CD8a antibody-functionalized biomimetic red blood cell membrane ectosomes delivering C646 reverse CD8⁺ T Cell exhaustion via H3K18la histone delactylation in gastric cardia adenocarcinoma

Abstract

The functional exhaustion of CD8⁺ T cells in the tumor microenvironment (TME) severely limits anti-tumor immunity in gastric cardia adenocarcinoma (GCA). Here, we developed CD8a antibody-functionalized biomimetic red blood cell membrane ectosomes (CD8a-NVEs) encapsulating the p300 inhibitor C646 to selectively target and reprogram exhausted CD8⁺ T cells. Single-cell RNA sequencing of human GCA tissues revealed lactate-driven epigenetic remodeling, characterized by elevated H3K18 lactylation (H3K18la) at the PDCD1 promoter, which correlated with impaired CD8⁺ T cell function. In vitro, C646 effectively reduced H3K18la, suppressed PDCD1 transcription, and restored effector molecule expression, including IFN-γ and GZMB. CD8a-NVEs@C646 exhibited superior targeting specificity, biocompatibility, and functional efficacy, markedly enhancing CD8⁺ T cell proliferation and cytotoxicity compared with free C646. In a humanized orthotopic GCA model, CD8a-NVEs@C646 significantly inhibited tumor growth, and its combination with anti–PD-1 therapy further enhanced T cell infiltration and tumor apoptosis. This biomimetic nanoplatform enables precise epigenetic reprogramming of tumor-infiltrating CD8⁺ T cells, overcoming lactate-induced histone modifications and reversing exhaustion. Collectively, these findings present a translational nanobiotechnology-based strategy to potentiate immunotherapy efficacy in GCA and potentially other malignancies driven by T cell dysfunction.

Graphic abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-025-03957-z.

Introduction

Gastric cardia adenocarcinoma (GCA) is a malignant tumor that arises at the gastroesophageal junction and exhibits increasing incidence and mortality worldwide. The disease poses a significant public health challenge, particularly in several Asian countries [1, 2]. Because of its anatomical location and subtle onset, GCA is often diagnosed at an advanced stage. Its aggressive biological behavior facilitates local invasion and distant metastasis [3, 4]. Although surgery, radiotherapy, and chemotherapy are conventional treatments, outcomes for patients with advanced GCA remain suboptimal, with low survival rates [5, 6]. Advances in tumor immunotherapy, especially through immune checkpoint blockade, have introduced new therapeutic strategies for GCA by reactivating immune responses against malignant cells [7, 8]. However, the complex tumor microenvironment (TME), characterized by immunosuppressive factors and elevated lactate accumulation, frequently limits the efficacy of such treatments, resulting in poor response rates among many patients [9, 10]. Therefore, overcoming the immunosuppressive microenvironment to activate the immune system and enhance anti-tumor effects is crucial to improving the efficacy of immunotherapy for GCA.

CD8⁺ T cells serve as key effector cells in anti-tumor immunity by recognizing and directly eliminating malignant cells. Upon encountering tumor-associated antigens, these cells become activated and release cytolytic proteins such as perforin and granzyme, which induce tumor cell lysis [11, 12]. However, within the TME, CD8⁺ T cell proliferation and activation are often suppressed, leading to progressive functional decline known as immune exhaustion [13, 14]. Lactic acid, a metabolic byproduct of tumor cells, accumulates within the TME and disrupts T cell metabolism, thereby suppressing their proliferation and cytotoxic activity [15–17]. Evidence indicates that reversing the immunosuppressive state of CD8⁺ T cells can effectively restore their function and enhance anti-tumor activity [18–20]. Enhancing CD8⁺ T cell responses represents a promising strategy for improving the efficacy of immunotherapy in GCA treatment.

Histone lactylation has emerged as a novel epigenetic regulatory mechanism, particularly significant in the TME, where elevated lactate levels critically influence T cell-mediated immune responses [21]. Histone H3K18la (lactylation of lysine 18 on histone H3) has been proposed as an epigenetic hallmark associated with CD8⁺ T cell dysfunction and exhaustion. In high-lactate environments, H3K18la contributes to impaired CD8⁺ T cell activation and expansion, ultimately diminishing their antitumor capabilities [22]. Recent findings demonstrate that inhibition of histone lactylation, particularly H3K18la, can partially restore T cell proliferation and cytotoxic function, providing an effective strategy to alleviate immune exhaustion [23]. This discovery provides a novel approach for antitumor immunotherapy by restoring the function of immune effector cells through epigenetic modification.

Biomimetic ectosomes derived from red blood cell (RBC) membranes have recently attracted significant attention as promising drug delivery vehicles due to their superior biocompatibility, low immunogenicity, and high structural stability [24]. RBC membrane-derived ectosomes can evade immune recognition and clearance, thereby prolonging systemic circulation in vivo [25, 26]. Additionally, these ectosomes can be specifically modified to target particular cells or tissues [27]. In the present work, an RBC membrane–mimicking ectosome was engineered to incorporate CD8α ligands on its surface for selective CD8⁺ T cell recognition and to encapsulate the p300 inhibitor C646 to promote H3K18la delactylation. This nanovesicle-based delivery system enhances intracellular drug accumulation within CD8⁺ T cells while reducing off-target distribution in vivo, effectively minimizing potential adverse effects. This nanovesicle design presents a novel targeted delivery platform for immunotherapy, offering a promising approach to potentiate antitumor immune responses by restoring T cell functionality..

The primary objective of this study is to investigate whether CD8a antibody-functionalized biomimetic RBC membrane ectosomes loaded with C646 (CD8a-NVEs@C646) can enhance the antitumor immune function of CD8⁺ T cells against GCA by inducing histone H3K18la delactylation in CD8⁺ T cells. Through the integration of single-cell RNA sequencing (scRNA-seq) and computational analyses, this study elucidates the intercellular interactions and metabolic regulation between immune and tumor cells in the TME of GCA. The engineered CD8α-NVEs@C646 exhibits excellent targeting and biocompatibility, effectively enhancing CD8⁺ T cell cytotoxic responses against GCA in both in vitro and in vivo experiments. Furthermore, C646 restores CD8⁺ T cell antitumor activity by suppressing H3K18la lactylation at the promoter region of the Programmed Cell Death Protein 1 (PDCD1) gene. In an orthotopic GCA mouse model, combination therapy using CD8α-NVEs@C646 and αPD-1 markedly inhibited tumor progression, exhibiting potent antitumor efficacy. This nanovesicle-based drug delivery approach provides a novel immunotherapeutic strategy that enhances T cell function and holds potential for improving clinical outcomes in patients with malignant tumors such as GCA.

Materials and methods

Acquisition of human GCA tissue samples

Tumor specimens were obtained from two individuals clinically diagnosed with GCA who underwent radical surgical resection at our hospital in 2024. Following excision, tissues were snap-frozen in liquid nitrogen and stored at − 130 °C for subsequent analyses. Written informed consent was obtained from all participants. All procedures complied with the Declaration of Helsinki and were approved by the Clinical Research Ethics Committee of our institution.

scRNA-seq

Tumor specimens were enzymatically dissociated into single-cell suspensions with trypsin (9002–07-7, Sigma-Aldrich, USA). Individual cells were isolated with the C1 Single-Cell Auto Prep System (Fluidigm, South San Francisco, CA, USA). Cell lysis and mRNA extraction were performed on the microfluidic chip, followed by reverse transcription and on-chip cDNA preamplification. Sequencing libraries were constructed and analyzed on the Illumina HiSeq 4000 platform using paired-end sequencing (2 × 75 bp) with an average depth of ~ 20,000 reads per cell.

Raw sequencing data were processed and filtered using the Seurat package (version 3.1) in R. Cells meeting the criteria 200 < nFeature_RNA < 5000 and mitochondrial gene percentage < 25% were retained. The top 2000 highly variable genes were identified for subsequent analysis. Dimensionality reduction was performed using principal component analysis (PCA), and the top 20 principal components were selected based on the ElbowPlot. Cell clustering was conducted with the “FindClusters” function (resolution = 1), and visualization was achieved through low-dimensional embedding for downstream interpretation.

T Cell subpopulations, cell communication, and metabolic analysis

T cell-specific data were extracted from the annotated scRNA-seq dataset using the subset function in Seurat. Clustering was performed by t-SNE with a resolution of 0.7. T cell subtypes were annotated based on established lineage markers and validated using the CellMarker database. Epithelial cells were similarly isolated, and datasets from T cells and epithelial cells were integrated to analyze intercellular communication with the “CellChat” R package. Metabolic profiling of T cell subsets was conducted using the “scMetabolism” package.

Transcriptome data acquisition

The GCA-related transcriptome dataset GSE2685 was retrieved from the GEO database (https://www.ncbi.nlm.nih.gov/gds). The dataset includes transcriptomic profiles from 22 gastric cancer samples and 8 paired adjacent normal tissues. Differentially expressed genes (DEGs) were identified using the “limma” R package, applying thresholds of |log₂ fold change|> 0.585 and p < 0.05. DEG visualization was performed with volcano plots generated using “ggplot2.”.

Weighted gene co-expression network analysis (WGCNA)

WGCNA was performed on the GSE2685 dataset with the “WGCNA” R package. Genes in the lowest 30% of Median Absolute Deviation (MAD) were removed to minimize noise. The “goodSamplesGenes” function was used to filter low-quality genes, and a scale-free co-expression network was constructed. A soft-threshold power (β) of 4 was selected using the “pickSoftThreshold” function, and the adjacency matrix was converted to a Topological Overlap Matrix (TOM). Hierarchical clustering was applied to group genes into modules with a minimum size of 120. Modules with similar expression patterns were merged at a dissimilarity threshold of 0.25. Module eigengenes (MEs) were calculated to summarize module expression, and correlations between MEs and biological traits were analyzed to identify key functional modules.

Selection of histone lactylation modification factors

Genes related to "CD8 T cell" and "Histone lactylation" were obtained from the GeneCards database, applying a filter for a score greater than 30. The DEGs from the GSE2685 dataset were then intersected with these gene sets, along with the key module genes identified through WGCNA analysis.

Isolation, activation, and expansion of human CD8⁺ T Cells

Human CD8⁺ T cells were isolated from peripheral blood mononuclear cells (hPBMCs, PCS-800–011, ATCC, USA) using the CD8 MicroBead Kit (130–045–201, Miltenyi Biotec, USA). Cell purity was verified by flow cytometry (FCM) with anti-human CD8α-eFluor®450 antibody (12–0088-80, Invitrogen, USA). Magnetic separation was performed according to the manufacturer’s protocol. Isolated CD8⁺ T cells were stimulated with 2 µg/mL anti-CD3 antibody (A19017, ABclonal, China) overnight, followed by 2 µg/mL anti-CD28 antibody (A20346, ABclonal, China) and 100 U/mL recombinant human IL-2 (#200–02, Peprotech, USA) for 48 h. Cells were cultured in RPMI-1640 medium supplemented with 10% FBS, antibiotics, HEPES, sodium pyruvate, L-glutamine, non-essential amino acids, and β-mercaptoethanol. The medium was replaced every 2–3 days to maintain a cell density above 2 × 10⁶ cells/mL. Expansion was continued for up to 14 days.

Lentiviral transduction and grouping of t cells for experimental assays

A PDCD1-overexpressing lentiviral vector (LV4, GenePharma, China) and its empty control vector were used to establish the PDCD1 overexpression system. Lentiviral plasmids targeting p300 (sh-p300#1 and sh-p300#2) and the control sh-NC were obtained from Sigma-Aldrich (USA). Virus packaging was conducted in 293 T cells via calcium phosphate transfection using expression constructs (PDCD1 or sh-p300), pCMV-dR8.91 (ZT1466, Fenghui Biotech, China), and pMD2.G (#12,259, Addgene, USA) at a ratio of 3:3:1. After 24 h, the medium was replaced with serum-free medium, and virus-containing supernatant was collected 24–30 h later, filtered (0.45 μm), and concentrated by ultracentrifugation (1600 g, 2 h, 4 °C). Viral titers were determined with a p24 ELISA kit (L00938, GenScript, China). The shRNA sequences were: sh-p300#1 (TRCN0000039886; target sequence: CCCGGTGAACTCTCCTATAAT), sh-p300#2 (TRCN0000009883; target sequence: CAATTCCGAGACATCTTGAGA), and sh-NC (CTCGCTTGGGCGAGAGTAA).

Activated CD8⁺ T cells were transduced at a multiplicity of infection (MOI) of 25 in the presence of 8 μg/mL puromycin (540,411, Sigma-Aldrich, USA), followed by centrifugation (850 × g, 80 min, 32 °C) and incubation for 9 h. The procedure was repeated on two consecutive days. Transduced cells were maintained in AIM V™ SFM medium (0870112DK, Gibco, USA) supplemented with 100 IU/mL IL-2.

Experimental groupings were defined as follows: (1) sh-NC group, sh-p300#1 group, and sh-p300#2 group; (2) Control (Ctrl) group, Lactate group (CD8⁺ T cells treated with 20 mM sodium lactate for 24 h), Lactate + p300 inhibitor group (CD8⁺ T cells treated with 20 mM sodium lactate and 30 μM p300 inhibitor C646 for 24 h), Lactate + p300 activator group (CD8⁺ T cells treated with 20 mM sodium lactate and 100 μM p300 activator CTB for 24 h), Lactate + sh-NC group, and Lactate + sh-p300 group. The p300 activator (cholera toxin B subunit, CTB; HY-134964) and p300 inhibitor (C646; HY-13823) were both purchased from MCE (USA); (3) PBS + vector group, CD8a-NVEs@C646 + vector group, and CD8a-NVEs@C646 + PDCD1 group, all treated with 20 mM sodium lactate for 24 h, with CD8a-NVEs@C646 treatment for 48 h. Cells from all groups were harvested for downstream analysis.

Detection of CD8⁺ T Cell proliferation by FCM

MKN-45 and SNU1 cells were cultured for 24 h, and the supernatants were collected as conditioned medium (CM) for subsequent treatment of CD8⁺ T cells for an additional 24 h. In separate experiments, CD8⁺ T cells were exposed to sodium lactate (1,614,308, Sigma-Aldrich, USA) at concentrations of 0, 10, or 20 mM for 24 h, or to 20 mM sodium lactate for 0, 6, 12, 24, or 48 h. In some groups, CD8⁺ T cells underwent additional transfection procedures. After treatment, cells were stained with 5 μM CellTrace CFSE (65–0850-84, Thermo Fisher Scientific, USA), and CFSE fluorescence was analyzed by FCM to evaluate proliferation.

Cell culture

The human gastric adenocarcinoma cell lines MKN-45 (RRID:CVCL_0434) and SNU1 (RRID:CVCL_0099) were obtained from Shanghai Enzyme Research Biotechnology Co., Ltd., while 293 T cells (RRID:CVCL_0063) were acquired from ATCC (USA) (https://www.atcc.org/). MKN-45 and SNU1 cells were cultured in RPMI-1640 medium (11,875,168, Gibco, Thermo Fisher Scientific, USA) enriched with 10% FBS (26,010,066, Gibco) and 1% penicillin–streptomycin (100 U/mL penicillin, 100 μg/mL streptomycin). The 293 T cells were maintained in DMEM (11,965,084, Gibco) under identical supplement conditions. All cells were maintained at 37 °C in a humidified incubator containing 5% CO₂ (BB15, Thermo Fisher Scientific, USA). All cell lines were authenticated by short tandem repeat (STR) profiling and tested negative for mycoplasma contamination.

For co-culture assays, activated CD8⁺ T cells subjected to different treatments were incubated with MKN-45 or SNU1 cells at a 1:5 ratio for 24 h. For cytotoxicity assays, cells were harvested, resuspended in PBS, and gastric cancer cells (MKN-45 or SNU1) were identified and separated by FCM using the epithelial marker EPCAM (14–9326-82, Thermo Fisher Scientific, USA).

Lactate quantification

Lactate concentrations in the culture supernatants of MKN-45 and SNU1 cells were measured at 0, 6, 12, and 24 h utilizing a commercial lactate assay kit (BC2235, Solarbio, Beijing, China). Absorbance was recorded at 530 nm with an Epoch microplate spectrophotometer (BioTek, Winooski, VT, USA).

Measurement of IFN-γ levels by ELISA

After treating CD8⁺ T cells under various conditions, culture supernatants were collected, and IFN-γ concentrations were assessed using an ELISA kit (PI511, Beyotime, Shanghai, China).

Assessment of tumor cell cytotoxicity

Cytotoxicity in MKN-45 and SNU1 cells was assessed using the LDH Cytotoxicity Assay Kit (MAK066-1, Sigma-Aldrich, MO, USA). After co-culture with CD8⁺ T cells, 75 μL of culture supernatant was mixed with 150 μL of LDH detection reagent and incubated at room temperature (RT) for 20 min. Absorbance was recorded at 490 nm using a microplate reader. LDH release was calculated as the ratio of supernatant LDH to total LDH (intracellular + extracellular fractions).

FCM detection of GCA Cell apoptosis

Apoptosis of GCA cells was quantified using an FCM apoptosis detection kit (APOAF, Sigma-Aldrich, USA). Following co-culture of MKN-45 cells with CD8⁺ T cells, single-cell suspensions (1 × 10⁵ cells/mL) were stained with 5 μL each of FITC–Annexin V and propidium iodide (PI) and incubated for 20 min at RT in the dark. Samples were analyzed using a Guava® easyCyte™ 6-2L flow cytometer (0500–5007, Luminex), and data were processed with CellQuest Pro software. The apoptosis rate was defined as the combined percentage of cells in quadrants 2 and 3.

ChIP-qPCR and chIP-seq analysis

CD8⁺ T cells treated with PBS or CD8α-NVEs@C646 for 48 h were collected for ChIP analysis using the Pierce™ Magnetic ChIP Kit (Thermo Fisher Scientific, USA). After crosslinking and chromatin fragmentation, samples were incubated overnight at 4 °C with 5 μg of anti-H3K18la (PTM-1427RM, PTM Bio), anti-p300 (ab275378, Abcam), or control IgG (ab172730, Abcam), followed by a 4-h incubation with Protein A/G magnetic beads. Purified DNA was analyzed by qPCR using PDCD1 promoter-specific primers: PDCD1 promoter forward primer: 5’-AGCCGATTAGCCATGGACAG-3’, PDCD1 reverse primer: 5’-CCACTCCCATTCTGTCGGAG-3’. For ChIP-seq, libraries were prepared using the Diagenode MicroPlex Library Preparation Kit v2 and sequenced on the Illumina NextSeq CN500 platform. Reads were aligned to the human genome (hg38) using BWA MEM, and BAM files were sorted with Samtools. Peak visualization and heatmap generation were performed with Deeptools. DEGs were identified based on thresholds of |logFC|> 1 and p < 0.05.

Co-IP assay

Protein interactions between H3K18la and p300 were examined by Co-IP. CD8⁺ T cells treated with PBS, CD8α-NVEs@C646 for 48 h, or transduced with sh-p300 or sh-NC lentivirus were lysed with IP lysis buffer (P0013, Beyotime, China). Lysates were centrifuged at 12,000 g for 20 min at 4 °C, and 200 μg of total protein was incubated overnight at 4 °C with anti-H3K18la (1:50, PTM-1406RM, PTM Bio) or rabbit IgG control (1:50, #3900, CST). Immune complexes were captured with Protein A/G Sepharose beads (sc-2003, Santa Cruz, USA), washed, and eluted by boiling in SDS loading buffer. Proteins were analyzed by Western blot (WB).

RNA extraction and sequencing analysis

Total RNA was extracted from CD8⁺ T cells exposed to PBS or CD8a-NVEs@C646 for 48 h utilizing Trizol reagent (15,596,026, Invitrogen, Thermo Fisher Scientific, USA). RNA integrity and concentration were evaluated via a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, USA).

For transcriptomic profiling, 5 μg of RNA from each sample was submitted to CapitalBio Technology (Beijing, China). Ribosomal RNA removal was carried out with the Ribo-Zero™ Magnetic Kit (MRZE706, Epicentre, USA), and library generation followed the NEBNext Ultra RNA Library Prep protocol (#E7775, NEB, USA). RNA was sheared to ~ 300 bp and converted into strand-specific cDNA using dUTP incorporation, followed by end polishing, adaptor attachment, and USER enzyme treatment (#M5508, NEB, USA). The resulting libraries were PCR-enriched, purified, and quantified with the KAPA Library Quantification Kit (KK4844, KAPA Biosystems), and fragment distribution was examined using the Agilent 2100 Bioanalyzer.

Raw reads were evaluated with FastQC (v0.11.8). Adaptor and poly(A) sequences were trimmed using Cutadapt (v1.18). Reads containing > 5% ambiguous bases were removed with a custom Perl filter, and those with ≥ 70% bases of Q ≥ 20 were retained using FASTX Toolkit (v0.0.13). Paired-end data were corrected with BBMap and aligned to hg38 using HISAT2 (v0.7.12). DEGs were identified in R (v4.2.1) using the “edgeR” package with criteria |log₂ FC|> 0.585 and p < 0.05.

Preparation of RBC membrane-derived ectosomes (NVEs)

According to the latest position statement from the International Society for Extracellular Vesicles (MISEV2023) [28], extracellular vesicles (EVs) can be classified by biogenesis or size. Because the vesicles in this study were artificially engineered from RBC membranes rather than naturally secreted, they are more accurately defined as ectosome-like nanovesicles. Whole blood was sourced from IPHASE (Suzhou, China). After centrifugation at 1,000 × g for 5 min at 4 °C to deplete leukocytes, the erythrocyte pellet was rinsed with PBS and subjected to lysis in hypotonic PBS (65 mOsm) on ice for 10 min. The lysate was clarified by centrifugation at 18,000 × g for 15 min at 4°C. The resulting pellet was incubated in 35 mOsm PBS on ice for 5 min to eliminate residual hemoglobin. The purified RBC membranes were resuspended in PBS and resealed at 37 °C for 1 h. The resealed membranes were extruded 20 times through a 0.4 μm polycarbonate membrane using an Avanti Mini Extruder (610,000, Avanti Polar Lipids, Merck, Germany) to generate NVEs, followed by extrusion through a 0.2 μm membrane (Z373427, Merck, Germany) to achieve uniform size reduction.

Ectosomes loaded with C646 (NVEs@C646)

The EP300 inhibitor C646 (HY-13823, MCE, USA) was dissolved in DMSO (HY-Y0320, MCE, USA) at 10 mg/mL under continuous stirring and diluted to 500 μg/mL in PBS containing 0.05% Triton X-100 (T8787, Sigma-Aldrich, Germany). The solution was maintained at 4 °C with gentle stirring for homogeneous dispersion. CM-DiI dye (HY-D1627, MCE, USA) was added to a final concentration of 12.5 μg/mL, and the mixture was combined with preformed NVEs at a 1:1 volume ratio. The suspension underwent ultrasonication (TL-ST150, Tenlin, China) at 50% amplitude in pulse mode (2 s on/off cycles, total 30 s) to facilitate C646 incorporation. The mixture was subsequently extruded 20 times through a 0.1 μm polycarbonate membrane (Z373419, Merck, Germany) using an Avanti Mini Extruder to obtain uniform NVEs@C646.

Encapsulation efficiency was determined by high-performance liquid chromatography (HPLC) based on the C646 content in NVEs@C646. Protein concentration was measured using a BCA Protein Assay Kit (Thermo Fisher Scientific, USA). Encapsulation efficiency was calculated as: encapsulation efficiency = (amount of C646 loaded/amount of C646 taken)*100.

Preparation of CD8⁺ T Cell-targeted NVEs

For antibody conjugation to the ectosome surface, DSPE-PEG2000-COOH (10 mg/mL, 600 μL in MES buffer; 880135P, Avanti Polar Lipids, Merck, Germany) was activated with 0.25 M NHS (67.5 μL; 56,480, Sigma-Aldrich) and 0.25 M EDC (67.5 μL; 39,391, Sigma-Aldrich) for 30 min at RT. The reaction pH was adjusted to 7.5 with sodium hydroxide. Subsequently, activated DSPE-PEG2000-COOH was then incubated overnight at 4 °C with anti-human CD8α antibody (clone 53–6.7; Cat. No. 300912, BioLegend, USA). Excess reactants were removed via two-step dialysis: first in PBS using a 3.5 kDa molecular weight cutoff (MWCO) membrane for 4 h, followed by dialysis with a 300 kDa MWCO membrane for another 4 h, refreshing PBS midway. The resulting anti-CD8α-PEG2000-DSPE conjugate was stored at 4 °C until use. To prepare CD8α-functionalized NVEs, the conjugate was incubated with preformed NVEs overnight at 4 °C, followed by centrifugation at 3,500 × g for 10 min at 4 °C to isolate CD8α-NVEs@C646.

Characterization of NVEs

The morphology of NVEs was examined using transmission electron microscopy (TEM). A 3 μL droplet of NVE suspension was applied to a carbon-coated copper grid (200 mesh) and allowed to adsorb for 5 min. The grid was stained with 1% uranyl acetate for 5 min, air-dried, and imaged with a TEM (TECNAI-10, Philips, The Netherlands) operated at 80 kV. Size distribution and ζ-potential were measured by dynamic light scattering (DLS) on a Nano ZS90 analyzer (Malvern Instruments, UK). All measurements were performed in triplicate.

Gold immunolabeling TEM (Au-immunogold TEM)

Carbon-coated 200-mesh copper grids were floated face-down on 3 μL of NVE suspension (NVEs@C646 or CD8α-NVEs@C646) for 10–20 min to allow vesicle adsorption. After PBS rinsing, the grids were incubated with a gold-conjugated secondary antibody (5–10 nm gold particles) for 30–60 min at RT. Unbound particles were removed by sequential PBS and distilled-water washes. The grids were then stained with 1% (w/v) uranyl acetate for 5 min and imaged at 80–120 kV using TEM to visualize gold particle distribution [29].

Stability of NVEs

The stability and dispersity (PDI) of NVEs were evaluated in PBS (pH 7.4) and 10% FBS using DLS. NVEs were suspended in the respective media and incubated at 37 °C. Particle size and distribution were measured at 12, 24, 48, and 72 h to monitor temporal changes.

For protein stability assessment, NVEs were isolated by ultracentrifugation at 100,000 × g for 1 h at 4 °C. The pellet was resuspended in 5 × SDS loading buffer, denatured at 90 °C, and subjected to SDS–PAGE on a 12% gel (90 V, 3 h). Gels were stained with Coomassie Brilliant Blue, and images were captured with a UV-P imaging system to assess protein retention following vesicle preparation.

In Vitro biocompatibility study

Hemolysis Assay: Hemocompatibility was evaluated by incubating 400 μL of NVE suspension at varying concentrations with 400 μL of 10% erythrocyte suspension in PBS at 37 °C for 2 h under gentle shaking. Samples were centrifuged at 2,000 × g for 5 min, and supernatant absorbance was measured at 540 nm. Erythrocytes treated with Triton X-100 and DPBS served as positive (PC) and negative (NC) controls, respectively. Hemolysis was calculated as: % hemolysis = (OD sample-OD NC)*100/(OD PC-OD NC).

Cell Compatibility: CD8⁺ T cells were seeded in 96-well plates (1 × 10⁴ cells/well) and treated with serial dilutions of NVEs in complete medium. After 48 h at 37 °C in a humidified 5% CO₂ atmosphere, the medium was replaced with 10 μL of CCK-8 reagent (Dojindo, Japan). Following a 1-h incubation, absorbance was recorded at 450 nm with an Epoch microplate reader (BioTek Instruments). Cell viability was calculated relative to untreated controls. All experiments were performed in triplicate.

In Vitro release of C646

To assess the release kinetics of C646 from NVEs, 600 μL of NVEs@C646 suspension was placed in a Micro Float-A-Lyzer dialysis device (MWCO 8,000 Da) and immersed in 400 mL of PBS (pH 7.4) under gentle stirring at 37 °C. At predetermined intervals, 10 μL samples were collected from inside the dialysis membrane, and C646 concentrations were quantified by HPLC. The percentage release of C646 was then calculated using the specified formula.

In vitro uptake of NVEs by CD8⁺ T Cells

Human CD8⁺ T cells were isolated from peripheral blood mononuclear cells (hPBMCs, PCS-800–011, ATCC, USA) using magnetic beads and seeded at 2 × 10⁵ cells per well in 12-well plates. Murine CD8⁺ T cells were obtained from wild-type (WT) mouse PBMCs isolated using a mouse PBMC isolation kit (Cat. No. 071E100.11, IPHASE, Beijing, China) followed by magnetic separation [30], and seeded at the same density. After 24 h, the culture medium was replaced with 500 μL of fresh medium containing CM-DiI-labeled NVEs. Cells were incubated for 30 min at RT, then washed three times with PBS to remove unbound vesicles. Surface-adsorbed NVEs were eliminated by briefly treating cells with 300 μL of 0.1% Tween-20 in PBS (Sigma-Aldrich, Merck) for 30 s, followed by three additional PBS washes. Cells were fixed in 4% paraformaldehyde (PFA) for 15 min, and nuclei were counterstained with DAPI. Fluorescence imaging was performed with a confocal laser scanning microscope (LSM 510 META, Carl Zeiss AG), with CM-DiI and DAPI signals visualized in red and blue, respectively. Uptake efficiency was quantified by FCM as the percentage of CM-DiI⁺ CD8⁺ T cells.

In vivo targeting of CD8⁺ T Cells by NVEs

NOD SCID immunodeficient mice (4–5 weeks old, 18–22 g; C001070, Cyagen, Jiangsu, China) were acclimated for one week before experimentation. Humanized immune system models were established by intraperitoneal injection of hPBMCs (PCS-800–011, ATCC, USA), leading to immune cell reconstitution within two weeks [31]. Mice were anesthetized with 3% isoflurane (792,632, Sigma-Aldrich, USA), and 5 × 10⁷ MKN-45 cells were injected subcutaneously into the ventral region. Tumor volume was calculated as length × width² × 0.5 and monitored until it reached approximately 400 mm³. Animals were randomly divided into two groups (n = 9; three mice per time point) and intravenously administered 200 μL of CM-DiI–labeled CD8α-NVEs@C646 or NVEs@C646 (containing 50 μg/kg C646). At 1, 24, and 48 h post-injection, tumors were excised, minced, and digested with type IV collagenase (50 U/mL, 17,104,019, Gibco, Thermo Fisher Scientific, USA) and DNase I (20 U/mL, 18,047,019, Invitrogen, Thermo Fisher Scientific, USA). Tumor homogenization was performed using a VIA Extractor™ tissue processor (Cytiva, China) for 37 s, followed by filtration through a 70 μm cell strainer. Single-cell suspensions were also prepared from the spleen and tumor-draining lymph nodes (TdLNs). Peripheral blood was collected via cardiac puncture, and erythrocytes were lysed using ACK buffer (11,814,389,001, Roche, Merck, Germany) [32]. FCM was used to quantify CM-DiI⁺ CD3⁺CD8⁺ and CD3⁺CD8⁻ T cells across tumor, spleen, TdLN, and peripheral blood samples, assessing the in vivo targeting specificity of NVEs.

Experimental model of orthotopic GCA transplantation in mice

A total of 24 humanized immune system NOD SCID mice were randomly assigned to four groups (n = 6 per group): (1) normal saline (NS), (2) CD8α-NVEs@C646, (3) αPD-1, and (4) CD8α-NVEs@C646 combined with αPD-1.

Donor NOD SCID mice were anesthetized with 3% isoflurane (792,632, Sigma-Aldrich, USA), and 5 × 10⁷ luciferase-labeled MKN-45 cells (IML-079, Immocell, Xiamen, China) were injected subcutaneously into the ventral region. After four weeks, when subcutaneous tumors reached approximately 1 cm³, mice were euthanized by isoflurane overdose. Tumors were excised under sterile conditions and dissected into 1 mm³ fragments using a #11 scalpel blade. Only viable peripheral tumor tissue was selected to avoid central necrosis. Recipient mice were anesthetized, and a small midline abdominal incision was made to expose the stomach. A small pocket was created at the gastric cardia (GC) using micro-scissors (RS-5610 VANNAS; Roboz, USA), and a tumor fragment was implanted into the site, secured with tissue adhesive. The stomach was repositioned, and the abdominal incision was closed in two layers with 4–0 absorbable sutures [33].

One week post-transplantation, CD8a-NVEs@C646 (150 μL, 2 mg/mL) was administered via tail vein injection every three days [34]. αPD-1 (6.6 mg/kg) was delivered intraperitoneally on the same schedule. For tumor imaging, fluorescein (150 mg/kg; 46,955, Sigma-Aldrich, USA) was injected intraperitoneally 10 min before bioluminescence imaging, conducted every two weeks using the IVIS-200 system (PerkinElmer, Waltham, MA, USA). After 11 weeks, mice were euthanized, and tumors were harvested, imaged, and weighed [35, 36]. Tumors were bisected: one portion was fixed in 4% PFA for histological analysis, and the other was snap-frozen in liquid nitrogen and stored at − 80 °C for molecular assays.

Histopathological analysis

Tumors and major organs were collected post-euthanasia, rinsed with saline, and fixed in 4% PFA for 30–50 min. Tissues were dehydrated through graded ethanol, cleared in xylene, embedded in paraffin, and sectioned. Sections were mounted on glass slides, dried at 45 °C, and deparaffinized with xylene followed by rehydration in descending alcohol concentrations. Hematoxylin staining was performed for 5 min, followed by differentiation in 1% hydrochloric acid–ethanol for 3 s and eosin counterstaining for 2 min. Slides were dehydrated, mounted, and examined under a light microscope to evaluate histological changes.

Immunohistochemistry (IHC) and terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) staining

Tumor tissues were fixed in 4% paraformaldehyde (PFA) for one week, embedded in paraffin, and sectioned at 4 μm thickness. For antigen retrieval, slides were immersed in 0.1 M citrate buffer (pH 6.0) and heated in a microwave at 100 °C for 10 min. Sections were incubated overnight at 4 °C with rabbit anti-Ki67 antibody (1:200, ab16667, Abcam, UK), followed by HRP-conjugated goat anti-rabbit secondary antibody (1:1000, ab6721, Abcam, UK) for 1 h at RT. Color development was achieved using DAB substrate (DA1010, Solarbio, China), and nuclei were counterstained with hematoxylin for 5 min, differentiated in 1% hydrochloric acid–ethanol for 4 s, and re-blued under running water. Immunostaining intensity was scored as 0 (negative), 1 (weak), 2 (moderate), or 3 (strong), and multiplied by the percentage of positive cells to yield a composite score. Images were captured using a Nikon ECLIPSE Ti microscope and analyzed with Nikon imaging software. All procedures were performed in triplicate.

TUNEL staining was conducted using a commercial apoptosis detection kit (C1089, Beyotime, China). After permeabilization, sections were incubated with the TUNEL reaction mixture and counterstained with Hoechst 33,342 (10 μg/mL, P0133, Beyotime, China) in antifade mounting medium. Apoptotic nuclei were visualized and quantified using an IX53 fluorescence microscope (Olympus, Japan).

Flow cytometric detection of CD8⁺ T Cells in tumor tissue

Blood, tumor, spleen, and TdLN samples were collected from the subcutaneous tumor-bearing mouse models. For orthotopic models, tumors were minced and enzymatically digested with 100 U/mL collagenase IV and 50 μg/mL DNase I at 37 °C for 30 min. Cell suspensions were filtered through 70 μm strainers, washed with PBS, and subjected to RBC lysis using ACK buffer (Thermo Fisher Scientific). Cells were resuspended in FACS buffer (PBS with 0.5% BSA) and stained with fluorochrome-conjugated antibodies: FITC-anti-CD45 (982,316, BioLegend), FITC-anti-CD3 (980,008, BioLegend), APC-anti-CD3 (17–0037–42, 1:100), PE-anti-CD8 (12–0088-80, 1:100), FITC-anti-IFN-γ (11–7319-82, 1:100), APC-anti-Granzyme B (GRB05, 1:100), FITC-anti-PD-1 (11–9969-42, 1:100), and FITC-anti-LAG-3 (11–2231-82, 1:500) (Thermo Fisher Scientific). Fluorescence-minus-one (FMO) controls were included for accurate gating. After 30 min of incubation at RT or 4 °C in the dark, cells were washed and analyzed using a FACSAria II flow cytometer (BD Biosciences). Data were processed with FlowJo software (Tree Star, USA).

real-time quantitative polymerase chain reaction (RT-qPCR)

Total RNA was extracted using Trizol reagent (15,596,026, Invitrogen, Thermo Fisher Scientific, USA), and RNA concentration and purity were evaluated with a NanoDrop 2000 spectrophotometer (1011U, Thermo Fisher Scientific, USA). Reverse transcription was performed with the PrimeScript RT reagent Kit (RR047A, Takara, Japan) at 42 °C for 30–50 min, followed by enzyme inactivation at 85 °C for 5 s. qPCR was conducted with the Fast SYBR Green PCR Kit (RR820A, Takara) on an ABI PRISM 7300 Real-Time PCR System. β-actin served as the internal control. Relative mRNA levels were determined using the 2^⁻ΔΔCt method. All assays were independently performed in triplicate. Primer sequences are listed in Table S1.

WB

Cells and tumor tissues were lysed on ice for 30 min in RIPA buffer supplemented with 1% PMSF (P0013B, Beyotime, Shanghai, China), followed by centrifugation at 14,000 × g for 10 min at 4 °C. Supernatants were collected, and protein concentrations were determined using the BCA Protein Assay Kit. Equal amounts of protein (50 μg) were mixed with 5 × SDS loading buffer, denatured at 100 °C for 10 min, separated by SDS–PAGE, and transferred onto PVDF membranes (FFP28, Beyotime). Membranes were blocked with 5% non-fat milk for 1 h at RT and incubated overnight at 4 °C with primary antibodies, including anti-pan-lysine lactylation (1:1000, PTM-1401RM), anti-H3K18la (1:2000, PTM-1406RM), and anti-H3K9la (1:2000, PTM-1419RM) (PTM Bio, China); anti-Histone H3 (1:5000, ab1791), anti-PDCD1 (1:1000, ab237728), anti-p300 (1:1000, ab259330), and anti-β-actin (1:1000, ab8226) (Abcam, UK). Histone H3 and β-actin served as loading controls. After PBST washes, membranes were exposed to HRP-conjugated goat anti-rabbit IgG (1:10,000, BA1054, BOSTER, China) under ambient conditions for 1 h. Blots were washed six times (5 min per wash) and developed using an ECL substrate (AR1172, BOSTER). Chemiluminescent signals were captured using an Amersham Imager 600 system. Band intensities were examined with ImageJ software. All assays were performed in triplicate.

Statistical analysis methods

All data are presented as mean ± standard deviation (SD) from at least three independent experiments. Comparisons between two groups were performed using unpaired two-tailed Student’s t-tests, while comparisons among three or more groups were analyzed by one-way ANOVA followed by Tukey’s post hoc test. Non-normally distributed data or data with unequal variances were analyzed using the Mann–Whitney U test or Kruskal–Wallis test as appropriate. Correlations were assessed using Spearman’s rank analysis. Statistical analyses were performed with GraphPad Prism 9 (GraphPad Software Inc.) and R (v4.2.1). Statistical significance was defined as p < 0.05 (two-tailed).

Results

Communication and metabolic profiling of cancer and immune cells in GCA using scRNA-seq analysis

To elucidate the TME landscape of GCA, we performed scRNA-seq on tumor samples from two patients (Fig. 1A). After stringent quality control, we retained high-quality single-cell transcriptomes and selected the top 2,000 highly variable genes for PCA. Initial PCA revealed sample-specific batch effects (Figure S1A), which were effectively corrected (Figures S1B-C). The first 20 principal components (PC1–PC20) captured the major transcriptional variance and were used for subsequent clustering and dimensionality reduction (Figure S1D).

Fig. 1.

scRNA-seq Analysis Reveals Changes in Cell Abundance, Cell Communication, and Metabolic Pathways in GCA. Note: (A) Schematic diagram of the scRNA-seq analysis workflow; (B) Visualization of cell annotation results based on t-SNE clustering, with each color representing a distinct cell subgroup; (C) Stacked bar chart showing the proportions of six cell types in the samples; (D) Visualization of T-cell subgroup annotation results based on t-SNE clustering, with each color representing a distinct T-cell subgroup; (E) Circos plots illustrating changes in cell communication, with line thickness in the upper plot representing the number of pathways and in the lower plot representing interaction strength; (F) Bubble plot of metabolic pathway scores across different cell types, with red boxes highlighting the scores for Pyruvate metabolism and the Citrate cycle (TCA cycle). n = 2

Using t-SNE based on the top 20 PCs, we identified 37 distinct cell clusters (Figure S2A). According to established lineage-specific markers and CellMarker database annotations, six major cell types were defined: epithelial cells (clusters 0, 1, 5, 7, 8, 10–12, 15–22, 26, 32), fibroblasts (clusters 24, 27, 34), endothelial cells (clusters 23, 25, 28, 36), mast cells (cluster 31), plasma cells (clusters 2, 9, 29, 30, 33), and T cells (clusters 3, 4, 6, 13, 14, 35) (Fig. 1B). Marker gene expression profiles for these cell types are shown in Figure S2B. Epithelial (cancer) cells and T cells represented the predominant cell populations (Fig. 1C).

To explore interactions between Epithelial cells and T cells, we isolated T cell data and performed further t-SNE clustering, yielding 19 T cell clusters (Figure S2C). Based on canonical markers, six T cell subsets were identified: CD4⁺ T cells (clusters 24, 27, and 34), CD8⁺ T cells (clusters 23, 25, 28, and 36), Central memory T cells (cluster 31), Exhausted T cells (clusters 2, 9, 29, 30, and 33), Natural killer T cells (clusters 2, 9, 29, 30, and 33), and Regulatory T cells (clusters 3, 4, 6, 13, 14, and 35) (Fig. 1D). Expression patterns of representative marker genes for each subset are shown in Figure S2D. Cell–cell communication analysis using the CellChat R package demonstrated extensive intercellular signaling between epithelial cells and T cell subtypes. Among these, the strongest interactions occurred between epithelial and CD8⁺ T cells, followed by Tregs and exhausted T cells (Fig. 1E; Figure S3).

Given the importance of immunometabolism in shaping antitumor responses, we next examined metabolic pathway activity across cell types. Both epithelial and CD8⁺ T cells exhibited high activity in pyruvate metabolism and the tricarboxylic acid (TCA) cycle (Fig. 1F). In tumor cells, pyruvate is preferentially converted to lactate rather than fully oxidized through the TCA cycle, even under normoxic conditions, a hallmark of the “Warburg effect” [37]. This metabolic shift enables rapid ATP generation while promoting substantial lactate accumulation. Elevated lactate levels within the tumor milieu have been shown to impair CD8⁺ T cell effector functions and facilitate immune evasion [38, 39].

To confirm lactate’s role as a mediator of crosstalk between GCA cells and CD8⁺ T cells, we cultured gastric adenocarcinoma cell lines MKN-45 and SNU1 and quantified lactate levels in their supernatants over 0–24 h. Both cell lines exhibited a time-dependent increase in lactate secretion (Figure S4A). CM collected after 24 h was then applied to purified CD8⁺ T cells (> 85% purity confirmed by FCM; Figure S4B). Treatment with CM significantly suppressed IFNG and GZMB expression at both mRNA and protein levels, accompanied by reduced CD8⁺ T cell proliferation (Figure S4C-F). Furthermore, direct exposure of CD8⁺ T cells to increasing concentrations of exogenous lactate recapitulated these effects, leading to a dose-dependent reduction in IFNG and GZMB expression as well as proliferation (Figure S4G-J).

Collectively, these findings demonstrate that GCA tumor cells communicate directly with CD8⁺ T cells via metabolic reprogramming and identify tumor-derived lactate as a critical immunosuppressive metabolite contributing to CD8⁺ T cell dysfunction within the GCA microenvironment.

Bioinformatics analysis reveals EP300 as a potential key regulator of histone lactylation in CD8⁺ T Cells

In the TME, elevated lactate accumulation can induce histone lactylation in CD8⁺ T cells, leading to their functional exhaustion or inactivation [40]. However, histone lactylation in GCA has not yet been reported. To identify potential regulators of this modification in GCA-infiltrating CD8⁺ T cells, we performed an integrative bioinformatics analysis (Fig. 2A). We first analyzed the GEO dataset GSE2685, which contains transcriptomic profiles of gastric cancer and adjacent normal tissues. Differential expression analysis identified 685 genes significantly upregulated and 637 genes downregulated in tumor tissues relative to normal controls (Fig. 2B). Using WGCNA, we identified disease-related gene modules, setting the soft-thresholding power to β = 4 (scale-free R² = 0.9) to ensure a scale-free topology (Figure S5A). Six co-expression modules were obtained (Figure S5B), among which the Blue module exhibited the strongest negative correlation with tumor status (correlation = –0.81, p = 4 × 10⁻⁸) (Fig. 2C). Gene importance analysis indicated that genes in the Blue module were the most significant (Figure S5C).

Fig. 2.

Bioinformatics Analysis Identifies Key Factors in Histone Lactylation Modification. Note: (A) Schematic workflow of bioinformatics analysis for identifying key factors; (B) Volcano plot of DEGs in tumor tissues and adjacent normal tissues from dataset GSE2685 (Normal = 8, Tumor = 22); (C) Heatmap showing the correlation of co-expression module genes with tumor and normal tissues, with each cell displaying the correlation coefficient and p-value; (D) Venn diagram illustrating the intersection of Blue module genes, DEGs, CD8⁺ T cell-related genes, and "Histone lactylation"-related genes; (E) Box plot of p300 differential expression; (F) The tSNE distribution map of EP300 in various cell types in the scRNA-seq data

Next, we intersected genes from the Blue module, DEGs, and genes related to CD8⁺ T cells and "Histone lactylation" obtained from the GeneCards database, identifying EP300 (p300) as a candidate gene (Fig. 2D). p300 is a well-known histone acetyltransferase that also catalyzes H3K18la and H3K9la modifications and its inhibition suppresses tumor growth in gastric cancer xenograft models (PMIDs: 38,711,083, 37,919,243, 33,731,686). Expression analysis further showed that p300 is significantly upregulated in gastric cancer tissues compared with adjacent normal tissues (Fig. 2E). Consistent with this, our scRNA-seq analysis demonstrated that p300 is expressed in both tumor epithelial cells and multiple T cell subsets, including CD8⁺ T cells (Fig. 2F).

These findings indicate that p300 may be a key regulator of histone lactylation in CD8⁺ T cells.

Inhibition of p300 reverses H3K18la histone lactylation in CD8⁺ T Cells and enhances cytotoxicity

To further investigate the regulatory effect of p300 on histone lactylation within CD8⁺ T cells, we conducted in vitro cell experiments (Fig. 3A). WB analysis revealed that exposure of CD8⁺ T cells to exogenous lactate led to a time- and dose-dependent increase in total protein lactylation (PKla) (Fig. 3B-C). Since histone H3K18la and H3K9la modifications are known to modulate CD8⁺ T cell phenotype and function under metabolic stress [41], we next quantified these specific marks. Both histone lactylation sites showed progressive upregulation in response to lactate stimulation (Fig. 3D-E).

Fig. 3.

p300 Regulates Histone Lactylation in CD8⁺ T Cells. Note: (A) Schematic representation of the experimental design, showing the workflow for detecting CD8⁺ T cells treated with lactate, p300 inhibitors, or activators. (B–C) WB analysis of PKla levels in CD8⁺ T cells over time (B) and under varying lactate concentrations (C). (D–E) WB analysis of the time-dependent (D) and dose-dependent (E) changes in H3K18la and H3K9la expression in CD8⁺ T cells following lactate treatment. ***p < 0.001, **p < 0.01, and *p < 0.05 compared to the 0-h or untreated lactate group. (F) WB analysis of H3K18la and H3K9la expression in CD8⁺ T cells following p300 knockdown, activation, or inhibition. (G) ELISA detection of IFN-γ levels in the supernatant of CD8⁺ T cells across different treatment groups. (H) FCM analysis of GZMB expression in CD8⁺ T cells. (I) FCM analysis of CD8⁺ T cell proliferation. (J) LDH release assay showing the cytotoxic effects of CD8⁺ T cells on MKN-45 and SNU1 cells. In panels (F–J), *p < 0.05, **p < 0.01, and ***p < 0.001 compared to the control group; #p < 0.05, ##p < 0.01, and ###p < 0.001 compared to the Lactate group; &p < 0.01 compared to the Lactate + sh-NC group. All cell-based experiments were performed in triplicate

Next, we knocked down p300 in lactate-treated CD8⁺ T cells or treated the cells with a p300 activator or inhibitor. Knockdown efficiency was confirmed by WB, with sh-p300#2 showing the most effective suppression and selected for subsequent analyses (Figure S6A-B). Compared with controls, lactate-treated CD8⁺ T cells displayed increased levels of H3K18la and H3K9la, which were markedly reduced following p300 inhibition via either C646 or sh-p300, but further enhanced by the p300 activator CTB (Fig. 3F). Among the two histone marks, H3K18la exhibited a more pronounced sensitivity to p300 modulation, suggesting it may represent the predominant p300-regulated lactylation site. Functionally, lactate exposure significantly decreased IFNG and GZMB expression and suppressed CD8⁺ T cell proliferation relative to control cells. In contrast, p300 inhibition (Lactate + C646 or Lactate + sh-p300) restored IFNG and GZMB expression and enhanced proliferative capacity, whereas p300 activation (Lactate + CTB) further impaired these effector functions (Fig. 3G-I; Figure S6C-D).

To evaluate the impact on cytotoxic activity, treated CD8⁺ T cells were co-cultured with MKN-45 and SNU1 gastric cancer cells, and target cell lysis was quantified by LDH release. Consistent with reduced effector function, lactate-treated CD8⁺ T cells exhibited lower cytotoxicity. Notably, p300 inhibition significantly increased LDH release, while p300 activation further suppressed cytotoxicity. The Lactate + sh-p300 group demonstrated the highest restoration of tumor cell killing in both MKN-45 and SNU1 models (Fig. 3J).

Collectively, these results demonstrate that p300 inhibition effectively reverses H3K18 histone lactylation in CD8⁺ T cells, thereby restoring their cytotoxic and effector functions against gastric tumor cells.

Construction of CD8a-NVEs@C646

C646, a small-molecule inhibitor of p300, provides rapid and dose-controllable inhibition of enzymatic activity. Unlike siRNA-based approaches, C646 can be efficiently encapsulated and released by ectosome-like nanovesicles, enabling effective suppression of p300 within a short time frame [42, 43]. To improve the C646-mediated cytotoxicity of CD8⁺ T cells against GCA cells, we synthesized NVEs that specifically target CD8⁺ T cells and encapsulate C646 (Fig. 4A). DLS analysis revealed average diameters of approximately 200 nm for NVEs, NVEs@C646, and CD8a-NVEs@C646, with zeta potentials of –12.56 ± 0.46 mV, –10.59 ± 0.58 mV, and –7.50 ± 0.35 mV, respectively. The PDIs were 0.168 ± 0.015 for NVEs, 0.182 ± 0.019 for NVEs@C646, and 0.191 ± 0.017 for CD8a-NVEs@C646, indicating good size uniformity (Fig. 4B–D). TEM further confirmed the membrane-enclosed spherical core–shell structure of NVEs (Fig. 4E). To verify the successful conjugation of anti-CD8a-PEG2000-DSPE to NVEs@C646 after overnight incubation at 4 °C, Au-immunogold TEM analysis was performed. In contrast to NVEs and NVEs@C646 controls, distinct gold nanoparticles were observed on the surface of CD8a-NVEs@C646, confirming successful surface coupling of CD8a antibodies (Fig. 4F).

Fig. 4.

Preparation and Characterization of NVEs. Note: (A) Schematic illustration of the construction of CD8a-NVEs@C646; (B) DLS analysis of nanoparticle sizes for NVEs, NVEs@C646, and CD8a-NVEs@C646; (C) Zeta potential measurements for NVEs, NVEs@C646, and CD8a-NVEs@C646, showing surface charge variations with different modifications; (D) PDI values of NVEs, NVEs@C646, and CD8a-NVEs@C646; (E) TEM images for NVEs, NVEs@C646, and CD8a-NVEs@C646; (F) Au-immunogold TEM images for NVEs, NVEs@C646, and CD8a-NVEs@C646 (scale bar = 200 nm); (G–H) Stability analysis of particle sizes for NVEs, NVEs@C646, and CD8a-NVEs@C646 after 12, 24, 48, and 72 h of incubation in PBS and 10% FBS; (I) SDS-PAGE analysis of membrane protein content for NVEs (2), NVEs@C646 (3), and CD8a-NVEs@C646 (4), with protein marker (1); (J) Encapsulation efficiency of NVEs@C646 and CD8a-NVEs@C646 at different C646 concentrations; (K) In vitro release profiles of C646 from NVEs@C646 and CD8a-NVEs@C646 over different time points. Experiments were repeated three times

The colloidal stability of NVEs, NVEs@C646, and CD8a-NVEs@C646 was assessed in PBS and 10% FBS at 12, 24, 48, and 72 h. DLS measurements showed no significant change in particle size or dispersity (Fig. 4G–H). SDS-PAGE confirmed the preservation of major membrane protein bands across all formulations. WB with anti-CD8a antibody revealed a distinct CD8a band only in CD8a-NVEs@C646, confirming successful antibody conjugation (Fig. 4I).

At an initial C646 concentration of 200 μM, the encapsulation efficiency reached 82–85% (Fig. 4J). In vitro release tests demonstrated that NVEs@C646 and CD8a-NVEs@C646 released around 5.5% of C646 within 1 h, approximately 30% within 2 h, and about 70% and 80% at 6 and 24 h, respectively (Fig. 4K).

These results demonstrate that CD8a-NVEs@C646 maintain structural stability and protein integrity under physiological conditions, providing an efficient and robust nanocarrier platform for targeted modulation of CD8⁺ T cell cytotoxicity.

CD8a-NVEs@C646 targets CD8⁺ T Cells both in vitro and in vivo

To evaluate the biocompatibility and targeting specificity of NVEs both in vitro and in vivo, we designed an experimental protocol as illustrated in Fig. 5A. Intravenous nanomedicine administration frequently results in direct contact with RBCs, potentially inducing hemolysis [44]. However, NVEs derived from RBC membranes retain native surface components, conferring intrinsic biocompatibility. When RBCs were exposed to NVEs (5–25 μg/mL) for 2 h, the hemolysis rate remained below 1% across all concentrations, markedly lower than that of the Triton X-100 positive control (Fig. 5B). We next assessed cytocompatibility by incubating CD8⁺ T cells with varying concentrations of NVEs for 48 h. CCK-8 analysis confirmed that viability remained unchanged even at the highest concentration (250 μg/mL), indicating excellent cytocompatibility (Fig. 5C). Cellular uptake studies revealed efficient internalization of both NVEs@C646 and CD8a-NVEs@C646 by CD8⁺ T cells after 4 h, with CD8a-NVEs@C646 showing significantly enhanced uptake. High-magnification imaging confirmed intracellular localization of CD8a-NVEs@C646 rather than mere surface attachment (Fig. 5D-E). To assess antibody specificity, we performed a cross-reactivity assay using mouse PBMC-derived CD8⁺ T cells. Immunofluorescence analysis revealed negligible binding and uptake of CD8a-NVEs@C646 by mouse CD8⁺ T cells, further confirming the species specificity of the antibody (Figure S7).

Fig. 5.

In Vivo and In Vitro Biocompatibility and Targeting Validation of NVEs. Note: (A) Schematic illustration of in vitro targeting and biocompatibility experiments of NVEs; (B) Hemolysis analysis of NVEs, NVEs@C646, and CD8a-NVEs@C646 after incubation with red blood cells (PC: positive control); (C) CD8⁺ T cell viability assessed using the CCK-8 assay after 48-h incubation with varying concentrations of NVEs; (D) Confocal microscopy and quantitative analysis of NVEs@C646 and CD8a-NVEs@C646 uptake in CD8⁺ T cells (Scale bars = 25 μm); (E) FCM analysis of the uptake efficiency of NVEs@C646 and CD8a-NVEs@C646 in CD8⁺ T cells; (F) Schematic illustration of the subcutaneous tumor model in mice and in vivo injection of NVEs; (G) FCM analysis of the percentage of CM-Dil-labeled CD8⁺ T cells in blood, spleen, tumor tissue, and TdLNs at different time points post-injection (*p < 0.05 compared with the NVEs@C646 group); (H) FCM analysis of CD3⁺CD8⁻ T cells binding with CM-Dil-labeled NVEs recovered from blood, spleen, tumor tissue, and TdLNs at different time points; (I) Quantification of CD8⁺ T cell numbers 48 h after NVEs injection. Cell-based experiments were performed in triplicate. Animal experiments included nine mice per group, with three mice per time point

Subsequently, we established a subcutaneous xenograft tumor model using the MKN-45 mouse cell line (Fig. 5F). CM-DiI-labeled NVEs@C646 and CD8a-NVEs@C646 were administered intravenously, and immune cells were collected from blood, spleen, tumors, and tumor-draining lymph nodes (TdLNs) at defined intervals. FCM at 1 h post-injection showed that over 90% of circulating, splenic, and tumor-infiltrating CD8⁺ T cells were labeled by CD8a-NVEs@C646. Although fluorescence intensity gradually decreased, detectable labeling persisted for up to 48 h (Fig. 5G; Figure S8). Minimal binding to CD3⁺CD8⁻ T cells further confirmed targeting specificity (Fig. 5H), and total CD8⁺ T cell counts remained unaffected (Fig. 5I).

These findings indicate that CD8a-NVEs@C646 can target CD8⁺ T cells both in vitro and in vivo, demonstrating excellent biocompatibility.

Inhibition of p300-mediated histone delactylation of key factor PDCD1 in CD8⁺ T cells

To elucidate the mechanism by which p300 inhibition modulates histone lactylation and influences CD8⁺ T cell cytotoxicity, CD8⁺ T cells treated or untreated with CD8a-NVEs@C646 were subjected to RNA-seq and ChIP-seq analyses using an anti-H3K18la antibody (Fig. 6A). ChIP-seq results revealed a marked reduction in histone lactylation enrichment at transcription start site (TSS) regions following CD8a-NVEs@C646 treatment (Fig. 6B-C). RNA-seq identified 158 genes potentially regulated by histone lactylation, with 140 genes upregulated and 18 downregulated (Fig. 6D). Integrative analysis intersecting these DEGs with ChIP-seq data and CD8⁺ T cell–associated genes (GeneCards score > 30) identified PDCD1 as a common target (Fig. 6E). Further ChIP experiments showed that both CD8a-NVEs@C646 and sh-p300 treatment significantly reduced the enrichment of p300 and H3K18la at the PDCD1 promoter region (Fig. 6F). Co-immunoprecipitation demonstrated a weakened interaction between H3K18la and p300 in both treatment groups (Fig. 6G). Consistently, RT-qPCR analysis showed a downregulation of PDCD1 mRNA in CD8⁺ T cells following CD8a-NVEs@C646 or sh-p300 treatment (Fig. 6H). Additionally, WB analysis confirmed a corresponding decrease in PDCD1 protein levels as well as reduced H3K18la abundance (Fig. 6I).

Fig. 6.

CD8a-NVEs@C646 Facilitates Histone Delactylation Modification in CD8⁺ T Cells. Note: (A) Schematic workflow of RNA-seq and ChIP-seq experiments for CD8⁺ T cells treated with CD8a-NVEs@C646; (B–C) ChIP-seq analysis showing signals at TSS regions in PBS-treated (n = 3) and CD8a-NVEs@C646-treated groups (n = 3); (D) RNA-seq volcano plot illustrating significantly upregulated and downregulated genes in the CD8a-NVEs@C646-treated group compared to PBS (n = 3); (E) Venn diagram of overlapping genes from ChIP-seq and RNA-seq analyses related to CD8⁺ T cells; (F) ChIP analysis of p300 and H3K18la enrichment at the PDCD1 promoter region; (G) The protein interaction between p300 and H3K18la was detected by co-IP assay; (H) RT-qPCR analysis of PDCD1 mRNA expression in CD8⁺ T cells following CD8a-NVEs@C646 treatment; (I) WB analysis of PDCD1 and H3K18la protein expression levels in CD8⁺ T cells treated with CD8a-NVEs@C646. *p < 0.05, ***p < 0.001 compared to the PBS or sh-NC group; experiments were conducted in triplicate

These findings suggest that CD8a-NVEs@C646 promotes histone delactylation of PDCD1 in CD8⁺ T cells by delivering C646, thereby inhibiting PDCD1 expression.

CD8a-NVEs@C646 enhances tumor cell killing by CD8.⁺ T Cells through histone delactylation of PDCD1

Recent studies have reported that the TME induces the upregulation of PD-1 (PDCD1) on tumor-reactive CD8⁺ T cells, leading to compromised antitumor immune responses [45, 46]. Therefore, suppressing PDCD1 expression is crucial for enhancing the efficacy of tumor immunotherapy. To determine whether the CD8a-NVEs@C646 nanoplatform modulates CD8⁺ T cell cytotoxicity through epigenetic regulation of PDCD1, we performed a series of gain-of-function experiments.

We overexpressed PDCD1 in CD8⁺ T cells via lentiviral transduction and treated them with CD8a-NVEs@C646 (Fig. 7A). RT-qPCR and WB analyses demonstrated that CD8a-NVEs@C646 markedly reduced PDCD1 and H3K18la expression in vector-transduced CD8⁺ T cells compared with PBS controls. In contrast, PDCD1 overexpression significantly elevated PDCD1 levels, whereas H3K18la expression remained unchanged relative to the CD8a-NVEs@C646 + vector group (Fig. 7B-C). Functionally, CD8⁺ T cells treated with CD8a-NVEs@C646 exhibited enhanced expression of IFNG and GZMB, along with increased proliferative activity compared with PBS + vector controls. PDCD1 overexpression reversed these effects, significantly reducing IFNG and GZMB expression and proliferation (Fig. 7D-F; Figure S9A-B).

Fig. 7.

CD8a-NVEs@C646-Mediated Regulation of PDCD1 Enhances the Tumor-Killing Ability of CD8⁺ T Cells. Note: (A) Schematic diagram of the experimental design: CD8⁺ T cells were transduced with PDCD1 overexpression lentivirus and treated with CD8a-NVEs@C646, followed by co-culture with MKN-45/SNU1 cells; (B) RT-qPCR analysis of PDCD1 mRNA expression in CD8⁺ T cells across different groups; (C) WB analysis of PDCD1 and H3K18la protein expression in CD8⁺ T cells across different groups; (D) ELISA detection of IFN-γ levels in the supernatant of CD8⁺ T cells across different groups; (E) FCM analysis of GZMB expression in CD8⁺ T cells across different groups; (F) FCM analysis of CD8⁺ T cell proliferation activity; (G) LDH release assay to evaluate the cytotoxicity of CD8⁺ T cells on MKN-45 cells; (H) FCM analysis of apoptosis in MKN-45 cells across different groups. * indicates p < 0.05 compared to the PBS + vector group, *** indicates p < 0.001, # indicates p < 0.05 compared to the CD8a-NVEs@C646 + vector group, ### indicates p < 0.001. All cellular experiments were repeated three times

To further validate these findings, we co-cultured the treated CD8⁺ T cells with MKN-45 and SNU1 cells. Compared with the PBS + vector group, the CD8a-NVEs@C646 + vector group demonstrated significantly reduced viability and increased apoptosis in MKN-45 and SNU1 cells. In contrast, PDCD1 overexpression resulted in increased cell viability and decreased apoptosis in both cell lines compared to the CD8a-NVEs@C646 + vector group (Fig. 7G-H; Figure S9C-D).

These findings indicate that CD8a-NVEs@C646 promotes histone delactylation of PDCD1, inhibiting its expression and thereby enhancing the tumor-killing efficacy of CD8⁺ T cells.

CD8a-NVEs@C646 combined with αPD-1 significantly enhances CD8⁺ T Cell-mediated antitumor immunity to suppress GCA progression

To further validate the antitumor immune effects of CD8a-NVEs@C646 targeting CD8⁺ T cells in GCA, we conducted in vivo experiments. Using luciferase-labeled MKN-45 cells and immunocompromised humanized mice, we established an orthotopic GCA model (Fig. 8A).

Fig. 8.

CD8a-NVEs@C646 Enhances CD8⁺ T Cell-Mediated Anti-Tumor Immunity. Note: (A) Schematic diagram of the experimental design, illustrating the establishment of the GCA orthotopic transplantation tumor model and treatment regimen; (B) In vivo fluorescence imaging of tumors in mice from different treatment groups and quantitative analysis of signal intensity over time; (C) Photographic images of tumor tissues from different treatment groups; (D) Tumor weight statistics for each group; (E–F) Immunohistochemistry and TUNEL staining showing cell proliferation and apoptosis in tumor tissues from each group (Scale bars = 50 μm); (G) Quantification of the proportion of infiltrating CD8⁺ T cells in tumor tissues from each group; (H-J) Expression levels of IFN-γ, TNFα, and GZMB in CD8⁺ T cells within tumor tissues from each group; (K) Proportion of PD-1⁺CD8⁺ T cells in tumor tissues from each group. Six mice per group; * indicates p < 0.05 compared to the NS group, and # indicates p < 0.05 compared to the αPD-1 group

Tumor progression was monitored biweekly via bioluminescence imaging, and terminal tumor weights were recorded at study completion. Treatment with CD8a-NVEs@C646 or αPD-1 monotherapy significantly suppressed tumor growth compared to the NS group. Moreover, the combination treatment (CD8a-NVEs@C646 + αPD-1) led to even greater reductions in tumor volume and weight compared to the αPD-1 group alone (Fig. 8B-D). Consistent with these findings, H&E staining confirmed smaller and less proliferative tumors in treated groups. Immunohistochemistry and TUNEL analyses revealed a marked reduction in Ki67-positive proliferating cells and a significant increase in apoptotic cells in both monotherapy groups, with the combination therapy yielding the most pronounced effects (Fig. 8E-F).

FCM further demonstrated robust immune activation in the tumor microenvironment. CD8⁺ T cell infiltration and the expression of effector molecules IFN-γ, TNFα, and Granzyme B were markedly elevated in both CD8a-NVEs@C646- and αPD-1-treated tumors compared to NS controls, and further enhanced in the combination group (Fig. 8G-J; Figure S10A). Furthermore, we evaluated the degree of T cell exhaustion in vivo by FCM, detecting the expression of LAG-3, a key exhaustion marker, in tumor-infiltrating CD8⁺ T cells. LAG-3⁺ T cells typically exhibit reduced cytokine production, cytotoxicity, and proliferative capacity [47]. The results showed that, compared with the NS group, the proportions of PD-1⁺CD8⁺ T cells and LAG-3⁺CD8⁺ T cells were significantly decreased in both the CD8a-NVEs@C646 and αPD-1 groups. Moreover, the CD8a-NVEs@C646 + αPD-1 group exhibited a further reduction in LAG-3⁺CD8⁺ T cells compared with the αPD-1 group (Figs. 8K and S10B), suggesting that T cell exhaustion was effectively alleviated.

Finally, biosafety evaluation via H&E staining revealed no observable histopathological abnormalities in major organs following systemic CD8a-NVEs@C646 administration, confirming favorable biocompatibility (Figure S11).

These findings indicate that CD8a-NVEs@C646 effectively targets CD8⁺ T cells, significantly enhances their anti-tumor immune functions, and, when combined with αPD-1, further suppresses the progression of GCA.

Discussion

CD8⁺ T cells are pivotal effector lymphocytes mediating antitumor immunity. However, within the TME, their cytotoxic function is frequently impaired by metabolic and epigenetic reprogramming. Among the key immunosuppressive metabolites, lactate has emerged as a major driver of T cell dysfunction by disrupting cellular metabolism, impairing proliferation, and suppressing effector cytokine production [22, 48]. Histone lactylation modifications are closely linked to immune cell exhaustion, particularly in TMEs with elevated lactate levels, where CD8⁺ T-cell function is significantly suppressed [49]. To overcome this challenge, we developed CD8a-NVEs@C646, a targeted biomimetic nanoplatform designed to enhance CD8⁺ T cell antitumor activity by promoting histone delactylation through p300 inhibition. This strategy integrates precise immune cell targeting with epigenetic metabolic regulation, offering a novel and clinically translatable approach for next-generation tumor immunotherapy.

scRNA-seq analysis revealed that GCA tumor cells profoundly reshape the metabolic landscape of CD8⁺ T cells through lactate secretion. This observation is consistent with prior reports demonstrating that lactate suppresses T cell proliferation and activation [50]. In the TME, excessive lactate accumulation exerts potent immunosuppressive effects, impairing T cell function and promoting tumor immune evasion [51, 52]. Previous studies have also indicated that lactate inhibits metabolic pathways in CD8⁺ T cells, such as oxidative phosphorylation and glycolysis, leading to immune cell exhaustion [53]. In this study, targeted delivery of the p300 inhibitor C646 effectively counteracted these suppressive effects, restoring CD8⁺ T cell antitumor activity. This approach not only reverses metabolic exhaustion but also underscores the potential of biomimetic ectosomes as efficient carriers for delivering anti-lactylation therapeutics to immune cells.

C646, a selective inhibitor of the histone acetyltransferase p300, was employed in this study to modulate H3K18la delactylation in CD8⁺ T cells. Unlike conventional immune exhaustion inhibitors, C646 exhibits higher specificity by targeting and modulating the H3K18la histone modification. Histone delactylation can profoundly reshape gene transcription, reprogram immune cell metabolism, and influence activation states. In this study, C646 specifically promoted the delactylation of PDCD1, thereby enhancing the antitumor activity of CD8⁺ T cells within the TME. These findings expand the therapeutic potential of C646 beyond its traditional applications, highlighting its promise as a tool for epigenetic metabolic reprogramming and providing new insights into histone modification–based strategies for cancer immunotherapy.

The CD8a-NVEs@C646 ectosomes developed in this study exhibited high targeting efficiency, structural stability, and excellent biocompatibility both in vitro and in vivo. Biomimetic RBC membrane-modified ectosomes naturally possess advantages in targeting and immune evasion, making them more suitable for delivery within the TME. Compared with conventional nanocarriers, this biomimetic design confers prolonged systemic circulation and enhanced tumor-specific accumulation, while reducing off-target distribution to non-tumor organs. Previous studies have found that the application of ectosomes in tumor immunotherapy is limited by targeting and stability [54]. However, this study addresses these challenges through biomimetic modification, laying the foundation for the future application of ectosomes in cancer therapy.

The combination of CD8a-NVEs@C646 with the PD-1 inhibitor αPD-1 demonstrated significant synergistic antitumor effects. Although PD-1 blockade has achieved remarkable success across multiple cancers, its capacity to fully reverse T-cell exhaustion remains limited. By enhancing the cytotoxic activity of CD8⁺ T cells, CD8a-NVEs@C646 complemented αPD-1 therapy, leading to stronger T-cell activation and effector function within the tumor microenvironment. This combined regimen markedly amplified CD8⁺ T cell–mediated immune responses, resulting in enhanced tumor suppression and improved overall immunotherapeutic efficacy. Collectively, these findings present a promising combinatorial strategy for the treatment of GCA and potentially other solid tumors.

In this study, the application of multi-omics technologies, including scRNA-seq, ChIP-seq, and RNA-seq, provided a comprehensive understanding of the regulatory mechanisms underlying H3K18la delactylation in CD8⁺ T cells, advancing insight into tumor immune regulation. Whereas previous studies have primarily relied on single-omics analyses, this multi-layered approach enabled a more holistic exploration of the intricate crosstalk between immune and tumor cells. Through single-cell multi-omics profiling, we identified critical regulatory factors within the TME and delineated the metabolic and activation dynamics of CD8⁺ T cells at high resolution. These findings offer new insights for optimizing nano-immunotherapy strategies.

Although this study demonstrates the potential of CD8a-NVEs@C646 in antitumor immunotherapy, several limitations remain. First, while CD8a-NVEs@C646 broadly targets CD8⁺ T cells, FCM analysis revealed no significant reduction in CD8⁺ T cell counts in peripheral blood, spleen, or lymphoid tissues, indicating minimal toxicity toward non-exhausted T cells. In addition, in vitro biocompatibility assessments, including hemolysis analysis after incubation of NVEs, NVEs@C646, and CD8a-NVEs@C646 with red blood cells, as well as the viability of CD8⁺ T cells following incubation with varying concentrations of NVEs, demonstrated that NVEs can be considered both hemocompatible and cytocompatible, exhibiting negligible cytotoxicity toward CD8⁺ T cells (Figs. 5B-C). However, this systemic targeting strategy inevitably affects the entire CD8⁺ T cell population. Therefore, long-term safety and off-target effects in non-tumor settings warrant further animal studies to ensure clinical safety. Second, although the animal model used in this study partially recapitulates the TME, inter-patient heterogeneity in TMEs may contribute to variable therapeutic responses. This variability could explain the relatively weaker tumor-suppressive effect observed in the single-agent treatment group. Future work will focus on optimizing therapeutic efficacy by developing CD8-targeting antibodies with improved cross-reactivity or enhanced human specificity. Third, this study did not directly demonstrate nuclear localization of C646. The membrane-derived nature of NVEs and their CD8⁺ T cell–targeted modification may facilitate endosomal escape. Previous studies have reported that exosome- or cell membrane–derived vesicles can enter the nucleus via nucleoporin-associated or cytoplasmic transport factor–mediated pathways [55, 56]. In future studies, we aim to confirm endosomal escape and nuclear transport of C646 using fluorescent labeling and cellular fractionation analyses. Finally, high production costs and challenges in large-scale manufacturing of nanovesicles may limit their clinical translation. Further preclinical and clinical trials are necessary to validate the broad applicability of this strategy.

In summary, this study developed CD8a-NVEs@C646, an innovative biomimetic nanoplatform that significantly enhances the cytotoxic activity of CD8⁺ T cells against GCA cells through H3K18la histone delactylation modification. The combination with αPD-1 therapy further amplified T cell-mediated antitumor immune responses, demonstrating remarkable antitumor effects. This research provides a novel strategy for cancer immunotherapy and underscores the effectiveness of metabolic regulation in enhancing T cell antitumor activity. This study holds significant scientific and clinical relevance by establishing a viable nanodrug delivery platform that augments the specificity and functional activity of immune cells, thereby offering substantial promise for future clinical applications. Nevertheless, further comprehensive investigations on long-term safety, efficacy, and optimization of ectosome targeting are necessary before clinical implementation. Future studies should also explore the integration of this nanoplatform with additional immunotherapeutic strategies and evaluate its potential applicability across diverse tumor types.

Conclusion

Based on the experimental findings, we can draw the following preliminary conclusions: A biomimetic RBC membrane ectosomes targeting CD8⁺ T cells and loaded with C646 was successfully designed. This nanoplatform promotes histone delactylation of PDCD1 in CD8⁺ T cells through C646 delivery, thereby enhancing CD8⁺ T cell antitumor immunity and inhibiting the progression of GCA (Graphic abstract).

This study not only elucidates the role of C646 in regulating H3K18la histone delactylation within CD8⁺ T cells but also introduces an innovative nanodelivery system, CD8a-NVEs@C646, capable of precisely targeting CD8⁺ T cells and boosting their cytotoxic activity. When combined with αPD-1 antibodies, this platform exerts a synergistic antitumor effect, offering a promising strategy to improve the efficacy of immune checkpoint blockade therapies. Furthermore, this work establishes a conceptual foundation for harnessing histone modification–based regulation of immune cell function to advance precision immunotherapy. However, several limitations remain. Although CD8a-NVEs@C646 exhibited favorable stability, targeting efficiency, and biocompatibility in vitro and in vivo, its long-term biosafety and potential off-target effects require further evaluation. In addition, the precise molecular mechanisms through which C646 regulates histone delactylation in CD8⁺ T cells warrant deeper investigation. Finally, as this study primarily focused on PDCD1 delactylation, future research should explore the involvement of other immune checkpoint molecules and metabolic pathways in this regulatory process.

Abbreviations

ACK Lysing Buffer

Ammonium-chloride-potassium lysing buffer

ANOVA

Analysis of variance

APC

Allophycocyanin

BCA

Bicinchoninic acid

BSA

Bovine serum albumin

CCK-8

Cell counting kit-8

CD8a-NVEs@C646

CD8a-modified biomimetic red blood cell membrane ectosomes loaded with C646

ChIP-qPCR

Chromatin immunoprecipitation quantitative polymerase chain reaction

ChIP-seq

Chromatin immunoprecipitation sequencing

DLS

Dynamic light scattering

DEGs

Differentially expressed genes

DNase I

Deoxyribonuclease I

ELISA

Enzyme-linked immunosorbent assay

FCM

Flow cytometry

FMO

Fluorescence minus one

FACS

Fluorescence-activated cell sorting

FITC

Fluorescein isothiocyanate

GCA

Gastric Cardia Adenocarcinoma

GZMB

Granzyme B

HEPES

4-(2-Hydroxyethyl)-1-piperazineethanesulfonic acid

H&E

Hematoxylin and Eosin

HPLC

High-performance liquid chromatography

hPBMCs

Human peripheral blood mononuclear cells

IHC

Immunohistochemistry

IFNG

Interferon gamma

LDH

Lactate dehydrogenase

MAD

Median absolute deviation

ME

Module eigengene

MWCO

Molecular weight cutoff

mOsm

Milliosmole

NC

Negative control

NVEs@C646

Ectosomes loaded with C646

OD

Optical density

PBS

Phosphate-buffered saline

PCA

Principal component analysis

PDCD1

Programmed cell death protein 1

PE

Phycoerythrin

PKla

Pan-lysine lactylation

PMSF

Phenylmethylsulfonyl fluoride

PVDF

Polyvinylidene fluoride

rhIL-2

Recombinant human interleukin-2

RIPA

Radio-immunoprecipitation assay

RPMI

Roswell park memorial institute

RBCs

Red blood cells

RT-qPCR

Real-time quantitative polymerase chain reaction

scRNA-seq

Single-cell rna sequencing

SNU1

Seoul national university-1

TdLN

Tumor-draining lymph node

TEM

Transmission electron microscopy

TME

Tumor microenvironment

TOM

Topological overlap matrix

TSS

Transcription start site

TUNEL

Terminal deoxynucleotidyl transferase dutp nick-end labeling

Tukey’s HSD

Tukeys honestly significant difference

t-SNE

T-Distributed stochastic neighbor embedding

WB

Western blot

WGCNA

Weighted gene co-expression network analysis

Author contributions

Z.X. and X.X.Z. designed and performed most experiments, analyzed data, and wrote the initial manuscript draft. X.L.L., X.H.L., Y.B.H., C.Z., and C.L.Z. contributed to experimental execution, data acquisition, and technical support. A.Q.W. and B.Z. assisted with methodology development, bioinformatic analyses, and data interpretation. W.W., G.D.C., and J.G.J. conceived and supervised the project, secured funding, and critically revised the manuscript. All authors reviewed and approved the final version of the manuscript.

Funding

This study was supported by Anhui Provincial Clinical Key Specialty Construction Project (No. 025), Anhui Provincial Health Commission Key Research Project (No. AHWJ2023A10093), Natural Science Key Project of Anhui Provincial Education Department (No. 2024AH051217), National Natural Science Foundation of China (No. 82403333) and Key Natural Science Project of Bengbu Medical University (2022byzd069).

Data availability

Data is provided within the manuscript or supplementary information files.

Declarations

Ethical statement

This study was approved by the Clinical Ethics Committee of The Second Affiliated Hospital of Bengbu Medical University (Ethics Approval No. 169 in 2024).

All animal experiments were approved by the Animal Ethics Committee of The Second Affiliated Hospital of Bengbu Medical University (Ethics Approval No. 169 in 2024).

Conflict of interest

The authors declare no competing interests.

Consent to participate

Written informed consent was obtained from all participants.

Consent to publish

Not applicable.