Bifunctional Labeling of Rabbit Mesenchymal Stem Cells for MR Imaging and Fluorescence Microscopy.

Purpose: Longitudinal imaging studies are important in the translational process of stem cell-based therapies. Small animal imaging models are widely available and practical but insufficiently depict important morphologic detail. In contrary, large animal models are logistically challenging and costly but offer greater imaging quality. In order to combine the advantages of both, we developed an intermediate-sized rabbit animal model for cartilage imaging studies.

Procedures: Rabbit mesenchymal stem cells (rMSC) were isolated as primary cultures from the bone marrow of New Zealand white rabbits. rMSC were subsequentially transduced lentivirally with eGFP and magnetically labeled with the iron oxide ferucarbotran. eGFP expression was evaluated by flow cytometry and iron uptake was analyzed by isotope dilution mass spectrometry and Prussian blue staining. Fluorescence microscopy of eGFP-transduced rMSC was performed. Viability and induction of apoptosis were assessed by XTT and caspase-3/-7 measurements. The chondrogenic potential of labeled cells was quantified by glycosaminoglycan contents in TGF-β3 induced pellet cultures. Labeled and unlabeled cells underwent magnetic resonance imaging (MRI) at 1.5 T before and after differentiation using T1-, T2-, and T2*-weighted pulse sequences. Relaxation rates were calculated. rMSCs were implanted in fibrin clots in osteochondral defects of cadaveric rabbit knees and imaged by 7 T MRI. T2* maps were calculated. Statistical analyses were performed using multiple regression models.

Results: Efficiency of lentiviral transduction was greater than 90 %. Fluorescence signal was dose dependent. Cellular iron uptake was significant for all concentrations (p < 0.05) and dose dependent (3.3-56.5 pg Fe/cell). Labeled rMSC showed a strong, dose-dependent contrast on all MR pulse sequences and a significant decrease in T2 and T2* relaxation rates. Compared with non-transduced or unlabeled controls, there were no adverse effects on cell viability, rate of apoptosis, or chondrogenic differentiation. MRI of labeled rMSCs in osteochondral defects showed a significant signal of the transplant with additional high-resolution anatomical information.

Conclusions: This intermediate-sized rabbit model and its bifunctional labeling technique allow for improved depiction of anatomic detail for noninvasive in vivo rMSC tracking with MRI and for immunohistological correlation by fluorescence microscopy.