Kynurenine promotes angiogenesis through mTOR signaling in head and neck squamous cell carcinoma

Abstract

Angiogenesis and its maturation play a crucial role in the progression and prognosis of head and neck squamous cell carcinoma (HNSCC). Tumor growth relies on an adequate blood supply to meet metabolic demands, while metabolic reprogramming within the tumor microenvironment actively reshapes vascular architecture. Meanwhile, the tumor microenvironment induces metabolic reprogramming that reshapes the vasculature. Alterations in tryptophan metabolism have been observed in HNSCC, with kynurenine (Kyn) significantly elevated in cancer tissues. Although Kyn has been reported to regulate vascular function in several diseases, its role in angiogenesis in HNSCC remains unclear. Metabolic profiling assessed alterations in tryptophan metabolism in HNSCC. Transcriptomic data from the TCGA HNSCC cohort were analyzed to examine correlations between IDO1 expression and angiogenesis-related genes. These findings were validated in tumor specimens from 28 HNSCC patients via immunohistochemistry. Based on these clinical observations, functional and mechanistic studies were conducted using human umbilical vein endothelial cells (HUVECs) and the chick embryo chorioallantoic membrane (CAM) model to evaluate the effects of Kyn on endothelial migration, angiogenic capacity, and neovascularization. A subcutaneous cell-derived xenograft (CDX) mouse model was further employed to assess the association between IDO1 expression and vascular markers in vivo. The involvement of IDO1-mediated Kyn production and mTOR signaling activation was further investigated. Tryptophan metabolism was markedly reprogrammed in HNSCC, with significantly elevated levels of Kyn in tumor tissues. Analysis of the TCGA HNSCC cohort revealed that IDO1, the key enzyme responsible for kynurenine production, was positively correlated with multiple angiogenesis-related genes. Immunohistochemical analysis revealed a positive correlation between IDO1 expression and endothelial markers CD31 and CD34 in patient samples, suggesting a potential link between Kyn and tumor angiogenesis. Consistently, Kyn enhanced endothelial cell migration and tube formation in vitro and promoted neovascularization in the CAM model. In line with these findings, analysis of tumor tissues from the CDX mouse model demonstrated positive correlations between IDO1 expression and multiple vascular markers, including CD31, CD34, and CD105. Mechanistically, IDO1-driven Kyn production activated the mTOR pathway in endothelial cells, leading to the upregulation of angiogenesis-related genes, pro-angiogenic factors, and their receptors. These findings reveal a novel pro-angiogenic role of Kyn in HNSCC, mediated through IDO1-dependent tryptophan metabolism and mTOR pathway activation. Targeting tryptophan metabolism may provide a potential therapeutic strategy for highly vascularized HNSCC.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-41141-x.

Introduction

According to the Global Cancer Observatory (GLOBOCAN), head and neck squamous cell carcinoma (HNSCC) ranked as the seventh most common cancer worldwide in 2020. Approximately 740,000 new cases of lip/oral, laryngeal, oropharyngeal, and hypopharyngeal cancers were reported globally, with around 360,000 deaths attributed to these malignancies¹. Due to the anatomical and physiological characteristics of the head and neck, HNSCC often presents with distant metastases and is prone to significant functional impairment, severely affecting aesthetics and quality of life². Currently, treatment for HNSCC primarily relies on a multimodal approach, with surgery being the cornerstone, often supplemented by chemotherapy and radiotherapy. Despite advances in technology, the five-year survival rate for HNSCC patients has shown limited improvement³. Consequently, there is an urgent need to explore more effective therapeutic strategies to address this unmet clinical need. Since the approval of the chimeric monoclonal antibody cetuximab combined with radiotherapy for the treatment of locally or regionally advanced metastatic tumors by the U.S. Food and Drug Administration in 2006, targeted therapies have gained significant attention due to their high selectivity and low toxicity⁴. Increasing evidence suggests that targeted therapies, when used in conjunction with other treatment modalities, have the potential to enhance survival rates and improve the quality of life for cancer patients⁵.

To sustain continuous proliferation and growth, tumor cells must undergo metabolic adaptations to balance their bioenergetic and biosynthetic demands, a process referred to as metabolic reprogramming in cancer⁶. Metabolic reprogramming is a hallmark of cancer, but it is not merely a biological feature—it also reveals therapeutic vulnerabilities. As tumor cells become dependent on specific metabolic pathways, the vulnerabilities induced by metabolic rewiring may present opportunities for targeted treatment⁷. Numerous metabolic molecules have shown promise as potential therapeutic targets in preclinical studies aimed at blocking tumor progression^8–11. A better understanding of the metabolic reprogramming in head and neck squamous cell carcinoma could enable the targeting of specific metabolic pathways to improve the precision of cancer therapies. In HNSCC, tryptophan metabolism is altered, leading to increased levels of kynurenine (Kyn). Our previous research indicates that tryptophan metabolism is reprogrammed in HNSCC, resulting in elevated levels of the metabolite Kyn in tumor tissues. Since IDO1 catalyzes the production of Kyn, using an IDO1 inhibitor can decrease Kyn levels, making it a potential target for therapy. Targeting tryptophan metabolism may, therefore, offer an effective approach to improving treatment outcomes for HNSCC. However, the failure of the IDO1 inhibitor in combination with immune checkpoint inhibitors in phase III clinical trials for melanoma¹² has prompted us to reconsider alternative combination therapies.

During tumorigenesis and progression, complex interactions between tumor cells and the surrounding tumor microenvironment (TME) play a critical role in shaping tumor biology. Increasing evidence suggests that the TME significantly influences the biological behavior of tumor cells¹³. The formation of new blood vessels within the TME has long been considered a fundamental aspect of the progression of many solid tumors^14,15. Tumor cells rely on the vascular system to supply oxygen and nutrients, while also assisting in the removal of metabolic waste and carbon dioxide. In early-stage tumors (diameter < 3 mm), the tumor primarily acquires nutrients via simple diffusion from adjacent tissues. As the tumor grows, however, the demand for oxygen and nutrients increases to sustain survival and proliferation, making the tumor’s growth increasingly dependent on vascular density. Continuous tumor growth necessitates sustained angiogenesis, and thus, the induction of angiogenesis has been recognized as one of the hallmarks of cancer¹⁶. In 1996, Hanahan et al.¹⁷ proposed the “angiogenic switch” hypothesis, which posits that the normally quiescent vasculature can be activated to form new capillaries. This process is regulated by the angiogenic switch mechanism, where the balance between pro-angiogenic and anti-angiogenic factors determines vascular formation. An imbalance in the expression of these factors in the tumor and surrounding tissues can trigger angiogenesis¹⁸. As the importance of tumor vasculature is increasingly recognized, targeting angiogenesis to “starve” tumors has emerged as a promising therapeutic strategy. By blocking the formation of new blood vessels, this approach aims to hinder or delay tumor growth and metastasis, becoming a focal point of recent cancer research. A study by Yang et al.¹⁹ demonstrated that overexpression of IDO1 in human lung adenocarcinoma A549 cells significantly enhanced HUVEC proliferation and capillary-like structure formation during co-culture, highlighting a role of IDO1 in tumor-associated angiogenesis. Similarly, Pan et al.²⁰ reported that silencing IDO1 in Lewis lung cancer cells markedly suppressed tumor cell migration, invasion, and vessel mimicry in vivo. Beyond cancer, accumulating evidence indicates that cellular metabolic status is a critical determinant of angiogenic capacity. In diabetic wounds, mitochondrial dysfunction has been shown to impair angiogenesis and delay tissue repair, underscoring the fundamental role of metabolic reprogramming in vascular regeneration²¹. In the context of cancer, tumor-specific metabolic or hypoxia-related regulators can also drive angiogenesis in a disease-dependent manner. For instance, GP73 was reported to promote angiogenesis in hepatocellular carcinoma by stabilizing HIF-1α and enhancing VEGFA production, as well as by activating mitogenic signaling in endothelial cells²². Despite these advances, whether tryptophan metabolic reprogramming—particularly through its key intermediate kynurenine—drives angiogenesis in head and neck squamous cell carcinoma remains largely unexplored. In the present study, altered tryptophan metabolism and elevated kynurenine levels were observed in HNSCC tissues, and immunohistochemical analysis of patient samples revealed a positive association between IDO1 expression and vascular markers, suggesting a potential role of kynurenine in tumor angiogenesis.

Guided by these tissue-level observations, we investigated the effects of kynurenine on endothelial cell behavior using human umbilical vein endothelial cells and the chick embryo chorioallantoic membrane model, which demonstrated that kynurenine enhances endothelial migration and angiogenic activity. Consistent with these findings, a subcutaneous xenograft mouse model further revealed positive correlations between IDO1 expression and multiple vascular markers (CD31, CD34, and CD105) within tumor tissues, providing in vivo evidence supporting the association between IDO1-driven tryptophan metabolism and tumor vascularization. Further analyses suggest that IDO1-dependent kynurenine production is associated with activation of the mTOR signaling pathway, leading to increased expression of angiogenesis-related genes, pro-angiogenic factors, and their receptors. These findings support a link between tryptophan metabolic reprogramming and vascular remodeling in HNSCC and provide a rationale for exploring metabolic modulation in combination with anti-angiogenic strategies.

Materials and methods

Obtaining tissue samples

Fresh samples of squamous cell carcinoma of the head and neck as well as normal tissues next to the cancer were acquired from Fujian Medical University’s First Affiliated Hospital. Each patient was informed about the conduct and gave written informed consent in accordance with the Helsinki Declaration’s requirements. The study was approved by the Ethics Committee of the Fujian Medical University School and Hospital of Stomatology (No. 2021-FMUSS-024).

Liquid chromatography-tandem mass spectrometry(LC-MS/MS)

Place 100 mg of a tissue sample in an EP tube after it has been crushed with liquid nitrogen. Before allowing it to sit in an ice bath for five minutes, add 500 µL of an 80% methanol aqueous solution with 0.1% formic acid, shake, and vortex. Centrifuge the mixture at 15,000 rpm and 4℃ for 10 min, and take an appropriate amount of the supernatant and dilute it with mass spectrometry grade water to a methanol concentration of 53%. After putting the mixture in a centrifuge tube and centrifuging it for 10 min at 15,000 g and 4 °C, gather the supernatant and add it to an LC-MS/MS for analysis.

Animals

All experimental protocols were approved by the Ethics Committee of the Fujian Medical University School and Hospital of Stomatology (approval no. 2021-FMUSS-024). All procedures were performed in accordance with the Chinese Regulations for the Administration of Laboratory Animals and institutional guidelines. Mice were housed in the SPF barrier facility of Fujian Medical University, which holds the Laboratory Animal Use Permit [SYXK(Min)2022-0005]. All methods are reported in accordance with the ARRIVE guidelines.

BALB/c nude mice (6-week-old, approximately 25 g at purchase) were obtained from GemPharmatech Co., Ltd. [SCXK(Yue)2023-0067]. IDO1-overexpressing Cal-27 cells were generated by transfecting Cal-27 cells with an IDO1 overexpression plasmid, followed by selection and expansion of stable clones. Control cells were transfected with an empty vector. For tumor induction, mice were randomly assigned to two groups and injected subcutaneously into the right flank with 5 × 10⁶ IDO1-overexpressing or control (empty-vector) Cal-27 cells suspended in 100 µl PBS under 2% isoflurane inhalation anesthesia. This established a Cell line-Derived Xenograft (CDX) model, which is an in vivo tumor model using human cancer cell lines in immunodeficient mice to study tumor growth and angiogenesis. When tumors reached approximately 100 mm³ (diameter ≈ 0.58 cm), the mice weighed 26.2 ± 1.6 g, and tumor size was measured with calipers every two days. The study endpoint was reached when tumor volume approached 1,000 mm³ (diameter ≈ 1.24 cm), at which point mice were euthanized by cervical dislocation.

Cell culture of human umbilical vein endothelial cells (HUVECs)

Human umbilical vein endothelial cells (HUVECs) were purchased from [iCell Bioscience, China]. Cells were cultured in endothelial cell growth medium (ECM, ScienCell) supplemented with 2% fetal bovine serum, vascular endothelial growth factor, and antibiotics (penicillin/streptomycin), and maintained at 37 °C in a humidified atmosphere containing 5% CO₂. Cells were subcultured using 0.25% trypsin-EDTA upon reaching 80–90% confluence. HUVECs at passages 3–10 were used for all experiments.

Reagents and treatments

L-kynurenine, a tryptophan metabolite, was purchased from Sigma-Aldrich (St. Louis, MO, USA) and dissolved in DMSO. The selective IDO1 inhibitor epacadostat was obtained from Selleck Chemicals (Shanghai, China) and prepared according to the manufacturer’s instructions. The concentrations and treatment durations used in each experiment are specified in the corresponding figure legends.

In vitro tube formation assay

To assess angiogenesis, HUVECs (1 × 10⁴ cells/well) were seeded onto growth factor-reduced Matrigel (Corning)-coated 96-well plates (100 µL/well) and incubated for 6–8 h at 37 °C. Tubular structures were imaged using an inverted phase-contrast microscope (Nikon). Network parameters (total Vessel’s length, branch points) were quantified using ImageJ with the Angiogenesis Analyzer plugin. Experiments were performed in triplicate, with three independent biological replicates.

Chick chorioallantoic membrane (CAM) assay

The eggs were first wiped with 1‰ benzalkonium bromide and then disinfected with 75% alcohol. They were randomly divided into a control group and an experimental group, with 10 eggs in each group. They were incubated in an electric heating incubator at 37.8℃, and a water tray was placed in the incubator to maintain a relative humidity of 50% − 70%. The eggs were turned 2–4 times a day to prevent embryo adhesion. The development of the embryos was checked using a concentrated light source, and dead or weak eggs will be removed. On the 8th day of egg development, well-developed eggs were selected to prepare the drug injection holes. The surface of the air cell (the large end) was disinfected with alcohol in a 2 cm × 3 cm area, and two holes were carefully poked with pointed forceps. The drug solution was prepared. The experimental group was physiological saline containing 1 µmol/L Kyn, and the control group NC was physiological saline without any metabolites. Starting from the 8th day of egg development, the drug solution was injected into the drug injection holes every 12 h using a disposable 1 ml syringe. On the 12th day of egg development, the eggshell of the air cell was removed, and the air cell membrane was gently peeled off to completely expose the chorioallantoic membrane of the chicken embryo. The embryo was observed and photographed under a stereomicroscope.

HUVEC-tumor cell adhesion assay

HUVEC adhesion to tumor cells was assessed using Cal-27 cells, a human tongue squamous cell carcinoma line. Briefly, Cal-27 cells were seeded in 24-well plates and cultured until 70–80% confluence. HUVECs (1 × 10⁵ cells/well) were pretreated with L-kynurenine (Sigma, St. Louis, MO, USA) or the IDO1 inhibitor epacadostat (Selleck, China) for 12 h, then added to Cal-27 monolayers and incubated for 2 h at 37 °C. Non-adherent HUVECs were removed by gentle washing with PBS. Adherent cells were stained with calcein-AM, and imaged under an inverted microscope (Nikon). Adhesion was quantified by counting adherent HUVECs in five randomly selected fields per well. Experiments were performed in triplicate with three independent biological replicates.

Immunohistochemistry (IHC)

Tumor xenografts from CDX mice and human tissue specimens were subjected to immunohistochemical analysis. Paraffin-embedded sections were deparaffinized, rehydrated, and subjected to antigen retrieval using either citrate buffer or EDTA buffer, as appropriate for each antibody. For CDX tumor sections, immunohistochemical staining was performed using antibodies against CD31 (Abcam, 1:200, citrate retrieval), CD34 (Boster Biological Technology, China, 1:200, EDTA retrieval), CD105 (Boster Biological Technology, China, 1:200, citrate retrieval), and α-smooth muscle actin (α-SMA; Abcam, 1:200, citrate retrieval) to evaluate tumor angiogenesis and vascular maturation. For human tumor tissue sections, immunohistochemical staining was performed using antibodies against CD31 (Abcam, 1:200, citrate retrieval), indoleamine 2,3-dioxygenase 1 (IDO1; Cell Signaling Technology, 1:300, EDTA retrieval), and CD34 (Boster Biological Technology, China, 1:200, EDTA retrieval). All sections were incubated with primary antibodies overnight at 4 °C, followed by incubation with appropriate secondary antibodies and visualization using standard diaminobenzidine (DAB) detection protocols. Hematoxylin was used for nuclear counterstaining. Stained sections were imaged under a light microscope, and immunoreactivity was evaluated according to established criteria.

Western blotting

HUVECs were lysed in RIPA buffer containing protease inhibitors. Protein concentrations were determined using a BCA assay (Thermo Fisher Scientific). Equal amounts of protein (20–30 µg) were separated by SDS–PAGE, transferred onto PVDF membranes, and blocked with 5% non-fat milk at room temperature. The blots were cut into 2 or 3 parts prior to incubation with primary antibodies when the expected bands, estimated by the molecular weight ladder and the manufacturer’s instructions of primary antibodies, were separated enough. Membranes were incubated overnight at 4 °C with primary antibodies, including anti-Hsp90 (1:1000), anti-AhR (1:1000; Protein tech), anti-IDO1, anti-phospho-mTOR, anti-mTOR, anti-phospho-AKT, and anti-AKT (all 1:1000; Cell Signaling Technology), followed by HRP-conjugated secondary antibodies (1:5000, 1 h). For nuclear–cytoplasmic fractionation, HUVECs were processed using a nuclear and cytoplasmic protein extraction protocol according to the manufacturer’s instructions. Briefly, cells were harvested, resuspended in cytoplasmic extraction buffer, and centrifuged to separate the cytoplasmic fraction. The remaining pellets were further lysed in nuclear extraction buffer to obtain nuclear proteins. Equal amounts of nuclear and cytoplasmic proteins were subjected to Western blot analysis as described above.

Reverse transcription quantitative PCR (RT-qPCR)

Total RNA was extracted using Trizol (Invitrogen), reverse-transcribed into cDNA (PrimeScript RT kit, Takara), and amplified using SYBR Green Master Mix (Roche) on a StepOnePlus system (Applied Biosystems). Primer sequences (table) were designed via NCBI Primer-BLAST. Relative gene expression was calculated using the 2^^−ΔΔCt method normalized to GAPDH.

ANG2-F(Homo)	5’-AACTTTCGGAAGAGCATGGAC-3’
ANG2-R(Homo)	5’-CGAGTCATCGTATTCGAGCGG-3’
EGF-F(Homo)	5’-GAGGCTTGGAACTGCTTGGT-3’
EGF-R(Homo)	5’-ATCAACTTCACCACCTGCCAA-3’
EGFR-F(Homo)	5’-TTGCCGCAAAGTGTGTAACG-3’
EGFR-R(Homo)	5’-GTCACCCCTAAATGCCACCG-3’
GAPDH-F(Homo)	5’-GCACCTGTTGGTTCATTGGTAA-3’
GAPDH-R(Homo)	5’-GAGTCCTTCCACGATACCAAAG-3’
HIF-1-F(Homo)	5’-GCAGCAACGACACAGAAACT-3’
HIF-1-R(Homo)	5’-TGCAGGGTCAGCACTACTTC-3’
PDGFB-F(Homo)	5’-CAAACTCGGGTGACCATTCG-3’
PDGFB-R(Homo)	5’-TCTCCTTCAGTGCCGTCTTG-3’
PTEN-F(Homo)	5’-TGCAGTATAGAGCGTGCAGA-3’
PTEN-R(Homo)	5’-TAGCCTCTGGATTTGACGGC-3’
VEGFA-F(Homo)	5’-ATCGAGTACATCTTCAAGCCAT-3’
VEGFA-R(Homo)	5’-GTGAGGTTTGATCCGCATAATC-3’
VEGFR-F(Homo)	5’-ACAGCAGGAATCAGTCAGTATCT-3’
VEGFR-R(Homo)	5’-TGTCCGTCTGGTTGTCATCT-3’

Bioinformatic analysis

Transcriptomic data of HNSCC patients were obtained from The Cancer Genome Atlas (TCGA) database. Gene expression levels of IDO1 and endothelial markers (PECAM1/CD31, ENG/CD105, CD34, VEGFC) as well as the pericyte marker ACTA2/α-SMA were extracted. Pearson correlation analysis was performed to assess the relationship between IDO1 and these vascular-related genes.

Results

Metabolomic analysis revealed a special metabolic characteristic in HNSCC

To examine the metabolic regulatory network in head and neck squamous cell carcinoma and the related metabolic changes, we collected and analyzed ten pairs of squamous cell carcinoma and adjacent non-cancerous tissue samples using LC-MS/MS. To predict the sample category, a relationship model between metabolite expression and sample category was established using partial least squares discrimination analysis (PLS-DA). In the positive ionization mode PLS-DA model, R2 values 0.96 and Q2 values 0.67, which indicates that the model is dependable and stable (Fig. 1A). In the case of negative ionization mode PLS-DA model, R2 values 0.87 and Q2 values 0.63 (Figure S1A).

Fig. 1.

Reprogramming of tryptophan metabolism in HNSCC. (A) partial least squares discriminant analysis (PLS-DA) in positive ion pattern, (B) volcano maps of differential metabolites in the positive ion pattern, (C) clustering analysis of differential metabolites in the positive ion pattern, (D) positive ion pattern KEGG-enriched bubble maps^23–25, (E) relative expression of Kyn in cancerous tissues of HNSCC and para-cancerous tissues in LC-MS/MS, n = 10. The analysis of the experimental group and the control group was conducted using paired t-tests.

The metabolite volcano plot demonstrates that differential metabolites in 83 are elevated in cancer tissues and 46 metabolites are down-regulated in the positive ion mode (Fig. 1B). Meanwhile, 18 down-regulated differential metabolites and 32 up-regulated metabolites were found in negative ion mode (Figure S1B).

To assess the possible role of these differentially expressed metabolites, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis and clustering algorithms were used (Fig. 1C,D and Figure S1C-D). It can be concluded from KEGG results that tryptophan metabolism undergoes reprogramming in HNSCC.

Our study found that metabolic intermediate Kyn was higher in HNSCC than in tumor-adjacent tissue, through using enrichment analysis of tryptophan metabolism, and the results were statistically significant (Fig. 1E).

The results mentioned above demonstrated that the metabolic intermediate Kyn was increased in HNSCC and that tryptophan metabolism was reprogrammed in the disease.

IDO1 expression is positively associated with tumor angiogenesis in HNSCC tissues

Given that IDO1 is the rate-limiting enzyme for kynurenine production, we first examined whether activation of the IDO1–Kyn metabolic axis is associated with angiogenesis-related signatures in HNSCC. Analysis of the TCGA HNSCC cohort showed that IDO1 expression was positively correlated with multiple endothelial- and angiogenesis-associated genes, including PECAM1 (CD31), CD34, ENG (CD105), and the pro-angiogenic factor VEGFC, whereas no significant correlation was observed with the pericyte marker ACTA2 (α-SMA) (Fig. 2A).

Fig. 2.

Association between IDO1 expression and angiogenesis-related features in HNSCC. (A) Heatmap showing correlations between IDO1 expression and angiogenesis-related genes in the TCGA HNSCC cohort, including endothelial markers (PECAM1/CD31, CD34, ENG/CD105) and the pro-angiogenic factor VEGFC, while no significant correlation was observed with the pericyte marker ACTA2 (α-SMA). (B) Representative immunohistochemical (IHC) staining of IDO1, CD34, and CD31 in tumor tissues from 28 HNSCC patients, illustrating low and high IDO1 expression groups (scale bar: see image). (C) Correlation analysis between IDO1 and CD34 expression in HNSCC tumor samples (Spearman’s correlation). (D) Correlation analysis between IDO1 and CD31 expression in HNSCC tumor samples (Spearman’s correlation).

These correlative findings raised the possibility that enhanced tryptophan metabolism toward Kyn may be linked to angiogenic programs in HNSCC, thereby providing a rationale for subsequent functional analyses to determine whether Kyn directly modulates endothelial angiogenic behavior. To validate these bioinformatic findings, immunohistochemical analyses were performed on tumor tissues from 28 HNSCC patients. Based on IDO1 staining intensity, samples were stratified into IDO1-low and IDO1-high expression groups. Representative IHC images showed that tumors with high IDO1 expression exhibited increased staining of CD34 and CD31 compared with those displaying low IDO1 expression (Fig. 2B). Consistently, correlation analyses demonstrated that IDO1 expression was positively associated with CD34-positive microvessel density (Fig. 2C) and CD31 expression (Fig. 2D). These results indicate a preferential association between elevated IDO1 expression, reflecting increased Kyn production, and enhanced angiogenic activity.

Regulation of vascular endothelial cell angiogenic capacity by Kyn

The effect of Kyn on HUVEC proliferation was first evaluated. Neither addition of Kyn (Fig. 3A) nor reduction via the IDO1 inhibitor epacadostat (Fig. 3B) significantly altered cell proliferation. Next, the impact on endothelial cell migration was assessed. Kyn enhanced migration in both wound healing (Fig. 3C) and Transwell assays (Fig. 3D), whereas epacadostat inhibited migration in the wound healing (Fig. 3E) and Transwell assays (Fig. 3F). The adhesion of HUVECs to Cal-27 tumor cells was then examined. Kyn treatment increased adhesion (Fig. 3G), while epacadostat decreased adhesion (Fig. 3H). Finally, vascular tube formation was assessed in Matrigel. Kyn promoted tube formation as indicated by increased node numbers and branch lengths (Fig. 3I), while epacadostat suppressed these effects (Fig. 3J).

Fig. 3.

Regulatory effects of Kyn on vascular endothelial cells in HNSCC. (A,B) The effect of Kyn on endothelial cell proliferation was observed by adding Kyn (A) or using the IDO1 inhibitor epacadostat (B), (C–F) Effect of Kyn or the IDO1 inhibitor epacadostat on HUVEC migration: (C) (Kyn) & (E) (Epacadostat): Wound healing assay; migrated area was quantified as the percentage of closure relative to the initial scratch. (D) (Kyn) & (F) (Epacadostat): Transwell migration assay; the number of cells migrated per field (unit area) was quantified, Scale Bar, 200 μm, (G,H) The effect of Kyn on the adhesion of HUVECs to tumor cells (Cal-27) was observed by adding Kyn (G) or using the IDO1 inhibitor epacadostat (H), Scale Bar, 200 μm; (I,J) The effect of Kyn on vascular tube formation was assessed by adding Kyn (I) or using the IDO1 inhibitor epacadostat (J), Scale Bar: 50 μm. Data are presented as mean ± SD from three independent biological replicates (n = 3), each with two technical replicates. Statistical significance was determined using a paired t-test. The data are represented as the means ± SEMs from dependent duplicates. ns: P > 0.05, *: P < 0.05, **: P < 0.01, ***:P < 0.001.

In vivo effects of Kyn and IDO1 overexpression on tumor angiogenesis

To investigate the in vivo effects of Kyn on angiogenesis, we first utilized the CAM model. Administration of Kyn resulted in increased total vessel length and branching points, indicating enhanced neovascularization compared with the control group (Fig. 4A–C).

Fig. 4.

In vivo effects of Kyn on angiogenesis and tumor progression. (A–C) The effect of Kyn on angiogenesis was assessed using the chorioallantoic membrane (CAM) vascularization model in chicken embryos, n = 5, Scale Bar: 50 μm, (D). Representative immunohistochemical staining of tumor xenografts for endothelial markers CD31, CD34, CD105, and pericyte marker α-SMA in control and IDO1-overexpressing groups. Scale bars as indicated in the images. (E) Quantification of the immunohistochemical staining for CD31, CD34, CD105, and α-SMA. (F) Tumor weight and mice weight of CDX model mice, (G) Immunohistochemistry of tumors stripped on the CDX model, indexes include IDO1 and Ki67 n = 5. The analysis of the experimental group and the control group was conducted using paired t-tests. The data are represented as the means ± SEMs from dependent duplicates. *:P < 0.05, ***:P < 0.001.

Next, IDO1-overexpressing Cal-27 cells and control cells were subcutaneously injected into nude mice to establish a CDX model. Immunohistochemical analysis of the resulting tumors revealed higher expression levels of endothelial cell markers CD31, CD34, and CD105 in the IDO1-overexpressing group, whereas α-SMA levels remained unchanged (Fig. 4D,E). Finally, analysis of tumor growth showed that tumors derived from IDO1-overexpressing cells were significantly larger and grew faster than those from control cells (Fig. 4F,G). These results collectively indicate that Kyn and IDO1 overexpression are associated with enhanced angiogenic features and accelerated tumor growth in vivo.

Kyn enhances expression of angiogenesis-related genes and factors in vascular endothelial cells

To delve deeper into the specific role of Kyn in vascular regulation, we further evaluated the transcription levels of angiogenesis-related genes, pro-angiogenic factors, and receptors in vascular endothelial cells under the influence of Kyn. The results indicate that following Kyn treatment, the transcription levels of HIF1A, EGF, EGFR, PDGFB, PTEN, and ANG2 in endothelial cells were elevated compared to the control group (Fig. 5A). Subsequent treatment with the IDO1 inhibitor epacadostat led to decreased transcription levels of EGF, PDGFB, HIF1, and EGFR (Fig. 5B). These findings were corroborated on the chorioallantoic membrane of chicken embryos, showing similar results (Fig. 5C).

Fig. 5.

Transcriptional regulation of angiogenesis-related markers by Kyn (A,B). Changes in angiogenesis-related markers were assessed after adding Kyn (A) and the IDO1 inhibitor epacadostat (B), (C) Changes in angiogenesis-related markers were assessed in the Kyn-treated chorioallantoic membrane. The analysis of the experimental group and the control group was conducted using paired t-tests. The data are represented as the means ± SEMs from dependent duplicates. *:P < 0.05, **: P < 0.01, ***:P < 0.001.

Kyn activates the mTOR signaling pathway to regulate endothelial cell angiogenic capacity

To explore the molecular mechanisms underlying the pro-angiogenic effects of Kyn on endothelial cells, we first examined whether Kyn activates the canonical AhR signaling pathway. Western blot analysis showed that neither exogenous Kyn treatment nor IDO1 knockdownmediated reduction of endogenous Kyn altered total AhR protein levels in HUVECs (Fig. 6A). Consistently, nuclearcytoplasmic fractionation assays revealed no apparent nuclear translocation of AhR following Kyn stimulation (Fig. 6B), suggesting that Kyn does not exert its effects through AhR activation in endothelial cells. We next investigated alternative signaling pathways potentially involved in Kyn-mediated endothelial regulation. Western blot analysis demonstrated that treatment with increasing concentrations of Kyn markedly enhanced the phosphorylation levels of mTOR and AKT in HUVECs in a concentration-dependent manner (Fig. 6C). Knocking down the IDO1 gene in HUVECs using RNA interference technology led to a decrease in the phosphorylation levels of intracellular mTOR and AKT (Fig. 6D). To further confirm whether the impact of Kyn on the migratory ability of vascular endothelial cells operates through the mTOR pathway, we observed the effects of the mTOR pathway inhibitor Rapamycin on both control HUVECs and Kyn-treated HUVECs. The results revealed that 0.3 µM Rapamycin inhibited the previously observed enhancement of HUVEC migration by Kyn (Fig. 6E). Additionally, tube formation capacity was also regulated by the mTOR signaling pathway (Fig. 6F), as Rapamycin reversed the increased tube length and branching points induced by Kyn in HUVECs.

Fig. 6.

Kyn Regulates the Angiogenic Capacity of Endothelial Cells through mTOR. (A) Western blot analysis showing AhR protein levels in HUVECs treated with exogenous Kyn or subjected to IDO1 knockdown to reduce endogenous Kyn. (B) Nuclear-cytoplasmic fractionation analysis showing the subcellular localization of AhR in HUVECs following Kyn treatment. (C,D) The regulation of the mTOR signaling pathway by increased Kyn (C) or decreased Kyn (D) in endothelial cells. Cropped blots are shown; original images are presented in Supplementary Information 2. (E) The effect of the mTOR inhibitor rapamycin on the migration of Kyn-treated HUVECs. Scale bar: 100 μm. (F) The effect of rapamycin on the angiogenic capacity of Kyn-treated HUVECs. Scale Bar: 100 μm. The analysis of the experimental group and the control group was conducted using paired t-tests. The data are represented as the means ± SEMs from dependent duplicates. ns: P > 0.05,*:P < 0.05,**: P < 0.01, ***:P < 0.001.

Discussion

Metabolic reprogramming, as a critical hallmark of cancer, plays a significant role in the initiation and progression of tumors. Tumor cells adeptly alter metabolic pathways to meet their demands for growth in terms of nutrients and energy²⁶. Various metabolic pathways have been identified to play crucial roles in tumorigenesis, metastasis, and drug resistance in various cancers^27,28. Targeting tumor metabolic reprogramming has emerged as a promising strategy in cancer therapy. In this study, we aimed to investigate metabolic reprogramming in HNSCC to uncover novel therapeutic strategies. Previous research has confirmed the reprogramming of tryptophan metabolism in HNSCC, with Kyn identified as an active metabolite exerting a promoting effect in tumor development. However, the precise mechanisms underlying its pro-carcinogenic effects remain elusive. Therefore, our goal was to elucidate the specific role of Kyn in HNSCC to enhance the efficacy and precision of targeting tryptophan metabolism in therapy.

Researchers like Zhang et al. have discovered that elevated Kyn stimulates cardiac angiogenesis by promoting AhR nuclear translocation and enhancing the expression of vascular endothelial growth factor A. Additionally, Ying et al. found that extracellular vesicles derived from ovarian cancer promote endothelial mitochondrial autophagy associated with tumor angiogenesis by upregulating SIRT3 expression in endothelial cells, with these vesicles being rich in Kyn²⁹. Therefore, these findings prompted us to investigate whether kynurenine is associated with angiogenic phenotypes in the context of HNSCC.

It should be noted that the LC-MS/MS measurement of Kyn levels in this study was performed on a relatively small sample set of ten paired tissues (Fig. 1E), which may limit statistical power. Nevertheless, recent studies have demonstrated that meaningful molecular insights can be obtained from limited sample sizes when appropriate analytical strategies are applied^30,31. These references support the rationale that targeted measurement of key metabolites, such as Kyn, in a well-characterized small cohort can still provide biologically relevant insights, while highlighting the need for future studies with larger cohorts to further validate these findings.

Angiogenesis stands as a pivotal factor in the progression of tumors, facilitating the supply of oxygen and nutrients to tumors through the surrounding capillary network. As tumors grow, their nutritional demands escalate, necessitating the generation of new blood vessels to sustain blood supply. In the absence of vascular support, tumors may undergo necrosis or apoptosis³². Furthermore, angiogenesis is essential for tumor invasion of surrounding tissues and metastasis³³. The significance of vasculature in tumors has made targeting tumor angiogenesis a hot topic in recent research endeavors³⁴. Therefore, combining strategies targeting metabolic reprogramming and angiogenesis in tumors may synergistically enhance therapeutic outcomes. Folkman categorized sprouting angiogenesis into three stages: initiation, invasion/proliferation, and maturation³⁵. During these stages, endothelial cells undergo morphological changes, proliferate extensively under the influence of growth factors, adhere to each other through cell surface adhesion molecules, detach from the extracellular matrix, and construct and expand new micro vessels.

Based on the angiogenic process, this study investigated the regulatory effects of Kyn and its key enzyme IDO1 on HUVEC function. Migration assays showed that Kyn significantly enhanced endothelial cell motility, indicating that kynurenine enhances endothelial cell migration, a fundamental component of angiogenic activity. Adhesion assays further demonstrated that Kyn increased HUVEC adhesion to squamous carcinoma cells, which may promote endothelial cell anchoring at tumor sites during neovascular sprouting. To evaluate the angiogenic effects of kynurenine across multiple experimental systems, endothelial angiogenic behaviors were examined using in vitro tube formation assays, the CAM model, and tumor xenografts with histological assessment of vascular markers. Kyn treatment markedly enhanced endothelial tube formation, as evidenced by increased branching points and capillary lengths, whereas the IDO1 inhibitor epacadostat significantly suppressed these angiogenic features. These findings indicate that IDO1–Kyn signaling positively regulates endothelial angiogenic capacity. Nevertheless, it should be acknowledged that in vitro angiogenesis models do not fully recapitulate tumor angiogenesis, as they reflect both vascular network formation and regression. Therefore, further validation in tumor-specific angiogenesis models is warranted.

Although HUVECs were employed as a standardized endothelial model in this study, tumor-associated endothelial cells (TECs) may exhibit distinct metabolic and signaling characteristics within the tumor microenvironment. To complement in vitro findings, we utilized the CAM assay, which, although primarily reflecting embryonic angiogenesis, provides a physiologically relevant microenvironment with dynamic tissue interactions, bridging the gap between in vitro and in vivo models³⁶. While CAM cannot fully replicate tum or-associated angiogenesis and mouse models remain the gold standard for tumor-host interactions, this intermediate system allows assessment of vascular formation and translational responses. Therefore, future studies using TECs, tumor-derived endothelial cells, or more complex in vivo models will be necessary to fully elucidate the relevance of Kyn-driven angiogenesis in HNSCC.

We also acknowledge a limitation of the current study: epacadostat reduces multiple tryptophan metabolites, and the observed effects may reflect the combined impact of these metabolites rather than Kyn alone. Importantly, our mechanistic experiments indicate that Kyn promotes endothelial angiogenesis primarily via mTOR signaling rather than AhR activation. Future studies employing receptor-specific antagonists or genetic approaches will be required to dissect the precise contributions of individual metabolites and their signaling pathways.

Pan et al. demonstrated that IDO1 knockdown by shRNA in tumor-bearing mice significantly reduced tumor angiogenesis, supporting a role for IDO1 in neovascular formation, consistent with our findings²⁰. Angiogenesis is initiated by multiple biological cues that drive endothelial cell migration and differentiation through receptor-mediated signaling cascades³⁷. In this study, transcriptional profiling of HUVECs revealed that kynurenine regulates several angiogenesis-related pathways, including HIF1A, EGF/EGFR, PDGFB, and ANG2. Although EGFR is not a classical angiogenic factor, accumulating evidence in HNSCC indicates its involvement in angiogenesis, cell survival, and cell cycle regulation^38,39. PDGFB, a key member of the PDGF family, promotes pericyte recruitment and vascular stabilization during angiogenesis and has been implicated in tumor-associated vascular remodeling^40,41. HIF-1α, a hypoxia-responsive transcription factor, enhances VEGF signaling and extracellular matrix remodeling during early angiogenic processes⁴².

Notably, PDGFB, HIF1A, and related angiogenic regulators have been reported to intersect with the mTOR signaling pathway, which serves as a central integrator of metabolic and inflammatory cues across diverse pathological contexts, including cancer and cardiovascular diseases^43–45, supporting the notion that kynurenine-driven metabolic reprogramming may promote vascular remodeling through mTOR pathway activation in HNSCC. Consistent with these reports, our data suggest that kynurenine promotes angiogenesis in HNSCC via mTOR pathway activation. Collectively, these findings reveal a mechanistic link between tryptophan metabolic reprogramming and vascular remodeling, highlighting combined targeting of tryptophan metabolism and angiogenesis as a potential therapeutic strategy for HNSCC.

mTOR signaling has been widely recognized as a central integrator of metabolic and inflammatory cues across diverse pathological conditions, including cardiovascular and metabolic disorders. Clinical studies have shown that systemic inflammatory and metabolic indices are closely associated with disease progression and adverse outcomes, highlighting the broad relevance of metabolism-linked signaling pathways such as mTOR in tissue remodeling processes.

Conclusion

This study unveils the pivotal role of the differential metabolite Kyn in promoting angiogenesis in head and neck squamous cell carcinoma through tryptophan metabolic reprogramming. Mechanistically, Kyn may enhance the expression of angiogenesis-related genes, pro-angiogenic factors, and receptors by activating the mTOR signaling pathway. This activation facilitates the expansion and settlement of endothelial cells within tumor tissues, ultimately fostering angiogenesis in head and neck squamous cell carcinoma and promoting malignant tumor progression (Fig. 7). Targeting IDO1 could represent a potential therapeutic strategy for head and neck squamous cell carcinoma. Moreover, a combination therapy involving vascular inhibition may enhance the therapeutic efficacy against head and neck squamous cell carcinoma.

Fig. 7.

Effect of Kyn-induced activation of the AKT-mTOR pathway on the regulation of angiogenesis in HNSCC.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

HNSCC

Head and neck squamous cell carcinoma

Kyn

Kynurenine

TME

Tumor microenvironment

CAM

Chorioallantoic membrane

LC-MS/MS

Liquid chromatography-tandem mass spectrometry

HUVECs

Human umbilical vein endothelial cells

Author contributions

The final manuscript was read and approved by all writers. The work has been sufficiently contributed to by each author. Shuoqi Lin: Conceptualization, Data curation, Investigation, Methodology, Data Analysis, Writing-original draft; Tesen Liao: Conceptualization, Data curation, Investigation, Writing-review & editing; Shijie Wang: Data curation; Investigation; Methodology; Writing-review & editing; Genggeng zheng: Data curation, Writing-review & editing; Bohua Su: Conceptualization; Project administration; Resources; Supervision; Writing-review & editing; Dali Zheng: Conceptualization; Project administration; Funding acquisition; Resources; Supervision; Writing-review & editing.

Funding

This work was financially supported by National Natural Science Foundation of China [Grant number: 82173180], the Special Project of Central Government for Local Science and Technology Development of Fujian Province [Grant number: 2023L3011], and the Natural Science Foundation of Fujian Province [Grant number: 2025J01142].

Data availability

The data that support the findings of this study are available on request from the corresponding author. The metabolomic data reported in this paper have been deposited in the OMIX, China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (OMIX010812) [46,47].

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of the Fujian Medical University School and Hospital of Stomatology (No. 2021-FMUSS-024). All procedures were followed in compliance with applicable rules and regulations. The written informed consent was obtained from the participants.

Consent for publication

The use and reporting of patient-derived tissue in this study were carried out with written informed consent obtained from the patient or their legally authorized representative, permitting its use for research and publication.

Competing interests

The authors declare no competing interests.

Conflict of interest

The authors declare no competing interest.