Cell therapies have transformed treatment for certain blood cancers and hold promise for autoimmune diseases and neurological conditions, but they share a fundamental problem: once therapeutic cells are injected into a patient, clinicians have almost no way to know where they go. Conventional imaging technologies including magnetic resonance imaging (MRI) and computed tomography (CT) cannot reliably visualize the small cell numbers involved in these treatments, leaving researchers without feedback on whether a therapy is reaching its target, how many cells arrive, or why some patients respond and others do not.
A study published in Science Advances from researchers at Johns Hopkins Medicine describes a potential path toward solving that problem. Using magnetic particle imaging (MPI) — a newer imaging technology that detects magnetically labeled cells with high sensitivity and no tissue background signal — the team tracked two types of therapeutic cells across the whole body in living mice, comparing how injection route, cell size, dose, and disease state influenced biodistribution. The results suggest MPI could eventually provide the real-time, quantitative cell tracking that cell therapy development has so far lacked.
In a press release, senior author Jeff Bulte, professor of radiology and radiological science and director of cellular imaging at the Johns Hopkins Institute for Cell Engineering, said that the work points toward using MPI to determine more precise cell therapy doses for individual patients.
What the imaging revealed
The team labeled two cell types of different sizes — mesenchymal stem cells (MSCs) at approximately 25 micrometers and neural precursor cells (NPCs) derived from induced pluripotent stem cells at approximately 10 micrometers — with superparamagnetic iron oxide (SPIO) nanoparticles, then injected them into normal mice and mice with experimental autoimmune encephalomyelitis (EAE), a well-established preclinical model for multiple sclerosis (MS).
MPI generates a positive signal proportional to the number of labeled cells present and is not attenuated by surrounding tissue — unlike MRI, which produces hypointense signals prone to artifacts and interference. That quantitative, whole-body readout allowed the team to assess cell distribution across multiple organs simultaneously without ionizing radiation or sacrificing the animal.
The most significant finding concerned delivery route. Intra-arterial injection resulted in substantially more cells reaching target organs — the brain and spleen — compared with intravenous injection. In EAE mice, therapeutic cells accumulated in the spleen, an organ implicated in MS pathology. In the press release, lead author Ali Shakeri-Zadeh, assistant professor of radiology and radiological science at Johns Hopkins, noted that the spleen is thought to be a source of harmful immune T cells in MS, and that the accumulation of therapeutic cells there suggested they could engage disease activity directly at that site.
The imaging also revealed off-target accumulation in the lungs and liver — relevant for biodistribution safety assessments. In normal mice, cells traveled to the lungs, liver, and brain, but were not detectable in the spleen, underscoring that disease state shaped organ distribution in ways healthy animal data alone could not predict.
Why this matters for cell therapy development
Dosing decisions for cell therapies are currently made with limited information about how many cells actually reach the intended site, conflating variability in delivery with variability in response. A quantitative imaging readout that decouples the two would allow developers to optimize injection routes, refine dose ranges, and identify sources of variability that make cell therapy outcomes difficult to predict or replicate.
Previous work established MPI's utility for tracking adoptively transferred T cells in vivo in brain tumor models, showing that magnetic labeling does not impair cell function and that MPI can detect labeled cells after intravenous or intracerebroventricular delivery. The Johns Hopkins study extends that precedent across a broader range of cell types, delivery routes, and disease contexts.
Human-scale MPI systems capable of brain imaging have been developed and validated in recent literature. As the technology moves toward clinical application, preclinical datasets like those generated here will be critical for interpreting what human MPI images of therapeutic cell distribution actually mean — and whether real-time imaging at the point of delivery can move cell therapy dosing toward the individualized model precision medicine requires.
