Assistant Professor Lehigh University Bioengineering, Pennsylvania, United States
Introduction: Mesenchymal stromal cells (MSCs) offer significant potential for treating a wide range of human musculoskeletal disorders, encompassing conditions such as cartilage defects, osteoarthritis, non-healing bone fractures, and tendon and muscle injuries. This is attributed to MSC multipotency, immunomodulatory properties, and ease of extraction and in vitro expansion. However, despite over 400 ongoing clinical trials involving MSCs listed on ClinicalTrials.gov, the FDA has yet to approve any MSC therapy. To address these challenges, CRISPR-based cell engineering approaches hold promise in producing next-generation stem cells with enhanced functionality, specificity, and responsiveness. However, CRISPR broader application in stem cell biomanufacturing is constrained by the absence of non-cytotoxic delivery vectors and concerns about off-target effects. Furthermore, the application of these therapies for the repair of musculoskeletal tissues is also limited by the lack of suitable delivery systems for localized and effective action at the injury site. Given these challenges, our main goal is to increase the therapeutic efficacy of MSCs for musculoskeletal applications through novel synthetic biology tools in vitro and in vivo. This talk will be centered on our research toward (1) developing safer and more effective strategies for stem cell engineering using different CRISPR systems, (2) their integration with machine learning to enhance their specificity and reduce side-effects, and (3) engineering biomaterial-guided CRISPR delivery strategies for the repair of musculoskeletal tissues.
Materials and
Methods: For enhanced CRISPR delivery to primary stem cells, we investigated the arginine-alanine-leucine-alanine (RALA) cell penetrating peptide (CPP) for the delivery of CRISPR machinery in different molecular formats for a wide range of editing modalities. First, we explored the knock-in of reporter genes into a safe harbor location of the MSC genome. Next, gene knock-out was performed through the RALA-mediated delivery of Cas9-gRNA ribonucleoprotein (RNP) complexes targeting the reporter gene sequences in the edited MSCs. Finally, gene activation (CRISPRa) was performed for the transcriptional activation of BMP2, TGFB3, and TNMD genes to enhance MSC differentiation into bone, cartilage and tendon. To identify CRISPR genetic targets and achieve reliable control over MSC fate, we designed a synthetic gene regulatory network (GRN) using publicly available RNA-seq data. Differential expression analysis elucidated the top differentially expressed genes (DEGs) in mature musculoskeletal tissues in comparison to MSCs. Then we identified gene candidates for MSC differentiation and together with key regulatory genes we designed a deep neural network (DNN) model. Following model training, CRISPR-mimetic perturbations were made in MSC RNA-seq datasets. These datasets were then input into the trained DNN and model predictions were compared to mature musculoskeletal tissue datasets to determine optimal CRISPR target genes for each application. To achieve localized and sustained CRISPR delivery, different biomaterials (alginate hydrogels and 3D printed polycaprolactone scaffolds) were assessed for their capacity to bind and release RALA-CRISPR nanoparticles. Their biomaterial loading efficiency and temporal release kinetics were tested to identify appropriate delivery platforms for controlled CRISPR delivery.
Results, Conclusions, and Discussions: In summary, we offer a novel perspective for the engineering of primary stem cells and their use for the repair of musculoskeletal tissues. We have systematically investigated and optimized more efficient cell delivery strategies, machine learning-based approaches for the selection of appropriate genetic targets, and promising biomaterial systems for in vivo administration. We demonstrate the potential of the RALA cell-penetrating peptide for CRISPR gene editing strategies in therapeutically relevant MSCs. We confirm that RALA can encapsulate and deliver CRISPR machinery in pDNA, mRNA, and RNP formats. Furthermore, we show that RALA preserves cell viability and sustains stem cell proliferation resulting in greater overall transfection yields and greater therapeutic promise than standard commercial vectors. We also employed RALA to facilitate gene knock-in, knock-out, and transcriptional regulation in MSCs demonstrating its therapeutic efficacy for a variety of gene editing modalities. Additionally, we have also developed, a tunable machine learning algorithm for the evaluation and discovery of CRISPR gene targets in a wide range of biomedical applications, from cell-specific differentiation to whole tissue disorders. To validate this machine learning model, we applied it to identify CRISPR-activation gene targets for the chondrogenic differentiation of MSCs, as well as the use of CRISPR-inhibition as a treatment for osteoarthritis (OA). In each application, CRISPR-GEM successfully identified both established and novel genes. All identified genes hold the potential to have a profound therapeutic impact for the given application, and future research will focus on the experimental validation of these genetic targets through high-throughput screening systems. Finally, the incorporation of RALA-CRISPR nanoparticles into biomaterial platforms resulted in spatiotemporal control over CRISPR action, highlighting the potential of this system for the regeneration of a wide range of musculoskeletal tissues.
