In Vivo Cellular Magnetic Imaging: Labeled vs. Unlabeled Cells

Abstract

Superparamagnetic iron oxide (SPIO)-labeling of cells has been applied for magnetic resonance imaging (MRI) cell tracking for over 30 years, having resulted in a dozen or so clinical trials. SPIO nanoparticles are biodegradable and can be broken down into elemental iron, and hence the tolerance of cells to magnetic labeling has been overall high. Over the years, however, single reports have accumulated demonstrating that the proliferation, migration, adhesion and differentiation of magnetically labeled cells may differ from unlabeled cells, with inhibition of chondrocytic differentiation of labeled human mesenchymal stem cells (hMSCs) as a notable example. This historical perspective provides an overview of some of the drawbacks that can be encountered with magnetic labeling. Now that magnetic particle imaging (MPI) cell tracking is emerging as a new in vivo cellular imaging modality, there has been a renaissance in the formulation of SPIO nanoparticles this time optimized for MPI. Lessons learned from the occasional past pitfalls encountered with SPIO-labeling of cells for MRI may expedite possible future clinical translation of (combined) MRI/MPI cell tracking.

Keywords: Magnetic resonance imaging; cell tracking; immune cells; magnetic particle imaging; stem cells; superparamagnetic iron oxide nanoparticles.

Figure 1:

Schematic outline of in vivo MPI/MRI cell tracking, as exemplified by monitoring of injected SPIO-labeled T cells in a breast cancer patient. (A) Cells are first magnetically labeled with SPIO nanoparticles through simple co-incubation in culture medium during normal expansion. After collecting and washing, cells are injected intravenously, and the patient is subjected to imaging follow-up. (B) Anticipated imaging findings are specific homing to the primary tumor (T) and a regional lymph node (LN) tumor metastasis, along with non-specific uptake in the lung (Lu), liver (Li), and spleen (S). While MRI cell tracking has entered the clinic, the use of MPI so far has only been pre-clinical.

Figure 2:

Methods for intracellular magnetic cell labeling. (A) Cells can endocytose SPIO nanoparticles by conjugating them to antibodies (AB) that induce receptor-mediated endocytosis, by synthesizing them with a cationic coating (CC), by complexing them with cationic transfection agents (TA), or by membrane permeabilization using magnetoelectroporation (MEP) or magnetosonoporation (MSP). (B) Upon invagination of the cell membrane and subsequent SPIO internalization into newly formed endosomes (E), the SPIO-containing endosomes rapidly fuse with lysosomes (L) to form endosomal-lysosomal complexes (EL). The SPIO particles are eventually degraded within the EL system to elemental iron (Fe), and incorporated into the normal body iron blood pool through natural iron-sequestering proteins.

Figure 3:

Symmetric vs. asymmetric cell division and dilution of SPIO particles among daughter cells. (A) In symmetrically dividing cells, the nanoparticles are distributed equally among daughter cells. (B) In asymmetrically dividing stem cells, the nanoparticles are not distributed equally. This can lead to a sharp decline in cell detectability. Progeny cells without particles are MR-invisible (solid crosses), while cells with only one particle are borderline detectable (dashed crosses). Asterisks represent parent cells that undergo self-renewal with an exact copy of itself, which is a hallmark feature of undifferentiated, noncommitted stem cells. (C-E) Intracerebroventricular transplantation of β-galactose-transduced C17.2 neural stem cells in mouse brain. Shown is a mismatch between high-resolution ex vivo MR images and histology, with a sharp boundary between MR-detectable and non-detectable cells as a result from asymmetric cell division. (C) Two weeks after transplantation, SPIO-labeled cells migrated vast distances toward the outer cortical layers of the cerebrum and olfactory bulb as revealed by anti-β-gal staining. This is in sharp contrast to the MRI pattern (D), which shows hypointense cells centered in and around the ventricles (site of transplantation), but not the cortical layers. (E) Merged histology/MR image, in which β-gal cells are visualized as red and MRI-hypointense cells are yellow, further illustrates this mismatch. Scale bar=1 mm. Panels Adapted/reproduced with permission.^[44]2007, Wiley.

Figure 4:

SPIO labeling of hMSCs inhibits chondrogenic differentiation. While SPIO-labeling does not affect labeled cells viability nor differentiation into adipocytes and osteocytes, labeled cells fail to produce a proper extracellular matrix that contains proteoglycans and collagen II. (A-C) hMSCs incubated with SPIO-PLL fail to generate (A) collagen II and (C) a safranin-O-positive extracellular matrix after 21 days of cell culture, with (B) Prussian Blue staining revealing SPIO-containing cells throughout the pellet. Partial inhibition of chondrogenesis in pellets made from 1:1 mixtures of SPIO-PLL-labeled and unlabeled hMSCs is shown in the next row. (D) Collagen II and (F) safranin O staining is limited to (E) unlabeled (Prussian Blue negative) cells. Chondrocytic differentiation is inhibited by SPIO and not PLL, as hMSCs exposed to PLL only (without SPIO) generate large amounts of (G) collagen II-rich and (I) safranin O-positive extracellular matrix, and (H) are unstained by PB. For comparison, hMSCs unexposed to either Feridex® or PLL are shown following staining for (J) collagen II, (L) safranin O, and (K) are unstained for PB. All pictures were taken at the same magnification. Bar in panel A represents 400 μm. Panels A-L reproduced with permission. ^[53a] 2004, Wiley.

Figure 5:

Detection of unlabeled hMSCs with MRI. Unlike other cells, hMSCs have an abundant expression of high-mannose N-linked glycans on the cell membrane, that can generate signal on MANw CEST MRI. Shown are T2-weighted (A,C) and MANw CEST MR (B,D) images for hMSCs (A,B) and (as representative example for other cells) human glial restrictor precursor cells (C,D). Panels A-D reproduced with permission.^[68] 2022, Nature Publishing Group.