Reduced sulfatide content in deferoxamine-induced senescent HepG2 cells

Noninvasive morpho-molecular imaging reveals early therapy-induced senescence in human cancer cells

“…By analyzing a large cancer cell dataset, using recent TIS biological models thoroughly validated elsewhere ( 25 , 54 ), we revealed notably reproducible quantitative morpho-chemical traits featured by TIS human cancer cells over early reprogramming stages. As the TIS markers here presented are visible via NLO microscopy and QPI, the community would benefit from the advantages of these easy-to-use, label-free, rapid, and quantitative optical technologies to foster further TIS exploration in cancer therapy screening, when using both in vitro cell culture models and patient-derived cell cultures.…”

“…1 . Our study is conducted based on a validated cellular model of senescence proposed by Ghislanzoni et al ( 25 ), which exploits deferoxamine (DFO)–mediated iron dysregulation with consequent mitochondrial dysfunction, as a trigger for the development of the TIS phenotype (see note S1). By analyzing a large population of such TIS cells, we could overcome the intrinsic, high biological variability of these experiments.…”

“… Human cancer cells from the HepG2 line were used as cellular models to investigate morpho-molecular traits of TIS onset and manifestation. In this part of the study, senescence was induced through treatment with 100 μM deferoxamine (DFO) ( 25 ). Five different TIS cultures and five control counterparts were prepared for each different time point and monitored over the treatment time.…”

Deep ensemble learning and transfer learning methods for classification of senescent cells from nonlinear optical microscopy images

“…The study by Oh et al associates lipids and proteins upregulation to the TIS cells phenotype, but lacks any time resolution, since they investigated only one time point and no measurements were conducted during the first hours after treatment, at the onset of senescence. In our previous work (Bresci et al, 2023), we successfully identified quantitative markers of TIS in vitro in human hepatic cancer cells, treated with a chemotherapeutic drug, via label-free multimodal NLO microscopy combining SRS and TPEF. We studied the onset and progression of TIS at consecutive time-points, up to 7-days post treatment, finding quantitative metabolic and biochemical indicators.…”

“…We also incorporate the Gradient-weighted Class Activation Mapping (Grad-CAM) visualization approach (Selvaraju et al, 2017) as an interpretability technique to provide insight into the features that contribute to accurate classification. We demonstrate the effectiveness of our system on an experimental dataset of human hepatic cancer cells, in which the TIS phenotype is induced both via drug-based and radiotherapy treatments (Ghislanzoni et al, 2023). We propose a comparison of our novel, unbiased, and automatic approach for the classification of TIS cancer cells over tumoral cells with other competing deep learning architecture specifically developed for this work.…”