In Vivo Tracking of Human Neural Progenitor Cells in the Rat Brain Using Magnetic Resonance Imaging Is Not Enhanced by Ferritin Expression

Abstract

Rapid growth in the field of stem cell research has generated a lot of interest in their therapeutic use, especially in the treatment of neurodegenerative diseases. Specifically, human neural progenitor cells (hNPCs), unique in their capability to differentiate into cells of the neural lineage, have been widely investigated due to their ability to survive, thrive, and migrate toward injured tissues. Still, one of the major roadblocks for clinical applicability arises from the inability to monitor these cells following transplantation. Molecular imaging techniques, such as magnetic resonance imaging (MRI), have been explored to assess hNPC transplant location, migration, and survival. Here we investigated whether inducing hNPCs to overexpress ferritin (hNPCs(Fer)), an iron storage protein, is sufficient to track these cells long term in the rat striatum using MRI. We found that increased hypointensity on MRI images could establish hNPC(Fer) location. Unexpectedly, however, wild-type hNPC transplants were detected in a similar manner, which is likely due to increased iron accumulation following transplantation-induced damage. Hence, we labeled hNPCs with superparamagnetic iron oxide (SPIO) nanoparticles to further increase iron content in an attempt to enhance cell contrast in MRI. SPIO-labeling of hNPCs (hNPCs-SPIO) achieved increased hypointensity, with significantly greater area of decreased T2* compared to hNPC(Fer) (p < 0.0001) and all other controls used. However, none of the techniques could be used to determine graft rejection in vivo, which is imperative for understanding cell behavior following transplantation. We conclude that in order for cell survival to be monitored in preclinical and clinical settings, another molecular imaging technique must be employed, including perhaps multimodal imaging, which would utilize MRI along with another imaging modality.

Figure 1

Transient ferritin expression in human neural progenitor cells (hNPCs^Trans-Fer). (A) Immunostained hNPCs^Trans-Fer showing robust transient expression of ferritin. (B) In vitro R2* maps of hNPCs^Trans-Fer incubated with 0, 2, 20, and 200 μM ferric ammonium citrate (FAC), compared with wild-type hNPCs (hNPCs^WT) incubated with 2, 20, and 200 μM FAC and hNPCs incubated with superparamagnetic iron oxide nanoparticles (hNPCs-SPIO). R2* instead of T2* maps shown for easier visualization. (C) Quantification of immunocytochemical analysis determines the percent of hNPCs^Trans-Fer (between days 1 and 10, p < 0.0001).

Figure 2

Stable ferritin expression in hNPCs (hNPCs^Fer). (A) Schematic of lentiviral construct consisting of the following: long terminal repeat (LTR), the central polypurine tract (cPPT), the mouse phosphoglycerate kinase 1 promoter (PGK), light (L-) ferritin gene, FLAG epitope tag, heavy (H-) ferritin gene, the posttranscriptional regulatory element of woodchuck hepatitis virus (WPRE), and self-inactivating (SIN). (B) Percentage of hNPCs stably expressing ferritin more than 7 months postinfection. (C) Western blot of hNPCs^Fer, hNPCs^WT, and hNPCs^WT incubated with 200 μM FAC lysates stained against H- and L-ferritin. (D) Comparison of hNPCs^Fer and hNPCs^WT in their capacity to proliferate (BrdU marker) and differentiate (GFAP and βIII tubulin markers). (E) In vitro R2* maps of hNPCs^Fer incubated with 0, 2, 20, and 200 μM FAC, compared with hNPCs^WT incubated with 2, 20, and 200 μM FAC and hNPCs-SPIO. R2* instead of T2* maps shown for easier visualization. (F) In vivo detection of luciferase-expressing hNPCs (hNPCs^Luc2, green box) and hNPCs^Fer (red box) in the rat striatum 1 week following transplantation. (G) Human cytoplasmic (hCyto) marker and (H) Prussian blue staining in the region of hNPCs^Luc2 transplantation. (I) hCyto staining and (J) Prussian blue staining in the region of hNPC^Fer transplantation. Data are given as mean ± SEM. MRI scale bars: 1 mm. Histological scale bars: 50 μm.

Figure 3

Analysis of ferritin-based hNPC tracking using MRI. (A) In vivo detection of needle insertion-induced hypointense signal in the right cortex and striatum at 1 week after treatment (yellow arrowhead). (B) In vivo detection of media injection-induced hypointense signal in the left cortex and striatum at 1 week following transplantation (yellow arrowhead). (C) In vivo detection of hNPCs^Fer (red box) and hNPCs^Luc2 (green box) induced hypointense signal at end point (11 weeks after transplantation). (D) Prussian blue, (E) human nuclear (hNuc) marker, (F) L-ferritin, and (G) H-ferritin staining from chronological sections within the hNPC^Fer transplant region. (H) Prussian blue, (I) hNuc, (J) L-ferritin, and (K) H-ferritin staining from chronological sections within the hNPC^Luc2 transplant region. (L) Schematics comparing hCyto-based histologically stained cell locations (left) and corresponding MRI hypointense signal (right). The same outline color depicts the histological and MRI hypointense signal in the same animal. MRI acquired at end point (week 11). hNPC^Fer signal shown in the left hemisphere of each schematic and hNPC^Luc2 signal shown in the right hemisphere of each schematic. All fluorescence immunostaining images were acquired with the same exposure times. MRI scale bars: 1 mm. Histological scale bars: 50 μm.

Figure 4

Analysis of iron-based hNPC tracking using MRI. In vivo detection of hNPCs incubated with (A) low (3 μg/ml), (B) medium (30 μg/ml), and (C) high (300 μg/ml) concentrations of SPIO at 1 week posttransplantation (yellow arrowheads). (D) Semiquantitative assessment of total areas of decreased T2* induced by various transplants 1 week following transplantation. Areas of decreased T2* and the number of animals for each condition: needle insertion (decreased T2*area = 85.33 ± 97.08; N = 3), media injection (decreased T2*area = 37.67 ± 7.64; N = 3), hNPCs^WT (decreased T2*area = 134.33 ± 25.50; N = 3), hNPCs^Luc2 (decreased T2*area = 168.67 ± 55.65; N = 3), hNPCs^Fer (decreased T2*area = 79.17 ± 77.46; N = 6), and hNPCs-SPIO (decreased T2*area = 1,091.67 ± 277.71; N = 3). Asterisk indicates significant difference compared to the rest of the conditions calculated using one-way ANOVA with Bonferroni correction. Individual p values comparing areas of decreased T2* between hNPCs-SPIO and the rest of the conditions are as follows: needle insertion (p = 0.0041), media injection (p = 0.0028), hNPCs^WT (p = 0.0040), hNPCs^Luc2 (p = 0.0049), and hNPCs^Fer (p < 0.0001). There was no significant difference between any of the other conditions. (E) Semiquantitative analysis of hNPC^Fer graft size and corresponding MR hypointensity ROI sizes in 14 animals through Pearson correlation (Pearson’s r = 0.1782). Data are given as mean ± SEM. MRI scale bars: 1 mm.

Figure 5

Detection of dead hNPCs^Fer. (A) In vivo imaging of hNPCs^Luc2 using bioluminescence imager, In Vivo Imaging System (IVIS), 1, 3, 5, and 9 weeks following transplantation. Cyclosporine (cyclo) was withdrawn at week 3 and presumed cell death occurred at week 5. (B) In vivo MR imaging of hNPCs^Fer (red arrowhead) and hNPCs^Luc2 (green arrowhead) in the same animal as in (A) 1, 3, 5, and 9 weeks after transplantation. Staining in the region of hNPC^Fer transplantation for (C) hNuc, (D) L-ferritin, and (E) Prussian blue. Staining in the region of hNPC^Luc2 transplantation for (F) hNuc, (G) L-ferritin, and (H) Prussian blue. All fluorescence immunostaining images were acquired with the same exposure times. IVIS scale bars: 1 cm. MRI scale bars: 1 mm. Histological scale bars: 50 μm.

Figure 6

Detection of dead hNPCs-SPIO. (A) In vivo detection of live hNPCs-SPIO 1 week posttransplantation (green box). Staining in the region of live hNPCs-SPIO for (B) hNuc and (C) Prussian blue. (D) In vivo detection of SPIO-media (white arrowhead) and dead hNPCs-SPIO (red box) 1 week following transplantation. Staining in the region of dead hNPCs-SPIO for (E) hNuc and (F) Prussian blue. All fluorescence immunostaining images were acquired with the same exposure times. MRI scale bars: 1 mm. Histological scale bars: 50 μm.