Associate Professor USC, California, United States
Introduction: Acute myeloid leukemia (AML) represents a significant challenge in oncology due to its vast heterogeneity, manifesting in varied genetic and molecular profiles as well as diverse clinical presentations and outcomes. The traditional treatment strategy predominantly relies on chemotherapy regimens, which, while often effective, are also associated with significant toxicity and a high risk of relapse. This underscores the urgent need for improved therapeutic options and highlights the necessity of developing novel biomarkers that can accurately classify leukemia subtypes and predict responses to treatment. Recent advancements in our understanding of cancer metabolism have shed light on how metabolic reprogramming is crucial to the development and progression of AML. This metabolic reprogramming in AML cells includes rewiring of glycolysis and mitochondria respiration, altered lipid metabolism, and changes in amino acid utilization. These metabolic alterations not only support the uncontrolled growth and survival of leukemic cells but also contribute to resistance against conventional therapies. Tracking and understanding these metabolic shifts represents new avenues for diagnostic and therapeutic interventions in AML. In the presented study, we utilized fluorescence lifetime imaging microscopy (FLIM) to track the intensity and lifetime changes of the autofluorescent metabolic coenzymes, NAD(P)H and FAD among different leukemia cell lines and patient samples. Machine-learning-assisted analysis revealed that imaging-based metabolic biomarkers can profile different leukemia subtypes as well as predict drug response in patient samples.
Materials and
Methods: Leukemia cell lines were cultured in RPMI + 10% FBS and patient leukemia cells were cultured on OP9 feeder layer in alpha-MEM + 20% FBS. Fluorescence intensity and lifetime images of NAD(P)H and FAD in leukemia cells were acquired by a two-photon microscope (Leica SP8 FALCON). RNA-seq data of leukemia cell lines were sourced from Cancer Cell Line Encyclopedia. Image processing and analysis were carried out in Python 3.5.
Results, Conclusions, and Discussions: We derived four channels of metabolic optical biomarkers (MOBs) from raw FLIM images to represent key metabolic phenotypes: fluorescence intensity of NAD(P)H, optical redox ratio, enzyme-bound NAD(P)H fraction, and fluorescence lifetime of enzyme-bound NAD(P)H. From these, we extracted 205 cellular and subcellular features, including morphology, signal strength, and metabolic compartmentalization in cytoplasmic and mitochondrial regions (Fig. 1A). This comprehensive dataset enabled profiling of leukemia phenotypes at high resolution. Using 7 leukemia cell lines representing different types of leukemia/lymphoma (including AML), we showed that the MOB dataset can distinguish different populations in downstream analysis (e.g. Uniform Manifold Approximation and Projection (UMAP) plot) (Fig. 1B). Three AML cell lines showed distinguishable profiles to other leukemia types. Moreover, compared to bulk RNA-seq data of these populations, we observed similar hierarchical distribution between MOB and transcriptome profiling (Fig. 1C). We further performed chemo drug-response assay on multiple AML cell lines and selected the drug-resistant and vulnerable lines. Representative MOB features that distinguish these two cell lines were identified and major pathways associated with these features were identified and may serve as candidates for combination therapy to overcome standard chemo-drug resistance (Fig. 1D, E). Currently, we are treating patient-derived leukemia samples with different chemo-drug cocktails and using MOB-based models to predict response. Overall, this study underscores MOBs as biomarkers for diagnosis and real-time monitoring of treatment responses, and may pave the way for the development of more effective and less toxic therapeutic strategies for leukemia.
