VEGF-expressing mesenchymal stem cells for improved angiogenesis in regenerative medicine : a bone tissue engineering approach

Rapid vascularization of tissue-engineered grafts is a major bottleneck in the development of regenerative medicine approaches. In order to overcome this limitation, we aimed to develop a bone tissue engineering strategy combining cell therapy with pro-angiogenic gene therapy.
Vascular Endothelial Growth Factor (VEGF) is the master regulator of physiological vascular growth and is commonly used as a therapeutic transgene for the induction of angiogenesis. However, uncontrolled and high levels of VEGF expression can lead to aberrant vascular growth. To achieve controlled expression in vivo, a high-throughput flow cytometry-based method has previously been developed in our group. Linking the VEGF cDNA to a cell-surface marker (a truncated version of CD8a) in a bicistronic construct enabled the rapid purification of genetically modified myoblasts secreting a desired VEGF level, using FACS sorting based on the intensity of CD8 expression in each cell. Controlled VEGF expression in skeletal muscle, achieved by implantation of these FACS-purified myoblast populations, induced only normal, stable and functional vascular networks and avoided any aberrant angiogenesis.
The aims of this thesis were to adapt this method to human adipose tissue- and bone marrow-derived mesenchymal stromal/stem cells (ASC and BMSC), and to apply these in a bone tissue engineering approach to increase the vascularization potential of osteogenic grafts.
As MSC gradually loose their regenerative potential during in vitro expansion, we first optimized our genetic engineering method for MSC, so as to enable high transduction efficiency and FACS-purification with minimal in vitro manipulation. Chapter 2 describes the generation of an optimized protocol allowing routine transduction efficiencies of > 90% of primary human ASC and BMSC already during the first plating, as well as flow cytometry purification of transduced cells at the time of the first passage. In addition we demonstrated that it was possible to FACS-purify specific sub-populations of transduced MSC homogeneously producing desired VEGF doses. Neither retroviral vector transduction, FACS-purification, nor the expression of the transgenes VEGF and CD8 impaired MSC proliferation and in vitro differentiation potential. Transgene expression was not lost during in vitro differentiation.
In Chapter 3, proof-of-principle was obtained by applying this platform to a bone tissue engineering approach. Human BMSC, transduced and rapidly FACS-purified to eliminate non-expressing cells, were seeded onto hydroxyapatite granules to generate non-critically sized constructs, and were implanted subcutaneously in nude rats. In vivo vascularization potential was significantly increased in VEGF-expressing BMSC. Although VEGF expression was heterogeneous, no aberrant angiogenesis was observed. Indeed, orderly vascular beds were induced, with flow-conducting arterioles feeding into extensive capillary networks, where metabolic exchanges can take place efficiently. The improvement in vascularization was not diminished by extensive in vitro expansion of the transduced BMSC up to 35 population doublings, showing that genetic modification conferred a stable angiogenic potential. As expected, these expanded BMSC lost their osteogenic potential. However, their sustained capacity to induce vascularization could be useful in other applications, where effective expansion of the vascular bed is required, but not progenitor differentiation, such as in cell-based approaches for therapeutic angiogenesis in peripheral or coronary artery diseases.
By minimizing cell expansion, both naïve and control transduced MSC generated abundant bone tissue in vivo. However, VEGF over-expression specifically caused a strong reduction in bone formation. This correlated with an increased recruitment of TRAP-positive osteoclasts specifically in VEGF-expressing constructs.
These data suggest that VEGF over-expression might impair bone formation by disrupting the balance between bone formation and resorption towards excessive degradation. To fully understand the underlying mechanism, further experiments will be needed.
The method described in chapter 2 provides a general platform to generate populations of genetically modified MSC, expressing specific levels of a therapeutic transgene, already at the time of the first passage. Therefore, it has the potential to be applied in other fields of regenerative medicine, beyond bone tissue engineering. We briefly describe two recently initiated projects, based on the results described in this thesis, which aim at either promoting or inhibiting angiogenesis in order to improve cardiac function after myocardial infarction, or cartilage tissue formation, respectively.