Aug 2026· Nature Biotechnology· 0 citations· 81 references
Medicine
TL;DR
A virus-like particle (VLP)-based toolkit that delivers diverse CRISPR editing modalities to human monocytes, macrophages and dendritic cells with high efficiency while preserving viability and innate immune responsiveness is presented.
Chimeric antigen receptor T cell immunotherapies are transforming therapies for hematological malignancies and solid tumors and can be enhanced by targeted gene knockout. Here, we report lentiviral-based virus-like particles that package and deliver Cas9 ribonucleoproteins to primary human T cells. Using distinct pseudotyping strategies for virus-like particles and for lentiviral or γ-retroviral vectors, we achieved chimeric antigen receptor expression and targeted gene disruption. Under optimized transduction conditions, more than 50% of T cells expressed a chimeric antigen receptor by flow cytometry, with vector copy numbers exceeding two. Editing efficiencies were above 70% at three different target loci tested: T cell receptor α constant chain, β2-microglobulin, and DNA methyltransferase 3α. When the editing efficiency of virus-like particles was directly compared to electroporation, electroporation achieved a higher editing efficiency (99% versus 70%–90%). However, virus-like particle treatment resulted in twice as many cells being recovered compared with electroporation with a 10% increase in cell viability. Furthermore, off-target editing in virus-like particle-treated cells was reduced compared to ribonucleoprotein electroporated cells. These results support the feasibility of using virus-like particle-mediated delivery of Cas9 ribonucleoprotein to disrupt genes of interest, enabling a more scalable and cost-effective process for generating T cell immunotherapies.
Francesca Ferrara, Matthew M Wielgosz, Jeoungeun J. Park et al.· Molecular therapy. Advances· 0 citations
A detailed protocol is described for a small-scale production of AsCas12a-VLPs using three distinct transfection methods and a large-scale production of VLPs using calcium-phosphate transfection, showing that both production scales yield comparable nuclease loading into VLPs and similar editing efficiencies.
N. Kruglova, S. E. Borovikova, M. Shepelev· Frontiers in Genome Editing· 0 citations
To systematically map cellular factors constraining nonviral genome editing, influencing uptake and intracellular trafficking, we develop a genome-wide CRISPR screening platform linking perturbation of 19,114 genes to editing outcomes in human cells. We identify six negative regulators of delivery whose depletion increases editing efficiency by up to six-fold across diverse payloads, loci, and cell types. We test the top two factors, GJB2 and BET1L, in two distinct human models: correction of a pathogenic adenine base mutation in KCNJ13 and introduction of a cytosine base mutation in the GABAA receptor gene. Depletion of either improves base-editing outcomes by 6-fold, potentially through effects on delivery. In a patient-derived model of retinal channelopathy, knockdown of either gene improves lipid nanoparticle base editing efficiency by over 3.5-fold. This enables functional restoration of Kir7.1 ion channels in a subset of edited cells, highlighting cellular barriers as actionable targets to enhance the potency of genetic therapies. Low editing efficiency of nonviral delivery in post mitotic tissues presents a challenge to the field of gene therapy. Here, authors dissect the genetic regulators of nonviral delivery in post mitotic retinal epithelial cells describe strategies for improved base editor delivery and editing.
Shivani Saxena, Meha Kabra, Amr A. Abdeen et al.· Nature Communications· 0 citations
A novel genome-wide CRISPR screening strategy that will facilitate the systematic engineering of novel nonviral genome editing delivery methods, where the identified novel gene hits can be further used to increase editing efficiency for other therapeutically relevant cell types.
Shivani Saxena, Meha Kabra, Amr A. Abdeen et al.· bioRxiv· 2 citations
A new method for surgically removing training examples from a model reveals that as datasets grow, the link between what a model learns and what it produces dissolves.
MIT News · Artificial Intelligence· news.mit.eduAug 17, 2026