Rare genetic disorders affect over 300 million people globally, yet roughly 95 percent lack specific treatments. For the many caused by haploinsufficiency, where a single functional copy of a gene no longer produces enough protein, therapeutic options have been sparse. Most attempts to upregulate the healthy allele’s protein output have relied on antisense oligonucleotides (ASOs) that block upstream open reading frames (uORFs). Now, a new computational and experimental approach suggests a simpler alternative: splice-switching that selectively removes exons containing uORFs.
In a December 2025 study, Beer Wells et al. investigated whether excluding uORF-containing exons via splice-switching could boost translation of haploinsufficient disease genes. The results suggest this method is feasible and generalizable. This opens a path toward therapy for dozens of rare monogenic disorders that have so far had no specific treatment.
The Problem: Why Haploinsufficiency Is Hard to Treat
Haploinsufficiency occurs when one functional gene copy produces less protein than the cell requires, causing disease. Unlike loss-of-function mutations (where there’s often a clear therapeutic target), haploinsufficient disorders require the opposite: boosting protein output from the remaining healthy copy.
Upstream open reading frames are short stretches of RNA sequence that precede the main protein-coding sequence. When translation machinery encounters a uORF, it typically translates that small protein first, then re-initiates or fails to initiate translation of the main protein. For many genes, uORFs are regulatory; they reduce translation of the main protein to a specific level. But when that one copy is all a cell has, low translation becomes a liability.
Several rare neurodevelopmental and genetic disorders are caused by haploinsufficiency. Examples include haploinsufficient variants in CTCF (a transcription factor), GRIN2B (a glutamate receptor subunit), KRIT1 (involved in vascular development), and TSC1 (a tumor suppressor). For these genes, even modest increases in protein output could potentially prevent or mitigate disease.
The Approach: Genome-Wide Screening for Druggable uORFs
Beer Wells and colleagues took a computational-first approach. They screened the genome for uORF-containing exons in 5’ UTRs of genes known to cause haploinsufficient monogenic disorders. Rather than designing a therapy for a single gene, they asked: how many disease genes have uORF-bearing exons that could theoretically be skipped by splice-switching?
The screening identified 79 candidate exons across dozens of disease genes. The logic is elegant: if you remove the exon containing the uORF, the mRNA sequence changes, potentially shifting how translation initiates. The main protein-coding sequence remains intact, but translation of the primary protein can proceed more efficiently.
They then tested this hypothesis experimentally. Using a luciferase reporter assay (a standard method where firefly luciferase protein expression serves as a quantifiable readout), they modeled six high-priority 5’ UTR exons from neurodevelopmental disorder genes (CTCF, GRIN2B, KRIT1, TSC1, PTEN, and DYNC1H1).
Key Results: Translation Boost Across Multiple Targets
The results were encouraging. In four of the six genes tested (CTCF, GRIN2B, KRIT1, and TSC1), removing the uORF-containing exon significantly increased luciferase expression. The fold-changes ranged from 1.4-fold to 5.5-fold. This represents a substantial gain in translation efficiency.
This increase is clinically meaningful. In haploinsufficient disorders, boosting protein output by even 1.5 to 2-fold can cross a critical threshold and prevent disease. The mechanisms differ: for CTCF and GRIN2B, exon skipping may shift the translation start site to a downstream ATG codon, bypassing the uORF entirely. For KRIT1 and TSC1, the mechanism likely involves altered secondary structure that changes ribosome scanning and initiation efficiency.
The two genes that did not show strong responses (PTEN and DYNC1H1) still provide useful negative data. Their lack of response is likely due to downstream regulatory elements or the complexity of their 5’ UTR structure, indicating that not all uORF-containing exons are equally amenable to this approach.
How This Compares to Existing Therapies
The field currently relies on antisense oligonucleotides as the primary tool for blocking uORFs. ASOs are effective but have limitations: they must be redesigned for each target, require repeated dosing, and don’t always achieve the required level of uORF blockade. Splice-switching using antisense oligonucleotides designed to modulate splicing (splice-switching ASOs) offers a different mechanism. Instead of blocking translation, they alter the mRNA structure itself, potentially offering better durability and on-target specificity.
The Beer Wells study did not directly compare splice-switching ASOs to traditional uORF-blocking ASOs, so the relative efficacy in cells or animals remains to be established. However, the computational screening itself is novel and could accelerate identification of candidates for future clinical trials.
Limitations and Important Caveats
This work is early-stage, and several caveats must be acknowledged:
First, all experiments used in vitro luciferase reporter assays, not cells or animal models. Reporter assays are useful for proof-of-concept but do not capture the complexity of endogenous gene regulation, cellular context effects, or organism-level physiology. Translation in primary patient cells or affected tissues might behave differently.
Second, only four of six tested targets showed substantial increases in translation. While this is a 67 percent success rate, it suggests that 5’ UTR exon skipping will not work universally. The authors did not explore why PTEN and DYNC1H1 failed to respond, which limits generalization.
Third, the study is computational screening followed by in vitro testing. No splice-switching oligonucleotides were synthesized or tested. The theoretical approach is sound, but confirmation that actual ASOs can achieve the desired splicing pattern and produce the predicted translation boost in cells awaits.
Fourth, off-target effects were not evaluated. Modulating splicing can affect multiple genes and transcript isoforms. The authors did not assess whether removing these exons might disrupt other regulatory functions or produce truncated proteins with unintended consequences.
What This Means in Practice
For patients with rare haploinsufficient disorders, this work opens a door. It identifies specific, targetable molecular features (uORF-containing exons) that could make dozens of previously untreatable conditions eligible for therapy development. If the approach generalizes, we could see splice-switching ASOs enter clinical trials within the next few years for conditions like CTCF-related disorders or GRIN2B-associated developmental delay.
For computational biologists and bioinformaticians, the screening pipeline itself is valuable. The authors demonstrate a tractable workflow: identify haploinsufficient disease genes, screen their 5’ UTRs for uORF-bearing exons, then prioritize based on conservation and predicted splicing patterns. This same approach could be applied to other therapeutic modalities (CRISPR activation, small molecules, etc.) where understanding genomic context is critical.
For the broader gene therapy field, this work is a reminder that therapeutic innovation often comes not from brute-force strength but from precise molecular engineering. Rather than asking “how do we block everything?” the Beer Wells study asks “which specific exon can we remove to achieve therapeutic benefit?” That precision is increasingly possible thanks to improvements in computational genome annotation and splice-site prediction.
What’s Next
The natural next step is to synthesize and test the candidate splice-switching ASOs in cultured cells derived from patients carrying haploinsufficient variants in these genes. Demonstrating restoration of protein levels in disease-relevant cell types would be a major validation. Animal models (especially for conditions like TSC1 deficiency, where mouse models are well-established) would follow.
Longer-term, if an ASO candidate shows sufficient safety and efficacy in preclinical work, Investigational New Drug (IND) enabling toxicology studies and eventually a first-in-human trial could follow. Given the unmet medical need in rare disorders, regulatory pathways like breakthrough designation are possible for candidates that show clear promise.
Source and Further Reading
Beer Wells, E. S., De Conti, L., Kim, H. C., et al., 2025, bioRxiv. “Modulating splicing in five prime untranslated regions to treat rare haploinsufficient disease.”