Subretinal delivery of lentiviruses bearing Cas9 mRNA and a guide RNA targeting the Vegfa gene reduces the development of choroidal neovascularization in a mouse model of wet age-related macular degeneration.
Lentiviruses and other retroviruses have been engineered for use in gene therapies that require the long-term expression of a foreign gene. Yet because of their capacity to integrate into the host genome, lentiviruses can cause adverse events. Integration-defective lentiviruses can, however, be engineered to package heterologous proteins (via their insertion within the structural viral polyprotein), as well as mRNA (bound to the packaged protein by means of aptamer sequences1; Fig.1a). Reporting in Nature Biomedical Engineering, Yujia Cai and colleagues now show that this strategy can be harnessed to generate lentivirus particles co-packaging mRNA encoding Cas9 and a guide RNA (gRNA; as part of the viral genome) targeting vascular endothelial growth-factor A (Vegfa)2. Delivered to retinal pigment epithelium in a mouse model of wet age-related macular degeneration, the engineered lentiviruses reduced the development of choroidal neovascularization2 (CNV). Such a ‘hit and run’ genome-editing approach (only traces of Cas9 protein were detected 96 hours after in vitro exposure of cells to the lentiviral particles) caused low levels of off-target editing and did not raise immune responses against the Cas9 protein.
Cai and co-authors generated mRNA-encoding plasmids coding for a coat protein of the MS2 phage fused to the N terminus of the lentivirus Gag–Pol polyprotein backbone with an HIV-1 protease-cleavage site inserted between the two fused proteins (Fig. 1a). Cas9 mRNA was packaged into the lentivirus particles (produced in HEK293T cells) via aptamer sequences binding to the expressed MS2 protein. The Vegfa-targeting gRNA was supplied by a U3-promoter-driven cassette in an integration-deficient genome. Lentiviral particles delivered by subretinal injection in C57BL/6J mice led to editing rates of up to 44% in the Vegfa locus of cells in retinal pigment epithelium (the only source of VEGFA protein in the posterior part of the eyes), and to a pronounced reduction of VEGFA levels two weeks after injection. No innate or adaptive immune responses to Cas9 were detected. In the mouse model of wet age-related macular degeneration, induction of CNV (via exposure to high-intensity laser pulses around the optic disc) 7 days after subretinal injection of the lentiviral particles led to a marked reduction of CNV, thus indicating that the reduction in VEGFA levels by this therapy is sufficient to prevent macular degeneration (Fig. 1b).
The direct delivery of either proteins or mRNAs by engineered lentiviral particles has been leveraged for genome-editing applications involving Cre recombinases3, transposases4, meganucleases5, zinc-finger nucleases, transcription activator-like effector nucleases6 and CRISPR7. Cai and colleagues show that a functional CRISPR toolbox can be delivered in vivo via a single engineered lentivirus particle, and that the gene therapy is safe and efficient in mice. The all-in-one viral-vector design has also now been applied as an antiviral therapy in mouse models of herpes simplex virus (HSV) infection8. Reported in Nature Biotechnology also by Cai and collaborators, the intrastromal injection of HSV-1 genome-targeting vectors led to the inhibition of viral replication, the prevention of herpetic stromal keratitis, and the depletion of the viral reservoir in the trigeminal ganglia. Concurrently, in cultured human corneas, the therapy eliminated HSV-1, and led to negligible off-target editing in the cellular genome, as verified by whole-genome sequencing.
The two studies compellingly demonstrate the prospects of in vivo hit-and-run CRISPR genome editing via lentiviral delivery of Cas9 mRNA and a gRNA expression cassette, and thus represent a significant step towards the development of safe and effective CRISPR therapies. However, aspects of this system will require optimization. First, the need for co-transfection of six plasmids for the production of lentiviral particles complicates up-scaling and thereby hampers applicability. Second, the non-selective packaging of cellular proteins and RNAs by enveloped viruses might raise concerns regarding vector safety and efficacy9,10. This is exemplified by Cai and colleagues’ observation of substantial levels of packaged Cas9 mRNA in the absence of the MS2–Gag–Pol protein fusion. Functionalized non-viral nanoparticles generated under chemically defined conditions may provide an attractive alternative. Furthermore, although the temporary expression of Cas9 in the hit-and-run strategy is beneficial, persistence of the episomal gRNA cassette may lead to random integration events (as occurs with adeno-associated viruses)11. Alternative strategies for efficient and timely restricted delivery of the gRNA will be needed.
Because lentiviral particles have a large packaging capacity, they could be used to deliver base editors encoded by mRNAs, and may also fit longer mRNAs encoding prime editors or CRISPR-associated transposases12. Moreover, higher editing efficiencies could be achieved by including stabilizing elements into the mRNA (or via RNA circularization)13 to elongate the expression of the Cas protein; any higher risks of off-target effects could then be alleviated by using Cas nucleases with higher specificity12. Also, engineered Cas variants with minimal immunogenicity may alleviate any partial depletion of the edited cells owing to pre-existing immunity to Cas9 proteins in organs other than the eye14, which is immune-privileged (partly explaining the absence of immune responses). Furthermore, pseudotyped lentiviral vectors with altered tropism may allow for enhanced viral transduction or the targeting of other cell types and organs, and thereby improve the therapeutic index.
Prel, A. et al. Mol. Ther. Methods Clin. Dev. 2, 15039 (2015).
Ling, S. et al. Nat. Biomed. Eng. https://doi.org/10.1038/s41551-020-00656-y (2021).
Michel, G., Yu, Y., Chang, T. & Yee, J. K. Mol. Ther. 18, 1814–1821 (2010).
Cai, Y. et al. Nucleic Acids Res. 42, e28 (2014).
He, C. et al. Gene Ther. 21, 759–766 (2014).
Cai, Y., Bak, R. O. & Mikkelsen, J. G. eLife 3, e01911 (2014).
Lyu, P., Wang, L. & Lu, B. Life (Basel) 10, 366 (2020).
Yin, D. et al. Nat. Biotechnol. https://doi.org/10.1038/s41587-020-00781-8 (2021).
Arthur, L. O. et al. Science 258, 1935–1938 (1992).
Rulli, S. J.Jr et al. J. Virol. 81, 6623–6631 (2007).
Nguyen, G. N. et al. Nat. Biotechnol. 39, 47–55 (2021).
Anzalone, A. V., Koblan, L. W. & Liu, D. R. Nat. Biotechnol. 38, 824–844 (2020).
Wesselhoeft, R. A., Kowalski, P. S. & Anderson, D. G. Nat. Commun. 9, 2629 (2018).
Li, A. et al. Mol. Ther. 28, 1432–1441 (2020).
D.G. is a co-founder and shareholder of AaviGen GmbH. F.B. and D.G. have filed patent applications related to CRISPR technologies and gene delivery via adeno-associated viruses.
About this article
Cite this article
Bubeck, F., Grimm, D. ‘Hit and run’ therapy averts macular degeneration. Nat Biomed Eng 5, 132–133 (2021). https://doi.org/10.1038/s41551-021-00690-4