m⁶A‐modified EHD1 controls PD‐L1 endosomal trafficking to modulate immune evasion and immunotherapy responses in lung adenocarcinoma

Abstract

Background

Eps15 homology domain (EHD) proteins, including EHD1 to EHD4, play vital roles in tumor progression. In this study, we aimed to investigate which specific EHD proteins, if any, are implicated in tumor immune evasion and immunotherapy response.

Methods

The immunotherapy responses of lung adenocarcinoma (LUAD) patients were predicted using tumor immune dysfunction and exclusion (TIDE) analysis. The T cell killing assay was performed by co‐culturing activated T cells with LUAD cells. The function of EHD1 as a regulator of programmed death‐ligand 1 (PD‐L1) endocytic recycling was determined by receptor internalization assays. Methylated RNA immunoprecipitation (MeRIP) was performed to investigate N6‐methyladenosine (m⁶A) modification of EHD1 mRNA. The protein‐protein interaction was revealed by the molecular docking analysis and validated by immunofluorescence (IF) and immunoprecipitation (IP) assays. RNA immunoprecipitation (RIP) was used to examine the interaction between YTH N6‐methyladenosine RNA‐binding protein 1 (YTHDF1) and EHD1 mRNA. The regulatory mechanism of YTHDF1 on EHD1 was investigated through the application of m⁶A‐binding site mutation analysis. The murine LUAD cells were employed to establish subcutaneous xenograft models within immunocompetent C57BL/6 mice to assess the immunomodulatory impact of EHD1 in vivo.

Results

TIDE algorithms and survival analysis identified that EHD1 promoted LUAD immune escape. EHD1 knockdown enhanced T cell cytotoxicity in killing LUAD cells across all effector‐to‐target (E/T) ratios. EHD1 overexpression exerted the opposite effect. The molecular docking analysis revealed an interaction between EHD1 and the PD‐L1 protein, verified by IF and IP. Furthermore, EHD1 knockdown inhibited PD‐L1 recycling, thereby promoting its lysosomal degradation. Disruption of the EHD1/PD‐L1 interaction impaired the regulatory function of EHD1 in tumor immune evasion. In an immune‐competent mouse model, we found that EHD1 silencing impeded tumor immune evasion and enhanced the efficacy of anti‑PD‑1 therapy. MeRIP‐qPCR confirmed obvious m⁶A modification of EHD1. Further, the EHD1 mRNA was found to bind to the YTHDF1 protein, an m⁶A reader. YTHDF1 overexpression up‐regulated EHD1 expression by enhancing its mRNA stability in an m⁶A‐dependent manner.

Conclusion

Our study illuminates the role of m⁶A‐modified EHD1 in tumor immune evasion and immunotherapy responses, thereby offering a novel avenue to potentially enhance immunotherapeutic sensitivity and improve the prognosis for patients with LUAD.

Abbreviations

Act D

Actinomycin D

Bax

Bcl‐2 Associated X

Bcl‐2

B‐cell lymphoma‐2

BCA

Bicinchoninic acid

CCK‐8

Cell counting Kit‐8

CD8

Cluster of differentiation 8

CD155

Cluster of differentiation 155

CFSE

Carboxyfluorescein diacetate succinimidyl ester

CTLA4

Cytotoxic T‐lymphocyte antigen 4

CHX

Cycloheximide

CTLs

Cytotoxic T lymphocytes

E/T

Effector‐to‐target ratio

EEs

Early endosomes

EH

Eps15 homology

EHD

Eps15 homology domain

EHD1‐FL

Full‐length EHD1

EHD1^MUT1

mutant EHD1, which displayed overexpression of EHD1 lacking the EH domain

EHD1^MUT2

mutant EHD1, which exhibited overexpression of EHD1 with mutations at the m6A peak sites

EHD1^OE

EHD1 overexpression

EHD1‐ΔEH

EHD1 variants lacking the EH domain

ELISA

Enzyme linked immunosorbent assay

ERC

Endocytic recycling compartment

FBS

Fetal bovine serum

FCM

Flow cytometry

FGL1

Fibrinogen‐like Protein 1

GAPDH

Glyceraldehyde 3‐phosphate dehydrogenase

GTPase

Guanosine Triphosphate ase

HE

Hematoxylin and eosin

ICB

Immune checkpoint blockade

IGF2BP1

Insulin‐like growth factor 2 mRNA binding protein 1

IF

Immunofluorescence

IFN‐γ

Interferon γ

IHC

Immunohistochemistry

IP

Immunoprecipitation

KRAS

Kirsten rat sarcoma viral oncogene homolog

LAMP1

Lysosome‐associated membrane protein 1

LDH

Lactate Dehydrogenase

LUAD

Lung adenocarcinoma

m⁶A

N6‐methyladenosine

MeRIP

Methylated RNA Immunoprecipitation

MOI

Multiplicity of infection

MHC

Major Histocompatibility Complex

PARP

Poly (ADP‐ribose) polymerase

PBS

Phosphate‐buffered saline

PBST

Phosphate‐buffered saline with tween

PD‐1

Programmed Cell Death 1

PD‐L1

Programmed Cell Death 1 Ligand 1

PD‐L1‐FL

Full‐length PD‐L1

PD‐L1‐ΔICD

EHD1 variants lacking the Intracellular domain

PE

Phycoerythrin

RT‐PCR

Real‐time Polymerase Chain Reaction

RAB11

Ras‐related protein Rab‐11

RAB11‐FIP2

RAB11 family interacting protein 2

RF

Random forest algorithm

RIP

RNA Immunoprecipitation

RIPA

Radioimmunoprecipitation assay

shCtrl

short hairpin RNA control

SDS‐PAGE

Sodium dodecyl sulfate polyacrylamide gel electrophoresis

shEHD1

short hairpin RNA for EHD1

shRNA

small hairpin RNA

shYTHDF1

short hairpin RNA for YTHDF1

STR

Short tandem repeat

STK11

Serine/threonine kinase 11

SVM

Support vector machine algorithm

TCGA

The Cancer Genome Atlas

TIDE

Tumor Immune Dysfunction and Exclusion

Total‐Ub

Total ubiquitin

UbK48

Ubiquitination at the lysine 48 residue

UbK63

Ubiquitination at the lysine 63 residue

WB

Western blotting

YTH

YT521‐B homology

YTHDF1

YTH N6‐Methyladenosine RNA Binding Protein F1

YTHDF1^MUT

mutant YTHDF1

YTHDF1^WT

wild‐type YTHDF1

1. BACKGROUND

Lung adenocarcinoma (LUAD) is one of the most dismal malignancies worldwide, accounting for substantial morbidity and contributing significantly to cancer‐related mortality rates [1, 2]. Immune evasion is increasingly acknowledged as a critical mechanism in tumorigenesis and progression of a variety of cancers, including LUAD [3]. The immune evasion mechanisms employed by human cancers are complex and multifaceted [4]. A prevalent mechanism of immune evasion involves the alteration of antigen expression by tumor cells, thereby reducing their visibility to therapeutic immune cells within the immune system [5]. Moreover, tumors may amplify the production of immunosuppressive molecules, including immune checkpoint proteins [e.g., programmed death‐ligand 1 (PD‐L1)], which act to suppress immune cell activation and thereby dampen the host's anti‐tumor immune response [6]. Other immune escape mechanisms comprise the secretion of immunosuppressive cytokines, the induction of T cell anergy and exhaustion, and so on [7, 8, 9, 10].

Reinvigorating antitumor immunity is a promising cancer treatment approach [11]. Therapeutic modalities such as immune checkpoint blockade (ICB), chimeric antigen receptor T‐cell, and cancer vaccines therapy are at the vanguard of a therapeutic revolution in the cancer immunotherapy [12, 13]. ICB therapies, exemplified by monoclonal antibodies targeting programmed cell death protein 1 (PD‐1), PD‐L1, and cytotoxic T‐lymphocyte antigen 4 (CTLA4), have been shown to extend survival in patients with diverse cancers [14]. Regrettably, only approximately 20% of unselected patients with advanced LUAD respond to ICB therapy [15]. Hence, it is critical to identify tumor‐intrinsic targets that modulate immune evasion and influence immunotherapy responsiveness to ICB therapy.

Endocytosis of receptors and cellular membranes represents an imperative mechanism in mammalian cellular physiology, essential for diverse intracellular processes, such as the uptake of nutrients and the modulation of cell surface receptor expression [16, 17]. Specific internalized receptors are destined for lysosomal degradation; in contrast, a discrete cohort is redirected back to the plasma membrane, where these proteins are capable of engaging in additional rounds of endocytosis. This reiterative mechanism is termed endocytic recycling [18, 19]. The mammalian eps15 homology domain (EHD) protein family, comprising EHD1‐EHD4, has been implicated in specific steps of the endocytic transport pathway [20, 21]. Among them, EHD1 has been the most investigated and shown to fulfill pivotal functions at multiple junctures in the endocytic recycling pathway [22, 23, 24, 25]. These functions include trafficking plasma membrane components from early endosomes (EEs, also called sorting endosomes) to the endocytic recycling compartment (ERC), their subsequent return to the plasma membrane, retrograde transport of cargo from EEs to the Golgi apparatus, and vesicular scission events originating from endosomes [25, 26, 27]. The interaction of EHD1 with the Ras‐related protein Rab‐11 (RAB11) effector, RAB11 family interacting protein 2 (RAB11‐FIP2), regulates the translocation of RAB11‐enriched vesicles from the ERC to the plasma membrane [25]. EHD1 plays a crucial role in influencing the signaling outputs of receptors. It modulates the trafficking of various endocytosed receptors, such as cluster of differentiation 44 [28], β1 integrin [29], the insulin‐like growth factor 1 receptor [23], and epidermal growth factor receptor [26]. In the field of oncology, EHD1 [30] can predict the prognosis of patients with various tumors. Moreover, EHD1 has been implicated in tumor progression in multiple cancers, influencing cancer cell proliferation [31, 32], migration, and invasion [28, 32, 33, 34]. However, the role of EHD proteins in tumor immune evasion and the response to immunotherapy remains largely unclear.

This study aimed to identify EHD proteins that can predict poor immunotherapy responses for LUAD patients, and to elucidate the functions and functional mechanisms of these identified EHD proteins in the immune evasion of LUAD. Based on computational predictions, we hypothesized that EHD1 plays a pivotal role as a regulator of immune escape in LUAD. Then, functional assays were performed to examine the effects of EHD1 on T cell‐mediated cytotoxicity. Mechanistic validation was used to explore the interaction between EHD1 and PD‐L1, EHD1's role in PD‐L1 recycling and the interaction between EHD1 mRNA and the N6‐methyladenosine (m⁶A) reader protein YTH N6‐methyladenosine RNA binding protein 1 (YTHDF1). in vivo xenograft models were established to confirm that EHD1 silencing affects tumor immune evasion and response to anti‐PD‐1 therapy.

2. MATERIALS AND METHODS

2.1. Bioinformatics analysis

The transcriptome RNA expression matrix and survival information of LUAD patients were downloaded from The Cancer Genome Atlas (TCGA, https://portal.gdc.cancer.gov/, project ID: TCGA‐LUAD). Survival analysis was plotted using Kaplan‐Meier curves based on the “survminer” R package (https://CRAN.R‐project.org/package = survminer). The immunotherapy responses were forecasted by tumor immune dysfunction and exclusion (TIDE) (http://tide.dfci.harvard.edu/) analysis [35]. TIDE directly models the functional status of anti‐tumor immunity by integrating gene expression signatures related to T cell dysfunction and exclusion. The dysfunction score is calculated as $\mathrm{Dysfunction}\mathrm{Score}={\sum }_{i=1}^n{s}_i(g_i)+\epsilon$. For poorly infiltrated tumors, the exclusion score is calculated as $ExclusionScore={\sum }_{j=1}^m{w}_j·g_j+b$. The final TIDE score is calculated as $TIDEScore=max(DysfunctionScore,ExclusionScore)$.

The best cut‐off values of the expression levels of EHD1 to EHD4 were determined by surv_cutpoint function in “survminer” R package.

EHD1‐PD‐L1 docking model was constructed using the DockCompound tool (https://www.home‐for‐researchers.com) for molecular docking based on the infrastructures of EHD1 (Protein Data Bank ID: 2JQ6) and PD‐L1 (Protein Data Bank ID: 7XYQ).

To predict EHD1 upstream m⁶A modification components, we used the m⁶A WER Target Gene Database (RM2target, http://rm2target.canceromics.org/), the RNA Modification Base (RMBase, https://rna.sysu.edu.cn/rmbase/), and the Database of functional variants involved in RNA modifications (RMVar, http://rmvar.renlab.org/). The resulting data were intersected using Venn diagrams. We used the Motif‐based sequence analysis tool (MEME Suite, https://meme‐suite.org/meme/) to predict the site where m⁶A might be modified in EHD1 mRNA. RPISeq (http://pridb.gdcb.iastate.edu/RPISeq/) was used to predict the possibility of protein and mRNA binding. The correlation of gene‐gene based on TCGA mRNA expression was evaluated by the GraphPad Prism 8.0.2 software (GraphPad Software, San Diego, CA, USA). Spearman correlation analysis was then conducted to identify epigenetic regulators associated with EHD1.

2.2. Cell lines and clinical tissue samples

The human LUAD cell lines A549, H1299, the human Jurkat T cell line, and the murine LUAD cell line LA795 were obtained from the American Type Culture Collection (Manassas, VA, USA). All cell lines were authenticated using short tandem repeat (STR) genotyping before use. These cells were cultured at 37°C in 5% CO₂ in a complete medium with RPMI‐1640 (Gibco, New York, NY, USA) and 10% fetal bovine serum (FBS, Gemini, West Sacramento, CA, USA).

Paraffin‐embedded tumor tissues were obtained from 100 LUAD patients who had undergone surgery in the Affiliated Cancer Hospital of Harbin Medical University (Harbin, Heilongjiang, China) and treated with ICB therapies. The Ethics Committee of the hospital approved this study (KY2023‐08). All included patients have signed informed consent to allow the use of their medical record data and biological specimens for all medical research purposes.

2.3. Construction of the plasmid and establishment of stable knockdown cell lines

Lentiviral‐based small hairpin RNAs (shRNA) targeting EHD1 and YTHDF1 were designed and obtained from GeneChem (Shanghai, China). A non‐targeting shRNA was included as a negative control to ensure the specificity of the results. A549 and H1299 cells (multiplicity of infection [MOI]  =  10) were harvested at 48 h post‐infection. HitransG P (REVG005, GeneChem) was used to enhance infection efficiency. Puromycin (1 µg/mL, 1299MG025, Biofroxx, Eibelstadt, Bavaria, Germany) was used to select stable cell clones for 1 week. Green fluorescent protein sorting was used to isolate successfully transfected cells. The shRNA sequences are displayed in Supplementary Table S1.

Recombinant plasmids were synthesized by General Biol (Chuzhou, Anhui, China). Transient transfection was performed using jetPRIME® Versatile DNA/siRNA transfection reagent (101000046, Polyplus, Strasbourg, France). The full‐length, truncated, and mutated sequences of EHD1 were cloned into the pcDNA3.1‐Flag vector at the appropriate cloning sites to construct the following expression plasmids: full‐length EHD1 (EHD1‐FL, amino acids 1‐534), C‐terminal EH domain deletion mutant (EHD1‐∆EH, EHD1^MUT1, amino acids 1‐439), and an EHD1 mutant with deletions at the m⁶A peak sites (EHD1 ^MUT2). To establish a gain‐of‐function model, we transfected the shEHD1#1 clone (shEHD1) with EHD1‐FL, generating a cell clone designated as EHD1 ^OE, which exhibited overexpression of EHD1. Similarly, we transfected shEHD1 with EHD1‐∆EH, resulting in the creation of a cell clone EHD1^MUT1, which displayed overexpression of EHD1 lacking the EH domain. Lastly, we transfected with EHD1^MUT2, generating a cell clone EHD1^MUT2, which exhibited overexpression of EHD1 with mutations at the m⁶A peak sites. The overexpression of EHD1 (wild‐type and mutants) was confirmed using Western blotting and qRT‐PCR.

Moreover, truncated sequences of PD‐L1 were inserted into the pcDNA3.1‐Myc vector at the designated cloning sites to construct the following expression plasmids: full‐length PD‐L1 (PD‐L1‐FL, amino acids 1‐290) and a C‐terminal intracellular domain deletion mutant (PD‐L1‐ΔICD, amino acids 1‐260). The wild‐type full‐length YTHDF1 and the full‐length mutant form with two critical amino acid mutations (K395A and Y397A), which has been reported to bind to mRNA through the m⁶A recognition pockets in the YTH domain [36], were cloned into the pcDNA3.1‐HA vector at the designated cloning sites to construct the YTHDF1 full‐length (YTHDF1^WT) and YTHDF1 mutant (YTHDF1^MUT) expression plasmids.

2.4. Western blotting

LUAD cells were lysed using radioimmunoprecipitation assay (RIPA) lysis buffer (AR0102, Boster, Wuhan, Hubei, China) to extract proteins, with 1 mmol/L phenylmethylsulfonyl fluoride (ST507, Beyotime, Shanghai, China), a phosphatase inhibitor.

Protein concentration was determined using the bicinchoninic acid (BCA) assay with the Pierce BCA Protein Assay Kit (23225, Thermo Fisher Scientific, Waltham, MA, USA) to ensure equal loading of samples. For sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS‐PAGE), 8%, 10%, or 12% gels (E302‐01, E303‐01, E304‐01, Vazyme, Nanjing, Jiangsu, China) were used, and 30 µg of protein was loaded per lane. After electrophoresis, the proteins were transferred to a polyvinylidene difluoride membrane (03010040001, Roche, Basel, Switzerland). The membrane was blocked with 5% non‐fat milk in PBS‐T for 1 h at room temperature. The primary antibodies were diluted at the recommended concentration and incubated overnight at 4°C. The secondary antibody was diluted at 1:10,000 and incubated for 1 h at room temperature. Signal detection was performed using an ultra‐sensitive chemiluminescent substrate (M2301, HaiGene, Shanghai, China), and the signals were visualized using the Tanon 5200 Multi imaging system (Tanon, Shanghai, China). The antibodies are displayed in Supplementary Table S2.

2.5. RNA extraction and quantitative real‐time PCR (qRT‐PCR)

Total RNA was extracted from LUAD cells using Trizol reagent (15596026CN, Thermo Fisher Scientific). Total RNA (1 µg) was reversely transcribed into cDNA using the FastKing kit (KR118‐02, TIANGEN, Beijing, China). The cDNA products were amplified by StepOne real‐time fluorescence quantitative PCR instrument (Applied Biosystems, Foster, CA, USA), and Talent qPCR Premix (SYBR Green) (FP209‐02, TIANGEN) was used to configure the reaction system (initial degeneration at 95°C for 3 min, PCR reaction lasted 40 cycles: degeneration 95°C for 5 s and extension 60°C for 15 s). Gene expression levels were normalized to glyceraldehyde 3‐phosphate dehydrogenase (GAPDH) as the internal reference gene. Gene expression levels were calculated using the ΔΔCt method. The total volume of the reaction system was 20 µL/well, with a primer concentration of 10 pmol/µL. The primer sequences are displayed in Supplementary Table S3.

2.6. T cell‐mediated tumor cell killing experiments

Jurkat T cells were treated with ImmunoCult Human CD3/CD28/CD2 T cell activator (10970, STEMCELL Technologies, Vancouver, British Columbia, Canada) for 24 h. Then, T cells were activated and collected [37]. Before co‐culturing with T cells, LUAD cells (1 × 10⁶ cells) were incubated with 1 µL 0.5 mmol/L carboxyfluorescein diacetate succinimidyl ester (CFSE, 65‐0850‐84, Invitrogen, Eugene, OR, USA) for 10 min at 37°C in a 5% CO₂ humidified atmosphere for labeling and were used as target cells. Activated T cells (effectors) and LUAD cells (targets) were co‐cultured for 24 h at effector‐to‐target (E/T) ratios of 1:1, 5:1, and 10:1 to evaluate T cell‐mediated cytotoxicity. The LUAD cell viability was evaluated based on the absorbance at 570 nm (A570) using the Cell Counting Kit‐8 (CCK‐8) kit (SC119, Seven, Beijing, China).

T cell‐mediated cytotoxicity was evaluated using a plate‐based colorimetric assay. LUAD cells were plated in 6‐well plates and cultured overnight to ∼80% confluence, followed by co‐culture with activated T cells at an E/T ratio of 5:1 for 48 h. The surviving LUAD cells adhering to the plate were washed with phosphate‐buffered saline (PBS, SW132, Seven), fixed with 4% paraformaldehyde (BL539A, Biosharp, Hefei, Anhui, China), and stained with 0.5% crystal violet (C8470, Solarbio, Beijing, China). The A570 of plates was measured using an iMark Microplate Reader (Bio‐Rad, Hercules, CA, USA). The liquid containing T cells and cell fragments was harvested. We used the Human IFN‐γ ELISA kit (EH008, ExCell Bio, Suzhou, Jiangsu, China) to detect the IFN‐γ secreting level by activated T cells in the liquid according to the manufacturer's instruction.

LUAD cells were plated and incubated in 96‐well plates overnight. Activated T cells were added at an E/T ratio of 5:1 for 48 h. The Lactate Dehydrogenase (LDH) Assay Kit (C0016, Beyotime) was used to assess T cell cytotoxicity by quantifying the release of endogenous LDH. We measured the release of LDH in the co‐culture system of T cells and LUAD cells and then calculated the cell death rate using the formula: Cell death (%) = (experimental LDH ‐ spontaneous LDH) / (maximum LDH ‐ spontaneous LDH) × 100%.

2.7. Apoptotic assay

After T cell‐mediated tumor cell killing experiments, adherent LUAD cells in the wells were washed, and trypsin solution (C0205, Beyotime) was applied to digest adherent cells. Cell counting was then performed, and the total reaction volume was adjusted to achieve a final concentration of 1×10⁶ cells/mL. Apoptotic cells were stained using the Annexin V‐FITC/PI apoptosis kit (FXP018‐100, 4A BIOTECH, Beijing, China) and determined by flow cytometry (FCM).

2.8. Immunofluorescence (IF)

Cells were seeded into confocal dishes at a density of approximately 1 × 10⁴ cells/cm² for confocal microscope and 3 × 10⁴ cells/cm² for fluorescence microscope and fixed with 4% paraformaldehyde. After permeabilization with 0.5% Triton X‐100 (BS084, Biosharp) for 30 min, cells were incubated with primary antibodies at 4°C overnight, followed by incubation with secondary antibodies at room temperature for 1 h. The nuclei were stained with 4',6‐diamidino‐2‐phenylindole (AR1176, Boster). Photos were captured using a laser confocal microscope (Olympus, Tokyo, Japan). Antibodies used in the experiment are listed in Supplementary Table S2.

2.9. Immunoprecipitation (IP)

For each IP reaction, 400 µL of total protein lysate was used. Protein A/G magnetic beads (HY‐K0202, MedChemExpress, Princeton, NJ, USA) were incubated with specific antibodies or isotype control IgG (5 µg/mL, 66360‐1‐Ig, Proteintech, Wuhan, Hubei, China), then mixed with whole cell lysates at 4°C for 2 h. Endogenous IP assays were performed by lysing untransfected cells with a specific lysis buffer. The lysates were then centrifuged at 14,000 rpm for 15 min at 4°C to eliminate cellular debris. The supernatant was incubated with antibodies, including IgG, anti‐EHD1 (24657‐1‐AP, Proteintech) and anti‐PD‐L1 (66248‐1‐Ig, Proteintech), targeting the endogenous protein at 4°C. Protein A/G beads were added, and the complex was pulled down and washed with phosphate‐buffered saline with tween (PBST), then analyzed by Western blotting. For exogenous IP assays, cells were transfected with plasmids encoding the overexpressed protein tagged with FLAG or Myc. After incubation, cells were lysed, and the lysate was incubated with antibodies, including IgG, anti‐Flag (20543‐1‐AP, Proteintech) and anti‐Myc (16286‐1‐AP, Proteintech), specific to the tag at 4°C. Protein A/G beads were added, and the complex was pulled down, washed with PBST, eluted, and analyzed by Western blotting.

2.10. Endocytic recycling assays

Detection of PD‐L1 on the cell surface was performed as previously described [38]. LUAD cells were harvested and washed with PBS to remove any residual media. LUAD cells were incubated with a fluorescently labeled anti‐PD‐L1 antibody (5 µL / 1 × 10⁶ cells in 100 µL PBS, PE‐65073, Proteintech) at 4°C for 1.5 h in the dark. After incubation, the cells were washed with 500 µL cold PBS for three times to remove any unbound antibodies. The stained cells were resuspended in 100 µL PBS and immediately analyzed using a flow cytometer to detect the fluorescence emitted by the antibody, which was proportional to the amount of PD‐L1 present on the cell surface. Internalization assay of PD‐L1 was conducted following the previously outlined procedure [39, 40]. Cell surface PD‐L1 was labeled with phycoerythrin (PE)‐conjugated anti‐PD‐L1 antibody at 4°C for 1.5 h (5 µL / 1 × 10⁶ cells in 100 µL 0.2% bovine serum albumin). After washing with 500 µL PBS for three times to remove unbound antibodies, LUAD cells were re‐incubated in the serum‐free medium for 0, 2, and 4 h at 37°C. At the indicated times, cells were fixed with 4% paraformaldehyde for 20 min and kept on ice for 5 min, washed with cold PBS to prevent further endocytosis of surface antibodies, and detected by FCM. The recycling assay of PD‐L1 was carried out as described earlier [39, 40]. The surviving LUAD cells were incubated with the anti‐PD‐L1 antibody (2 µg/mL, 28076‐1‐AP, Proteintech) for 1 h on ice. After washing three times with 500 µL PBS, LUAD cells were re‐incubated in a prewarmed medium and incubated for 0, 2, and 4 h at 37°C to allow antibody‐labeled PD‐L1 to undergo recycling or degradation. After went through 2 rounds of resuspension in 500 µL formulated low pH buffer (pH 2.7, 0.5% acetic acid, A116166, Aladdin, Shanghai, China) for 2 min on ice to strip remaining surface bound antibody, cells were lysed with RIPA lysis buffer on ice. IP assays were used to capture antibody‐combined PD‐L1. Proteins were eluted in the SDS loading buffer (P0015A, Beyotime), denatured, and detected by Western blotting.

2.11. Protein degradation assay

For the cycloheximide (CHX) chase assay, cells were treated with 10 µg/mL CHX (HY‐12320, MedChemExpress) for 0, 1, 2, 3, and 4 h. For the protein degradation pathway assay, cells were treated with the proteasome inhibitor MG132 (10 µmol/L, HY‐13259, MedChemExpress) or lysosomal inhibitor NH₄Cl (20 mmol/L, 12125‐02‐9, Aladdin) for 8 h. Then cells were collected and lysed on ice, subsequently verified by Western blotting.

For the ubiquitination (Ub) assay, LUAD cells were transfected with UbK63‐, UbK48‐, or total‐Ub‐expressing plasmids. The cells were cultured in the presence of MG132 to increase the level of ubiquitinated PD‐L1. IP assays were used to capture ubiquitinated PD‐L1, which was then immunoprecipitated by anti‐PD‐L1 antibody.

2.12. Methylated RNA Immunoprecipitation (MeRIP)

The m⁶A MeRIP assay was performed using RiboMeRIP^TM m⁶A transcriptome analysis kit (C11051‐1, Ribobio, Guangzhou, Guangdong, China). A total of 100 µg of RNA was extracted and purified by Trizol reagent. One‐tenth of the RNA was separated as an input group and stored at ‐80°C. For the IP step, RNA was mixed with Protein A/G magnetic beads (25 µL) of anti‐m⁶A (5 µg) or isotype control IgG (5 µg). The mixture was incubated at 4°C for 2 h with gentle rotation. The beads were washed three times with the 1× IP Buffer, which was prepared by adding 3.2 mL of enzyme‐free water to 800 µL of 5× IP buffer. After incubation, the beads were magnetically separated, and the bound m⁶A‐modified RNA was eluted using RNA fragmentation buffer. The eluted RNA was then processed and purified using the Hipure Serum/Plasma miRNA Kit (R431402, Magen Biotech, Cambridge, MA, USA). The quality and quantity of the m⁶A‐modified RNA were measured using qPCR. MeRIP‐qPCR primers for EHD1 were predicted by SRAMP (http://www.cuilab.cn/sramp). The primer sequences are displayed in Supplementary Table S3.

2.13. RIP

A RIP assay was performed using a RIP Kit (Bes5101, BersinBio, Guangzhou, Guangdong, China) according to the manufacturer's instructions, with a YTHDF1‐specific antibody (66745‐1‐Ig, Proteintech) to capture RNA‐protein complexes. Cells were harvested and lysed in an appropriate polysome lysis buffer, protease inhibitor and RNase inhibitor for 10 min on ice, followed by DNase treatment to remove genomic DNA and incubation at 37°C for 10 min to ensure complete DNA degradation. After the removal of DNase, the samples were divided into IP, IgG, and Input groups in the ratio 2: 2: 1, and the Input group was stored at ‐80°C. The IP and IgG groups were incubated with 4 µg of the corresponding antibody (YTHDF1 or isotype control IgG) at 4°C overnight, with gentle rotation, followed by addition of Protein A/G magnetic beads to both groups, and incubated at 55°C for 1 h to capture the antibody‐RNA complexes, which were then magnetically separated. The supernatants were collected, and added with 200 µL of the phenol (108‐95‐2, Merck KGaA, Darmstadt, HE, Germany) ‐ chloroform (67‐66‐3, Merck KGaA) ‐ isopentanol (540‐84‐1, Fuyu Chemical, Tianjin, China) mixed solution at the ratio of 25:24:1, and clarified by centrifugation at 13,000 g, 4°C for 10 min, then the upper water phase was collected. The immunoprecipitated RNA was isolated by precipitation using a mixture (1 µL glycogen, 10 µL sodium acetate and 500 µL 100% ethanol), followed by incubation at ‐80°C overnight. RNA from both the Input and IP groups was isolated, quantified, and analyzed by qPCR to detect specific RNAs immunoprecipitated by YTHDF1, with quantification compared between the IP, IgG, and Input groups to assess the enrichment of specific RNAs.

2.14. Agarose electrophoresis

The agarose gel was prepared with a 1% agarose (BS081‐100g, Biosharp), with the concentration determined by the size of the DNA fragments to be separated. The electrophoresis buffer used was 1× Tris‐acetate‐EDTA buffer (BL533A, Biosharp). The DL500 DNA marker (2 µL, 3590A, Takara, Dalian, Liaoning, China) was loaded in the appropriate lane(s) for size reference. The reaction product (10 µL) from qPCR and 10× DNA Loading Buffer (9157, Takara, Tokyo, Japan) were mixed and loaded into the gel wells, with 5 µL DL500 DNA Marker (3590A, Takara). Electrophoresis was conducted under constant voltage of 120 V for 45 min to allow proper separation of DNA fragments. After electrophoresis, the gel was stained with GelRed (41003, Biotium, Fremont, CA, USA) and visualized under ultraviolet light.

2.15. RNA stability assay

LUAD cells were counted and seeded in 6‐well plates, with 2 × 10⁵ cells per well, to achieve 50% confluence after 24 h of incubation. Prior to treatment, the cells were pre‐incubated for 6 h under low standard growth conditions (medium without serum) for starving cells. Following pre‐incubation, the cells were treated with 10 µg/mL of actinomycin D (Act D, HY‐17559, MedChemExpress) for 0, 3, and 6 h. At the specified time points, the total RNA was extracted from the cells, and the mRNA abundance of the target gene was quantified using RT‐PCR. Equal volumes of cDNA were selected for reverse transcription, and the mRNA abundance of 0 h served as the internal reference, and the data were analyzed using the 2^−ΔΔCT method.

2.16. Immunohistochemistry (IHC) and Hematoxylin and eosin (HE) staining assay

For the IHC assay, tumor tissue sections from LUAD patients or mouse xenograft tumors were treated with hydrogen peroxide for antigen retrieval, then incubated with primary antibodies at 4°C overnight, and finally with secondary antibodies. Diaminobenzidine (ZLI‐9018, ZhongshanJinqiao, Beijing, China) was used to visualize target proteins. Pictures were captured using an inverted microscope (Leica, Wetzlar, HE, Germany). Two experienced pathologists, blinded to the patient's clinical information, independently double‐evaluated the staining intensity. Any discrepancies were resolved through consensus. The percentage of microscopically positive cells was scored as follows: 0 point for 0%‐25%, 1 point for 26%‐50%, 2 points for 51%‐75%, and 3 points for 76%‐100%. In addition, a score of 0‐1 is called the low expression group, and a score of 2‐3 is called the high expression group. The antibodies used are listed in Supplementary Table S2.

For the HE staining assay, deparaffinization and rehydration of mouse tumor sections were performed using xylene and ethanol, respectively. The sections were stained with hematoxylin (AR1180‐1, Boster) for the nucleus for 4 min and eosin (AR1180‐2, Boster) for the cytoplasm for 1 min at room temperature. Then, the sections were rinsed with flowing water for 10 min. After staining, the sections were photographed under a microscope. Images were then subjected to cell morphology analysis.

2.17. Animal experiments

Six‐week‐old female C57BL/6 mice were used to construct a transplanted tumor model. Mice were housed in clean, well‐ventilated cages under specific pathogen‐free conditions with appropriate temperature and humidity, and humane endpoints included signs of severe distress or significant health deterioration. Euthanasia methods involved the use of injectable anesthetics, followed by cervical dislocation. A total of 1 × 10⁶ LA795 (shCtrl and shEHD1) cells were injected into different axillae of the same mouse, with shCtrl cells injected into one side and shEHD1 cells injected into the opposite side of each mouse. When the tumors reached approximately 100 mm³, the mice were intraperitoneally injected with 100 µg of mouse anti‐PD‐1 (RMP1‐14, MedChemExpress) or anti‐IgG (BE0089, BioXcell, Lebanon, NH, USA) per mouse in 100 µL of PBS on Days 8, 11, and 14. Tumor sizes were measured by vernier caliper every 2 days. Bioluminescence imaging was performed on Day 18. Then, the mice were euthanized, and the tumors were removed. Tumor volumes were estimated using the formula: ½ × (length × width²). All animal experiments were approved by the Harbin Medical University Animal Protection and Use Committee (KY2023‐08).

2.18. Statistical analysis

In this study, graphs were statistically drawn and analyzed by GraphPad Prism 8.0.2 software. Differences between two groups for normally distributed continuous data were tested by the Student's t‐test, and one‐way analysis of variance was employed for comparisons involving more than two groups, with P < 0.05 considered statistically significant. All experiments were performed with three independent biological replicates to ensure the reproducibility and reliability of the results.

3. RESULTS

3.1. EHD1 was identified and validated to reinforce the immune evasion of LUAD cells

We used the computational framework, TIDE, to identify EHD proteins that modulate tumor immune evasion. TIDE assessment indicated that elevated levels of EHD1, EHD2, and EHD4 were significantly related to high TIDE scores (Figure 1A), which suggested an increased risk of immune evasion. Tumor immune escape plays a pivotal role in the progression of cancer and influences the prognosis of patients with LUAD. Further, Kaplan‐Meier survival analyses revealed that elevated expression of EHD1 was a consistent indicator of short OS in patients with LUAD, in contrast to EHD2 to EHD4, which did not exhibit the predictive value (Figure 1B). These results prompted us to choose EHD1 as a candidate gene to further investigate its role in tumor immune evasion.

FIGURE 1.

EHD1 knockdown in LUAD cells increased sensitivity to T‐cell‐induced cytotoxicity. (A) Violin plots depicting the distribution of TIDE scores among high‐ versus low‐ expression groups of individual members within the EHD gene family. (B) Kaplan‐Meier survival curves, using the TCGA‐LUAD dataset, illustrating the OS stratification categorized based on the level of gene expression in the EHD gene family. (C‐D) Comparative analysis of the mRNA (C) and protein (D) levels of EHD1 in the untreated group and A549 cells transfected with shCtrl, shEHD1#1, or shEHD1#2. (E) Flowchart illustrating the process of co‐culturing activated T cells with LUAD cells. T cells were activated after 24 h of cytokine treatment and then co‐cultured with LUAD cells for an additional 24 h. (F) CCK‐8 assay results showing the influence of EHD1 knockdown on the killing of A549 cells by T cells. Activated T cells and LUAD cells were co‐cultured at E/T ratios of 1:1, 5:1, or 10:1. (G) Crystal violet‐stained images of surviving LUAD cells treated with activated T cells across the designated groups of cells. Representative pictures (left panel) and the quantification chart (right panel) are displayed. (H) Evaluation of T‐cell‐mediated apoptosis in the designated LUAD cells using FCM. The right bar graph summarizes the statistical evaluation of these results. (I) Western blotting analysis revealed the influence of EHD1 knockdown on the protein levels of apoptosis‐related markers in T‐cell‐treated LUAD cells. (J) Histogram showing the influence of EHD1 knockdown on the T‐cell‐mediated cell death rate of A549 cells. The E/T ratio is 5:1. (K) Histograms showing the levels of LDH released by LUAD cells alone or following treatment with T cells. (L) Histograms quantifying the levels of IFN‐γ secreted by T cells co‐cultured with LUAD cells. Data were collected from three independent experiments. Two‐tailed unpaired Student's t test; The data are shown as the mean ± SD. ns, not significant; * P < 0.05, ** P < 0.01, *** P < 0.001. Bax, Bcl‐2 Associated X; Bcl‐2, B‐cell lymphoma‐2; CCK‐8, Cell counting Kit‐8; CD2, Cluster of Differentiation 2; CD28, Cluster of Differentiation 28; CD3, Cluster of Differentiation 3; E/T, Effector‐to‐target ratio; EHD, Eps15 homology domain; GAPDH, Glyceraldehyde‐3‐phosphate dehydrogenase; IFN‐γ, Interferon γ; LDH, Lactate Dehydrogenase; LUAD, Lung adenocarcinoma; PI, Propidium iodine; OS, Overall survival; TIDE, Tumor Immune Dysfunction and Exclusion; FCM, Flow cytometry; shCtrl, short hairpin RNA control; shEHD1, short hairpin RNA for EHD1; SD, Standard deviation.

According to a previous study, A549 and H1299 cells were used to establish a loss‐of‐function model [28]. We successfully constructed EHD1‐deficient clones (shEHD1#1 and shEHD1#2) using lentiviral constructs, resulting in downregulated EHD1 mRNA and protein levels in LUAD cells, as evidenced by comparisons with shRNA control clones (shCtrl; Figure 1C‐D, Supplementary Figure S1A‐B). T cell‐mediated tumor cell killing assay was conducted to ascertain the influence of EHD1 on tumoral immune evasion (Figure 1E). Activated human T cells were co‐cultured with human LUAD cells at various E/T ratios. CCK‐8 assay showed that, compared with shCtrl, EHD1 knockdown sensitized LUAD cells to T cell‐mediated killing at E/T ratios of 5:1 and 10:1 (Figure 1F, Supplementary Figure S1C). The E/T ratio of 5:1 was accordingly selected for subsequent experiments. Compared to the untreated cells without co‐culture, the untreated cells co‐cultured with activated Jurkat T cells showed a decrease in cancer cell survival (Figure 1G, Supplementary Figure S1D), indicating that activated Jurkat T cells have tumor‐killing ability. The plate‐based colorimetric assay further revealed a reduction in the number of viable LUAD cells with EHD1 knockdown after being co‐cultured with activated human T cells (Figure 1G, Supplementary Figure S1D). Then, FCM analysis and Western blotting were performed to evaluate the effects of EHD1 depletion on T cell‐mediated apoptosis. EHD1 silencing noticeably increased the proportion of apoptotic LUAD cells after T‐cell treatment, which coincided with a decreased expression of the anti‐apoptotic protein B‐cell lymphoma‐2 (Bcl‐2), an increased expression of the pro‐apoptotic protein Bax, and elevated levels of cleaved Caspase‐3 and cleaved poly(ADP‐ribose) polymerase (PARP) (Figure 1H‐I, Supplementary Figure S1E‐G). LDH assays revealed that the knockdown of EHD1 increased the T‐cell‐mediated cell death of LUAD cells (Figure 1J, Supplementary Figure S2A). EHD1 depletion enhanced the release of LDH, which indicates LUAD cell damage (Figure 1K, Supplementary Figure S2B). Furthermore, our observations indicated an elevation in the secretion of interferon‐gamma (IFN‐γ) by activated T cells when co‐cultured with LUAD cells following the knockdown of EHD1 (Figure 1L, Supplementary Figure S2C). Our data suggest that silencing EHD1 in LUAD cells enhances their susceptibility to T cell‐mediated cytotoxicity and promotes the cytotoxic function of T cells.

The EHD1 ^OE cell clone was characterized by EHD1 overexpression, which was validated by qRT‐PCR and Western blotting analyses (Supplementary Figure S3A‐B). Compared with the control vector, T‐cell‐mediated tumor cell‐killing assays revealed that EHD1 overexpression in LUAD cells increased the number of surviving LUAD cells (Supplementary Figure S3C). FCM revealed that EHD1 overexpression inhibited T‐cell‐induced apoptosis (Supplementary Figure S3D). In line with our expectations, EHD1 overexpression impaired the cell‐killing ability of T cells and diminished LDH and IFN‐γ release (Supplementary Figure S3E‐G). Collectively, we indicated that EHD1 attenuates the cytotoxic capacity of T cells against LUAD cells and promotes immune evasion of LUAD cells within the tumor microenvironment.

3.2. EHD1 interacted with PD‐L1 and upregulated its expression

The immune checkpoint protein PD‐L1, acting as a receptor on tumor cells, constitutes a pivotal mechanism underlying immune evasion (Figure 2A) [41]. Encouragingly, molecular docking analysis revealed an interaction between EHD1 and PD‐L1 (Figure 2B). Utilizing IF staining, co‐localization assays, and confocal microscopy, we documented the co‐localization of EHD1 and PD‐L1 in the A549 and H1299 cell lines (Figure 2C). Additionally, the interaction between EHD1 and PD‐L1 was confirmed through IP assays (Figure 2D‐E). In LUAD cells concurrently expressing Flag‐tagged EHD1 and Myc‐tagged PD‐L1, reciprocal binding was observed between exogenous EHD1 and PD‐L1 (Figure 2F). The demonstrated physical interaction between EHD1 and PD‐L1 implies a potential regulatory influence of EHD1 on PD‐L1. Western blotting analyses revealed that PD‐L1 protein expression was decreased following ehd1 knockdown (Figure 2G), whereas major histocompatibility complex (MHC)‐I, Fibrinogen‐like protein 1 (FGL1) and cluster of differentiation 155 (CD155) remained constant (Supplementary Figure S4A). The qRT‐PCR and IF analyses revealed that PD‐L1 protein expression was diminished following EHD1 knockdown, whereas PD‐L1 mRNA levels remained constant (Figure 2H‐I), implying that EHD1 functions as a post‐transcriptional modulator of PD‐L1.

FIGURE 2.

EHD1 interacted with PD‐L1 and upregulated its expression. (A) Diagram indicating the mechanism by which LUAD cells escape CD8⁺ T‐cell‐mediated cytotoxicity through PD‐L1. (B) Docking models of the EHD1‐PD‐L1 complex. (C) IF analysis of the colocalization of EHD1 and PD‐L1 in LUAD cells. (D‐E) IP assays validated the molecular interaction between endogenous EHD1 and PD‐L1. (F) IP assays confirmed the interaction between exogenous EHD1 and PD‐L1. (G) Western blotting analysis was used to detect the regulatory influence of EHD1 silencing on PD‐L1 protein expression. (H) Influence of EHD1 silencing on PD‐L1 mRNA levels in LUAD cells, as detected by qPCR. (I) IF analysis showing the PD‐L1 localization and the influence of EHD1‐knockdown on PD‐L1 expression in LUAD cells. Data were collected from three independent experiments. Two‐tailed unpaired Student's t test; The data are shown as the mean ± SD. ns, not significant; *** P < 0.001. CD8, Cluster of Differentiation 8; DAPI, 4',6‐diamidino‐2‐phenylindole; EHD, Eps15 homology domain; LUAD, Lung adenocarcinoma; GAPDH, Glyceraldehyde‐3‐phosphate dehydrogenase; IP,Immunoprecipitation; IF, Immunofluorescence; IgG, Immunoglobulin G; PD‐1, Programmed cell death 1; PD‐L1, Programmed Cell Death 1 Ligand 1; shCtrl, short hairpin RNA control; shEHD1, short hairpin RNA for EHD1; SD, Standard deviation.

3.3. EHD1 promoted the endocytic recycling of PD‐L1, thereby inhibiting its lysosomal degradation

In light of our observations and the well‐established role of EHD1 in facilitating the recycling of internalized receptors during endocytic processes [42], we postulated that EHD1 may also be involved in the endocytic recycling pathway as well as in the lysosomal degradation of the membrane receptor PD‐L1. RAB11, a small guanosine triphosphatase (GTPase) involved in the regulation of recycling endosomes, and lysosome‐associated membrane protein 1 (LAMP1), a marker of lysosomal membranes, are key players in the trafficking and degradation of membrane receptors [18, 43]. Upon EHD1 knockdown, a reduced co‐localization of PD‐L1 with RAB11‐positive vesicles, whereas an enhanced co‐localization of PD‐L1 with LAMP1, was evident (Figure 3A, Supplementary Figure S4B), suggesting that EHD1 downregulation inhibits the endocytic recycling of PD‐L1 and promotes its lysosomal degradation. Using FCM to detect PD‐L1 on the cell surface, we found that, in comparison to shCtrl, shEHD1 reduced surface PD‐L1 expression in LUAD cells (Figure 3B, Supplementary Figure S4C), implying that shEHD1 might weaken receptor‐mediated endocytic recycling. Using internalization assay, we found that the surface remaining PD‐L1, which underwent endocytic recycling and returned to the cell membrane, was decreased after EHD1 knockdown (Figure 3C). Recycling assay of PD‐L1 showed that EHD1 silencing decreased the surface remaining PD‐L1 (Figure 3D), suggesting that surface PD‐L1 is increasingly internalized and degraded. The findings indicate that EHD1 enhances the recycling of PD‐L1 to the cell surface.

FIGURE 3.

EHD1 promoted the endocytic recycling of PD‐L1 and inhibited its lysosomal degradation. (A) Colocalization analysis of PD‐L1 with RAB11 and LAMP1 in shCtrl and shEHD1 LUAD cells. (B) The relative PD‐L1 expression in LUAD cell surface regulated by EHD1 was detected by FCM. Representative FCM histograms (left), and quantification data (right) are displayed. (C‐D) The surface level of remaining PD‐L1 after EHD1 knockdown was detected by (C) FCM analysis and (D) IP experiments at the specified time points. (E) A CHX chase assay was performed to analyze the stability of the PD‐L1 protein in shCtrl and shEHD1 LUAD cells. The cells were treated with CHX (10 µg/mL) for the designated time. (F‐G) Western blotting analysis was performed to measure PD‐L1 expression in shCtrl and shEHD1 LUAD cells treated with the proteasome inhibitor MG132 (F) or the lysosomal inhibitor NH₄Cl (G) for 8 h. (H) Ubiquitination IP assays validated the degradation of PD‐L1 in LUAD cells transfected with the total‐Ub, UbK63 or UbK48 plasmid. NH₄Cl was used for treatment before cell lysis. Data were collected from three independent experiments. Two‐tailed unpaired Student's t test; The data are shown as the mean ± SD. ns, not significant; * P < 0.05, ** P < 0.01, *** P < 0.001. CHX, Cycloheximide; DAPI, 4',6‐diamidino‐2‐phenylindole; EHD, Eps15 homology domain; GAPDH, Glyceraldehyde‐3‐phosphate dehydrogenase; HA‐Ub, Hemagglutinin ubiquitin; IP, Immunoprecipitation; LAMP1, Lysosome‐associated membrane protein 1; LUAD, Lung adenocarcinoma; PD‐L1, Programmed cell death 1 ligand 1; PE, Phycoerythrin; RAB11, Ras‐Related Protein 11; shCtrl, short hairpin RNA control; shEHD1, short hairpin RNA for EHD1; Total‐Ub, Total ubiquitin; UbK48, Ubiquitination at the lysine 48 residue; UbK63, Ubiquitination at the lysine 63 residue; FCM, Flow cytometry.

We conjectured that the diminished re‐localization of PD‐L1 on the surface of EHD1‐knockdown cells could signify an escalation in the lysosomal degradation pathway of PD‐L1. A CHX chase assay demonstrated that EHD1 knockdown substantially reduced the half‐life of the PD‐L1 protein (Figure 3E). Consistent with predictions, the application of a proteasome inhibitor, MG132, failed to restore the EHD1 knockdown‐induced decrease in PD‐L1 protein levels (Figure 3F), whereas the lysosomal inhibitor NH₄Cl successfully restored PD‐L1 expression (Figure 3G). UbK63 is pivotal for promoting lysosomal degradation pathways [44], whereas UbK48 is instrumental in targeting proteins for degradation via the proteasome [45]. The silencing of EHD1 augmented the total ubiquitination and UbK63 of PD‐L1 without impacting the UbK48 in A549 and H1299 cells (Figure 3H), indicating that EHD1 confers protection against lysosomal degradation of PD‐L1, yet does not prevent its proteasomal degradation. The findings indicate that EHD1 enhances the endocytic recycling of PD‐L1, consequently impeding its degradation via lysosomal pathways.

3.4. Disruption of the EHD1/PD‐L1 interaction impaired the regulatory function of EHD1 in tumor immune evasion

Previous studies showed that the EH domain of EHD1 mediates interaction with the intracellular domain of membrane receptors [28, 46]. In the present investigation, we engineered plasmids to express Flag‐tagged full‐length EHD1 (EHD1‐FL, also refer to EHD1^OE) as well as an EHD1 variant lacking the EH domain (EHD1‐ΔEH, also refer to EHD1^MUT1) (Figure 4A). These vectors were then co‐transfected with Myc‐tagged full‐length PD‐L1 (PD‐L1‐FL). IP assays revealed that EHD1‐FL was successfully precipitated by the Myc tag, while EHD1‐ΔEH was not (Figure 4B), indicating that the EH domain is essential for EHD1's interaction with PD‐L1. Similarly, Myc‐tagged PD‐L1‐FL and PD‐L1 variants lacking the intracellular domain (PD‐L1‐ΔICD) were established (Figure 4C) and co‐expressed them with Flag‐tagged EHD1‐FL. We found that the cytoplasmic domain of canonical PD‐L1 was engaged in molecular interaction with the EHD1 protein (Figure 4D). Subsequently, EHD1^OE or EHD1^MUT1 was transfected into shEHD1 cells. EHD1^OE was able to upregulate the protein expression of PD‐L1 in EHD1‐knockdown cells, whereas EHD1^MUT1 could not (Figure 4E). Compared with EHD1 ^OE, EHD1^MUT1 failed to reverse the promoting effect of shEHD1 on T cell‐mediated cytotoxicity against LUAD cells, as evidenced by T cell‐mediated tumor cell killing assay, T cell‐mediated apoptosis assay, LDH assays, IFN‐γ release assays (Figure 4F‐J, Supplementary Figure S5). Then, we down‐regulated PD‐L1 in EHD1 ^OE with siRNA plasmids. We found that PD‐L1 knockdown reversed the role of EHD1 in modulating T cell‐mediated cytotoxicity against LUAD cells, as evidenced by T cell‐mediated tumor cell killing assay, T cell‐mediated apoptosis assay, LDH assays, IFN‐γ release assays (Supplementary Figures S6‐S7). These results underscore the crucial role of PD‐L1 in EHD1‐mediated tumor immune evasion.

FIGURE 4.

Disruption of the EHD1/PD‐L1 interaction hindered the regulatory role of EHD1 in tumor immune evasion. (A) Diagrammatic representation of the structure of full‐length EHD1 and the variant without the EH domain constructed with the plasmid. (B) IP analysis showing the ability of Myc‐tagged PD‐L1 to precipitate EHD1 proteins with or without EH domains. (C) Schematic diagrams of full‐length PD‐L1 and variants lacking the ICD domain constructed with plasmids. (D) IP analysis showing the structure of PD‐L1 that physically interacts with EHD1. (E) Western blotting analysis revealed the influence of disrupting the EHD1/PD‐L1 interaction on PD‐L1 expression. An antibody targeting the aa515‐534 region of the EHD1 protein was utilized in this Western blotting analysis. (F) Crystal violet‐stained images of surviving A549 cells treated with activated T cells across the designated groups. (G) Evaluation of T‐cell‐mediated apoptosis in the designated cells by FCM. (H) The influence of disrupting the EHD1/PD‐L1 interaction on the ability of T cells to inhibit LUAD cells was detected in an LDH release assay. (I) Histograms quantifying the levels of endogenous LDH released by LUAD cells alone or treated with T cells. (J) Histograms quantifying the levels of IFN‐γ secreted by T cells coincubated with LUAD cells. Data were collected from three independent experiments. Two‐tailed unpaired Student's t test; The data are shown as the mean ± SD. ns, not significant; * P < 0.05, ** P < 0.01, *** P < 0.001. ECD: Extra cellular domain; EHD: Eps15 homology domain; EHD1‐FL: Full‐length EHD1; EHD1‐ΔEH: EHD1 variants lacking the EH domain; EHD1^MUT1: Mutant EHD1#1, which displayed overexpression of EHD1 lacking the EH domain; EHD1^OE: Overexpression of full‐length EHD1; FCM: Flow cytometry; IFN‐γ: Interferon γ; IP: Immunoprecipitation; LDH: Lactate Dehydrogenase; LUAD: Lung adenocarcinoma; PD‐L1: Programmed cell death 1 ligand 1; PD‐L1‐FL: Full‐length PD‐L1; PD‐L1‐ΔICD: EHD1 variants lacking the intracellular domain; PI: Propidium iodine; shEHD1: short hairpin RNA for EHD1; SP: Signal peptide; TM: Transmembrane domain.

3.5. EHD1 silencing impeded tumor immune evasion and enhanced the efficacy of anti‑PD‑1 therapy

To assess the immunomodulatory impact of EHD1 in vivo, the murine LUAD cell line LA795 was used to establish subcutaneous xenograft models in immunocompetent C57BL/6 mice. Efficient reduction of EHD1 mRNA and protein expression in LA795 cells was attained in EHD1‐depleted clones (Figure 5A‐B). C57BL/6 murine models were inoculated with control and EHD1‐knockdown LA795 cells and subsequently administered with either PD‐1 monoclonal antibody (anti‐PD‐1) or the corresponding IgG isotype control (anti‐IgG) (Figure 5C). The xenograft tumors in the shEHD1 group exhibited significantly slower growth compared to the shCtrl group, with the most pronounced suppression observed in mice receiving combined EHD1 depletion and PD‐1 blockade therapy (Figure 5D). Additionally, the combined therapeutic approach of EHD1 depletion and PD‐1 blockade yielded a substantial diminishment in tumor mass relative to controls, as demonstrated by the analysis of luciferase signals (Figure 5E), tumor size (Figure 5F), tumor weight (Figure 5G) and tumor volume (Figure 5H). No body weight reduction was observed in any cohort (Figure 5I).

FIGURE 5.

EHD1 suppression hampered tumor immune evasion and increased the effectiveness of anti‐PD‐1 therapy. (A‐B) qRT‐PCR and Western blotting were performed to detect EHD1 expression in engineered clones derived from the mouse LUAD cell line LA795 transfected with shCtrl, shEHD1#1, or shEHD1#2. (C) Schematic showing the construction of xenograft models in C57BL/6 mice bearing shCtrl or shEHD1 tumors and the treatment regimen with anti‐PD‐1/anti‐IgG. (D) Mice were allocated into four experimental cohorts: shCtrl + anti‐IgG, shEHD1 + anti‐IgG, shCtrl + anti‐PD‐1, and shEHD1 + anti‐PD‐1. n = 5 mice per group. The diagram shows growth curves of xenograft tumor volume in the different treatment groups. (E) Bioluminescence images were obtained on Day 18. The bioluminescence signals were quantified and are presented in bar graphs, which depict the statistical analysis of the luciferase activity in the xenografts for each group (n = 5). Additionally, at the end of the experiment, photographic documentation of thexenograft tumors from all the groups was performed. (F‐H) Subsequent statistical assessments were conducted to compare tumor sizes (F), mean tumor weights (G), and tumor volumes (H) across the groups. (I) Statistical assessment of the body weights of the mice across the groups. (J) Representative images of H&E staining and IHC staining for EHD1, PD‐L1, and CD8 in tumor tissue slices (left) and data statistics of the IHC scores of corresponding proteins in different groups (right) are displayed. Data were collected from three independent experiments. Two‐tailed unpaired Student's t test; The data are shown as the mean ± SD. ns, not significant; ** P < 0.01, *** P < 0.001. CD8, Cluster of Differentiation 8; EHD1, Eps15 homology domain 1; H&E, Hematoxylin and eosin; IgG, Immunoglobulin G; IHC, Immunohistochemistry; L, Left; LUAD, Lung adenocarcinoma; PD‐L1, Programmed cell death 1 ligand 1; PD‐1, Programmed cell death 1; R, Right; shCtrl, short hairpin RNA control; shEHD1, short hairpin RNA for EHD1; SD, Standard deviation.

An increase in CD8⁺ cytotoxic T cells (CTLs) may serve as a sign of a positive response to ICB therapy [47, 48]. IHC showed less PD‐L1 but more CD8 staining in xenograft tumor tissues from the shEHD1 group than in those from the shCtrl group (Figure 5J). Specifically, the concomitant administration of anti‐PD‐1 and EHD1 knockdown resulted in a pronounced augmentation of activated CD8⁺ T cells within neoplastic regions (Figure 5J). Collectively, these results indicated that EHD1 contributed to tumor immune escape and ICB treatment resistance by upregulating PD‐L1 expression.

3.6. The expression and mRNA stability of EHD1 were regulated by YTHDF1

m⁶A is considered the most common epigenetic modification of mRNA [49]. m⁶A modification can influence tumor progression by regulating the stability or translation efficiency of target mRNAs, yet it remained unexplored whether EHD1 is regulated through this pathway. Using the online bioinformatic tools SRAMP and MEME, the typical m⁶A motif RRACH (D = A, G or U; R = A or G; H = A, U or C) was identified in EHD1 mRNA (Figure 6A‐B), suggesting the potential of m6A modification in EHD1. To test the presence of m⁶A methylation in EHD1, MeRIP for the enrichment of methylated EHD1 was performed. MeRIP‐qPCR experiments revealed that m⁶A modification occurred in the EHD1 mRNA (Figure 6C). By integrating the data of RMVAR, RM2target, and RMBase online databases, we screened out two m⁶A regulators that might modify EHD1 mRNA m⁶A methylation (Figure 6D). RPISeq predicted potential interactions of YTHDF1 protein and insulin‐like growth factor 2 mRNA binding protein 1 (IGF2BP1) protein with EHD1 mRNA (Figure 6E). A significant correlation was observed between mRNA expression levels of EHD1 and YTHDF1 among these m⁶A regulators (R = 0.21, P < 0.001, Figure 6F, Supplementary Figure S8A). Thus, YTHDF1 was selected for further investigation.

FIGURE 6.

YTHDF1 regulated EHD1 expression and its mRNA stability. (A‐B) The typical m⁶A motif RRACH in EHD1 mRNA was identified using SRAMP (A) and MEME (B) (R = A or G; H = A, U or C). (C) Enrichment of m⁶A modifications on EHD1 mRNA was detected by MeRIP‐qPCR. (D) Venn diagram showing the regulators of m⁶A transcriptional modifications upstream of EHD1, which were predicted by RM2target, RMBase, and RMVAR. (E) The binding of EHD1 mRNA to YTHDF1 or IGF2BP1 was predicted using RPISeq. RF and SVM classifiers are used to predict RNA‐protein interaction probabilities based on sequence features; values above 0.5 suggest a potential interaction. (F) The correlation between the m⁶A regulators YTHDF1 and EHD1 was characterized by Spearman correlation analysis based on mRNA expression levels using the TCGA‐LUAD dataset. (G‐H) EHD1 mRNA and protein levels in LUAD cells with YTHDF1 knockdown were measured by qRT‐PCR and Western blotting. (I‐J) An Act D pulse‐chase experiment was performed to determine the stability of EHD1 mRNA in the indicated cells treated with 5 µg/mL Act D. (K‐L) EHD1 expression in shCtrl and shYTHDF1 LUAD cells treated with (K) MG132 or (L) NH₄Cl for 8 h was detected by Western blotting analysis. Data were collected from three independent experiments. Two‐tailed unpaired Student's t test; The data are shown as the mean ± SD. ns, not significant; * P < 0.05, ** P < 0.01, *** P < 0.001. Act D, Actinomycin D; EHD1, Eps15 homology domain 1; IGF2BP1, Insulin like growth factor 2 MRNA binding protein 1; IgG, Immunoglobulin G; LUAD, Lung adenocarcinoma; m⁶A, N6‐methyladenosine; MeRIP‐qPCR, Methylated RNA immunoprecipitation‐quantitative polymerase chain reaction; RF, Random forest algorithm; shCtrl, short hairpin RNA control; shYTHDF1, short hairpin RNA for YTHDF1; SVM, Support vector machine; YTHDF1, YTH N6‐Methyladenosine RNA Binding Protein F1; SD, Standard deviation.

RT‐PCR and Western blotting analyses revealed that the depletion of YTHDF1 (shYTHDF1#1 and shYTHDF1#2) resulted in a subsequent decrease of EHD1 expression (Figure 6G‐H, Supplementary Figure S8B‐C). Next, we wondered whether the YTHDF1‐mediated effect on EHD1 expression occurs at the mRNA or protein level. Act D chase assay showed that silencing YTHDF1 diminished the mRNA stability of EHD1 in LUAD cells (Figure 6I‐J). However, the proteasome inhibitor MG132 and lysosome inhibitor NH₄Cl did not reverse the reduction in the EHD1 protein level elicited by YTHDF1 silencing (Figure 6K‐L). Together, our data suggest that YTHDF1 binds to EHD1 mRNA, promoting its stability and up‐regulating EHD1 expression.

3.7. The regulation of EHD1 by YTHDF1 exhibited a dependency on m⁶A modification

Recent studies indicated that YTHDF1 could bind to mRNA, thereby increasing its stability [50, 51, 52]. RIP assays followed by qPCR and gel electrophoresis assays demonstrated that EHD1 mRNA was greatly enriched by YTHDF1 antibody in A549 and H1299 cell lines (Figure 7A‐B). Subsequently, we questioned whether the YTHDF1‐regulated EHD1 mRNA stability depended on m⁶A modification. K395 and Y397 are key amino acids in the YTH domain that control the binding affinity to mRNA via the m⁶A‐binding pockets [36, 53]. Thus, we constructed an HA‐tagged mutant YTHDF1 vector (YTHDF1^MUT) with the two critical amino acid mutations (K395A and Y397A) (Figure 7C). RIP assay followed by qPCR showed that EHD1 mRNA was efficiently co‐immunoprecipitated with YTHDF1^WT in A549 and H1299 cells, whereas the amount of EHD1 mRNA co‐immunoprecipitated with YTHDF1^MUT was minimal (Figure 7D). Western blotting analysis revealed that overexpression of YTHDF1^WT markedly upregulated the expression of EHD1 protein, while YTHDF1^MUT exerted a minimal influence on the expression levels of the EHD1 protein (Figure 7E, Supplementary Figure S9). Act D chase assay showed that overexpression of YTHDF1^WT induced an increase in the half‐life of EHD1 mRNA, while overexpression of YTHDF1^MUT, compared to YTHDF1^WT, demonstrated a shorter half‐life of EHD1 mRNA degradation in LUAD cells (Figure 7F‐G).

FIGURE 7.

EHD1 is regulated by YTHDF1 in an m⁶A‐dependent manner. (A‐B) RIP‐qPCR and agarose electrophoresis results showing the enrichment of EHD1 mRNA. (C) Schematic description showing the construction of wild‐type (YTHDF1^WT) and mutant (YTHDF1^MUT) YTHDF1 vectors. (D) RIP‐qPCR was used to detect the precipitation of EHD1 mRNA by YTHDF1 in the YTHDF1^WT or YTHDF1^MUT group. (E) Western blotting analysis showing the effects of YTHDF1^WT and YTHDF1^MUT on the protein expression level of EHD1. (F‐G) An Act D chase assay was used to investigate the effect of 5 µg/mL Act D on EHD1 mRNA in the indicated cells. (H) Schematic representation of an EHD1 mutant variant (EHD1 ^MUT2) with alterations at the YTHDF1‐related m⁶A peak site of EHD1 mRNA. (I) Western blotting analysis revealed the influence of YTHDF1‐mediated m⁶A modification on the expression of Flag‐tagged EHD1. Data were collected from three independent experiments. Two‐tailed unpaired Student's t test; The data are shown as the mean ± SD. ns, not significant; * P < 0.05, ** P < 0.01, *** P < 0.001. Act D, Actinomycin D; CDS, Coding sequence; EHD1, Eps15 homology domain 1; EHD1^MUT2, Mutant EHD1 mutant EHD1#2, which exhibited overexpression of EHD1 with mutations at the m⁶A peak sites; EHD1^OE, EHD1 overexpression; HA, Hemagglutinin; IgG, Immunoglobulin G; RIP‐qPCR, RNA immunoprecipitation‐quantitative polymerase chain reaction; t_1/2, Half life time; UTR, Untranscribed region; YTHDF1, YTH N6‐methyladenosine RNA binding protein F1; YTHDF1^MUT, Mutant YTHDF1; YTHDF1^WT, Wild type YTHDF1; SD, Standard deviation.

To further delineate the role of the m⁶A peak within EHD1 in the regulatory activity of YTHDF1, we engineered Flag‐tagged EHD1 mutants (EHD1 ^MUT2) bearing alterations at the m⁶A peak sites, guided by the profiling of YTHDF1‐associated m⁶A‐modified transcripts in A549 cells (Figure 7H). Western blotting analysis demonstrated a reduction in protein expression levels of EHD1 ^MUT2 when compared to EHD1 ^OE, suggesting diminished responsiveness to the overexpression of YTHDF1^WT (Figure 7I). Relative to YTHDF1^WT, YTHDF1^MUT exhibited no discernible impact on the expression of Flag‐tagged EHD1 protein of either isoform (Figure 7I). To summarize, the data presented herein confirm that the stability of EHD1 mRNA, as governed by YTHDF1, is in an m⁶A modification‐dependent manner.

3.8. Expression levels of EHD1 were associated with those of PD‐L1 and YTHDF1 in clinical LUAD specimens

To elucidate the potential regulatory mechanisms of EHD1 in patients with LUAD, relevant patient‐derived tissue specimens were selected for study. IHC analysis showed that EHD1, PD‐L1, and YTHDF1 all exhibited a concordant upregulation trend in LUAD patients (Figure 8A‐B), suggesting that EHD1 might upregulate PD‐L1 expression and potentially interact with YTHDF1 in a regulatory manner. Using real‐world data, we found that high EHD1 expression conferred a worse prognosis for LUAD patients treated with ICB therapies (Figure 8C). These findings combined with the experimental data, derived from LUAD cell lines and xenograft models, revealing that m⁶A‐modified EHD1 can contribute to immune evasion through PD‐L1 endosomal trafficking.

FIGURE 8.

EHD1 expression levels were associated with those of PD‐L1 and YTHDF1 in clinical LUAD samples. (A) Representative images of EHD1, PD‐L1, and YTHDF1 IHC staining of tumor tissue samples from patients with LUAD (left). Bar graphs (right) showing the distributions of PD‐L1 and YTHDF1 expression in different EHD1 expression groups as determined by IHC in clinical LUAD samples. (B) Expression levels of PD‐L1 and YTHDF1 in patients with low (Patient 1) or high (Patient 2) EHD1 expression. (C) Kaplan‐Meier survival curves illustrating the PFS stratification categorized based on the level of EHD1 expression based on a publicly available dataset (GSE135222) from the GEO database. (D) Diagram of the mechanism by which m⁶A‐modified EHD1 promotes immune evasion and ICB resistance through PD‐L1 endosomal transport in LUAD cells. Data were collected from three independent experiments. Two‐tailed unpaired Student's t test; The data are shown as the mean ± SD. ns, not significant; *** P < 0.001. EHD1, Eps15 homology domain 1; GEO, Gene Expression Omnibus; ICB, Immune checkpoint blockade; IHC, Immunohistochemistry; LUAD, Lung adenocarcinoma; PD‐1, Programmed cell death 1; PD‐L1, Programmed cell death 1 ligand 1; PFS, Progression free survival; RAB11, Ras‐related protein 11; Ub, Ubiquitin; YTHDF1, YTH N6‐methyladenosine RNA binding protein F1; SD, Standard deviation.

4. DISCUSSION

In this study, EHD1 was identified and found to reinforce the immune evasion of LUAD cells. In detail, YTHDF1 increased EHD1 mRNA stability and expression via m⁶A methylation. EHD1 interacts with the PD‐L1‐ICD, promoting PD‐L1 recycling, preventing lysosomal degradation, and upregulating PD‐L1 expression. Increased PD‐L1 expression impairs the killing ability of cytotoxic T cells, thereby enhancing tumor immune escape (Figure 8D).

In the present study, EHD1 has been identified for the first time as a factor related to immune evasion. The rapid advancement of next‐generation, high‐throughput RNA sequencing technologies has led to the development of tremendous amounts of high‐dimensional “omics” data and powerful bioinformatics tools, facilitating comprehensive studies across various aspects of cancer [54, 55, 56]. The therapeutic administration of immune ICB in oncology can confer long‐lasting clinical advantages; however, the response to such treatment is limited to a subset of patients [57]. To predict ICB response, a computational method TIDE was developed. This bioinformatics tool encapsulates the modeling of two cardinal mechanisms of tumor immune evasion: firstly, the elicitation of T cell dysfunction within tumors possessing pronounced CTL infiltration, and secondly, the impediment of T cell infiltration in tumors exhibiting diminished CTL infiltration [35]. It is reported that TIDE prognosticated the clinical outcomes of cancer patients subjected to first‐line anti‐PD1 or anti‐CTLA4 therapy with greater precision than alternative biomarkers, including PD‐L1 expression and mutational burden [35]. Analogous to our screening approach, TIDE algorithms are used to identify biomarkers that possess significant potential for immune escape and may be indicative of effective immunotherapy outcomes [58, 59, 60].

Endocytic recycling occurs when receptors undergo sorting within EEs, alternatively known as sorting endosomes, and are subsequently trafficked to the plasma membrane via a direct, expedited pathway (constituting rapid recycling), or through an intermediate organelle, referred to as the endocytic recycling compartment (ERC), which represents a more protracted recycling process [61, 62]. EHD1 has been elucidated to fulfill pivotal functions at multiple junctures of the endocytic recycling pathway, including the trafficking of plasma membrane constituents from EEs to the ERC, as well as their subsequent return to the plasma membrane, the retrograde conveyance of cargo from EEs to the Golgi apparatus, and the vesicular scission events originating from endosomes [22, 23, 24, 25, 63]. For example, the interaction of EHD1 with the RAB11 effector, RAB11‐FIP2, regulates the translocation of RAB11‐enriched vesicular compartments from the ERC to the plasma membrane [25]. Particularly, delineating the precise step(s) at which PD‐L1 endocytic recycling is modulated by EHD1 represents an attractive avenue for future investigation.

In this study, EHD1 exerts regulatory control endosomal trafficking of PD‐L1 to enhance PD‐L1 expression, contributing to immune surveillance. PD‐L1 has previously been shown to be trafficked to the plasma membrane, which prevents its subsequent degradation in the lysosome. Manipulation of PD‐L1 endosomal trafficking plays a significant role in immune escape and sensitivity to immunotherapy [38, 40, 64]. The triazine derivative 6J1 impedes the endosomal transport of PD‐L1, thereby rendering tumor cells more vulnerable to the cytolytic action of T cells. Moreover, the concomitant application of 6J1 with an anti‐PD‐1 monoclonal antibody markedly enhances the immune‐mediated antineoplastic response [64]. Canagliflozin treatment induces PD‐L1 degradation in endocytic recycling, significantly diminishing its expression, thereby hindering tumor progression and correlating with enhanced antitumor activity of cytotoxic T cells [40]. Trafficking protein particle complex subunit 4 facilitates the interaction between PD‐L1 and RAB11 in recycling endosomes, promoting RAB11‐mediated recycling of PD‐L1 and surface replenishment on tumor cells, while its depletion reduces PD‐L1 levels, enhancing T cell‐mediated cytotoxicity and sensitizing tumors to checkpoint therapy in murine models [38]. Consequently, in combination with our findings, the strategic modulation of PD‐L1 endosomal trafficking emerges as a viable strategy to potentiate anticancer immune responses, complementing ICB therapy.

YTHDF1, a member of the evolutionarily conserved YT521‐B homology (YTH) domain‐containing protein family, modulates gene expression profiles by specific interactions with m⁶A‐modified mRNAs via its carboxy‐terminal YTH domain, thus altering the post‐transcriptional outcomes of these genes [53, 65]. YTHDF1 promotes the translation of m⁶A‐modified mRNAs by interacting with translation machinery [66, 67]. Moreover, recent research showed that YTHDF1 enhanced the stability of m⁶A‐modified mRNA [50, 51, 52]. When the YTHDF1 is abnormally expressed, the translation or mRNA stability of target genes is disorganized, and tumor progression is propelled [49]. Therapeutic intervention targeting YTHDF1 has been demonstrated to enhance antitumor immunity and potentiate the efficacy of anti‐PD‐1 therapy through distinct mechanisms [68, 69]. The YTHDF1‐m⁶A‐p65‐CXCL1 signaling cascade has been elucidated as the underlying pathway through which YTHDF1 exerts its immunosuppressive effects, resulting in the chemotactic recruitment of myeloid‐derived suppressor cells that subsequently inhibit effector cell function [68]. YTHDF1 facilitates the partial stabilization of PD‐L1 transcription, a requisite process for tumor cells to circumvent the cytotoxicity of effector T cells and mitigate CD8⁺ T cell‐induced ferroptosis [69]. In this study, we found that the expression and mRNA stability of EHD1 is elevated by YTHDF1 in an m⁶A‐dependent manner. Due to the limitations of the experimental apparatus in our laboratory, we did not investigate whether YTHDF1 modulates the translation efficiency of EHD1 mRNA, a subject that warrants future study. Moreover, it is fascinating to investigate the presence of other epigenetic modifications on EHD1, including DNA methylation, histone modification, chromatin remodeling, non‐coding RNA association, and various types of RNA methylation, such as N1‐methyladenosine, 5‐Methylcytosine, and N7‐methylguanosine.

As a limitation of our study, A549 is a LUAD cell line with distinct molecular characteristics. This cell line carries a mutation in Kirsten rat sarcoma viral oncogene homolog (KRAS) and a co‐mutation in serine/threonine kinase 11 (STK11). While KRAS mutations are often associated with a better response to immunotherapy due to a higher mutational burden, STK11 mutations are generally linked to resistance to immunotherapy. The combination of these mutations typically indicates poor prognosis and reduced response to immunotherapy. Therefore, using this model to analyze the effects of immunotherapy may not be the most suitable choice. It is worth mentioning that the findings of this study involve the use of activated Jurkat T cells to perform T cell‐mediated tumor cell killing experiments in vitro. Activated Jurkat cells can kill tumor cells [70, 71, 72, 73, 74], but their efficacy may be weaker compared to autologous T cells. Although activated Jurkat T cells do not fully replicate the real T cell‐mediated tumor killing response, in our study, animals can model this response, which helps to partially compensate for the limitations of activated Jurkat cells.

5. CONCLUSION

Together, we identified and verified that EHD1 modulates immune surveillance and anti‐PD‐1 efficacy. Our study provides not only biological functional and mechanistic insights into the role of EHD1 in tumor immune evasion and ICB response but also uncovers a potential regulatory function of YTHDF1 in modulating immunotherapy sensitivity. These findings indicate that targeting EHD1 or its m⁶A modification is a potential strategy to sensitize LUAD to ICB therapy.

AUTHOR CONTRIBUTIONS

FLT, LC and YX designed the study. FLT, JH and WNF carried out the development of the methodology. YNZ, KXZ, XH, XL, XW, XYW and JR supported analysis and interpretation of data. FLT conducted experiments. FLT, HJ, WNF, LC and YX wrote, reviewed, and edited the manuscript. All authors read and approved the final manuscript.

CONFLICTS OF INTEREST STATEMENT

The authors declare that they have no competing interests.

ETHICS APPROVAL AND CONSENT TO PARTICIPATE

The study was carried out in accordance with the principles of the Declaration of Helsinki and was approved by the Ethics Committee of Harbin Medical University Cancer Hospital (KY2023‐08). The patients/participants provided their written informed consent to participate in this study.

CONSENT FOR PUBLICATION

Not applicable.

Supporting information

Supporting information

Tian F, Huang J, Fan W, Li X, Zhan Y, Zhu K, et al. m⁶A‐modified EHD1 controls PD‐L1 endosomal trafficking to modulate immune evasion and immunotherapy responses in lung adenocarcinoma. Cancer Commun. 2025;45:1285–1308. 10.1002/cac2.70052

DATA AVAILABILITY STATEMENT

Data is available on reasonable request. All data generated or analyzed during this study are available from either the supplementary information or the corresponding author.