Breast Cancer Research and Treatment

, Volume 105, Issue 2, pp 157–167

Mesenchymal stem cells as a vehicle for targeted delivery of CRAds to lung metastases of breast carcinoma

Authors

  • Mariam A. Stoff-Khalili
    • Division of Human Gene Therapy, Departments of Medicine, Surgery, Pathology and the Gene Therapy CenterUniversity of Alabama at Birmingham
    • Department of Obstetrics and Gynecology, Medical CenterUniversity of Duesseldorf
  • Angel A. Rivera
    • Division of Human Gene Therapy, Departments of Medicine, Surgery, Pathology and the Gene Therapy CenterUniversity of Alabama at Birmingham
  • J. Michael Mathis
    • Gene Therapy Program, Department of Cellular Biology and AnatomyLouisiana State University Health Sciences Center
  • N. Sanjib Banerjee
    • Department of Biochemistry and Molecular GeneticsUniversity of Alabama at Birmingham
  • Amanda S. Moon
    • Animal Resources ProgramUniversity of Alabama at Birmingham
  • A. Hess
    • Department of Obstetrics and Gynecology, Medical CenterUniversity of Duesseldorf
  • Rodney P. Rocconi
    • Division of Human Gene Therapy, Departments of Medicine, Surgery, Pathology and the Gene Therapy CenterUniversity of Alabama at Birmingham
  • T. Michael Numnum
    • Division of Human Gene Therapy, Departments of Medicine, Surgery, Pathology and the Gene Therapy CenterUniversity of Alabama at Birmingham
  • M. Everts
    • Division of Human Gene Therapy, Departments of Medicine, Surgery, Pathology and the Gene Therapy CenterUniversity of Alabama at Birmingham
  • Louise T. Chow
    • Department of Biochemistry and Molecular GeneticsUniversity of Alabama at Birmingham
  • Joanne T. Douglas
    • Division of Human Gene Therapy, Departments of Medicine, Surgery, Pathology and the Gene Therapy CenterUniversity of Alabama at Birmingham
  • Gene P. Siegal
    • Division of Human Gene Therapy, Departments of Medicine, Surgery, Pathology and the Gene Therapy CenterUniversity of Alabama at Birmingham
  • Zeng B. Zhu
    • Division of Human Gene Therapy, Departments of Medicine, Surgery, Pathology and the Gene Therapy CenterUniversity of Alabama at Birmingham
  • Hans Georg Bender
    • Department of Obstetrics and Gynecology, Medical CenterUniversity of Duesseldorf
  • Peter Dall
    • Department of Obstetrics and Gynecology, Medical CenterUniversity of Duesseldorf
  • Alexander Stoff
    • Division of Human Gene Therapy, Departments of Medicine, Surgery, Pathology and the Gene Therapy CenterUniversity of Alabama at Birmingham
    • Department of Plastic and Reconstructive SurgeryDreifaltigkeits-Hospital
  • Larissa Pereboeva
    • Division of Human Gene Therapy, Departments of Medicine, Surgery, Pathology and the Gene Therapy CenterUniversity of Alabama at Birmingham
    • Division of Human Gene Therapy, Departments of Medicine, Surgery, Pathology and the Gene Therapy CenterUniversity of Alabama at Birmingham
Preclinical Study

DOI: 10.1007/s10549-006-9449-8

Cite this article as:
Stoff-Khalili, M.A., Rivera, A.A., Mathis, J.M. et al. Breast Cancer Res Treat (2007) 105: 157. doi:10.1007/s10549-006-9449-8

Abstract

Purpose

Alternative and complementary therapeutic strategies need to be developed for metastatic breast cancer. Virotherapy is a novel therapeutic approach for the treatment of cancer in which the replicating virus itself is the anticancer agent. However, the success of virotherapy has been limited due to inefficient virus delivery to the tumor site. The present study addresses the utility of human mesenchymal stem cells (hMSCs) as intermediate carriers for conditionally replicating adenoviruses (CRAds) to target metastatic breast cancer in vivo.

Experimental design

HMSC were transduced with CRAds. We used a SCID mouse xenograft model to examine the effects of systemically injected CRAd loaded hMSC or CRAd alone on the growth of MDA-MB-231 derived pulmonary metastases (experimental metastases model) in vivo and on overall survival.

Results

Intravenous injection of CRAd loaded hMSCs into mice with established MDA-MB-231 pulmonary metastatic disease homed to the tumor site and led to extended mouse survival compared to mice treated with CRAd alone.

Conclusion

Injected hMSCs transduced with CRAds suppressed the growth of pulmonary metastases, presumably through viral amplification in the hMSCs. Thus, hMSCs may be an effective platform for the targeted delivery of CRAds to distant cancer sites such as metastatic breast cancer.

Keywords

Breast cancerCell vehicleCRAdsMetastasesStem cellsVirotherapy

Introduction

In the United States, breast cancer remains the most common malignancy in women. In some women, breast cancer is a local disease without spread. Such early breast cancers are usually diagnosed by screening mammography and are highly curable with local or regional treatment alone. However, most women with primary cancer have subclinical metastases, and in a high percentage of those treated with apparently curative surgery, distant metastases ultimately develop. The clinical course of metastatic breast cancer is variable. Chemotherapy, hormonal therapy, radiotherapy, and limited surgery are all used in the treatment of women with metastatic breast cancer, although the overwhelming majority of these women will die of their disease. In view, of the limited success of available treatment modalities for metastatic breast cancer, alternative and complementary strategies need to be developed.

In this regard, virotherapy is an exciting therapeutic approach for the treatment of cancer in which the replicating virus itself is the anticancer agent. Among various viruses, the adenovirus-based vector has emerged as a leading candidate for in vivo cancer virotherapy. Conditionally replicative adenovirus based agents (CRAds) have been designed to replicate in tumor cells whereby the virus can self-amplify and spread in the tumor from an initial infection of only a few cells. However, highly effective use of CRAd agents in tumors clinically has been heretofore hindered by three main factors: (1) low viral infectivity, (2) suboptimal replicative specificity and (3) inefficient viral agent delivery to the tumor site [10]. With respect to breast cancer, transduction efficacy by adenovirus serotype 5 (Ad5) is often suboptimal due to the highly variable and often low expression pattern of the primary adenovirus receptor, coxsackie adenovirus receptor (CAR) [4, 9]. To circumvent this, genetic alterations of the virus fiber protein that utilizes CAR-independent entry pathways have been identified, thus bypassing CAR deficiency on cancer cells and enhancing tumor transduction. In parallel, strategies have been developed to enhance the transcription selectivity of current vector systems toward tumor cells by using tumor specific promoter (TSP) that limit ectopic expression in non-tumor cells and decrease treatment-associated toxicities. However, efficient virus delivery to the tumor site is a central mandate of virotherapy.

In this regard, cell carriers exhibiting endogenous tumor homing activity have been recently exploited to chaperone virus delivery to the tumor site. Although the utility of cells as vehicles for toxic genes, anti-angiogenic molecules, and immunostimulatory genes has been suggested in several studies, there have been only limited studies whereby cells have been exploited as carriers for virotherapeutic agents. In this regard, we have proposed human mesenchymal stromal cells (hMSCs) as carriers of oncolytic viruses in vitro [15]. MSCs are bone marrow-derived non-hematopoietic precursor cells that when systemically administered, home to the tumor, preferentially survive and proliferate in the presence of malignant cells and become incorporated into the tumor architecture as stromal fibroblasts [20]. In this regard, it has been recently shown that systemically administered hMSCs home to breast cancer metastasis of the lung [21].

Based upon these findings, we hypothesized that hMSCs could be used as a targeting strategy for CRAds in the treatment of breast cancer metastasis of the lung. Our results, demonstrating that systemic administration of hMSC carriers can target CRAds to metastatic disease, constitute a novel therapeutic paradigm for breast cancer that couples cell therapy with virotherapy.

Materials and methods

Adenoviral vectors

The CRAd Ad5/3.CXCR4 was constructed as follows: the plasmid pBSKCAT/CXCR4, which contains a 279 bp sequence from the human CXCR4 promoter (−191 to +88), was a kind gift of Dr. Nelson L. Michael [25]. The CXCR4 promoter sequence (NCBI Accession Number AY728138, from 1780 to 2059 bp) containing the 279 bp CXCR4 promoter and the simian virus 40 (SV40) polyadenylation (poly A) signal was cloned by PCR into the NotI/XhoI site of pScsE1 plasmid [12, 17] (a kind gift from Dr. Dirk Nettelbeck, Department of Dermatology, University Medical Center-Erlangen, Erlangen, Germany), resulting in pScsE1CXCR4 that contained the E1A gene downstream of the CXCR4 gene promoter. The Ad vector, pAdback 5/3 was a kind gift from Dirk Nettenbeck and contains both the E3 gene and a capsid modified F5/3 [12, 17]. After cleavage with PmeI, the shuttle vector, pScsE1CXCR4, was recombined with pAdback5/3 to generate the CRAd Ad5/3.CXCR4, where the human Ad knob serotype 5 is replaced by human Ad knob serotype 3. The Ad vector, pVK503c, was a kind gift from Dr. V. Krasnykh (M.D. Anderson, Houston, TX), and contains both the E3 gene and a capsid modified RGD4C [23]. After cleavage with PmeI, the shuttle vector, pScsE1CXCR4, was recombined with Cla I linearized pVK503c to generate the CRAd Ad5RGD.CXCR4 with a fiber protein incorporating an RGD-motif in the HI-loop of Ad5 fiber. The recombinant plasmids were linearized with PacI and transfected into 293 cells using superfect reagent (Qiagen; Valencia, CA) to generate the Ad5/3.CXCR4 and Ad5R6D.CXCR4 adenoviruses. The other replication competent adenoviruses used in this study were: wild type Ad5wt with unmodified fiber [22], Ad5/3wt with a chimeric fiber having the knob of Ad5 fiber replaced by knob of Ad3 fiber [12, 17], Ad5.RGDwt with a fiber protein incorporating an RGD-motif in the HI-loop of Ad5 fiber. Ad.CXCR4Luc [26] was used for replication negative control. The adenoviruses were propagated in the A549 cells (a lung cancer cell line in which the CXCR4 gene is overexpressed), and purified by double CsCl density gradient centrifugation, followed by dialysis against phosphate buffered saline (PBS) containing 10% glycerol. The concentration of total viral particle numbers (PN) was determined by measuring absorption at 260 nm. Infectious PNs were determined by measuring the concentration of viral hexon protein-positive 293 cells after a 48-h infection period, using an Adeno-X Rapid Titer Kit (Clontech; Mountain View, CA).

Cell line and cell culture

MDA-MB-231 breast cancer cell lines were obtained from the American Type Culture Collection (ATCC) and cultured as described. In brief, the cells were maintained in DME/F-12 medium (Life Technologies, Inc., Grand Island, NY), containing 10% fetal bovine serum (FBS: Gemini Bio-Products, Woodland, Ca), and 1% antibiotic–antimycotic solution (penicillin–streptomycin–fungizone, Sigma Chemicals Co., St. Louis, MO). The cells were maintained in T-175 flasks at 37°C and 5% humidified CO2, and were sub-cultured using 1% trypsin–EDTA (Gibco BRL, Life Technologies).

Human mesenchymal stem cells and labeling

Human mesenchymal stem cells were obtained from the Tulane Center for Gene Therapy (Tulane University Health Sciences Center, New Orleans, LA, USA) and cultured according to the protocol provided. Carbocyanine dye (CellTracker CM-Dil; Molecular Probes Inc, Eugene, Ore) was used to label the human mesenchymal stem cells according to the manufacturer’s standard protocol as previously described [3].

In vitro cytotoxicity assay of CRAds

For determination of virus-mediated cytotoxicity, 1 × 104 MDA-MB-231 cells were seeded in 48-well plates and infected with adenoviruses at MOI of 1–1000 or were mock-infected [17]. To visualize cell killing, cells were fixed and stained with 1% crystal violet in 70% ethanol for 20 min followed by washing with tap water to remove excess dye at day 3, 5, 7 and 9. The plates were dried and images were captured with a Kodak DC260 digital camera (Eastman Kodak, Rochester, NY, USA).

In vitro cytotoxicity assay of human mesenchymal stem cells loaded with CRAd

About 3 × 105 hMSCs were plated in 6 well plates and infected with Ad5/3.CXCR4 at MOI of 1–1000 in 2% media. After 18 h the Ad5/3.CXCR4 infected hMSCs, now called hMSC-Ad5/3.CXCR4, were washed three times with PBS and trypsinized. Next, 5 × 104 MDA-MB-231 breast cancer cells plated in 48 well plates were co-cultured with 5 × 104 hMSCs carrying Ad5/3.CXCR4 at different MOIs. To visualize cell killing, cells were fixed and stained with 1% crystal violet in 70% ethanol for 20 min followed by washing with tap water to remove excess dye at day 3, 5, 7, and 9. The plates were dried and images were captured with a Kodak DC260 digital camera (Eastman Kodak, Rochester, NY, USA).

Quantitating virus replication

Purification of the DNA and quantitative real-time PCR for E4 was performed as previously described [6]. About 1.5 × 105 cells were seeded per well in a six-well plate. The next day cells were infected with the indicated viruses at different MOIs or mock infected and growth medium was collected at the indicated time points. Negative controls without templates were performed for each reaction series, and an internal control (human GAPDH) was used to normalize the copy number for the E4 genes. Comparison of replication rates of different treatment groups were performed with a Student’s t-test.

Mouse xenograft model for experimental metastatic breast cancer to the lung

Female C.B-17 SCID mice (6-weeks old) were obtained from Charles River Laboratories, Inc (Wilmington, MA, USA). Mice were used according to approved institutional protocols. The mouse xenograft model for metastatic breast cancer to the lung and the route of hMSC injection was performed as described [21]. Mice were injected intravenously in the lateral tail vein with 2 × 106 MB-MDA-231 suspended in 200 μl of PBS [21]. In preliminary experiments, we determined that all mice injected with 2 × 106 MB-MDA-231 developed macroscopic tumor nodules in their lungs at 14 days after tumor cell injection (data not shown). Fourteen days later, treatment was started. The following preparations were made prior intravenous injection: 106 hMSCs were infected with Ad5/3.CXCR4 at MOI of 1000 in 2% media. After 18 h the Ad5/3.CXCR4 infected hMSCs, now called hMSC-Ad5/3.CXCR4, were washed three times with PBS and trypsinized. Then hMSC-Ad5/3.CXCR4 and hMSCs alone were labeled with dye (CellTracker CM-Dil; Molecular Probes Inc., Eugene, Ore) at 37°C in complete media for 30 min followed by three washes with PBS and resuspended in 100 ul of PBS. Then, at day 14, mice were we injected intravenously in the lateral tail vein with either 106 hMSCs which had been infected with Ad5/3.CXCR4 of MOI 1000 (n = 8), hHMSC alone (n = 8) or Ad5/3.CXCR4 of MOI of 1000 alone (n = 8) suspended in 100 μL PBS. Non-treated MDA-MB-231 tumor bearing mice (n = 8) and healthy mice (n = 8) served as a control.

Determination of effect of CRAd loaded hMSCs on MB-MDA-231 tumor weight in mouse lung

Fourteen days after MB-MDA-231 tumor cell injection (as described above), the mice obtained treatment intravenously in the lateral tail vein with either 106 hMSCs which had been infected with Ad5/3.CXCR4 of MOI 1000 (n = 4), MSC alone (n = 4) or Ad5/3.CXCR4 of MOI of 1000 alone (n = 4) suspended in 100 μl PBS. Mice injected with MB-MDA-231 tumor cells alone (n = 4) and healthy mice with no tumor cell injection (n = 4) served as controls. Mice were sacrificed by asphyxiation with CO2 30 days after tumor cell injection. We measured the weight of whole lungs in all groups of mice and used whole lung weight as a surrogate endpoint of MD-MBA-231 tumor burden in the lung and to assess the effect of hMSCs, Ad5/3.CXCR4 on tumor growth [21]. All mice were followed daily until euthanasia. None of the mice had to be sacrificed because of excessive bleeding, open wound infection, moribund status, or prostration with weight loss of more than 25% of initial body weight.

Determination of effect of CRAd loaded hMSCs on survival in mice bearing metastatic breast cancer

Fourteen days after MD-MBA-231 tumor cell injection (as described above), the mice obtained treatment intravenously in the lateral tail vein either 106 hMSCs which had been infected with Ad5/3.CXCR4 of MOI 1000 (n = 10), hMSC alone (n = 10) or Ad5/3.CXCR4 of MOI of 1000 alone (n = 10) suspended in 100 μl PBS. Mice injected with MB-MDA-231 tumor cells alone (n = 10) and healthy mice with no tumor cell injection (n = 10) served as controls. All mice were followed daily until day 150. None of the mice had to be sacrificed because of excessive bleeding, open wound infection, moribund status, or cachexia.

MDA-MB-231 lung metastases

To assess the effect of transient transfection of hMSC-Ad5/3.CXCR4 on MDA-MB-231 lung metastasis, lungs were harvested 30 days later, fixed in 10% neutral-buffered Formalin, longitudinally trisected, paraffin embedded, and three 5–7 μm thickness sections were cut at 200 μm intervals from each embedded block. Tissue sections were stained with hematoxylin and eosin, examined for the presence of tumor nodules by a pathologist unaware of the treatments.

Tissue processing and imaging studies

Lungs from each group of mice were harvested 30 days after treatment, fixed in 10% neutral-buffered formalin, longitudinally trisected, paraffin embedded, and three 5–7 μm thickness sections were cut at 200 μm intervals from each embedded block. Tissue sections were stained with hematoxylin and eosin, examined for the presence of tumor nodules by a pathologist. In addition, lungs for each group of mice were embedded in Tissue TEK OTC compound (Miles, Elkhart, IN), snap-frozen in liquid nitrogen, and stored at −80°C.

Immunofluorescence

GFP labeled hMSCs were identified in the lung as following: Frozen tissue sections were air dried, fixed in 4% paraformaldehyde for 1 h, permeabilized in PBS with 0.2% Triton X-100, washed 5 min in PBS and blocked in 25% goat serum for 30 min. The sections were washed in PBS three times for 5 min, mounted with VectaShield mounting medium with 4′,6-diamidino-2-phenylindole (H-1200; Vector Laboratories) and then analyzed by fluorescence microcopy. Adenoviral hexon was identified in the lung as following: Frozen tissue sections were air dried, fixed in 4% paraformaldehyde for 1 h, permeabilized in PBS with 0.2% Triton X-100, washed 5 min in PBS and blocked in 25% goat serum for 30 min. Then the tissue sections were treated overnight with goat anti-hexon antibody (Chemicon) at 4°C. The sections were washed in PBS three times for 5 min. The tissue sections were treated with Alexa 534 labeled anti-goat secondary antibody (Alexa 534, Molecular Probes, Invitrogen) for 1 h, washed three times in PBS. All sections were mounted with VectaShield mounting medium with 4′,6-diamidino-2-phenylindole (H-1200; Vector Laboratories). The images were captured with either a Texas Red or a FITC filter in an Olympus AX70 fluorescence microscope equipped with a Zeiss Axiocam camera (Carl Zeiss, Oberkochem, Germany). Individual images were processed and merged using Adobe Photoshop 5.5 application software.

Statistical methods

We used the Wilcoxon rank sum test to perform pairwise comparisons of treatment effect on lung weight between all groups. Survival was measured from the day of MDA-MB-231 cell injection until the day of 130. For the survival data, the log-rank test was used to assess differences in survival among the four treatment groups. Because this overall test showed that the difference between MSC-Ad5/3.CXCR4–treated and control mice was statistically significant (P < 0.001), pairwise log-rank tests were performed. All statistical tests were two-sided; a P value of less than 0.05 was considered statistically significant. Statistical analyses were performed by using GraphPad Prism software (GraphPad Software, San Diego, CA).

Results

Identification of a CRAd allowing efficient loading of hMSCs and maximal oncolysis of breast cancer cells

One prerequisite of exploiting hMSCs as cell carriers for CRAds in vivo was to identify a CRAd agent, which combined efficient loading of the hMSCs with maximal killing potency of tumor cells. Therefore, our goal was to identify a CRAd possessing a limited oncolytic activity in the carrier cells in vitro while exhibiting efficient cytotoxicity in breast cancer cells. In this regard, previous studies have demonstrated that efficient transduction of hMSCs as well as breast cancer cells were hampered due to the paucity of the primary adenoviral receptor CAR. Thus, we determined the transduction efficiency of a panel of replication competent Ads, which utilize CAR-independent viral entry pathways. As shown in Fig. 1, we tested the cytotoxicity in hMSCs and in the metastatic breast cancer cell line MDA-MB-231. The hMSCs and MDA-MB-231 cells were infected with replication competent Ads that have: wild-type Ad5 fiber (Ad5wt), an RGD peptide incorporated into the HI loop of the Ad5 fiber knob domain (Ad5RGDwt and Ad5RGD.CXCR4) or a serotype switching of the Ad5 knob with that of Ad3 (Ad5/3wt and Ad5/3.CXCR4). In these experiments, Ad.CXCR4Luc served as a non-replicating control. While the crystal violet staining-based cell-killing assay showed that hMSCs were most sensitive to RGD peptide-containing Ads (Ad5RGDwt and Ad5RGD.CXCR4), the Ad5wt, Ad5/3wt, and Ad5/3.CXCR4 viruses had an attenuated cytopathic effect on hMSCs. In MDA-MB-231 cells, the most prominent oncolysis was attributed to Ad5/3 viruses (Ad5/3wt and CRAd Ad5/3.CXCR4). Thus, the Ad5/3.CXCR4 exhibited sufficient hMSC infectivity with limited cytotoxic effect, while it was substantially more cytopathic in MDA-MB-231 cells. Based on these data, we selected the CRAd Ad5/3.CXCR4 for our subsequent in vitro and in vivo experiments. Cells carrying this CRAd were designated as hMSC-Ad5/3.CXCR4.
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-006-9449-8/MediaObjects/10549_2006_9449_Fig1_HTML.gif
Fig. 1

Cytopathic effect of infectivity enhanced CRAds in breast cancer cells and human mesenchymal stem cells as carrier cells in vitro. (A) MDA-MB-231 cells and (B) human mesenchymal stem cells (hMSCs) were infected with Ad.CXCR4Luc (as a non-replicating control), the CRAd Ad5/3.CXCR4, the CRAd Ad5RGD.CXCR and the replication competent vectors Ad5RGDwt and Ad5wt at different MOIs. Cytotoxic activity was evaluated by crystal violet staining. Control without any viral infection is indicated as C

Ad5/3.CXCR4 loaded human mesenchymal stem cells display oncolysis of breast cancer cells in vitro

Next, we tested the proof-of-principle that Ad5/3.CXCR4 loaded hMSCs (hMSC-Ad5/3.CXCR4 cells) could affect oncolysis of MDA-MB-231 breast cancer cells in vitro. In addition, we evaluated the time course by which hMSC-Ad5/3.CXCR4 could produce oncolysis of MDA-MB-231 cells in vitro. As shown in Fig. 2, MDA-MB-231 cells were co-cultured with hMSCs loaded with Ad5/3.CXCR4 at MOI ranging from 0 to 1000 oncolysis, and oncolysis was assessed by crystal violet staining on days 3, 5, 7, 9 and 11. To exclude any toxic effects of the adenovirus infection itself that may also contribute to cell lysis, tumor cells were also mixed with HMSC carrier cells infected with non-replicative virus. In addition, any oncolytic effect from adenovirus infection alone was assessed by direct infection with non-replicative virus. These control experiments demonstrated that non-replicative Ad does not cause cell lysis of breast cancer tumor cells when applied directly or carried in hMSCs.
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-006-9449-8/MediaObjects/10549_2006_9449_Fig2_HTML.gif
Fig. 2

Ad5/3.CXCR4 loaded human mesenchymal stem cell display oncolysis in breast cancer cells in vitro. (A) Co-culture of Ad5/3.CXCR4 loaded human mesenchymal stem cells (hMSC-Ad5/3.CXCR4) and MDA-MB-231 breast cancer tumor cells at days 3, 5, 7, 9 and 11. Oncolytic activity was evaluated by crystal violet staining. (B) Direct oncolytic effect of Ad5/3.CXCR4 on MDA-MB-231 cells by infection with Ad5/3.CXCR4 at same MOIs at day 3, 5, 7, 9 and 11. Both experiments (A) and (B) were started in parallel at the same time point. Oncolytic activity was evaluated by crystal violet staining

At day 3 of the co-culture experiment, hMSC-Ad5/3.CXCR4 cells showed evidence of initial oncolysis in MDA-MB-231 cells at MOI of 1000, which was completed at day 7. Interestingly, between days 7 and 9 of co-culture, an increase of oncolysis of MDA-MB231 of more than 3 orders of magnitude was achieved, resulting in complete oncolysis at MOI of 1. These results were correlated with PCR assays of Ad DNA in the culture media, which showed a significant increase in viral copy number between days 7 and 9 corresponding to enhanced viral oncolysis (data not shown). In the aggregate, co-culturing of MDA-MB-231 cells with hMSC-Ad5/3.CXCR4 resulted in increased oncolysis of MDA-MB-231 cells with time, suggesting viral amplification in hMSCs. The killing of MDA-MB-231 was protracted, thus indicating a sufficient time window to manipulate hMSCs with CRAds.

MSC-Ad5/3.CXCR4 homes to breast cancer metastases in the lung in vivo

We investigated the homing capacity of hMSCs loaded with CRAds to breast cancer metastases in the lung in vivo. In this experiment, we injected MDA-MB-231 cells intravenously into the tail veins of SCID mice to establish pulmonary metastases as previously described [21]. Fourteen days later, the mice were injected intravenously with Ad5/3.CXCR4 loaded hMSCs and labeled with fluorescent dye, with Ad5/3.CXCR4, or with fluorescent dye labeled hMSCs alone. At day 3 after treatment the mice were sacrificed and lung samples were assessed for the presence of hMSCs loaded with CRAds. Fluorescence microcopy was performed to identify cell populations in lung sections. Dye labeled hMSCs were identified by green fluorescence and cells immunostained against Ad hexon (as an indicator of viral replication [1]) were identified by red fluorescence. HMSCs carrying CRAd would appear as orange overlay. As shown in Fig. 3A, fluorescent microscopy demonstrated orange fluorescent cells indicative of hMSCs positive for Ad hexon localized to tumor nodules in lung tissue from tumor bearing mice injected with Ad5/3.CXCR4 loaded hMSCs. In lung tissue from tumor bearing mice injected with Ad5/3.CXCR4, only red fluorescent cells positive for Ad hexon immunostaining were observed (Fig. 3B). In tumor bearing mice injected with HMSCs alone, these cells were detected in the tumor nodules as indicated by green fluorescence (Fig. 3C). Interestingly, no evidence of hMSCs could be found in non-MDA-MB-231 bearing lungs (data not shown). Thus, these results suggest that hMSCs loaded with Ad5/3.CXCR4 are capable to home to breast cancer metastases in the lung after systemic injection.
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-006-9449-8/MediaObjects/10549_2006_9449_Fig3_HTML.jpg
Fig. 3

Evaluation of homing capacity of systemically administered CRAd loaded human mesenchymal stem cells to metastatic breast cancer to the lungs by fluorescence microscopy. Established pulmonary metastases of MDA-MB231 carcinoma intravenously injected with hMSC-Ad5/3.CXCR4 (A), Ad5/3.CXCR4 (B) or hMSC (C). Fluorescence microscopy was performed to detect hexon of Ad5/3.CXCR4 (red, see B), GFP label of hMSCs (green, see A and C) and hMSC carrying hexon (orange, see A →) in established pulmonary metastatses of MDA-MB231 carcinoma at day 3 after treatment

Systemically administered MSC-Ad5/3.CXCR4 reduces the growth of MDA-MB-231 cell derived lung metastases in vivo

Next, we investigated the in vivo anti-tumor activity of hMSC-Ad5/3.CXCR4. We injected MDA-MB-231 cells intravenously into the tail veins of SCID mice to establish pulmonary metastases (Fig. 4A). Fourteen days later, we injected 1 × 106 hMSCs loaded with Ad5/3.CXCR4 (Fig. 4B) or Ad5/3.CXCR4 alone (Fig. 4C) intravenous. Control mice received either no treatment (Fig. 4A) or intravenous injection of 1 × 106 hMSCs (Fig. 4D). Thirty days after tumor cell injection, the mice were sacrificed, and the weights of whole lungs were measured. A group of healthy mice that received no cell injection served as a reference for measurement of normal lung weight. As shown in Fig. 5, the mean lung weight of mice injected with MDA-MB-231 tumor cells was statistically significantly greater than the mean lung weight of healthy mice. Much of this weight difference was due to the tumor tissue occupying substantial portions of the lungs of the mice injected with the tumor cells. Therefore, we used whole lung weight as a surrogate endpoint of tumor burden in lungs and as an assessment of treatment, as previously described [21]. Mice injected with tumor cells and intravenously with hMSC-Ad5/3.CXCR4 had smaller mean lung weight than Ad5/3.CXCR4 treated mice or control untreated mice injected with tumor cells only (P < 0.05). Thus, the tumor burden of breast cancer metastases in the lungs was significantly less in animals treated with CRAd loaded hMSCs than with the CRAd alone.
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-006-9449-8/MediaObjects/10549_2006_9449_Fig4_HTML.gif
Fig. 4

Treatment schedules of MDA-MB-231 cell derived lung metastases in vivo. Pulmonary breast cancer metastases were established in mice (A). Mice were treated on day 14 after intravenous injection with MDA-MB231 with one dose of intravenous injection of hMSC-Ad5/3.CXCR4 (B), Ad5/3.CXCR4 alone (C), hMSC alone (D), or no treatment (E). Representative lung sections are shown, lungs were analyzed by histology and H&E staining

https://static-content.springer.com/image/art%3A10.1007%2Fs10549-006-9449-8/MediaObjects/10549_2006_9449_Fig5_HTML.gif
Fig. 5

Effect of systemically administered hMSC-Ad5/3.CXCR4 on the weight-growth of MDA-MB-231 cell derived lung metastases in vivo. Lung weights after treatment with hMSC-Ad5/3.CXCR4, Ad5/3.CXCR4 or hMSC compared with those of untreated mice with MDA-MB231 metastasis at day 30 after tumor cell injection. Lung weights of healthy animals with no tumors are included for comparison. Each bar presents the mean of four experiments ± SD. *, P < 0.05, hMSC-Ad5/3.CXCR4 versus Ad5/3.CXCR4

Treatment with hMSC-Ad5/3.CXCR4 improved survival of mice bearing breast cancer metastases in the lungs

Finally, we examined whether hMSCs loaded with Ad5/3.CXCR4 improved the survival of mice with pre-established pulmonary metastases derived from MDA-MB-231 cells. Mice were treated either with hMSC-Ad5/3.CXCR4, Ad5/3.CXCR4, or with hMSCs alone. Untreated mice served as a control group. Among mice bearing pulmonary metastases derived from MDA-MB-231 cells, the group treated with hMSC-Ad5/3.CXCR4 survived significantly longer (P < 0.05) than Ad5/3.CXCR4 treated, hMSC treated or untreated mice (Fig. 6). Thus, these results indicate that the delivery of CRAds with hMSCs resulted in a greater survival benefit compared to CRAd injection alone.
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-006-9449-8/MediaObjects/10549_2006_9449_Fig6_HTML.gif
Fig. 6

Survival analysis. Survival of mice with established pulmonary metastases of MDA-MB231 breast carcinoma intravenously injected with hMSC-Ad5/3.CXCR4, Ad5/3.CXCR4 or hMSC compared with those of untreated mice

Discussion

In the present study, we have exploited the utility of hMSCs as cellular carriers of CRAds for the therapy of metastatic breast cancer to the lung. We have demonstrated both in vitro and in vivo the capability of hMSCs as intermediate carriers for CRAds. Importantly, we have shown the ability of systematically administered CRAd loaded hMSCs to home to and to kill breast cancer pulmonary metastases. The exploitation of mammalian cells represents a novel approach of coupling virotherapy with cell therapy. In this regard, delivery of conditional replication HSV type 1 viruses by neural precursor cells has been endeavored for the treatment of glioma [5]. Tumor cells have also been used as cellular carriers of replication competent parvoviruses for the treatment of hepatic cancer [16]. Recently, the ability of mesenchymal stem cells to deliver replication competent adenoviruses to ovarian and cervical cancer cells in vitro was investigated [15]. Herein, our present study extends this paradigm to an in vivo approach to distant metastatic cancer.

In the design of our study we investigated a combined cell vehicle therapy and virotherapy approach by exploiting a CRAd agent designed with enhanced tumor infectivity and specificity in vitro as well as anti-tumor potency in vivo. With respect to virotherapy, our CRAd was designed to address the biological requirements of breast cancer. Tumor infectivity enhancement of this CRAd was achieved by incorporation of the knob domain from the human adenovirus serotype 3 (Ad5/3) to provide CAR-independent tropism [19]. Enhanced specificity was achieved by placing the CRAd E1A gene under the transcriptional control of the tumor selective promoter CXCR4. In this regard, the CXCR4 gene promoter revealed a superior “breast cancer—on/liver-off” profile in a preclinical evaluation of transcriptional targeting strategies for carcinoma of the breast in a tissue slice model system [18]. Furthermore, recent evidence points to the SDF-1α-CXCR4 complex as having an important role in progression to metastasis in several tumor contexts [13, 24]. This CRAd combining transductional and transcriptional targeting strategies has been recently shown to exhibit superior infectivity and tumor selective replication in breast cancer (Stoff-Khalili et al., unpublished observations).

The present study addresses the utility of hMSCs as intermediate carriers for CRAds to target metastatic breast cancer in vivo. The choice of hMSCs as cellular vehicles was based on their described tumor homing capacity and the observation that intravenously administered hMSCs do not engraft in healthy organs (i.e. lung, liver, spleen, kidney and muscle) [7, 14, 21]. Indeed, fluorescence microscopy of lung tissue sections showed homing of labeled hMSCs to the breast cancer metastases, while no hMSCs could be identified in healthy regions of the lung. It has been suggested that signals mediating increased turnover and proliferation of connective stromal cells in tumors may induce the homing of MSCs to this site [20, 21]. In this regard, it has been recently suggested that this process may be related to high local concentrations of paracrine growth factors such as fibroblast growth factor, platelet-derived growth factor, epidermal growth factor, transforming growth factor-β, or other mediators within the tumor microenvironment [2].

In order to exploit hMSCs as cellular carriers for CRAds, it was important to consider how to limit viral activity in the carrier cells until the hMSCs would reach the metastatic breast cancer cells in the lung. We selected the Ad5/3.CXCR4 CRAd as optimal for our strategy because this agent exhibited limited cytotoxicity in hMSCs and maximal cytotoxicity in breast cancer cells. Indeed, the identification of labeled hMSCs loaded with the CRAd at metastatic tumor sites in the lungs after systemic injection suggested a sufficient time window for the hMSCs to carry this agent to the disease. The prevention of cytotoxicity of viruses on cell carriers is an important practical issue relevant to our strategy [8, 11]. In this regard, several approaches have been designed to attenuate the cytotoxicity of viruses on cell carriers, such as mimosine treatment to avoid premature destruction of neural precursors serving as cellular carriers for conditionally replicative herpes simplex virus type 1 vectors [5]

We demonstrated here that hMSC based viral delivery enhanced the oncolytic effect of virotherapeutic treatment and increased the survival of tumor-bearing animals. This result validates the feasibility of using hMSCs as cellular carriers for replication competent adenoviruses. The amount of virus loaded into hMSCs was measured at the time of in vivo delivery and corresponded to the titer of CRAd injected alone. Therefore, the superior therapeutic response of metastatic breast cancer in the lung from treatment with CRAd loaded hMSCs cannot be attributed to a higher dose of the CRAd initially introduced. This response may result from the amplification of viral load in the hMSC carriers or from targeted virus delivery and release of the virus in the vicinity of tumor cells at multiple foci. Further studies are in progress to investigate our hypothesis that hMSCs may function as local factories of CRAds to explain the superior oncolysis of CRAd loaded hMSCs compared to CRAd alone in vivo. In our study, we used an immune-compromised animal model. In future studies, we will investigate if hMSCs could also serve as a “Trojan horse” for CRAd delivery by evading an antiviral immune response. As a result of the transport of CRAds inside hMSCs, CRAds may be protected not only from untargeted trapping but also from inactivation through hemaglutination and neutralization through preexisting antiviral immunity.

In conclusion, this study represents to the best of our knowledge, the first attempt to demonstrate the proof of principle that hMSCs can serve as cellular carriers to deliver CRAds to distant tumors such as breast cancer metastases in the lungs and mediate oncolysis in vivo after systemic administration. This work provides the first evidence to suggest that hMSCs as carrier cells represent a promising mode of administration of CRAds for the treatment of distant neoplastic metastatic disease.

Acknowledgements

This work was supported by Grant of the Deutsche Forschungsgemeinschaft Sto 647/1-1 (to M. A. Stoff-Khalili), by grants from the National Institutes of Health 5T32CA075930 (NIH Training Grant) and Department of Defense: W81Xwh-05-1-035 (to D. T. Curiel). R01CA93796, R01CA98543 and AR46031 (to G. P. Siegal) and from the Louisiana Gene Therapy Research Consortium, Inc. (to J. M. Mathis) and R01CA108585 (to J. T. Douglas).

Copyright information

© Springer Science+Business Media, LLC 2006