Worldwide, haematological malignancies are responsible for roughly 6% of all the cancer-related deaths. Leukaemias are one of the most severe types of cancer, as only about 40% of the patients have an overall survival of 10 years or more. Myelodysplastic Syndrome (MDS), a pre-leukaemic condition, is a blood disorder characterized by the presence of dysplastic, irregular, immature cells, or blasts, in the peripheral blood (PB) and in the bone marrow (BM), as well as multi-lineage cytopenias.
We have created a detailed, lineage-specific, high-fidelity in-silico erythroid model that incorporates known biological stimuli (cytokines and hormones) and a competing diseased haematopoietic population, correctly capturing crucial biological checkpoints (EPO-dependent CFU-E differentiation) and replicating the in-vivo erythroid differentiation dynamics. In parallel, we have also proposed a long-term, cytokine-free 3D cell culture system for primary MDS cells, which was firstly optimized using easily-accessible healthy controls. This system enabled long-term (24-day) maintenance in culture with high (>75%) cell viability, promoting spontaneous expansion of erythroid phenotypes (CD71+/CD235a+) without the addition of any exogenous cytokines. Lastly, we have proposed a novel in-vitro in-silico framework using GC-MS metabolomics for the metabolic profiling of BM and PB plasma, aiming not only to discretize between haematological conditions but also to sub-classify MDS patients, potentially based on candidate biomarkers. Unsupervised multivariate statistical analysis showed clear intra- and inter-disease separation of samples of 5 distinct haematological malignancies, demonstrating the potential of this approach for disease characterization.
The work herein presented paves the way for the development of in-vitro in-silico technologies to better, characterize, diagnose, model and target haematological malignancies such as MDS and AML.