Delivering larger genes in gene therapy

Gene therapy has captured the imagination of scientists and science fiction fans alike. But as its development gains momentum, scientists have encountered one major problem: some genes are just too big to deliver into cells. Gene delivery by adeno-associated virus (AAV), the gold standard, can only deliver genes up to five kilobases long, but several genetic mutations associated with disease involve much longer genes.

In a recent study in Nature Communications, University of Zurich geneticist Elvir Becirovic and his collaborators developed a new method called REVeRT (reconstitution via mRNA trans-splicing) that splits longer genes into shorter fragments, delivers the fragments into cells on separate AAV vectors, and then reassembles the genes back into their full length during transcription (1). Literally expanding the capabilities of gene therapy, their approach could be used to treat a broad range of genetic diseases.

Elvir Becirovic smiles while wearing a gray blazer over a checkered blue and white shirt.

Elvir Becirovic develops therapies for retinal genetic diseases at the University of Zurich.

credit: Elvir Becirovic

“Currently we have little choice, but this [work] will expand our toolbox for both gene editing and gene therapy,” said Guangping Gao, a molecular geneticist at UMass Chan Medical School who was not involved in the study.

Becirovic’s team first tested their REVeRT platform in vitro, delivering a cyan fluorescent protein called Cerulean with two separate plasmid vectors that each only contained a fragment of the Cerulean coding sequence. The team appended each fragment with additional DNA sequences called binding domains, each of which were complementary to the binding domain on the other fragment. The binding domains also included a target site of splicing proteins that eventually cut the binding domain off to leave a continuous, full-length mRNA. Becirovic recalled that they spent years figuring out where exactly to split the protein, how to design their binding domains, and where to add additional sequence elements on each fragment of the protein that would better guarantee reassembly. After all the refining, delivering the fragments on two separate plasmids led to as much fluorescence within the cell as delivering both fragments on the same plasmid.

They next used ReVERT to split and deliver luciferase, another fluorescent protein, into wildtype mice and human retinal organoids in vivo. The mice and organoids lit up as much as they did when injected with the whole protein. In mice, the heart lit up thousands of times more than any other organ. “Really, I didn't expect this finding,” said Becirovic. “Maybe in the future, people focusing on the heart will find out what exactly is the driver for this, but honestly I don't know.”

Next, Becirovic’s team wanted to demonstrate the versatility of their new tool and see whether their REVeRT technology was compatible with CRISPR activation (CRISPRa). CRISPRa is another popular gene therapy which switches on or increases the expression of otherwise silent genes (2). They used REVeRT to deliver a “dead” Cas9 (dCas9) protein fused with a series of transcription activators (VP64, p65, and Rta) collectively known as VPR, all of which add up to around six kilobases. They first treated a mouse model of Usher syndrome, the leading cause of genetic deafblindness. Usher syndrome is currently incurable and involves point mutations in the gene Myo7a that are too long to deliver with typical AAV vectors. The team used REVeRT to activate Myo7b, a gene that is typically silent in the retina but performs the same function as Myo7a elsewhere in the body. Western blots revealed Myo7b expression in the retina of treated mice, and no expression in that of untreated mice. To take it a step further, they used the REVeRT CRISPR system to simultaneously knock out the Rhodopsin (Rho) gene, which is prone to mutations, and activate an M-opsin encoding gene, Opn1mw, which compensates for loss of Rhodopsin function. In all their treated mice, Rho expression was effectively erased, while Opn1mw expression significantly increased.

Becirovic recounted being surprised at how smoothly the in vivo studies went. “Before we started this study, I would always say the in vivo part was the most challenging, but in this particular case, it was the in vitro part,” said Becirovic. “We were of course happy because of this, but it was completely different to what I would have expected.”

Currently we have little choice, but this will expand our toolbox for both gene editing and gene therapy.
- Guangping Gao, UMass Chan Medical School

Finally, Becirovic’s team used their gene delivery platform to treat a mouse model of Stargardt disease, the most common inherited retinal disease caused by mutations in a single gene. Patients with Stargardt disease have mutations in the Abca4 gene, which, at 6.8 kilobases long, exceeds the packaging capacity of normal AAV vectors. They injected Abca4 knockout mice with REVeRT vectors expressing split Abca4 and successfully recovered Abca4 expression to wildtype levels in the retinae of three out of four mice.

As for the mouse in which ABCA4 levels were not recovered, Becirovic suspected it might be due to his team’s choice of injection method. Subretinal injections deliver genes directly to photoreceptor cells, resulting in a higher success rate but also a higher risk of damage to the retina. Becirovic's team opted for the much less invasive intravitreal injections, where gene vectors have to penetrate several cell layers and biological barriers to reach photoreceptor cells. “Sometimes, if the intravitreal injection is not placed correctly, the AAV vectors will simply be distributed all over the place and won't reach the photoreceptor cells,” said Becirovic. Still, Becirovic noted that this was the first large gene to be delivered intravitreally and remained optimistic about the successful delivery in the other three mice.

Gao added, “I would want to see the authors make this process more controllable and regulated.” He said that the efficiency and flexibility of the ReVERT system was impressive, but he also suggested that Becirovic’s team show that the system can be modulated to avoid both short-term side effects from triggering the immune system as well as long-term side effects where the transported genes are inherited.

Ultimately, Becirovic hopes that their work will help broaden molecular geneticists’ perspectives as much as ReVERT has broadened gene therapy’s potential applications. He noted that many scientists like to stick to their own systems, himself included. However, he hopes newcomers will be more open to weighing their options. “Hopefully [they will] find out that our ReVERT technology is, at least currently, the most suitable one, particularly for clinical trials,” he said. Becirovic's team is currently working to bring their Stargardt disease treatment to clinical trials within the next year.