Biological properties of iron oxide nanoparticles for cellular and molecular magnetic resonance imaging

Abstract

Superparamagnetic iron-oxide particles (SPIO) are used in different ways as contrast agents for magnetic resonance imaging (MRI): Particles with high nonspecific uptake are required for unspecific labeling of phagocytic cells whereas those that target specific molecules need to have very low unspecific cellular uptake. We compared iron-oxide particles with different core materials (magnetite, maghemite), different coatings (none, dextran, carboxydextran, polystyrene) and different hydrodynamic diameters (20-850 nm) for internalization kinetics, release of internalized particles, toxicity, localization of particles and ability to generate contrast in MRI. Particle uptake was investigated with U118 glioma cells und human umbilical vein endothelial cells (HUVEC), which exhibit different phagocytic properties. In both cell types, the contrast agents Resovist, B102, non-coated Fe(3)O(4) particles and microspheres were better internalized than dextran-coated Nanomag particles. SPIO uptake into the cells increased with particle/iron concentrations. Maximum intracellular accumulation of iron particles was observed between 24 h to 36 h of exposure. Most particles were retained in the cells for at least two weeks, were deeply internalized, and only few remained adsorbed at the cell surface. Internalized particles clustered in the cytosol of the cells. Furthermore, all particles showed a low toxicity. By MRI, monolayers consisting of 5000 Resovist-labeled cells could easily be visualized. Thus, for unspecific cell labeling, Resovist and microspheres show the highest potential, whereas Nanomag particles are promising contrast agents for target-specific labeling.

Keywords: cellular localization; electron microscopy; iron oxide nanoparticles; iron staining; magnetic resonance imaging (MRI); toxicity; uptake kinetics.

Figure 1

Internalization of different iron oxide nanoparticles (see Table 1) in human U118 glioma cells (A, C, D) and human umbilical vein endothelial cells (HUVEC, B) after 24 h at an iron concentration of 0.3 mg/mL; (A, B) The microspheres (0.31 μm and 0.85 μm) were absorbed best, Resovist, B102 and non-coated Fe₃O₄ showed an equally strong high uptake, whereas NanomagN20, N70 and N100 were not measurably internalized; (C) Iron uptake after 24 h as a function of the iron/particle concentrations applied; (D) Different uptake kinetics were observed for the different particles at the same iron concentration (0.2 mg Fe/mL); however, the strongest uptake was mostly observed between 24 and 36 h; (A–D) n = 3 ± S.D.

Figure 2

Long term stability of cell-labeling by nanoparticles. U118 glioma cells were incubated with different iron oxide nanoparticles for 24 h at an iron concentration of 0.2 mg/mL, washed and further cultivated. The cellular iron content was measured after different periods of time. Resovist showed the longest resting time, B102 the shortest. Initial iron concentration varied due to different initial labeling efficiencies, see Figure 1. n = 3 ± S.D.

Figure 3

Toxicity of the particles as measured by the release of the cytosolic enzyme lactate dehydrogenase (LDH) in vitro. U118 cells were exposed to nanoparticles for 24 h at an iron concentration of 0.2 mg/mL at 1% fetal calf serum (FCS, heat-inactivated), and released LDH was quantified in the culture supernatants; LDH-activity in lysed cells yielded the total activity (100%). All particles produced only a slight LDH-release comparably to the effect of 4% dimethyl sulfoxide (DMSO). n = 3 ± S.D.

Figure 4

Aggregation of N100-particles after 1 h exposure to human plasma, human fibrinogen (0.21 mg/mL) or phosphate-buffered saline (control), revealed by electron microscopy.

Figure 5

Visualization of the internalized particles by light microscopy after Prussian blue staining (top) and electron microscopy (bottom). U118 glioma cells were exposed to the particles for 24 h at 0.2 mg iron/mL. All particles were internalized (examples marked with arrows) except Nanomag-particles N20, N70 and N100. Resovist, B102 and non-coated Fe₃O₄ clustered in the cells but the microspheres (M 850 und M 310) did not. Typical results from three individual incubations are shown for each.

Figure 6

MRI visualization of nanoparticles-labeled U118 glioma cells grown as a monolayer on membrane filters. Cells grown on polyethylene terephthalate (PET) membrane filters were labeled for 24 h with nanoparticles (0.2 mg Fe/mL) and washed. T2*-weighted images were generated with a 3 Tesla MRI at room temperature. Controls with Resovist or 100,000 unlabelled cells are negative (arrow). After labeling with Resovist, already 5000 cells caused a visible signal reduction. Similar results were obtained with microspheres (850 und 310 nm), B102 and non-coated Fe₃O₄ (shown only in one example). In contrast, Nanomag particles N20, N70 and N100 yielded no signals, even at high cell numbers. Typical results from three individual incubations are shown for each.