LC-MS has emerged as a powerful analytical technique for profiling CYP450 enzymes, particularly due to its ability to resolve complex protein mixtures and provide high-resolution data on sequence coverage and peptide identification [9, 10, 13]. One of the key advantages of LC-MS is its unmatched precision in distinguishing highly homologous CYP450 isoforms, which pose challenges for traditional antibody-based detection methods. Consequently, some CYP450 isoforms have not previously been unequivocally detected in their protein form due to the difficulty in generating specific assays and monoclonal antibodies. However, GEL-LC-MS overcomes this challenge by combining the high-resolution separation of gel electrophoresis with the precise peptide mass detection of LC-MS, enabling the differentiation of closely related isoforms [10]. While DIA or other advanced LC-MS approaches offer higher throughput, they often struggle to detect low-abundance CYP450 isoforms due to signal suppression by abundant proteins. The Gel-LC-MS approach enriches the 45–65 kDa fraction, enhancing isoform-specific detection and enabling reliable semi-quantitative comparisons across samples.

The analysis of CRC, head and neck squamous cell carcinoma (HNSCC), and breast cancer models revealed distinct CYP450 expression profiles, highlighting tissue-specific metabolic adaptations that reflect the unique biochemical demands and microenvironmental conditions of each cancer type [14, 15]. In CRC models, the predominant expression of CYP2W1, CYP2S1, CYP1B1, and CYP2J2 suggests their potential functional involvement in tumour metabolism, xenobiotic processing and disease progression [16,17,18,19]. CYP2W1, in particular, has been previously associated with CRC and is known for its ability to metabolize pro-carcinogens into active compounds [16, 19]. The presence of CYP3A5 across all CRC patient biopsy stages T1 to T4, suggests its role in xenobiotic metabolism, possibly influencing drug response in CRC patients [20].

In head and neck squamous cell carcinoma (HNSCC) models, CYP1B1 was among the most abundant enzyme in SCC9 cells. CYP1B1 is known to metabolize a wide range of xenobiotics and is often overexpressed in cancer, making it a potential biomarker for HNSCC and a target for selective inhibitors [21,22,23]. The presence of CYP2W1 in HNSCC further reinforces its potential functional role in cancerous cells, expanding its relevance beyond CRC [23], while CYP2S1 has been implicated in inflammatory responses [24], potentially contributing to the aggressive nature of HNSCC tumours. Breast cancer models, particularly MCF-7 cells, exhibited a different CYP450 profile, with CYP2J2, CYP2S1, CYP2A6 and CYP2A13 being the most prominent isoforms. CYP2J2 has been implicated in the metabolism of chemotherapeutic agents, raising the possibility that its expression could influence drug resistance and treatment response [25]. Differential expression of CYP450 isoforms across cancer types has potential clinical implications, as these enzymes modulate drug metabolism, efficacy and toxicity. For example, elevated CYP2W1 in CRC and HNSCC could inform prodrug activation strategies, while CYP1B1 overexpression in HNSCC may guide selective inhibitor development. Linking these expression patterns to potential drug metabolism enhances the translational relevance of our dataset and may inform personalized therapeutic strategies.

The study also highlighted the influence of xenograft and tissue microenvironments on CYP450 expression. For example, CYP1A1 expression increased significantly in CHO-1A1 xenograft models, reaching 51% sequence coverage compared to 22% coverage in transfected cells. This suggests that the in vivo tumour microenvironment may enhance CYP1A1 stability or metabolic activity, possibly due to stromal interactions or extracellular matrix remodelling [26]. Similarly, CYP4F3 and CYP20A1 were prominently expressed in advanced-stage HNSCC patient biopsies, indicating that the tumour microenvironment may upregulate these isoforms to support cancer progression. These findings underscore the importance of considering the tissue context when studying CYP450 expression and designing therapeutic interventions.

GEL-LC-MS provides a highly sensitive and effective approach for enriching specific protein subgroups that are often undetectable in conventional total proteomic analyses. Notably, a search of the PRIDE database (https://www.ebi.ac.uk/pride/) did not reveal any datasets reporting the detection of CYP2W1 and CYP2S1 in colon or head and neck cancer models, underscoring the novelty of our findings. To further contextualize our results, we compared our GEL-LC-MS dataset with publicly available ProteomicsDB data (https://www.proteomicsdb.org/). This comparison identified 17 CYP isoforms consistently detected across seven common models (HepG2, SCC9, DOK, HT55, MCF-7, DLD1, HCC2998). Importantly, we detected CYP isoforms with established roles in drug metabolism; CYP2W1, CYP2J2, and CYP1A1, expressed uniquely in additional cancer cell lines (A253, C106, COLO-205, Detroit-562, FADU, LS174T, OSC-19, SSC9), transfected lines (CHO-1A1, SW480-2W1, SW480-mock), and xenografts (DLD-1-Xeno, HT55-Xeno, CHO-1A1-Xeno), which have not been previously reported. These findings demonstrate the ability of GEL-LC-MS to uncover low-abundance, drug-metabolizing CYP isoforms that may be missed in standard shotgun proteomics. However, a limitation of this approach is reduced accuracy in absolute protein quantification, primarily due to variability in trypsin digestion and peptide extraction from gels. Overall, our results highlight the unique strength of GEL-LC-MS for comprehensive, isoform-specific CYP450 profiling, providing novel insights into cancer-associated metabolic pathways and potential drug metabolism targets.

Log₂-transformed abundances and categorical expression levels (Low, Medium, High) enabled semi-quantitative comparisons of CYP450 expression across cancer models. Differential expression patterns suggest tissue-specific metabolic adaptations and have potential implications for drug metabolism. For example, CYP2W1, highly expressed in CRC and HNSCC models, represents an attractive target for prodrug activation strategies [16, 27], whereas CYP1B1 overexpression in HNSCC may guide selective inhibitor development. Similarly, CYP2J2 and CYP2S1 expression in breast cancer models could influence drug metabolism and treatment response. These findings highlight how isoform-specific CYP450 profiling can inform personalized therapeutic strategies. Microsomal enrichment and GEL-LC-MS provided sensitive detection of low-abundance isoforms, while comparison with public datasets (PRIDE, ProteomicsDB) confirmed the novelty of key drug-metabolizing CYPs uniquely detected in our models.

Microsomal isolation is a critical step in achieving in-depth CYP450 profiling [28,29,30]. Microsomes, vesicles derived from the endoplasmic reticulum, are enriched with CYP450 enzymes and other drug-metabolizing enzymes. Isolating microsomes allows researchers to focus on the functional fraction of the proteome, thereby enhancing the detection of low-abundance CYP450 isoforms. For instance, this study identified low-abundance CYPs such as CYP4A10 and CYP4A14 in mouse liver microsomes, demonstrating the utility of microsomal enrichment in overcoming detection challenges associated with whole-cell lysates. CYP450 profiles generated by GEL-LC-MS compared favourably with the previously established GEL-MALDI-MS strategy [10], with 22 CYP450 isoforms being common between the two approaches. Notably, CYP2C39, CYP3A25, and CYP8B1 were detected exclusively by GEL-LC-MS, while CYP17A1, CYP1A1, CYP2B10, CYP2C44, CYP2D11, CYP3A16, CYP4A10, CYP4F3, and CYP51A1 were only observed with GEL-MALDI-MS. These differences underscore the complementary nature of these techniques and their potential for comprehensive CYP profiling.

This study offers valuable insights into CYP450 expression across cancer models but has limitations. While GEL-LC-MS is effective for detecting specific isoforms, its sensitivity remains constrained by the dynamic range of protein expression, potentially leading to the omission of very low-abundance CYP450 enzymes. Targeted proteomic approaches, such as parallel reaction monitoring (PRM) or multiple reaction monitoring (MRM), could further enhance sensitivity and quantification accuracy. Additionally, this study focused on protein expression without assessing enzymatic activity, which is critical for functional relevance. The influence of post-translational modifications (PTMs), as well as genetic and epigenetic factors affecting CYP450 regulation, were not explored. Future studies incorporating activity assays, PTM analysis, and transcriptomic integration will provide a more comprehensive understanding of CYP450 function in cancer progression and drug metabolism.