Figure 1.
Optimization of labeling and in vitro detection limits.
A–C: 19F MRS of NSCs incubated with 15 µL (A), 50 µL (B), and 100 µL (C) PFPE emulsion per mL medium together with a KF solution as an internal standard, KF frequency set to −120.9 ppm, main PFPE peak at −93.2 ppm, signal intensity in arbitrary units. The cell signal was maximized at a concentration of 100 µL/mL. D–F: MRI of 20,000, 10,000, and 5,000 labeled, and 65,000 control cells plated in gelatin, D shows the merged 1H and 19F, E the 19F image only. The 20,000 and 10,000 cells spots were clearly detectable, whereas a significant signal was not detected for the 5,000 cells, probably due to cells spreading over many voxels, thus not exceeding the critical detection limit. From the summed SNR of the two spots with 19F signal we estimate that approximately 1,000 cells need to accumulate in one image voxel to overcome the detection limit, as indicated by the crossing of the linear fit in F with the dashed line illustrating an SNR of 3.5. For the fit we assumed 0 SNR for 0 cells. Error bars in F are over n = 3 technical replicates. Note: The surface coil was adjusted parallel to the paper plane.
Figure 2.
Effect of the labeling on cell viability, proliferation, and differentiation capacity.
Photomicrographs of in vitro samples stained for Ki67 (A, B, I and J), nestin (C, F E and H), GFAP (D, E G, H, K and N), and βIII-tubulin (L, M, O and P). There were no major qualitative differences in the viability or proliferation and differentiation capacity of the labeled (B, F, G, H, J, N, O and P) versus non-labeled control cells (A, C, D, E, I, K, L and M), both during proliferation (A–H), i.e. directly after the labeling procedure, and after 9 days of differentiation (I–P). The percentage of proliferating Ki67+ cells during proliferation or after differentiation was not significantly altered by the labeling (Q). Quantification of the number of living cells before, 0 days after, and 7 days after incubation with the 19F marker revealed that the viability was significantly decreased compared to controls at 0 days but normalized at 7 days after labeling (R). The scale bar represents 25 µm for M and P, 50 µm for all others.
Figure 3.
In vivo 19F MRI and correlation with immunohistochemistry.
1H, 19F, and merged MR images of a mouse (Animal 1), which had been injected with non-labeled control cells into the left striatum and labeled NSCs into the right hemisphere (A). Only the labeled cells generated a 19F signal whereas hunu staining confirmed the presence of cell grafts on both sides as indicated by the arrows (B). MRI of another mouse (Animal 2) 2 days (C) and 6 days (E) after grafting showed no major signal loss in the 19F images over time. This animal had received two deposits of labeled cells in the left striatum and one deposit in the right striatum. The location and intensity of 19F signal from cell clusters, marked with white arrows, correlated well with hunu staining on histological sections. Note that the 19F resolution allows the distinction of the two clusters on the left hemisphere (B, F). Only cells that were clearly immunoreactive to both hoechst and hunu were considered as grafted human NSCs (D). Scale bars are 50 µm for D, 1 mm for all others.
Figure 4.
Effect of the 19F marker on cell phenotypes in vivo.
Photomicrographs of non-labeled control (A, C, E) and PFPE labeled NSCs (B, D, F) on tissue sections from transplanted animals. The majority of the implanted cells were, at this early time point after injection, still neural stem and progenitor cells as confirmed by nestin/hunu and GFAP/hunu stainings (A–D). Presence of DCX+/hunu+ (E, F) neuroblasts showed that NSCs were capable of neuronal differentiation. By qualitative analysis we did not detect major differences in the PFPE labeled cell population compared to control cells. Scale bars are 20 µm.