Sometimes all it takes is one malfunctioning part to make a whole machine go haywire. But other times, it’s hard to know where the problems started. Just look at cells, where a single error in the genetic code can render a gene impotent and lead to debilitating illness. 

Gene therapies and gene editing tools are the mechanics that parachute into the cell to fix faulty parts or to tack on a working version of the gene. For example, inserting a correct copy of the RPE65 gene into people with the eye disease Leber congenital amaurosis staves off progressive blindness. That’s because in those individuals, RPE65 is the only piece of faulty machinery. When more than one piece malfunctions, it poses a challenge for gene therapies. 

Polygenic or complex diseases characterized by multiple mutations across many impaired genes have recently forced biologists to come up with more creative tactics. There’s the potential to make a big impact if they succeed; Polygenic diseases are much more common than diseases caused by single gene mutations, and they include recognizable names such as cancer, heart disease, and diabetes, which lead to millions of deaths annually.

“The real opportunity, I think, is for the more common diseases,” said Keith Martin, a gene therapy researcher at the Centre for Eye Research Australia. “That’s the next big thing.”

A pathway to treatment

For complex diseases driven by thousands of mutations sprinkled across the genome, targeting every genetic error may not be possible. But there may be patterns that can be targeted instead. 

In age-related macular degeneration and geographic atrophy — the more advanced form of macular degeneration that damages cells in the retina — genetic studies have revealed mutations scattered throughout a set of immune genes encoding complement proteins (1). These proteins normally work together sequentially to trigger immune responses against invading pathogens, but mutations may activate them in inappropriate contexts to trigger retina-damaging inflammation. 

To target the whole complement pathway, Gyroscope Therapeutics developed a gene therapy that uses adeno-associated viral vectors targeted to the retina to deliver more copies of the gene encoding complement factor I (CFI), a molecule that reduces complement activity (2). Preliminary data from Gyroscope Therapeutics’s Phase I/II trial in humans show that the treatment reduces the levels of some complement proteins by almost 50% — even in patients whose disease isn’t driven by CFI mutations.

“The complement pathway is part of the mechanism,” said Martin, who is running clinical trials for Gyroscope Therapeutics’s drug. “But it’s not replacing a single missing gene.”

Targeting an upstream molecule like CFI is “the simplest approach,” said Krishanu Saha, a biomedical engineer at the University of Wisconsin-Madison. “But sometimes that’s just not possible.” Practical considerations can get in the way, such as when the cell types of interest are hard to target. The eye has many advantages since it is small, compartmentalized, and easy to access. But if the cells that need to be targeted have already died, or if the gene that controls the pathway is active too early in development to modulate without affecting many other cells, it might be too late for gene therapy interventions to be effective. 

One solution is to equip cells with new powers that help them fight disease. Take genetically engineered chimeric antigen receptor (CAR) T cells, for example. Inserting genes that help T cells recognize tumor markers offers a relatively straight forward solution for targeting cancers driven by complex mutations throughout the genome. Martin wants to use gene therapy to “toughen up cells,” making neurons more resilient to the type of damage that a barrage of polygenic mutations may cause.

In a study published in Science Advances last year, his team devised a way to deliver a pair of proteins to improve neuronal function in glaucoma (3). This eye condition is particularly challenging to treat because it can arise from many different mishaps in the eye. 

“What we're trying to find is something that works at the level of the nerve cell to make it tougher, irrespective of how it's being affected by its environment [in the eye],” Martin said.

For his strategy to work, the team needed to be able to deliver two proteins: a signaling molecule called brain-derived neurotrophic factor (BDNF) and its receptor, which promote healthy nerve signaling in the eye. Delivering the gene for BDNF didn’t work because the cells became overwhelmed by the flood of BDNF and reduced expression of the receptor in response. The researchers weren’t eager to deliver the two genes separately because it was harder to control how much of each was produced.

To get around this, they built a single piece of DNA that encoded the two proteins with a small connector in between that created a gap when the cell’s protein-making machinery encountered it. This allowed the proteins to separate and go off on their own missions. BDNF traveled into the space around the cell, while the newly formed receptors migrated to the cell’s surface. Because they were delivered together as a self-cleaving protein, both the molecule and the receptor were present to bind to each other.

Packing the long piece of DNA encoding these two proteins into a viral vector was no small task; it required strategic choices to remove other components of the vector. The double-delivery system prompted better nerve function in rodent models of glaucoma or Alzheimer’s disease-like neurodegeneration than delivering BDNF or its receptor alone. Martin’s team is now designing similar self-cleaving peptides for macular degeneration, and they next hope to deliver three genes to counter diabetic macular edema.

“Now that we've demonstrated that this technology can work in one situation, there's a real opportunity to use it in others as well,” Martin said.

Escalating edits

Saha expects genome editing to have a place in the search for gene-based treatments for polygenic diseases, but that will require a better understanding of how editing proteins operate when there’s more than one cut to make in the genome.

CRISPR-based gene editors use a piece of RNA to guide them to their targets, where they change the genetic sequence to fix a disease-causing mutation, activate, or repress a gene. When there are multiple mutations, these editors have to target multiple sequences in the genome to produce a therapeutic effect, which poses a challenge when a single molecule needs to navigate long expanses of the genome. “Making one cut and then fixing mutations that are more than thousands of bases apart is very difficult,” Saha said.

Another option is to deliver multiple base editors, each tasked with fixing one mutation. This also isn’t easy. Similar to the struggles that Martin encountered when delivering two genes, delivering two base editors requires a spacious vector. Base editors also aren’t known for precision; they often fail to edit their target or edit the wrong part of the genome. As scientists add more editors, the chances of getting the right combination of edits decreases.

“As an engineer, we have all these different knobs that we can turn to customize a genome editing strategy to a particular genotype that we want to fix,” Saha said. “Going into this world, there was a need for more quantitative modeling to understand the effects of these different knobs that we can turn.”

These uncertainties prompted Saha to take a closer look at what happens when two base editors are set loose in the genome. He tested this approach in cells from patients with a condition called Pompe Disease, which isn’t polygenic but it still has some tricky genetic characteristics.

Pompe Disease is caused by mutations localized in one particular gene called GAA, which encodes a protein called acid alpha-glucosidase that breaks down glycogen, the complex assembly of many glucose molecules. When mutations inactivate GAA, glycogen builds up to potentially fatal levels in cells. The disease is recessive, but a person can inherit different mutations from each parent, which leaves two mutations to fix in the gene.

Published in Nature Communications, Saha’s findings illustrate the complex negotiations happening between base editors in a single cell. His team developed a computational model that predicted how effectively they could edit liver cells in infants by taking into account factors such as the base editors’ precision, how much of each base editor was delivered, when it was delivered, and what kind of cell it reached.

Precision is key, especially when editing at two positions. If the first base editor is not precise, it may make edits that prevent the second editor from recognizing its target sequence. Editors that are more precise didn’t always edit efficiently, sometimes editing as few as 5% of cells. But Saha isn’t deterred because using a different delivery mechanism such as a lipid nanoparticle and repeated dosing can boost the number of cells it reaches. “These gene correction strategies that are quite precise still have promise as a therapeutic,” he said.

For a disease like Pompe Disease, targeting both mutations isn’t necessary to treat the patient; a single fix still works. To make multiple edits, however, there’s still a lot of work to be done to understand how cells react to having multiple parts of their genomes edited and not always edited correctly. As scientists try to make three, four, or more edits, efficiency will continue to drop. According to Saha, organs like the eye may be good places to start because established delivery systems ensure higher efficiency.

“We have our hands full with single gene disorders,” Saha said. “But certainly the next logical step could be trying to administer multiple of those editors to try to fix multiple variants.”

References

1. Fritsche, L. G. Age-Related Macular Degeneration: Genetics and Biology Coming Together. Annu Rev Genomics Hum Genet  15, 151-171 (2014).
2. Kiss, S. Interim Results from a First-inHuman Phase I/II Gene Therapy Study (FOCUS) of GT005, an Investigational AAV2 Vector Encoding Complement Factor I (CFI), in Patients with Geographic Atrophy (GA). Presented at the Retina Society. September 30, 2021.
3. Khatib, T. Z. et al. Receptor-ligand supplementation via a self-cleaving 2A peptide–based gene therapy promotes CNS axonal transport with functional recovery. Sci Adv  7(14), eabd2590 (2021).
4. Carlson-Stevermer, J. et al. Design of efficacious somatic cell genome editing strategies for recessive and polygenic diseases. Nat Commun  11(1), 6277 (2020).