The T-cell acute lymphoblastic leukemia cell lines, Jurkat and CCRF-CEM and acute lymphocytic leukemia cell line, Reh were purchased from American Type Culture Collection (Manassas, VA) and cultured in RPMI-1640 (Life Technologies, Grand Island, NY) medium supplemented with 10% fetal bovine serum (FBS; Life Technologies), 1% HEPES (Life Technologies), 100 U/ml of penicillin (Life Technologies), and 100 mg/ml of streptomycin (Life Technologies). These cell lines were maintained in a humidified 5% CO₂ atmosphere at 37°C.

Reh, Jurkat and CCRF-CEM cell lines were transduced by addition of 1 ml of viral stock consisting of pLenti6/V5-CMV viral vector (Life Technologies) encoding firefly luciferase. To facilitate the entrance of the viral vector into the cells, 8 μg/ml of polybrene (Santa Cruz Biotechnologies, Santa Cruz, CA) was added. Viable cells were washed with excess volume of phosphate-buffered saline (PBS; Biowest, Nuaille, France) 4 times and incubated directly in RPMI-1640 with 10% FBS, 1% penicillin/streptomycin, 1% HEPES and 5 μg/ml blasticidine (Sigma-Aldrich, St. Louis, MO) for clonal selection. To confirm successful transduction of f-luciferase gene, 1×10⁶ clone cells were seeded in each well of a 6-well plate (Nalge Nunc, Naperville, IL) for bioluminescent imaging. Bioluminescent clones with the most luciferase activity of each cell lines were selected for injection and designated these bioluminescent leukemia cell lines as Reh/fLuc, Jurkat/fLuc and CCRF-CEM/fLuc.

NOD.CB17/scid Arc (NOD/SCID) mice were purchased from The Animal Resources (Canning Vale, WA, Australia) and maintained at the Laboratory Research Animal Center of the Samsung Biomedical Research Institute according to AAALAC approved protocols. Bioluminescent leukemia models were prepared by injecting 7-to 8-week old NOD/SCID mice with 1x10⁶ CCRF-CEM/fLuc via three different injection routes; intraperitoneal, intravenous (tail-vein) and intra-bone marrow (tibia) injection.

These three types (IP, IV and IBMT) of leukemia in vivo models were compared through continuous bioluminescent monitoring using IVIS 100 imaging system (Xenogen Corporation, Alameda, CA). D-luciferin (150 mg/kg) (Xenogen) was injected intraperitoneally to each mouse prior to imaging. Mice were anesthetized with vaporized isofurane (Abbott Laboratory, Abbott Park, IL) and placed in imaging chamber. After 5.5 min, each animal was imaged alone in supine and prone positions with an exposure time of 1 min for each position weekly for 6 weeks. All bioluminescent image data were provided by Living Image software (version 1.0, Xenogen). Photons detected from leukemia models were converted to average radiance (photon/sec/cm²/sr). Average radiance values are quantitative data obtained from region of intensity (ROI) where photons emitted by bioluminescent cells of assigned rectangular area over the whole body of each mouse. Both luminescence and image data were analyzed using Living Image software.

To validate bioluminescent correlation with peripheral leukemic cells of leukemia burden in vivo model, mice were sacrificed at week 6 after bioluminescent images were obtained. Peripheral blood (PB), bone marrow (BM) aspirates, and spleen were obtained to isolate human leukocytes using Ficoll-Paque™ Plus (GE Healthcare, Uppsala, Sweden) solution. Residual red blood cells were removed using erythrocyte lysis buffer (Qiagen, Hilden, Germany) prior to double staining the isolated cells with phycoerythrin (PE)-conjugated anti-human CD45 (hCD45) (BD Pharmigen™, San Diego, CA) and fluorescein isothiocyanate (FITC)-conjugated anti-mouse CD45 (mCD45) (BD Pharmigen™) antibodies. PB and BM of the mice treated with vincristine and methylprednisolone were also analyzed to determine the correlation between bioluminescent changes in responses to vincristine or methylprednisolone within leukemia burdened mice. Mononuclear cells isolated from PB of the second generation IBMT leukemia model was evaluated with hCD45 and mCD45 antibodies once again to validate that human leukemic cell of the first IBMT model were responsible for the development of bioluminescent leukemia model.

In vitro sensitivity of CCRF-CEM/fLuc cell line to vincristine and methylprednisolone was validated using Alamarblue^® (Life Technologies) assay and luminescence assay prior to transplantation. CCRF-CEM/fLuc cells (1×10⁵) were seed in each well of 96-well. Five columns of 4 wells were treated with vincristine at concentrations of 0.1, 0.5, 1, 5, and 10 ng/ml. Cells were exposed to vincristine for 48 h before endpoint data analysis were conducted using Alamarblue assay. Replicate vincristine in vitro assay was done and analyzed with bioluminescent imaging, 2 μl (300 μg/ml) of D-luciferin was added to each well, and image of the in vitro assay plate was obtained. Another 96-well plate seeded with CCRF-CEM/fLuc cells of the identical experimental setting as vincristine was prepared to determine effective cytotoxic concentration of methylprednisolone, 100, 250, 500 and 1000 μg/ml of methylprednisolone were treated for 48 h. Alamarblue and bioluminescent assay were also used to assess in vitro responses of CCRF-CEM/fLuc to methylprednisolone.

Mice in groups of three were treated with two different vincristine (Hospira, Mulgrave, Australia) concentrations starting at week 3. Vincristine (0.1 and 0.5 mg/kg) in 100 μl PBS were injected intravenously into tail vein 3 times in 7 days interval. Control mice were injected with 100 μl PBS every week as well. The effects of vincristine on leukemia mouse models were monitored once a week from the initial treatment for 21 days. Prior to administration of vincristine, bioluminescent images of the mice were obtained. Standard bioluminescent imaging protocol was used to monitor changes in vincristine treated mice. Another set of IBMT leukemia mouse model was prepared according to methods described previously. These mice were treated weekly with two different concentrations of methylprednisolone (Pfizer, Puurs, Belgium) starting at week 3. Methylprednisolone (2 and 10 mg/kg) in 100 μl PBS were administered intravenously into tail vein. Monitoring responses to methylprednisolone was done by weekly imaging and images were obtained according to the protocol described in bioluminescent imaging section of Materials and methods.

CCRF-CEM/fLuc cells were isolated from peripheral blood of IBMT-leukemia mouse model via density gradient removal of RBC using Ficoll solution. Cells were maintained in RPMI-1640 with 10% FBS, 1% antimycotic-antibiotic and 1% HEPES. To assess luciferase expression level of ex vivo CCRF-CEM/fLuc, the number of cells ranging between 10³ and 10⁷ cells were suspended in 100 μl of PBS and plated on flat-bottom 96-well plate. 2 μl (300 μg/ml) of D-luciferin was added to each well, then bioluminescent signal were detected for 30 sec with CCD camera. Bioluminescent leukemia mouse model was reconstituted by injecting 10 μl of 1×10⁶ of the isolated CCRF-CEM/fLuc cells directly into tibia of the secondary recipient mice (n=5). Leukemia development was monitored weekly. Mice were given an intraperitoneal injection of D-luciferin (150 mg/ml) and imaged 5.5 min later. Each mouse in prone and supine positions according to previously described in bioluminescent imaging section of Materials and methods.

Statistical analysis of the data was performed using Prism 5 (Graphpad Software, La Jolla, CA). In vitro data represent mean of triplicates and correlation between bioluminescent signals and number of viable leukemic cells was determined via Spearman correlation analysis. Kruskal-Wallis analysis was performed with Dunn Multiple Comparison test to compare experimental groups to controls.

Bioluminescent leukemia cell lines were established using a firefly luciferase (fLuc) gene-encoding vector delivered via lentiviral infection into human ALL cell lines Reh, Jurkat and CCRF-CEM. Stable expression of fLuc in these cell lines was confirmed by analyzing the bioluminescent intensity of 10⁶ cells of each, quantified in average radiance (photon/sec/cm²/sr) (Fig. 1A). The bioluminescent intensity was measured in serial dilutions (10⁷–10³) of CCRF-CEM/fLuc cells (Fig. 1B), and the signal intensity was converted into average radiance (photon/sec/cm²/sr) (Fig. 1C). Bioluminescent intensity was directly proportional to the number of viable cells (p<0.001). To assess purity of the established human cell lines, CCRF-CEM and CCRF-CEM/fLuc cells were analyzed by flow cytometry for expression of common human leukocyte antigen marker, CD45 using PE-conjugated hCD45 and FITC-conjugated mCD45 antibodies (Fig. 1D). Double staining analysis ensures that only human-origin cell lines were used for leukemia model development. Both cell lines were positive for hCD45 and negative for mCD45. These results established the correlation between viable, luciferase-expressing CCRF-CEM/fLuc cells and bioluminescent signal intensity.

NOD/SCID mice were inoculated with 10⁶ CCRF-CEM/fLuc cells via IP, IV or IBMT routes (Fig. 2A) and examined using the bioluminescent imaging system every week for 6 weeks. For mice injected with IP, bioluminescent signals were localized and detected only at the site of injection (Fig. 2B); however, in mice inoculated with IV, the average radiance from luciferase activity of viable CCRF-CEM/fLuc was within the detectable range but was insufficient for visualization during the first three weeks after transplantation. Bioluminescent signals from the lower border of the sternum on the ventral side, head and upper sternum on the dorsal side became visible at week 4 (Fig. 2C). Unlike leukemia models prepared via IP and IV injections, bioluminescent signals of mice inoculated via IBMT were detectable in the tibia as early as one week after the inoculation of CCRF-CEM/fLuc cells. At week 4, strong bioluminescent signals were detected in mice in both supine and prone positions. The highest intensity signals were mainly detected in the head and backbone in the prone position, and in the upper sternum and hind legs in the supine position (Fig. 2D). Fig. 2E shows the mean radiance values for each inoculation group (n=15) over 6 weeks of monitored time. At week 6, IBMT model exhibited significantly stronger bioluminescent signals (p<0.001) than IP or IV mediated leukemia models. These results indicate that the visualization and quantification of leukemia progression can vary with the leukemia cell transplantation method.

To compare leukemia burden level in IP, IV and IBMT leukemia models with accuracy, calibrated signal intensity ranges representing the optimal range of signals above noise but below saturation level detected by CCD camera were assigned to each type of leukemia mouse model. In particular, images of week 4, 5 and 6 were compiled and compared with calibrated signal intensity ranges within each type leukemia model (Fig. 3A). The gradual increase in bioluminescent signal intensity from week 4 through 6 images was observed. In addition, images from one representative mouse at week 5 in each of the IP, IV and IBMT groups were pooled and compared within the calibrated signal intensity range (Fig. 3B). The IBMT leukemia model exhibited the strongest signal intensity; this represents the highest degree of leukemia burden among the three leukemia models. The corresponding bioluminescent intensity of isolated organs (brain, heart, lung, liver and spleen) from these leukemia model mice at week 5 was assessed (Fig. 3C). Organ images revealed that only those from the IBMT leukemia model had strong bioluminescent signals. It was investigated by flow cytometry whether the bioluminescent areas observed corresponded to leukemic cells in the organs from the IBMT leukemia model. Disease burden, i.e., human leukemic cells, in PB, BM and spleen cells from IP, IV and IBMT leukemia model mice sacrificed at week 5 post-injection was quantified by flow cytometric analysis of hCD45 and mCD45 expression (Fig. 3D). BM aspirates from the tibia (LBM) of IBMT mice contained mainly hCD45⁺ cells (96.89%), PB contained 79.06% hCD45⁺ cells and the spleen contained 15.08% hCD45⁺ cells. However, in the IP model, less than 1% of cells in PB, BM and spleen were hCD45⁺, and in the IV model, less than 2.26% of cells in PB, BM and spleen were hCD45⁺. Overall, results from comparative analyses of IP, IV and IBMT leukemia models indicate that detectable bioluminescent leukemia development in NOD/SCID can be effectively achieved using an IBMT route of administration.

To demonstrate that the IBMT bioluminescent leukemia model can be used to show the correlation between in vivo responses to anti-leukemic treatments and its corresponding bioluminescent intensity changes, IBMT leukemia mice were monitored by whole body bioimaging after treatment with anti-cancer drugs, vincristine and methylprednisolone. Response to the drugs was first tested in vitro. Viability of the leukemia cell line decreased as the concentration of vincristine increased (Fig. 4A). Bioluminescent evaluation of the viability assay confirmed a decrease in bioluminescent viable leukemic cells with an increase of vincristine concentration (Fig. 4B and C). Next, bioluminescent imaging of the IBMT leukemia model was evaluated during treatment with vincristine at two concentrations; 0.1, 0.5 mg/kg and PBS (Fig. 4D). Where bioluminescence at week 3 was initially detected pre-treatment, the intensity reduced significantly (p<0.01) after three weekly doses of vincristine at 0.5 mg/kg (Fig. 4E). The control group mouse (no vincristine) became moribund shortly after the image was taken at week 6. The reduction of whole body bioluminescent signals observed with the higher dose of vincristine was also observed in isolated organ images. The higher concentration of vincristine resulted in a reduction in leukemic cell infiltration of organs (Fig. 4F). The reduction in bioluminescent signals from vincristine-treated mice was validated by flow cytometric evaluation of isolated PB and BM cells (Fig. 4G). The percentage of hCD45⁺ cells in PB of 0.1 and 0.5 mg/kg vincristine-treated mice was lower than in PB from a control mouse. BM aspirates taken at the site of injection (left tibia) of the control IBMT mice and 0.1 mg/kg vincristine-treated IBMT mice were greater than 90% hCD45⁺. In contrast, analysis of BM aspirates from the injection site of a 0.5 mg/kg vincristine-treated mouse revealed a significantly lower percentage of hCD45⁺ cells (25.66%). In BM aspirates from the right tibia, compared to the control (92.23%), the percentage of hCD45⁺ cells in 0.1 mg/kg (10.79%) and 0.5 mg/kg (0.58%) vincristine-treated mice was significantly lower. In addition, a viability assay of CCRF-CEM/fLuc with methylprednisolone was conducted. The endpoint analyses from Alamarblue assay (Fig. 5A) and bioluminescent evaluation (Fig. 5B and C) showed that the viability of the leukemic cells decreased with an increase in methylprednisolone concentration. The mice were treated with PBS (control, n=9), 2 mg/kg (n=9) or 10 mg/kg (n=9) methylprednisolone once a week for 3 weeks. Each dotted line represents average bioluminescent intensity values for each treatment group (Fig. 5D). Methylprednisolone treatments (10 mg/kg) resulted in significantly (p<0.01) lower bioluminescence than 2 mg/kg or PBS treatment groups at week 6. Before (week 3) and after (week 6) images of representative mice treated with methylprednisolone show that the IBMT leukemia model can capture sensitive treatment responses (Fig. 5E). The results indicate that bioluminescent images of the IBMT in vivo leukemia model can provide an accurate representation of sensitive responses to candidate drugs.

To demonstrate that the IBMT route of administration is an effective method for generating a reproducible in vivo leukemia model, leukemia cells from PB of the first generation leukemia model mice at week 6 (Fig. 6A) were isolated. PB mononuclear cells from an IBMT leukemia model mouse that comprised 71.83% hCD45⁺ cells along with 25.63% mCD45⁺ cells (Fig. 6B). The luciferase expression level of the ex vivo leukemic cells was evaluated to ensure that these cells maintained characteristics of the human leukemia cell line, CCRF-CEM/fLuc, which was originally transplanted. Bioluminescent images of 10⁵, 10⁶ and 10⁷ isolated, ex vivo expanded cells were obtained and their average radiance values were plotted (Fig. 6C and D). These ex vivo cells proliferated and stably expressed luciferase during long-term in vitro culture (data not shown). After ex vivo expansion, the isolated leukemic cells were xenografted into the left tibia of secondary recipient NOD/SCID mice (Fig. 6E). The second generation model mice showed a similar course of bioluminescent intensity to that of the first (Fig. 6F). Bioluminescent activity of human leukemic cells in PB from the second generation was quantified by flow cytometric analysis 6 weeks after transplantation of the first generation leukemic cells. PB comprised 64.36% hCD45⁺ cells and 24.70% mCD45⁺ cells (Fig. 6G). These results indicate that the IBMT route is able to generate a reproducible in vivo leukemia model.