In vivo human adipose-derived mesenchymal stem cell tracking after intra-articular delivery in a rat osteoarthritis model

Abstract

Background: Human adipose-derived mesenchymal stem cells (haMSCs) have shown efficacy in treating osteoarthritis (OA) both preclinically and clinically via intra-articular (IA) injection. However, understanding the mode of action of the cell therapy has been limited by cell tracking capability and correlation between the pharmacokinetics of the injected cells and the intended pharmacodynamics effect. This study aims to explore methodology and to understand in vivo biodistribution of clinical-grade haMSCs labeled with fluorescent dye and injected into an immunocompetent OA rat model.

Methods: haMSCs labeled with fluorescent dye were investigated for their proliferation and differentiation capabilities. Labeled cells were used to establish detection threshold of a noninvasive biofluorescent imaging system before the cells (2.5 × 10⁶) were injected into a conventional rat OA model induced by medial meniscectomy for 8 weeks. We attempted to reveal the existence of labeled cells in vivo by imaging and a molecular biomarker approach, and to correlate with the in vivo efficacy and physical presence over a follow-up period up to 10 weeks.

Results: In vitro proliferation and differentiation of haMSCs were not affected by the labeling of DiD dye. Detection thresholds of the labeled cells in vitro and in vivo were determined to be 10⁴ and 10⁵ cells, respectively. When 2.5 × 10⁶ haMSCs were injected into the joints of a rat OA model, fluorescent signals (or >10⁵ cells) lasted for about 10 weeks in the surgical knee joint at the same time as efficacy was observed. Signals in nonsurgical rats only lasted for 4 weeks. The human MSCs were shown to engraft to the rat joint tissues and were proliferative. Human FOXP2 gene was only detected in the knee joint tissue, suggesting limited biodistribution locally to the joints.

Conclusions: The current study represents the first attempt to correlate cell therapy efficacy on OA with the physical presence of the injected haMSCs in the OA model, and demonstrates that human adipose-derived mesenchymal stem cells persisted for 10 weeks locally in the rat joint, coinciding with the efficacy observed. It is postulated that persistence and/or proliferation of the haMSCs in the joint is required in order to exert their functions on promoting joint regeneration and/or cartilage protection, further supporting the safety and feasibility of IA injection of MSCs for the treatment of OA patients.

Keywords: Adipose-derived mesenchymal stem cells; Biodistribution; Carbocyanine dyes; Intra-articular injection; Osteoarthritis.

Fig. 1

Characterization of DiD-labeled human adipose-derived mesenchymal stem cells (haMSCs) showing the labeling efficiency and unaltered surface markers and proliferation in vitro. a The high efficiency of DiD labeling, comparing cells positive in the fluorescent image with those in the bright field (scale bars = 50 μm). b After labeling, haMSCs were positive for CD90 and CD105, and negative for the cocktail including CD34, CD11b, CD19, CD45, and HLA-DR, the same as the unlabeled haMSCs. c Cell viability assay by OD measurement with CCK8 kit confirmed no significant variation in proliferation capacity of labeled haMSCs at days 1, 3, 5, 7, and 9 compared with unlabeled haMSCs (no statistical significance)

Fig. 2

DiD labeling had no influence on differentiation of human adipose-derived mesenchymal stem cells (haMSCs) in vitro. a Unlabeled and labeled haMSCs differentiated similarly under osteogenic, adipogenic, and chondrogenic conditions, producing robust mineralized matrix-large lipid droplets, and condensing chondrocytes detecting by Alizarin Red S staining, Oil Red O staining, and Alcian Blue staining, respectively. Differentiated haMSCs retained the DiD label, as fluorescence was detected at the end of each differentiation (scale bar = 100 μm). b At the end of culturing, the osteogenic, adipogenic, and chondrogenic gene markers, OSTEOCALCIN, ADIPONECTIN, and CAL2A1 in labeled haMSCs were upregulated during osteogenesis, adipogenesis, and chondrogenesis, respectively. No significant difference was seen between the labeled and unlabeled haMSCs for the differentiation potential, except for the adiponectin expression where labeled haMSCs showed slightly less expression than the unlabeled cells. The P values were obtained using one-way ANOVA analysis of variance (*P < 0.05). Data were collected from three repetitions (n = 3). ns nonsignificant

Fig. 3

Determination of tracking sensitivity of biofluorescent imaging in vitro and in vivo. a Among the serial dilutions, 10⁴ labeled haMSCs displayed a marginal fluorescence signal that could be detected above the phosphate-buffered saline (PBS) background. b Quantification of the fluorescence signal in a showing that the detection threshold in vitro is between 10³ and 10⁴ cells (10⁷ and 10⁸ photons/s/cm²). c Remarkable fluorescent signals were displayed in rats injected with 10⁶ and 10⁵ cells, while signals of 10⁴ cells were undetectable, suggesting that the in vivo detection threshold of DiD-haMSCs was between 10⁵ and 10⁴ cells. d Quantification of the fluorescence signal in c showing that the detection threshold in vivo is between 10⁴ and 10⁵ cells (8 × 10⁷ and 5 × 10⁸ photons/s/cm²). Data were representative of two repetitions (n = 2)

Fig. 4

Signal of DiD-haMSCs residing in the knee joint. a Detection of florescent imaging of representative rats in both nonsurgery and surgery groups over time. When DiD-haMSCs (2.5 × 10⁶ cells) were injected into the right knee joint of the nonsurgery rats (top panel) and surgery rats (bottom panel), signals remained relatively stable in surgery joints and became undetectable after 70 days of injection while in the nonsurgery joints the signal diminished rather quickly at day 28. b Quantitative fluorescent intensity of the nonsurgery and surgery groups. Data show that the average (n = 2) fluorescent intensity appeared to fluctuate as time went on. Signals in the surgery group lasted much longer than in nonsurgery group

Fig. 5

Proliferative human cells were detected in the rat knee joint tissues. a Human skin tissues were used for positive (with monoclonal antihuman mitochondria antibody) and negative (IgG isotype) IHC staining controls. Positive signals were detected in both nonsurgery and surgery groups with human adipose-derived mesenchymal stem cell (haMSC) treatment, but not in the free DiD group (no haMSC treatment). The human cells were mainly detected in the rat meniscus and cartilage. b Human hepatocellular carcinoma cell line HepG2 were used for positive and negative controls with regard to a monoclonal antihuman ki67 antibody IHC staining. Human cells undergoing proliferation were detected in the positive control as well as in the joint tissues of the DiD-labeled haMSC treatment groups, but not in the free DiD group. Samples of the surgery group (n = 2) and the nonsurgery group (n = 2) were both collected at the end of the experiments (day 70 and day 28), respectively

Fig. 6

Threshold of human DNA in rat tissue was determined by quantitative real-time PCR using the human-specific FOXP2 gene probe. DNA from haMSCs were mixed with rat genomic DNA in different concentrations for quantitative PCR showing that 0.032 % human DNA (0.01056 ng human DNA in a total of 33 ng DNA) is the lowest amount detected before reaching the plateau

Fig. 7

Therapeutic efficacy of IA injection of human adipose-derived mesenchymal; stem cells (haMSCs) in the OA model. a In the normal group, HE staining showed natural histology of joints with a thick layer of cartilage and subchondral bone, and chondrocytes in organized lacuna (top panel, 4×, and lower panel, 20×). In contrast, the thickness of the cartilage and especially the subchondral bone in the surgery-induced groups were significantly reduced, and chondrocytes appeared to be more disorganized. haMSC treatment restored the thickness of the cartilage and subchondral bone (scale bar = 500 μm). Note that the meniscus in the surgery groups all displayed damage, confirming the successful surgery of meniscectomy for the model creation. b Safranin-O/Fast green staining for proteoglycan (PG)/collagen content revealed the significant loss of PG (red staining) and increase of fibrillated collagen (green staining), a characteristic of degenerative OA phenotypes in the DiD control group compared with the normal group. Treatment with the DiD-haMSCs exhibited increased proteoglycan expression and decreased type I collagen expression (scale bar = 500 μm), thus restoring the overall cartilage thickness. c Cartilage thickness measured by ImageJ software increased significantly in the haMSC-treatment group compared with the DiD group. d The modified O’Driscoll histological score for morphology and structure characteristics quantified the changes, and supported that haMSC treatment showed significant morphological and structural improvement of cartilage. The P values were obtained using one-way ANOVA analysis of variance (**P < 0.01). All the experiments were repeated at least two times (n = 3)