SLC1A5-mediated kynurenine metabolism drives AHR-FANCD2 axis to remodel chromatin and induce T cell exhaustion in lung adenocarcinoma

Abstract

Background

Kynurenine, a byproduct of tryptophan breakdown, is linked to immune suppression during cancer development. This study explores the involvement of the amino acid transporter solute carrier family 1 member 5 (SLC1A5) in kynurenine-mediated T cell exhaustion in LUAD and delves into its functional mechanism.

Methods

RNA-sequencing analysis was employed to identify transcriptome differences between T progenitor and terminal exhausted T cells (TPEX vs. TEX). The SLC1A5 expression was detected in T cells following L-kynurenine (L-ky) treatment. Mouse LUAD cells LLC were implanted into wild-type (WT), SLC1A5 knockout (SLC1A5^−/−), SLC1A5^flox/flox (SLC1A5^fl/fl), or CD8⁺ T cell-specific SLC1A5 conditional knockout (SLC1A5^cko) mice, followed by L-ky treatment, to examine the effect of SLC1A5^cko on L-ky-mediated tumorigenesis and T cell exhaustion. Interacting proteins of AHR, a core transcription factor in the kynurenine pathway, were explored by liquid chromatography/mass spectrometry and bioinformatics.

Results

SLC1A5 is upregulated in TEX, and its expression in CD8⁺ T cells was increased by L-ky treatment dose-dependently. The tumorigenic activity of LLC cells, under L-ky treatment stimulation, was suppressed in both SLC1A5^−/− and SLC1A5^cko mice, accompanied by increased T cell activity within tumors. CD8⁺ T cells extracted from SLC1A5^cko mice also showed reduced L-ky uptake and increased cytotoxicity in vitro. Mechanistically, AHR recruits the chromatin modifying enzyme FANCD2 to enhance SLC1A5 expression, promoting chromatin accessibility in T cells and cell exhaustion.

Conclusion

This study suggests that SLC1A5 is upregulated in TEX, which modulates kynurenine metabolism and induces T cell exhaustion through the AHR-FANCD2 axis-mediated chromatin remodeling.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12964-026-02732-3.

Introduction

Lung cancer remains a leading global health challenge, with approximately 2 million new diagnoses and 1.8 million fatalities each year, making it the most prevalent and deadly form of cancer [1]. Non-small cell lung cancer (NSCLC) accounts for roughly 85% of these cases, with lung adenocarcinoma (LUAD) comprising over half of NSCLC diagnoses [2, 3]. Over the past ten years, remarkable progress in LUAD treatment has significantly extended patient survival [4]. Notably, therapies targeting programmed death 1 (PD-1) and its ligand (PD-L1) with specific antibodies have transformed outcomes, offering prolonged survival for many LUAD patients in both initial and subsequent treatment lines [5, 6]. However, these treatments yield response rates below 20% in non-selected patients and come with high costs [7]. This underscores the urgent need to better understand tumor immune dynamics and broaden the group of patients who can benefit from immunotherapy.

Tumor-growth cytokines drive cancer progression by triggering self-sustaining signaling within tumor cells and fostering interactions with the surrounding tumor microenvironment (TME) [8]. The TME, a complex network of immune cells, endothelial cells, fibroblasts, and tumor cells, is linked to varied clinical outcomes [9]. The presence of cytolytic immune cells, particularly CD8⁺ T cells, is a key indicator of favorable prognosis across multiple cancers due to their ability to trigger cancer cell death [10, 11]. These CD8⁺ T cells attack malignant cells by releasing granzymes, perforin, and pro-inflammatory cytokines [12]. However, their efficacy is frequently hindered by suppressive mechanisms within the TME [13]. Despite their promise, most patients do not achieve lasting responses to immunotherapy, and even tumors with highly immunogenic neoantigens may progress after initial immune activity [14]. Strategies to boost CD8⁺ T cell function and restore their cancer-killing capacity hold significant potential for improving anti-tumor immunity.

Metabolites from tumor, immune, or stromal cells within the TME fulfill critical functions in modulating metabolic responses, often contributing to immune suppression and tumor growth [15–17]. Kynurenine, a metabolite produced through tryptophan breakdown via indoleamine 2,3-dioxygenase (IDO) and tryptophan 2,3-dioxygenase (TDO) pathways, is essential for cellular energy production and immune regulation [18]. Blocking tryptophan-to-kynurenine conversion by targeting IDO1 or TDO2 has shown promise in enhancing immune responses in LUAD [19, 20]. Kynurenine acts as a powerful immunosuppressive molecule, promoting regulatory T cell (Treg) differentiation while inhibiting cytotoxic T cell development [21]. It also activates the aryl hydrocarbon receptor (AHR), a transcription factor that influences immune cell behavior and supports inducible Treg differentiation [22]. Elevated kynurenine levels and metabolism are observed in multiple solid tumors, including lung cancer [23].

In cancer, amino acid transporters are frequently overexpressed to support rapid tumor growth by securing sufficient amino acid supplies [24]. Solute carrier family 1 member 5 (SLC1A5), plays a critical role in amino acid transport and is linked to unfavorable prognoses in various cancers [25], positioning it as a promising candidate for targeted cancer treatments [26]. Elevated SLC1A5 expression correlates with enhanced tryptophan metabolism and kynurenine production, which drives cancer cell proliferation [27]. Our research identifies SLC1A5 as a significantly upregulated gene in tumor-infiltrating exhausted T cells (TEX). The function of SLC1A5 in modulating immune responses within the LUAD TME, its interplay with the kynurenine pathway, and the underlying mechanisms remain largely unexplored. This study seeks to investigate these aspects, aiming to uncover new insights into the immunosuppressive processes driving LUAD progression.

Materials and methods

T cell isolation, activation, and exhaustion model construction

C57BL/6 mice (6–8 weeks old, 20–22 g) were procured from Charles River (China) and used in protocols approved by the Animal Ethics Committee of the Institute. The animals were sacrificed by overdosed nembutal, and the spleens were harvested to prepare single-cell suspensions. CD8⁺ T cells were extracted and purified using CD8α MicroBeads (Miltenyi Biotec, #130–049–401) magnetic bead sorting. The cells were seeded at 1 × 10⁶ cells/mL in 96-well flat-bottom plates pre-coated with anti-CD3ε (5 μg/mL, BioLegend, #100,314), and stimulated with soluble anti-CD28 (2 μg/mL, BioLegend, #102,112) and IL-2 (50 U/mL, Peprotech, #212–12). To establish a chronic stimulation model, fresh stimuli were replenished every 48 h for 5–7 d to simulate T cell exhaustion after continuous antigen exposure. Flow cytometry (FACS, BD LSRFortessa) was employed fluorescence-activated cell sorting (FACS) based on PD-1 (BioLegend, #135,210) and TIM-3 (BioLegend, #119,704) expression, isolating PD-1⁺TIM-3⁻ (T progenitor exhausted, TPEX) and PD-1⁺TIM-3⁺ (terminal exhausted T, TEX) subsets for subsequent experiments.

RNA-sequencing (RNA-seq) sample preparation and analysis

After chronic stimulation of the activated CD8⁺ T cells, TPEX and TEX subsets were sorted and harvested, and total RNA was extracted using TRIzol reagent (Invitrogen, #15,596,018). RNA quality was analyzed using Nanodrop and Agilent 2100 Bioanalyzer. Following mRNA enrichment and library preparation, samples were sequenced using the Illumina NovaSeq 6000 platform with paired-end 150 bp reads. Raw data were processed for quality control using fastp and adapter sequences were removed. STAR alignment was performed against the mouse reference genome (mm10), followed by gene quantification using featureCounts. Criteria for differential expression analysis were set at |log₂FC|> 1 and adj p value < 0.05. Volcano plots and heatmaps were generated using ggplot2. A subset of differentially expressed genes (DEGs) was further analyzed for pathway enrichment, functional validation, and integration with other omics data.

mRNA and protein detection

For mRNA extraction, total cellular RNA was extracted, and cDNA was then synthesized via reverse transcription (Vazyme, #R223-01). Quantitative polymerase chain reaction (qPCR) was then carried out employing ChamQ SYBR qPCR Master Mix (Vazyme, #Q711-02) on the ABI QuantStudio 6 Flex platform. GAPDH was employed as the endogenous loading, and results were determined by the 2^−ΔΔCt method.

For protein quantification, cells were lysed using RIPA lysis buffer (Beyotime, #P0013B) to extract total proteins, with concentrations determined utilizing the bicinchoninic acid method. Equal amounts of protein were then separated by SDS-PAGE and loaded to membranes. The membranes were probed overnight with the following antibodies: anti-SLC1A5 (Abcam, #ab237704), anti-AHR (Abcam, #ab84833), anti-IDO1 (Cell Signaling Technology, #86,630), anti-FANCD2 (Abcam, #ab108928), and anti-GAPDH (Proteintech, #60,004–1-Ig). The next day, after 1-h incubation with HRP-conjugated IgG, protein bands were developed using enhanced chemiluminescence, and band signals were analyzed using Image J.

T cell proliferation analysis

The purified CD8⁺ T cells from mouse spleen tissues were labeled with 1 μM carboxyfluorescein diacetate succinimidyl ester (CFSE) dye (Invitrogen, #C34554) in PBS at 37℃ for 10 min, followed by supplementation of 5 times the volume of complete medium (RPMI-1640 + 10% FBS) to stop the staining. The CFSE-labeled T cells were plated at 1 × 10⁶ cells/mL in 96-well plates coated with anti-CD3ε, and stimulated with soluble anti-CD28 (2 μg/mL) and IL-2 (50 U/mL) as described above. L-kynurenine (L-ky, Sigma, #K8625) was added at 0, 25, 50, or 100 μM to the treatment group. After 48 h, cells were collected. CFSE fluorescence intensity was determined by FACS (BD LSRFortessa), and cell proliferation was assessed by analyzing fluorescence dilution. FlowJo v10 was employed for data analysis.

Immunofluorescence analysis of the subcellular localization of AHR

After L-ky treatment for 48 h, the CD8⁺ T cells were washed, fixed on poly-L-lysine-coated 12-well coverslips. Following fixing and permeabilization, the cells were blocked with 5% goat serum and then incubated overnight with anti-AHR (Santa Cruz, #sc-133088, 1:100) at 4℃. The other day, cells were incubated with Alexa Fluor 488-conjugated IgG (Invitrogen, #A11008, 1:200) for 1 h, followed by DAPI staining (Sigma, #D9542, 300 nM) for 5 min. After mounting, images were captured utilizing a confocal microscope (Zeiss LSM880) in the 405 nm (DAPI) and 488 nm (AF488) channels. AHR nuclear localization was analyzed using Zen Blue software.

Chromatin immunoprecipitation (ChIP) and qPCR/ChIP-seq

Following the protocols of the SimpleChIP Enzymatic Kit (Cell Signaling Technology, #9003), the L-ky-treated CD8⁺ T cells were crosslinked with 1% formaldehyde for 10 min, which was stoped by the addition of glycine. After cell lysis, chromatin was truncated to fragments of approximately 200–1000 bp using micrococcal nuclease, and IP was carried out with antibodies: anti-AHR (Abcam, #ab84833), anti-H3K27ac (Abcam, #ab4729), anti-H3K9ac (Abcam, #ab10812), anti-FANCD2 (Abcam, #ab108928). Immunoprecipitated DNA was purified using column-based methods and used for the following experiments: ChIP-qPCR: SYBR Green qPCR Mix (Vazyme, #Q141) was used on a QuantStudio 5 system (Thermo Fisher) to assess enrichment in the promoter regions (2–3 kb upstream) of SLC1A5, PD-1, TIM-3, and CTLA4. Specific primers were designed and experiments were conducted in triplicate. ChIP-seq: Libraries were prepared using the VAHTS Universal DNA Library Prep Kit (Vazyme, #ND607) and sequenced on the Illumina NovaSeq platform. Data were aligned to the mm10 genome using Bowtie2, and peaks were identified with MACS2 (v2.2.7.1). Results were visualized using IGV (Broad Institute).

Chromatin accessibility sequencing (ATAC-seq)

CD8 + T cells were cultured under chronic stimulation conditions and treated with L-kynurenine (100 μM) for 48 h before harvesting for ChIP/ATAC-seq library preparation. All procedures were conducted adhering to the protocols of the TruePrep DNA Library Prep Kit V2 (Vazyme, #TD503). For each sample, 5 × 10⁴ CD8⁺ T cells were washed and then lysed in ice-cold lysis buffer (10 mM Tris–HCl, 10 mM NaCl, 3 mM MgCl₂, 0.1% IGEPAL). Nuclei were collected and incubated with the Tn5 transposase mixture at 37℃ for 30 min. After DNA purification, libraries were prepared and subjected to high-throughput sequencing using the NovaSeq platform. Sequencing data were quality-controlled with FastQC, aligned to the mm10 genome using Bowtie2, and open chromatin regions were identified using MACS2. Differential analysis was performed using DiffBind (v3.8). Functional annotation of differentially accessible regions was carried out using clusterProfiler (v4.6) for GO/KEGG enrichment analyses, and motif analysis was performed using HOMER (v4.11).

Plasmid construction and cell transfection

cDNAs of human AHR, FANCD2, and APEX2 were amplified by PCR and cloned into pcDNA3.1-HA, pcDNA3.1-Flag, pGEX-4 T-1, or pCDH-APEX2 vectors to construct HA-AHR, Flag-FANCD2, GST-AHR, and APEX2-AHR vectors. Sequences were verified by sequencing. Constructed plasmids were transfected into HEK293T cells using Lipofectamine 2000 (Invitrogen, #11,668,019) adhering to the recommended protocol. T cells were transfected using the Lonza Nucleofector system (Kit V, Program X-001). Cells were harvested 48 h post-transfection for subsequent Co-IP, pull-down, or APEX experiments.

Co-immunoprecipitation (Co-IP)

HEK293T cells were plated at 2 × 10⁵ cells/well in six-well plates and transfected 24 h later with HA-AHR and Flag-FANCD2 expression plasmids (1 μg/plasmid) utilizing Lipofectamine 2000. After 48 h, cells were collected and lysed in ice-cold RIPA buffer for 30 min. After centrifugation, the supernatant was incubated with anti-HA (CST, #3724S, 1:100) or anti-Flag (Sigma, #F1804, 1:100) at 4℃ for 6 h. Protein A/G magnetic beads (Thermo Fisher, #88,802) were employed, and the mixture was incubated overnight at 4℃. The beads were washed RIPA buffer, and protein samples were eluted in 2 × SDS sample buffer. After boiling at 95℃ for 10 min, proteins were collected for WB analysis using anti-Flag or anti-HA to analyze protein interactions.

Glutathione S-Transferase (GST) pull-down experiments

The GST-AHR expression plasmid was introduced into E. coli BL21 (DE3) cells, which were grown at 37℃ until reaching an optical densityof about 0.6. Protein production was triggered by adding 0.2 mM IPTG and incubating at 16℃ overnight. Following cell collection, the bacteria were resuspended in PBS and disrupted using ultrasonication. The GST-AHR fusion proteins were isolated with GSH-Sepharose 4B magnetic beads (GE Healthcare, #17–0756-01). The purified GST-AHR protein, bound to the beads, was mixed with HEK293T cell lysates (containing 200–300 μg of protein in a 500 μL total volume) and incubated at 4℃ for 6 h with gentle agitation. The beads were washed five times, and the bound proteins were eluted using SDS sample buffer for analysis via SDS-PAGE. Protein interactions were visualized through Coomassie Blue staining or WB analysis. For protein identification, samples were submitted for liquid chromatography/mass spectrometry (LC/MS) analysis.

APEX proximity labeling and mass spectrometry (APEX Proximity Labeling.⁺ LC/MS)

After constructing the APEX2-AHR fusion expression plasmid, it was electroporated into T cells (using the Lonza 4D Nucleofector system and the Primary Cell P3 Kit, Program DN100). After 48 h, cells were incubated in pre-warmed RPMI 1640 complete medium with 500 μM biotin-phenol (Iris Biotech, #LS-3500.500) at 37℃ for 30 min. The labeling reaction was triggered by adding 1 mM hydrogen peroxide (H₂O₂) for 1 min and promptly stopped using a quenching solution containing 10 mM sodium ascorbate, 5 mM Trolox, and 10 mM sodium azide. The cells were then washed three times with ice-cold PBS and lysed in RIPA buffer. Following centrifugation, the supernatant was collected and incubated with Dynabeads Streptavidin C1 magnetic beads (Invitrogen, #65,001) at 4℃ for 6 h to capture biotinylated proteins. After extensive washing, the proteins were eluted by boiling in SDS sample buffer and analyzed using SDS-PAGE. For mass spectrometry analysis, samples were sent to a professional platform (e.g., PTM Biolab) for LC/MS analysis, and data were analyzed using MaxQuant. Chromatin-related proteins associated with AHR were filtered using the BIOGRID database.

Kynurenine treatment and transport/metabolism analysis

To examine kynurenine uptake by T cells, the activated CD8⁺ T cells were treated with 14C-labeled L-ky (American Radiolabeled Chemicals, #ART 0653) for 30 min. After the reaction was terminated, cells were rinsed, lysed, and the intracellular radioactivity was evaluated using a liquid scintillation counter (Beckman LS6500) to reflect L-ky uptake. For metabolite detection, T cell lysates were precipitated with methanol to remove proteins, and LC/MS analysis was performed to measure the concentrations of kynurenine metabolites, 3-hydroxykynurenine (3-HK) and kynurenic acid (KYNA). Targeted metabolomics platforms (Shanghai Synchrotron Radiation Facility) were used for the measurements.

For the delayed treatment assay, CD8 + T cells were first activated with anti-CD3/CD28 for 48 h. Subsequently, the medium was replaced with fresh medium containing L-kynurenine (100 μM) or vehicle, and cells were cultured under chronic stimulation conditions as described above.

Enzyme-linked immunosorbent assays (ELISA)

The supernatants of CD8⁺ T cells treated with L-ky or the FANCD2 inhibitor TNG463 (10 μM) were collected, centrifuged, and harvested for ELISA detection. Commercial ELISA kits (eBioscience, IFN-γ: #88–7314-88, GZMB: #88–8021-88, TNFA: #88–7324-88) were used adhering to the manufacturer's protocol. After adding samples and standards, the plates were incubated at 37℃ for 1–2 h. After washing, biotin-labeled antibodies, HRP-conjugates, and TMB substrate were added for color development, and the reaction was terminated with stop solution. Absorbance at 450 nm was read utilizing a microplate reader (BioTek Synergy HTX). The concentrations of IFN-γ, GZMB, and TNFA were evaluated based on a standard curve to evaluate T cell functional status.

T cell cytotoxicity assessment

After polarization, T cells were pre-treated for 24 h with L-ky (100 μM) or FANCD2 inhibitor TNG463 (10 μM). The effector-to-target cell ratio was set at 10:1, and the T cells were added to pre-seeded LLC target cells (5 × 10⁴ cells/well). After 24 h of co-culture, supernatants were collected and analyzed using a lactate dehydrogenase (LDH) release assay kit (Promega, #G1780) following the manufacturer's guidelines. The supernatant was mixed with the reagent and incubated for 30 min. Absorbance was measured at 490 nm. LDH release rates were calculated based on the high-value control (complete lysis group) and blank control, reflecting T cell-mediated cell lysis and cytotoxicity.

Animal experiments

The animal models used in this study included C57BL/6 wild-type (WT) mice, SLC1A5 knockout mice (SLC1A5^−/−), SLC1A5^flox/flox (SLC1A5^fl/fl) mice, CD8-Cre transgenic mice, and CD8⁺ T cell-specific SLC1A5 conditional knockout (SLC1A5^cko) mice, all purchased from Charles River (China). All mice were 8–10 weeks old males, housed in SPF animal rooms with approval from the Institutional Animal Ethics Committee. Mouse LLC cells (VCM00009, Vicell, China) were transfected with overexpression plasmids of IDO1 using Lipofectamine 2000. In tumor implantation models, LLC cells stably overexpressing IDO1 were suspended in PBS and inoculated on the right flank of WT, SLC1A5^−/−, SLC1A5^fl/fl, or SLC1A5^cko mice subcutaneously (5 × 10⁵ cells/mouse). Tumor volume was measured on day 3, 7, 14, and 28 using the formula: π/6 × length × width². At endpoint (day 28), animals were sacrificed, and tumors were excised, weighed, and prepared for histological and molecular biological analysis. To model lung cancer metastasis, a tail vein injection lung metastasis model was established by injecting LLC cells overexpressing IDO1 (5 × 10⁵ cells), with or without green fluorescent protein (GFP) labeling, into the tail veins of SLC1A5^fl/fl or SLC1A5^cko mice. After 28 d, lungs were harvested, the number of metastatic nodules was recorded, and GFP fluorescence intensity was assessed using the IVIS Spectrum imaging system. Tissue samples were collected for FACS, immunohistochemistry (IHC), and gene expression analysis. To investigate the impact of tryptophan metabolism on kynurenine production, mice were fed with tryptophan-deficient diet (Research Diets, TD.90229) starting from day 7 after tumor inoculation for 7 d to restrict kynurenine sources. Body weight and physiological status were monitored. In the immunotherapy experiment, mice were randomly assigned to receive the following treatments: PBS, SLC1A5 small molecule inhibitor V-9302 (60 mg/kg, oral gavage daily), IgG control, PD-1 monoclonal antibody (mAb) alone (BioXcell, BE0146, 200 μg, intraperitoneal injection every other day), or the combination of V-9302 and the PD-1 mAb. L-ky and the FANCD2 inhibitor TNG463 (15 mg/kg, intraperitoneal injection daily) were employed for experimental needs. Treatment duration was 7–10 d, and tumor growth, T cell exhaustion, and functional molecule expression were monitored.

Tumor tissue single-cell preparation and FACS analysis

At endpoint, tumor tissues from euthanized mice were collected, minced, and placed in RPMI-1640 complete medium. Collagenase IV (1 mg/mL, Sigma-Aldrich, C5138) and DNase I (0.1 mg/mL, Roche, 11,284,932,001) were added, and the tissue was digested by shaking at 37℃ for 30 min. The digested mixture was filtered to obtain a single-cell suspension. Red blood cells were lysed using 150 mM NH₄Cl, and the suspension was washed and resuspended in FACS buffer (PBS ⁺ 2% FBS) for subsequent staining. First, the LIVE/DEAD Fixable Aqua dye (Thermo Fisher, L34966) was used to exclude dead cells, followed by surface antigen staining with fluorescent-labeled antibodies: anti-CD45 (BioLegend, 30-F11), anti-CD3 (BioLegend, 145-2C11), anti-CD8a (BioLegend, 53–6.7), anti-PD-1 (BioLegend, 29F.1A12), anti-TIM-3 (BioLegend, RMT3-23), and anti-SLC1A5 (R&D Systems, MAB5696). The staining was performed at 4℃ in the dark for 30 min. For functional T cell analysis, cells were stimulated with PMA (50 ng/mL) + ionomycin (1 μg/mL) for 4 h, and BFA (BD GolgiPlug) was added. After fixation/permeabilization using BD Cytofix/Cytoperm buffer, intracellular staining was performed with anti-GZMB (BioLegend, QA16A02), anti-TNFA (BioLegend, MP6-XT22), and anti-IFN-γ (BioLegend, XMG1.2). All samples were analyzed using the BD LSRFortessa flow cytometer and analyzed with FlowJo v10.8 software.

IHC assays

Paraffined tumor tissues were sectioned into 4 μm slices, baked at 60℃ for 1 h, and subjected to routine dewaxing and graded alcohol rehydration. Antigen retrieval was performed utilizing Tris–EDTA buffer (pH 9.0) in a microwave for 15 min. After cooling, non-specific binding was prevented using normal goat serum, and endogenous peroxidase activity was blocked by 3% H₂O₂. The sections were incubated overnight with anti-Cleaved caspase-3 (C-Cas-3, Cell Signaling Technology, 9661, diluted 1:200) at 4℃, followed by incubation with HRP-labeled IgG. DAB reagent (ZSGB-BIO) was employed for color development, followed by counterstaining with hematoxylin, dehydration, and mounting. Images were captured with a Leica microscope.

Public databases and protein-interaction analysis

To analyze the correlation between SLC1A5 expression and T cell exhaustion in lung adenocarcinoma, gene expression (FPKM) and clinical data from The Cancer Genome Atlas (TCGA)-LUAD dataset were downloaded. The correlation between SLC1A5 and immune checkpoint molecules PDCD1 (PD-1), CTLA4, and HAVCR2 (TIM3) was evaluated using the limma and ggplot2 packages in R (v4.2.1). Additionally, gene expression matrices from the GEO datasets GSE3141 and GSE30219 were downloaded using the GEOquery package. Specifically, the GSE3141 dataset contains gene expression profiling data from 111 lung adenocarcinoma patients, while the GSE30219 dataset includes transcriptomic data from 307 non-small cell lung cancer (NSCLC) samples. These datasets were utilized to validate the correlation between SLC1A5 expression, immune infiltration, and clinical prognosis. After normalization, the CIBERSORT algorithm was utilized to estimate the level of cytotoxic T lymphocyte (CTL) infiltration, and TIDE was applied to compute the T dysfunction score, further evaluating the correlation between SLC1A5 expression and T cell dysfunction. Survival analysis was performed using the survival and survminer packages. Kaplan–Meier survival curves were plotted based on the median expression of SLC1A5, and log-rank tests were conducted to assess its prognostic value. To identify potential interacting factors mediating AHR regulation of chromatin conformation, known AHR protein interactors were extracted from the BIOGRID database (v4.4). Focus was placed on molecules with chromatin regulation functions, and these interactors were cross-referenced with LC/MS data obtained from APEX2-AHR proximity labeling experiments. FANCD2 was identified as a specific chromatin-regulating protein that binds to AHR. This candidate molecule was further validated through co-immunoprecipitation and ChIP-qPCR to confirm its cooperative regulation of SLC1A5 transcription with AHR in subsequent experiments.

Statistical analysis

All statistical analyses were performed using GraphPad Prism 9.0 software. Unless otherwise stated, all experiments were repeated a minimal of six times, with data presented as mean ± standard deviation. Differences were compared by an unpaired two-tailed Student’s t-test, or by the one-way or two-way analysis of variance (ANOVA) with Tukey's post-hoc examinations, as appropriate. For animal experiments, each group consisted of six samplesto ensure statistical power. In survival analysis, the log-rank test was utilized to assess differences between Kaplan–Meier survival curves. Correlation analysis was performed using Spearman's or Pearson's correlation coefficients. A p-value < 0.05 was considered statistically significant.

Results

SLC1A5 is linked to T cell exhaustion and upregulated following L-ky administration

Chronic stimulation was applied to induce CD8⁺T cell exhaustion, followed by RNA-seq analysis to analyze DEGs in (TPEX-like) (PD1⁺TIM3⁻) and TEX (PD1⁺TIM3⁺). The results revealed a significant upregulation of the amino acid transporter SLC1A5 in TEX cells (Fig. 1A-C). Additionally, the mean fluorescence intensity (MFI) of SLC1A5 was markedly higher in TEX compared with TPEX (Fig. 1D), with consistent patterns observed in mRNA and protein levels (Fig. 1E-F). To further explore the correlation between SLC1A5 expression and T cell exhaustion in LUAD, TCGA-LUAD system-based single-gene analyses were performed, showing positive correlations with classic exhaustion markers PD-1, CTLA4, and TIM3 (Fig S1A). Additionally, in the GSE3141 dataset, SLC1A5 expression presented an inverse correlation with CTL infiltration while a positive correlation with T cell dysfunction (Fig S1B-C). Similarly, in the GSE30219 dataset, a positive link was observed between SLC1A5 expression and T cell exhaustion (Fig S1D-E). Moreover, Kaplan–Meier analysis indicated SLC1A5 as a distinct risk factor in LUAD linking to unfavorable prognosis in patients (Fig S1F).

Fig. 1.

SLC1A5 is linked to T cell exhaustion and upregulated following L-ky administration. A Mouse CD8⁺ T cells were activated through CD3/CD28/IL-2 stimulation, followed by chronic stimulation to induce T cell exhaustion. TPEX (PD1⁺TIM3⁻) and TEX (PD1⁺TIM3⁺) cells were then sorted with FACS for RNA-seq analysis. B-C Volcano plots (B) and heatmap (C) for DEGs between TPEX and TEX identified using RNA-seq analysis. D FACS analysis of the MFI of SLC1A5 in TEX and TPEX. E–F. qPCR (E) and WB (F) analysis of SLC1A5 mRNA and protein levels in TEX and TPEX. L-ky was supplemented to culture system of T cells. FACS analysis of the PD1⁺TIM3.⁺CD8 T cell population after L-ky treatment. I-J qPCR (I) and WB (J) analysis of SLC1A5 mRNA and protein levels in cells after L-ky treatment. K Immunofluorescence detection of nuclear translocation of AHR in T cells after L-ky treatment. L ChIP-qPCR analysis of AHR binding to the SLC1A5 promoter using anti-AHR antibody. Cell-based experiments repeated six times. Data are presented as dot plots and bars. Statistical significance was set at p < 0.05

As an amino acid transporter, SLC1A5 facilitates the transport of tryptophan and the generation of kynurenine, which is linked to immunosuppression [28]. When L-ky was supplemented to the culture system, a significant increase in TEX cell numbers within 48 h was observed (Fig. 1G-H). Interestingly, L-ky treatment led to a dose-dependent increase in SLC1A5 expression in T cells (Fig. 1I-J). To position SLC1A5 within the broader context of amino acid transport, we compared its responsiveness to L-ky treatment against other known kynurenine transporters, SLC7A5 and SLC7A11 [29, 30]. While L-ky treatment elicited a broad upregulation of these transport systems, SLC1A5 demonstrated the most significant induction, with a fold-change markedly surpassing that of Slc7a5 and Slc7a11 (Figure S2A-B). This identifies SLC1A5 as the most sensitive responder to kynurenine accumulation in CD8 + T cells, supporting its pivotal role in the observed metabolic reprogramming. It is hypothesized that L-ky uptake activates the kynurenine endogenous receptor AHR in cells, promoting AHR nuclear translocation and subsequently enhancing SLC1A5 expression. To test this, immunofluorescence was applied to assess AHR nuclear translocation, revealing a significant increase following L-ky supplementation (Fig. 1K). Additionally, AHR was found to bind to the SLC1A5 promoter (Fig. 1L). To rule out the possibility that the increased TEX proportion was due to L-Ky interfering with initial T cell priming, we modified the experimental protocol by adding L-Ky only after 48 h of initial anti-CD3/CD28 activation. Consistent with our initial observations, this delayed L-Ky administration still resulted in a significant expansion of the TEX population and a concomitant reduction in the non-exhausted (PD-1-Tim-3-) fraction (Figure S2C-D). This confirms that kynurenine metabolism specifically drives the progression of T cell exhaustion post-activation.

SLC1A5 knockout diminishes L-ky-mediated T cell exhaustion in mice

In the kynurenine pathway, tryptophan is converted to kynurenine via IDO or TDO and subsequently exported from the cell. To investigate this process, IDO1, the rate-limiting enzyme, was artificially overexpressed in mouse LLC cells (Fig S3A-B), which, as anticipated, resulted in increased kynurenine secretion in cells (Fig S3C). The supportive role of IDO1 in tumor cell growth has been well established [31, 32]. In this study, we had LLC cells overexpressing IDO1 or not subcutaneously implanted into either WT or SLC1A5^−/− mice. The growth rate of IDO1-overexpressing cells was significantly accelerated in both WT and SLC1A5^−/− mice, though the growth rate was notably higher in WT mice compared to SLC1A5^−/− mice (Fig. 2A-C). The number of total CD8⁺ T cells was substantially reduced within tumors formed by IDO1-overexpressing cells (Fig. 2D). However, the populations of both TEX and TPEX cells were significantly elevated by IDO1 overexpression in tumors in WT mice, effects absent in SLC1A5^−/− mice (Fig. 2D-E). The SLC1A5 MFI in the tumors in WT mice was increased by IDO1 overexpression as well, which was absent in SLC1A5^−/− mice (Fig. 2F). Consistent with the expansion of exhausted subsets (TPEX and TEX), we observed a concomitant depletion of the functional cytotoxic population. Specifically, the frequency of GzmB + CD8 + T cells was significantly reduced in IDO1-OE tumors compared to controls (Fig. 2G).

Fig. 2.

SLC1A5 knockout diminishes L-ky-mediated T cell exhaustion in mice. A Mouse LLC cells overexpressing IDO1 or not implanted into either WT or SLC1A5-/- mice subcutaneously, followed by animal sacrifice on day 28. B Growth rate of subcutaneous tumors in mice. C Diagram and weight of tumors formed by LLC cells on day 28. D FACS analysis of the number of CD45⁺CD8⁺ T cells in tumors formed by LLC cells. E FACS analysis of the number of TEX (PD1⁺TIM3⁺) and TPEX (PD1⁺TIM3⁻) cells in tumors formed by LLC cells. F FACS analysis of the number of SLC1A5⁺PD-1⁺CD8⁺ T cells in tumors formed by LLC cells. G FACS analysis of the number of GZMB + CD8 + T cells in tumors formed by LLC cells. H CD8 + T cells from either WT or SLC1A5-/- mice were extracted, activated by CD3/CD28 stimulation, labeled with CFSE, and then treated with L-ky. Groups included: 'T activated' (stimulation in pure medium, system control), 'WT + PBS' (stimulation with vehicle, experimental control), and 'WT + L-Ky' (treatment). Note that statistical comparisons for the L-Ky effect were performed against the WT + PBS vehicle control to account for handling variations. I FACS analysis of CFSE-positive cells to assess T cell proliferation. J WT mice bearing IDO1-overexpressing LLC cells were subjected to tryptophan-deprivation diets or normal diets. K Analysis of LLC cell growth rate in mice after tryptophan deprivation. L Weight of tumors formed by LLC cells after tryptophan deprivation. M FACS analysis of the number of TEX (PD1⁺TIM3⁺) and TPEX (PD1⁺TIM3.⁻) cells in tumors formed by LLC cells after tryptophan deprivation. Each group contained 6 mice. Data are presented as dot plots and bars. Statistical significance was set at p < 0.05

To further delve into the effects of kynurenine-SLC1A5 interaction on T cell function, CD8⁺ T cells from either WT or SLC1A5^−/− mice were extracted, activated by CD3/CD28 stimulation, labeled with CFSE, and then treated with L-ky (Fig. 2H). Notably, CD8⁺ T cells derived from SLC1A5^−/− mice showed increased proliferation versus those derived from WT mice. Additionally, the L-ky treatment substantially reduced proliferation of WT CD8⁺ T cells, but it had no notable impact on the proliferation of SLC1A5^−/− CD8⁺ T cells (Fig. 2I). Furthermore, the WT CD8⁺ T cells were supplemented with an AHR-specific antagonist (IK-175), while SLC1A5^−/− CD8⁺T cells were treated with an AHR agonist (AHR agonist 2), followed by the supplementation of L-ky to the culture system (Fig S4A). Treatment with IK-175 in WT CD8⁺ T cells significantly enhanced T cell proliferation and markedly reduced L-ky-induced T cell exhaustion, whereas AHR agonist treatment led to opposite effects in SLC1A5^−/− CD8⁺T cells (Fig S4B-D). It is worth noting that while L-ky treatment primarily accumulated terminal TEX cells in WT mice (Fig. 1H), the AHR agonist treatment in SLC1A5-/- T cells led to an increase in both TPEX and TEX populations. We postulate that this difference reflects the higher potency of the synthetic agonist, which may drive a broader expansion of the entire exhaustion lineage compared to the metabolite-driven acceleration of TPEX-to-TEX differentiation observed with L-ky.

To further investigate, WT mice bearing IDO1-overexpressing LLC cells were subjected to tryptophan-deprivation diets to limit kynurenine production (Fig. 2J). Following tryptophan deprivation, a marked reduction in the growth of IDO1-overexpressing LLC cells was observed in vivo (Fig. 2K-L), along with a substantial decrease in the populations of TPEX and TPEX (Fig. 2M).

SLC1A5 knockout in T cells inhibits LLC cell tumorigenesis in mice

To eliminate potential interference from other cell types, CD8⁺ T cell-specific SLC1A5 conditional knockout (SLC1A5^cko) mice were generated, to which IDO1-overexpressing LLC cells were implanted (Fig. 3A). Analysis demonstrated that the growth rate of IDO1-overexpressing cells was significantly reduced in SLC1A5^cko mice versus that in SLC1A5^fl/fl mice (Fig. 3B-C). Additionally, the populations of both TEX and TPEX in tumors formed in SLC1A5^cko mice were markedly lower (Fig. 3D), while the population of functional T cells (GZMB⁺) was increased (Fig. 3E).

Fig. 3.

SLC1A5 knockout in T cells inhibits LLC cell tumorigenesis in mice. A CD8⁺ T cell-specific SLC1A5 conditional knockout (SLC1A5^cko) mice were generated, to which IDO1-overexpressing LLC cells were implanted. B Growth rate of subcutaneous tumors formed by LLC cells. C Diagram and weight of subcutaneous tumors formed by LLC cells on day 28. D FACS analysis of TEX (PD1⁺TIM3⁺) and TPEX (PD1⁺TIM3⁻) in LLC-formed tumors. E FACS analysis of GZMB⁺CD8⁺ T cells in LLC-formed tumors. F Immunofluorescence detection of AHR nuclear translocation in CD8⁺ T cells within tumor tissue. G CD8⁺ T cells were isolated from tumor tissues in SLC1A5^cko mice and SLC1A5^fl/fl mice, followed by the treatment of 14C-labeled kynurenine. H Uptake of 14C-labeled kynurenine by T cells. I Quantification of kynurenine metabolites (14C-3-HK and KYNA) in T cells. J CD8⁺ T cells extracted from SLC1A5^cko mice were treated with the AHR agonist 2, followed by FACS analysis of TEX (PD1⁺TIM3⁺) and TPEX (PD1⁺TIM3⁻) cells. L ELISA measurement of IFN-γ, GZMB, and TNF-α production by CD8.⁺ T cells. Animal experiments included six mice per group, and cell-based experiments repeated six times. Data are presented as dot plots and bars. Statistical significance was set at p < 0.05

CD8⁺ T cells were subsequently isolated from tumor tissues in SLC1A5^cko mice and SLC1A5^fl/fl mice, with substantially reduced nuclear translocation of the AHR observed in cells isolated from SLC1A5^cko mice (Fig. 3F). Subsequently, these CD8⁺ T cells were treated with 14C-labeled kynurenine (Fig. 3G). Notably, the kynurenine uptake, as well as the levels of kynurenine metabolites (3-HK and KYNA) were substantially reduced in SLC1A5^cko-derived CD8⁺ T cells (Fig. 3H-I). Furthermore, treatment of these SLC1A5^cko CD8⁺ T cells with the AHR agonist 2 significantly increased T cell exhaustion (Fig. 3J-K), with notable suppression of IFN-γ, GZMB, and TNF-α production (Fig. 3L).

L-ky enhances chromatin accessibility in T cells

To investigate the epigenetic regulation of T cell function by L-ky, ATAC-seq analysis was performed on T cells following L-ky administration. The results demonstrated that the L-ky treatment markedly enhanced overall chromatin accessibility in T cells (Fig. 4A), indicating that L-ky may influence T cell fate by altering chromatin structure. Further analyses of the differentially accessible regions showed that these regions were enriched for binding sites of transcription factors linked to T cell activation and exhaustion, including BATF, TOX, NFATC1, and RUNX3 (Fig. 4B). Crucially, motif enrichment analysis also identified a significant enrichment of the AHR consensus binding sequence (XRE) within these L-ky-induced open chromatin regions (Figure S5). Furthermore, genes associated with these open regions showed a significant overlap with known AHR target genes, supporting the direct role of AHR in mediating this chromatin remodeling. Additionally, GO enrichment analysis of genes near these differentially accessible regions identified pathways linked to T cell exhaustion, cellular stress response, redox regulation, and metabolic adaptation (Fig. 4C).

Fig. 4.

L-ky enhances chromatin accessibility in T cells. A ATAC-seq analysis illustrating changes in chromatin accessibility in in vitro chronically activated CD8⁺ T cells following 48 h of L-ky treatment (100 uM). B Enrichment analysis of differentially accessible regions for genes significantly affected by L-ky treatment. C GO enrichment analysis of genes near differentially accessible regions. D-E ChIP-seq analysis using H3K27ac and H3K9ac antibodies to assess the occupancy of these modifications in CD8.⁺ T cells after L-ky treatment. F-G Distribution of H3K27ac and H3K9ac modifications across the genome following L-ky treatment. H Integrative Genomics Viewer analysis of H3K27ac and H3K9ac occupancy in the 2–3 kb upstream region of SLC1A5, PD-1, TIM3, and CTLA4 promoters. I ChIP analysis using H3K27ac and H3K9ac antibodies, followed by qPCR to quantify enrichment of these modifications at the SLC1A5, PD-1, TIM3, and CTLA4 promoters. Cell-based experiments repeated six times. Data are presented as dot plots and bars. Statistical significance was set at p < 0.05

To further investigate, ChIP-seq analysis was conducted using anti-H3K27ac and anti-H3K9ac antibodies. The analysis showed a significant increase in H3K27ac and H3K9ac enrichment within ± 3 kb of transcription start sites following L-ky treatment (Fig. 4D-E). Meanwhile, the global genomic distribution of H3K27ac and H3K9ac modifications remained largely comparable, with only a slight increase observed in the 2–3 kb region of promoters after L-ky treatment (Fig. 4F-G). However, a significant and specific enrichment of these active histone marks was observed at the promoters of key exhaustion-associated genes (Fig. 4H). These results were validated through ChIP experiments, which confirmed increased recruitment of H3K27ac and H3K9ac around 2–3 kb upstream of the SLC1A5, PD-1, TIM3, and CTLA4 promoters in T cells following L-ky treatment (Fig. 4I). These findings support the hypothesis that L-ky promotes the activity of critical transcription factors through AHR while also amplifying target gene expression by improving chromatin openness, thereby synergistically promoting T cell dysfunction.

HR recruits chromatin modifying enzyme FANCD2 to enhance SLC1A5 expression

ChIP experiments from Fig. 1L demonstrated the binding of AHR to the SLC1A5 promoter. Combined with findings from Fig. 4, it is hypothesized that AHR recruits regulatory enzymes to modulate chromatin accessibility, thus affecting SLC1A5 transcription. To investigate this, an APEX-AHR fusion protein was constructed and transfected into T cells. APEX2 is an engineered peroxidase that, in the presence of biotin-phenol and hydrogen peroxide, catalyzes the biotinylation of endogenous proteins within a ~ 20 nm radius, thereby allowing for the capture and identification of spatially proximal interacting partners [33] Biotin-phenol and H₂O₂ were then added, followed by streptavidin-mediated IP/MS to identify AHR-interacting proteins. Using the Biological General Repository for Interaction Datasets (BioGRID), a comprehensive public repository of known protein–protein interactions [34], chromatin-related proteins interacting with AHR were filtered, with FANCD2 identified as a promising target (Fig. 5A).

Fig. 5.

AHR recruits chromatin modifying enzyme FANCD2 to enhance SLC1A5 expression. A An APEX-AHR fusion protein was constructed and transfected into T cells. Biotin-phenol and H₂O₂ were then added, followed by streptavidin-mediated IP and mass spectrometry to identify AHR-interacting proteins, with chromatin-related proteins filtered using the BIOGRID database. B A GST-AHR fusion vector was constructed, loaded in T cells, and subjected to affinity chromatography using GSH. Enriched products were analyzed via SDS-PAGE. C-D HA-AHR and Flag-FANCD2 vectors were transfected into HEK293T cells. IP with anti-HA or anti-Flag was followed by WB analysis to detect Flag and HA expression. E HA-AHR was loaded in T cells treated with varying L-ky doses. IP with anti-HA antibody was followed by WB analysis to detect AHR and FANCD2 expression. F ChIP analysis using anti-FANCD2 assessed FANCD2 enrichment at the SLC1A5, PD-1, TIM3, and CTLA4 promoters via qPCR. G L-ky-treated T cells were treated with FEN1-IN-1. H-I qPCR and WB analysis of SLC1A5 mRNA and protein expression in T cells following L-ky and FEN1-IN-1 treatments. J ELISA analysis of IFN-γ, GZMB, and TNF-α secretion by T cells following L-ky and FEN1-IN-1 treatments. Cell-based experiments repeated six times. Data are presented as dot plots and bars. Statistical significance was set at p < 0.05

FANCD2, a core component of the Fanconi anemia (FA) pathway, is widely involved in DNA damage repair, chromatin stability, and transcriptional regulation [35]. Here, we hypothesized that FANCD2 collaborates with AHR in the epigenetic regulatory network. To further explore this interaction, a GST-AHR fusion vector was constructed, loaded in T cells, and subjected to affinity chromatography using GSH. SDS-PAGE analysis of the enriched products revealed a band corresponding to FANCD2 (Fig. 5B), providing preliminary evidence of an AHR-FANCD2 interaction.

Furthermore, HA-AHR and Flag-FANCD2 vectors were constructed and transfected into HEK293T cells. The Co-IP experiments confirmed the AHR-FANCD2 interaction (Fig. 5C-D). To assess whether L-ky influences this interaction, HA-AHR was loaded in T cells, which were then treated with varying doses of L-ky. Increased L-ky doses led to greater FANCD2 recruitment to AHR, as detected by anti-AHR IP (Fig. 5E). These results indicate that L-ky promotes nuclear translocation of AHR, which in turn recruits FANCD2, modulating the epigenetic state and contributing to T cell exhaustion.

ChIP analysis using anti-FANCD2 revealed FANCD2 recruitment to the 2–3 kb upstream regions of the SLC1A5, PD-1, TIM3, and CTLA4 promoters (Fig. 5F). To further investigate, L-ky-treated T cells were further treated with FEN1-IN-1, a FANCD2 inhibitor (Fig. 5G). This FEN1-IN-1 treatment markedly reduced SLC1A5 expression in T cells (Fig. 5H-I). The secretion of IFN-γ, GZMB, and TNF-α by T cells, initially suppressed by L-ky, were also restored following FEN1-IN-1 treatment (Fig. 5J).

Targeting FANCD2 restores T cell function

To further evaluate the role of FANCD2 in kynurenine-mediated T cell exhaustion, WT mice implanted with IDO1-overexpressing LLC cells were further treated with TNG463, a FANCD2 antagonist (Fig. 6A). Treatment with TNG463 substantially inhibited the growth rate of LLC cells in vivo (Fig. 6B-C). Additionally, the population of SLC1A5⁺CD8⁺ T cells in tumor tissues was considerably reduced following TNG463 treatment (Fig. 6D), accompanied by decreased populations in in TEX and TPEX cells (Fig. 6E). Conversely, the populations of GZMB⁺ and TNFα⁺ CD8⁺ T cells within the tumors were considerably increased following TNG463 treatment (Fig. 6F-G). Furthermore, IHC assays showed that the positive staining of C-Cas-3 in tumor tissues was enhanced under TNG463-mediated FANCD2 inhibition (Fig. 6H).

Fig. 6.

Targeting FANCD2 restores T cell function. A WT mice implanted with IDO1-overexpressing LLC cells were further treated with the FANCD2 antagonist TNG463. B Growth rate of LLC cells in vivo. C Diagram and weight of LLC-formed tumors on day 28. D FACS analysis of the number of SLC1A5⁺CD8⁺ T cells in tumor tissues. E FACS analysis of the number of TEX (PD1⁺TIM3⁺) and TPEX (PD1⁺TIM3⁻) cells in LLC-induced tumors. F-G FACS analysis of the number of GZMB⁺CD8⁺ (F) and TNFA⁺ CD8⁺ (G) T cells in tumor tissues. H IHC of C-Cas-3 staining in tumor tissues. I CD8⁺ T cells treated with L-ky alone or alongside TNG463 were co-cultured with LLC cells. J LDH release from LLC cells after co-culture with T cells, measured by LDH assay kit. K FACS analysis of CFSE.⁺ T cell counts. Animal experiments involved six mice per group, with cell-based experiments repeated six times. Data are presented as dot plots and bars. Statistical significance was set at p < 0.05

To confirm that targeting FANCD2 enhances T cell function, WT CD8⁺ T cells were treated with L-ky alone or alongside TNG463, which were then co-cultured with LLC cells (Fig. 6I). Notably, the L-ky treatment substantially suppressed T cell-mediated killing of LLC cells, but subsequent TNG463 treatment increased LDH release from LLC cells, indicating enhanced cytotoxicity (Fig. 6J). Additionally, TNG463 treatment significantly attenuated L-ky-induced inhibition of T cell proliferation (Fig. 6K). This evidence suggests that kynurenine induces T cell exhaustion through FANCD2 activation.

SLC1A5 knockout in T cells enhances the efficacy of anti-PD1 in mice

The findings above suggest that targeting SLC1A5 might improve the clinical efficacy of immunotherapy in LUAD. To test this, IDO1-overexpressing LLC cells were injected into SLC1A5^fl/fl or SLC1A5^CKO mice through the tail vein, followed by administration of either anti-PD-1 mAb or control IgG (Fig. 7A). As anticipated, the metastatic nodules formed by LLC cells in the lungs of SLC1A5^CKO mice were greatly reduced compared to SLC1A5^fl/fl mice (Fig. 7B). These SLC1A5^CKO mice also had longer survival period (Fig. 7C). Additionally, anti-PD-1 mAb treatment showed a greater therapeutic effect in SLC1A5^CKO mice as well (Fig. 7B-C).

Fig. 7.

SLC1A5 knockout in T cells enhances the therapeutic efficacy of anti-PD1 in mice. A IDO1-overexpressing LLC cells were injected into SLC1A5^fl/fl or SLC1A5^CKO mice through the tail vein, followed by treatment with anti-PD-1 monoclonal antibody (mAb) or control IgG. B Number of metastatic nodules formed by LLC cells in the lungs of mice. C Kaplan–Meier survival analysis of the mice. D GFP-IDO1-overexpressing LLC cells were implanted into C57BL/6 mice, followed by treatment with V-9302 (an SLC1A5 inhibitor), anti-PD-1 mAb, or a combination of both. E–F Analysis of bioluminescence intensity in the mice at 7, 14, and 28 d post-treatment. G-H. FACS analysis of the number of GZMB⁺CD8⁺ (G) and TNFA⁺ CD8⁺ (H) T cells in tumor tissues. I FACS analysis of the number of TEX (PD1⁺TIM3⁺) and TPEX (PD1⁺TIM3.⁻) cells in LLC-induced tumors. J Kaplan–Meier survival analysis of the mice. Animal experiments involved six mice per group. Data are presented as dot plots and bars. Statistical significance was set at p < 0.05

To further elucidate the effect of targeting SLC1A5 on anti-PD-1 mAb efficacy under standard conditions, GFP-IDO1-overexpressing LLC cells were implanted into C57BL/6 mice, followed by treatment with V-9302 (an SLC1A5 inhibitor), anti-PD-1 mAb, or a combination of both (Fig. 7D). Treatment with V-9302 restricted LLC cell growth in vivo and enhanced the tumor-suppressive effect of anti-PD-1 mAb (Fig. 7E-F). Furthermore, V-9302 treatment significantly increased the populations of GZMB⁺ and TNF-α⁺ CD8 T cells in lung metastatic nodules (Fig. 7G-H), while markedly reducing TEX and TPEX cell populations (Fig. 7I). V-9302 showed synergistic effects with anti-PD-1 mAb in improving immune activity in mice (Fig. 7G-I). Still, the survival of mice was prolonged by the treatment of V-9302 or anti-PD-1 mAb, particularly by the combination regimen.

Discussion

The uptake or release of metabolites within the TME significantly alters the behavior and functionality of its components, impacting tumor progression [36]. This study reveals that the amino acid transporter SLC1A5 enhances L-ky metabolism, activating the AHR. This activation collaborates with the chromatin-modifying enzyme FANCD2 to increase chromatin accessibility in T cells, upregulating SLC1A5 expression. This process forms a self-reinforcing cycle that drives CD8⁺ T cell exhaustion.

Prolonged antigen exposure and chronic inflammation in the TME can exhaust CD8⁺ T cells, transforming them into dysfunctional exhausted T cells [14]. These cells display increased inhibitory receptor expression, reduced self-renewal, and a gradual loss of effector functions, rendering them ineffective against tumors [37]. Exhausted CD8⁺ T cells are categorized into progenitor TPEX (TIM3-) and terminal TEX (TIM3⁺), each with distinct properties [38]. TPEX cells remain responsive to immune checkpoint inhibitors, aiding tumor control, while TEX cells, with limited proliferative capacity, resist most immunotherapies [39–42]. In our experiments, chronic stimulation of mouse CD8⁺ T cells induced TPEX and TEX subsets, with TEX showing significantly elevated SLC1A5 expression, suggesting its role in T cell exhaustion.

Prior studies have linked SLC1A5 to energy metabolism and tumor development [25, 43, 44]. Its association with immunosuppression is also evident, as seen in breast cancer, where SLC1A5 overexpression correlates with Foxp3⁺ Treg cells and poor survival [45]. Conversely, some evidence suggests that boosting SLC1A5 expression may enhance anti-tumor immunity by improving nutrient uptake in cytotoxic T cells or natural killer cells [46]. A bioinformatics and cell-based study by Zheng et al. linked SLC1A5 expression to CD8⁺ T cell exclusion in squamous cell carcinomas, including lung squamous cell carcinoma [47]. Our analysis of TCGA-LUAD and GEO datasets further confirms that SLC1A5 expression correlates with PD-1, CTLA4, and TIM3 expression, T cell exhaustion, and worse prognosis in LUAD.

Further investigation revealed that L-ky supplementation dose-dependently upregulated SLC1A5 expression in T cells. As a key transporter, SLC1A5 regulates tryptophan transport and kynurenine metabolism [27, 48]. Notably, tumorigenic activity in IDO1-overexpressing LLC cells was significantly reduced in SLC1A5^−/− and SLC1A5^cko mice, accompanied by enhanced intratumoral T cell activity. Kynurenine pathway activation requires AHR nuclear translocation, which forms heterodimers with the aryl hydrocarbon receptor nuclear translocator to drive downstream gene transcription [49, 50]. In SLC1A5^cko mice, CD8⁺ T cells showed reduced AHR nuclear translocation, and an AHR agonist (AHR agonist 2) promoted T cell exhaustion while decreasing cytotoxic cytokine production. These findings highlight SLC1A5’s critical role in kynurenine metabolism, AHR activation, and T cell exhaustion.

Furthermore, ATAC-seq analysis showed that L-ky treatment significantly increased chromatin accessibility in CD8⁺ T cells, with heightened H3K27ac and H3K9ac enrichment—markers of open chromatin and active transcription [51]—in the 2–3 kb promoter regions of SLC1A5, PD-1, TIM3, and CTLA4. This suggests that L-ky influences T cell fate by modifying chromatin structure. Additionally, AHR was found to bind the SLC1A5 promoter, indicating a positive feedback loop involving kynurenine, AHR, and SLC1A5. Using an APEX-AHR fusion protein and IP/MS, we identified FANCD2, a chromatin-modifying enzyme, as an AHR partner in regulating SLC1A5 transcription. FANCD2, a key component of the FA pathway, supports DNA damage repair through ubiquitination of the FANCD2-FANCI heterodimer, which facilitates DNA crosslinking and homologous recombination [52–54]. FA pathway defects, particularly in FANCD2, promote genomic instability and tumorigenesis [55]. A pan-cancer analysis by Xie et al. revealed frequent FANCD2 upregulation in solid tumors, correlating with immune infiltration, immune checkpoints, and tumor mutation burden [56]. In LUAD, Ye et al. linked FANCD2 expression to higher tumor immune dysfunction scores and poor prognosis [57]. Our study found that AHR recruits FANCD2 to the nucleus, enhancing chromatin accessibility and SLC1A5 transcription. Treatment with the FANCD2 inhibitor TNG463 reduced T cell exhaustion and tumorigenesis in mice with IDO1-overexpressing LLC cells. While our study primarily utilized TIM-3 as a marker to delineate T cell subsets, we acknowledge that the TPEX population is more comprehensively defined by markers such as TCF-1 and Slamf6. However, previous research in LLC models has consistently shown that the PD-1 + TIM-3- population corresponds to the TCF-1 + progenitor pool that sustains the anti-tumor response. Our findings regarding SLC1A5 upregulation specifically in the TIM3 + terminal subset suggest that kynurenine metabolism is a key driver of the transition toward terminal dysfunction.

Further in vivo experiments demonstrated that SLC1A5 loss synergized with anti-PD-1 mAb in mice implanted with LLC cells via the tail vein. These findings collectively suggest that the kynurenine-AHR-FANCD2 axis increases chromatin accessibility and SLC1A5 transcription in CD8⁺ T cells, amplifying kynurenine metabolism and forming a feedback loop that drives T cell exhaustion. Targeting SLC1A5 could thus offer a promising approach to mitigate T cell exhaustion and enhance immunotherapy efficacy in LUAD.

Authors’ contributions

Conceptualization: Xinnan Xu, Li Xu, Hongdou Ding; Methodology: Yifei Zhu, Yanxi Yao and Haoji Wang Investigation: Yifei Zhu and Yanxi Yao Visualization: Yanxi Yao and Haoji Wang Funding Acquisition: Xinnan Xu Project Administration: Li Xu Supervision: Li Xu, Hongdou Ding, Xinnan Xu Writing – Original Draft: Yifei Zhu, Yanxi Yao and Haoji Wang Writing – Review & Editing: Li Xu, Hongdou Ding and Xinnan Xu Resources: Yanxi Yao, Haoji Wang and Yifei Zhu.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

All animal studies have been approved by the Ethics Committee of Shanghai Pulmonary Hospital, Tongji University School of Medicine (Permit number: K25-263Y) and performed in accordance with the ethical standards.

Competing interests

The authors declare no competing interests.