Scientists at Synlogic Inc. have engineered bacteria to produce phenylalanine hydroxylase (PAH), the gene that is missing or dysfunctional in individuals with phenylketonuria (PKU). Currently in a phase I trial in healthy volunteers and adult phenylketonurics, the bacteria sit at an intersection between synthetic biology and gene correction.

PKU, the disease those bacteria are ultimately meant to treat, is also an illustration of how the optimal form of gene correction can differ even in different forms of the same disease.

Amongst the rare diseases, PKU is a common one, with more than 15,000 cases in the U.S. It is an autosomal recessive disorder, leaving individuals with two nonfunctional copies of PAH unable to metabolize phenylalanine, which builds up in the blood and leads to severe mental retardation if the condition is not detected within weeks after birth.

Kuvan (sapropterin dihydrochloride, Biomarin Pharmaceutical Inc.), a PAH activator, can help some patients keep their phenylalanine blood levels in check. The enzyme replacement therapy Palynziq (pegvaliase, Biomarin Pharmaceutical Inc.) was approved for adults with PKU this year, albeit with a complicated induction schedule and a boxed warning about anaphylaxis, which affected close to 10 percent of those treated in clinical trials. (See BioWorld, May 29, 2018.)

But currently, dietary restriction remains the only treatment for well over 80 percent of adults with PKU.

Bug as drug

In work published in the Aug. 13, 2018, online issue of Nature Biotechnology, corresponding author Vincent Isabella and his team described testing Escherichia coli Nissle, which expressed phenylalanine ammonia lyase (PAL) in response to anoxic conditions in the small intestine, on both a mouse model of phenylketonuria and nonhuman primates fed labeled phenylalanine.

PAL “is an enzyme derived from another bacteria that breaks down phenylalanine to a nontoxic product, and complements the PAH mutation characteristic of PKU patients,” Isabella, who is a principal scientist at Synlogic, explained.

E. coli Nissle did not colonize either mice or monkeys – indeed that lack of colonization was part of clinical development criteria. The bacterium is “passed through stool,” Isabella told BioWorld Insight. “While it’s in the host, it’s degrading phenylalanine, [acting] within the gut to remove any phenylalanine it encounters… We can dose our bacteria and we can prevent [dietary] phenylalanine from making its way into the bloodstream.”

Because E. coli Nissle does not colonize the human gut, the treatment is not curative, Synlogic interim CEO Aoife Brennan noted.

“Our approach is something that would need to be taken daily,” she told BioWorld Insight.

Delivering genes to a patient’s cells via a viral vector, or correcting faulty genes via gene editing, on the other hand, has the potential to cure patients.

Letting biology lead

Not that gene therapy guarantees a cure.

Current methods that deliver replacement genes most often do not lead to correcting the gene in the cell’s own genome. Instead, a functional copy is delivered in a structure called an episome, explained Albert Seymour, Homology Medicines Inc.’s chief scientific officer.

“The challenge with gene therapy is that each time a cell divides, you start to lose those episomes,” which are not copied along with the cell’s own DNA, he told BioWorld Insight.

Homology is working on both gene therapy and gene editing approaches. In July, researchers from Homology and the City of Hope Medical center co-published a paper in the Proceedings of the National Academy of Sciences describing a gene editing approach that was able to use homologous recombination, the most precise of cellular DNA repair tools, to repair a gene without the use of external nucleases to induce DNA breaks.

Seymour said the company’s approach is to “let disease biology dictate” whether a gene therapy or gene editing approach is more appropriate.

In diseases where there is expected to be a lot of cell division after the therapy is administered, editing is preferable, to prevent loss of the therapeutic gene from a large proportion of cells over time. Pediatric PKU, where the therapy would be delivered to “very small livers” that have a lot of cell division ahead of them, is an example. That could be a potential cure for the approximate 300 newborns diagnosed with PKU each year.

Adult PKU, on the other hand, is an example of a disease where target tissue does not turn over very quickly, and gene therapy could be successful. In adult patients with PKU, gene therapy can be applied to deliver a normal copy of the gene that is missing to the liver.

Beyond the specifics of which approach is best under which circumstances, “it’s a fantastic time to be in this field, because you have a variety of tools,” Seymour said.

The work published this week by Isabella and his colleagues illustrates that gene therapy, gene correction and gene replacement are distinct concepts.

Initial gene therapy attempts were straightforward in their focus on delivering a working copy of a gene that was defective in an individuals, such as the IL2RG gene, which, when missing, leads to X-linked severe combined immunodeficiency (X-SCID).

The first gene therapy to make it across the finish line, Kymriah (tisagenlecleucel-t, Novartis AG) – approved almost exactly a year ago, on Aug. 29, 2017 – did not replace a missing or defective gene. Instead, it consisted of the patient’s own T cells, transfected with a gene that encoded for the chimeric antigen receptor (CAR), which does not normally exist in those patients, or anywhere else in nature – hence “chimeric.”

An FDA spokesperson told BioWorld when Kymriah was approved that “the function of the CAR T-cell product depends on the genetic material transferred to the patient’s cells. Therefore, the agency considers CAR T cells to be a type of cell-based gene therapy.”

Synlogic’s approach to PKU has some similarities. It adds a gene not in the cells where that gene would normally be expressed, or any human cells at all. Instead, it’s E. coli Nissle to the rescue.

The Synlogic team’s work also describes something that might otherwise be considered a contradiction in terms: short-term gene correction.

Brennan noted that in the long run, with an improved understanding of bacterial colonization of the gut and how to predict how much PAL would be produced by a given dose of bacteria in a given patient with their own unique genetic and microbiome makeup, developing strains that would colonize the gut after administration and produce PAL over longer periods of time is “certainly very possible in the future.”

For now, though, she said, “we really don’t currently understand the rules of colonization… I just don’t know that the science is there yet.”