Abstract

Accurate detection of lymph node metastasis (LNM) is critical for colorectal cancer (CRC) staging and treatment planning, yet current histopathological assessment based on lymph nodes remains labor-intensive and operator-dependent. Here, we developed a tissue metabolic fingerprinting platform leveraging label-free ferric nanoparticle-enhanced laser desorption/ionization mass spectrometry (FELDI-MS) to directly acquire colorectal cancer tissue metabolic fingerprints (CRC-TMFs) from 276 primary CRC tissue samples (138 non-metastatic/LNM-, 138 metastatic/LNM+). Based on CRC-TMFs, we constructed a machine learning-based diagnostic model for LNM detection, achieving area under the curve (AUC) of 0.914. Furthermore, metabolic profiling revealed cysteine deficiency in LNM + tissues, concomitant with upregulation of glutamate-cysteine ligase catalytic subunit (GCLC), which catalyzes the rate-limiting step in glutathione biosynthesis from cysteine. Functional validation demonstrated that GCLC knockdown inhibited CRC cell proliferation and migration, underscoring its role in metastatic reprogramming. Our work not only introduces a rapid, operator-independent tool for precise LNM assessment but also highlights dysregulated cysteine-GCLC-glutathione metabolism as a key feature of metastatic reprogramming in CRC.

Fig. 1. Workflow depicting metabolic fingerprinting for the evaluation of lymph node metastasis (LNM) in primary colorectal cancer (CRC) tissues. (A) Pathologically confirmed CRC tissues from the biobank were analyzed via the ferric nanoparticle-enhanced laser desorption/ionization mass spectrometry (FELDI-MS) platform to acquire metabolic fingerprints, establishing the comprehensive CRC tissue metabolic fingerprint (CRC-TMFs) database. (B) Machine learning pipelines, incorporating feature selection and algorithm optimization, were utilized to develop a diagnostic model for CRC LNM prediction. (C) Mechanism analysis of metabolic reprogramming through identification of dysregulated pathways, followed by biological validation including enzyme expression analysis and functional knockdown experiments in CRC cell lines to confirm roles in cell migration and metastasis.

Fig. 2. Performance characterization of the FELDI-MS platform for metabolomic profiling. (A) Scanning electron microscopy (SEM) image of ferric oxide nanoparticles, revealing rough surface that facilitates the capture of small molecule metabolites. Sale bar: 100 nm. (B) Elemental mapping of ferric oxide nanoparticles, with red representing ferric (Fe), green representing oxygen (O), and yellow representing the overlay of Fe and O distributions. Sale bar: 100 nm. (C) Selected-area electron diffraction pattern of ferric oxide nanoparticles. Sale bar:5 nm−1. (D) Carbon elemental mapping of nanoparticle-molecule hybrids in mixed solutions containing glucose (Glc) and bovine serum albumin (BSA), highlighting molecular selectivity for small metabolites. Tolerance of the FELDI-MS. Signal intensity remains unsuppressed under (E) high salt (0.5 M NaCl) and (F) protein (5 mg mL−1 BSA) conditions, enabling direct analysis of small metabolites, including glycine (Gly), methionine (Met), pyroglutamic acid (Pyro), spermine (Spe), and taurine (Tau). Reproducibility assessment of (G) Fe3O4 nanoparticles and (H) 1,5-diaminonaphthalene (DAN) for standard metabolites (methionine, Met; taurine, Tau; glucose, Glc; cysteine, Cys). The coefficients of variation (CVs) ranged from 7.7 % to 14.2 % for Fe3O4 nanoparticles and from 35.8 % to 58.4 % for DAN (n = 8). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

https://doi.org/10.1016/j.mtbio.2025.102712