Cancer Gene Therapy (2014) 21, 200–208 & 2014 Nature America, Inc. All rights reserved 0929-1903/14 www.nature.com/cgt

ORIGINAL ARTICLE

Human placenta mesenchymal stem cells expressing exogenous kringle1-5 protein by fiber-modified adenovirus suppress angiogenesis Y Chu1,3, H Liu2,3, G Lou2, Q Zhang1 and C Wu1 Anti-angiogenesis gene therapy is considered a promising treatment for excessive vascularization. Mesenchymal stem cell (MSC)-based gene therapy may enhance the effect of anti-angiogenesis by maintaining a long therapeutic period in vivo. However, transduction efficiencies and transgene expression in MSC-based gene therapy should be improved. Here we report human placenta-derived MSC (HPMSC)-based gene therapy using a fiber-modified adenoviral vector carrying the kringle1-5 gene to maintain long-term survival and effectively suppress angiogenesis both in vitro and in vivo. HPMSCs infected by the adenoviral vector were transduced at high efficiency with a low multiplicity of infection, and the infected HPMSCs expressed exogenous kringle1-5 protein in vitro and in vivo. Infected HPMSCs were detected at 2 weeks in vivo by fluorescence imaging and immunohistochemistry of reporter gene expression. Importantly, the microvessel growth of aortic rings in vitro was inhibited by administration of infected HPMSCs expressing kringle1-5 protein (K1-5-HPMSCs) at day 6. In Matrigel plugs combined with K1-5-HPMSCs, microvessel density was decreased as detected by immunohistochemistry and blood flow was decreased as detected by the power Doppler contrast enhanced at day 14. The fiber-modified adenovirus is an effective gene vector for HPMSC-based gene therapy, which may be a promising strategy for cancer anti-angiogenesis. Cancer Gene Therapy (2014) 21, 200–208; doi:10.1038/cgt.2014.19; published online 23 May 2014

INTRODUCTION Angiogenesis theory was first proposed by Folkman in 1971. Angiogenesis inhibitors are essential for some diseases, such as the spread and growth of malignant tumors, and are considered a promising target for treating excessive angiogenesis. Angiogenesis inhibitors arrest neovascularization in tumors and in some chronic diseases by blocking endothelial cell (EC) activity, and gene therapy can enhance their effect.1–3 Among angiogenesis inhibitors, kringle1-5 (K1-5) protein, a kind of plasminogen degradation fragment, shows stronger inhibition of proliferation and angiogenesis of ECs than other inhibitors such as endostatin and angiostatin.4,5 K1-5 is a 55-KDa protein containing five plasminogen fragments and has a longer half-life than other kringle fragments including angiostatin (kringle1-4).4 However, administration of angiogenesis inhibitors alone, including K1-5, does not maintain their effects for a long period.6 We propose that human mesenchymal stem cell (MSC)-based gene therapy may enhance the effect of K1-5 protein on anti-angiogenesis by maintaining a long therapeutic period in vivo. Stem cells include hematopoietic and epithelial stem cells and MSCs. MSCs can be obtained from various tissues and cultured abundantly without specialized culture medium. Importantly, MSCs can self-renew for long-term survival in vivo and are relatively nonimmunogenic without immunomodulation or subsequent immunosuppressive therapy for the recipient.7 Human MSC transplantation into patients can easily overcome the difficulties related to the immune rejection of transplanted

cells.8,9 Thus, human MSC-based gene therapy may be an effective approach for long-term delivery of therapeutic agents for treating human diseases. However, this gene delivery system has failed to reach clinical use because of low-efficiency gene transfer and short-term transgene expression. The extended life span of MSCs makes them particularly interesting as cell factories for the cell-based gene therapy. Presently, adenoviral vectors have been widely used for gene transfer because they can be produced at high titers and can infect various cell types including quiescent and non-quiescent cells.10 Among several adenoviral vectors, the non-replicative subgroup C adenovirus serotype 5 (Ad5) vector is relatively safe and well studied in animal models.11 Infection by Ad5 is mediated by high-affinity binding of the fiber knob to the coxsackie– adenovirus receptor (CAR).12 However, it has reported that the commonly used Ad5 does not efficiently transduce MSCs, both in vitro and in vivo, because MSCs lack expression of the adenovirus-binding receptor CAR.13,14 An alternative strategy to improve Ad5 vector tropism in CAR-negative cells is to alter the fiber shaft.15,16 Members of adenovirus subgroup B, such as Ad11, bind somatic cells via adhesion molecules CD46 and desmoglein 2, more efficiently infect MSCs compared with that of Ad5.17,18 We proposed that Ad5 with an Ad11 fiber modification may improve the clinical utility of Ad5 vectors for gene therapy. Indeed, we and other researchers have previously shown that Ad5F11B, a Ad5 vector that is fiber modified by the Ad11 fiber, enhances the transduction of MSCs.19,20

1 Department of Ultrasound, First Affiliated Hospital of Harbin Medical University, Harbin, China and 2Department of Gynecological Oncology, Harbin Medical University Cancer Hospital, Harbin, China. Correspondence: Professor C Wu, Department of Ultrasound, First Affiliated Hospital of Harbin Medical University, Youzheng Street No. 23, Nangang District, Harbin 150001, China. E-mail: [email protected] 3 These authors contributed equally to this work Received 6 October 2013; revised 9 April 2014; accepted 14 April 2014; published online 23 May 2014

Human placenta mesenchymal stem cells Yu Chu et al

201 In this study, we explored the efficacy of the fiber-modified adenovirus Ad5F11B, carrying K1-5 gene and enhanced green fluorescent protein (EGFP) genes, for human placenta-derived MSC (HPMSCs)-based anti-angiogenesis gene therapy.

MATERIALS AND METHODS Cell isolation and culture HPMSCs were isolated from placental tissue using collagenase II (Invitrogen, Carlsbad, CA, USA) digestion. Human placental tissues were obtained from healthy donor mothers (n ¼ 13; 37–40 weeks’ gestation), who provided informed consent. The tissue specimen of human placenta were cut into pieces as small as possible B1 mm3 and were washed in phosphate-buffered saline. The fragmentation tissues were digested with 1% collagenase II (Invitrogen) at 37 1C for 30 min. Single-cell suspensions of human placental tissues were made by flushing the tissue parts through a 100-mm nylon filter (Falcon, Becton, Dickinson, San Jose, CA, USA) using washing medium. Cells were cultured in Dulbecco’s modified Eagle’s medium-nutrient mixture F12 (DMEM-F12, Invitrogen) supplemented with 10% fetal bovine serum (GIBCO, Carlsbad, CA, USA), 100 U ml  1 penicillin and 100 mg ml  1 streptomycin (Invitrogen). Adherent cells were harvested at confluence and used at passage 5–7 for our experiments. Human placental tissue-derived cells at passage 5–7 were cultured in six-well plates in osteogenic (0.1 mM dexamethasone, 10 mM b-glycerol phosphate and 50 mM ascorbate) or adipogenic (1 mM dexamethasone, 5 mg ml  1 insulin, 0.5 mM isobutylmethylxanthine and 60 mM indomethacin) medium. Adipogenic differentiation was assessed using oil red O staining after 14 days, and osteogenic differentiation was evaluated using 2% silver nitrate staining after 28 days. All chemicals for differentiation assays were purchased from Sigma-Aldrich (St Louis, MO, USA).

Flow cytometry HPMSCs were cultured to passage 5–7 and then, CD molecule expression was detected by flow cytometry (BD Biosciences, Franklin Lake, NJ, USA). Cells were detached using 0.04% EDTA (Invitrogen), washed and resuspended at 1  105 cells per 100 ml phosphate-buffered saline þ 2% bovine serum albumin, followed by incubation with specific antibodies at 4 1C. Then, cells were washed with phosphate-buffered saline þ 2% bovine serum albumin and analyzed by flow cytometry. HPMSCs were characterized using the following antibodies: anti-CD73, anti-CD166, anti-CD31 (BD Biosciences), anti-CD105 and anti-CD14 (BioLegend, San Diego, CA, USA).

Western blotting HPMSCs were seeded in T-75 culture flasks (1  106 cells per flask) and infected with rAd-Mock or rAd-K1-5 at an MOI of 50 for 2 h. Then, the medium was changed with serum-free medium. After 48 h, conditioned medium of HPMSCs, Mock-HPMSCs and K1-5-HPMSCs were harvested. K1-5 protein expression in HPMSCs was assessed by western blotting. Briefly, 10 ml conditioned medium from cultures of various types of HPMSCs was subjected to 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then transferred to nitrocellulose membranes. Blots were probed with a mouse monoclonal anti-angiostatin antibody (1:500; BD Biosciences), followed by a horseradish peroxidase-conjugated secondary antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) and then visualized by enhanced chemiluminescence (Western Lightning; PerkinElmer Life Sciences, Waltham, MA, USA). Membranes were stripped by incubating in western blot stripping buffer (Restore; Pierce, Rockford, IL, USA) at 37 1C and then re-blocked in 1% bovine serum albumin for further probing. Membranes were developed with an alkaline phosphatase solution (Promega, Madison, WI, USA) according to the manufacturer’s instructions. The 55-KDa protein band was scanned to measure the density of staining of K1-5 protein and the 42-KDa protein band was scanned to measure the density of staining of the actin as the correct internal control.

Aortic ring assay All animal experiments were performed in accordance with the guidelines of and with approval from the institutional animal care and use committee of all foundations. Growth factor-enriched Matrigel (BD Biosciences) was prepared just before use under sterile conditions. Matrigel was thawed for 24 h at 4 1C. Forty-eight-well plates were coated with 100 ml Matrigel and allowed to form a gel for 45 min at 37 1C with 5% CO2. In our experiment, to prepare aortic rings, we have chosen one abdominal aorta from the same rat in each test. All assays were performed six times. The abdominal aorta was excised from 8-week-old female Sprague-Dawley rats (Shanghai Experimental Animal Center, Shanghai, China), weighing 210–270 g, and the fibroadipose tissue was removed. Aortas were cut into 1 mm cross-sections and then rinsed four times with phosphate-buffered saline. Sections were placed in Matrigel-coated wells, covered with an additional 50 ml Matrigel and allowed to gel for 45 min at 37 1C with 5% CO2. Aortic rings were cultured for 24 h in DMEM-F12 supplemented with 2% fetal bovine serum. Then, the medium was removed and replaced with fresh medium supplemented with 2% fetal bovine serum and mixed with 100 ml cell suspensions (1  106 cells) of K1-5-HPMSCs, Mock-HPMSCs, HPMSCs or DMEM-F12 alone as the control. Aortic rings were photographed on day 6. Analysis of microvessel growth was performed by calculating tube length using the NIH Image software program (NIH, Bethesda, MD, USA). For each analysis, ten randomly chosen fields were measured and averaged.

Gene transduction A fiber-modified recombinant adenoviral vector Ad5F11B carrying the human K1-5 gene and EGFP gene (rAd-K1-5) or EGFP gene alone (rAdMock) were purchased from Shanghai Yiyuan Biology Company (Shanghai, China). The fiber-modified vectors Ad5F11B-EGFP and Ad5F11B-Kringle1-5EGFP used in this study have been described as following. In the Ad5F11BKringle1-5-EGFP vectors, E1 was replaced by an expression cassette consisting of the cytomegalovirus gene promoter and the open reading frame encoding the kringle1-5 protein, and E3 was replaced by an expression cassette consisting of the cytomegalovirus gene promoter and the open reading frame encoding the EGFP. In the Ad5F11B-EGFP vectors, E1 was replaced by an expression cassette consisting of the cytomegalovirus gene promoter, and E3 was replaced by an expression cassette consisting of the cytomegalovirus gene promoter and the open reading frame encoding the EGFP. HPMSCs were seeded in T-75 culture flasks (1  106 cells per flask) and incubated overnight. The cells were washed three times with DMEM-F12, and the medium was replaced with culture medium containing the vector at an multiplicity of infection (MOI) of 50. After 2 h, the virus-containing culture medium was removed. The cells were washed three times with DMEM-F12 and then cultured in DMEM-F12 containing 10% fetal bovine serum for 48 h. The transduction efficiency was evaluated by EGFP expression under a fluorescence microscope (Olympus, Tokyo, Japan), and the number of EGFP-positive cells was analyzed by flow cytometry (BD Biosciences). HPMSCs, including uninfected HPMSCs, infected with rAd-Mock (Mock-HPMSCs) and HPMSCs infected with rAd-K1-5 (K1-5-HPMSCs) were used in the following experiments. DMEM-F12 alone was used as the control group. & 2014 Nature America, Inc.

Matrigel plug assay All animal experiments were approved by the Experimental Animal Committee of Harbin Medical University. Twenty-four female BALB/c nude mice (5–6-weeks-old, weighing 15–17 g) were purchased from the Shanghai Experimental Animal Center. Matrigel was prepared as described above. Matrigel solutions (500 ml) mixed with 1  106 K1-5-HPMSCs, MockHPMSCs, HPMSCs or DMEM-F12 alone were injected subcutaneously into the left waist side of nude mice.

Luminescence imaging Nude mice, which were injected Matrigel mixed with various types of HPMSCs, were monitored persistently by a LT-9500 Illumatool TLS Imaging System (Lightools Research, Encinitas, CA, USA) from day 1 to 14. At day 14, the implanted Matrigel plug was removed and also imaged using the LT9500 Illumatool TLS Imaging System. The images were photographed with a digital camera (Canon EOS500D, Tokyo, Japan). Imaging was performed every 3 days.

Blood flow ultrasound imaging and analysis Twenty-four female BALB/c nude mice (Shanghai Experimental Animal Center) implanted with Matrigel plugs were studied at day 14. An Esoate Mylab 90 system and LA435 transducer (Esoate, Genova, Italy) were used for ultrasound imaging. The frequency of the transducer was 6–18 MHz. Matrigel plugs in animals were first individually imaged in B-mode (brightness mode). Then, the transducer was positioned to image the Cancer Gene Therapy (2014), 200 – 208

Human placenta mesenchymal stem cells Yu Chu et al

202 Matrigel plug at the largest transverse cross-sections. Next, the ultrasound system was set to power Doppler with 100% transmit power, and a baseline pre-contrast image was recorded. The second generation contrast agent SonoVue (Bracco Diagnostics, Milan, Italy) was diluted with 5 ml sterile saline, and 0.02 ml ultrasound contrast agent was injected via the tail vein with an insulin syringe. Contrast-enhanced images were recorded within 1 min after injection. The percentage of contrast-enhancement area of blood flow in Matrigel plug was calculated. The cross-sectional area of each region of interest (cm2) was extracted from a pre-contrast B-mode image of each plug. From post-contrast images, contrast-enhancement area of blood flow in Matrigel plug was calculated. (The fraction of enhanced pixels ¼ number of enhanced pixels/ number of pixels in region of interest). More than one post-contrast image was saved for each animal, and results were averaged from saved images (n ¼ 3) to reduce bias in image selection. The number of samples in every group was six. Images were analyzed by Image-Pro Plus v6.0 software (Media Cybernetics, Bethesda, MD, USA).

Expression of K1-5 protein and EGFP in Matrigel plugs After imaging at day 14 post implantation, mice were killed and Matrigel plugs were removed along with the underlying abdominal muscle and overlying skin for better histological orientation. The plugs were formalinfixed first, then they were paraffin embed. Routine paraffin-embedded histological sections were cut at 5 mm intervals. A primary mouse monoclonal anti-K1-5 antibody (BD Biosciences, San Jose, CA, USA) and a primary rabbit polyclonal anti-EGFP antibody (Abcam, Cambridge, UK) were used, followed by a horseradish peroxidase-conjugated secondary antibody (Santa Cruz Biotechnology). Sections were visualized by enhanced diaminobenzidine.

Histological observations and neovascularization study The paraffin-embedded plugs were cut at 5-mm intervals, and sections were stained with hematoxylin and eosin (HE). Detection of the specific ECs’ marker vWF was performed by immunohistochemistry using an antivWF antibody obtained from DAKO Novus Biologicals (Glostrup, Denmark) with a biotinylated anti-rabbit secondary antibody from Zhongshan Goldenbridge Biotechnology (Beijing, China). The reaction was developed with diaminobenzidine substrate (Zhongshan Goldenbridge Biotechnology), and sections were counterstained with Mayer’s hematoxylin (Zhongshan Goldenbridge Biotechnology). Sections were reviewed by Image-Pro Plus 6.0 software (Media Cybernetics) that quantified the number of microvessels per 10 high-power fields for each implanted Matrigel plug. The number of samples in every group was six.

Statistics Analysis of variance and student–Newman–Keuls method were used to compare measurement data among different groups. The Mann–Whitney test was used to analyze numeration data among groups. The w2 test was used to analyze the infection rate among groups. Results were considered statistically significant at Pp0.05. Sigmaplot software was used for analysis (Systat Software, Chicago, IL, USA).

RESULTS Properties, infection efficiency and transgene expression of HPMSCs We investigated whether the cells obtained from human placental tissue have the properties of MSCs (Figure 1a). Cells were adherent with a fibroblastic shape. Differentiation of cells into adipogenic and osteogenic lineages was verified by oil red O and silver nitrate staining, respectively (Figures 1b and c). Cells constitutively expressed CD166 and CD73 and did not express CD14, CD31 and CD105 as determined by flow cytometry (Figures 1d–h). The above results demonstrated that cells obtained from human placental tissue were MSCs. We used passage 5–7 human placental tissue-derived MSCs (HPMSCs) for the following experiments. Next, we detected the transgene expression efficiency of HPMSCs infected with the fiber-modified adenovirus Ad5F11B vector. We observed that EGFP were expressed both in MockHPMSCs and K1-5-HPMSCs groups (Figure 1i and j). As shown in Cancer Gene Therapy (2014), 200 – 208

Figure 1i and j, infection efficiency exceeded 90% at a MOI of 50 in the two groups and there was no difference in EGFP expression between Mock- and K1-5-HPMSCs, respectively (P40.05). K1-5 protein, 55KD, was markedly increased in K1-5-HPMSCs group compared with Mock-HPMSCs group at a MOI of 50 (Po0.01) (Figure 1k and l). HPMSCs in our study were efficiently infected by Ad5F11B and expressed exogenous K1-5 protein. We used the infected HPMSCs at a MOI of 50 for subsequent gene therapy experiments. Detection of infected HPMSCs in Matrigel plugs Next, we observed the activity of infected HPMSCs combined with a Matrigel plug, which transplanted subcutaneously into nude mice. At day 14, implanted Matrigel plugs were evaluated by optical imaging and exhibited bright fluorescence in Matrigel plugs combined with K1-5-HPMSCs and Mock-HPMSCs. In contrast, there was no fluorescence detected in Matrigel plugs combined with HPMSCs or DMEM-F12 alone (Figure 2a). These results also demonstrated that infected HPMSCs maintained survival for at least 2 weeks. At 14 days after implantation, Matrigel plugs were resected and immunohistochemistry showed that infected HPMSCs were present in resected Matrigel plugs as EGFP-positive cells (Figure 2b). The expression of EGFP was positive in K1-5-HPMSCs and Mock-HPMSCs groups, and negative in HPMSCs and DMEM-F12 groups. However, as shown in Figure2c, K1-5 protein was positive only in the K1-5-HPMSCs group, suggesting that therapeutic protein K1-5 can be expressed in the infected HPMSCs by fiber-modified adenovirus AD5F11B and the infected HPMSCs may express therapeutic protein in vivo for a long period. Anti-angiogenesis effects of infected HPMSCs in vitro and in vivo We first observed the effects of infected HPMSCs on the tube outgrowths of ECs in an aortic ring assay. As shown in Figure 3a, there was an obvious expansion of tubules in Mock-HPMSC, HPMSC and DMEM-F12 groups at day 6. In contrast, the expansion was suppressed by incubation with K1-5-HPMSCs. The percentage of the total tube length of neovessels of the K1-5-HPMSC group was decreased compared with those of all control groups (Figure 3b). The q-values are 7.657, 6.845, 5.058, 6.761, 5.120 and 4.933 between K1-5-HPMSCs and HPMSCs, HPMSCs and control, HPMSCs and Mock-HPMSCs, Mock-HPMSCs and K1-5-HPMSCs, Mock-HPMSCs and control, control and K1-5-HPMSCs, respectively. Po0.05 for all comparison. We investigated blood flow inside Matrigel plugs treated with various types of HPMSCs at 14 days after implantation. As shown in Figure 3c, all Matrigel plugs were distinct from the surrounding tissue in the B-mode ultrasound (Figure 3c). Power Doppler ultrasonography showed a little blood flow inside Matrigel plugs at day 14, and there were no obvious differences among four groups (Figure 3d). After the contrast media administration, blood flow was observed in all Matrigel plugs and was generally more prominent (Figure 3e). The K1-5-HPMSCs’ group showed significantly less enhancement than Matrigel plugs’ group combined with Mock-HPMSCs, HPMSCs or DMEM-F12 alone (Figure 3e). As shown in Figure 3f, the percentage of average contrast-enhancement area of blood flow within each plug was 46.0±5.8, 52.6±6.9 or 46.3±7.2% for Mock-HPMSC, HPMSC or DMEM-F12 Matrigel plugs, respectively, and 24.1±7.3% for K1-5 Matrigel plugs. The percent of average contrast-enhancement area of blood flow in the K1-5-HPMSCs’ group was obviously decreased compared with that in other groups (Po0.01), which suggest that HPMSCs expressing K1-5 protein can inhibit the migration of ECs and original blood flow in vitro and in vivo. The number of samples in every group was six. Next, we observed neovessels in Matrigel plugs at 14 days after implantation. Neovascularization was observed in all Matrigel plugs by HE staining (Figure 3g). In many & 2014 Nature America, Inc.

Human placenta mesenchymal stem cells Yu Chu et al

203

Figure 1. Properties, infection efficiency and exogenous protein expression of HPMSCs. (a) Cells obtained from human placenta were adherent fibroblast-like cells. (b and c) Cells derived from human placenta appeared positive in oil red O staining and silver nitrate assays. Lipid-containing vacuoles inside cells can be seen and the cell surface appeared black with calcium mineralization. (d–h) The CD adhesion molecules of cells obtained from human placental tissue were detected by flow cytometry. (i and j) Infection efficiency was tested for infected HPMSCs at MOI of 50 by fluorescence microscopy. (k and l) After transduction with the fiber-modified adenovirus Ad5F11B for 48 h in vitro, K15 gene and protein expressions in HPMSCs were determined by western blotting, respectively. Actin was used as the correct internal control. Scale bar ¼ 50 mm (a–c and i). Values are means±s.e. Ad5F11B, fiber-modified Ad5 vector that contained the Ad5 knob and Ad11 fiber; HPMSCs, human placenta-derived mesenchymal stem cells; MOI, multiplicity of infection.

vessels, intravascular erythrocytes were observed, indicating the presence of blood flow in these vessels. Average microvessel densities in Matrigel plugs combined with K1-5-HPMSCs, MockHPMSCs, HPMSCs or DMEM-F12 alone were 4.4±1.8, 9.3±2.6, 5.4±2.0 or 1.1±0.5, respectively. As shown in Figure 3h, the microvessel densities of K1-5-HPMSCs were lower than that of other groups (Po0.01). The number of samples in every group was six. DISCUSSION Angiogenesis has an important role in the spread and growth of malignant tumors and in some chronic diseases.1,2 Cell-based antiangiogenesis gene therapy has become a promising strategy for treating these diseases. Our study confirmed that HPMSCs & 2014 Nature America, Inc.

expressing exogenous K1-5 protein by infection with the fibermodified adenovirus Ad5F11B can maintain significant antiangiogenesis activity for a long period both in vitro and in vivo. The characteristics of transgenic cells, such as life span in vivo and the ability to self-renew, have important roles in gene therapy. As transgenic cells in gene therapy, MSCs have advantages including a longer life span and higher proliferation, compared with those of other cells.7 Our previous studies have shown that human adipose tissue-derived MSCs can express exogenous protein and maintain biological activity in vitro for more than 2 weeks.20 However, obtaining adipose tissue by liposuction is a harmful and invasive process, and we found that there were a small quantity of stem cells obtained from human adipose tissues. Importantly, MSCs are present in a low frequency in many adult Cancer Gene Therapy (2014), 200 – 208

Human placenta mesenchymal stem cells Yu Chu et al

204 tissues. Placenta is a discarded tissue and contains quite a high number of stem cells that have a strong ability to self-renew.21–24 In the present study, we easily extracted a large number of HPMSCs from a small quantity of placental tissue and found that the stem cells were more suitable as transgenic cells, which possess a longer life span and higher proliferation than stem cells derived from adult tissue and bone marrow (data not shown). This result was consistent with previous studies of the extraction and cultivation of human placenta derived MSCs.21–24 The choice of transgene vector is important for the application of effective gene therapy. Viral vectors are capable of infection with high efficiency. There are several viral vectors for gene transfer, such as retrovirus, adenovirus, lentivirus and so on. To

date, in clinical trials, adenoviral vectors have been most extensively used as gene delivery vehicles because they can transfer genes into various cell types.10 Importantly, the safety of adenoviral vectors has been extensively evaluated in many clinical trials. However, adenovirus-mediated gene therapy has shown insufficient transduction of MSCs.19,25 Low levels of transgene expression obtained with traditional adenoviruses can be compensated for by a high MOI that may induce cytotoxicity in MSCs in vitro and in vivo.26,27 In addition, transient expression of the transgene by adenovirus may be an issue that needs to be overcome when MSCs are used as therapeutic protein-release cells over a long period. Thus, searching of a new generation of adenoviral vectors should be accomplished for effective therapeutic application in gene therapy protocols. We found that a new generation of Ad5 including a fiber from the Ad3 vector, such as serotypes 11, 16 and 35, had displayed high efficiency for gene transfer.16,28,19 This new generation of Ad5 vector does not recognize the ubiquitous CAR, instead binding cells via the attachment receptor CD46/80.16,17 CD46 is a complement regulatory molecule and is present on many cell types including MSCs.25 We and other researchers have previously shown more effective infection of stem cells using a fibermodified Ad5 vector containing the Ad5 knob and Ad11 vector fiber.19,20 An additional advantage of using the fiber-modified Ad vector is the ability to maintain high gene expression for a longer period than the traditional Ad vector in MSC-based gene therapy. Some paper show that fiber-modified Ad vectors showed superior transduction of MSCs compared with conventional Ad5 vectors, such as their high foreign gene delivered by Ad5FBs in a large fraction of the cells for B3 weeks and mediated 130-fold transgene activity, compared with conventional Ad5 vectors.29,30 It has been demonstrated that the fiber-modified Ad5 vector can efficiently infect MSCs at a low virus titer.19,20 Importantly, infection with the fiber-modified Ad5 vector allows transgenic cells to express exogenous protein for a long period, compared with that of other adenoviruses.19,20 In the present study, we have described the use of Ad5F11B, a fiber-modified Ad5 vector that contains the Ad5 knob and Ad11 fiber for efficient gene transfer into HPMSCs in vitro. The fiber-modified Ad5 vector carrying the subgroup B Ad11 fiber, Ad5F11B-K1-5-EGFP, used in our study was E1 and E3 deleted, with the E3 region replaced by an EGFP expression cassette and the E1 region replaced by a K1-5 gene expression cassette. A mock vector, Ad5F11B-EGFP carrying EGFP only, was used as a control. We confirmed that the fiber-modified Ad5 vector not only infected HPMSCs, with high efficiency at a low MOI (50), but also showed high efficiency gene transfer into

Figure 2. Activity of infected HPMSCs in Matrigel plugs implanted in nude mice. Mice were injected the Matrigel plugs combined with infected HPMSCs, HPMSCs or Dulbecco’s modified Eagle’s mediumnutrient mixture F12 (DMEM-F12) alone. HPMSCs were transduced by the fiber-modified adenovirus Ad5F11B encoding K1-5 and EGFP genes or the EGFP gene alone. (a) The survival of infected HPMSCs in the Matrigel plug was determined by reporter gene imaging at day 14 post implantation. (a1–a4) are the fluorescence images of the Matrigel plug in vivo, (a5–a8) are the fluorescence images of the Matrigel plug ex vivo. The fluorescence image of representative mice (total, n ¼ 6 mice per group). (b) Expression of EGFP in Matrigel plugs, combined with DMEM-F12, HPMSCs, K1-5-HPMSCs or MockHPMSCs, implanted in mice was detected by immunohistochemistry. (c) Expression of K1-5 protein in Matrigel plugs, combined with DMEM-F12, HPMSCs, K1-5-HPMSCs or Mock-HPMSCs, implanted in mice was detected by immunohistochemistry. The plugs were stained by diaminobenzidine and hematoxylin and eosin in (b and c). Scale bar ¼ 200 mm (b and c). Ad5F11B, fiber-modified Ad5 vector containing the Ad5 knob and Ad11 fiber; EGFP, enhanced green fluorescent protein; HPMSCs, human placenta-derived mesenchymal stem cells; K1-5, kringle1-5. Cancer Gene Therapy (2014), 200 – 208

& 2014 Nature America, Inc.

Human placenta mesenchymal stem cells Yu Chu et al

205

Figure 3. Evaluation of angiogenesis in vitro and vivo. (a) Expansion of the aortic ring treated with infected HPMSCs. (b) The percentage of the tube length of neovessels among groups in aortic ring angiogenesis test. (c) Matrigel plug imaging under a B-mode. (d) The power Doppler displayed the vascularization in the Matrigel plug. (e)The power Doppler contrast-enhanced imaging displayed the vascularization in the Matrigel plug. (f ) Angiogenesis is depicted as the percentagement of average enhancement area. (g and h) Neovessels in Matrigel plugs at day 14 post implantation. The formation of new blood vessels was observed in all plugs by HE staining. The number of samples in every group was six. Scale bar ¼ 100 mm (a and g). HE, hematoxylin and eosin; HPMSCs, human placenta-derived mesenchymal stem cells.

& 2014 Nature America, Inc.

Cancer Gene Therapy (2014), 200 – 208

Human placenta mesenchymal stem cells Yu Chu et al

206 HPMSCs. Moreover, we demonstrated that these infected HPMSCs can express exogenous K1-5 protein for at least 2 weeks in vitro and in vivo. These results suggest that the Ad5F11B adenoviral vector can accomplish efficient gene transfer into HPMSCs that may then achieve improved exogenous protein expression in vivo. The studies demonstrated that Ad5F11B may be a promising gene vector for MSC-based gene therapy. It has been shown that a noninvasive method to monitor the fate of transplanted MSCs in vivo is necessary to evaluate the therapeutic continuity.31–34 The activity of transgenic MSCs is important for maintaining a long-term therapeutic effect, such as their longevity and function. In this study, we evaluated the longevity of HPMSCs by reporter gene imaging and reporter gene immunohistochemistry staining. Numerous imaging techniques such as optical imaging, magnetic resonance imaging, positron emission tomography, single-photon emission computed tomography and ultrasound have been used to monitor the in vivo behavior of MSCs.31–35 An ideal imaging technique for monitoring transplanted cells in vivo should be non-harmful, and optical imaging is mostly applicable to preclinical studies because of rapid imaging with less side effects. Moreover, in animal models, optical imaging including bioluminescence and fluorescence imaging is able to persistently detect transplanted cells expressing a reporter gene.36,37 In recent years, reporter genes have been used for imaging in vivo.37,38 After transplantation, infected stem cells can be tracked by expressing the reporter gene. In contrast, there is little tracking of the reporter gene after cell death. In our study, optical imaging was used to investigate the fate of transplanted HPMSCs. Moreover, our results demonstrated that a strong excitation signal with little optical noise was detected in mice carrying infected HPMSCs. Importantly, fluorescence imaging is specific, sensitive, convenient and a real-time method to observe transplanted cells in animals. In fluorescence imaging, there was bright fluorescence in Matrigel plugs with K1-5-HPMSCs and Mock-HPMSCs during the 14-day experimental period, although the intensity of fluorescence decreased slightly over time (data not shown). In addition, we demonstrated that expression of the EGFP reporter gene in the infected HPMSCs was maintained for at least 2 weeks, as shown by the immunohistochemical assay. Importantly, we have also confirmed that Matrigel plugs combined with K1-5-HPMSCs that were transplanted into nude mice show detectable K1-5 protein at day 14. These studies demonstrated that infected HPMSCs expressing exogenous K1-5 protein can maintain long-term survival and may be a promising cellar tool for gene therapy. MSCs can carry some therapeutic genes and express these proteins for a long period in cancer therapy. However, MSCs have been reported that they have some disadvantage in tumor gene therapy, such as promoting angiogenesis, growth and recurrence rate of cancer and so on. The purpose of our experiment is to inhibit the tumor neovascularization by reversing character of its promoting neovascularization. In this study, we further evaluated the function of infected HPMSCs by detecting Kringle1-5 (K1-5) protein. We demonstrated the effects of infected HPMSCs expressing exogenous K1-5 protein on angiogenesis in vitro. Some studies suggest that the use of rat aortic rings in an angiogenesis assay shows similar angiogenesis to that in humans and provides a more physiological model as tissue sections rather than cell lines alone.39,40 They demonstrated that mouse aorta rings generate microvessel outgrowths in fibrin or collagen gels and provide a sensitive assay to study angiogenic antagonists in a defined environment.39,40 Our experiments examined the effects of K1-5-HPMSCs on the microvessel outgrowths of aortic rings. In the aortic ring assay, there was an obvious expansion of tubules that were presumably a combination of endothelial, mesenchymal and perivascular cells.39,40 Tube formation was lower in the K1-5HPMSC group than that in other groups. This result was in Cancer Gene Therapy (2014), 200 – 208

accordance with our previous study.20 Our current study suggests that K1-5-HPMSCs expressing K1-5 protein may arrest angiogenesis in vitro. In our previous studies, the effects of MSC-based antiangiogenesis gene therapy were mainly assessed by the activity of ECs in vitro.20 However, it is complex to study angiogenesis in vivo. Various cell types such as endothelial and stromal cells and some cytokines are also involved in angiogenesis. Objective evaluation of angiogenesis should be conducted by an in vivo assay. It has been demonstrated that the Matrigel plug assay is an ideal model in vivo to evaluate blood vessel formation.41,42 Once Matrigel is injected subcutaneously, ECs migrate into the Matrigel plug to form new blood vessels. This angiogenic process can be accelerated by the addition of various growth factors, such as basic fibroblast growth factor and vascular endothelial growth factor, into Matrigel. Newly formed vessels can contain erythrocytes, indicating that they are functional capillaries.41,43 A large number of studies have shown that angiogenesis inhibitors can inhibit angiogenesis in a Matrigel plug assay.41,43 In this study, we used Matrigel with angiogenic factors for the angiogenesis assay, and analyzed the microvessel density in the Matrigel plug to indicate therapeutic effects. Consistent with our in vitro results, transgenic HPMSCs expressing K1-5 protein showed an antiangiogenic effect on Matrigel plug of nude mice containing the angiogenesis factor at 2 weeks. Microvessel densities measured by HE staining were lower in the K1-5-HPMSCs group than in the other groups. The effect was evaluated in terms of the number of functional capillaries containing red blood cells. This result suggested that infected HPMSCs expressing K1-5 protein suppressed capillary vessel formation in the Matrigel angiogenesis assay in vivo, which was consistent with our previous study using growth factor-reduced Matrigel.20 These results suggest that transplanted transgenic HPMSCs expressing the exogenous K1-5 protein can repress angiogenesis in vivo over a long period. Although the above studies have demonstrated the role of K1-5-HPMSCs in angiogenesis in vivo, HE staining can only detect vessels in one plane and cannot detect entire internal angiogenesis at the same time. Noninvasive techniques for imaging angiogenesis, such as magnetic resonance imaging, computed tomography, positron emission tomography and contrastenhanced ultrasonography, can monitor angiogenesis at diverse angles.31–35 Among these methods, ultrasonography is an ideal method because it is economical, rapid and easily applied. Importantly, it is harmless, does not release ionizing radiation and can provide real-time imaging.35,44 Contrast-enhanced ultrasound remains the most commonly available and used method because of its noninvasive imaging modality to evaluate angiogenesis for clinical detection.45 Power Doppler imaging is sensitive to small quantities of ultrasound contrast agent and it can detect microvasculature successfully with contrast agent microbubbles. Microbubbles, as a kind of ultrasound contrast agent, can enhance ultrasound imaging by entry into the intravascular space throughout circulation. In this study, we used the contrast agent SonoVue (sulfur hexafluoride microbubbles) via sonography as a sensitive tool for quantitative monitoring of neovessels. We monitored angiogenesis in Matrigel plugs with growth factors combined with transgenic HPMSCs in nude mice. In this model, neovascularization occurs both in the outer border and in the center of the plug. The average contrast-enhanced area of blood flow quantifies the blood stream in every direction, and we used the measurement indexes to evaluate the effect of K1-5-HPMSCs on anti-angiogenesis in Matrigel plugs. In the pre-contrast image, there was a low power Doppler signal of the blood stream in all groups. However, when the contrast agent was administrated, there were increasing power signal of the blood stream in all groups. Importantly, we found that the percentage of average contrast-enhancement area of blood flow was reduced in the K1-5-HPMSC group, compared with those of control groups. & 2014 Nature America, Inc.

Human placenta mesenchymal stem cells Yu Chu et al

207 Our results further demonstrate that K1-5-HPMSCs can arrest neovascularization in vivo at diverse angles. There are several limitations in our study. Reporter geneinduced cytotoxicity should be considered, although recent data suggest that the introduction of reporter genes does not appear to significantly alter the biological properties and differentiation capacity of stem cells.20,27 Also, the model of Matrigel angiogenesis in vivo may not be fully representative of clinical cell therapy of angiogenesis in cancer or other excessive angiogenesis disease. Here we present the first evidence that the fiber-modified adenovirus Ad5F11B may become a promising transgene vector to efficiently infect HPMSCs that may then maintain exogenous protein expression over a long period. Importantly, this result is a significant finding with important implications for the field of MSC-based gene therapy by evaluating a fiber-modified adenovirus, whereas previous attempts to use traditional adenoviral vectors have encountered limitations owing to low vector transduction efficiency and limited gene expression. Our research shows HPMSCs expressing exogenous K1-5 protein can maintain long-term longevity in vivo and induce an anti-angiogenic effect both in vivo and in vitro. Our present study provides a new strategy involving HPMSCs expressing K1-5 protein by local gene delivery to effectively suppress angiogenesis by a fiber-modified adenovirus.

10 11

12

13 14

15 16

17

18 19

CONFLICT OF INTEREST The authors declare no conflict of interest. 20

ACKNOWLEDGEMENTS This work was supported by grants from Specialized Research Fund for the Doctoral Program of Higher Education (No. 20092307110018); The Key Program of the Heilongjiang Provincial Science and Technology Committee (No. GC10C302); Programs Foundation of the Department of Education of Heilongjiang Province (No. 12511330); Programs Foundation of the Affiliated Tumor Hospital of Harbin Medical University (No. JJZ2011-06); Programs Foundation of the First Affiliated Hospital of Harbin Medical University (No. 2013B19); Programs Foundation of the Department of Health Heilongjiang Province (No. 2012661); and Programs Foundation of Natural Science of the Heilongjiang (No. D201272).We thank Yong Wang (Esaote Shenzhen Medical Equipment) for the help in the contrast imaging of the Matrigel plug models.

21

22

23

24

25

REFERENCES 1 Quesada AR, Mun˜oz-Cha´puli R, Medina MA. Anti-angiogenic drugs: from bench to clinical trials. Med Res Rev 2006; 26: 483–530. 2 Bodempudi V, Ohlfest JR, Terai K, Zamora EA, Vogel RI, Gupta K et al. Blood outgrowth endothelial cell-based systemic delivery of antiangiogenic gene therapy for solid tumors. Cancer Gene Ther 2010; 17: 855–863. 3 Kim KS, Kim DS, Chung KH, Park YS. Inhibition of angiogenesis and tumor progression by hydrodynamic cotransfection of angiostatin K1-3, endostatin, and saxatilin genes. Cancer Gene Ther 2006; 13: 563–571. 4 Yokoyama Y, Dhanabal M, Griffioen AW, Sukhatme VP, Ramakrishnan S. Synergy between angiostatin and endostatin: inhibition of ovarian cancer growth. Cancer Res 2000; 60: 2190–2196. 5 Cao R, Wu HL, Veitonma¨ki N, Linden P, Farnebo J, Shi GY et al. Suppression of angiogenesis and tumor growth by the inhibitor K1-5 generated by plasminmediated proteolysis. Proc Natl Acad Sci USA 1999; 96: 5728–5733. 6 Cao Y. Endogenous angiogenesis inhibitors and their therapeutic implications. Int J Biochem Cell Biol 2001; 33: 357–369. 7 Kode JA, Mukherjee S, Joglekar MV, Hardikar AA. Mesenchymal stem cells: immunobiology and role in immunomodulation and tissue regeneration. Cytotherapy 2009; 11: 377–391. 8 Farquhar C, Marjoribanks J, Basser R, Hetrick S, Lethaby A. High dose chemotherapy and autologous bone marrow or stem cell transplantation versus conventional chemotherapy for women with metastatic breast cancer. Cochrane Database Syst Rev 2005; 20: CD003142. 9 Spitzer TR, Ambinder RF, Lee JY, Kaplan LD, Wachsman W, Straus DJ et al. Dosereduced busulfan, cyclophosphamide, and autologous stem cell transplantation

& 2014 Nature America, Inc.

26

27

28

29

30

31

32

for human immunodeficiency virus-associated lymphoma: AIDS Malignancy Consortium study 020. Biol Blood Marrow Transplant 2008; 14: 59–66. Volpers C, Kochanek S. Adenoviral vectors for gene transfer and therapy. J Gene Med 2004; 6(Suppl(1)): S164–S171. Sheets RL, Stein J, Bailer RT, Koup RA, Andrews C, Nason M et al. Biodistribution and toxicological safety of adenovirus type 5 and type 35 vectored vaccines against human immunodeficiency virus-1 (HIV-1), Ebola, or Marburg are similar despite differing adenovirus serotype vector, manufacturer’s construct, or gene inserts. J Immunotoxicol 2008; 5: 315–335. Bergelson JM, Cunningham JA, Droguett G, Kurt-Jones EA, Krithivas A, Hong JS et al. Isolation of a common receptor for Coxsackie B viruses and adenoviruses 2 and 5. Science 1997; 275: 1320–1323. Kawabata K, Sakurai F, Koizumi N, Hayakawa T, Mizuguchi H. Adenovirus vectormediated gene transfer into stem cells. Mol Pharm 2006; 3: 95–103. Olmsted-Davis EA, Gugala Z, Gannon FH, Yotnda P, McAlhany RE, Lindsey RW et al. Use of a chimeric adenovirus vector enhances BMP2 production and bone formation. Hum Gene Ther 2002; 13: 1337–1347. Sakurai F, Kawabata K, Mizuguchi H. Adenovirus vectors composed of subgroup B adenoviruses. Curr Gene Ther 2007; 7: 229–238. Yu L, Shimozato O, Li Q, Kawamura K, Ma G, Namba M et al. Adenovirus type 5 substituted with type 11 or 35 fiber structure increases its infectivity to human cells enabling dual gene transfer in CD46-dependent and -independent manners. Anticancer Res 2007; 27: 2311–2316. Segerman A, Atkinson JP, Marttila M, Dennerquist V, Wadell G, Arnberg N. Adenovirus type 11 uses CD46 as a cellular receptor. J Virol 2003; 77: 9183–9191. Wang H, Li ZY, Liu Y, Persson J, Beyer I, Mo¨ller T et al. Desmoglein 2 is a receptor for adenovirus serotypes3, 7, 11 and 14. Nat Med 2011; 17: 96–104. Knaa¨n-Shanzer S, van de Watering MJ, van der Velde I, Gonc¸alves MA, Valerio D, de Vries AA. Endowing human adenovirus serotype 5 vectors with fiber domains of species B greatly enhances gene transfer into human mesenchymal stem cells. Stem Cells 2005; 23: 1598–1607. Liu HX, Chu YZ, Lou G. Fiber-modified adenovirus can mediate human adipose tissue-derived mesenchymal stem cell-based anti-angiogenic gene therapy. Biotechnol Lett 2010; 32: 1181–1188. In ’t Anker PS, Scherjon SA, Kleijburg-van der Keur C, de Groot-Swings GM, Claas FH, Fibbe WE et al. Isolation of mesenchymal stem cells of fetal or maternal origin from human placenta. Stem Cells 2004; 22: 1338–1345. Brooke G, Rossetti T, Pelekanos R, Ilic N, Murray P, Hancock S et al. Manufacturing of human placenta-derived mesenchymal stem cells for clinical trials. Br J Haematol 2009; 144: 571–579. Barlow S, Brooke G, Chatterjee K, Price G, Pelekanos R, Rossetti T et al. Comparison of human placenta- and bone marrow-derived multipotent mesenchymal stem cells. Stem Cells Dev 2008; 17: 1095–1107. Sabapathy V, Ravi S, Srivastava V, Srivastava A, Kumar S. Long-term cultured human term placenta-derived mesenchymal stem cells of maternal origin displays plasticity. Stem Cells Int 2012; 2012: 174328. Tuve S, Wang H, Ware C, Liu Y, Gaggar A, Bernt K et al. A new group B adenovirus receptor is expressed at high levels on human stem and tumor cells. J Virol 2006; 80: 12109–12120. Conget PA, Minguell JJ. Adenoviral-mediated gene transfer into ex vivo expanded human bone marrow mesenchymal progenitor cells. Exp Hematol 2000; 28: 382–390. Dayoub H, Dumont RJ, Li JZ, Dumont AS, Hankins GR, Kallmes DF et al. Human mesenchymal stem cells transduced with recombinant bone morphogenetic protein-9 adenovirus promote osteogenesis in rodents. Tissue Eng 2003; 9: 347–356. Knaa¨n-Shanzer S, Van Der Velde I, Havenga MJ, Lemckert AA, De Vries AA, Valerio D. Highly efficient targeted transduction of undifferentiated human hematopoietic cells by adenoviral vectors displaying fiber knobs of subgroup B. Hum Gene Ther 2001; 12: 1989–2005. Knaa¨n-Shanzer S, van de Watering MJ, van der Velde I, Gonc¸alves MA, Valerio D, de Vries AA. Endowing human adenovirus serotype 5 vectors with fiber domains of species B greatly enhances gene transfer into human mesenchymal stem cells. Stem Cells 2005; 23: 1598–1607. Mizuguchi H, Sasaki T, Kawabata K, Sakurai F, Hayakawa T. Fiber-modified adenovirus vectors mediate efficient gene transfer into undifferentiated and adipogenic-differentiated human mesenchymal stem cells. Biochem Biophys Res Commun 2005; 332: 1101–1106. Jendelova´ P, Herynek V, Urdzı´kova´ L, Glogarova´ K, Kroupova´ J, Andersson B et al. Magnetic resonance tracking of transplanted bone marrow and embryonic stem cells labeled by iron oxide nanoparticles in rat brain and spinal cord. J Neurosci Res 2004; 76: 232–243. Stojanov K, de Vries EF, Hoekstra D, van Waarde A, Dierckx RA, Zuhorn IS. [18F]FDG labeling of neural stem cells for in vivo cell tracking with positron

Cancer Gene Therapy (2014), 200 – 208

Human placenta mesenchymal stem cells Yu Chu et al

208 33 34

35 36 37 38

emission tomography: inhibition of tracer release by phloretin. Mol Imaging 2012; 11: 11–12. Fu Y, Azene N, Xu Y, Kraitchman DL. Tracking stem cells for cardiovascular applications in vivo: focus on imaging techniques. Imaging Med 2011; 3: 473–486. Gildehaus FJ, Haasters F, Drosse I, Wagner E, Zach C, Mutschler W et al. Impact of indium-111 oxine labelling on viability of human mesenchymal stem cells in vitro, and 3D cell-tracking using SPECT/CT in vivo. Mol Imaging Biol 2011; 13: 1204–1214. Leong-Poi H. Molecular imaging using contrast-enhanced ultrasound: evaluation of angiogenesis and cell therapy. Cardiovasc Res 2009; 84: 190–200. Zhang SJ, Wu JC. Comparison of imaging techniques for tracking cardiac stem cell therapy. J Nucl Med 2007; 48: 1916–1919. Lin Y, Molter J, Lee Z, Gerson SL. Bioluminescence imaging of hematopoietic stem cell repopulation in murine models. Methods Mol Biol 2008; 430: 295–306. Wu JC, Cao F, Dutta S, Xie X, Kim E, Chungfat N et al. Proteomic analysis of reporter genes for molecular imaging of transplanted embryonic stem cells. Proteomics 2006; 6: 6234–6249.

Cancer Gene Therapy (2014), 200 – 208

39 Masson V VE, Devy L, Grignet-Debrus C, Bernt S, Bajou K, Blacher S et al. Mouse aortic ring assay: a new approach of the molecular genetics of angiogenesis. Biol Proced Online 2002; 4: 24–31. 40 Baker M, Robinson SD, Lechertier T, Barber PR, Tavora B, D’Amico G et al. Use of the mouse aortic ring assay to study angiogenesis. Nat Protoc 2011; 7: 89–104. 41 Cattaneo MG, Pola S, Deho` V, Sanguini AM, Vicentini LM. Alprostadil suppresses angiogenesis in vitro and in vivo in the murine Matrigel plug assay. Br J Pharmacol 2003; 138: 377–385. 42 Hernot S, Klibanov AL. Microbubbles in ultrasound-triggered drug and gene delivery. Adv Drug Deliv Rev 2008; 60: 1153–1166. 43 Kragh M, Hjarnaa PJ, Bramm E, Kristjansen PE, Rygaard J, Binderup L. In vivo chamber angiogenesis assay: an optimized Matrigel plug assay for fast assessment of anti-angiogenic activity. Int J Oncol 2003; 22: 305–311. 44 Furlow B. Contrast-enhanced ultrasound. Radiol Technol 2009; 80: 547S–561S. 45 Leong-Poi H. Molecular imaging using contrast-enhanced ultrasound: evaluation of angiogenesis and cell therapy. Cardiovasc Res 2009; 84: 190–200.

& 2014 Nature America, Inc.