Factory neovessels: engineered human blood vessels secreting therapeutic proteins as a new drug delivery system


Several works have shown the feasibility of engineering functional blood vessels in vivo using human endothelial cells (ECs). Going further, we explored the therapeutic potential of neovessels after gene-modifying the ECs for the secretion of a therapeutic protein. Given that these vessels are connected with the host vascular bed, we hypothesized that systemic release of the expressed protein is immediate. As a proof of principle, we used primary human ECs transduced with a lentiviral vector for the expression of a recombinant bispecific αCEA/αCD3 antibody. These ECs, along with mesenchymal stem cells as a source of mural cells, were embedded in Matrigel and subcutaneously implanted in nude mice. High antibody levels were detected in plasma for 1 month. Furthermore, the antibody exerted a therapeutic effect in mice bearing distant carcinoembryonic-antigen (CEA)-positive tumors after inoculation of human T cells. In summary, we show for the first time the therapeutic effect of a protein locally secreted by engineered human neovessels.


Recent studies showed that a network of human mature blood vessels can be formed by co-implantation of primary human endothelial cells (EC) and human mesenchymal stem cells (MSCs) in immunodeficient mice.1, 2, 3, 4, 5 The vessel walls engineered under these conditions were found to be substantially similar to native vessels at both molecular and cellular levels.2, 3, 4, 5 These findings have generated considerable interest in the potential application of engineered blood vessels in regenerative medicine and cell-based revascularization therapies. Furthermore, the possibility of establishing a functional vascular network in a surgically accessible location provides an unprecedented opportunity for therapeutic intervention that goes far beyond tissue engineering.

In this context, the genetic modification of ECs would ensure the secretion of a therapeutic protein into the systemic circulation for a prolonged period of time. We have previously reported that vasculature generated from lentivirally transduced human umbilical vein endothelial cells (HUVECs) expressing firefly luciferase (HUVECLuc) and co-implanted with human bone marrow MSC can be assessed quantitatively by in vivo whole-body bioluminescence imaging (BLI) for more than 120 days.2 Luciferase activity correlated with the formation of a network of mature, functional blood vessels of human nature inside the implant, as assessed by human CD34 and anti-smooth muscle actin (anti-smooth muscle actin) staining.2 The presence of MSC is essential for sustained luciferase activity, suggesting a key role of MSC in regulating vessel maturation and functionality, providing efficient recruitment and investment of mural cells to the engineered vessels. The presence of erythrocytes within the lumen of these neovessels was highly indicative of a functional vasculature, integrated into the recipient's vascular bed. To further assess the connection to the mouse circulatory system, we intravenously (i.v.) injected TRITC-conjugated lectin UEA-1 and showed binding to neovessels lined by human ECs inside the implants.2

We have previously shown the feasibility of in vivo production and systemic delivery of therapeutic antibodies by gene-modified human cells.6, 7, 8 Bispecific antibodies are non-natural immunoglobulin-based molecules that contain two distinct binding specificities.9 Bispecific antibodies redirect the cytolytic activity of a variety of immune effector cells toward tumor cells by binding to cell activation molecules with one domain and to specific tumor-associated antigen on cancer cells with the other. T-cell-activating bispecific antibodies have shown great potential for the treatment of malignant diseases in both animal models7, 8, 11 and humans.12 Recently, it was reported that systemically administered recombinant anti-CD19 × anti-CD3 antibody induced partial and complete tumor regressions in patients with non-Hodgkin’s lymphoma.12 In previous works we have shown the antitumoral effect of a small recombinant bispecific antibody (diabody, dAb)13, 14 directed against the human carcinoembryonic antigen (CEA, CD66) and the T-cell co-receptor CD3 (anti-CEA × anti-CD3), secreted by different human cells, including T lymphocytes7, 8 and MSCs.10 However, activated T lymphocytes possess a short life span, and this implies an obvious limitation to their application in a gene therapy strategy. On the other hand, MSCs are endowed with a greater expansion capacity but their proangiogenic and immunomodulatory properties may raise concern over its potential use in cancer patients.15

The stability of tissue-engineered human blood vessel, the permissiveness of HUVEC to be transduced by lentiviral vectors and the comparative tolerance of its use encouraged us to use engineered blood vessels as therapeutic protein factories. Given that these neovessels connect to the host vascular bed, we hypothesized that the therapeutic protein would be directly released into the bloodstream and exert a systemic effect.


Transduction of human endothelial cells with a lentiviral vector encoding a bispecific αCEA/αCD3 diabody

To validate the concept of human neovessels as cell factories and its therapeutic potential, we used a small anti-CEA × anti-CD3 diabody (αCEA/αCD3).11 Early passage HUVECs were transduced at a vector multiplicity of infection of 10, with lentivirus encoding the αCEA/αCD3 diabody and enhanced green fluorescent protein (EGFP) (LentidAb), or with a lentivirus harboring a luciferase-IRES-EGFP cassette (LentiLuc).7, 8 More than 90% of transduced HUVEC (HUVECdAb and HUVECLuc) expressed EGFP for at least 30 days in vitro (Figure 1a) and showed normal morphology, phenotype and function (data not shown). Interestingly, the secretion of functional αCEA/αCD3 diabody molecules remained stable during this period with levels around 100 ng ml−1 per 105 cells per 72 h at day 30 (Figure 1b). A suspension of HUVECLuc cells was serially diluted in a 96-well plate and was imaged with BLI.2 The light emitted from HUVECLuc in the presence of the substrate luciferin was approximately 10 photons per second per cell (data not shown).

Figure 1

In vitro transgene expression stability in human umbilical vein endothelial cells (HUVECs) after infection with dAb.EGFP-encoding lentivirus (HUVECdAb). (a) Expression of enhanced green fluorescent protein (EGFP) and (b) secretion of αCEA/αCD3 diabody (ng/ml per 105 cells per 72 h).

Effect of engineered blood vessels on tumor growth

Mixtures (ratio 4:1) of either HUVECdAb and MSC or HUVECLuc and MSC were embedded in Matrigel and inoculated subcutaneously (s.c.; vascular implants) in the ventral area of nude mice. Vascular implants containing HUVECdAb and MSC (‘factory’ neovessels) were inoculated in the left flank and vascular implants containing HUVECLuc and MSC (control neovessels) were inoculated contralaterally for BLI monitoring of cell engraftment and persistence. As shown in Figure 2a, control vascular implants showed stable luciferase activity throughout the study period indicating a good connection to the mouse circulatory system. Functional αCEA/αCD3 diabodies were detected in the plasma of factory neovessels-bearing mice but not in samples from mice only bearing control neovessels. Following an initial peak of expression of about 90 ng ml−1, αCEA/αCD3 diabody plasma levels persisted around 50 ng ml−1 at least up to 7 weeks (Figure 2b). Interestingly, diabody levels expressed by HUVEC are much more stable than those previously reported with MSC,10 which dropped more than 80% in the same lapse of time.

Figure 2

Inhibition of tumor growth by angio-secreted αCEA/αCD3 diabodies. The persistence of viable functional human umbilical vein endothelial cells (HUVECs) infected with luciferase-encoding lentivirus (HUVECLuc) was assessed by bioluminescence imaging (BLI) at the indicated times after implantation of HUVECLuc/MSC (1) and HUVECdAb/MSC (2) vascular implants (a). BLI of a representative mouse (total, n=4 mice per group). (b) Plasma concentration of αCEA/αCD3 diabody was determined by enzyme-linked immunosorbent assay (ELISA). HCT-116 human colon carcinoma cells (2 × 106) were injected s.c. into the dorsal skin of nude mice (n=4 mice per group). Five days after tumor implantation, either HUVECLuc/MSC vascular implants (control neovessels) or HUVECLuc/MSC and HUVECdAb/MSC vascular implants (factory neovessels) were inoculated s.c. in the ventral area. After 2 days animals received an i.v. injection of 2 × 106 human peripheral blood lymphocytes (PBLs). The mean tumor volume±s.d. (b) and the Kaplan-Meier survival curves (c). Mice from group B showed significant tumor growth inhibition (P=0.035) and survived significantly longer (P=0.0481) than control mice.

To determine the in vivo antitumor activity of the angio-secreted diabodies, we used a human colon carcinoma xenograft model by injecting CEA+ HCT-116 tumor cells s.c. into the dorsal space of nude mice on day 0. The animals then were divided into three cohorts (A, B and C) and 5 days after tumor inoculation vascular implants were injected s.c. in the ventral area. Group A received one vascular implant (control neovessels) inoculated s.c. in the left ventral area; groups B and C received two vascular implants (control neovessels and factory neovessels) injected s.c. in opposite flanks in the ventral area. The experiment was performed three times with 4–5 animals per group (n=12/15). In two experiments (nos. 1 and 3) mice from groups A and B received a single i.v injection of 2 × 106 human T cells through the tail vein on day 7 (Figure1; Supplementary Figure 1). In an attempt to improve the therapeutic effect, mice from groups A and B received two i.v. injections (days 10 and 25) of 2 × 106 human T cells (Figure 3; experiment 2). Mice from group C did not receive T cells.

Figure 3

The mean tumor volume±s.d. (a) and the Kaplan–Meier survival curves (b) of nude mice bearing human colon carcinoma xenografts after implantation of control or factory neovessels. HCT-116 cells (2 × 106) were injected s.c. into the dorsal skin of nude mice (n=5 mice per group). At 5 days after cell inoculation, either control neovessels (group A) or factory neovessels (groups B and C) were implanted s.c. in the ventral area. At days 10 and 25 mice from groups A and B received an i.v. injection of 2 × 106 human peripheral blood lymphocytes (PBLs). Mice from group C received phosphate-buffered saline (PBS) instead of lymphocytes. Mice from group B showed significant tumor growth inhibition (P=0.02) and survived significantly longer (P=0.0333) than control mice.

The group of mice with αCEA/αCD3 diabody-secreting neovessels (group B) showed consistently slower tumor growth rate and statistically significant difference in tumor volume compared with the group of mice bearing control neovessels (group A) in three independent experiments (Figures 2b and 3; Supplementary Figure 1). Moreover, in the three experiments the survival times between factory neovessels-bearing and control mice differed significantly (P<0.05). Overall, mice with diabody-secreting neovessels survived an average of 23 days longer than control mice. No therapeutic effect of angio-secreted αCEA/αCD3 diabodies was observed in the absence of human T cells (group C) (Figure3; data not shown).

Tumor homing of human T cells

To show tumor homing of human T cells, we retrieved tumors 2 days after T-cell injection and analyzed immunohistochemically. T cells could be clearly identified with a mouse mAb anti-human CD3 (Supplementary Figure 2). Previously, circulating T human lymphocytes with the original phenotype had been detected in peripheral blood by fluorescence-activated cell sorting 24 h after administration (Figures 4a and b). Human T cells localized to tumors, as further assessed by the detection of human CD3 transcripts by real-time–PCR. Percentage of CD3 RNA with respect to total RNA in tumors peaked at day 2 after inoculation and subsequently decreased, being still detectable at day 15 (Figure 4c). Most remarkably, expression of CD3 RNA was far more abundant in mice bearing factory neovessels, indicating that T-cell infiltration was higher than in the control group (Figure 4c).

Figure 4

(a) Flow cytometry analysis of in vitro expanded human T-cell phenotype before inoculation. (b) Flow cytometry detection of circulating human T cells 24 h after i.v. injection. (c) Percentage of human CD3ɛ RNA transcripts with respect to total RNA from tumors of mice receiving human T cells and bearing control or factory neovessels.


Here, we show for the first time the capability of engineered human blood vessels to secrete a recombinant bispecific antibody. This strategy allows systemic release of sustainedly high antibody levels, and induces tumor growth inhibition in vivo after i.v. injection of human T lymphocytes. T-cell-activating bispecific antibodies are able to target polyclonal T-cell populations on tumor cells and are considered powerful reagent for cancer immunotherapy. Indeed, several groups have used bispecific antibodies successfully in both animal models16, 17 and humans.12 Recently, it was reported that systemically administered recombinant αCD19/αCD3 antibodies induced partial and complete responses in patients with refractory non-Hodgkin's lymphoma.12 However, practical use of systemically administered recombinant antibody fragments is limited, by problems related to large-scale production and biodistribution.9, 17

In situ secretion of therapeutic antibodies is an attractive alternative to systemic administration of antibody fragments. In situ expression potentially circumvents problems of tumor penetration and compensates for the rapid blood clearance of antibody fragments and could make the antibody less immunogenic and better tolerated.9, 17 The use of gene therapy methods offers additional benefits by achieving sustained and effective concentrations of therapeutic antibodies directly at points of target intervention. In fact, we have previously shown that human T lymphocytes, transduced ex vivo to secrete the αCEA/αCD3 diabody in autocrine manner, inhibited the growth of CEA-positive tumors.8 However, the short life span of activated human T lymphocytes and their inefficient transduction constitute important drawbacks for their application in cancer therapy.

Another strategy to deliver therapeutic genes to tumors is to exploit the tumor-targeting capabilities of certain cells, as bone-marrow-derived stem cells, that after systemic administration are supposed to migrate to and infiltrate primary and metastatic solid tumors.18 However, incorporation into the tumor microenvironment is not always apparent. What is more, the role of recruited MSC in the tumor microenvironment remains unclear, given that recent evidence also suggests that they could have a role in tumor growth and metastasis.10, 19 This implies a potential risk in the use of MSCs in cell-based anticancer therapies,15 especially when MSCs are inoculated in the vicinity of tumor cells or allowed to migrate toward them.

In fact, for strategies where cells are used as therapeutic factories, their ability to disseminate throughout the body is not required, or is even undesirable. Producer cells can be confined to scaffolds that keep cells at the implantation site, although the therapeutic protein may act at distance if secreted into circulation.20 Recent studies established the feasibility of using gene-modified MSCs embedded in matrix scaffolds, as cell factories for the production of therapeutic protein.10, 21 According to our histological study data MSCs in factory ‘plugs’ do not organize or differentiate into recognizable structures (unpublished results). Conversely, the co-implantation of HUVECs and MSCs in Matrigel produces mature blood vessels that are connected with the host vascular bed,2 allowing immediate systemic release of the expressed protein. This could explain why plasma concentration of αCEA/αCD3 diabody in mice bearing factory neovessels dropped less than 30% with respect to the peak of expression. Interestingly, plasma levels of diabody produced by in mice implanted with MSCs in the same lapse of time dropped more than 80% with respect to the peak of expression.10 Even after a single administration of human T cells to factory neovessels-bearing mice, the in vivo secretion of diabody is correlated with a significant therapeutic effect. We could hypothesize that this effect should be significantly improved in a human host, with a continuous supply of human T cells.

The ‘factory neovessel’ approach offers the added advantage of gene-modifying terminally differentiated EC that will remain quiescent after blood vessel formation, avoiding the potential risks of stem cells. MSCs in this model are not only the source of mural cells, but also show immunomodulatory properties that may protect producer ECs from immune responses. Subcutaneous delivery of ECs would provide an easily accessible vascular implant that could be retrieved once the therapeutic effect is fulfilled, increasing tolerance and facilitating the application in a clinical setting.

Factory neovessels secreting other types of therapeutic proteins could easily be engineered encompassing a wide range of applications (inherited or acquired diseases). The neovessels formed in this way can serve as a continuous drug delivery system to replace periodic administration of the needed protein. In a different setting, VEGF-transduced ECs could enhance angiogenesis in revascularization therapies.

Materials and methods

Cells and culture conditions

Human umbilical vein endothelial cells, isolated from human umbilical cord veins by collagenase treatment, were kindly provided by M Feijóo (Hospital Universitario de la Princesa, Madrid, Spain) and cultured in EGM-2 medium containing 2% fetal bovine serum and EC growth supplements (Cambrex, Baltimore, MD, USA). Human MSCs were obtained from bone marrow samples from healthy donors after informed consent. MSCs were purified, expanded and characterized as previously described.2, 10 MSCs were used for experiments between the third and fourth passages. Human peripheral blood lymphocytes were isolated from healthy donors peripheral blood by density-gradient centrifugation. Cells were expanded and characterized by flow cytometry as described.8 Human colon carcinoma cell line HCT-116 (CCL-247; ATCC, Rockville, MD, USA) was grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum.

Lentiviral vectors and cell transduction

The vector pRRL.dAb. EGFP promotes the expression of a bispecific αCEA/αCD3 diabody and EGFP8 and the vector pRRL-Luc-EGFP promotes the expression of firefly luciferase and EGFP.2 Lentiviral particles were produced by co-transfection of 293T cells through calcium phosphate precipitation method. HUVEC passage 1 were seeded at a density of 7.5 × 103 cells per cm2, allowed to adhere and infected overnight with the lentivirus dAb-EGFP (HUVECdAb) or Luc-EGFP (HUVECLuc) at a final multiplicity of infection of 10. EGFP transgene expression was monitored by flow cytometry. Conditioned medium from transduced HUVECs was analyzed for αCEA/áCD3 diabody secretion by enzyme-linked immunosorbent assay as described.8 Transgene expression in vitro was monitored for more than 50 days. Transduced HUVECs were used in vivo in passages 3–5.


Five-week-old female athymic nude mice (Hsd: athymic Nude/Nude; Harlan Iberica, Barcelona, Spain) were used. All animals were maintained in a sterile environment. Cages, bedding, food and water were autoclaved. Anesthesia was induced by intraperitoneal (i.p.) injection of a combination of ketamine (Ketolar; Pfizer, New York, NY, USA) and diazepam (Valium; Roche, Nutley, NJ, USA). All mice were handled in accordance with the guidelines of institutional animal care and use committee and the experiment was performed in accordance with Spanish legislation.

Vascular implants

A mixture of either HUVECLuc or HUVECdAb (3 × 105) with MSCs (7.5 × 104) was suspended in 200 μl Matrigel (BD Biosciences, Erembodegen, Belgium) and inoculated s.c. in opposite flanks in the ventral area of nude mice, as described.2 All batches of Matrigel were adjusted to 8 mg ml−1 by addition of phosphate-buffered saline and supplemented with heparin (128 U ml−1 final concentration; Sigma-Aldrich, St Louis, MO, USA), human VEGF (50 ng ml−1) and human bFGF (150 ng ml−1) (both from PeproTech, London, UK). Mice were bled by retro-orbital puncture and αCEA/αCD3 diabody plasma concentration was determined by enzyme-linked immunosorbent assay.8

In vivo effect of vascular implants

HCT-116 tumor cells (2 × 106) were inoculated s.c. into the dorsal space of nude mice. At 5 days after tumor implantation, mice received HUVECLuc implant and HUVECdAb implant injected s.c. in opposite flanks in the ventral area. Control group received only the HUVECLuc implant. After 2 days, mice received i.v. injection of human T lymphocytes (2 × 106). Tumor volumes were determined at various time points using the formula: width2 × length × 0.52. In a follow-up study mice received two i.v. injections (days 10 and 25) of human T lymphocytes (2 × 106). Some animals were killed at different time points to analyze the systemic distribution and tumor localization of human T lymphocytes.

In vivo bioluminescence imaging

Mice were imaged using the high-resolution charge-coupled device cooled digital camera ORCA-2BT (Hamamatsu Photonics France, Massy, France) and Wasabi software (Hamamatsu Photonics). Animals were injected i.p. with 125 mg kg−1 (150–200 μl) D-luciferin (Promega, Madison, WI, USA) 10 min before imaging. BLI was collected with 1 min integration time and pseudocolor representations of light intensity were superimposed over the gray scale reference image acquired at low light (exposure time 20 ms). An average of six kinetic BLI acquisitions were collected after substrate injection to confirm a peak of photon emission. For quantification of the detected light, regions of interest were drawn and the light emitted from each region was assessed by recording the total number of photons per second (total flux) after background subtraction.

Flow cytometric identification of circulating human T lymphocytes in recipient mice

Blood samples (100 μl) were obtained by retro-orbital puncture before and after (days 1 and 7) i.v. injection of in vitro expanded human T lymphocytes. Blood samples were collected aseptically using EDTA as anticoagulant. Circulating human T cells were identified using anti-human CD45-FITC (clone B3821F4A), CD4-RD1 (clone SFC112T4D11), CD8-ECD (clone SFCI21Thy2D3) and CD3-PC5 (clone UCHT1) mAbs (Beckman Coulter Inc., Galway, Ireland). Erythrocyte lysis procedure was performed using the TQ-Prep workstation (Coulter Electronics, Hialeah, FL, USA), and samples were analyzed using an EPICS XL flow cytometer (Coulter Electronics).

RNA isolation

At different time points after i.v. injection of human T cells, animals were killed. Blood, tumor and organs (lung, liver and spleen) were removed and processed for total RNA isolation. Approximately 1 ml EDTA blood was mixed with 5 ml of ACK lysis buffer (BioWhittaker, Lonza, Walkersville, MD USA) and incubated for 10 min at room temperature. After centrifugation the buffer was removed and the cell pellet was resuspended in 350 μl RLT buffer (Qiagen, Basel, Switzerland), total RNA was isolated with the RNeasy Plus Micro Kit (Qiagen), according to the manufacturer's instructions. Total RNA isolation of tumor and organs was performed using the RNeasy Plus Mini Kit (Qiagen). For disruption and homogenization of samples, MagNA Lyser Green Beads (Roche Diagnostics, IN, USA) were used according to the manufacturer's protocol.

Reverse transcription, primers and real-time polymerase chain reaction

cDNA was synthesized using 0.5 μg of total RNA from blood, tumor and organs, by random primed reverse transcription with First Strand cDNA Synthesis kit for RT–PCR (Roche Diagnostics) according to the recommended protocol. Real-time PCR (RT-PCR) was performed in a LightCycler 480 apparatus (Roche Diagnostics) using the LightCycler FastStartPLUS DNA Master SYBR Green I kit (Roche Diagnostics). mRNA expression in each sample was measured as a ratio against the geometric average of the reference housekeeping human gene succinate dehydrogenase complex subunit A (hSDHA). The relative concentrations of the target and the reference genes were calculated by interpolation using a standard curve of each gene plotted from the same serial dilution of cDNA. For human CD3ɛ (hCD3) detection in tumors, a standard curve was generated mixing, at different percentages, RNA from in vitro expanded human T cells and tumor RNA from mice that did not receive human T cells. The quantitative mRNA analysis was performed in duplicate. PCR primers for hCD3, hSDHA and murine SDHA (mSDHA) were designed with Primer Express software (Applied Biosystems, Foster City, CA, USA) and were placed in two exons to eliminate amplification for genomic DNA. The following primers (5′–3′) were used: hCD3, forward (F)-IndexTermCTACCCCAGAGGAAGCAAACC and reverse primer (R)-IndexTermgacatcacatccatctccatgc; hSDHA (F)-IndexTermTGGGAACAAGAGGGCATCTG and (R)-IndexTermCCACCACTGCATCAAATTCATG and mSDHA (F)-IndexTermGCAGTTTCGAGGCTTCTTCG and (R)-IndexTermAAGTGAAAGCCGCAGGTCTG. The resulting amplicons had a size of approximately 100 pb.

Histology and immunohistochemical study of tumors

Tumors were retrieved 2, 7 and 15 days after i.v. injection of human T lymphocytes, and at the end of study were formalin-fixed and paraffin-embedded. Sections were stained with hematoxylin and eosin according to standard protocols. Human T lymphocytes were visualized with mouse anti-human CD3ɛ mAb, clone F7.2.38 (Dako, Glostrup, Denmark) using the Mouse on Mouse (M.O.M) peroxidase kit (Vector, Burlingame, CA, USA).

Statistical analysis

The SPSS version 14. 0 software was used for statistical analysis. Repeated-measures analysis of variance model was used to compute the statistical significance of differences between groups. All P-values were two sided and values of 0.05 or less were considered to indicate statistical significance.


  1. 1

    Koike N, Fukumura D, Gralla O, Au P, Schechner JS, Jain RK . Tissue engineering: creation of long-lasting blood vessels. Nature 2004; 428: 138–139.

  2. 2

    Sanz L, Santos-Valle P, Alonso-Camino V, Salas C, Serrano A, Vicario JL et al. Long-term in vivo imaging of human angiogenesis: critical role of bone marrow-derived mesenchymal stem cells for the generation of durable blood vessels. Microvasc Res 2008; 75: 308–314.

  3. 3

    Au P, Tam J, Fukumura D, Jain RK . Bone marrow-derived mesenchymal stem cells facilitate engineering of long-lasting functional vasculature. Blood 2008; 111: 4551–4558.

  4. 4

    Melero-Martin JM, Khan ZA, Picard A, Wu X, Paruchuri S, Bischoff J . In vivo vasculogenic potential of human blood-derived endothelial progenitor cells. Blood 2007; 109: 4761–4768.

  5. 5

    Melero-Martin JM, De Obaldia ME, Kang SY, Khan ZA, Yuan L, Oettgen P et al. Engineering robust and functional vascular networks in vivo with human adult and cord blood-derived progenitor cells. Circ Res 2008; 103: 128–130.

  6. 6

    Sanz L, Kristensen P, Blanco B, Faceteau S, Russell SJ, Winter G et al. Single-chain antibody-based gene therapy: inhibition of tumor growth by in situ production of phage-derived human antibody fragments blocking functionally active sites of cell-associated matrices. Gene Therapy 2002; 15: 1049–1053.

  7. 7

    Blanco B, Holliger P, Vile RG, lvarez-Vallina L . Induction of human T lymphocyte cytotoxicity and inhibition of tumor growth by tumor-specific diabody-based molecules secreted from gene-modified bystander cells. J Immunol 2003; 171: 1070–1077.

  8. 8

    Compte M, Blanco B, Serrano F, Cuesta AM, Sanz L, Bernad A et al. Inhibition of tumor growth in vivo by in situ secretion of bispecific anti-CEA x anti-CD3 diabodies from lentivirally transduced human lymphocytes. Cancer Gene Ther 2007; 14: 380–388.

  9. 9

    Holliger P, Hudson PJ . Engineered antibody fragments and the rise of single domains. Nat Biotechnol 2005; 23: 1126–1136.

  10. 10

    Compte M, Cuesta AM, Sánchez-Martín D, Alonso-Camino V, Vicario JL, Sanz L et al. Tumor immunotherapy using gene-modified human mesenchymal stem cells loaded into synthetic extracellular matrix scaffolds. Stem Cells 2009; 27: 753–760.

  11. 11

    Holliger P, Manzke O, Span M, Hawkins RE, Fleischmann B, Qinghua L et al. Carcinoembryonic antigen (CEA)-specific T-cell activation in colon carcinoma induced by anti-CD3 × anti-CEA bispecific diabodies and B7 × anti-CEA bispecific fusion proteins. Cancer Res 1999; 59: 2909–2916.

  12. 12

    Bargou R, Leo E, Zugmaier G, Klinger M, Goebeler M, Know S et al. Tumor regression in cancer patients by very low doses of a T cell-engaging antibody. Science 2008; 321: 974–977.

  13. 13

    Holliger P, Prospero T, Winter G . ‘Diabodies:’ small bivalent and bispecific antibody fragments. Proc Natl Acad Sci USA 1993; 90: 6444–6448.

  14. 14

    Perisic O, Webb PA, Holliger P, Winter G, Williams RL . Crystal structure of a diabody, a bivalent antibody fragment. Structure 1994; 2: 1217–1226.

  15. 15

    Lazennec G, Jorgensen C . Concise review: adult multipotent stromal cells and cancer: risk or benefit? Stem Cells 2008; 26: 1387–1394.

  16. 16

    Segal DM, Weiner GJ, Weiner LM . Bispecific antibodies in cancer therapy. Curr Opin Immunol 1999; 11: 558–562.

  17. 17

    Sanz L, Blanco B, Alvarez-Vallina L . Antibodies and gene therapy: teaching old ‘magic bullets’ new tricks. Trends Immunol 2004; 25: 85–91.

  18. 18

    Aboody KS, Najbauer J, Danks MK . Stem and progenitor cell-mediated tumor selective gene therapy. Gene Therapy 2008; 15: 739–752.

  19. 19

    Karnoub AE, Dash AB, Vo AP, Sullivan A, Brooks MW, Bell GW et al. Mesenchymal stem cells within tumour stroma promote breast cancer metastasis. Nature 2007; 49: 557–563.

  20. 20

    Roth JC, Curiel DT, Pereboeva L . Cell vehicle targeting strategies. Gene Therapy 2008; 15: 716–729.

  21. 21

    Eliopoulos N, Francois M, Boivin MN, Martienau D, Galipeau J . Neo-organoid of marrow mesenchymal stromal cells secreting interleukin-12 for breast cancer therapy. Cancer Res 2008; 68: 4810–4818.

Download references


This work was supported by grants from the Ministerio de Ciencia e Innovación (BIO2008-03233), the Comunidad Autónoma de Madrid (S-BIO-0236-2006) and the European Union (Immunonet–SUDOE) to LA-V; and from the Fondo de Investigación Sanitaria (PI061621) to LS. MC was supported by Instituto de Salud Carlos III (Contrato Rio Hortega, CM06/00055). DS-M was supported by a Comunidad Autónoma de Madrid/European Social Fund training grant (FPI-000531). LS is an investigator from the Ramón y Cajal Program (Ministerio de Ciencia e Innovación), co-financed by the European Social Fund. VAC is a predoctoral fellow from the Gobierno Vasco (BFI07.132).

Author information

Correspondence to L Sanz or L Álvarez-Vallina.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Additional information

Supplementary Information accompanies the paper on Gene Therapy website

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Compte, M., Alonso-Camino, V., Santos-Valle, P. et al. Factory neovessels: engineered human blood vessels secreting therapeutic proteins as a new drug delivery system. Gene Ther 17, 745–751 (2010) doi:10.1038/gt.2010.33

Download citation


  • cell-based gene therapy
  • human endothelial cells
  • human mesenchimal stem cells
  • lentivirally transduced ECs
  • blood vessels
  • cancer

Further reading