Biodistribution of 89Zr-oxine-labeled human bone marrow-derived mesenchymal stem cells by micro-PET/computed tomography imaging in Sprague-Dawley rats.

PURPOSE: To develop a method for labeling human bone marrow mesenchymal stem cells (hMSCs) with 89Zr-oxine to characterize the biodistribution characteristics of hMSCs in normal Sprague-Dawley (SD) rats in real-time by micro-PET-computed tomography (micro-PET/CT) imaging.
METHODS: 89Zr-oxine complex was synthesized from 89Zr-oxalate and 8-hydroxyquinoline (oxine). After hMSCs were labeled with the 89Zr-oxine complex, the radioactivity retention, viability, proliferation, apoptosis, differentiation, morphology, and phenotype of labeled cells were assessed. The biodistribution of 89Zr-oxine-labeled hMSCs in SD rats was tracked in real-time by micro-PET/CT imaging.
RESULTS: The cell labeling efficiency was 52.6 ± 0.01%, and 89Zr-oxine was stably retained in cells (66.7 ± 0.9% retention on 7 days after labeling). Compared with the unlabeled hMSCs, 89Zr-oxine labeling did not affect the biological characteristics of cells. Following intravenous administration in SD rats, labeled hMSCs mainly accumulated in the liver (7.35 ± 1.41% ID/g 10 days after labeling, n = 6) and spleen (8.48 ± 1.20% ID/g 10 days after labeling, n = 6), whereas intravenously injected 89Zr-oxalate mainly accumulated in the bone (4.47 ± 0.35% ID/g 10 days after labeling, n = 3).
CONCLUSION: 89Zr-oxine labeling and micro-PET/CT imaging provide a useful and non-invasive method of assessing the biodistribution of cell therapy products in SD rats. The platform provides a foundation for us to further understand the mechanism of action and migration dynamics of cell therapy products.

Introduction

Cell therapy plays an important role in the treatment of a variety of diseases by repairing damaged tissue or replacing defective cells. The application of cell therapy requires a greater understanding of the fate of cells after transplantation. The technical guidelines for the study and evaluation of cell therapy products indicate that biodistribution and survival time after transplantation of cell therapy products are important factors affecting the effectiveness and safety of cell therapy products, and one or more appropriate cell tracing methods should be used to assess cell distribution, migration, and homing characteristics. The technical options for cell tracing include imaging, quantificational real-time PCR (Q-PCR), and immunohistochemistry. The latter two techniques require the sacrifice of animals after cell transplantation and the removal of corresponding tissues for testing, which has certain limitations. Molecular imaging technology can access the biodistribution and migration characteristics of transplanted cells in real-time after cell transplantation [1,2]. Many radionuclide cell labeling methods based on single-photon emission computed tomography (SPECT) [3-5] and PET [6-12] imaging have been used for cell tracking studies. Radionuclide labeled cells can detect picomolar concentrations or even lower concentrations of radiolabeled substances, therefore, radionuclide labeling is usually one of the methods of cell tracing. Compared to PET, SPECT itself has lower sensitivity and resolution, so SPECT requires higher radioactivity, which can lead to cell damage. PET is at least 10 times more sensitive than SPECT, reducing the amount of radiation exposure of cells by a logarithm [13], and therefore, PET offers the potential to reduce cytotoxicity [14,15]. To achieve reduced radiation exposure while still achieving high sensitivity, high resolution, and sufficient duration to track cells for several days, a long-lived positron-emitting radioisotope is required. Zirconium 89 (89Zr) has become an attractive radionuclide by monitoring labeled cells up to 2–3 weeks by PET imaging.

There are two commonly used methods for cell labeling with 89Zr. One is to synthesize the 89Zr-DBN complex by 89Zr with p-NCS-Bz-DFO cheating, and the 89Zr-DBN complex easily binds to cell membrane surface proteins to complete cell labeling. The other is the synthesis of the 89Zr-oxine complex by 89Zr-oxalate and 8-hydroxyquinoline (oxine), and the high lipid soluble oxine serves as the carrier of 89Zr, making 89Zr easy to penetrate cell membrane and firmly bind to the protein or cytoplasmic components within the cell. In this study, we used the second cell labeling method, which employed a two-step procedure (Fig. 1): (1) preparation of 89Zr-oxine complex, (2) labeling of human bone marrow-derived mesenchymal stem cells (hMSCs) using 89Zr-oxine complex. We assessed the radioactivity retention in cells after labeling as well as the viability, proliferation, apoptosis, differentiation, morphology, and phenotype post-labeling. In addition, the biodistribution and migration of 89Zr-oxine-labeled hMSCs in normal Sprague–Dawley (SD) rats were traced by micro-PET–computed tomography (micro-PET/CT) imaging. The objective of this study was to establish a method for labeling hMSCs with 89Zr-oxine complexes to characterize the biodistribution characteristics of hMSCs in normal SD rats in real-time by micro-PET/CT imaging.

Fig. 1

Schematic for the synthesis of 89Zr-oxine and cell labeling.

Materials and methods

Rats and cells

All animal experiments were approved by the institutional animal care and use committee of our institution. SD rats were purchased from Zhejiang Weitonglihua Experimental Animal Technology Co., Ltd (Animal license No. SYXK Jiangsu 2021-0054). hMSCs were obtained from a commercial source (Cyagen, Suzhou, China) and cultured in hMSCs complete medium (hMSCs basal medium + 10% fetal bovine serum + 1% penicillin-streptomycin + 1% glutamine) after resuscitation. Cell cultures were maintained in a humidified atmosphere of 95% air with 5% carbon dioxide (CO2) at 37 °C.

Synthesize of 89Zr-oxine complex

89Zr-oxine complex was generated by conjugating oxine to 89Zr-oxalate at room temperature [6,8]. 89Zr-oxalate (PerkinElmer, 150 μl, 160 MBq) was neutralized to pH 7.0–7.5 with 40 μl 1 M HEPES (Gibco) and 180 μl 1 M Na2CO3. Oxine in 1 M HEPES (500 μl, 1 mg/ml; Sigma-Aldrich, Saint Louis, USA) was added and spun in a vortex for 60 minutes at room temperature. Then, added 500 μl of chloroform, and extracted with shaking for 20 minutes, collected the lower layer (chloroform phase) liquid and dried it completely with a vacuum centrifuge concentrator. Finally, 100 μl of dimethyl sulfoxide (DMSO; Sigma-Aldrich) was added, sonicated to fully dissolved to obtain 89Zr-oxine complex (100 μl, 95 MBq).

Cell labeling with 89Zr-oxine complex

Cells were detached under mild conditions in 0.25% Trypsin-0.04% EDTA (Cyagen) for 3 minutes and centrifuged at 1200 rpm for 5 minutes. The cell pellet was resuspended in HHBS culture solution (Gibco, Hank’s Balanced Salt Solution with 20 mM HEPES buffer) instead of complete medium, because resuspension of cells using complete medium would decrease labeling efficiency [8]. Diluted 50 μl of 89Zr-oxine solution (40.0 MBq) to 5 ml (for adjusting DMSO ≤2%) with HHBS solution, then added to 4.5 ml hMSCs cell suspension (2.2 × 107 cells) and incubated for 30 minutes at room temperature for 30 minutes on a shaker (650 rpm). Then centrifuged at 1200 rpm for 5 minutes and determined the radioactivity in the supernatant and cell pellet with a radioactivity meter (CAPINTEC, CRC-55tR), respectively. Cells were then washed three times with 3 ml HHBS solution until the supernatant radioactivity/cell pellet radioactivity ≤10%. The cell labeling efficiency of 89Zr-oxine was calculated by dividing the final cell pellet radioactivity by the total radioactivity used for labeling.

Determination of efflux of 89Zr from 89Zr-oxine-labeled cells

To assess retention of 89Zr in cells, 89Zr-oxine-labeled hMSCs were seeded into a 24-well culture plate at a density of 1 × 105 cells/well and cultured in an incubator at 37 °C, 95% air, and 5% CO2. One (1), 2, 3, 4, 5, and 7 days after seeding, three wells were randomly selected each time point. The supernatant and cell pellet were collected separately, with a γ counter (Wizard2 3470) to detect the radioactivity count of each well supernatant and cell pellet separately. The percentage retention of 89Zr in the labeled cells was calculated by cell pellet radioactivity count/(cell pellet radioactivity count + supernatant radioactivity count) × 100%.

Determination of cell viability and proliferation of 89Zr-oxine-labeled cells

89Zr-oxine-labeled and unlabeled hMSCs were each seeded in 24-well plates at a density of 1 × 105 cells/well. At 1, 2, 3, 4, 5, and 6 days after seeding, two wells were randomly selected from labeled and unlabeled cells each day, respectively, and cell viability was assessed by trypan blue (Solarbio) staining. Also, 89Zr-oxine labeled and unlabeled hMSCs were inoculated into a 96-well plate at a density of 1 × 104 cells/well, respectively. One (1), 2, 3, 4, 5, and 7 days after inoculation, 10 μl of CCK-8 (Cell Counting Kit-8; Abcam, Cambridge, UK) reagent was added into three wells of labeled and 3 wells of unlabeled cells, and incubated in a CO2 incubator for 2 h. The absorbance values of each well at 450 nm were then read using a microplate reader (SpectraMax i3x; Molecular Devices, Shanghai, China) to calculate the cell viability at each time point.

Determination of apoptosis of 89Zr-oxine-labeled cells

Annexin V-FITC and propyl iodide (PI; Sigma-Aldrich) were used for fluorescent labeling of labeled and unlabeled hMSCs. Apoptosis was assessed by flow cytometry (CytoFLEX S; Beckman, Suzhou, China). Annexin V+/PI− cells were considered to be early apoptotic cells and Annexin V+/PI− cells were dead cells.

Determination of differentiation functionality of 89Zr-oxine-labeled cells

To assess the effect of 89Zr-oxine labeling on the ability of hMSCs to differentiate into osteoblast lineages, 89Zr-oxine labeled and unlabeled hMSCs were inoculated into two six-well plates at a density of 2 × 105 cells/well respectively. Three wells selected in each plate were incubated with hMSCs osteogenic differentiation medium (Cyagen) while the other three were incubated with hMSCs complete medium as un-induced control. The osteogenic differentiation medium consisted of FBS, penicillin-streptomycin, glutamine, ascorbate, β-Glycerophosphate, and dexamethasone. Osteogenic differentiation was determined using alizarin red staining.

The effect of 89Zr-oxine labeling on the ability of hMSCs to differentiate into the adipogenic cell lineage was also tested using hMSCs adipogenic differentiation medium consisted of FBS, penicillin-streptomycin, glutamine, insulin, IBMX, rosiglitazone, and dexamethasone. Change the medium, which every 2–3 days, and after 3 weeks of incubation, the adipogenic differentiation effect of the cells was determined with oil-red O dye.

Analysis of Phenotype of 89Zr-oxine-labeled cells

89Zr-oxine-labeled and unlabeled hMSCs were evaluated for their surface expression of CD14, CD19, CD73, CD90, CD105, CD34, CD45, and HLA-DR with flow cytometry (CytoFLEX S; Beckman). Cells (106 cells in volume of 98 µl) were incubated with antibodies (2 µl) that had been labeled with different fluorescent dyes (CD73-PE-Vio770, 130-111-910; CD105-APC, 130-112-166; CD90-PE, 130-114-860; CD14-FITC, 130-110-518; CD19-FITC, 130-113-645; CD34-PE, 130-120-515; CD45-FITC, 130-110-631; HLA-DR-FITC, 130-111-788) for 30 minutes, and then washed cells twice with DPBS (washed off fluorescent antibodies that were not bound to cells). Added 500 μl of DPBS to resuspend the cell pellet, and the expression of surface markers for 89Zr-oxine-labeled and unlabeled hMSCs was detected with flow cytometry. All antibodies were purchased from the Miltenyi Biotechnology Company.

Micro-PET/CT imaging of 89Zr-oxine-labeled human bone marrow mesenchymal stem cells and 89Zr-oxalate

A group of 3 male and 3 female rats were administered 3 × 106 89Zr-oxine-labeled hMSCs (approximately 3.7 MBq). An additional of 3 female rats were administered 89Zr-oxalate (89Zr-oxalate was first adjusted to neutrality with 1M Na2CO3 and then diluted with HHBS to the desired radioactivity) of the same radioactivity. Micro-PET/CT (Inveon, SIMENS, Berlin, Germany) scans of rats were performed sequentially at 1 h, 1 day, 3 days, 7 days, and 10 days after intravenous injection to obtain whole-body distribution images of radioactive signals at different time points. Before scanning, the rats were anesthetized with an isoflurane gas anesthesia machine (a mixture of 3% isoflurane and oxygen), and then the rats were held in a prone position and fixed in a micro-PET/CT scanner, and the rats were supplied with an anesthesia mask to inhale 1.5% isoflurane (a mixture of 1.5% isoflurane and oxygen) to maintain the rat’s anesthesia state during imaging. The voltage of the CT scan was 80 kV. Inveon Acquisition Workplace software was used for image acquisition and Inveon Research Workplace software was used for data processing. Image reconstruction used an ordered-subset expectation maximization algorithm. Image processing was performed by PMOD software and then the 3D regions of interest (ROIs) were manually plotted around the heart, liver, spleen, lungs, kidneys, brain, and bone tissue based on the contrast of CT soft tissues. The percentage of % injection dose per gram tissue (% ID/g) (see Tables S1 and S2, Supplemental digital content 1, http://links.lww.com/NMC/A218) and standardized uptake values (see Tables S3 and S4, Supplemental digital content 1, http://links.lww.com/NMC/A218) were calculated using the attenuation corrected ROIs value.

The management and use of animals during the experiments complied with the requirements in the guide of the care and use of laboratory animals issued by the US National Research Council, the administration of laboratory animals issued by the Chinese Science and Technology Commission in 1988, and the Jiangsu Provincial approach to the management of laboratory animals issued in 2008.

Statistical analysis

Statistical analysis and plotting were performed using SPSS 23.0 and GraphPad Prism 9 software. All data were expressed as mean ± SD. Statistical analysis was performed using Student’s t-test and the one-way analysis of variance. A value of P < 0.05 or P < 0.01 was considered statistically significant difference.

Results

Labeling efficiency and cellular 89Zr retention

2.2 × 107 hMSCs were incubated with 40.0 MBq 89Zr-oxine for 30 minutes to obtain an average cell labeling efficiency of 52.6 ± 0.01% (n = 2), sufficient for imaging needs. Labeling efficiency of 89Zr-oxine was higher than the previous methods [8], possibly due to the use of different cell types or higher cell concentrations in this study, resulting in the difference in the amount of intracellular proteins available for 89Zr binding. The procedure resulted in an initial specific activity of approximately 1.3 MBq/106 cells. The percentage retention of 89Zr-oxine in the cells on days 1, 2, 3, 4, 5, and 7 after labeling was 87.3 ± 1.9%, 86.7 ± 0.2%, 83.3 ± 0.8%, 81.1 ± 1.0%, 78.1 ± 0.9%, and 66.7 ± 0.9%, respectively (Fig. 2). The retention rate of 89Zr-oxine in cells decreased in a time-dependent manner.

Fig. 2

Retention of 89Zr in hMSCs at 1, 2, 3, 4, 5, and 7 days after labeling. Each point was representative of the mean ± SD of three determinations.

Cell viability, proliferation, and apoptosis of 89Zr-oxine-labeled human bone marrow mesenchymal stem cells

To assess 89Zr-oxine cytotoxicity, hMSCs were labeled with different radiation doses of 89Zr-oxine (2.2, 3.7, and 5.6 MBq/106 cells). Labeling efficiency correlated negatively with 89Zr-oxine amount, with the highest labeling efficiency (59.4%) and cell viability (84.9%) at the lowest doses of 89Zr-oxine (see Table S5, Supplemental digital content 1, http://links.lww.com/NMC/A218). High labeling efficiency and viability could still be achieved when the amounts of cells and 89Zr-oxine were scaled up according to the ratio of 2.2 MBq/106 cells (see Table S6, Supplemental digital content 1, http://links.lww.com/NMC/A218), and this ratio of cells and 89Zr-oxine was used for subsequent experiments. Within 6 days after labeling, 89Zr-oxine-labeled hMSCs exhibited cell viability similar to unlabeled cells (Fig. 3, P = 0.78). The results of the CCK-8 assay showed that the proliferation level of 89Zr-oxine-labeled hMSCs was slightly reduced compared to unlabeled cells 3 days after labeling (Fig. 4, P < 0.01). Apoptosis detection by flow cytometry showed that the apoptosis rate of 89Zr-oxine-labeled hMSCs (12.0%) was slightly higher than that of unlabeled cells (6.4%) (Fig. 5).

Fig. 3

Determination of viability of 89Zr-oxine-labeled hMSCs. 89Zr-oxine-labeled cells demonstrated similar survival as compared with unlabeled cells up to 7 days after labeling (P = 0.78). Data were representative of two independent experiments. All data were expressed as the mean ± SD.

Fig. 4

Determination of cell proliferation of 89Zr-oxine-labeled hMSCs. The proliferative level of the 89Zr-oxine-labeled hMSCs was significantly reduced from 3 to 7 days after labeling, as compared with that of unlabeled. Data were representative of three independent experiments. All data were expressed as the mean ± SD. **The difference was significant at the 0.01 level.

Fig. 5

Determination of apoptosis of 89Zr-oxine-labeled cells. Apoptosis of 89Zr-oxine-labeled and unlabeled hMSCs was detected by flow cytometry, which showed that radioisotope labeling slightly increased the degree of apoptosis. P > 0.05, there was no statistically significant difference.

89Zr-oxine labeling did not affect the characterization of human bone marrow mesenchymal stem cells

Inverted phase-contrast microscopy showed that 89Zr-oxine-labeled hMSCs had the same typical fibroblast-like morphology as unlabeled hMSCs. (see Fig. S1, Supplemental digital content 1, http://links.lww.com/NMC/A218). In addition, we found that 89Zr-oxine-labeled hMSCs had the same differentiation function as unlabeled hMSCs. After 3 weeks of adipogenic differentiation induction, a large number of adipocytes were produced in both 89Zr-oxine-labeled and unlabeled hMSCs, and lipid-filled vesicles were stained red with oil-red O (Fig. 6a and c). After 3 weeks of induction culture for osteogenesis differentiation, there were large amounts of calcium nodule deposition in both the 89Zr-oxine labeled and unlabeled hMSCs, stained red by alizarin red (Fig. 6e and g), and the uninvited cultured cells acted as a negative control (Fig. 6b, d, f, and h).

Fig. 6

Determination of differentiation functionality of 89Zr-oxine-labeled hMSCs. 89Zr-oxine-labeled hMSCs were differentiated towards both the osteogenic (c) and the adipogenic (g) lineages as did as unlabeled cells (a, e), with uninduced cells (b, d, f, h) serving as negative controls. Bar = 100 μm.

The effect of 89Zr-oxine labeling on hMSCs specific surface marker expression was evaluated by flow cytometry according to the ISCT-approved minimum criteria for identifying mesenchymal stromal cells (MSCs) [16]. 89Zr-oxine-labeled hMSCs showed the same surface marker expression as unlabeled cells (positive expression of CD73, CD90, and CD105; negative expression of CD14, CD19, HLA-DR, CD34, and CD45) (Fig. 7).

Fig. 7

89Zr-oxine-labeled hMSCs showed surface marker expression same as that of unlabeled cells (positive expression of CD73, CD90, and CD105; negative expression of CD14, CD19, HLA-DR, CD34, and CD45).

In-vivo tracking of 89Zr-oxine-labeled human bone marrow mesenchymal stem cells with micro-PET/CT

The biodistribution of radioactive signal was visualized and quantified by micro-PET/CT at 1 hour and 1, 3, 7, 10 days after injection of 89Zr-oxine-labeled hMSCs and 89Zr-oxalate (Fig. 8, Figs. S2 and S3, Supplemental digital content 1, http://links.lww.com/NMC/A218). Consistent with previous studies [8,12], the distribution of 89Zr-oxine-labeled hMSCs was mainly concentrated in the lungs (17.90 ± 3.63% ID/g) 1 hour after injection (Fig. 9) and migrated to the liver (10.55 ± 2.17% ID/g) and spleen (9.25 ± 1.17% ID/g) 1 day after injection (Fig. 10). From 1 h to 1 day after injection, pulmonary uptake decreased from 17.90 ± 3.63% ID/g to 0.66 ± 0.18% ID/g. Up to 10 days after injection, there was still a large intake in the liver (7.35 ± 1.41% ID/g) and spleen (8.48 ± 1.20% ID/g) (Fig. 11). Unlike the biodistribution of hMSCs, the biodistribution of 89Zr-oxalate was relatively dispersed 1 h after injection, with a relatively high uptake in the heart (3.77 ± 0.75% ID/g). Migration to bone (3.57 ± 0.06% ID/g) was mainly seen 1 day after injection, and large uptakes (4.77 ± 0.35% ID/g) were still visible in bone 10 days after injection. The biodistribution of 89Zr-oxalate indicated that the low radioactivity seen in the bone (0.96 ± 0.33% ID/g, 10 days after labeling) of rats injected 89Zr-oxine-labeled hMSCs might be derived from free 89Zr released from dead hMSCs.

Fig. 8

Micro-PET/CT images of 89Zr-oxine-labeled hMSCs and 89Zr-oxalate in SD rats. (a) PET/CT whole-body images of the biodistribution of 89Zr-oxine-labeled hMSCs in SD rats over the following ten days (n = 6). (b) PET/CT whole-body images of the biodistribution of 89Zr-oxalate in SD rats over the following ten days (n = 3).

Fig. 9

The biodistribution of 89Zr-oxine-labeled hMSCs and 89Zr-oxalate at 1 hour post-injection. 89Zr-oxine-labeled hMSCs were mainly distributed in the lung (17.90 ± 3.63% ID/g). The distribution of 89Zr-oxalate was relatively discrete at 1 hour after injection, with signal uptake in the heart, liver, spleen, lung, kidney, and bone, and relatively higher uptake in the heart (3.77 ± 0.75% ID/g). *: the difference from that of the control was significant at the P < 0.05 level. **: the difference from that of the control was significant at the P < 0.01 level. All data were expressed as means ± SD.

Fig. 10

The biodistribution of 89Zr-oxine-labeled hMSCs and 89Zr-oxalate at 1 day post-injection. 89Zr-oxine-labeled hMSCs were mainly distributed in the liver (10.55 ± 2.17% ID/g) and spleen (9.25 ± 1.17% ID/g). 89Zr-oxalate was mainly distributed in the bone (3.57 ± 0.06% ID/g). **The difference was significant from that of the control at P < 0.01 level. All data were expressed as means ± SD.

Fig. 11

The biodistribution of 89Zr-oxine-labeled hMSCs and 89Zr-oxalate at 10 days post-injection. 89Zr-oxine-labeled hMSCs were mainly distributed in the liver (7.35 ± 1.41% ID/g) and spleen (8.48 ± 1.20% ID/g). 89Zr-oxalate was mainly distributed in the bone (4.77 ± 0.35% ID/g). **The difference was significant from that of the control at P < 0.01 level. All data were expressed as means ± SD.

Discussion

As we know, MSCs, whether from humans or from other species, have a low immunogenic cellular phenotype [17-19]. MSCs therapy is considered an effective immunomodulatory and regenerative therapy in a variety of animal models, including transplant models [20], experimental colitis [21], rheumatoid arthritis [22], bone/cartilage injury [23,24], and a number of other models. Although previous studies have described a series of immune responses related to the immunomodulatory effects of MSCs, such as inducing the production of cells with regulatory phenotypes, inhibiting inflammatory cells and secreting anti-inflammatory molecules [25-28], the mechanism of action of MSCs has not been fully determined, and the fate of cells after transplantation has not been fully understood [29]. Therefore, cell tracking technique is needed to assess cell biodistribution, which is helpful to further understand the mechanism of many emerging cell-based therapies. One of the most promising ways to monitor the biological distribution of MSCs is cell labeling with tracers such as 89Zr-oxine. The effects of the 89Zr-oxine marker on cell viability and proliferation have been reported [6,8,10-12], and these effects may vary depending on cell types. In this study, we evaluated the effects of the 89Zr-oxine labeling on the cellular functionality and biological characteristics of hMSCs, analyzed the radioactivity retention, cell viability, proliferation, apoptosis, differentiation, morphology, and phenotype in cells after the 89Zr-oxine labeling in vitro, and used micro-PET/CT to initially characterize the biodistribution and migration of cells in rats.

The trypan blue staining test showed that 89Zr-oxine was incubated with hMSCs for 30 min at a labeled dose of 2.2 MBq/106 cells, with no negative effect on the viability of hMSCs. However, cell viability decreased significantly at labeled doses of 3.7 and 5.6 MBq/106 cells, suggesting that cell viability decreased in a dose-dependent manner after 89Zr-oxine in-vitro labeling. The effect of the different radiation doses of 89Zr-oxine on radioactivity retention in cells was also examined (Table S7, Supplemental digital content 1, http://links.lww.com/NMC/A218), and as the same effect on cell viability, 89Zr-oxine had a significant dose-dependent effect on radioactivity retention in cells. The retention of 89Zr in cells was the key to evaluate the success of labeling. Once 89Zr-oxine enters the cell, the 89Zr remains stably inside the living cell, and once the labeled cell dies, it causes the free 89Zr to be released rapidly from the cell [8]. Therefore, cell mortality can be reflected according to the outflow rate of 89Zr. We also tried labeling cells with a lower radioactive dose of 89Zr-oxine (1.5 MBq/106 cells), but the specific radioactivity of the cells after labeling was too low to complete a continuous 10-day in-vivo tracer. Therefore, the labeling dose of 2.2 MBq/106 cells was used for subsequent study. Labeling efficiency of 52.6 ± 0.01% was achieved, which was sufficient for imaging needs. The retention of 89Zr in cells decreased time-dependently, with intracellular retention of 66.7 ± 0.9% at 7 days after labeling (Fig. 2), compared with a recent study using 89Zr-oxine labeling MSCTRAIL, and the retention of 89Zr in cells was greatly improved in this study [12]. Within 6 days after labeling, 89Zr-oxine-labeled hMSCs showed simar viability to unlabeled cells (Fig. 3, P = 0.78). Compared to unlabeled cells, the 89Zr-oxine labeling had a slight effect on cell proliferation, and cell proliferative activity decreased from 3 days after labeling (Fig. 4). Cells are known to be most sensitive to radioactivity during the division phase and lowest in the DNA synthesis phase, so it can be assumed that cells are in the DNA synthesis phase 1–2 days after seeding, and begin to enter the cell division phase 3 days after seeding. At this time, the sensitivity of cells to radioactivity is the highest, resulting in a large number of cells being killed, and a decrease in cell proliferation. 89Zr-oxine labeling slightly increased the level of apoptosis (Fig. 5, P > 0.05), which may be related to repeated cleaning and blowing of cells during the cell labeling process, or may be related to radiotoxicity or chemical toxicity of oxine, but the overall effect was small. We also found that the 89Zr-oxine labeling did not affect the basic stem cell properties [30]. 89Zr-oxine-labeled hMSCs showed the persistence capability to differentiate into the adipogenic and osteogenic lineages (Fig. 6). In addition, 89Zr-oxine labeling did not affect either the normal cell morphology of hMSCs or the expression of stem cell-specific surface markers (positive expression of CD73, CD90, and CD105; negative expression of CD14, CD19, HLA-DR, CD34, and CD45). If 89Zr-oxine labeling interferes with cell function, it may result in differences in the actual distribution of labeled cells versus unlabeled cells. In conclusion, the labeling method used in this study did not interfere with the functional and biological characteristics of hMSCs.

We also observed the biodistribution and migration dynamics of intravenous 89Zr-oxine-labeled hMSCs in normal SD rats with micro-PET/CT imaging. As expected, the overall biodistribution observed with 89Zr-oxine-labeled hMSCs (Fig. 8a, initially uptake in the lungs followed by migration to the liver, spleen, and bone) was consistent with earlier reports in mice imaging studies [12,31,32], as well as in clinical trials [33]. One hour after injection, most (66%) radioactive signals were found in the lungs, and small amounts were also found in the liver (18%) and spleen (10%). After 1 day, the level of radioactivity in the lungs (3%) decreased significantly, while the level of radioactivity in the liver (46%) and spleen (40%) increased significantly. Ten days after the injection, there was still a large amount of radioactivity in the liver (37%) and spleen (43%), in addition, a small amount of radioactivity (5%) was also found in the bone. Radioactivity in other organs was around background levels. As a control, we found that intravenously injected free 89Zr-oxalate resulted in a relatively discrete distribution of 89Zr-oxalate at 1 hour, with radioactivity uptake in heart, liver, spleen, lung, kidney, and bone tissue but relatively high uptake in heart (3.77 ± 0.75% ID/g), which was not seen with labeled hMSCs. From 1 day to 10 days after injection, signals of bone uptake gradually accumulate, this bone-targeting phenomenon of 89Zr was consistent with that reported previously [34]. Thus, the loss of 89Zr in cells could be estimated by the radioactivity of the bones. Eggenhofer et al. demonstrated that some exogenous MSCs survived in the lungs within 24 hours of intravenous injection. After 24 hours, living MSCs disappeared from the lungs, but these lived MSCs did not reappear in the liver, spleen, kidneys, or bones, suggesting that they could not survive long-term in the recipient animals [31]. In this study, radioactivity signals were also found in the liver and spleen from 1 day to 10 days after injection, but we were unable to detect hMSCs in these organs by Q-PCR at 10 days post-injection (below the minimum quantitative limit), suggesting that Q-PCR analysis required a higher detection threshold compared to radiolabeling (e.g. a 10-fold increase in the number of hMSCs required for cardiac Q-PCR detection [35]). According to Eggenhofer et al., even if the number of hMSCs reaches the threshold required for Q-PCR analysis, the radioactive signals in the liver and spleen might originate from hMSCs fragments or other endogenous cells that engulf hMSCs fragments, rather than live hMSCs. Because living hMSCs could not pass through the capillary bed of the lungs after intravenous injection. Immune cells may well be involved in this process. It had been hypothesized that apoptosis of infused cells could trigger an immunomodulatory response [36] and recently it was demonstrated that macrophages adapted an immunoregulatory function after phagocytosis of the dead hMSCs [37]. Further research must reveal which signals conduct this response through the body.

One potential limitation of direct cell labeling techniques is that cells inevitably lose a portion of the radiolabeled substance, resulting in in-vivo imaging results that cannot determine whether the signal is from the labeled cell or from an endogenous cell that engulfs the outflow signal, or from dead labeled cell debris, a common problem with all direct cell labeling techniques. To overcome this limitation, genetic labeling strategies may be more appropriate for long-term tracking of transplanted cells in vivo. This method requires genetic modification of donor cells in vitro to stably express a reporter protein that can selectively bind to the tracer [38]. Currently, although in-vivo imaging modalities have been largely used in studies of small animals, our ultimate goal is to apply this groundbreaking technology to the clinic to aid in understanding the distribution and fate of different stem cells in the human body. Large animals (pigs, dogs, monkeys, etc.) are most similar to humans in terms of body size, disease progression, body weight, and overall anatomy [39]. Therefore, a lot of researches are still needed on these animal models. In summary, we still have a lot of work to do in order to elucidate the mechanism and fate of hMSCs after transplantation and provide more reliable information for clinical applications.

Conclusion

This study showed that 89Zr-oxine labeling preserved the functionality and characteristics of hMSCs, demonstrating the feasibility of this platform for tracking systemically delivered cells. This labeling method could be a valuable tool for the preclinical noninvasive assessment of various cell-based therapies. To further explore the potential of this approach, more work is needed to test this strategy in a variety of model systems and diseases treatments relevant to cell tracking and cell therapy.

Acknowledgements

The authors sincerely thank all members of the Small Animal live Imaging and Flow Cytometry Research team at Innostar biotechnology Nantong Co., Ltd. We also thank Jiangsu Huayi Technology Co. Ltd. provided the micro-PET/CT intravital scanning instrument and the testing site for this experiment.

This work was partly funded by the China’s National Key R&D Program ‘Stem Cell and Transformation Research’ Special Project 2020YFA0112600 and Special project of engineering technology research center of Shanghai Science and Technology Commission 17DZ2252900. Our institution has filed a patent 202111459230.0 assigned to the Shanghai Bixing Law firm on a method to label hMSCs with 89Zr.

Shuzhe Wang, Yan Wang, Bohua Xu, and Yunliang Qiu designed and coordinated the research. Shuzhe Wang, Yan Wang, Tian Qin, Yupeng Lv, Heng Yan, Yifei Shao, Yangyang Fang, and Shaoqiu Zheng performed the research. Shuzhe Wang and Yan Wang analyzed experimental data and write the initial draft. All authors contributed to the final manuscript.

Conflicts of interest

There are no conflicts of interest.