor liver were predominant LCACs, among which, four commercially available LCACs including LC, MC, PC, and SC were accumulated (Fig. 5A and B). LCACs have been reported to evoke cytosolic calcium accumulation and mitochondrial calcium buffering released from ER stores (35), which could drive mitochondrial dysfunction and human cell senescence (36, 37). Consistently, exogenous fluorochrome-labeled PC, a representative LCAC accumulated in HCC, was mainly located in the cytoplasm of the lymphocytes, and a frequent colocalization of PC and the mitochondrial tracker was observed (Fig. 5C; Supplementary Fig. S8). This suggested that accumulated PC was readily to enter the cells and locate to their mitochondria.

Figure 5.

Accumulation of LCACs in tumor tissue coincides with aberrant lipid metabolism in iNKT from patients with HBV-related HCC. Differentially lipid components from wide-targeted lipidomic profiling and transcripts from RNA-seq data of tumor tissue (T1-T7) and nontumor liver (NT1-NT8) were compared, respectively. A, The volcano plot shows the upregulated and downregulated lipids in the tumor tissue with top 10 upregulated lipids labeled. B, Heatmap shows relative levels of indicated lipids in different samples. C, Confocal images show the localization of FITC-PC in the lymphocytes that were stained with CD45 or mitochondrial deep red (MTDR). Arrowheads, colocalization of FITC-PC and MTDR. D, GSEA plot of the transcriptional signature of “Fatty acid degradation” KEGG pathway (left) in HCC tumor tissue versus nontumor liver, and heatmap (right) shows the relative mRNA levels of related genes in indicated samples. E, The histogram shows the log2-fold change value of indicated genes, with grandient color bars showing P values in tumor tissue versus nontumor liver. F, The schematic diagram shows the partial process and key enzymes involved in fatty acid oxidation in mitochondria. G, The bubble chart shows GSEA-GO enriched lipid metabolism–related biological processes of HCC iNKT versus CHB iNKT. H, Bar graphs with scatter plots show related mRNA levels of CD36, ACSL1, and ACSL4 in iNKT cells from indicated groups. Data are shown as mean ± SEM. **, P < 0.01; ***, P < 0.001, according to the result of DESeq.

Figure 5.

Accumulation of LCACs in tumor tissue coincides with aberrant lipid metabolism in iNKT from patients with HBV-related HCC. Differentially lipid components from wide-targeted lipidomic profiling and transcripts from RNA-seq data of tumor tissue (T1-T7) and nontumor liver (NT1-NT8) were compared, respectively. A, The volcano plot shows the upregulated and downregulated lipids in the tumor tissue with top 10 upregulated lipids labeled. B, Heatmap shows relative levels of indicated lipids in different samples. C, Confocal images show the localization of FITC-PC in the lymphocytes that were stained with CD45 or mitochondrial deep red (MTDR). Arrowheads, colocalization of FITC-PC and MTDR. D, GSEA plot of the transcriptional signature of “Fatty acid degradation” KEGG pathway (left) in HCC tumor tissue versus nontumor liver, and heatmap (right) shows the relative mRNA levels of related genes in indicated samples. E, The histogram shows the log2-fold change value of indicated genes, with grandient color bars showing P values in tumor tissue versus nontumor liver. F, The schematic diagram shows the partial process and key enzymes involved in fatty acid oxidation in mitochondria. G, The bubble chart shows GSEA-GO enriched lipid metabolism–related biological processes of HCC iNKT versus CHB iNKT. H, Bar graphs with scatter plots show related mRNA levels of CD36, ACSL1, and ACSL4 in iNKT cells from indicated groups. Data are shown as mean ± SEM. **, P < 0.01; ***, P < 0.001, according to the result of DESeq.

Close modal

To explore the reason for LCACs accumulation in tumor tissue, we analyzed the critical pathway and enzymes involved in LCAC metabolism. GSEA-KEGG revealed predominant downregulated transcripts in the “Fatty acid degradation” pathway in tumor tissue (Fig. 5D). In particular, SLA25A20 and CPT2, genes encoding carnitine-acylcarnitine translocase transferring LCAC into mitochondrial matrix and carnitine acyltransferase II (CPTII) converting LCAC to long-chain fatty acid (LCFA)-CoA, respectively, were downregulated in tumor tissue (Fig. 5D and E), which has been reported to promote the accumulation of LCACs (38, 39). In addition, ACADL and ACADVL, encoding long-chain fatty acyl-CoA dehydrogenase (LCAD) and very long-chain fatty acyl-CoA dehydrogenase (VLCAD), respectively, were also downregulated (Fig. 5D and E), which could result in fatty acid oxidation defects and a reverse reaction to produce LCACs from LCFA-CoA (Fig. 5F). These changes would ultimately lead to LCAC accumulation in the tissue and plasma. ACSL3/4 encoding LCFA-CoA ligase 3/4 (ACSL3/4) that preferentially activates degradation of arachidonic acid [FFA (20:5)] and eicosapentaenoic acid [FFA (20:4)] were upregulated (Fig. 5E). Consistently, reduction in FFA (20:4)/(20:5) were observed in tumor tissue (Fig. 5B). Given that ACSL3/4 also activates LCFA to LCFA-CoA (Fig. 5F), ACSL3/4 upregulation and ACADL/ACADVL downregulation in tumor tissue were supposed to synergistically enhance LCFA-CoA–dependent LCACs accumulation. Collectively, our data showed the disrupted lipid metabolism associated with LCACs accumulation in tumor tissue.

To access whether iNKT from patients with HCC had aberrant lipid metabolism, we further analyzed DEGs between HCC iNKT and CHB iNKT for lipid metabolism-related pathways (Fig. 5G). As shown, HCC iNKT cells upregulated lipid metabolism–related biological processes, such as “lipid localization,” “lipid transport,” “lipid catabolic process,” etc. (Fig. 5G). Transcripts of some crucially involved lipid transporter– or lipid degradation–related enzymes were increased, including CD36 that transports LCFA into cells and ACSL1/4 that activates fatty acid oxidation procession (Fig. 5H). Collectively, our data showed that LCACs accumulation in HCC coincided with disturbed lipid metabolism in HCC iNKT, which might relate to iNKT cell senescence.

Exogenous LCACs impair α-GalCer–induced iNKT cell expansion and contribute to the development of iNKT cell senescence

To investigate the role of LCACs in iNKT cell senescence, we set up an α-GalCer/IL2-induced iNKT cell expansion system in the presence of different LCACs, including LC, MC, PC, and SC that accumulated in HCC (Fig. 6A). Notably, all the four types of LCACs inhibited the α-GalCer–induced iNKT cell expansion from PBMCs in a dose-dependent manner (Fig. 6B), regardless the absence or presence of exogenous IL2 (Supplementary Fig. S9). PC and SC had IC50 lower than 100 μmol/L and 50 μmol/L, respectively (Fig. 6B). PC (100 μmol/L) and SC (50 μmol/L) inhibited iNKT cell proliferation and expansion to a greater degree than LC and MC (100 μmol/L; Fig. 6C and D; Supplementary Fig. S10A and S10B). To investigate whether the toxicity of the LCACs contributed to the decreased iNKT cell expansion, we treated purified α-GalCer/IL2-expanded iNKT cells with LCACs (Fig. 6E; Supplementary Fig. S10C). None of the selected LCACs caused significant apoptosis and reduction of iNKT cell in the given concentration (Fig. 6F and G; Supplementary Fig. S10D and S10E). LCAC-treated groups had similar ratios and CD1d-expressing levels of monocytes and B cells (two major APCs in PBMCs) as compared with controls (Supplementary Fig. S11A–S11C). In addition, in another iNKT cell expansion system stimulated by PBS57/hCD1d tetramer-coated microbeads, LCACs also exhibited significant inhibition (Supplementary Fig. S11D–S11F). These results showed the inhibitory effects of LCACs on iNKT cell expansion were less likely attributed to the impaired α-GalCer presentation by LCACs treatment. Neither exogenous LCACs impaired IFNγ production of iNKT cells (Supplementary Fig. S12A–S12E). Of note, diminished IFNγ production was not a typical SASP of senescent T cells (23, 24, 40). Given the complicated microenvironment changes in HCC, there should be other factors than LCACs affecting IFNγ production by iNKT cells. Together, our results suggested an apoptosis- and IFNγ-independent mechanism for inhibition of iNKT cell expansion by LCACs.

Figure 6.

LCACs impair α-GalCer–induced iNKT cell proliferation and promote their senescent phenotype. AD, PBMCs from healthy donors were stimulated with α-GalCer/IL2 for 7 days in the presence of vehicle (Veh), LC, MC, PC, or SC. A, The flow chart shows the process of the assay. B, The line chart shows the percent of inhibition of iNKT cell proliferation with different LCACs in indicated concentration (n = 2–4/group). C, Representative dot plots and summary bar graph with scatter plots show the percent of iNKT cells in PBMCs treated with vehicle (Veh), PC (100 μmol/L), or SC (50 μmol/L; n = 10/group). D, Representative histograms and summary bar graph with scatter plots show the percent of CFSEhigh iNKT cells in indicated groups (n = 9/group). EI, PBMCs from healthy donors were isolated and stimulated with α-GalCer for 7 days, and then treated with vehicle (Veh), PC (100 μmol/L), or SC (50 μmol/L) for 3 days. E, The flow chart shows the process of the assay. F, Representative dot plots and bar graph with scatter plots show the percent of iNKT cells in PBMCs in indicated groups (n = 6/group). G, The bar graph with scatter plots shows the percent of AnnexinV+ iNKT cells in indicated groups (n = 6/group). H, Representative dot plots and summary bar graph with scatter plots show the SA-β-Gal levels of iNKT cells in indicated groups (n = 8–16/group). I, The summary bar graphs with scatter plots show the relative mRNA expression of p16INK4a and p53 of iNKT cells in indicated groups (n = 3–8/group). Data are shown as mean ± SEM and the statistical significance was determined by paired two-tailed Student t test. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 6.

LCACs impair α-GalCer–induced iNKT cell proliferation and promote their senescent phenotype. AD, PBMCs from healthy donors were stimulated with α-GalCer/IL2 for 7 days in the presence of vehicle (Veh), LC, MC, PC, or SC. A, The flow chart shows the process of the assay. B, The line chart shows the percent of inhibition of iNKT cell proliferation with different LCACs in indicated concentration (n = 2–4/group). C, Representative dot plots and summary bar graph with scatter plots show the percent of iNKT cells in PBMCs treated with vehicle (Veh), PC (100 μmol/L), or SC (50 μmol/L; n = 10/group). D, Representative histograms and summary bar graph with scatter plots show the percent of CFSEhigh iNKT cells in indicated groups (n = 9/group). EI, PBMCs from healthy donors were isolated and stimulated with α-GalCer for 7 days, and then treated with vehicle (Veh), PC (100 μmol/L), or SC (50 μmol/L) for 3 days. E, The flow chart shows the process of the assay. F, Representative dot plots and bar graph with scatter plots show the percent of iNKT cells in PBMCs in indicated groups (n = 6/group). G, The bar graph with scatter plots shows the percent of AnnexinV+ iNKT cells in indicated groups (n = 6/group). H, Representative dot plots and summary bar graph with scatter plots show the SA-β-Gal levels of iNKT cells in indicated groups (n = 8–16/group). I, The summary bar graphs with scatter plots show the relative mRNA expression of p16INK4a and p53 of iNKT cells in indicated groups (n = 3–8/group). Data are shown as mean ± SEM and the statistical significance was determined by paired two-tailed Student t test. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Close modal

We next determined whether LCACs contributed to the development of the senescence-associated phenotype of iNKT cells. Exogenous PC increased protein level of SA-β-Gal and transcript levels of cyclin-dependent kinase inhibitors p16INK4a (also known as CDKN2A) and p53 (also known as TP53) in iNKT cells (Fig. 6H and I), while MC and LC showed little effect (Supplementary Fig. S12F and S12G). In addition, SC treatment also slightly increased SA-β-Gal and p53 levels in iNKT cells. These results supported the role of PC and SC in promoting premature senescence of iNKT cells to different degrees. Altogether, our data indicated that exogenous LCACs contributed to the development of iNKT cell senescence.

CD1d-restricted iNKT cells are one of the predominant lymphocyte populations in the liver (41). They can mount strong antitumor responses, thus becoming a promising focus in the development of effective anticancer immunotherapy. In the current study, we showed that iNKT cells from patients with HBV-related HCC suffered severe defects in ratio and α-GalCer responsiveness. Although α-GalCer and IL2 synergistically helped to expand the residual responsive clones, they displayed exhausted and senescent phenotype accompanied with reduced TCR diversity. LCACs, the intermediate metabolites of fatty acid oxidation, were accumulated in the tumor tissue and related to the premature senescence of iNKT cells in patients with HBV-related HCC (Supplementary Fig. S13). Our study implies the challenges and intervention targets for α-GalCer–based and iNKT cell–based immunotherapy in HBV-related HCC.

α-GalCer is a potent activator of iNKT cells. Following treatment of α-GalCer, iNKT cells produce large amounts of cytokines, undergo clonal expansion, and subsequently activate natural killer (NK) cells, T cells, neutrophils, macrophages, and DCs (42). Antitumor effects of iNKT cells are largely relied on the production of Th1 cytokines, especially IFNγ, even though they could directly lyse tumors that express CD1d (43). However, iNKT cell defects have been widely reported in several cancers with dampened IFNγ production, which could potentially decrease the IFNγ-dependent antitumor activity of NK cells and conventional CD8+ T cells (44, 45). Here, iNKT cell deficits in ratio and α-GalCer responsiveness, as well as impaired IFNγ production were found in the peripheral blood, liver tissue surrounding tumors, and in tumors themselves in patients with HBV-related HCC (Fig. 1). Given the correlation between cell senescence and loss of α-GalCer–induced iNKT cell proliferation (Fig. 6), elevated senescence profiles in HCC iNKT were supposed to be associated with defects in ratio and α-GalCer responsiveness of iNKT cells in patients with HCC. However, most senescent T cells and LCAC-induced senescent iNKT cells retained IFNγ production (Supplementary Fig. S12; ref. 40), suggesting other mechanism than cell senescence might be involved. Senescent T cells contribute to inflammaging that provides a chronic inflammation environment and increases the risk of cancer (46). Consistently, HCC iNKT cells had SASP with elevated transcripts of IL6, CXCL8, and IL1A/B (Fig. 3), which are typical mediators of an inflammatory milieu that enhance the expression of tumor-promoting factors. Moreover, increased anti-inflammatory cytokine IL10 by HCC iNKT cells was supposed to additionally lead to the suppression of effective antitumor immune responses (2). The senescent state with dampened cytotoxic capacity endows HCC iNKT clones with tumorigenesis potential with reduced immune surveillance, which needs to be considered during NKT cell–based immunotherapy for different sets of malignant diseases.

However, there are few reports about the iNKT cell senescence and related mechanism. Recently, many studies have proposed the contribution of lipid metabolism and altered metabolites to T-cell senescence (47). Lipids are an important fuel source and create crucial biological intermediates that involve in cellular signaling, cell and organelle structure, and cellular homeostasis. When the tumor develops, liver tissue always has disturbed lipid metabolism, accompanied with mutative lipids constituents including intermediate metabolites. The changed lipid composition in tumor tissue will alter lipid uptake and metabolism in infiltrated T cells, thus determining their fates such as cell division, apoptosis, and cellular senescence. Here, we found accumulated LCACs in tumor tissue of the patients with HBV-related HCC and altered lipid metabolism of HCC iNKT (Fig. 5). LCACs are a kind of intermediate metabolites of fatty acid oxidation and are reported to be involved in the induction of stresses in mitochondria and endoplasmic reticulum (35). A previous study has shown that accumulation of LCFA in the tumor microenvironment impairs mitochondrial function and drives dysfunction in intraphancreatic CD8+ T cells (48). Positively correlation between LCACs and cell senescence has also been proposed previously (49). Consistently, we found HCC iNKT displayed cell senescence-related mitochondrial dysregulation and aberrant calcium transportation process. In addition, LCACs efficiently loaded to the lymphocyte mitochondria and induced premature senescence status of iNKT cells from the healthy donors (Fig. 6). Our current study collectively indicated the senescence status of iNKT from the patients and the contribution of LCACs to iNKT cell senescence. This provides a new target for harnessing the iNKT cell–based antitumor effect by reshaping LCAC metabolism. However, whether inhibition of LCACs accumulation could prevent or delay cellular senescence needs to be further investigated.

Despite the prototypical semi-invariant TCR usage, iNKT cells have TCR variations in Vβ-Jβ pairing and CDR3β loop (50). The diversity of the TCR β chain in iNKT cells ultimately impacts the antigen recognition capacity and consequently the functional outcomes (51). In this study, we found the diversity of CDR3β sequences was significantly reduced in iNKT cells from patients with CHB and HCC when compared with those from healthy donors (Fig. 2). The reduction in TCR diversity is also a hallmark for T-cell senescence and suggested the depletion of activated functional iNKT cell clones, which could explain iNKT cell defects in the patients from different aspects. The residual iNKT cells from patients with CHB and HCC exhibited differential changes in TCR repertoire, suggesting the selection of iNKT cells by the HBV and tumor-generated factors in the patients. The biased TCR usage coincided with the premature senescence phenotype of HCC iNKT cells suggested a relationship between iNKT cell senescence and its TCR repertoire. This would be beneficial for targeting and manipulating the dysfunctional or senescent iNKT cells. Given that senescent iNKT cells had inflammation- and tumor-associated secretome with protumorigenesis potential, engineering CAR-T cells to target and clear senescent iNKT cells might offer alternative therapeutic options for inflammation-associated tumor development.

As a dysfunctional status of T cells in chronic infections and cancers, T-cell senescence has been reported to dampen effective antitumor immunity and sustain the suppressive tumor microenvironment (46). Our study reported iNKT cell defects and senescence status in patients with HBV-related HCC. The validation of distinct TCR repertoire of residual iNKT cells in patients with HBV-related HCC and the identification of LCACs’ roles in iNKT cell senescence add to our understanding of the mechanisms underlying iNKT cell dysregulation and unravel promising targets for anti-HCC immunotherapy. Because of the monopolymorphism of CD1d, iNKT cells show a low risk of graft-versus-host reaction (GVHR) in allogeneic cell fusion, which is different from the high risk of GVHR by allogeneic MHC class I/II-restricted conventional T cells. Therefore, in combination with senescent iNKT cell depletion and/or LCAC metabolism intervention, the application of allogeneic iNKT cells from healthy donors or iNKT-TCR–modified T cells might be attractive to achieve a favorable response in iNKT cell–based immunotherapy against HBV-related HCC.

No disclosures were reported.

X. Cheng: Data curation, software, formal analysis, validation, investigation, visualization, methodology, writing–original draft, project administration. X. Tan: Data curation, formal analysis, funding acquisition, validation, investigation, visualization, methodology, writing–original draft, project administration. W. Wang: Validation, investigation, project administration. Z. Zhang: Formal analysis, validation, investigation. R. Zhu: Conceptualization, resources. M. Wu: Investigation, project administration. M. Li: Project administration. Y. Chen: Project administration. Z. Liang: Conceptualization. P. Zhu: Conceptualization, resources, supervision, writing–review and editing. X. Wu: Conceptualization, supervision, writing–review and editing. X. Weng: Conceptualization, data curation, software, formal analysis, supervision, funding acquisition, methodology, writing–review and editing.

The authors thank the NIH Tetramer Facility for PBS57/hCD1d tetramers.

This work was supported by National Nature Science Foundation of China (NSFC grant 81871235 to X. Weng), Hubei Natural Science Foundation for Distinguished Young Scholars (grant no. 2020CFA074), and China Postdoctoral Science Foundation (no. 2019M662642 to X. Tan).

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.

Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

1.
El-Serag
HB
.
Epidemiology of viral hepatitis and hepatocellular carcinoma
.
Gastroenterology
2012
;
142
:
1264
73
.
2.
Ringelhan
M
,
Pfister
D
,
O'Connor
T
,
Pikarsky
E
,
Heikenwalder
M
.
The immunology of hepatocellular carcinoma
.
Nat Immunol
2018
;
19
:
222
32
.
3.
Jenne
CN
,
Kubes
P
.
Immune surveillance by the liver
.
Nat Immunol
2013
;
14
:
996
1006
.
4.
Jenkins
MK
,
Moon
JJ
.
The role of naive T cell precursor frequency and recruitment in dictating immune response magnitude
.
J Immunol
2012
;
188
:
4135
40
.
5.
Syn
WK
,
Oo
YH
,
Pereira
TA
,
Karaca
GF
,
Jung
Y
,
Omenetti
A
, et al
.
Accumulation of natural killer T cells in progressive nonalcoholic fatty liver disease
.
Hepatology
2010
;
51
:
1998
2007
.
6.
Tan
X
,
Ding
Y
,
Zhu
P
,
Dou
R
,
Liang
Z
,
Yang
D
, et al
.
Elevated hepatic CD1d levels coincide with invariant NKT cell defects in chronic hepatitis B virus infection
.
J Immunol
2018
;
200
:
3530
8
.
7.
Natori
T
,
Morita
M
,
Akimoto
K
,
Koezuka
Y
.
Agelasphins, novel antitumor and immunostimulatory cerebrosides from the marine sponge Agelas-mauritianus
.
Tetrahedron
1994
;
50
:
2771
84
.
8.
Zhang
H
,
Bai
L
.
Challenges of iNKT cell-based antitumor immunotherapies
.
Cell Mol Immunol
2021
;
18
:
1077
8
.
9.
Zhang
Y
,
Springfield
R
,
Chen
S
,
Li
X
,
Feng
X
,
Moshirian
R
, et al
.
alpha-GalCer and iNKT Cell-based cancer immunotherapy: realizing the therapeutic potentials
.
Front Immunol
2019
;
10
:
1126
.
10.
Waldowska
M
,
Bojarska-Junak
A
,
Rolinski
J
.
A brief review of clinical trials involving manipulation of invariant NKT cells as a promising approach in future cancer therapies
.
Cent Eur J Immunol
2017
;
42
:
181
95
.
11.
Heczey
A
,
Courtney
AN
,
Montalbano
A
,
Robinson
S
,
Liu
K
,
Li
M
, et al
.
Anti-GD2 CAR-NKT cells in patients with relapsed or refractory neuroblastoma: an interim analysis
.
Nat Med
2020
;
26
:
1686
90
.
12.
Zhu
Y
,
Smith
DJ
,
Zhou
Y
,
Li
YR
,
Yu
J
,
Lee
D
, et al
.
Development of hematopoietic stem cell-engineered invariant natural killer T cell therapy for cancer
.
Cell Stem Cell
2019
;
25
:
542
57
.
13.
Exley
MA
,
Friedlander
P
,
Alatrakchi
N
,
Vriend
L
,
Yue
S
,
Sasada
T
, et al
.
Adoptive transfer of invariant NKT cells as immunotherapy for advanced melanoma: a phase I clinical trial
.
Clin Cancer Res
2017
;
23
:
3510
9
.
14.
Nakano
S
,
Eso
Y
,
Okada
H
,
Takai
A
,
Takahashi
K
,
Seno
H
.
Recent advances in immunotherapy for hepatocellular carcinoma
.
Cancers
2020
;
12
:
775
.
15.
Passos-Castilho
AM
,
Carvalho
VM
,
Cardozo
KH
,
Kikuchi
L
,
Chagas
AL
,
Gomes-Gouvea
MS
, et al
.
Serum lipidomic profiling as a useful tool for screening potential biomarkers of hepatitis B-related hepatocellular carcinoma by ultraperformance liquid chromatography-mass spectrometry
.
BMC Cancer
2015
;
15
:
985
.
16.
Patterson
AD
,
Maurhofer
O
,
Beyoglu
D
,
Lanz
C
,
Krausz
KW
,
Pabst
T
, et al
.
Aberrant lipid metabolism in hepatocellular carcinoma revealed by plasma metabolomics and lipid profiling
.
Cancer Res
2011
;
71
:
6590
600
.
17.
Ma
X
,
Xiao
L
,
Liu
L
,
Ye
L
,
Su
P
,
Bi
E
, et al
.
CD36-mediated ferroptosis dampens intratumoral CD8(+) T cell effector function and impairs their antitumor ability
.
Cell Metab
2021
;
33
:
1001
12
.
18.
Ringel
AE
,
Drijvers
JM
,
Baker
GJ
,
Catozzi
A
,
García-Cañaveras
JC
,
Gassaway
BM
, et al
.
Obesity shapes metabolism in the tumor microenvironment to suppress anti-tumor immunity
.
Cell
2020
;
183
:
1848
66
.
19.
Jimeno
R
,
Lebrusant-Fernandez
M
,
Margreitter
C
,
Lucas
B
,
Veerapen
N
,
Kelly
G
, et al
.
Tissue-specific shaping of the TCR repertoire and antigen specificity of iNKT cells
.
Elife
2019
;
8
:
e51663
.
20.
Zhao
Y
,
Shao
Q
,
Peng
G
.
Exhaustion and senescence: two crucial dysfunctional states of T cells in the tumor microenvironment
.
Cell Mol Immunol
2020
;
17
:
27
35
.
21.
Liu
X
,
Hoft
DF
,
Peng
G
.
Senescent T cells within suppressive tumor microenvironments: emerging target for tumor immunotherapy
.
J Clin Invest
2020
;
130
:
1073
83
.
22.
Liu
Y
,
Zhu
P
,
Wang
W
,
Tan
X
,
Liu
C
,
Chen
Y
, et al
.
Mucosal-associated invariant T cell dysregulation correlates with conjugated bilirubin level in chronic HBV infection
.
Hepatology
2021
;
73
:
1671
87
.
23.
Calcinotto
A
,
Kohli
J
,
Zagato
E
,
Pellegrini
L
,
Demaria
M
,
Alimonti
A
.
Cellular senescence: aging, cancer, and injury
.
Physiol Rev
2019
;
99
:
1047
78
.
24.
Lian
J
,
Yue
Y
,
Yu
W
,
Zhang
Y
.
Immunosenescence: a key player in cancer development
.
J Hematol Oncol
2020
;
13
:
151
.
25.
Erickson
S
,
Sangfelt
O
,
Heyman
M
,
Castro
J
,
Einhorn
S
,
Grandér
D
.
Involvement of the Ink4 proteins p16 and p15 in T-lymphocyte senescence
.
Oncogene
1998
;
17
:
595
602
.
26.
Schwarze
SR
,
Shi
Y
,
Fu
VX
,
Watson
PA
,
DF
J
.
Role of cyclin-dependent kinase inhibitors in the growth arrest at senescence in human prostate epithelial and uroepithelial cells
.
Oncogene
2001
;
20
:
8184
92
.
27.
Maddika
S
,
Ande
SR
,
Panigrahi
S
,
Paranjothy
T
,
Weglarczyk
K
,
Zuse
A
, et al
.
Cell survival, cell death and cell cycle pathways are interconnected: implications for cancer therapy
.
Drug Resist Updat
2007
;
10
:
13
29
.
28.
Sandhu
C
,
Donovan
J
,
Bhattacharya
N
,
Stampfer
M
,
Worland
P
,
Slingerland
J
.
Reduction of Cdc25A contributes to cyclin E1-Cdk2 inhibition at senescence in human mammary epithelial cells
.
Oncogene
2000
;
19
:
5314
23
.
29.
Lee
BY
,
Han
JA
,
Im
JS
,
Morrone
A
,
Johung
K
,
Goodwin
EC
, et al
.
Senescence-associated beta-galactosidase is lysosomal beta-galactosidase
.
Aging Cell
2006
;
5
:
187
95
.
30.
Callender
LA
,
Carroll
EC
,
Bober
EA
,
Akbar
AN
,
Solito
E
,
Henson
SM
.
Mitochondrial mass governs the extent of human T cell senescence
.
Aging Cell
2020
;
19
:
e13067
.
31.
Desdín-Micó
G
,
Soto-Heredero
G
,
Aranda
JF
,
Oller
J
,
Carrasco
E
,
Gabandé-Rodríguez
E
, et al
.
T cells with dysfunctional mitochondria induce multimorbidity and premature senescence
.
Science
2020
;
368
:
1371
6
.
32.
Miwa
S
,
Kashyap
S
,
Chini
E
,
von Zglinicki
T
.
Mitochondrial dysfunction in cell senescence and aging
.
J Clin Invest
2022
;
132
:
e158447
.
33.
Huang
J
,
Meng
P
,
Wang
C
,
Zhang
Y
,
Zhou
L
.
The relevance of organelle interactions in cellular senescence
.
Theranostics
2022
;
12
:
2445
64
.
34.
Weng
X
,
Kumar
A
,
Cao
L
,
He
Y
,
Morgun
E
,
Visvabharathy
L
, et al
.
Mitochondrial metabolism is essential for invariant natural killer T cell development and function
.
Proc Nat Acad Sci U S A
2021
;
118
:
e2021385118
.
35.
Berezhnov
AV
,
Fedotova
EI
,
Nenov
MN
,
Kasymov
VA
,
Pimenov
OY
,
Dynnik
VV
.
Dissecting cellular mechanisms of Long-chain Acylcarnitines-driven cardiotoxicity: Disturbance of calcium homeostasis, activation of Ca(2+)-dependent phospholipases, and mitochondrial energetics collapse
.
Int J Mol Sci
2020
;
21
:
7461
.
36.
Ziegler
DV
,
Vindrieux
D
,
Goehrig
D
,
Jaber
S
,
Collin
G
,
Griveau
A
, et al
.
Calcium channel ITPR2 and mitochondria-ER contacts promote cellular senescence and aging
.
Nat Commun
2021
;
12
:
720
.
37.
Wiel
C
,
Lallet-Daher
H
,
Gitenay
D
,
Gras
B
,
Le Calvé
B
,
Augert
A
, et al
.
Endoplasmic reticulum calcium release through ITPR2 channels leads to mitochondrial calcium accumulation and senescence
.
Nat Commun
2014
;
5
:
3792
.
38.
Pereyra
AS
,
Rajan
A
,
Ferreira
CR
,
Ellis
JM
.
Loss of muscle carnitine palmitoyltransferase 2 prevents diet-induced obesity and insulin resistance despite long-chain acylcarnitine accumulation
.
Cell Rep
2020
;
33
:
108374
.
39.
Knottnerus
SJG
,
Bleeker
JC
,
Wüst
RCI
,
Ferdinandusse
S
,
IJ
L
,
Wijburg
FA
, et al
.
Disorders of mitochondrial long-chain fatty acid oxidation and the carnitine shuttle
.
Rev Endocr Metab Disord
2018
;
19
:
93
106
.
40.
Mittelbrunn
M
,
Kroemer
G
.
Hallmarks of T cell aging
.
Nat Immunol
2021
;
22
:
687
98
.
41.
Rossjohn
J
,
Pellicci
DG
,
Patel
O
,
Gapin
L
,
Godfrey
DI
.
Recognition of CD1d-restricted antigens by natural killer T cells
.
Nat Rev Immunol
2012
;
12
:
845
57
.
42.
Neparidze
N
,
MV
D
.
Harnessing CD1d-restricted T cells toward antitumor immunity in humans
.
Ann N Y Acad Sci
2009
;
1174
:
61
7
.
43.
Berzofsky
JA
,
Terabe
M
.
NKT cells in tumor immunity: opposing subsets define a new immunoregulatory axis
.
J Immunol
2008
;
180
:
3627
35
.
44.
Molling
JW
,
Kolgen
W
,
van der Vliet
HJ
,
Boomsma
MF
,
Kruizenga
H
,
Smorenburg
CH
, et al
.
Peripheral blood IFN-gamma-secreting Valpha24+Vbeta11+ NKT cell numbers are decreased in cancer patients independent of tumor type or tumor load
.
Int J Cancer
2005
;
116
:
87
93
.
45.
Tahir
SM
,
Cheng
O
,
Shaulov
A
,
Koezuka
Y
,
Bubley
GJ
,
Wilson
SB
, et al
.
Loss of IFN-gamma production by invariant NK T cells in advanced cancer
.
J Immunol
2001
;
167
:
4046
50
.
46.
Chibaya
L
,
Snyder
J
,
Ruscetti
M
.
Senescence and the tumor-immune landscape: Implications for cancer immunotherapy
.
Semin Cancer Biol
2022
;
86
:
827
45
.
47.
Liu
X
,
Hartman
CL
,
Li
L
,
Albert
CJ
,
Si
F
,
Gao
A
, et al
.
Reprogramming lipid metabolism prevents effector T cell senescence and enhances tumor immunotherapy
.
Sci Transl Med
2021
;
13
:
eaaz6314
.
48.
Manzo
T
,
Prentice
BM
,
Anderson
KG
,
Raman
A
,
Schalck
A
,
Codreanu
GS
, et al
.
Accumulation of long-chain fatty acids in the tumor microenvironment drives dysfunction in intrapancreatic CD8+ T cells
.
J Exp Med
2020
;
217
:
e20191920
.
49.
Setoyama
D
,
Lee
HY
,
Moon
JS
,
Tian
J
,
Kang
YE
,
Lee
JH
, et al
.
Immunometabolic signatures predict recovery from thyrotoxic myopathy in patients with Graves' disease
.
J Cachexia Sarcopenia Muscle
2022
;
13
:
355
67
.
50.
Chen
Y
,
Song
Y
,
Du
W
,
Gong
L
,
Chang
H
,
Zou
Z
.
Tumor-associated macrophages: an accomplice in solid tumor progression
.
J Biomed Sci
2019
;
26
:
78
.
51.
Cameron
G
,
Pellicci
DG
,
Uldrich
AP
,
Besra
GS
,
Illarionov
P
,
Williams
SJ
, et al
.
Antigen specificity of type I NKT cells is governed by TCR beta-chain diversity
.
J Immunol
2015
;
195
:
4604
14
.