Human adipose tissue–derived mesenchymal stem cells (AT-MSC) are considered to be a promising source of autologous stem cells in personalized cell-based therapies. Tumor tracking properties of MSC provide an attractive opportunity for targeted transgene delivery into the sites of tumor formation. In the present study, we addressed whether the suicide gene introduction into human AT-MSC could produce a tumor-specific prodrug converting cellular vehicle for targeted chemotherapy. We prepared yeast fusion cytosine deaminase::uracil phosphoribosyltransferase gene–expressing cells [cytosine deaminase (CD)–expressing AT-MSC (CD-AT-MSC)] by retrovirus transduction. We explored their therapeutic potential on a model of human colon cancer in the presence of prodrug 5-fluorocytosine (5-FC). Gene manipulation of human AT-MSC did not sensitize CD-AT-MSC to 5-FC, thus overcoming the inherent disadvantage of suicide effect on cellular vehicle. CD-AT-MSC in combination with 5-FC augmented the bystander effect and selective cytotoxicity on target tumor cells HT-29 in direct coculture in vitro. We confirmed directed migration ability of AT-MSC and CD-AT-MSC toward tumor cells HT-29 in vitro. Moreover, we achieved significant inhibition of s.c. tumor xenograft growth by s.c. or i.v. administered CD-AT-MSC in immunocompromised mice treated with 5-FC. We confirmed the ability of CD-AT-MSC to deliver the CD transgene to the site of tumor formation and mediate strong antitumor effect in vivo. Taken together, these data characterize MSC derived from adipose tissue as suitable delivery vehicles for prodrug converting gene and show their utility for a personalized cell-based targeted cancer gene therapy. [Cancer Res 2007;67(13):6304–13]
- mesenchymal stem cells
- colon carcinoma cells HT-29
- mouse xenograft model
- cell-directed enzyme prodrug therapy
Colorectal cancer is one of the most common malignancies. Major cause of death associated with this type of cancer can be attributed to the development of metastasis in liver, abdominal lymph nodes, and lung ( 1). Prognosis remains poor for most patients although treated with the most effective chemotherapy of 5-fluorouracil (5-FU), which is a cornerstone of metastatic colorectal cancer treatment ( 2). This widely used anticancer agent has its limits: serious side effect and high dose levels required for the response. Therefore, it seems as though the metastatic colon cancer disease therapy requires development of novel strategies and could be a good target for molecular chemotherapy approaches ( 3).
High systemic toxicity of 5-FU could be circumvented by introducing gene-directed enzyme prodrug therapy relying on the ability of bacterial and/or yeast cytosine deaminase (CD) enzyme to convert far less toxic substrate 5-fluorocytosine (5-FC) to 5-FU. 5-FU is capable of nonfacilitated diffusion into and out of cells resulting in significant bystander effect of CD/5-FC ( 4). Antitumor effect of CD/5-FC combination on colon carcinoma was shown both in vitro and in vivo, and clinical trials reported safety of the CD/5-FC combination ( 5). We extended these observations to the breast cancer cells in our previous experiments and observed high sensitivity of retroviral-transduced breast cancer cells with bacterial CD to 5-FC as well as a significant bystander effect in vitro ( 6). However, promising experimental results were not translated into significant curative effect in patients mostly due to low expression of therapeutic transgene bringing only slight therapeutic benefit ( 7).
There were several attempts to increase the efficiency of CD/5-FC therapy by achieving higher conversion of prodrug 5-FC to 5-FU and the toxic metabolites production. Yeast CD was shown to produce 15-fold higher amount of 5-FU compared with bacterial CD ( 8). Moreover, construction of bifunctional fusion gene CD::uracil phosphoribosyltransferase (CD::UPRT) was reported to shortcut rate-limiting enzymatic steps of the 5-FC/5-FU conversion, thus resulting in 10,000-fold sensitization of transgene-expressing cells to 5-FC compared with unmanipulated counterparts ( 9). As expected, fusion yeast CD::UPRT gene-mediated prodrug conversion therapy led to effective direct and bystander tumor cell killing translated to significant antitumor effect in experimental animals in vivo ( 10, 11).
Nevertheless, successful gene therapy technology relies on the delivery of the therapeutic product into appropriate target cells. Introduction of transgene of interest into autologous stem cell types poses an attractive cell-based delivery strategy. Unique biology of mesenchymal stem cells (MSC) predetermines them to become valuable cytoreagents for gene therapy approaches in future ( 12). MSC can be relatively easily transduced with adenoviral, oncoretroviral, and lentiviral vectors, which is a key prerequisite for the introduction and durable expression of marker and/or therapeutic genes in MSC ( 13, 14).
Furthermore, MSC on systemic administration were reported to possess intrinsic preferential migratory ability toward some tumor types ( 15– 19). Human bone marrow–derived MSC after systemic i.v. administration could migrate and engraft into microscopic tumor lesions during early stage of development as shown for the s.c. tumors originated from human colon cancer cells HT-29 ( 20). Possible mechanism of preferential MSC homing and engraftment into the tumor mass can be attributed to the necessity of tumor stroma formation including tissue remodeling with high proliferation of MSC during the process of tumor growth ( 21). Exogenously administered MSC could form a significant proportion of tumor mass ( 19, 20); thus, their gene modification could be exploited for both tracking the sites of metastasis formation and, more importantly, for the tumor site-specific delivery of gene products to eliminate tumor.
Strong antitumor effect mediated by MSC producing therapeutic cytokines, such as IFN-β ( 16, 18, 19) and interleukin-2 ( 17, 22), was reported. MSC were tested as delivery vehicles for oncolytic adenoviruses in an ovarian and breast metastatic model, and authors report decrease in tumor burden ( 23, 24). Taken together, these studies have shown the endogenous tumor tracing capability of BM-MSC and their therapeutic potential for specific targeting of invasive and malignant tumors.
Adipose tissue, like bone marrow, is a mesodermally derived organ that was shown to contain stem cells ( 25). This population termed adipose tissue–derived MSC (AT-MSC) shares many of the characteristics of their bone marrow counterpart, including morphology, extensive proliferation potential, and the ability to undergo multilineage differentiation ( 26– 28). The success rate of AT-MSC isolation was 100% with the highest colony frequency observed ( 26). The yield of MSC from adipose tissue was ∼40-fold higher compared with the bone marrow ( 26). Last but not least to consider with the regard to the source of autologous stem cells for personalized cell-based therapy is the minimal risk to the donor and no ethical concerns ( 29).
In the present study, we address the capability of human AT-MSC to serve as vehicles for a cell-based gene-directed enzyme prodrug conversion approach to targeted chemotherapy. Here, we present a generally applicable strategy to specifically track the sites of tumor formation even at the stage of micrometastasis and target them by site-specific production of chemotherapeutic agent. In this pilot study, we show that MSC derived from adipose tissue expressing fusion yeast CD::UPRT gene [CD-expressing AT-MSC (CD-AT-MSC)] in combination with prodrug 5-FC augment potent cytotoxic effect over tumor cells HT-29 in vitro. Moreover, CD-AT-MSC were capable of significant inhibition of tumor growth in a therapeutic paradigm in vivo.
Materials and Methods
Cell lines and chemicals. Human colon adenocarcinoma cell line HT-29 was purchased from European Collection of Animal Cell Cultures (ECACC no. 91072201). HT-29 and human fibroblasts (kindly provided by Dr. J. Bizik, Cancer Research Institute Slovak Academy of Sciences, Bratislava, Slovakia) were cultured in RPMI 1640 and retroviral packaging mouse cell lines GP+E-86 and GP+envAM12 (kindly provided by Dr. J. Bies, National Cancer Institute, NIH, Bethesda, MD) were cultured in DMEM supplemented with 5% FCS and Antibiotic-Antimycotic mix (Life Technologies) in humidified atmosphere and 5% CO2 at 37°C.
All chemicals were purchased from Sigma if not stated otherwise.
Construction of retroviral vector containing CD. Construction of recombinant retrovirus was done by standard PCR cloning. Synthetic fusion yeast CD::UPRT gene was digested from expression vector pORF5-Fcy::Fur (InvivoGen) and inserted into retrovirus vector backbone from plasmid pJZ308 ( 30). Retrovirus vector obtained is a bicistronic construct with fusion yeast CD::UPRT gene separated by internal ribosome entry site sequence from neo gene. Resulting plasmid pST2 was sequenced to verify the correct reading frame and DNA sequence.
Preparation of virus-producing cells. GP+E-86 cells were transfected with 1 μg pST2 plasmid using Effectene (Qiagen). Virus-containing medium from G418-resistant GP+E-86/pST2 cells supplemented with 100 μg/mL protamine sulfate was used to transduce GP+envAM12 cells. Medium from G418-resistant cells GP+envAM12/pST2 was used to transduce back GP+E-86/pST2. Three to five rounds of repeated transduction were done to obtain cells producing recombinant retrovirus particles with mixed envelope glycoproteins as described ( 6). Activity of reverse transcriptase in a virus-containing medium was measured by standard reverse transcriptase assay ( 6).
Virus-containing medium for transduction was collected from semiconfluent cultures of GP+envAM12/pST2 cells incubated in fresh culture medium for 24 h, filtered through 0.45-μm filters, and used either fresh or kept frozen at −80°C until use.
MSC isolation from human adipose tissue, culture, and retrovirus transduction. AT-MSC cells were isolated from lipoaspirate using a collagenase type VII digestion and plastic adherence technique as described ( 31). Material was obtained from healthy persons undergoing elective lipoaspiration, who provided an informed consent. Cells were plated in low-glucose (1,000 mg/L) DMEM supplemented with 10% MSC-stimulatory supplement (human; StemCell Technologies) and antibiotic-antimycotic at a density of 2 × 105 to 5 × 105 nucleated cells/cm2. Adherent cells were split on reaching confluence and AT-MSC were used for the experiments up to passage 5.
To prepare MSC expressing CD (CD-AT-MSC), subconfluent cultures of AT-MSC were transduced thrice in 3 consecutive days with virus-containing medium from GP+envAM12/pST2 cells supplemented with 100 μg/mL protamine sulfate. The same method was used to prepare human fibroblasts expressing CD (CD-Fib).
Expression of CD, glyceraldehyde-3-phosphate dehydrogenase, and octomer-binding transcription factor 4 (Oct-4) in MSC. Total RNA was isolated from 2 × 106 cells by RNeasy mini kit (Qiagen) and treated with RNase-free DNase (Qiagen). RNA was reverse transcribed with RevertAid H Minus First-Strand cDNA Synthesis kit (Fermentas). cDNA (200 ng) was subject to standard PCR with 35 cycles and gel was resolved on 2% agarose.
Real-time PCR was done in 1× PCR Master Mix (Fermentas), 2.5 mmol/L MgCl2, 0.16 μmol/L primers, 1× SYBR Green (Molecular Probes), and 200 ng template cDNA on RotorGene 2000 (Corbett Research) and analyzed by RotorGene software version 4.6. Gene expression was compared using Δ cycle threshold (ΔCt = CtOct-4 − CtGAPDH) values for cells examined, where glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression was taken as endogenous reference gene. Change in gene expression was evaluated by Δ-Δ (ΔΔ)Ct method according to the following formula: ΔΔCt = [CtOct-4(CD-AT-MSC) − CtGAPDH(CD-AT-MSC)] − [CtOct-4(AT-MSC) − CtGAPDH(AT-MSC)]. Analysis was done twice in triplicates and data were expressed as mean ± SE.
The following primers were used: 5′-ATGGACATTGCCTATGAGGA-3′ (FcyFor) and 5′-TTCTCCAGGGTGCTGATCTC-3′ (FcyRev) for CD (167 bp); 5′-GAAGGTGAAGGTCGGAGTC-3′ (GAPDHFor) and 5′-GAAGATGGTGATGGGATTTC-3′ (GAPDHRev) for GAPDH (226 bp); and 5′-ACATCAAAGCTCTGCAGAAAGAACT-3′ (Oct-4For) and 5′-CTGAATACCTTCCCA AATAGAACCC-3′ (Oct-4Rev) for transcription factor (Oct-4; 133 bp).
Spectrophotometric measurement of 5-FC conversion. CD enzyme activity was measured spectrophotometrically ( 6). Cells (5 × 105) were resuspended in 50 μL PBS and lysed by five cycles of rapid freeze-and-thaw in liquid nitrogen. Lysates were centrifuged at 15,000 rpm for 30 min at 4°C. Clear supernatant (30 μL) was mixed with 30 μL of 0.4 mg/mL 5-FC in PBS and incubated at 37°C. Aliquots (10 μL) were taken at each time point, quenched with 90 μL of 0.1 mol/L HCl, spectrophotometrically evaluated on NanoDrop-1000, and analyzed by NanoDrop software ND-1000 3.3.0. Enzymatic conversion of 5-FC was calculated as percentage of 5-FC remaining in cell lysate based on following equation: 5-FC remaining (%) = [0.119 × A290(0) − 0.025 × A255(0)] − [0.119 × A290(x) − 0.025 × A255(x)] / [0.119 × A290(0) − 0.025 × A255(0)], where A290(0) and A255(0) are starting absorbance values (time point 0) at 290 and 255 nm, respectively, and A290(x) and A255(x) are absorbance values at x hours at 290 and 255 nm, respectively. Conversion of 5-FC is expressed as mean of triplicates ± SE and indicates percentage of starting 5-FC concentration that was set to 100% by default.
Immunophenotyping. MSC and HT-29 were labeled with the following antihuman antibodies conjugated to FITC: CD14-FITC, CD34-FITC, CD44-FITC, CD45-FITC, and CD105-FITC (purchased from Santa Cruz Biotechnology) and CD29-FITC and CD90-FITC (purchased from Chemicon). Mouse isotype antibodies IgG1 and IgG2a (Santa Cruz Biotechnology) served as respective controls. Labeled cells were analyzed using an EPICS ALTRA flow cytometer (Beckman Coulter) equipped with Expo 32 program.
MSC differentiation. Cultured AT-MSC and CD-AT-MSC were tested for the ability to differentiate into adipogenic cell lineage as described by Colter et al. ( 32) with modifications. Briefly, 5 × 103 cells per well in 96-well plates were grown in standard culture medium for 48 to 72 h. Cells were switched to differentiation medium composed of α-MEM medium supplemented with 15% FCS, antibiotics, 0.5 μmol/L hydrocortisone, 0.5 mmol/L isobutylmethylxanthine, and 60 μmol/L indomethacin for 28 days. Cells were washed with PBS, fixed in 4% formalin for 1 h, and stained for 15 min with fresh Oil Red-O solution (Fisher Scientific). Adipogenic differentiation was confirmed based on the detection of cells containing well-stained oil droplets.
For osteogenic differentiation, 103 cells per well in 96-well plates were grown in standard culture medium for 48 to 72 h followed by incubation in osteogenic medium using Osteogenic stem cells kit (StemCell Technologies). Medium α-MEM was supplemented with 15% osteogenic-stimulatory supplements (human), 10−8 mol/L dexamethasone, 0.2 mmol/L ascorbic acid, and 10 mmol/L β-glycerolphosphate. Medium was replaced every 3 to 4 days for 28 days. Cultures were washed with PBS, fixed in 4% formalin for 1 h, and stained for 10 min with 1 mL of 40 mmol/L Alizarin red (pH 4.3). Osteogenic differentiation was confirmed by detection of red-stained calcium deposits.
Cell invasion assays. The tropism of MSC for tumor cells was determined using a Cell invasion assay kit (Chemicon). HT-29 cells or fibroblasts (1.5 × 104) were loaded in the lower well of the 24-well plates 24 h before the experiment start. MSC (105 cells) in serum-free medium were placed onto the 8-μm pore size inserts coated with ECMatrix. Lower wells containing cells were washed with PBS, filled with serum-free medium, assembled for the migration assay, and incubated for 48 h. Noninvading cells and ECMatrix were wiped away from the inside of the insert, cells were stained for 20 min, and invading cells were photographed through the microscope. Results were evaluated by directly counting the number of migrated cells in five fields and by calculating the mean. Experiments were repeated twice with similar results.
Cell proliferation assays. Triplicates of MSC and HT-29 (3,000 cells per well) for each treatment were plated into 96-well plates and incubated overnight at 37°C. Culture medium was replaced for medium containing 0 to 500 μg/mL 5-FC or 0 to 1,000 μg/mL 5-FU 24 h later. Five days later, plates were subjected to the CellTiter 96 Aqueous One Solution Cell Proliferation Assay (Promega) and read out spectrophotometrically at 490 nm. Results were expressed as the percentage of proliferation, where the proliferation of cells in culture medium without 5-FC or 5-FU was set to 100%. Similar results were obtained in three independent experiments.
Direct coculture of HT-29 cells and MSC in vitro. Quadruplicates of HT-29 cells (5 × 103) were plated in 96-well plates. Increasing amounts of either CD-AT-MSC or AT-MSC (50, 100, 250, 375, 500, 1,000, and 2,500) were added to the tumor cells on day 1. 5-FC (100 μg/mL) was added to the mixed cultures on day 2. Cells were incubated with or without 5-FC for 5 days and cell viability was measured by standard proliferation assay. HT-29 cell viability without MSC in medium with or without 5-FC was set to 100%. Experimental values are expressed as mean percentage of control viability ± SE.
In prolonged direct coculture experiments, HT-29 tumor cells (105/6-cm dish) and CD-AT-MSC (2.5 × 104/6-cm dish) were cultured either alone or mixed in the medium supplemented with 100 μg/mL 5-FC for 20 days. Medium was changed every 3 days. Cells were then fixed with 70% ethanol, stained with H&E, and photographed through microscope. Experiments were repeated thrice with similar results.
Animal experiments. Six- to eight-week-old female athymic nude mice (BALB/c-nu/nu) were used for the in vivo experiments in accordance with institutional guidelines under the approved protocols. Cells were administered as suspension of 2.5 × 105 HT-29 cells in 100 μL PBS s.c. into the flank in the control animals. In coinjection studies, the following cell suspensions were injected s.c.: 2.5 × 105 HT-29 + 2.5 × 105 CD-AT-MSC, 2.5 × 105 HT-29 + 2.5 × 104 CD-AT-MSC, 2.5 × 105 HT-29 + 2.5 × 105 AT-MSC, and 2.5 × 105 HT-29 + 2.5 × 104 AT-MSC (in 100 μL PBS/each animal).
Suspension of 106 AT-MSC, 106 CD-AT-MSC, or 106 CD-Fib in 200 μL PBS or 200 μL PBS/each animal was injected i.v. into the lateral tail vein for the systemic administration.
Animals were treated with 500 mg/kg/d 5-FC diluted in PBS i.p. in two rounds of 5 consecutive days with 2-day break where indicated.
Tumors were measured by caliper, and tumor volume was calculated according to formula volume = length × width2 / 2 ( 33). None of the animals had to be sacrificed during experiment due to tumor ulceration, bleeding, or moribund state with excessive weight loss exceeding 25% of initial weight. At the experiment end point, animals were sacrificed, and tumors were excised and weighted for tumor burden. Results were evaluated as mean of tumor volume or weight ± SE. To detect CD transgene in tissues, some animals were sacrificed on days 9 and 20 to excise tumors and tissues for DNA isolation ( 34).
Analysis of tumor targeting. To detect MSC by fluorescence, AT-MSC were labeled with fluorescent dye CFDA-SE using Vybrant CFDA-SE Cell Tracer kit (Molecular Probes). AT-MSC suspension (106 cells/mL) in serum-free medium was supplemented with CFDA-SE (10 mmol/L), incubated for 15 min, and washed away. Cells were used for i.v. injections (106 per animal) of mice with preestablished tumors induced 3 days in advance by inoculation of HT-29 cells (2.5 × 105) s.c. Six days later, tumors and tissues were snap frozen, and fluorescent AT-MSC were detected on fresh cryosections (6–8 μm) of tumor samples under the fluorescent microscope.
To detect CD transgene delivery, 500 ng DNA isolated from tumors and tissues was subjected to PCR amplification for 35 cycles. Primers FcyUP (5′-ACCATGGTCACAGGAGGCAT-3′) and FcyANTI (5′-CTTGTTCCTGATGATGGTGTAG-3′) were used for the primary PCR to give 558-bp product, and primers FcyUP and FcyREV were used to amplify primary product in heminested PCR to detect 213-bp product. Template integrity was checked either with primers specific for mouse and human β-actin [5′-CCTTCTACAATGAGC-3′ (Act1) and 5′-ACGTCACACTTCATG-3′ (Act2); 594 bp] or with primers specific for human repetitive Alu sequences [5′-GTCAGGAGATCGAGACCATCCC-3′ (AluS) and 5′-TCCTGCCTCAGCCTCCCAAG-3′(AluA)] to obtain 124-bp product.
For CD quantification, 200 ng of isolated DNA were subjected to quantitative real-time PCR with primers FcyFor and FcyRev specific for CD gene under same conditions as described for cDNA. Standard curve for CD-positive cells was prepared from DNA samples with known proportion of CD-AT-MSC cells (10%, 1%, 0.5%, 0.1%, and 0.05%) mixed with HT-29 cells. The proportion of CD-positive cells in tumors was extrapolated from this linear standard curve and the Ct values for each sample. Analysis was done thrice in triplicates and data were expressed as mean ± SE.
Statistics. Values were calculated as mean ± SE or expressed as mean percentage of control ± SE. Significant differences were determined using Mann-Whitney U test. P < 0.05 was considered significant.
Transgene expression and enzymatic activity in CD-AT-MSC. Retrovirus-transduced human MSC derived from adipose tissue expressing CD::UPRT transgene (CD-AT-MSC) were prepared. CD was expressed in CD-AT-MSC and not in cultured untransduced AT-MSC by reverse transcription-PCR (RT-PCR; ref. Fig. 1A ). Enzymatic conversion of prodrug 5-FC was confirmed in cell lysates prepared from CD-AT-MSC ( Fig. 1C). There was not any 5-FC converting activity in cell lysates from either unmanipulated AT-MSC or HT-29 tumor cells. We estimated that ∼60% of 5-FC was converted within 24 h, thus showing effective enzymatic activity of transgene.
Stem cell properties of CD-AT-MSC are not affected by retrovirus transduction and transgene expression. To confirm that culture expansion, manipulation, and retroviral transduction did not change stem cell properties of CD-AT-MSC, experiments were done to compare CD-AT-MSC with cultured AT-MSC. First, we found no change in the expression of pluripotent embryonic transcription factor Oct-4 on transduction (ΔΔCt = 0.35 ± 0.214). Both CD-AT-MSC and AT-MSC express Oct-4 indicative of undifferentated state and high proliferative capacity preservation ( Fig. 1A and B). Moreover, we analyzed AT-MSC and CD-AT-MSC immunophenotype by standard flow cytometry to search for the potential surface marker expression change. Both cell types were found to be consistently positive for CD29, CD44, CD90, and CD105 ( Fig. 2A ) and negative for CD14, CD34, and CD45. Both CD-AT-MSC and cultured AT-MSC were capable of differentiation into adipogenic ( Fig. 2B) and osteogenic lineages ( Fig. 2C) with no significant difference. Taken together, these results suggest that transgene introduction does not affect any of stem cell characteristics examined in CD-AT-MSC.
Migration toward tumor cells HT-29 in vitro. We hypothesized that the migration capability toward the tumor cells was a prerequisite for the AT-MSC to specifically target tumor cells. Directional migration of AT-MSC through ECMatrix was significantly stimulated by human colon adenocarcinoma cells HT-29 cells in vitro [72.3 ± 12.2 migrated cells compared with control (25 ± 9.3)]. Migration capacity was not affected by transgene expression in CD-AT-MSC (67 ± 13 cells; Fig. 2D). The same results were obtained, when the assay was done with tumor cells HT-29 conditioned serum-free medium used as chemoattractant.
Effect of 5-FC and 5-FU on cell proliferation. Chemosensitivity of AT-MSC, CD-AT-MSC, and HT-29 cells to 5-FC and 5-FU was analyzed by cell proliferation assay. 5-FC is relatively nontoxic, and even at high concentrations up to 500 μg/mL, it did not affect HT-29 or MSC cell proliferation ( Fig. 3A ). CD expression sensitized CD-AT-MSC to 5-FC at highest concentration only, suggesting for the existence of endogenous mechanism in AT-MSC to eliminate a suicide effect of transgene expression. Intrinsic chemoresistance of AT-MSC to 5-FC metabolites was confirmed by direct cultivation in 5-FU–containing medium. Tumor cells HT-29 were very sensitive to 5-FU with IC50(HT-29) of 7 μg/mL 5-FU. AT-MSC were 128.6-fold more resistant with the IC50(AT-MSC) of 900 μg/mL 5-FU. Transgene expression sensitized CD-AT-MSC to 5-FU 2.5-fold with IC50(CD-AT-MSC) of 360 μg/mL 5-FU. However, even at high concentration of 100 μg/mL 5-FU, proliferation of CD-AT-MSC remained 56.6% and AT-MSC 58.6% in contrast to only 6.7% in HT-29 cells ( Fig. 3B). This finding indicated the capacity of MSC to overcome the inherent disadvantage of suicide effect on cellular vehicles.
Selective cytotoxicity toward tumor HT-29 cells mediated by CD-AT-MSC in the presence of 5-FC in vitro. Therapeutic potential of CD-AT-MSC was evaluated in direct coculture experiments in vitro. CD-AT-MSC were cocultured with tumor HT-29 cells in the presence of prodrug 5-FC. They exert strong selective cytotoxic effect over HT-29 cells, even if the ratio was as low as 1:100 (P = 0.05; Fig. 3C). No inhibitory effect was observed without 5-FC or CD in cocultures. Prolonged direct coculture (20 days) showed complete eradication of tumor cells ( Fig. 3D, left). However, CD-AT-MSC could withstand the effect of 5-FU and its toxic metabolites produced ( Fig. 3D, middle). Chemoresistance of the CD-AT-MSC could not be attributed to the loss of transgene because the presence of CD-specific transcripts in these cells was confirmed by RT-PCR (data not shown). Moreover, HT-29 cells could proliferate intensively under these culture conditions ( Fig. 3D, right).
CD-AT-MSC mediated tumor growth inhibition in vivo. To test the proposed approach in vivo, mixtures of tumor cells HT-29 with 10% and 50% MSC were injected in nude mice s.c. treated with 5-FC (500 mg/kg/d) i.p. ( Fig. 4A ). Significant inhibition of tumor growth was observed in animals treated with 10% or 50% CD-AT-MSC ( Fig. 4B). The outcome shows that only 10% CD-positive cells within the tumor mass could inhibit tumor growth by ∼79% by the experiment end point time showing strong bystander effect in vivo. All animals injected with HT-29 alone or mixed with AT-MSC developed tumors ( Fig. 4C, left flank). In contrast, CD-AT-MSC–treated animals did not always develop tumor and if present, it was always smaller at the experiment end point ( Fig. 4C, right flank). Comparison of tumor burden showed 73.8% and 78.4% tumor inhibition in the therapeutic regimen CD-AT-MSC+HT-29/5-FC 1:1 and 1:10, respectively. ( Fig. 4D).
Tumor targeting and growth inhibition on systemic administration of CD-AT-MSC in vivo. To test the tumor-directed migratory ability of AT-MSC, fluorescently labeled AT-MSC were injected into mice with preestablished HT-29–induced tumors i.v. Six days after treatment, fluorescent MSC were detected located within the tumor parenchyma ( Fig. 5A ). Thus, AT-MSC were capable of localizing into tumors after systemic injection. Moreover, to analyze other possibilities of AT-MSC homing in vivo, animals with preestablished tumors were injected with CD-AT-MSC i.v. and sacrificed at two different time points. Based on PCR detection of CD transgene, CD-AT-MSC were found in lungs and liver of some animals although at low amounts on day 9 ( Fig. 5B). CD-AT-MSC were cleared away at 20 days although consistently found in s.c. tumors.
Next, we tested whether the capability of exogenously administered CD-AT-MSC to reach the sites of s.c. tumor formation could influence the tumor growth in a therapeutic setup in vivo. Mice with preestablished tumors were injected with CD-AT-MSC, AT-MSC, CD-Fib, and saline i.v. and 5-FC treated ( Fig. 6A ). Significant inhibition of tumor growth up to 68.5% was observed in all animals injected with CD-AT-MSC by day 18 ( Fig. 6B). None of the animals exhibited any signs of toxic side effects of therapeutic regimen. No effect was achieved in the absence of CD or if CD-Fib were used as delivery agents. This was ascribed to the failure of human fibroblasts to home within the tumor mass, as confirmed by analysis of tumor samples for the presence of CD transgene ( Fig. 6C). Detailed quantitative analysis showed that 0.87% to 3.89% and 0.23% to 0.55% CD-AT-MSC had been incorporated within the tumors by days 9 and 20, respectively ( Fig. 6D), which translated into 91% and 68.5% tumor inhibition at given time points.
In this pilot study, we present the capability of human MSC derived from adipose tissue to serve as cellular vehicles for targeted cancer chemotherapy. CD-AT-MSC were prepared by retrovirus transduction. High transduction efficiency estimated to be close to 98% based on the G418 selection 1 is advantageous for the optimal therapeutic application. It eliminates lengthy selection requirement and/or undesired prolonged AT-MSC expansion ex vivo. Moreover, transgene-bearing vector was designed as a replication-deficient retrovirus for safety reasons. Retention of AT-MSC properties on gene manipulation ex vivo was investigated to some extent. Experiments published thus far ( 13, 19) and our data did not show any negative effect of retrovirus transduction and transgene expression on MSC differentiation capabilities, immunophenotype, migration properties, or pluripotency transcription factor expression.
Yeast fusion CD::UPRT gene chosen for our study was reported to exhibit improved 5-FC conversion efficiency and higher bystander effect in vitro and in vivo ( 9– 11). Because there is neither enzyme nor virus produced from therapeutic cells, our strategy relies on bystander cytotoxic effect of therapeutic CD-AT-MSC combined with 5-FC on target cells. Tumor targeted delivery and site-specific production of 5-FU mediated by CD-expressing cells can exert strong cytotoxic effect on rapidly proliferating tumor cells ( 35– 37). In contrast to other stem cell–based approaches, MSC derived from adipose tissue seem to be intrinsically highly resistant to 5-FU. The “suicide” gene CD::UPRT introduction moderately changed sensitivity CD-AT-MSC to 5-FC and 5-FU; however, they still seem to be much more resistant compared with untransduced target tumor cells HT-29. This may spare the cellular vehicle from the “suicide” effect and pronounce further the bystander killing effect by prolonged production of 5-FU and/or achieving higher local concentrations herewith increasing the therapeutic efficiency. AT-MSC sensitivity to 5-FU increased with number of passages, suggesting gradual loss of resistance during forced proliferation ex vivo. 1 We hypothesize that potential uncontrolled CD-AT-MSC proliferation would be eliminated by suicide effect of CD::UPRT transgene based on these observations. This presents additional safety feature of therapeutic CD-AT-MSC in vivo. The mechanism underlying chemoresistance AT-MSC properties was not the objective of this study, and it remains under further investigations. However, be it a general property of MSC, it would predetermine MSC to become valuable vehicles for the delivery of various “suicide” transgenes to target the cells of interest and spare the vehicle cells by their endogenous resistance mechanism.
Here, we show for the first time the capability of MSC derived from the adipose tissue to migrate actively toward tumor cells. This is consistent with the reports describing tumor-directed migratory abilities of MSC derived from bone marrow ( 16). Migratory properties have to be further evaluated for the tumor type specificity as well as possible signaling molecules to be involved due to the relevance for potential therapeutic applications. It has been reported that epidermal growth factor, platelet-derived growth factor, SCF/c-Kit, stromal cell–derived factor-1/CXCR4, and vascular endothelial growth factor (VEGF)/VEGF receptor (VEGFR) 1 and VEGFR2 may play a role in these intrinsic tumor-specific migration capability and interactions ( 16, 36).
MSC were reported to support s.c. tumor growth in a different model ( 18, 38). Coinjection of 50% and 10% AT-MSC+HT-29 mixtures led to initial tumor growth support ( Fig. 4B). Supportive effect was not observed on AT-MSC systemic administration ( Fig. 6B), which is the therapeutic approach more relevant for future clinical studies. The proportion of CD-AT-MSC in tumors achieved <4% CD-AT-MSC after single i.v. injection. We hypothesize that such proportion did not suffice to support tumor growth. In contrast, combination of targeted CD-AT-MSC homing into tumor with inherent CD-AT-MSC resistance to “suicide” effect pronounced neighboring killing effect still to achieve significant tumor growth inhibition in vivo.
Systemic administration of MSC was extensively evaluated in terms of organ distribution and possible unwanted adverse effect of MSC accumulation ( 19, 23). To this end, all studies published thus far and our experiments showed no accumulation in other organs and no other side effects were observed thus far.
None of the animals remained tumor-free or cured in our experimental setting. It is generally acknowledged that the successful long-term effective therapy or curative approach would require combination of more treatment strategies. In our approach, the 5-FU produced at the site of tumor formation is expected to eliminate rapidly proliferating tumor cell population based on its mechanism of action ( 4). However, quiescent and nondividing tumor cells present within the tumor mass might withstand the toxicity. Our future studies will be directed toward optimized protocols for dosage and timing of the therapy to achieve maximum tumor growth suppression. We also intend to design combination strategy of targeting nondividing quiescent cells within the tumor mass. It is obvious from the recent data that those cells may in fact be responsible for the recurrence of the disease and metastasis outgrowth and they certainly are the targets for the long-term effective cancer therapy ( 39, 40).
Taken together, here, we present human MSC derived from adipose tissue as cell-based delivery vehicles for the site-specific enzyme prodrug conversion approach to targeted chemotherapy. Engineered CD-AT-MSC expressing yeast fusion CD::UPRT gene combined with 5-FC were efficient in suppression of s.c. human colon cancer xenograft growth in our pilot study in vivo. MSC from adipose tissue are easily obtainable with no ethical concerns about the stem cell source, rapidly expandable to desired amount, and can be gene manipulated ex vivo. AT-MSC possess specific chemoresistance and migration properties and therefore should be considered valuable adult stem cell types for autologous use in cancer therapy. It is worth to test AT-MSC further as cellular vehicles for other prodrug converting gene systems in treatment of various tumor types. This would open new ways to AT-MSC exploitation for personalized cell-based therapeutic approaches in future.
Grant support: VEGA grant 2/5028/26, state program of Slovak Republic “Use of cancer genomics to improve the human population health,” League against Cancer grant (V. Altanerova), and Centre of Excellence of the SAS “Molecular Medicine.”
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank M. Dubrovcakova, M. Fischerova, and R. Bohovic for technical assistance; K. Hlubinova and V. Frivalska for animal maintenance; J. Jakubikova and J. Bodo for help with flow cytometry analysis; A. Pastorakova for valuable discussions; J. Fabry for providing us with material; and R. Kucera for help with image processing and statistical analysis.
↵1 M. Matuskova, et al. Human mesenchymal stromal cells expressing cytosine deaminase mediate strong cytotoxic effect on cancer cells. Stem Cells, submitted.
- Received October 31, 2006.
- Revision received April 11, 2007.
- Accepted May 1, 2007.
- ©2007 American Association for Cancer Research.