journal article·2024-12-05Open access

Machine Learning and Metabolomics Predict Mesenchymal Stem Cell Osteogenic Differentiation in 2D and 3D Cultures

Michail E. Klontzas, Spyros I. Vernardis, Aristea Batsali, Fotios Papadogiannis, Nicki Panoskaltsis, Athanasios Mantalaris

Journal of Functional Biomaterials· Volume 15· Issue 12

Abstract

Stem cells have been widely used to produce artificial bone grafts. Nonetheless, the variability in the degree of stem cell differentiation is an inherent drawback of artificial graft development and requires robust evaluation tools that can certify the quality of stem cell-based products and avoid source-tissue-related and patient-specific variability in outcomes. Omics analyses have been utilised for the evaluation of stem cell attributes in all stages of stem cell biomanufacturing. Herein, metabolomics in combination with machine learning was utilised for the benchmarking of osteogenic differentiation quality in 2D and 3D cultures. Metabolomics analysis was performed with the use of gas chromatography–mass spectrometry (GC-MS). A set of 11 metabolites was used to train an XGboost model which achieved excellent performance in distinguishing between differentiated and undifferentiated umbilical cord blood mesenchymal stem cells (UCB MSCs). The model was benchmarked against samples not present in the training set, being able to efficiently capture osteogenesis in 3D UCB MSC cultures with an area under the curve (AUC) of 82.6%. On the contrary, the model did not capture any differentiation in Wharton’s Jelly MSC samples, which are well-known underperformers in osteogenic differentiation (AUC of 56.2%). Mineralisation was significantly correlated with the levels of fumarate, glycerol, and myo-inositol, the four metabolites found most important for model performance (R2 = 0.89, R2 = 0.94, and R2 = 0.96, and p = 0.016, p = 0.0059, and p = 0.0022, respectively). In conclusion, our results indicate that metabolomics in combination with machine learning can be used for the development of reliable potency assays for the evaluation of Advanced Therapy Medicinal Products.
Source: crossref · Quality: baseline

Research context

OpenAlex topics: Mesenchymal stem cell research · 3D Printing in Biomedical Research · Metabolomics and Mass Spectrometry Studies

SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.

Affiliations

Artificial Intelligence and Translational Imaging (ATI) Lab, Department of Radiology, School of Medicine, University of Crete, 71003 Heraklion, Greece

Computational Biomedicine Laboratory, Institute of Computer Science, Foundation for Research and Technology (ICS-FORTH), 70013 Heraklion, Greece

The Francis Crick Institute, London NW1 1AT, UK

Haemopoiesis Research Laboratory, School of Medicine, University of Crete, 71003 Heraklion, Greece

BioMedical Systems Engineering Laboratory, Panoz Institute, School of Pharmacy and Pharmaceutical Sciences, Trinity College, D02 PN40 Dublin, Ireland

National Institute for Bioprocessing Research and Training (NIBRT), Foster Avenue, Mount Merrion, Blackrock, A94 X099 Dublin, Ireland

Funding

European Union

References (43)

Show source-provided references
1. Dimitriou, R., Jones, E., McGonagle, D., and Giannoudis, P.V. (2011). Bone Regeneration: Current Concepts and Future Directions. BMC Med., 9.10.1186/1741-7015-9-66 ↗
2. Oxidized Alginate Hydrogels with the GHK Peptide Enhance Cord Blood Mesenchymal Stem Cell Osteogenesis: A Paradigm for Metabolomics-Based Evaluation of Biomaterial Design10.1016/j.actbio.2019.02.017 ↗
3. Quality Cell Therapy Manufacturing by Design10.1038/nbt.3525 ↗
4. Understanding and Leveraging Cell Metabolism to Enhance Mesenchymal Stem Cell Transplantation Survival in Tissue Engineering and Regenerative Medicine Applications10.1002/stem.3079 ↗
5. Cell Culture Metabolomics: Applications and Future Directions10.1016/j.drudis.2010.06.012 ↗
6. Tunable Supramolecular Hydrogels for Selection of Lineage-Guiding Metabolites in Stem Cell Cultures10.1016/j.chempr.2016.07.001 ↗
7. Migdadi, L., Sharar, N., Jafar, H., Telfah, A., Hergenröder, R., and Wöhler, C. (2023). Machine Learning in Automated Monitoring of Metabolic Changes Accompanying the Differentiation of Adipose-Tissue-Derived Human Mesenchymal Stem Cells Employing 1H-1H TOCSY NMR. Metabolites, 13.10.3390/metabo13030352 ↗
8. Development of a Robust Consensus Modeling Approach for Identifying Cellular and Media Metabolites Predictive of Mesenchymal Stromal Cell Potency10.1093/stmcls/sxad039 ↗
9. Bone Grafts and Bone Graft Substitutes in Orthopaedic Trauma Surgery. A Critical Analysis10.2106/00004623-200703000-00026 ↗
10. Metabolomics Analysis of the Osteogenic Differentiation of Umbilical Cord Blood Mesenchymal Stem Cells Reveals Differential Sensitivity to Osteogenic Agents10.1089/scd.2016.0315 ↗
11. RGD-Functionalized Polyurethane Scaffolds Promote Umbilical Cord Blood Mesenchymal Stem Cell Expansion and Osteogenic Differentiation
12. Standardizing GC–MS Metabolomics10.1016/j.jchromb.2008.04.049 ↗
13. Data Correction Strategy for Metabolomics Analysis Using Gas Chromatography-Mass Spectrometry10.1016/j.ymben.2006.08.001 ↗
14. Vernardis, S.I., Terzoudis, K., Panoskaltsis, N., and Mantalaris, A. (2017). Human Embryonic and Induced Pluripotent Stem Cells Maintain Phenotype but Alter Their Metabolism after Exposure to ROCK Inhibitor. Sci. Rep., 7.10.1038/srep42138 ↗
15. Feature Selection with the Boruta Package10.18637/jss.v036.i11 ↗
16. Differential Expression of Cell Cycle and WNT Pathway-Related Genes Accounts for Differences in the Growth and Differentiation Potential of Wharton’s Jelly and Bone Marrow-Derived Mesenchymal Stem Cells10.1186/s13287-017-0555-9 ↗
17. Distinct Differentiation Potential of “MSC” Derived from Cord Blood and Umbilical Cord: Are Cord-Derived Cells True Mesenchymal Stromal Cells?10.1089/scd.2011.0414 ↗
18. Rybkowska, P., Radoszkiewicz, K., Kawalec, M., Dymkowska, D., Zabłocka, B., Zabłocki, K., and Sarnowska, A. (2023). The Metabolic Changes between Monolayer (2D) and Three-Dimensional (3D) Culture Conditions in Human Mesenchymal Stem/Stromal Cells Derived from Adipose Tissue. Cells, 12.10.3390/cells12010178 ↗
19. Osteogenic Commitment of Wharton’s Jelly Mesenchymal Stromal Cells: Mechanisms and Implications for Bioprocess Development and Clinical Application10.1186/s13287-019-1450-3 ↗
20. WISP-1 Is a Wnt-1- and Beta-Catenin-Responsive Oncogene10.1101/gad.14.5.585 ↗
21. The Use of ’Omics Technology to Rationally Improve Industrial Mammalian Cell Line Performance10.1002/bit.25673 ↗
22. From Omics to Cellular Mechanisms in Mammalian Cell Factory Development10.1016/j.coche.2021.100688 ↗
23. Capelli, C., Cuofano, C., Pavoni, C., Frigerio, S., Lisini, D., Nava, S., Quaroni, M., Colombo, V., Galli, F., and Bezukladova, S. (2023). Potency Assays and Biomarkers for Cell-Based Advanced Therapy Medicinal Products. Front. Immunol., 14.10.3389/fimmu.2023.1186224 ↗
24. ATMP development and pre-GMP environment in academia: A safety net for early cell and gene therapy development and manufacturing10.1016/j.iotech.2022.100099 ↗
25. Manzini, P., Peli, V., Rivera-Ordaz, A., Budelli, S., Barilani, M., and Lazzari, L. (2022). Validation of an automated cell counting method for cGMP manufacturing of human induced pluripotent stem cells. Biotechnol. Rep., 33.10.1016/j.btre.2022.e00708 ↗
26. Quality Control Assays for Clinical-Grade Human Mesenchymal Stromal Cells: Methods for ATMP Release10.1007/978-1-4939-3584-0_19 ↗
27. Lechanteur, C., Briquet, A., Bettonville, V., Baudoux, E., and Beguin, Y. (2021). MSC Manufacturing for Academic Clinical Trials: From a Clinical-Grade to a Full GMP-Compliant Process. Cells, 10.10.3390/cells10061320 ↗
28. Quality Control Assays for Clinical-Grade Human Mesenchymal Stromal Cells: Validation Strategy10.1007/978-1-4939-3584-0_20 ↗
29. Analyzing cell-type-specific dynamics of metabolism in kidney repair10.1038/s42255-022-00615-8 ↗
30. Integrating metabolome dynamics and process data to guide cell line selection in biopharmaceutical process development10.1016/j.ymben.2022.03.015 ↗
31. Metabolomics—The link between genotypes and phenotypes10.1023/A:1013713905833 ↗
32. Real-time health monitoring through urine metabolomics10.1038/s41746-019-0185-y ↗
33. Metabolic Regulation of Stem Cells and Differentiation: A Forkhead Box O Transcription Factor Perspective10.1089/ars.2020.8126 ↗
34. The metabolic programming of stem cells10.1101/gad.293167.116 ↗
35. Metabolomic profiling of human pluripotent stem cell differentiation into lung progenitors10.1016/j.isci.2022.103797 ↗
36. Rajput, D., Wang, W.J., and Chen, C.C. (2023). Evaluation of a decided sample size in machine learning applications. BMC Bioinform., 24.10.1186/s12859-023-05156-9 ↗
37. Dhiman, P., Ma, J., Qi, C., Bullock, G., Sergeant, J.C., Riley, R.D., and Collins, G.S. (2023). Sample size requirements are not being considered in studies developing prediction models for binary outcomes: A systematic review. BMC Med. Res. Methodol., 23.10.1186/s12874-023-02008-1 ↗
38. Automatic early stopping using cross validation: Quantifying the criteria10.1016/S0893-6080(98)00010-0 ↗
39. Noise injection for training artificial neural networks: A comparison with weight decay and early stopping10.1118/1.3213517 ↗
40. Foundations of Machine Learning-Based Clinical Prediction Modeling: Part II-Generalization and Overfitting10.1007/978-3-030-85292-4_3 ↗
41. Snell, K.I.E., Allotey, J., Smuk, M., Hooper, R., Chan, C., Ahmed, A., Chappell, L.C., Von Dadelszen, P., Green, M., and Kenny, L. (2020). External validation of prognostic models predicting pre-eclampsia: Individual participant data meta-analysis. BMC Med., 18.10.1186/s12916-020-01766-9 ↗
42. TRIPOD+AI statement: Updated guidance for reporting clinical prediction models that use regression or machine learning methods10.1136/bmj-2023-078378 ↗
43. Recommendations for Reporting Machine Learning Analyses in Clinical Research10.1161/CIRCOUTCOMES.120.006556 ↗