Science to Practice: Can Stem Cells Be Labeled Inside the Body Instead of Outside?

Abstract

Mesenchymal stem cells (MSCs) were labeled in vivo by intravenous injection of ferumoxytol, a Food and Drug Administration–approved intravenous iron supplement, and after their isolation and processing from bone marrow, the same MSCs were injected in rats with an osteochondral defect, allowing MR monitoring of their engraftment for at least 4 weeks; this straightforward labeling approach, avoiding several regulatory issues, may accelerate clinical translation of MR imaging for stem cell tracking.

Summary

Instead of conventional labeling ex vivo in cell culture, mesenchymal stem cells (MSCs) were labeled in vivo with intravenous injection of ferumoxytol (Feraheme; AMAG Pharmaceuticals, Lexington, Mass), a Food and Drug Administration (FDA)-approved intravenous iron supplement. After their isolation and processing from bone marrow, the same MSCs were injected in rats with an osteochondral defect, allowing MR monitoring of their engraftment for at least 4 weeks. This straightforward labeling approach, avoiding several regulatory issues, may accelerate clinical translation of magnetic resonance (MR) imaging for stem cell tracking.

The Setting

Because of their ease of isolation and versatile curative properties, MSCs are currently undergoing clinical trials in a large variety of degenerative diseases. For further optimization of treatment paradigms, it is of crucial importance to be able to noninvasively monitor the accuracy of cell injection of MSCs, their immediate engraftment pattern, and their long-term retention at the target site. MR imaging for cell tracking is the only modality suited for this purpose, and all clinical studies to date have relied on labeling cells ex vivo with superparamagnetic iron oxide (SPIO) during their normal expansion in culture (1). For instance, MSCs labeled with ferumoxides (Feridex; Bayer HealthCare Pharmaceuticals, Wayne, NJ) were found to home to the occipital horn and upper spinal cord following intravenous injection in patients with multiple sclerosis, providing direct evidence of local versus systemic cell-mediated immunosuppression (2). However, the clinically FDA-approved SPIO formulations such as ferumoxides (Feridex, Bayer HealthCare Pharmaceuticals; Endorem, Guerbet, Roissy, France) and ferucarbotran (Resovist; Bayer Schering Pharma, Berlin, Germany) used for clinical MR cell tracking were discontinued in 2008. Since then, ferumoxytol, an FDA-approved ultrasmall SPIO (USPIO) formulation for treatment of iron deficiency anemia, has been explored as a potential clinical alternative, although it is a less efficient MR imaging tracking label (3,4). In the current issue of Radiology, Khurana et al (5) have exploited the natural phagocytic properties of MSCs for their in vivo labeling with ferumoxytol following intravenous injection. With in vivo labeling, iron uptake was found to be superior to comparative ex vivo labeling, and labeled, collected MSCs that were subsequently implanted into the knees of rats with an induced osteochondral defect could be readily detected on T2-weighted MR images for at least 4 weeks after transplantation.

The Science

USPIOs have long been known to accumulate in bone marrow cells following intravenous injection (6). In the study by Khurana et al (5), a dose of 28 mg of iron per kilogram of ferumoxytol was injected in nine normal Sprague-Dawley donor rats. Two days after injection, bone marrow cells were harvested, processed, and placed in culture. As bone marrow contains a mixture of non-MSC cell types (including macrophages), cells were plated for 7 days to isolate the adherent MSCs. Immunophenotyping showed that most of the labeled cells were indeed MSCs. Further analysis revealed a distribution of ferumoxytol in endosomes and lysosomes, with an overall iron content of 4.3 pg of iron per cell, which was found to have no effect on cell viability and the differentiation of MSCs into chondrocytes.

MSCs were then mixed with agarose gels as scaffolding material, and 1 × 10⁶ cells were bilaterally implanted into the knee joints of seven athymic rats having a circular, mechanically induced osteochondreal defect (7). For each animal, one knee received ferumoxytol-labeled cells, and the other, unlabeled cells. Three MR images were obtained at 7.0 T, immediately after and at 2 and 4 weeks after implantation. The mean T2 values for the ferumoxytol-labeled MSCs and unlabeled cells on the day of transplantation were 15.5 versus 24.4 milliseconds, respectively (P < .05), proving that a sufficient decrease of signal-to-noise ratio can be obtained to detect labeled cells. At 2 and 4 weeks, the differences between labeled and unlabeled cells became less pronounced, as unlabeled cells showed a progressive decrease of T2 values, believed to be a result from degradation of the hydrogel scaffold. Histopathologic examination revealed colocalization of cells that were positive for the Prussian blue (iron) and Alcian blue (cartilage) stains in the knee joint, demonstrating that cells retained ferumoxytol in vivo while being able to induce chondrogenesis, which is the ultimate goal of their therapeutic application.

The Practice

Clinical use: While the authors have chosen a mechanical osteochondral defect as the therapeutic target (7), it is important to realize that this is just an example application. Being able to track MR-labeled MSCs safely and effectively will have myriad clinical applications, including but not limited to stroke, myocardial infarct, and a range of autoimmune diseases including multiple sclerosis and type I diabetes mellitus. As of today, the results of seven clinical trials in which the researchers used cells labeled ex vivo with ferumoxides (Feridex, Bayer HealthCare Pharmaceuticals; Endorem, Guerbet) have been published. All these studies, facing stringent FDA-controlled regulatory issues, have been performed outside the United States (1).

With ferumoxytol now left as the only FDA-approved clinical formulation (albeit for treatment of iron deficiency anemia, and not as an MR labeling agent), it only seems logical to have optimized the hitherto used ex vivo labeling approaches for this particular agent (3,4). However, this approach requires the use of an additional transfection agent (ie, protamine sulfate). While protamine sulfate is also FDA-approved for a different application (to counteract heparin toxicity), the combination of the off-label use of several agents, together with the necessary ex vivo manipulation of labeled cells (including washing procedures to remove excess label), represents an extra hurdle for regulatory approval of such a uniform labeling procedure. Instead, the in vivo labeling approach presented here (5) circumvents those issues, with the injected intravenous dose of 28 mg of iron per kilogram of ferumoxytol only being twice the currently approved clinical dose of 1.02 g of iron per patient.

Future opportunities and challenges: This in vivo labeling approach will be limited to MSCs or other stem cell types that are phagocytic in nature and that can be readily isolated. Neural stem cells have been labeled in vivo by intracerebroventricular injection of micron-sized iron particles (8), but the nonapproved formulation and the impossibility of isolating stem cells from the subventricular zone safely preclude clinical translation. This approach will also only be applicable to allogeneic MSC transplants, as in an autologous scenario the MSC donor will have a ubiquitous presence of ferumoxytol-labeled macrophages that cannot be discriminated from the transplanted cells. This does not prevent its clinical use; many clinical trials are based on allogeneic MSC therapy because not all donors yield sufficient numbers of MSCs. To obtain enough labeled MSCs for clinical dosing, they may have to be expanded well beyond the current 7 days, with many cell divisions, diluting the ferumoxytol label to uncertain cellular detection levels.

It will be furthermore important to assure that the labeled cells are indeed MSCs and not macrophages, to be pursued with more detailed immunophenotyping studies. Finally, inherent to labeling cells with any sort of particle is the inability to distinguish live from dead cells, the possibility of false-positive results in the case of a macrophage influx and secondary particle uptake, and long-term monitoring of cells owing to either label dilution (if cells would divide) or biodegradation. However, the ultimate clinical utility of MR imaging for cell tracking will be monitoring the accuracy of cell injection (in real time by using MR-compatible catheters) and the immediate cell engraftment, which will be possible with the use of this approach. One can envision that this use may be extended to other imaging strategies for cell tracking, for instance, “hot spot” fluorine 19 MR imaging (9), where the first patient was injected with fluorinated (Cell Sense CS-1000; CelSense, Pittsburgh, Pa ) labeled cells in April 2013.