Improving the genetic engineering of human mesenchymal stromal cells with HAdV-5 vectors: a toolbox for new therapies

The genetic engineering of human mesenchymal stromal cells (hMSCs) is a promising strategy to improve their therapeutic potential and clinical translation. Viral vectors based on human adenovirus type 5 (HAdV-5) are a versatile tool for this purpose. However, due to a lack of coxsackie and adenovirus receptor (CAR) expression, transduction of hMSCs with HAdV-5 vectors is naturally very inefficient, and strategies to overcome this limitation are essential for therapeutic application. In the present study, we identified and characterized several transduction enhancers which significantly improved the transduction of hMSCs with HAdV-5 vectors. Especially with the polyamines spermine and spermidine, expression levels of a reporter transgene were increased more than 1000 fold, and significantly elevated expression and secretion of a potential therapeutic protein (Tumor necrosis factor-inducible gene 6 protein (TSG-6)) was achieved. Furthermore, we found that the genetically charge-modified vector HAdV-5-HexPos3 (exchange of 13 negatively charged amino acids by four lysines in the hypervariable region 1 (HVR1) of Hexon) enabled efficient transduction of hMSCs (and several tumor cell lines) without the need for enhancing molecules. Transduction with HAdV-5-HexPos3 was even significantly higher than with the chimeric vector HAdV-5/3, which has been shown to be one of the most efficient adenovirus-based vectors for the transduction of hMSCs to date. We found that HAdV-5-HexPos3 uses a CAR-independent uptake mechanism that involves binding to heparan sulfate proteoglycans (HSPGs), while the interactions with non-cellular blood components were unaltered. In addition, we evaluated the enhanced transduction of hMSCs in two pre-clinical models of hMSC-based oncolytic virotherapy, in which hMSCs were used as carrier cells for oncolytic viruses to the tumor tissue. However, neither in two human head and neck squamous cell carcinoma (HNSCC) xenograft tumor models established in mice nor in a chorioallantoic-membrane (CAM) model migration of transduced hMSCs towards the tumor was detected. Nevertheless, in in vitro analyses, we showed that both, transduced or infected hMSCs, migrated towards several cell lines, including HNSCC cell lines. Moreover, we confirmed efficient replication of wildtype and life cycle modified HAdV-5 viruses in hMSCs, which was significantly accelerated by a knock-out mutation of the adenoviral E1B 19K protein. Strikingly, replication-competent vectors carrying the HexPos3 modification showed significantly improved replication, highlighting the potential of HAdV-5-HexPos3 for hMSC-based oncolytic approaches. In summary, transduction enhancers and the novel HAdV-5-HexPos3 mutant are valuable tools for the efficient genetic engineering of hMSCs either to overexpress therapeutic transgenes or in the context of oncolytic virotherapy. However, in vivo testing of hMSC-based oncolytic approaches requires further investment in suitable animal models.