Background:
Intravenous (IV) administration of human mesenchymal stem cells (hMSCs) is clinically preferred for its simplicity and systemic distribution [1]. IV-injected hMSCs can exert remote immunomodulatory effects [2]. In contrast, intraarterial (IA) injection may enhance delivery to the central nervous system [3]. To elucidate the dynamics of each route, we employed magnetic particle imaging (MPI) to perform dynamic whole-body tracking with minute-scale temporal resolution of superparamagnetic iron oxide (SPIO)-labeled hMSCs following IV or IA injection in a multiple sclerosis (MS) mouse model.

Hypothesis/Objective:
We hypothesized that the injection route (IV vs. IA) critically determines the biodistribution of hMSCs, and aimed to quantitatively characterize the biodistribution of injected hMSCs using whole-body MPI in an MS mouse model.

Methods:
Experimental autoimmune encephalomyelitis (EAE) was induced in C57BL/6 mice using an MOG₃₅–₅₅/CFA emulsion kit. EAE mice were catheterized via the tail vein or internal carotid artery for IV and IA injection, respectively (n=3 per cohort). hMSCs were labeled with ferucarbotran (25 µg Fe/mL) complexed with poly-L-lysine (3 µg/mL). Mice received IV injections totaling 5.0×10⁵ hMSCs or IA injections totaling 1.2×10⁵ hMSCs, each delivered as four equal sequential doses to enable dynamic MPI tracking. The biodistribution of injected cells was monitored using a Momentum MPI scanner. A calibration curve for signal quantification was generated using two different fiducials containing labeled hMSCs. CT was performed on an IVIS Spectrum/CT scanner for anatomical reference. Prussian blue (PB) staining and anti-human nuclear antigen (anti-HuNa) immunostaining were performed on tissue sections to confirm the presence of SPIO and the localization of injected human cells in tissues.

Results:
Pre-injection validation confirmed successful ferucarbotran labeling of hMSCs, with intracellular iron uptake verified by PB staining (Fig. 1). Sequential imaging after each of the four incremental IV injections showed progressive signal intensification within the upper-thoracic region (Fig. 2). Longitudinal in vivo MPI/CT co-registration (Fig. 3) demonstrated gradual clearance of pulmonary, hepatic, and splenic signals over seven days post IV injection. These findings suggest that IV-injected hMSCs predominantly localize in the lungs initially, with delayed secondary accumulation in the reticuloendothelial system (Fig. 4). Sequential imaging after each of the four incremental IA injections demonstrated progressive enhancement of signal intensity along the brain and carotid regions (Fig. 5). 3D MPI/CT revealed strong localization of MPI signal in the brain with secondary accumulation in the liver and spleen (Fig. 5). These results highlight that IA injection achieves substantially greater brain delivery of hMSCs compared to IV administration.

Conclusion:
MPI can be used for real-time visualization of hMSC biodistribution following IV and IA delivery. Both routes led to splenic accumulation, while IA injection enhanced brain-targeted delivery. These findings highlight the value of MPI for optimizing stem cell delivery strategies in neuroinflammatory diseases.

Author
Asif Itoo
Johns Hopkins