Exosomal Pparα derived from cancer cells induces CD8 + T cell exhaustion in hepatocellular carcinoma through the miR-27b-3p /TOX axis.

Introduction
Hepatocellular carcinoma (HCC), accounting for 75–85% of all liver cancer cases, is ranked as the sixth most common cancer globally in terms of incidence and the fourth leading cause of cancer-related deaths.[1] Although early identification and treatment can improve the prognosis of patients, HCC is highly aggressive, fast-growing, and insidious disease that is very hard to diagnose in its early stages. The majority of patients are diagnosed at an advanced stage, and the delayed diagnosis undermines the effectiveness of treatment and decreases the probability of a favorable prognosis. The molecular mechanisms of HCC pathogenesis and the development of drug resistance are still unknown.[2–5] Consequently, research into tumor development and progression is urgently needed to improve treatment strategies for HCC.
The effectiveness of treatment for various cancers is considerably influenced by the characteristics of the tumor microenvironment.[6,7] Communication among tumor cells and the surrounding cellular components within the tumor microenvironment plays a crucial role in the progression and growth of cancer. These interactions govern the cancer progression by orchestrating complex signaling pathways and molecular exchanges.[8] Cluster of differentiation 8 positive (CD8+) T cells with tumor specificity serve as the key immune effectors responsible for eradicating tumor cells within the tumor microenvironment. These cells are considered key targets for immunotherapy and have showed the potential to generate favorable clinical outcomes.[9–11] However, a substantial number of cytotoxic T lymphocytes (CTLs) become exhausted due to the continuous tumor antigen stimulation, exhibit overexpression of inhibitory receptors, and demonstrate reduced effector function.[12–14] Exhausted CD8+ T cells appear as progenitor effector-like and terminally exhausted T cells.[15,16] Progenitor exhausted CD8+ T cells exhibit a strong ability against tumor and respond well to the immune checkpoint blockade (ICB) therapy, but terminally exhausted tumor-infiltrating lymphocytes (TILs) do not.[17] T cell exhaustion, especially the terminally exhausted T cells, is difficult to prevent or reverse, hindering tumor immunotherapy.[18] Thymocyte selection-associated high mobility group box (TOX), a critical regulator of immune cell development, is vital in the differentiation of various immune cell populations.[19] It exerts its function by engaging with nuclear DNA in a structure-dependent manner, resulting in disrupted T cell function and promoting tumor progression.[20–22]
Peroxisome proliferator-activated receptor alpha (PPARα) is a member of the nuclear hormone receptor (NR) superfamily, whose biological function is regulated by many factors, such as stress, growth hormone, glucocorticoid, and insulin.[23] The primary site of PPARα metabolism is the liver, and the stimulation of PPARα in hepatocytes prompts the expression of acetyl-CoA synthase (ACS) and fatty acid β-oxidation system rate-limiting enzymes, such as carnitine palmitoyltransferase 1A (CPT1A), promoting the fatty acid transport to mitochondria where essential proteins are expressed. PPARα does not only enhance the fatty acid metabolizing enzymes’ expression but also promotes the levels of liver peroxisomes and the generation of HCC.[24] The expression level of Pparα messenger RNA (mRNA) is elevated in HCC tissues compared to noncancerous tissues. Additionally, continuous administration of peroxisome proliferators in mice can result in the development of HCC.[25]
Pparα is primarily involved in carcinogenesis through the regulation of downstream target genes expression. However, its association with disease progression and immune escape in HCC and the molecular mechanism by which Pparα promotes tumor cell growth remain unclear. MiR-27b-3p inhibits Pparα expression in the liver,[26,27] and in T cells, it suppresses TOX expression.[28] The upregulation of Pparα may be associated with T cell dysfunction in HCC.
Exosomes, which are small membrane-bound vesicles, play a critical role in the transfer of information and substances among cells within the tumor microenvironment.[29,30] These extracellular vesicles secreted by the specific cells contain several molecules, including proteins, RNA, DNA, and lipids, which are transported to the recipient cells to regulate their function.[31–34] Tumor-derived exosomes (TDEs) promoted angiogenesis, metastasis, and immune escape by altering the phenotype and function of target cells.[35–37] Specifically, exosomes of different tumor cells directly or indirectly caused morphological and functional changes associated with T-cell exhaustion, a process that contribute to tumor progression independently of direct contact with cancer cells.[38–42] Exosomes derived from HCC cell may promote HCC development by inducing T-cell exhaustion through immunosuppression. However, the underlying mechanisms of tumor progression are unknown. This study aimed to investigate the mechanisms by which cancer cell-derived exosomal Pparα contributes to CD8+ T cell exhaustion in hepatocellular carcinoma.

Methods

Ethical approval
All animal procedures were approved by the Institutional Animal Care and Use Committee guidelines at the experimental animal center of the Air Force Medical University (No. 20230796).

Tumor-bearing animal models
Male C57BL/6 mice aged 8–12 weeks and weighing between 21–24 grams were maintained and treated according to the Institutional Animal Care and Use Committee guidelines at the experimental animal center of the Air Force Medical University. A subcutaneous HCC model was established by injecting 3 × 106 Hepa1-6 cells into the mice’s left dorsal region. Tumor growth was monitored over time, with tumor volume calculated using the formula: volume = length × width × height × π/6.

Immunohistochemistry (IHC)
Sections of tissue undergo a 10-min treatment with 3% hydrogen peroxide (Fisher Chemical, Waltham, Massachusetts, USA) to suppress endogenous peroxidase activity, followed by rehydration through an alcohol gradient and dewaxing in xylene. For antigen retrieval, each section was heated in a 0.01 mol/L sodium citrate buffer solution (pH 6.0; Thermo Fisher Scientific, Waltham, Massachusetts, USA) at 100°C for half an hour. Afterward, sections were rinsed three times for 5 min each with phosphate-buffered saline (PBS) and then incubated with a TOX antibody from mouse origin (1:500; Abcam, Cambridge, UK), diluted in PBS, for an hour. This was followed by three more 5-min PBS rinses, after which the sections were incubated with a horseradish peroxidase (HRP)-labeled immunoglobulin (Sigma-Aldrich, St. Louis, USA) at room temperature for another hour. Chromogen diaminobenzidine (DAB; Sigma-Aldrich) was used to display the peroxidase activity following three more washings at room temperature. A brown precipitate of oxidized DAB was visualized. All tissue on microscopic slides underwent independent inspection and scoring for immunoreactivity and histological appearance.

Bioinformatics analysis
The target genes for micro RNA (miRNA) were predicted using TargetScan (https://www.targetscan.org/vert_80/) and miRDB (http://www.mirdb.org/). The correlations of differential molecules were analyzed through TIMER2.0 (http://timer.cistrome.org/) database. UALCAN (http://ualcan.path.uab.edu/) was used to determine the association between clinical characteristics, immune cell infiltration, differential molecular expression levels, and T cell infiltration in HCC tissues and patient survival prognosis. The Kaplan–Meier plotter method was employed to assess the correlation between PPARα expression in HCC tumor tissues and patient survival prognosis. The association between miR-27b-3p in HCC and the indicators of T cell exhaustion, such as cytotoxic T lymphocyte-associated antigen-4 (CTLA4), T cell immunoreceptor with Ig and ITIM domains (TIGIT), programmed death-1 (PD-1), TOX, TIM-3, lymphocyte-activation gene 3 (LAG3), and LAYN (Layilin), was revealed using Starbase database (http://starbase.sysu.edu.cn/). The TOX molecule was subjected to the gene ontology (GO) function enrichment analysis and the Kyoto encyclopedia of genes and genomes (KEGG) pathways enrichment analysis based on data obtained from the DAVID database.(https://davidbioinformatics.nih.gov/tools.jsp). Additionally, the differential expression of molecules in tumor and normal adjacent tissues was evaluated using the online database of differentially expressed miRNAs in human cancers (dbDEMC) (https://www.biosino.org/dbDEMC/index).

Cell culture
In this study, three cell lines were procured from the American Type Cell Culture Collection (ATCC, Manassas, Virginia, USA, https://www.atcc.org/), namely, Alpha mouse liver 12 (AML12) cells, mouse liver cancer (Hepa1-6) cells, and the human acute T cell leukemia (Jurkat) cells. The Jurkat cells were cultured in Roswell Park Memorial Institute 1640 (RPMI-1640) medium with 1% penicillin–streptomycin and 10% heat-inactivated fetal bovine serum (FBS, Hyclone Lab. Inc., Logan, UT, USA). Both Hepa1-6 and AML12 cells were nurtured in Dulbecco’s modified Eagle’s medium (DMEM, Hyclone Lab. Inc) supplemented with identical concentrations of penicillin–streptomycin and heat-inactivated FBS as the Jurkat cells. The cell lines were subcultured bi-daily and incubated at 37°C in an atmosphere containing 5% CO2.

Luciferase assay
Ppara 3′untranslated regions (3′UTR) and Tox 3′UTR sequence-containing dual luciferase reporter gene plasmids were generated. Jurkat cells were co-transfected with these plasmids along with the miR-27b-3p mutant type (MUT) or miR-27b-3p full-length wild type (WT). The luciferase activity was quantified.

Isolation and characterization of exosomes
AML12 and Hepa1-6 cells were washed for twice with PBS and subsequently clutured in a serum-free medium, followed by incubation at 37°C with 5% CO2 for 48 hours. Thereafter, the supernatant was collected for exosome isolation. Multi-stage ultracentrifugation was used for the separation of exosomes. Initially, we spun the culture medium at 250 × g for 10 minutes at 4°C to discard dead cells. This was followed by an additional centrifugation at 10,000 × g for 15 minutes at 4°C, effectively removing any remaining cellular debris. We then filtered the supernatant through a 0.22 μm filter, and the filtrate was subjected to ultracentrifugation at 100,000 ×g for two hours at 4°C. The isolated exosomes were resuspended in PBS and stored at –80°C for later use.
The Pierce BCA Protein Assay Kit (Thermo Fisher Scientific) was utilized to quantify the exosome concentration by measuring the total protein content. The exosomes’ size was evaluated using a NanoPlus instrument (Otsuka Electronics, Tokyo, Japan) after dilution to 1 mg/mL. Finally, we quantified the expression levels of exosomal proteins such as tumor susceptibility gene 101 (TSG101), cluster of differentiation 81 (CD81), and Golgi matrix protein 130 (GM130) using Western blotting analysis.

Tracking exosomes
Specifically, exosomes (1 μg/μL, 300 μL) were labeled with 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate (DiI, Invitrogen, Waltham, MA, USA) and isolated as previously described. About 50 μg of DiI-tagged exosomes were then incubated with Jurkat cells in a glass-bottom dish for 6 h. Post-incubation, cells were fixed with 4% paraformaldehyde for 10 min and rinsed thrice with PBS. Cell nuclei were counter-stained with 1 μg/mL Hoechst 33342 (Beyotime) in a dark chamber at 37°C for 10 min. Cells were washed three times with PBS before examination using a Nikon A1 spectral confocal microscope (Nikon, Tokyo, Japan).
For in vivo tracking, roughly 200 μg of exosomes were labeled with 1,1′-dioctadecyltetramethyl indotricarbocyanine Iodide (DiR)/DiI (Invitrogen) by incubating them with the 1 mmol/L dye in a 500:1 ratio. These marked exosomes were then administered to tumor-bearing mice via the tail vein. After 6 h, mice were euthanized, and their organs and tumors were collected. The distribution of fluorescent exosomes in different organs was observed.
To evaluate exosome uptake by T cells in tumor tissue, tumors were embedded in optimal cutting temperature (OCT) compound and stored at −80°C. Cryosections, 8 μm thick, were obtained using a cryostat, fixed with 4% paraformaldehyde for 10 min, and washed three times with PBS. Sections were then incubated with 5% bovine serum albumin (BSA) for 1 h, followed by overnight incubation at 4°C in a moist, dark chamber with the primary antibody (anti-CD3; 1:800; Servicebio Inc. Wuhan, China). Subsequently, sections were treated with the secondary antibody (FITC-goat anti-rabbit; 1:2000; Invitrogen) at room temperature for 1 h. To visualize cell nuclei, Hoechst 33342 (1 μg/mL) was used to stain the sections for 10 min. Finally, tissue section images were captured using a Nikon A1 spectral confocal microscope.

Construction and characterization of ExomiR-27b-3p
AML12 cell-derived exosomes were used and ExomiR-27b-3p was constructed using the electroporation technique. To characterize the miR-27b-3p functionalized exosomes (ExomiR-27b-3p), Western blotting analysis, nanoparticle tracking analysis (NTA), and transmission electron microscopy (TEM) were used.

Loading of miRNA into exosomes
Exosomes, quantified to a protein concentration of 1 mg/mL with the BCA Protein Assay Kit (Thermo Fisher Scientific), underwent electroporation at 700 V/150 mF with 1 optical density (OD) miRNA mimics within 0.4 cm electroporation cuvettes. Post-electroporation, the mixture was cooled for a minimum of 30 min to facilitate membrane recovery. To purge any unbound nucleic acids from the exosomes, we undertook a sequence of procedures including RNase (Invitrogen) application, washing with cold PBS (Invitrogen), and ultracentrifugation. Details of the miRNA mimics/inhibitor used in this study can be found in Supplementary Table 1, http://links.lww.com/CM9/C682.

Plasmid construction
Pparα 3′UTR was synthesized by GenScript Biotech (Piscataway, NJ, USA) and subcloned into pcDNA3.1(-) vector using restriction enzymes BamHI and HindⅢ, with the resulting clone known as pcDNA3.1(-)-Pparα 3′UTR. Tox was synthesized by GenScript Biotech and subcloned into pcDNA3.1(+) vector using restriction enzymes EcoRV and HindⅢ, with the resulting clone known as pcDNA3.1(+)-Tox. Sequencing confirmed all clones and the correct clones were kept at –80°C for the subsequent procedure.

Transfection with plasmids
We employed Lipofectamine 2000 transfection reagent (Invitrogen) to facilitate the integration of Ppara 3′UTR mRNA into exosomes via the AML12 cells, adhering strictly to the instructions provided by the manufacturer. Once the AML12 cells achieved a density of 60–70% within a 100-mm culture dish, we replaced the existing culture medium with an antibiotic-free variant. Subsequently, we prepared two tubes, each containing 1 mL of DMEM medium: one with 10 μg of the desired plasmid and the other with 20 μL of Lipofectamine 2000. After a 5-min incubation at room temperature, both mixtures were introduced to the AML12 cells in the culture dish.

Functional validation of miR-27b-3p on T cells in vivo
HCC model was randomly allocated into different experimental groups. For the miR-27b-3p mimic intervention, mice received either exo-NC (Negative Control) or exo-miR-27b-3p-mimic via intravenous injection (6 injections in total, administered every 3 days). Tumors were harvested 3 days after the final injection. For the miR-27b-3p antagonism intervention, mice received either antagomir-NC or exo-miR-27b-3p-antagomir (9 injections in total, administered every 2 days), and tumors were harvested 2 days following the last dose. Tumor volume changes in each group were recorded to evaluate the intervention effects.

Functional validation of exosomes on T cells in vivo
A subcutaneous HCC model was established by injecting 3 × 106 Hepa1-6 cells into the left dorsal region of mice. Tumor growth was monitored continuously until the tumor volume reached approximately 100 mm3 (day 7), at which point the mice were randomly allocated into two experimental groups (PBS group and GW4869 group, GW4869 is an inhibitor of exosome secretion). Treatment was initiated via tail vein injection of 0.2 mL PBS or GW4869 solution every three days, for a total of six administrations. Tumor volume changes in each group were recorded to evaluate the intervention effect. Tumor tissues were collected on the third day after the final administration for subsequent research.

Functional validation of HCC-derived exosomes on T cells
To determine how exosomes derived from HCC affect T cell functions, Jurkat cells were co-cultured with Hepa1-6-exo (EXO), AML-12-exo (EXO-NC), AML-12-exo-Pparα3′UTR (EXO-Pparα3′UTR), and AML-12-exo-Pparα3′UTR-miR27 (EXO-Pparα3′UTR-miR27) for 48 h, and stimulated with phytohemagglutinin-L (PHA-L). In vivo, a subcutaneous HCC model was established by injecting 3 × 106 Hepa1-6 cells into the mice’s left dorsal region. Tumor growth was monitored over time, when tumors reached an approximate size of 100 mm3, mice were randomly allocated to various experimental groups (PBS, AML-12-exo, Hepa1-6-exo, AML-12-exo-Pparα3′UTR [exo-3′UTR], AML-12-exo-miR27 [exo-miR27], and AML-12-exo-Pparα3′UTR-miR27 [exo-3′UTR-miR27]). On days 0, 3, 6, 9, 12, and 15, 0.2 mL PBS, AML-12-exo, Hepa1-6-exo, and exosomes (109 particles/mL, derived from AML-12 cells) containing Pparα 3′UTR (2.5 µg/µL, loading rate 68.56%) and/or 0.5 OD miR-27b-3p mimic (5 µg/µL, loading rate 70.08%) were loaded with 2 × 109 particles per kilogram of body weight and administered via the tail vein into the mice bearing subcutaneous tumor. The tumors’ sizes were recorded over time to assess the effectiveness of the intervention in each group. Tumor tissues were collected on the second day following the final administration for additional investigation.

Flow cytometry analysis
Tumor samples were sectioned and subjected to digestion using collagenase type IV (1 mg/mL; Gibco, Waltham, Massachusetts, USA) for an hour at 37°C. This process yielded single-cell suspensions, which were subsequently treated with a red blood cell lysis solution (Beyotime, Shanghai, China) to eliminate red blood cells. Next, the cells underwent staining with a cocktail of antibodies, including the allophycocyanin/fire 750-conjugated anti-cluster of differentiation 3 antibody (anti-APC/Fire750-CD3) (0.5 µg per million cells in 100 µL; BioLegend, San Diego, California, USA), anti-fluorescein isothiocyanate programmed cell death protein 1 antibody (anti-FITC-PD-1) (1 µg per million cells in 100 µL; BioLegend), anti-phycoerythrin/cyanine7 conjugated anti-CD8 antibody (anti-PE/Cyanine7-CD8) (0.25 µg per million cells in 100 µL; BioLegend), and anti-phycoerythrin-T-cell immunoglobulin and mucin-domain containing-3 antibody (anti-PE-TIM3) (0.25 µg per million cells in 100 µL; BioLegend). This staining was conducted at 4°C over a 30-min period, preparing the cells for surface marker analysis via flow cytometry (CytoFLEX; Beckman Coulter, Brea, California, USA). The resulting data were then analyzed using CytExpert software (Beckman Coulter, Brea, California, USA).

Western blotting
Proteins were harvested from samples using the radio immunoprecipitation assay (RIPA) lysis buffer (Thermo Fisher Scientific, Waltham, Massachusetts, USA) under cold conditions for half an hour. The Pierce BCA Protein Assay Kit (Thermo Fisher Scientific) was employed to gauge protein concentrations. The primary antibodies included mouse anti-tumor-specific glycoprotein 101 (anti-TSG101, 1:200; Santa Cruz Biotech, Shanghai, China), rabbit anti-CD81 (1:2000; Proteintech, Chicago, USA), rabbit anti-golgi matrix protein 130 antibody (anti-GM130) (1:1000; Abcam, Cambridge, UK), rabbit anti-TOX (1:500; Abcam), and mouse anti- glyceraldehyde-phosphate dehydrogenase (anti-GAPDH, 1:20,000; Proteintech). Visualization of protein bands was achieved using an enhanced chemiluminescence (ECL) analyzer (GE Healthcare, Pittsburgh, USA). Details of the reagents used could be found in Supplementary Table 2, http://links.lww.com/CM9/C682.

Quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR)
Using the TRIzol reagent (Invitrogen), total RNA was isolated from the samples. For mRNA evaluation, the switching mechanism at 5′ end of RNA template (SMART)-moloney murine leukemia virus (MMLV) reverse transcriptase enzyme (Takara Bio Inc., Shiga, Japan) was used to convert 2 μg of RNA into complementary DNA (cDNA). Conversely, the miRcute Plus miRNA qPCR detection kit (Tiangen, Beijing, China) was employed for reverse transcription of an equivalent amount of RNA into cDNA for miRNA examination.
The qPCR reaction mix was assembled using the FastStart Essential DNA Green Master Kit (Roche, Basel, Switzerland). The 2(–ΔΔCt) method was utilized to ascertain the relative expression of mRNA and miRNA after normalization to Gaphd and U6, respectively. Details of the PCR primer sequences could be found in Supplementary Table 3, http://links.lww.com/CM9/C682.

Statistical analysis
Statistical analyses were conducted using GraphPad Prism 6 (GraphPad Software, San Diego, USA). Data were expressed as means ± standard error of means (SEM). One-way analysis of variance (ANOVA) or the unpaired t-test was used for data analysis. A value of P <0.05 was considered statistically significant.

Results

T-cell exhaustion in HCC
T-cell exhaustion can occur because of continuous exposure to antigens and/or inflammatory signals in chronic infection and cancer progression.[43,44] The exhaustion is initially reversible, but the CD8+ T cell dysfunction later becomes irreversible.[45] Early removal of persistent antigen stimulation allows the recovery and differentiation of CD8+ T cells primed during chronic infections into fully functional memory T cells. However, prolonged exposure to antigen stimulation leads to stable and irreversible exhaustion.[12] C57BL/6 mice bore Hepa1-6 cells for 2, 3, 4, and 5 weeks to determine whether the T cells are functionally altered by persistent tumor-bearing [Figure 1A]. As mice carry tumors for a longer period, the tumor size tended to be larger [Figure 1B and Supplementary Figure 1, http://links.lww.com/CM9/C682, P <0.001], and the proportion of CD3+CD8+ T cells in tumor tissue was also decreased [Figure 1C]. However, the T-cell surface inhibitory receptors, such as TIM-3 and PD-1, were elevated in CD8+ T cells and the expression levels of TOX, Havcr2, Pdcd1 in tumor tissues were also increased, leading to impaired T-cell function [Figure 1C–F and Supplementary Figure 2, http://links.lww.com/CM9/C682]. The CD8+ T cells develop immune exhaustion, resulting in their dysfunction when the duration of the tumor bearing increases.

Downregulation of miR-27b-3p in HCC and its negative association with T cell exhaustion
According to accumulating evidence, T-cell exhaustion serves as a primary mechanism employed by cancer cells to escape the immune system.[46] The differential expression of miR-27b-3p between tumor and normal tissues was assessed using the dbDEMC database, derived from the gene expression omnibus (GEO) database. MiR-27b-3p expression in tumor tissues was lower compared to normal tissues (P = 0.0162; P <0.05) [Figure 2A]. Analysis of CTLA4, TIGIT, PD-1, TOX, TIM-3, LAG-3, and LAYN using the Starbase database revealed that these markers, which are extensively documented in the literature and serve as indicators of T cell exhaustion, exhibit a negative correlation with miR-27b-3p in HCC (P <0.05) [Supplementary Figure 3, http://links.lww.com/CM9/C682]. These findings suggest a potential association between the downregulation of miR-27b-3p and T cell exhaustion.
Additionally, the activity of miR-27b-3p’s downstream gene effectors in HCC was investigated. The Pparα and Tox genes’ 3′ untranslated regions (UTRs) have extremely potent and highly conserved miR-27b-3p binding sites, based on the TargetScan algorithm-based bioinformatics analysis. The miR-27b-3p has complementarity on the sequences of the 3′UTRs of Pparα (positions 4516–4522 nt) and Tox (positions 253–259 nt) [Figure 2B]. To investigate the direct interactions between potential downstream targets and miR-27b-3p, a luciferase assay utilizing the 3′UTR of the luciferase gene was performed. Jurkat cells were treated with Tox or Pparα 3′UTR and miR-27b-3p WT or miR-27b-3p MUT to overexpress the miRNA. The overexpression of miR-27b-3p WT resulted in a decrease in luciferase activity of both Tox and Pparα 3′UTR, with a more pronounced effect observed for Tox [Figure 2C] (P <0.01). The suppressive effect of miR-27b-3p WT was eliminated when the 3′UTR of Pparα, Tox luciferase vectors had the altered binding sites of miR-27b-3p MUT. These results provide confirmation of the molecular communications between miR-27b-3p and the Pparα or Tox. The miR-27b-3p is associated with T cell exhaustion in HCC.

PPARα high expression correlates with advanced cancer stage
Paired samples of peritumoral liver tissue and tumor tissue were used for analyzing PPARα expression using the UALCAN database. HCC-derived exosomes were also used for analyzing PPARα expression. Pparα exhibited high expression levels in both HCC tumor tissues and HCC-derived exosomes [Figure 3A, B] (P <0.001).The expression of PPARα differed with respect to the patient’s gender, age, weight, race, tumor type, and cancer stage [Supplementary Figure 4, http://links.lww.com/CM9/C682]. Tumor tissues with higher PPARα expression exhibited reduced immune cell infiltration [Figure 3C].

MiR-27b-3p suppresses the T cell exhaustion via downregulation of TOX
Immune checkpoint molecules, including PD-1, CTLA4, HAVCR2, LAG-3, TIGIT, and LAYN were positively correlated to the TOX expression (P <0.001) [Supplementary Figure 5A, http://links.lww.com/CM9/C682]. Functional analyses indicate that TOX is involved in critical immune-related processes [Supplementary Figure 5B, C, http://links.lww.com/CM9/C682]. The miR-27b-3p mimic or inhibitor is transfected into T cells [Figure 4A]. The construction and characterization of ExomiR-27b-3p are presented in Supplementary Figure 6, http://links.lww.com/CM9/C682. As expected, these data demonstrate that exosomes were successfully extracted and their properties, size, and morphology were not affected after loading miR-27b-3p. The miR-27b-3p mimic led to a decrease in TOX protein expression, while the miR-27b-3p inhibitor increased TOX protein expression [Figure 4B]. These results confirm that TOX was regulated by miR-27b-3p. The expression levels of PD-1 and TIM-3 in T cells were increased following the inhibition of miR-27b-3p, whereas they are decreased upon treatment with a miR-27b-3p mimic [Figure 4C]. Furthermore, the mimic reduced the expression levels of Pdcd1 (P <0.01, t = 5.448; P <0.01, t = 5.193) and Havcr2 (P <0.05, t = 3.594; P <0.01, t = 5.132), whereas the inhibitor increased their expression [Figure 4D and E]. By injecting Hepa1-6 cells, a xenograft mouse model was established. The mice were administered with either EXO-miR-27b-3p mimic, miR-27b-3p antagomirs, or negative controls to examine their effects [Supplementary Figure 7A, http://links.lww.com/CM9/C682]. The miR-27b-3p antagomir treatment resulted in accelerated tumor growth, while the EXO-miR-27b-3p mimic treatment suppressed tumor growth [Supplementary Figure 7B, http://links.lww.com/CM9/C682]. The miR-27b-3p antagomirs increased the levels of Tox (P <0.001, t = 11.68; P <0.001, t = 9.952), Havcr2 (P <0.05, t = 3.041; P <0.01, t = 6.779), and Pdcd1 (P <0.05, t = 4.153; P <0.05, t = 2.576), while the EXO-miR-27b-3p mimic decreased their levels [Supplementary Figure 7C, D, http://links.lww.com/CM9/C682]. Moreover, the EXO-miR-27b-3p mimic increased the infiltration of CD3 and CD8+ T cells into tumor tissue, whereas the miR-27b-3p antagomirs decreased their infiltration (P <0.01) [Supplementary Figure 7E, http://links.lww.com/CM9/C682]. These results collectively indicate that miR-27b-3p regulates T cell immune function by targeting Tox.

Exosome tracing and verification of endocytosis of exosomes by T cells
To investigate whether exosomes released by HCC were absorbed by Jurkat cells, DiI-labeled exosomes were co-cultured with Jurkat cells for 6 h, and the exosomes that were taken up by Jurkat cells were viewed under a confocal fluorescence microscope [Figure 5A]. Then, to further verify whether HCC-derived exosomes can be endocytosed by tumor-infiltrating T cells in vivo, equal amounts of DiR/DiI-labeled exosomes or control PBS were injected into the tumor via tail vein. After 6 h, noninvasive fluorescence imaging was performed using DiR dye to track the distribution of exosomes [Figure 5B]. Then the tumor tissue was harvested to determine whether exosomes were internalized in tumor-infiltrating CD3+ T cells. Confocal imaging showed that CD3 and DiI were co-expressed in mice tumors, confirming the uptake of exosomes by tumor-infiltrating CD3+ T cells [Figure 5C].

Hepatocellular carcinoma-derived exosomes drive the T cell immune dysfunction
Exosomes were crucial in transporting essential cellular cargo, enabling the interaction between cancer cells and immune cells.[47] Therefore, how exosomes affect the roles of T cells in the immune escape process of tumor cells were elucidated herein. To determine how exosomes derived from HCC affects T cell functions, the exosomes were extracted from cell lines using differential centrifugation, incubated with Jurkat cells for 48 h, and stimulated with phytohemagglutinin-L (PHA-L) [Figure 6A]. Compared to incubation with PBS or control exosomes (Con-Exo), the presence of HCC-derived exosomes resulted in elevated levels of Ppara, Pdcd1, Tox, and Havcr2 [Figure 6B–C] (P <0.001). Further, the exosome secretion inhibitor GW4869 was used [Figure 6D]. Subcutaneous tumors size in the PBS control group or the groups that receive GW4869 treatment differed [Figure 6E]. The GW4869 treatment group exhibited increased infiltration of CD3 and CD8+ T cells into tumor tissue (P <0.05, t = 2.799). Interestingly, the expression of TOX (P <0.01, t = 7.754), TIM-3 (P <0.01, t = 7.919; P <0.01, t = 6.696), and PD-1 (P <0.01, t = 5.495; P <0.05, t = 3.505) in GW4869 group was reduced in tumor-infiltrating CD8+ T cells compared to control group [Figure 6F–J]. To summarize, these findings indicate that the HCC-derived exosomes cause the T cells’ immune dysfunction.

HCC-derived exosomal Pparα promotes T cell exhaustion via the miR-27b-3p/TOX axis
To investigate the role of HCC-derived exosome-transferred Pparα in regulating T cell effector functions, we conducted flow cytometry, Western blotting, and qRT-PCR analyses to assess Tox and Pdcd1 levels in Jurkat cells co-cultured with EXO, EXO-NC, EXO-Pparα 3′UTR, and EXO-Pparα 3′UTR-miR27 for 48 h [Supplementary Figure 8A, http://links.lww.com/CM9/C682]. The Jurkat cells co-cultured with EXO overexpressing Pparα 3′UTR showed a significant increase in TOX (P <0.001), TIM-3 (P <0.001), and PD-1 (P <0.001) levels, while miR-27b-3p (P <0.01) expression was decreased. However, the levels of TIM-3 (P <0.01), PD-1 (P <0.001), and TOX (P <0.001) in Jurkat cells from the EXO-Pparα 3′UTR-miR27 group did not exhibit significant changes compared with EXO and EXO-NC groups [Supplementary Figure 8B–D, http://links.lww.com/CM9/C682]. Overall, our findings indicate that HCC-derived exosome-transferred Pparα effectively suppressed miR-27b-3p expression in Jurkat cells, leading to the upregulation of TOX, TIM-3, and PD-1 levels in Jurkat cells. Animal studies are conducted to further determine the function of HCC-derived exosomal Pparα on immune escape via the miR-27b-3p/Tox axis. The mice were treated with Hepa1-6-EXO, EXO-Pparα3′UTR, EXO-miR27, EXO-Pparα3′UTR-miR27, PBS, and AML-12 EXO every 3 days, using PBS and AML-12 EXO as control [Figure 7A]. The EXO-Pparα3′UTR aggravated the tumor growth whereas the EXO-Pparα3′UTR-miR27 slowed the tumor growth (P <0.001) [Figure 7B]. The upregulation of Pparα3′UTR resulted in increased levels of TOX (P <0.001), PD-1 (P <0.001), and TIM-3 (P <0.001) in mouse tumor tissues, accompanied by reduced infiltration of CD3 and CD8+ T cells (P <0.001). Additionally, it led to elevated PD-1 and TIM-3 expression specifically in CD8+ T cells (P <0.001). However, the concurrent increase in miR-27b-3p expression attenuated or inhibited these observed changes (P <0.001) [Figure 7C–F]. These findings suggest that HCC-derived exosomal Pparα promoted T cell exhaustion via the miR-27b-3p/TOX axis.

Discussion
Immune checkpoint-based therapy might be a potential option for managing HCC, a malignant tumor, which is aggressive with unfavorable outcomes.[48,49] HCC’s immune evasion is driven by the interaction of TME components and HCC cells.[50,51] T cells become exhausted as a result of exposure to ongoing antigen and/or inflammatory signals due to persistent infection or cancer progression.[52] Effective immunotherapy for HCC is severely hampered by the exhaustion of CD8+ T cells.[53] We demonstrate that HCC-derived exosomes promote CD8+ cell exhaustion in HCC via the Pparα/miR-27b-3p/TOX axis [Figure 8].
T cells play a crucial role in eliminating cancer cells.[54] A group of immune checkpoint molecules, also known as co-inhibitory receptors, and their ligands regulate this process, resulting in tumor immune escape.[55–60] Among the epigenetic transcription factors associated with the immune checkpoint molecule expression in many malignancies, the TOX was an important therapeutic target.[61–64] In recent years, miRNAs have emerged as regulators of TOX expression in T cells, exerting influence on its levels.[21,65–67] As a member of the miR-27 family, miR-27b-3p serves as a tumor-suppressive miRNA and is frequently downregulated in solid malignancies.[68,69] It directly interacts with the 3′UTR region of Tox and impedes its expression, along with immune checkpoint molecules, across various tumor types, including HCC. In our investigation, we discover that the 3′UTR sequence of Pparα mRNA derived from HCC facilitates the binding of miR-27b-3p, thereby alleviating the inhibitory effect of miR-27b-3p on Tox. Consequently, this leads to the upregulation of TOX and immune checkpoint molecules in T cells infiltrating the tumor microenvironment. Enhanced expression of TOX in activated T cells is induced by tumor cells, ultimately leading to T cell exhaustion. However, the precise mechanism underlying this process remains elusive. Our investigation reveals a novel finding, demonstrating that cancer cells secrete Pparα-containing exosomes that carry Pparα mRNA. Subsequently, these exosomes are internalized by T cells, resulting in increased Tox expression through the sequestration of miR-27b-3p. This study elucidates the role of tumor cell-derived mRNA in modulating Tox and immune checkpoint molecules in T cells through competitive binding of miRNAs.
This study has several limitations. First, we cannot formally exclude potential off-target effects resulting from the molecular tools employed, nor can we rule out the involvement of alternative mechanisms regulating the expression of TOX. Second, the clinical relevance of this axis in human HCC remains to be fully established and warrants further validation using patient-derived samples and larger clinical cohorts. Future studies should aim to optimize the specificity of genetic and pharmacological tools to minimize off-target confounders. Third, comprehensive multi-omics approaches using human HCC specimens may help clarify the translational potential of these findings and facilitate the development of biomarker-driven therapeutic strategies.
In summary, the upregulation of Pparα is observed in both cancer cell-derived exosomes and HCC tissues. These exosomes released by cancer cells facilitate the transfer of Pparα to T cells, leading to enhanced expression of Tox by sequestering miR-27b-3p within T cells. These findings indicate that Pparα holds promise as a potential immunotherapeutic target and a diagnostic/prognostic biomarker for HCC. Notably, Pparα plays a crucial role in promoting the expression of TOX and immune checkpoint molecules in T cells.

Acknowledgments
We appreciate Zhenzhen Hao and Jing Zhang of the Department of Biochemistry and Molecular Biology, Air Force Medical University for their technical assistance.

Funding
This study was supported by grants from the Shaanxi Provincial Science and Technology Department, Social Development Fund Project (No. 2024SF-YBXM-140) and the Tangdu Hospital Major Clinical Technology Innovation Project (No. 2024LCJS005).

Conflicts of interest
None.

Supplementary Material