PREMIER rewrites the genome without double-strand breaks: human TP53 replacement and protein expression confirmed
Background: Limitations of conventional homologous recombination and integrase technologies, and the bottleneck in large-scale sequence-replacement data for cancer and genetic-disease R&D
Conventional homology-directed repair (HDR) methods that rely on double-strand breaks (DSBs), along with classical integrase systems, have exposed a critical blind spot: they cannot control severe chromosomal translocations, large deletions, or unwanted residual-sequence noise at insertion sites in therapeutic target cells. In particular, structural dropout noise arising during cell dissociation and the limited extension capacity of reverse transcriptase (RT) have made precise insertion of large foreign DNA fragments of 100 bp or more an intractable challenge for computational control and design in silico. This has created a major data barrier to maintaining effective engraftment concentrations of transplanted cells and preventive concentrations of target proteins, while entrenching metabolic and genetic data bottlenecks in therapeutic-baseline design arising from interspecies variation patterns and complex feedback fluxes.
Discovery: Activation of the PREMIER algorithm based on microhomology-arm design and demonstration of synchronized genome-replacement tensors at the cellular scale
To overcome these challenges, the PREMIER platform adopted a proactive in silico design that installs single-stranded microhomology arms at donor–target genome junctions using a prime editor (PE), without inducing double-strand breaks. Integration efficiency was increased through systematic optimization of reverse-transcriptase (RT) activity, prime-editing guide RNA (pegRNA) structure, and donor design. At the same time, a computational algorithm was used to remove batch effects, achieving disruptive gene-replacement efficiencies of 63.4% on average (65.2% median) across diverse cell populations, reaching up to 85.9%. This was 10–20 times higher than conventional HDR and reduced off-target integration by more than 100-fold compared with non-homologous end joining (NHEJ), but these findings cannot be generalized as complete proof of genomic stability. In PREMIER configurations with nick spacing below 100 bp, increased γ-H2A.X signal and decreased Ki67 signal were observed, indicating increased DNA damage and reduced cell proliferation.
Tuning specific pathways and structures and establishing a reversible, finely stratified model of biological homeostasis
The researchers replaced exon 3 of Trp53 in mouse embryonic stem cells (mESCs), including its translation initiation site, with the full coding sequence (CDS) of human TP53 fused to a 3xFLAG tag, enabling human p53 expression under the endogenous mouse promoter. They then confirmed human p53 protein expression by Western blotting in the heart, liver, spleen, and kidneys of the resulting chimeric mice. This result does not demonstrate therapeutic restoration of impaired TP53 function or complete functional validation. It resulted from tuning the rate constants of rate-limiting steps in cancer-development pathways upward and downward in silico, enabling flexible tracking of topological variation curves in transcriptomic networks even under conditions of high cellular stress. By reversibly tuning key nodes in cellular damage-repair pathways, the researchers established a control backbone that allows effective biological homeostasis to return autonomously to the normal range without collapsing under abnormal microenvironmental stimuli.
Outlook: Establishing programmable genome-coding standards and activating next-generation IND digital governance
This large-scale genome-rewriting platform heralds a complete departure from conventional static, linear, post hoc symptomatic biotechnology research and a major transition to programmable infrastructure based on AI computational multidimensional tensor models. PREMIER showed improved efficiency and specificity over existing methods, but further research is needed, particularly to optimize delivery for in vivo applications. Eliminating production-batch variability in company pipelines, meeting companion-diagnostic (CDx) standards, obtaining Investigational New Drug (IND) and cGMP approvals, and shortening drug-development approval timelines were not demonstrated in this study.
Precise and efficient replacement of large genomic DNA segments without inducing double-strand breaks (DSBs) remains a central challenge in genome engineering. Traditional homologous recombination relies on DSBs and long homologous arms, yet it remains inefficient, while recombinase or integrase systems suffer from residual sequences at integration sites. Prime editing (PE), limited by the processivity of reverse transcriptase, struggles to integrate large fragments (>100 bp). To address this challenge, we introduce Prime Editing-Microhomology-Enabled Replacement (PREMIER), a DSB-free platform by installing single-stranded microhomology arms at donor and genomic junctions via PE. In cell lines, PREMIER achieved a mean efficiency of 63.4% (median 65.2%) in diverse target sites, with peak efficiencies reaching 85.9%, exceeding homology-directed repair by 10-20-fold and reducing off-target integrations by over 100-fold compared to nonhomologous end joining. It bypasses the need for long homology arms, simplifies donor preparation, achieves targeted replacement of sequences up to 10.3 kb. In vivo, PREMIER integrates a 6.2-kb oncogene cassette into the mouse liver. Additionally, PREMIER replaces murine Trp53 with human TP53 CDS, generating functional humanized mice. Altogether, PREMIER provides a precise, high-efficiency, and DSB-free strategy for large-scale genome rewriting, offering a powerful tool for complex modeling and therapeutic genome editing.
The PREMIER platform’s large-scale genome-rewriting findings represent preclinical technical validation in cultured cells and animal models; they are not results of direct application to the global finished-drug supply chain or precision personalized-medicine businesses.
RAD51 expression is suppressed in several cell types in vivo, which may limit PREMIER integration efficiency, and improving in vivo performance remains an unresolved challenge. No results were presented demonstrating that patient-specific treatment effects had been secured in clinical settings or that off-target adverse effects and cytotoxicity had been eliminated at their source.
Clinical-trial simulations linked to ClinVar and gnomAD, real-time back-calculation of donor-DNA concentrations, and implementation of companion-diagnostic (CDx) panels were not demonstrated in this study.
The study also did not demonstrate that TP53 replacement results could be used to eliminate efficacy and safety variability across production batches or improve the likelihood of obtaining IND and cGMP authorizations. Further research is particularly needed to optimize delivery for in vivo applications.