Are base editing therapies ready for clinical translation?

Can base editors be delivered to the right cells in the body?

Delivery is the single biggest barrier to clinical translation. Base editors are large proteins, and getting them into the right cells—without affecting others—is difficult. Most studies here focus on solving this problem. For example, one team used a retrograde injection technique (RIDI) to deliver lipid nanoparticles directly into the pancreas of mice, achieving ~20-30% gene editing and 5% base editing in pancreatic cells [1]. Another group used focused ultrasound to burst microbubbles carrying base editor DNA in the mouse liver, achieving ~2.5% editing [8]. While these numbers are low, they prove the concept works in living animals.

For brain diseases, a team screened a library of lipid nanoparticles and found one formulation (C3 LNP) that delivered base editors to the brains of fetal and neonatal mice, improving the biochemical defect in a lysosomal storage disease model [9]. They also showed mRNA delivery in a nonhuman primate fetus, a key step toward human use. For the liver, a dual-functional polymer that targets hepatocytes directly delivered adenine base editors after a single injection, leading to durable editing of the ANGPTL-3 gene and sustained cholesterol reduction in a mouse model of atherosclerosis [3]. These studies collectively show that delivery is being solved organ by organ, but no single method works for all tissues yet.

How safe and precise are base editors?

Base editors were designed to avoid the DNA double-strand breaks caused by traditional CRISPR-Cas9, which can lead to large deletions or chromosomal rearrangements. The evidence confirms this advantage: in human blood stem cells, adenine base editors achieved up to 90% editing efficiency without triggering the DNA damage response or reducing cell growth, unlike CRISPR-Cas9 [12]. Similarly, in T cells, cytosine base editors enabled quadruple gene edits without causing karyotypic abnormalities or impairing cell proliferation [2]. These are critical safety improvements.

However, base editors have their own safety issues—off-target edits elsewhere in the genome or transcriptome. One study found that standard cytosine base editors caused genome-wide and transcriptome-wide off-target mutations, but they engineered a 'transformer' base editor (tBE) that stays inactive until it binds its target, reducing off-target edits to background levels [7]. Another group inserted small peptides into the deaminase enzyme to narrow its editing window, minimizing bystander edits (edits at unintended bases near the target) [10]. A newer adenine base editor variant, TadA8r, was evolved to work efficiently on a wider range of target sequences (RA context) and corrected disease-relevant mutations in PCSK9 and ABCA4 genes with controlled off-target effects [11]. These engineering advances are steadily improving precision.

Where do the studies agree—and where do they conflict?

There is strong agreement across studies that base editing is more precise and safer than traditional CRISPR-Cas9 for making single-nucleotide changes. Multiple papers highlight the absence of double-strand breaks as a key advantage [2][4][5][6][12]. There is also broad consensus that delivery remains the primary bottleneck, with each study tackling a different organ or cell type [1][3][8][9].

However, there is disagreement about the extent of off-target effects. One study using focused ultrasound delivery found that non-viral delivery revealed greater off-target base exchange in vivo than in vitro [8], suggesting that delivery method itself can influence safety. Another study found that cytosine base editors, but not adenine base editors, triggered a p53 DNA damage response in blood stem cells, albeit milder than CRISPR-Cas9 [12]. This indicates that different types of base editors (cytosine vs. adenine) may have different safety profiles. Finally, while most studies focus on correcting disease-causing mutations, one paper proposes a different strategy—editing signaling proteins to treat heart failure—which could expand the therapeutic scope but introduces new unknowns about long-term effects [5].

About These Sources

This answer is built on 12 peer-reviewed studies — published from 2021 to 2026, 7 from 2024 or later, 11 in Q1 journals, collectively cited 330 times — selected as the most relevant from 15 studies that passed quality screening, drawn from 75 papers retrieved from a database of over 500 million.

Sources used in this answer

Related research questions