Probing stem cells on the prowl: Set them (label)-free


How did this get started?

Since about 30 years ago, we have been developing magnetic resonance imaging (MRI) techniques to locate and interrogate the presence of injected therapeutic cells in a living body, without the need to perform biopsies or extract tissues (which can be harmful to patients). Why is this important? At present, once stem cells are injected in patients, we don't know where they go, how long they stay and where and for how long they survive1. At the time, we were able to label immune cells magnetically by loading them with iron oxide nanoparticles, which disturb the magnetic field and make labelled cells appear “as black holes” on MRI scans2-3. Following proof-of-principle that we could track stem cells in vivo4, a dozen or so clinical trials using iron oxide-loaded cells have been performed over the last 15 years5.

Is there room for improvement?

The number of clinical trials has been few in number, indicating that there are hurdles with implementing magnetically labelled cell tracking. Chief amongst them are the bespoke manufacturing and clinical approval processes of the clinical imaging agents, dilution of label upon cell division, the inability to discriminate live from dead cells, and the presence of other substances that mimic the black spots of cells complicating image interpretation. Ideally, tracking transplanted cells with MRI uses the innate properties of cells without the need of labelling them.

How did we get the label-free idea?

Previously we were able to distinguish malignant tumor cells from their benign precursor cells by virtue of changes in cell membrane glycosylation6. Our colleagues also demonstrated that glucose can be used as a diamagnetic chemical exchange saturation transfer (CEST) MRI contrast agent, in both pre-clinical7 and clinical8 studies. Shortly thereafter, Yue Yuan joined my lab through the Pearl and Yueh-Heng Yang foundation, a foundation that brings talented Chinese scientists into our department with the goal to get trained and eventually return back home. While applying for a fellowship application with the Maryland Stem Cell Research Fund, she stumbled upon a paper that reported that mesenchymal stromal cells (MSCs) are rich in high-mannose (HM) N-linked glycans, the level of which was markedly reduced in differentiated osteoblasts9. Since mannose resembles glucose, we hypothesized that we can detect unlabelled MSCs by mannose-weighted (MANw) CEST MRI. And thus, the idea was born (Figure 1).


Figure 1: Tracking of unlabelled MSCs by mannose-weighted chemical exchange saturation transfer magnetic resonance imaging (MANw CEST MRI). Unlike most other cells, high-mannose N-linked glycans are abundant on the MSC cell membrane. The hydroxyl groups on mannose residues (green) can be selectively manipulated with an off-resonance frequency saturation pulse, which changes the MRI water proton signal of the transplanted cells. Image courtesy of co-author Shreyas Kuddannaya.

What were the results?

We first confirmed that mannose exhibits indeed a distinct CEST MRI signal, alone and when mixed in complex systems containing other biologically relevant metabolites. We then compared the MANw CEST MRI signal of high- and low-mannose expressing cell lines in vitro.

Among the 10 cell lines tested, MSCs clearly stood out in terms of MANw CEST MRI signal. To confirm that this was indeed a result of high HM N-glycan expression, we stained cells with fluorescent galanthus nivalis lectin (GNL), which binds specifically to mannose. MSCs exhibited a strongly positive cellular membrane staining, which could not be detected in the other cell lines. As further proof of validation, we differentiated MSCs into osteoblasts, which have reduced numbers of HM N-linked glycans9. Compared to undifferentiated MSCs, osteoblasts had both a lower MANw CEST MRI signal and GNL-FITC positive mannose expression.


Figure 2: Yue Yuan’s first baby shower.

Encouraged by these results, we performed in vivo studies of transplanted MSCs and three other types of cells.  In the midst of the experiments, Dr. Yue Yuan got her first baby (Figure 2) but fortunately Congxiao Wang, a Chinese Scholarship Council student in our lab at the time, was able to step in and continue to make progress. On the MANw CEST MR images, a much more pronounced contrast was observed for transplanted hMSCs compared to the other cell types. To validate that the changes in MANw CEST MRI signal was primarily a result from changes in the local mannose content associated with the expression of HM N-glycans, GNL-FITC staining demonstrated an excellent match between the local area of contrast on MRI and histopathology.

Importantly, we also found that the CEST MRI signal could be used as a surrogate biomarker of cell survival. When we performed bioluminescence imaging (BLI) of luciferase-transduced cells, where only live cells emit signal, we found a strong correlation between the signals of the two imaging modalities. We hypothesized here that when cells die, the HM N-linked glycans are quickly broken down, and are no longer be able to provide CEST MRI contrast.

What is next?

As of January 2022, over 1000 clinical trials are registered that use human MSCs, mostly for immunosuppressive purposes in a variety of neurodegenerative diseases including multiple sclerosis, amyotrophic lateral sclerosis or Lou Gehrig’s disease, and ischemic stroke. Since MRI is already routinely performed to diagnose these diseases initially and for follow-up monitoring of therapeutic intervention, tracking unlabelled MSCs using MANw CEST MRI can be easily implemented, as it needs a simple adjustment of MRI pulse sequence parameters. This would just call for a local IRB amendment of an existing MRI protocol, if any. We are excited to pursue this avenue with MSCs, previously also termed as medicinal signaling cells10, and now here as “mannosylated stem cells”.


  2. Bulte, J. W.; Hoekstra, Y.; Kamman, R. L.; Magin, R. L.; Webb, A. G.; Briggs, R. W.; Go, K. G.; Hulstaert, C. E.; Miltenyi, S.; The, T. H.; et al., Specific MR imaging of human lymphocytes by monoclonal antibody-guided dextran-magnetite particles. Magn Reson Med 1992, 25 (1), 148-57.
  3. Bulte, J. W.; Ma, L. D.; Magin, R. L.; Kamman, R. L.; Hulstaert, C. E.; Go, K. G.; The, T. H.; de Leij, L., Selective MR imaging of labeled human peripheral blood mononuclear cells by liposome mediated incorporation of dextran-magnetite particles. Magn Reson Med 1993, 29 (1), 32-7.
  4. Bulte, J. W.; Douglas, T.; Witwer, B.; Zhang, S. C.; Strable, E.; Lewis, B. K.; Zywicke, H.; Miller, B.; van Gelderen, P.; Moskowitz, B. M.; Duncan, I. D.; Frank, J. A., Magnetodendrimers allow endosomal magnetic labeling and in vivo tracking of stem cells. Nat Biotechnol 2001, 19 (12), 1141-7.
  5. Bulte, J. W. M.; Daldrup-Link, H. E., Clinical Tracking of Cell Transfer and Cell Transplantation: Trials and Tribulations. Radiology 2018, 289 (3), 604-615.
  6. Song, X.; Airan, R. D.; Arifin, D. R.; Bar-Shir, A.; Kadayakkara, D. K.; Liu, G.; Gilad, A. A.; van Zijl, P. C.; McMahon, M. T.; Bulte, J. W., Label-free in vivo molecular imaging of underglycosylated mucin-1 expression in tumour cells. Nat Commun 2015, 6, 6719.
  7. Chan, K. W.; McMahon, M. T.; Kato, Y.; Liu, G.; Bulte, J. W.; Bhujwalla, Z. M.; Artemov, D.; van Zijl, P. C., Natural D-glucose as a biodegradable MRI contrast agent for detecting cancer. Magn Reson Med 2012, 68 (6), 1764-73.
  8. Xu, X.; Yadav, N. N.; Knutsson, L.; Hua, J.; Kalyani, R.; Hall, E.; Laterra, J.; Blakeley, J.; Strowd, R.; Pomper, M.; Barker, P.; Chan, K.; Liu, G.; McMahon, M. T.; Stevens, R. D.; van Zijl, P. C., Dynamic Glucose-Enhanced (DGE) MRI: Translation to Human Scanning and First Results in Glioma Patients. Tomography 2015, 1 (2), 105-114.
  9. Heiskanen, A.; Hirvonen, T.; Salo, H.; Impola, U.; Olonen, A.; Laitinen, A.; Tiitinen, S.; Natunen, S.; Aitio, O.; Miller-Podraza, H.; Wuhrer, M.; Deelder, A. M.; Natunen, J.; Laine, J.; Lehenkari, P.; Saarinen, J.; Satomaa, T.; Valmu, L., Glycomics of bone marrow-derived mesenchymal stem cells can be used to evaluate their cellular differentiation stage. Glycoconjugate J 2009, 26 (3), 367-384.
  10. Caplan, A. I., Medicinal signalling cells: they work, so use them. Nature 2019, 566 (7742), 39.


Please sign in or register for FREE

If you are a registered user on Nature Portfolio Bioengineering Community, please sign in