Gene therapy is emerging at a rapid pace. In December 2017, it achieved a significant first when the U.S. Food and Drug Administration (FDA) approved Luxturna for the treatment of Leber’s congenital amaurosis, a retinal dystrophy that is one of the most common causes of blindness in children. Never before had a directly administered gene therapy for a disease caused by mutations in a specific gene been approved in the United States. A little more than a year later, in January 2019, the FDA indicated that it had been weighing more than 800 INDs for directly administered or cell-based gene therapies. The regulator also offered the encouraging prediction that by 2025, approvals of cell and gene therapies would reach 10 to 20 each year.
As interest in gene therapy grows, manufacturers face physical, biological, and engineering challenges to developing and producing drugs on a larger scale. “For early-stage Phase I trials, little material is required, but for Phase II and Phase III trials, or a licensed product, you need large-scale production equipment,” says Farzin Farzaneh, PhD, professor of molecular medicine at King’s College London.
Farzaneh’s lab participates in extensive academic and industrial collaborations, including projects to accelerate the development and clinical translation of viral vectors. According to him, a major challenge is the limited availability (and expense) of facilities that satisfy current good manufacturing practice (cGMP) requirements. Part of his lab’s collaborative work is developing closed manufacturing systems, which avoid problems with environmental and other contamination.
“When we started out, almost everything was done in isolators, like glove boxes where you open a flask to add something,” he recalls. But with closed systems, he notes, reagents may be added or removed while the systems stay closed, allowing work to be done in a general-purpose lab.
Manufacturing challenges are being addressed not only by academically oriented collaborations, such as those involving the Farzaneh lab, but also by collaborations between companies. For example, Symbiosis Pharmaceutical Services and Cobra Biologics recently completed a project to bring together expertise in viral vector and drug manufacturing. The project, which benefited from a $5.9 million (£4.8 million) investment, positions the companies to offer “gene to vial” services.
Scaling up, not out
Gene therapies have shorter development times than small molecules or biopharmaceuticals, says Henry George, head of viral gene therapy cell lines and bioprocessing, MilliporeSigma. Whereas traditional drug development (from discovery to regulatory approval) can take anywhere between five and seven years, gene and cell therapies can take three to five years—especially as they often fulfill unmet medical needs. “The pressure to produce enough viral vector to treat larger patient populations with short lead times is intense,” he emphasizes. “[But] the development of improved and more efficient production processes has not kept pace with current demand.”
One major issue, he points out, is the tendency for manufacturers to rely on cell cultures that are physically attached to the bottom of growth vessels, such as large T flasks. These adherent cultures can be grown at higher volumes only by increasing the number of flasks, that is, by “scaling out” the process. This approach, however, requires more factory space, labor, and equipment than using a larger bioreactor or other production vessel.
George describes a system for growing cell cultures in suspension. This system, which could allow for thousands of liters to be produced in a single vessel, is designed to work with existing equipment and technologies as well as a propriety version of the industry-standard human embryonic kidney (HEK293T) cell line. This cell line has been adapted to accept viral genes with high efficiency via a low-cost, commonly used polyethylenimine transfection (PEI) reagent.
MilliporeSigma’s longstanding viral vector contract manufacturing service is starting to offer the platform for lentivirus production, George explains, and several customers are already beta testing the platform at their own production laboratories. The company will offer off-the-shelf chemically defined cell culture media, compatible with PEI-based transfection processes later this year, as well as cGMP cell banks based on the technology.
“Viral titers recovered from our suspension-based platform are equal to or higher than those from the industry-standard process of using adherent cultures,” George states. Although he allows that overall titer may vary with the type of gene sequence and other factors, he says the MilliporeSigma platform still offers advantages. “Our process,” he asserts, “allows for faster and more cost-effective manufacturing compared to existing techniques.”
Adopting new cell lines
The challenge of adherent cultures is also being tackled by Voyager Therapeutics. “The classical system that people have used over the past decades is doing a triple transfection in HEK293 cells, but multiple companies are now working on other systems,” observes Luis Maranga, PhD, the company’s chief technical operations officer. “Our preferred system uses a baculovirus to drive adeno-associated virus expression in Sf9 cells.”
Voyager Therapeutics is working on using adeno-associated viruses (AAVs) to deliver DNA to patient cells. AAVs, like gammaretroviruses and lentiviruses, are popular vectors for gene delivery. For example, AAVs are used in an approved gene therapy, namely, Luxturna. “The data the FDA has published show exponential growth in gene therapy applications,” Maranga points out, “and AAV is pretty much one of the vectors of choice.”
According to Maranga, gene therapies using recombinant AAVs have so far tackled rare diseases, but he sees this changing over the next few years. “We are working on therapies for rare diseases,” he details, “but also on therapies for diseases affecting large populations, such as Parkinson’s and Alzheimer’s.”
Dealing with larger numbers of patients means acquiring larger quantities of viral vector, but the systems used at small scale may be unsuitable for high-demand applications. “We’ve invested in a Sf9/baculovirus system because it’s easier to scale up,” he notes. The system enables a process that resembles the CHO-based processes that are used to manufacture antibodies. “The cells are in suspension,” he continues, “and they perform the same in 50-, 200-, and even 2000-L bioreactors.”
In addition, the Sf9 cells produce a higher yield of virus per liter of culture. “When you scale up a HEK293 system, you also have a lower percentage of AAV particles that contain the transgene DNA (that is, full particles),” he asserts. “And you have a lower purity—because purity follows yield.”
“Two of the key challenges we face in making gene therapies are consistency and throughput,” notes Franz Schnetzinger, director of analytical development at Gyroscope Therapeutics. “We’ve been able to show that one way we can address these challenges is through automation of the qPCR and ELISA assays.”
Schnetzinger indicates that Gyroscope has implemented a liquid-handling robot, the OT-2 from Opentrons. The OT-2 offers a flexible design in terms of custom consumables and easy change of pipette sizes. The pipetting steps can be fine-tuned by coding their sequence using a Python script. “It’s this last step that allows us to maximize the power of the platform by fine-tuning our assays’ liquid handling steps.”
Schnetzinger believes that automation could eliminate variability between human operators, improve the repeatability of an assay, and increase throughput of samples while freeing up scientists to do other important work.
“Keep in mind that automation is not just limited to liquid handling,” he advises. “It extends to assay protocols and how we collect and analyze data. There are many areas where it can really streamline the development work that we put into gene therapies.”
Chromosomal debris or leftover chromatin is a major problem in the manufacture of gene therapies. “There are dozens, maybe hundreds of startups needing to get AAVs into clinical trials quickly, but their development efforts are being hampered by an enemy they can’t see,” says Pete Gagnon, PhD, chief scientific officer, BIA Separations.
Chromatin makes up about 10% of the total contaminant load after cell culture harvest but is responsible for 90% of the serious challenges in purification. It forms a super-sticky mass of DNA and histone proteins that depresses both product purity and recovery. This is true even for antibodies, but with AAV being produced at much lower levels, the ratio of chromatin to product goes up, and the magnitude of its adverse effects increases in proportion.
“To date, there have been no chromatin-directed sample preparation methods in the field of AAV purification,” notes Gagnon. Such methods are beginning to emerge, however, now that a solid-phase material designed for chromatin extraction has become available. Gagnon asserts that using this material to extract chromatin in advance of chromatography lets BIA Separations develop clinical-quality purification procedures for any AAV within weeks.
Advance extraction simplifies purification so much that it can be done with non-affinity methods that support higher capacity, better purity, faster processing, and better process economics compared to affinity-based techniques, according to Gagnon. “Compare sailboat racing with a barnacle- and algae-covered hull to a hull with a fresh coat of Teflon,” he suggests. “If you liberate your tools from fouling, they can do what they’re designed to do.”
Keith Carson, founder and content chair of the International Society for BioProcess Technology (ISBioTech), is working on a new lentivirus reference material. “This is very important because there’s no reference material for lentiviruses, even though there has been a huge uptick in their use in production,” he explains. “We’re talking about probably one of the most exciting applications in biotech.”
Although most lentiviruses are used in chimeric antigen receptor (CAR) T-cell applications, most of the interest in the new reference virus is, Carson says, from companies working on viral gene vectors. These companies hope to use the new virus to compare the infectivity titers and particulate counts of their viruses.
These data provide useful safety information for drug regulators, such as the FDA, which can compare the amount of virus that companies are using to dose potential patients. “The FDA doesn’t require that someone use a reference material, but the agency highly recommends it,” Carson notes. This can help prevent patients dying from too much of a vector, especially if they’re immunocompromised.
ISBioTech is currently coordinating a working group to manufacture the new reference virus. “Lentivirus material is being produced by the National Research Council in Canada,” Carson points out. “We’re hoping to be making some pilot-scale batches in late summer/early fall.” He hopes to move to a 200-L production run by next summer with 13 organizations involved in characterizing the virus and preparing to distribute it to interested companies.
Looking to the future
“As we move onto diseases where higher doses of vector are required, there’ll be a move away from adherent to 3D cell cultures,” predicts Jonathan Appleby, PhD, chief scientific officer of the UK’s Cell and Gene Therapy Catapult. “The next stage after that will be continuous bioprocessing.”
Manufacturers will use smaller bioprocessing vessels, he continues, but will draw media off continuously. He also anticipates that companies will move from using three or four plasmids to transfer viral DNA into cells to using only one to two. “That will reduce costs,” he says. “But at the moment, the yields aren’t as good.”
He predicts that synthetic biologists will help gene therapy manufacturers increase yields. For example, synthetic promoters could boost the signaling pathways that cell lines use to produce viral vectors, while shutting down unnecessary pathways. If such production-enhancing innovations succeed, they will help ensure a bright future for gene therapy.