A Step Towards Making Customized Blood Vessels


Johns Hopkins University scientists have directed stem cells to form networks of new blood vessels, and successfully transplanted those laboratory-made blood vessels into laboratory mice.

The stem cells in this experiment were made by reprogramming ordinary cells. Thus this new technique could potentially be used to make blood vessels that are genetically matched to individual patients and have a very low chance of being rejected by the patient’s immune system.

“In demonstrating the ability to rebuild a microvascular bed in a clinically relevant manner, we have made an important step toward the construction of blood vessels for therapeutic use,” said Sharon Gerecht, associate professor in the Johns Hopkins University Department of Chemical and Biomolecular Engineering, Physical Sciences-Oncology Center and Institute for NanoBioTechnology. “Our findings could yield more effective treatments for patients afflicted with burns, diabetic complications and other conditions in which vasculature function is compromised.”

Gerecht’s research group and others have previously grown blood vessels in the laboratory using stem cells, but there are problems with using these blood vessels in human patients. For example, in a paper published by Gerecht’s group in Stem Cells Translational Medicine earlier this year (Stem Cells Trans Med April 2013 vol. 2 no. 4 297-306), ECFCs or endothelial colony-forming cells from human umbilical cords were grown and used to make networks of blood vessels in culture. Those blood vessels were then embedded in blocks of “hyaluronic acid.” Hyaluronic acid is a component of human connective tissue, and when the ECFCs were embedded into it, they were then placed on the skin of mice that had received third-degree burns. On day 3 following transplantation, white blood cells called macrophages degraded the hyaluronic acid gel rather quickly. Between days 5–7, the mouse’s blood vessels infiltrated the implant and connected with the human blood vessels in the wound area. The growth of the human blood vessels peaked at day 7 and then decreased by the end of the proliferation stage. As the wound reached the remodeling period during the second week after transplantation, the blood vessels, including the transplanted human vessels generally regressed, and a few microvessels, wrapped by mouse smooth muscle cells and with a vessel area less than 200 square micrometers (including the human ones), remained in the healed wound.

This is a fascinating experiment, but making blood vessels this way is a heck of a lot of work, and even though the umbilical cord ECFCs are less likely to be rejected by the immune system of the patient, the chances of immune system rejection are still present. is there a better way?

In this current study, Gerecht and her team tried to streamline the new growth process. Where other experiments used chemical cues to differentiate stem cells into the desired cell type, Sravanti Kusuma in Gerecht’s laboratory devised a way to instruct stem cells to exclusively form the two cell types required for blood vessel construction (smooth muscle cells and endothelial cells).

According to Kusuma, “It makes the process quicker and more robust if you don’t have to sort through a lot of cells you don’t need to find the ones you do, or grow two batches of cells,”

Derivation of EVCs from hPSCs. (A) Schema for self-assembled vascular derivatives. (i) hPSCs are differentiated toward EVCs that can be matured into functional ECs and pericytes. (ii) Derived EVCs are embedded within a synthetic HA matrix that facilitates self-organization into vascular networks. (B) VEcad expression in day 12 differentiated hiPSC-MR31 and hESC-H9 cell lines comparing the three differentiation conditions tested (flow cytometry analysis; n = 3). (C and D) Flow cytometry plots (n = 3) of EVC derivatives assessing the expression of pluripotent markers TRA-1-60 and TRA-1-81 (C) and CD105 and CD146 (D). (E) EVC differentiation efficiency from hPSC lines per 1 million input hPSCs. (F) Flow cytometry plots (n = 3) of EVC derivatives assessing expression of VEcad double-labeled with CD105 or PDGFRβ. (Left) Isotype controls. (G) Quantitative RT-PCR analysis of EC and perivascular marker expression by EVCs and sorted VEcad+ and VEcad− cells. # denotes not detected. Data are normalized to EVCs of each specific hPSC type. (H) Flow cytometry plots (n = 3) of hematopoietic marker CD45 (hiPSC-BC1). (I) Quantitative RT-PCR of H9-EVCs for the expression of SMMHC and peripherin, compared with undifferentiated cells (d0) and mature derivatives (13, 37). Isotype controls for flow cytometry are in gray. Flow cytometry results shown are typical of the independent experiments. *P < 0.05; **P < 0.01; ***P < 0.001
Derivation of EVCs from hPSCs. (A) Schema for self-assembled vascular derivatives. (i) hPSCs are differentiated toward EVCs that can be matured into functional ECs and pericytes. (ii) Derived EVCs are embedded within a synthetic HA matrix that facilitates self-organization into vascular networks. (B) VEcad expression in day 12 differentiated hiPSC-MR31 and hESC-H9 cell lines comparing the three differentiation conditions tested (flow cytometry analysis; n = 3). (C and D) Flow cytometry plots (n = 3) of EVC derivatives assessing the expression of pluripotent markers TRA-1-60 and TRA-1-81 (C) and CD105 and CD146 (D). (E) EVC differentiation efficiency from hPSC lines per 1 million input hPSCs. (F) Flow cytometry plots (n = 3) of EVC derivatives assessing expression of VEcad double-labeled with CD105 or PDGFRβ. (Left) Isotype controls. (G) Quantitative RT-PCR analysis of EC and perivascular marker expression by EVCs and sorted VEcad+ and VEcad− cells. # denotes not detected. Data are normalized to EVCs of each specific hPSC type. (H) Flow cytometry plots (n = 3) of hematopoietic marker CD45 (hiPSC-BC1). (I) Quantitative RT-PCR of H9-EVCs for the expression of SMMHC and peripherin, compared with undifferentiated cells (d0) and mature derivatives (13, 37). Isotype controls for flow cytometry are in gray. Flow cytometry results shown are typical of the independent experiments. *P < 0.05; **P < 0.01; ***P < 0.001

Another difference from previous experiments was the use of induced pluripotent stem cells rather than bone marrow-derived endothelial precursor cells or umbilical cord-derived endothelial colony-forming cells. Gerecht’s team collaborated with Linzhao Cheng from the Institute for Cell Engineering to co-opt her expertise with induced pluripotent stem cells (iPSCs), which are made from adult cells and are de-differentiated through genetic engineering techniques to become embryonic-like stem cells.

Cheng said that this experiment is “an elegant use of human induced pluripotent stem cells that can form multiple cell types within one kind of tissue or organ and have the same genetic background [as the patient].” Cheng continued” “In addition to being able to form blood cells and neural cells as previously shown, blood-derived human induced pluripotent stem cells can also form multiple types of vascular network cells.”

To grow blood vessels, Cheng, Gerecht and others placed the stem cells into a scaffolding made of hydrogel (hyaluronic acid and water). This hydrogel was full of chemical cues that directed the cells to differentiate in to endothelial and smooth muscle cells and form a network of blood vessels. This constitutes the first time human blood vessels had been made from human pluripotent stem cells in a synthetic material.

Self-assembly of EVCs to multicellular networks in a 3D matrix. (A) Network formation from BC1-EVCs in collagen (i) and HA hydrogels (ii). (B) Sorted VEcad+ and VEcad− cells encapsulated within collagen gels were unable to form networks. (Insert) Example of a cell with typical stellate morphology, with phalloidin in green and nuclei in blue. (Scale bars: 100 μm.) (C) Vacuole formation was observed after one day as evidenced by light microscopy (LM) (i) and confocal images (ii) of vacuole vital stain, FM4-64, in red and nuclei in blue. (Scale bar: 10 μm.) (D) On day 2, network formation with enlarged lumen (i and ii) and cell sprouting (iii and iv) were visualized by LM images (i and iii) and confocal images (ii and iv) of FM4-64 in red and nuclei in blue. (Scale bars: 10 μm in i and iii; 20 μm in ii; 50 μm in iv.) (E) On day 3, complex networks were observed with enlarged and open lumen, as indicated by confocal z-stacks and orthogonal sections of FM4-64 in red and nuclei in blue. (Scale bar: 20 μm.) (F) After 3 d, multilayered structures were also detected, as demonstrated by a 3D projection image of NG2 (green), phalloidin (red), and nuclei (blue) showing NG2+ pericytes integrated onto hollow structures. Images shown are typical of the independent experiment. (Scale bars: 50 µm.)
Self-assembly of EVCs to multicellular networks in a 3D matrix. (A) Network formation from BC1-EVCs in collagen (i) and HA hydrogels (ii). (B) Sorted VEcad+ and VEcad− cells encapsulated within collagen gels were unable to form networks. (Insert) Example of a cell with typical stellate morphology, with phalloidin in green and nuclei in blue. (Scale bars: 100 μm.) (C) Vacuole formation was observed after one day as evidenced by light microscopy (LM) (i) and confocal images (ii) of vacuole vital stain, FM4-64, in red and nuclei in blue. (Scale bar: 10 μm.) (D) On day 2, network formation with enlarged lumen (i and ii) and cell sprouting (iii and iv) were visualized by LM images (i and iii) and confocal images (ii and iv) of FM4-64 in red and nuclei in blue. (Scale bars: 10 μm in i and iii; 20 μm in ii; 50 μm in iv.) (E) On day 3, complex networks were observed with enlarged and open lumen, as indicated by confocal z-stacks and orthogonal sections of FM4-64 in red and nuclei in blue. (Scale bar: 20 μm.) (F) After 3 d, multilayered structures were also detected, as demonstrated by a 3D projection image of NG2 (green), phalloidin (red), and nuclei (blue) showing NG2+ pericytes integrated onto hollow structures. Images shown are typical of the independent experiment. (Scale bars: 50 µm.)

While these networks of blood vessels looked like the real thing, would they work within a living creature? The answer that question, Gerecht and her group transplanted them into mice. After two weeks the lab-grown blood vessels had integrated with the mouse’s own blood vessels and the hydrogel had dissolved and been degraded. “That these vessels survive and function inside a living animal is a crucial step in getting them to medical application,” Kusama said.

An important follow-up to these experiments is to examine the three-dimensional structure of these blood vessels to determine if truly have all the characteristics of human blood vessels that can deliver blood to damaged tissues and help those tissues recover from injury or trauma.