Cancer is a complex genetic disease with reduced treatment alternatives due to tumor heterogeneity and drug multiresistance emergence. The sphingolipid (SL) metabolic pathway integrates different responses of cellular stress signals and defines cell survival. Therefore, we suggest studying the perturbations on the sphingolipid pathway (SLP) caused by chemotherapeutic drugs using a systems biology approach to evaluate its functionality as a drug response sensor. We used a sphingomyelin-BODIPY (SM-BOD) sensor to study SL metabolism by flow cytometry and live cell imaging in different cancer models. To decode pathway complexity, we implemented Gussian mixture models, ordinary differential equations models, unsupervised classification algorithms and a fuzzy logic approach to assess its utility as a chemotherapy response sensor. Our results show that chemotherapeutic drugs perturb the SLP in different ways in a cell line-specific manner. In addition, we found that few SM-BOD fluorescence features predict chemosensitivity with high accuracy. Finally, we found that the relative species composition of SL appears to contribute to the resulting cytotoxicity of many treatments. This report offers a conceptual and mathematical framework for developing personalized mathematical models to predict and improve cancer therapy.
Tumor heterogeneity leads to drug resistance in cancer treatment with the crucial role of sphingolipids in cell fate and stress signaling. We analyzed sphingolipid metabolism and autophagic flux to study chemotherapeutic interactions on the A549 lung cancer model. Loaded cells with fluorescent sphingomyelin analog (BODIPY) and mCherry-EGFP-LC3B were used to track autophagic flux and assess cytotoxicity when cells are exposed to chemotherapy (epirubicin, cisplatin, and paclitaxel) together with sphingolipid pathway inhibitors and autophagy modulators. Our cell model approach employed fluorescent sphingolipid biosensors and a Gaussian Mixture Model of cell heterogeneity profiles to map the influence of chemotherapy on the sphingolipid pathway and infer potential synergistic interactions. Results showed significant synergy, especially when combining epirubicin with autophagy inducers (rapamycin and Torin), reducing cell viability. Cisplatin also synergized with a ceramidase inhibitor. However, paclitaxel often led to antagonistic effects. Our mapping model suggests that combining chemotherapies with autophagy inducers increases vesicle formation, possibly linked to ceramide accumulation, triggering cell death. However, the in silico model proposed ceramide accumulation in autophagosomes, and kinetic analysis provided evidence of sphingolipid colocalization in autophagosomes. Further research is needed to identify specific sphingolipids accumulating in autophagosomes. These findings offer insights into potential strategies for overcoming chemotherapy resistance by targeting the sphingolipid pathway.
Transcriptomics enables comprehensive analysis of RNA molecules, with RNA-sequencing providing key insights into gene expression. However, bioinformatics analyses remain limited by the chosen strategy, an aspect that can be improved through machine learning integration. This study compared three RNA-seq pipelines (step-by-step, nf-core, and OneStop) implemented in HPC and GUI environments, using data from BioProject GSE130437 of breast cancer MCF7 cell lines resistant and susceptible to palbociclib. All pipelines identified a comparable number of genes ($\sim 14000$ per condition), but each detected unique subsets not shared across methods. Differential gene expression analysis revealed distinct transcriptomic profiles, with limited overlap among pipelines. LUM, KCNK2, and LINC00052 transcripts were consistently identified across pipelines and in the reference study, while others appeared only in a pipeline-specific context, highlighting biases introduced by preprocessing steps, statistical models, and default parameters. Performance assessments demonstrated trade-offs, the step-by-step approach offered flexibility but required high expertise and longer runtimes; nfcore achieved the fastest and most reproducible results through containerized workflows; and OneStop provided the most user-friendly interface, although scalability and parameter customization were limited. Overall, pipeline selection substantially impacts DEG outcomes and downstream interpretations. Considering scalability, reproducibility, and machine learning integration, nf-core proved the most practical choice for large datasets, while OneStop lowered entry barriers for non-specialists. These findings emphasize the need for careful pipeline selection and the development of ML-ready outputs to strengthen reproducibility and expand transcriptomics applications toward predictive modeling and personalized medicine.