What evidence gaps are holding back RNA editing therapies?

How big is the off-target editing problem, and can we fix it?

Off-target editing — unintended changes to RNA molecules that were not the intended target — is a major safety concern that could cause unintended side effects, including activating immune responses or disrupting normal gene function. A 2022 review of site-directed RNA editing systems notes that current tools still suffer from off-target effects that need to be addressed before clinical use [5]. Similarly, a 2026 review on RNA editing in cancer therapy explicitly lists 'off-target effects' as an enduring hurdle that has not yet been resolved [2]. These off-target edits are not just theoretical; they can occur because the guide RNAs used to direct editing enzymes to the correct spot sometimes bind to similar sequences elsewhere in the transcriptome.

A promising solution comes from a 2025 study that introduced MIRROR, a new guide RNA design that mimics natural, highly edited RNA structures found in human cells. This approach boosted editing efficiency up to 5.7-fold in multiple human cell types and in primary mouse liver cells, while presumably reducing off-target activity by using more selective substrates [3]. However, this study was done in cell cultures and animal models, not in humans, so the real-world precision gain remains unproven in patients.

Can we get RNA editing tools to the right cells in the body?

Delivery is arguably the biggest practical barrier. RNA editing tools are large molecules (enzymes plus guide RNAs) that need to reach specific cells without being degraded or causing immune reactions. The evidence shows a stark contrast between tissues that are easy to reach and those that are not. For example, a 2023 study used a single injection of an enhanced RNA base editor into the inner ear of newborn mice and achieved ~80% editing efficiency, restoring hearing to near-normal levels for at least 7 months [1]. The ear is a small, enclosed space that is relatively easy to inject directly. Similarly, a 2024 study delivered a compact CRISPR-Cas13 system via an AAV virus into the retina of mice and human retinal organoids, achieving effective knockdown of a target gene (VEGFA) linked to vision loss [4]. The eye is also an immune-privileged, accessible site.

In contrast, a 2026 review on cancer therapy bluntly states that 'limited antitumor activity was observed in solid tumors in preclinical studies due to insufficient delivery of editing tools' [2]. Solid tumors are dense, poorly vascularized, and often protected by a physical barrier, making it hard for viral or non-viral delivery vehicles to penetrate and reach all cancer cells. This delivery gap means that even if the editing tool works perfectly in a dish, it may fail in a patient with a solid tumor. The review calls for 'realistic translation expectation' until better delivery methods are developed [2].

Does RNA editing work consistently, and how long does the effect last?

Editing efficiency — the percentage of target RNA molecules that get successfully edited — varies wildly depending on the tool, target tissue, and disease model. This inconsistency is a critical evidence gap because a therapy that works 80% of the time in one setting might work only 10% in another. In the hearing loss study, the emxABE editor achieved approximately 80% A-to-I conversion efficiency in humanized mice, and the hearing improvement persisted for at least 7 months [1]. That is a strong result for a specific mutation in a specific tissue. However, a 2023 review on RNA editing for cancer therapy notes that 'variable editing efficiency' remains a problem, and that the field has yet to validate these tools in larger, multi-site clinical trials [2]. A 2021 review similarly highlights that different site-directed RNA editing approaches have different efficiencies and that optimizing these is an open challenge [7].

The transient nature of RNA editing is both a strength and a weakness. Because RNA editing does not alter the genome, its effects are reversible — which is desirable for conditions like acute pain or inflammation where permanent changes are unwanted [6]. But for chronic diseases like inherited deafness or cancer, the effect may wear off, requiring repeated dosing. The durability data are limited: the hearing study tracked mice for 7 months [1], but most other studies report only short-term outcomes in cell lines or animal models [3][4]. No human data on long-term durability exist yet.

About These Sources

This answer is built on 7 peer-reviewed studies — published from 2021 to 2026, 3 from 2024 or later, 5 in Q1 journals, collectively cited 282 times — selected as the most relevant from 10 studies that passed quality screening, drawn from 55 papers retrieved from a database of over 500 million.

Sources used in this answer

Related research questions