Mesenchymal stromal/stem cells (MSCs) are a multipotent cell population characterised by self-renewal and multilineage differentiation towards bones, adipose tissues and myocytes. Given their multipotency and immunomodulatory properties, MSCs have been investigated intensively for applications in the treatment of several diseases, such as bone defects, degenerative nerve diseases and infectious diseases. However, MSCs progressively lose their biological functions after isolation, long cell culture and massive expansion. They suffer from a low survival rate and a high apoptosis rate because of harsh environments in tissue lesions. Introducing exogenous genes to MSCs has emerged as the main bioengineering method to enhance their therapeutic performance. The key to the success of gene transfer depends on gene delivery strategies, which aim to increase DNA uptake and protein expression as much as possible. The availability of gene delivery methods to engineering MSCs is affected not only by transfection efficiency, biocompatibility and cytotoxicity but also by the chemical composition of gene carriers and the physical disruption that modulates differentiation potential. However, the effects of chemical carriers and microinjection on MSC differentiation are poorly understood.

This thesis was conducted to investigate the availability of four delivery methods (calcium phosphate coprecipitation, polyethylenimine transfecton, liposome-based delivery and robotic microinjection) for engineering MSCs in terms of transfection efficiency, cytotoxicity and differentiation potential preservation. A high-throughput robotic microinjection system was used to inject a large number of cells effectively. CRISPR/Cas9 components were delivered into human adult MSCs to knock out the PPARγ gene and modify MSCs genetically. The effects of the four delivery methods on the cell fate of MSCs and the function of therapeutic genes were comprehensively explored using biochemical and histological measurements.

In the first part of the current work, the transfection efficiency and cytotoxicity of the four selected delivery vehicles in human MSCs are compared. Different delivery methods possessed various advantages and disadvantages. As such, transfection efficiency and cytotoxicity should be balanced to prevent undesired effects on MSCs. The capacity of robotic microinjection fabricated in our laboratory is more than 3,000 cells per h with more than 90% viability. Our robotic microinjection system achieves high transfection efficiency with 50%–70% when an eGFP plasmid is injected into MSCs. Microinjection treatment preserves the cellular morphology and the ability of MSCs to differentiate towards osteogenic and adipogenic pathways after transfection. By contrast, the three other deliveries exhibit a low transfection efficiency of 10%–25%, and differently disrupt the multipotency of MSCs. PEI treatment causes severe cytotoxicity, whereas PEI and calcium phosphate coprecipitation further weaken the multipotency of MSCs. These observations reveal the effects on the cell fate of MSCs significantly differ amongst various delivery methods.

In the second part of the work, the effects of gene delivery methods on the function of therapeutic genes after the genetic modification of MSCs are demonstrated. The delivery of the CRISPR/Cas9 system to knock out the PPARγ gene in human adult MSCs is made through robotic microinjection and the three other chemical carriers. The robotic microinjection system injects the plasmid DNA expressing Cas9 and special sgRNAs into human MSCs and achieves high gene editing with 71% indel efficiency. The similar editing efficiency is observed in the three other chemical carriers. After induction, PPARγ-KO MSCs derived by robotic microinjection and liposome conduct differentiation towards the osteogenic pathway, whereas PEI compels the PPARγ-KO MSCs to differentiate towards the adipogenic pathway. Furthermore, calcium phosphate nanoparticles exaggerate the osteogenic function of PPARγ deficiency. Several differentiation biomarkers are detected using qRT-PCR. Our results reveal inevitable effects on exogenous gene function, and the extent of these effects strongly depends on delivery methods to which the cell are exposed.

In summary, the findings of this work suggest that the choice of delivery vehicles leads to various biases in the differentiation potential in MSCs and affects the function of therapeutic gene. The bias towards osteogenic or adipogenic commitment strongly depends on exposure to different delivery vehicles. Thus, the influence of gene delivery methods on engineered MSCs should be sufficiently investigated before infusion into patients. Notably, the influence of each delivery approach seems to show inherent features. PEI, not inorganic particles, causes severe morphological changes, although inorganic particles and PEI lead to permanent damage to multipotency and gene function. After the injection of a target gene into MSCs, robotic microinjection exhibits high transfection efficiency, low cell damage and ability to preserve MSC multipotency and therapeutic gene function compared with those in the three other chemical delivery methods. As such, robotic microinjection combined with CRISPR/Cas9 shows potential for applications that generate engineered MSCs without disrupting differentiation potential and gene function.