Draft:Minimal versatile genetic perturbation technology

mvGPT components and functions

Minimal versatile genetic perturbation technology (mvGPT) is a novel platform for modifying the genome with the ability to perform gene editing, activation, and repression in human cells in a single package. It incorporates elements of prime editing, guide RNA, and short hairpin RNA in order to perform these functions. mvGPT is unique among other genome modification tools in that it performs all the functions of genetic editing, activation, and repression orthogonally, meaning different effects can be directed to particular genes. Additionally, it has been engineered to be capable of single package transmission by viral vector or mRNA, minimizing the effort needed to perform these separate actions on one target organism.

History

Over the last decade, there has been rapid progress achieved in genome editing - encompassing a wide array of novel tools and techniques. One notable example is the CRISPR-Cas system, which induces double-stranded DNA breaks at sequence specific points in the genome. However, this system is inefficient and prone to unintended errors, leading to growing concerns for safety for therapeutic application.

Base editing is an alternative method of engineering gene expression. Single base modification can provide a safer, albeit limited, alternative to techniques reliant on nucleases and endogenous DNA repair. Base editing is ultimately unable to be effectively incorporated into a versatile, multiplexed gene editing and activation platform.

Recently, a new technology known as prime editing (PE) has built upon previous genetic perturbation technology to address prior concerns. Prime editing uses impaired Cas9 endonuclease fused to a reverse transcriptase enzyme to direct site specific insertions, deletions and base conversions.

The last hurdle that faced researchers was the ability to scale up efficient gene editing. Developed recently, the direct and process (DAP) array provided a novel multiplex technique useful for gene editing, activation and repression. The steps forward in each area of genetic engineering directly laid the groundwork for the combinatorial platform - mvGPT. Minimal versatile genetic perturbation technology was developed at the University of Pennsylvania in 2025 is the latest in genetic engineering technology.

Components of genetic perturbation platform

Genome editing

mvGPT uses a drive-and-process (DAP) array in order to enact genome editing. The DAP array consists of an engineered human cysteine transfer RNA (hCtRNA) serving as an RNA Polymerase III promoter, which in turn facilitates the expression of small RNAs programmed and contained within the complex. Prime editor with advanced kernel, or PEAK, contains a truncated Moloney Murine Leukemia Virus reverse transcriptase (MMLV-RT) engineered for increased efficiency and reduced size, an engineered N-terminal VirD2 nuclear localization sequence (NLS), and a C-terminal SV40 NLS.

Structure of the Drive-and-Process (DAP) array
Component	Function
N-terminal and C-terminal NLS	Facilitate nuclear localization of the array
Human cysteine transfer RNA (hCtRNA)	RNA Polymerase III promoter and spacer
engineered pegRNA	Combine to form prime editor with advanced kernel (PEAK), which is directed to target loci and directs genome editing.
nicking guide RNA

Mechanism

The DAP array carries the RNA load and coordinates the endogenous production of RNAs with unique functions. Nuclear localization sequences (NLS) guides the transfer array into the nucleus. hCtRNA act as promoters and spacers between other elements and recruit endogenous elements for transcription and translation.

Prime editor

Once the PEAK component of DAP has been transcribed, the fusion protein encoded by ngRNA creates a nick in the target DNA sequence, exposing a 3’-hydroxyl group that functions as the starting point (primer) for reverse transcription of the RT template section of the pegRNA. This generates a branched intermediate with two DNA flaps: a 3’ flap carrying the newly synthesized (edited) sequence, and a 5’ flap containing the unedited, redundant DNA sequence. The 5’ flap is then removed by structure-specific endonucleases or 5’ exonucleases. This enables ligation of the 3’ flap, forming a heteroduplex DNA with one edited strand and one unedited strand. The reformed double-stranded DNA has nucleotide mismatches where the editing occurred. To correct these mismatches, the cells use the inherent mismatch repair (MMR) mechanism, which can lead to two possible outcomes: (i) the edit is permanently incorporated into the complementary strand; or (ii) the original nucleotides are restored in the edited strand, eliminating the edit.

Endogenous gene activation

Truncated antigene RNAs (agRNA) integrates PEAK with the synergistic activation mediator (SAM) to recruit the transcription fusion activator MS2-p65-HSF1 (MPH) to promote gene activation.

agRNA-SAM system
Component	Function
Catalytically dead Cas9 (dCas9)	Deliver complex and bind to specific genomic location
Truncated agRNA containing MS2-binding stem loops	Precise gene targeting and recruits MPH activation complex to the SAM system
MS2-p65-HSF1 (MPH) transcriptional activator	Recruits transcription factors and chromatin remodelling complexes

Mechanism

To develop a gene activation system using prime editing, the SAM system was included. The SAM system is an RNA-guided gene activator consisting of a catalytically dead Cas9 (dCas9), activation sgRNA (agRNA) with MS2-binding stem loops, and an MS2-p65-HSF1 (MPH) transcriptional activator. The system works by recruiting MPH to the target loci through the MS2-binding stem loops. This, in turn, attracts transcription factors and chromatin remodeling complexes, leading to upregulation of gene expression.

Endogenous gene silencing

shRNA-mediated gene silencing

Gene repression is achieved through RNA interference (RNAi) via the DAP array-generated short hairpin RNA.

Mechanism

Once the DAP array has been delivered into the host nucleus, hCtRNA promotes the transcription of shRNA by polymerase III. The resulting transcript resembles pri-microRNA (pri-miRNA) and is processed by Drosha. This generates pre-shRNA, which is then exported from the nucleus by Exportin 5. The pre-shRNA undergoes further processing by Dicer before being incorporated into the RNA-induced silencing complex (RISC). The sense (passenger) strand is discarded, while the antisense (guide) strand guides RISC to target mRNA with a complementary sequence. If there is perfect complementarity, RISC cleaves the mRNA. In cases of imperfect complementarity, RISC inhibits translation of the mRNA. Both processes lead to the silencing of the target gene.

Platform delivery

Since mvGPT is an RNA construct, it can be delivered through several methods. As with traditional viral vector delivery, mvGPT can be delivered by retroviruses such as Lentivirus, or other viral delivery methods such as adeno-associated virus. Since mvGPT is specifically engineered to be as minimal in size as possible, several of its elements can be delivered with a viral vector via mRNA; However, the DAP array itself requires a viral delivery platform due to its size

Implications

Strengths

The capability of mvGPT hinges on the integration of several strengths that enhance its efficiency and flexibility. One of the key features of this system is its compact and robust nature, minimizing the size of each of its individual components and reducing the steps for achieving genetic perturbation. mvGPT is also a generalizable genetic perturbation system allowing for adaptations to a wide range of cell types, organisms, and experimental setups, useful broadly in genetic research, and in future therapeutic application,

This method allows for a precise genome editing function combined with the ability to simultaneously modulate gene expression. This dual capability ensures that researchers can not only edit specific genes but also control their activity levels, facilitating a deeper understanding of gene function and regulation. The precision of the system minimizes off-target effects, enhancing the accuracy of the perturbations made.

Another advantage is its ability to operate completely endogenously upon introduction, requiring no further proteins or RNA. This reduces the potential for interference with the host genome, particularly important in applications where maintaining the natural state of the cell or organism is essential.

mvGPT also offers flexibility with viral and non viral delivery modalities for introducing the system into target cells. Viral delivery allows for highly efficient gene transfer, while non-viral methods offer safer alternatives that may be preferable in clinical settings.

Limitations

While the new genetic perturbation platform offers significant advancements, several limitations remain, particularly as the system progresses through early stages of in vivo testing. One of the primary concerns is the uncertainty surrounding the success of expression changes in more complex environments. It is well established that without tissue specific targeting, genetic perturbation can produce varied response. Broad applicability of the system requires adaptability to a variety of cell types, while minimizing unwanted off target effects in non-target tissues. Developing a truly modular and flexible system for diverse tissues is a key hurdle that will need to be overcome before the system can be widely adopted in research, and therapeutics are considered.

As the technique moves from in vitro to in vivo, the complexity of the organism increases, making it harder to predict and control off-target interactions. Despite the platform's precision, there is always the possibility that unintended genetic alterations could occur. These off-target modifications can have unforeseen consequences, making it critical to further refine the system to minimize these risks.

Ultimately, the lack of engineered activation or deactivation of the current mvGPT system functionally limits its utility.

Applications

Metabolic engineering

With mvGPT’s stated success in vitro, it could likely first be applied to current metabolic engineering systems. Synthetic biology is often limited due to its inability to create multiple perturbations in a minimalistic organism. mvGPT provides an endogenous system, capable of transient and permanent gene expression changes to fine-tune complex networks independently. In a general use case, mvGPT can edit a gene encoding a rate-limiting enzyme to improve catalytic activity, while concurrently activating genes upstream of the pathway to increase precursor supply and repressing competing pathways that consume intermediates. Normally to optimize a metabolic pathway would require extensive trial and error, but mvGPT carries out multiplexed selection in parallel reducing time and cost. Technology capable of combinatorial gene expression changes can lead to expanded product portfolios through simultaneous activation of multiple biosynthesis pathways. It also has the potential for swift shifts in production targets, improving efficiency.

Genetic screening

Genetic screens in complex organisms can be hard to discriminate due to redundancies. Using a combination genetic perturbation tool can help to better understand the intricacies of synergistic or antagonistic molecular interactions. Specifically, simultaneous perturbation can provide a rapid screen of both gain of function and loss of function phenotypes. Additionally, this endogenous, minimalistic system preserves natural physiological conditions ideal for understanding the dynamic genetic regulation with alterations to gene expression.

Cell therapy

Redesigning cell therapies to include a system capable of permanent and transient transcriptional modulation could be highly beneficial as part of a treatment course. One example could be the use of hematopoietic stem cell therapy for the purpose of tissue regeneration. In an allogeneic transplant, mvGPT could edit cells to promote differentiation to a specific cell type. The system could also temporarily repress host immune cell response mechanisms to mitigate graft-versus-host disease, while simultaneously activating genes responsible for cell homing to tissues of interest. This process can be incorporated on a variety of tissues and stages of differentiation including reincorporating immune cells after ablation and modified mature cell transplantation.

Specifically, CAR-T cell therapies can greatly benefit from the level of cellular manipulation achieved through a multi-component genetic perturbation system like mvGPT. Utilizing the system’s ability to endogenously confer transient or permanent genetic alterations of the CAR T constructs – such as mitigating loss of inhibitory receptors, introducing novel tissue targets and modify resistance to differentiation and exhaustion. This is especially useful to alleviate changes in the T cells during the manufacturing process.

Genetic disease

mvGPT can reduce current therapy burden by employing gene modification strategies to target a mutation, promote recovery and suppress any affiliated negative side effects. This system could be a drastic improvement to the recent advancements in monogenic disease treatments. However, polygenic diseases such as cardiovascular disease, type 1 diabetes and neurological diseases often require a more complex, multivariate approach. One clear use case would be that explored as mvGPT proof of concept – successful editing of a mutation causing Wilson’s disease, while upregulating a gene linked to type 1 diabetes treatment and suppressing another associated with transthyretin amyloidosis.

mvGPT can enhance modelling for complex genetic interaction associated with polygenic disease. By targeting multiple variables at once, researchers can identify synergistic or antagonistic effects between genes. This could immediately contribute to more accurate polygenic risk scores – assessing influence of genetic variation on disease risk and severity, and refine targeted therapies for these diseases.