Both small and mighty: novel CRISPRa enables tuneable gene control
CRISPR activation promises reversible gene therapy, yet design constraints have limited its clinical translation. A novel inducible system now addresses these long-standing barriers.
Size and inducibility have long limited the in vivo use of CRISPR activation (CRISPRa), a bacterial defence system converted into a potent gene activator. By combining Cas miniaturisation with light and chemical inducibility, Wan et al, writing in nature, introduce HEAL: a CRISPRa system designed for tuneable, spatiotemporal regulation of gene expression in living organisms.
Gene expression is regulated at multiple levels by complex programs, including epigenetic modification and RNA processing. Dysregulation at any of these stages can lead to aberrant protein production and is implicated in disease progression. Since transcription is dynamic, CRISPRa must not only control which genes are activated, but how strongly and for how long. Such control could enable therapeutic restoration of genes lost to haploinsufficiency and large-scale studies of gene function.
However, most potent CRISPRa systems have relied on large Cas proteins fused directly to transcriptional activators. Although this substantially increases gene activation, it comes at the cost of increased construct size and a need to constitutively express strong activators, a combination that complicates viral delivery and risks transcriptional toxicity. Thus, HEAL design focused on addressing limitations of earlier CRISPRa systems: in vivo delivery, transcriptional activation and inducibility.
Rather than designing an entirely novel system, researchers repurposed components of CRISPR-Cas12f, a gene editing system. Here, a catalytically inactive Cas12f is directed by a single-guide RNA (sgRNA) to a specific DNA sequence. A short neighbouring sequence, known as the protospacer adjacent motif (PAM), helps ensure the complex binds to the correct location. Once docked, the system recruits transcriptional activators, which engage the cell’s transcriptional machinery to switch the gene on.
Firstly, most in vivo CRISPR applications rely on adeno-associated viral (AAV) vectors due to low immunogenicity and long-term expression. However, AAV vectors are limited by a packaging capacity of 4.7kb. Conventional platforms often exceed this limit to account for bulky fused activators, requiring multiple vectors. In response, HEAL uses dCas12f, a catalytically inactive Cas variant with a compact size (1.2-1.5 kb) which retains AAV compatibility. Notably, dCas12f contains a triple mutation that enhances DNA binding affinity, compensating for the reduced binding strength typically associated with smaller Cas proteins.
The architecture of many CRISPRa systems introduces a second challenge. The fusion and constitutive expression of transcriptional activators to Cas may induce toxicity (Liang, 2025). This may be due to sustained transcriptional stress or ‘squelching’, in which fused activators titrate endogenous transcription factors and disrupt global gene regulation (Matis, 2001). HEAL circumvents this by decoupling activators from dCas12f, and recruiting them instead, when and where they are needed. To do this, small molecular structures called MS2 hairpin aptamers are embedded within the guide RNA; when the RNA-Cas complex docks onto its DNA target, these hairpins grab the activator proteins and pull them into position. Here, sgRNA acts as a recruitment scaffold, reducing construct size for AAV delivery while allowing amplified gene expression and flexible swapping of activator domains. This strategy outperformed fusion based systems such as dCasMINI in cross comparative studies, suggesting that direct activator fusion may no longer represent the default strategy for robust CRISPRa. Although HEAL employs HSF1–p65, long considered the gold standard, recent protein engineering studies suggest that alternative activator combinations may achieve stronger activation with reduced toxicity (Giddins, 2025), highlighting opportunities for further refinement.
A third layer of risk arises from the output of the system itself: toxicity induced by excessive gene activation. Thus, therapeutic gene control requires not just precision in targeting but also control over timing and dosage. The development of inducible HEAL variants addresses this need. OptoHEAL harnesses red light responsive dimerization between photoreceptor DrBphP and binding partner LDB3 to reversibly recruit transcriptional activation domains, providing a switchable ON-OFF control for manipulating gene expression. Given the limitations of light delivery in deep tissues, the authors also developed ChemHEAL, which uses small molecules to recruit activators.
Figure 1 | Incremental development of HEAL. Previous models such dCas9-VPR utilised fused activators and a large Cas9. Newer models such as dCasMINI and dCpf1 utilise fused activators but replaced Cas9 with smaller Cas12. Overall, previous models prioritised strong gene activation at the expense of increased toxicity and dual vector delivery. HEAL tackles this by recruiting activators when needed and using optimally mutated Cas12. Inducible variants, ChemHEAL and OptoHEAL, were designed to enable controlled, reversible gene activation for those prone to toxicity when overexpressed.
Functionally, the system demonstrates striking potency. HEAL induced up to a 100,000-fold increase in gene expression, exceeding many existing CRISPRa platforms. In a mouse model of acute kidney injury, HEAL-mediated activation of IL-10 alleviated renal damage. OptoHEAL drove light-dependent, reversible activation of Oct4, a gene that is epigenetically silenced in liver tissue. ChemHEAL achieved controlled induction of Tslp, a cytokine prone to toxicity when overexpressed, improving adiposity and metabolic parameters in obese mice. Together, these findings suggest that compact, recruitment-based architectures can achieve both strong gene activation and inducible control.
However, the inducible variants reintroduce a familiar trade-off. The addition of large regulatory modules increases construct size, exceeding single AAV packaging limits and necessitating dual-vector strategies. To address this, Wan et al. propose a tetracycline-inducible alternative, in which doxycycline triggers the incremental transcription of HEAL components. However, leaky expression, where partial gene activation occurs without the drug, risks recreating the very toxicity the inducible design was built to prevent. Although tetracycline-inducible CRISPR systems have proven effective in cancer models (Peng, 2018), whether tetracycline-inducible HEAL variants would retain similar activation efficiency, remains to be determined. Similarly, genome-wide transcriptomic analysis, to confirm that off-target genes are not activated, would be necessary to strengthen therapeutic translations of HEAL.
In summary, HEAL represents a conceptual advance in CRISPRa design. By combining compact architecture, recruitment-based activation, and inducible control, it overcomes key limitations of previous platforms while offering a versatile toolkit for in vivo gene regulation. With further refinement, HEAL and its variants lay a foundation for precise therapeutic interventions.
Bibliography
Giddins, M., Kratz, A.F., De, M.B., Forget, A., Tiwari, R., Jang, G., Blazejewski, T., Qin, C., Huang, Y., Lao, Y.-H., Falconer, T., Leong, K.W., Nevan Krogan, Staller, M., Wang, H., Wei, L., Chavez, A., Giddins, M., Kratz, A.F. and De, M.B. (2025). Combinatorial protein engineering identifies potent CRISPR activators with reduced toxicity. Nature Communications. [online] doi:https://doi.org/10.1038/s41467-025-65986-4.
Liang, Z., Maddineni, A., Ortega, J.A., Magdongon, C.B., Jambardi, S., Roy, S., Tycko, J., Patil, A., Manzano, M., Bartom, E.T. and Gottwein, E. (2025). Cytotoxicity of activator expression in CRISPR-based transcriptional activation systems. Nature Communications, [online] 16(1). doi:https://doi.org/10.1038/s41467-025-63570-4.
Matis, C., Chomez, P., Picard, J. and Rezsohazy, R. (2001). Differential and opposed transcriptional effects of protein fusions containing the VP16 activation domain. FEBS Letters, 499(1-2), pp.92–96. doi:https://doi.org/10.1016/s0014-5793(01)02532-7.
Peng, L., Pan, P., Chen, J., Yu, X., Wu, J. and Chen, Y. (2018). A tetracycline‑inducible CRISPR/Cas9 system, targeting two long non‑coding RNAs, suppresses the malignant behavior of bladder cancer cells. Oncology Letters. doi:https://doi.org/10.3892/ol.2018.9157.
Comments