Throughout the U.S. presidential election season and into the Trump presidency, the high cost of drugs has made for heated discussion. A popular topic for pundits is the amount drug companies spend on marketing compared to the generally smaller sum they spend on research.
Case study #1: Clovis takes Lonza up on a dedicated plant for its cancer drug.
Case study #2: Immunomedics and Johnson Matthey connect for antibody-drug conjugates.
Case study #3: AMRI can do Nemus’s cannabinoid chemistry.
Little attention, though, is paid to what it costs them to manufacture their products. That’s in part because manufacturing is less glamorous than marketing and in part because it’s a hard question to answer. Reports on the topic put manufacturing at anywhere from 15 to 50% of the overall cost of getting a drug to market.
What is clear is that molecules have to be built before they can be tested, approved, and sold. For small and start-up firms that don’t have molecule-building assets of their own, contract service partners are essential to getting this job done.
In the pages that follow, C&EN presents stories of three small companies working on new drugs with outsourcing service partners. One of those drugs, Clovis Oncology’s Rubraca, was approved by the Food & Drug Administration in December and is now selling for close to $14,000 per month.
People can argue over the price of the drug, but there is no disputing that Clovis owes some of its success to a pharmaceutical outsourcing partner.
The Swiss fine chemicals firm applies a biologics strategy to small-molecule manufacturing
By Rick Mullin
The Swiss fine chemicals firm Lonza is given credit for pioneering contract manufacturing of active pharmaceutical ingredients (APIs). Leading the way among its European cohort, Lonza pushed an industry in which APIs were made by drug companies in factories dedicated to a single product toward one in which they are produced under contract by chemical companies operating in versatile plants that can quickly switch capacity from one project to another.
Lonza’s sprawling, Alps-nestled manufacturing site in Visp, Switzerland, has risen as a paragon of the multipurpose approach to manufacturing that now defines pharmaceutical outsourcing.
But it seems the company is now pointing in a new direction, roughly in reverse, as it begins building capacity dedicated to one product manufactured for one customer.
In October, Lonza announced plans to build a plant for a current client, Clovis Oncology, at which it will make the API for rucaparib, a drug approved last year to treat late-stage ovarian cancer. The new plant, which will go on-line in 2019, will consolidate disparate manufacturing processes and will include advanced automation, allowing for more-rapid release of materials than the batch-release process under which the product is currently made at Visp.
The partners will not discuss financial details or provide information on the volume of material being produced. But the drug, which won U.S. approval in December under the name Rubraca, is awaiting approval in Europe for the same indication. It’s also in trials for broader application in ovarian cancer and in development for prostate cancer. Every indication is that volume will be higher by 2019.
Lonza already operates dedicated plants for biologic APIs in Singapore and Portsmouth, N.H. The Clovis project will be its first such plant for a small molecule.
Patrick Mahaffy, chief executive officer of Clovis, says Lonza has been a dependable supplier ever since his company acquired rucaparib from Pfizer in 2011. But Clovis anticipates that increased demand for the product will tax the current approach of using multipurpose manufacturing assets at Visp.
“One issue is that part of the process requires being in a containment facility,” Mahaffy says. “There is always a lot of competition to get in there.” Above all, he says, Clovis needs Lonza to be able to respond to changes in demand that haven’t had to be managed until recently.
“We went to Lonza and said, ‘Let’s see a proposal that addresses lead time and gets the cost of goods down,’ ” he recalls, “and they came back with a dedicated approach, which they have never done for a small molecule.”
According to Christian Dowdeswell, Lonza’s head of commercial development for chemical and microbial manufacturing, the request for a proposal came at a good time because Lonza was reviewing its approach to serving key customers. “We wanted to develop a better understanding of what our customers need and really start to tailor solutions to those needs,” he says.
In Clovis’s case, the needs were access to assets, security of supply, and production flexibility for variable but likely increasing demand. “We took a look and came up with this solution,” Dowdeswell says. “The concept is a highly automated plant available to Clovis all year round with guaranteed access.”
The plan calls for consolidating processes that are currently dispersed at the huge Visp site. “We have a wide breadth of high-potency and high-containment assets, and I think Clovis has touched most of those throughout the period of development,” Dowdeswell says.
Engineering an integrated plant will lead to obvious efficiency improvements, he says, and the opportunity to introduce a higher level of automation will also allow “real-time release”—automated quality testing in the manufacturing line during production—of material that is currently released after finished batches are tested.
Converting from batch testing to real-time release can in some cases reduce production time from several months to a matter of weeks, Dowdeswell says.
Mahaffy agrees that automated testing will deliver key time and cost savings. “The use of real-time release will eliminate or greatly reduce the analytical testing necessary for each batch of drug substance and the associated quality review of that testing after the conclusion of processing,” he says. “These costs are not a large component of the cost of goods, but they do contribute over time and large numbers of batches.”
The more important benefit, he says, is inventory management. “Being able to move material quickly through the manufacturing facility without the need to hold material pending testing and release will allow more plant time to be dedicated to production. This could translate into more batches per year and correspondingly lower cost of goods.”
Once production is completed, Mahaffy says, Clovis will be able to move the API to finished-dose manufacturing operations more quickly than from a batch-released operation, and ultimately to a product ready for sale. “This reduces the duration of time that we hold inventory in work-in-progress status,” he says.
Dowdeswell concedes that building the plant for Clovis contradicts the “standard model” of multipurpose contract manufacturing that Lonza helped establish. “However, the multipurpose concept doesn’t address some of our customers’ fundamental needs in today’s environment,” he points out. “We took time to gain an understanding of what Clovis needed and to develop a concept for how to best address those needs and add value to our customer in ways that traditional business models cannot.”
Lonza debuted the dedicated plant model for biologic drugs, most recently deploying it in a planned mammalian cell culture manufacturing facility for Sanofi in Visp. Lonza is now promoting it as an option for small molecules.
There are no guarantees in any pharmaceutical chemical manufacturing venture, and some question whether designing a plant for one customer’s small-molecule drug makes sense.
James Bruno, head of the consulting firm Chemical & Pharmaceutical Solutions, points out that the volumes of API found in new drugs are generally trending downward, meaning large amounts of capacity are typically not needed. And although FDA requirements for biologic drugs support a dedicated manufacturing model, the versatile multipurpose chemical plant may still be best for a large contract manufacturer such as Lonza, Bruno says.
There is little question, however, that engineering a plant to produce a single product will result in more efficient manufacturing, and Lonza and Clovis are optimistic that rucaparib will keep the new plant busy. It may also have a future beyond rucaparib.
“The plant is dedicated to this drug for now, but if the need emerges, it could be applied or modified for additional molecules as well,” Mahaffy says. For the time being, however, Clovis is enthusiastic about the new approach to making rucaparib. “It was a great proposal from Lonza,” Mahaffy says, “a really creative response.”
Special handling delivers drug and linker combo for creating tumor-targeting antibody-drug conjugate
After 35 years of toil with few rewards, good things are finally coming to Immunomedics: It has a drug development deal worth up to $2 billion and a solid contract-manufacturing relationship. But the New Jersey-based firm also has an impatient shareholder that thinks it knows what’s best for the company.
Although it got started in diagnostic imaging, Immunomedics has long tried to evolve into a drug company focused on targeted cancer therapies. It has built a pipeline of humanized monoclonal antibodies to be used alone as therapies or conjugated with radioactive isotopes, chemotherapeutics, cytokines, or cytotoxins.
The company’s most advanced antibody-drug conjugate is sacituzumab govitecan, or IMMU-132. In the ADC, a long-chain chemical linker called CL2A connects SN-38, the active drug form of the chemotherapy agent irinotecan, to an antibody that targets the cell surface receptor TROP-2.
SN-38 is about 1,000 times as potent as irinotecan, but it can’t be administered systemically because of its toxicity and poor solubility. The ADC is designed to deliver the chemotherapeutic payload directly to cancer cells while sparing healthy ones. So far, it seems to be working.
Immunomedics has reported good patient responses to IMMU-132 in Phase II clinical trials for several solid-tumor cancers. FDA has designated IMMU-132 as a breakthrough therapy for patients with triple-negative breast cancer who did not respond to other therapies and given it fast-track approval status for that indication as well as two forms of lung cancer. All this helped secure a licensing agreement worth up to $2 billion with ADC-specialist Seattle Genetics to finish development and commercialize IMMU-132.
Immunomedics got IMMU-132 to this point largely by itself. For early-stage work, “we manufactured the antibody, the linker plus the drug, and did the conjugation,” says Cynthia Sullivan, Immunomedics’s CEO. An outside manufacturer did the final formulation.
Because TROP-2 is present on a variety of solid tumors, Immunomedics moved to a “basket approach” for clinical trials, testing the drug for different indications simultaneously. “We are looking at about 13 different types of cancer,” Sullivan says. The need for drug material has risen as a result of this approach, preparations for late-stage trials, and the fast-track approval status.
In 2013, Immunomedics started working with the contract manufacturer Johnson Matthey Fine Chemicals to scale up production. Johnson Matthey’s custom pharma solutions unit in Devens, Mass., is tasked with making the linker and attaching it to SN-38. For now, Immunomedics still provides the antibody, but eventually Lonza will produce it at large scale in Singapore, Sullivan says. Italy’s BSP Pharmaceuticals conjugates the pieces and completes the final formulation.
Though using multiple suppliers is logistically complicated, it is common in ADC development. “We haven’t had any real issues in terms of supply,” Sullivan says. As scale-up progresses, comparability testing of the original components made by Immunomedics and those from its contractors has been under way. Eventually all three suppliers must undergo their own preapproval inspections for Immunomedics to be able to market its product.
“Working with a supplier in the same time zone makes direct communication more streamlined and has helped us build a closer working relationship, but it was not a critical factor,” Sullivan says about partnering with Johnson Matthey. “It was more important to identify a high-quality manufacturing organization with the right set of technology and expertise.”
Garrett Dilley, a senior director at Johnson Matthey, says the company already had the know-how and most of the equipment needed to support Immunomedics. Top on the list are production suites that can handle highly potent compounds such as SN-38. The firm was also familiar with developing and scaling up processes for linking small-molecule payloads to long-chain linkers and polymers for drug delivery applications.
Assembling the drug-linker combo presents some challenges, largely around protecting and derivatizing hydroxyl groups on the SN-38 molecule, says Serengulam V. Govindan, senior director for conjugation chemistry at Immunomedics. Most important, the CL2A linker is attached to SN-38’s lactone ring to prevent the ring from opening and the molecule converting to an inactive carboxylate form. A short polyethylene glycol segment in CL2A helps solubilize the insoluble SN-38.
Given the complex nature of the drug-linker combo, a relatively large molecular weight “small molecule,” according to Dilley, the project needs sophisticated analytical methods to understand parameters impacting quality, especially during scale-up.
Whereas Immunomedics was producing tens of grams of material, Johnson Matthey is making hundreds of grams and will reach kilograms. To prepare for full-scale manufacturing, Matthey is working to boost productivity. It is also using engineering techniques to increase efficiency in handling the molecule, given its potency, Dilley says.
This capability may be useful again because other ADCs in Immunomedics’s pipeline—such as labetuzumab govitecan (IMMU-130), which is in Phase II trials for metastatic colorectal cancer—are based on the same drug-linker duo but conjugated to other antibodies.
The company’s approach “differs from the current paradigm for making ADCs in the sense that the drug and the linker characteristics are quite different,” Govindan says. Many ADCs in development use ultrastable linkers to avoid premature release of extremely toxic drugs before getting to a target. “We use a moderately potent drug” and instead conjugate more of it to the antibody, he says.
Immunomedics’s chemistry, which involves a maleimide group on one end of the linker, allows for attaching up to eight drug molecules per antibody site-specifically at reduced interchain disulfides while avoiding the formation of dimers and aggregates. The pH-sensitive, cleavable linkage is moderately stable, with the potential to release SN-38 in the acidic tumor microenvironment and increase the drug’s bioavailability. Together, these features allow for reduced toxicity, higher antibody doses, repeated therapy cycles, and a better therapeutic window, Immunomedics says.
Despite IMMU-132’s success so far, not everyone is waiting patiently. Immunomedics’s largest shareholder, the venture capital firm venBio, has been in a long and contentious battle with the company for control. It even opposes the recent deal with Seattle Genetics.
VenBio is concerned about the ability of Immunomedics’s board and management to “realize the significant potential of IMMU-132.” If the drug is approved, sales for four cancer indications could reach $7.5 billion per year by 2025, according to Jefferies stock analyst Matthew Andrews.
Immunomedics and venBio have tried to negotiate but are also suing each other. The decision over control came down to a shareholder vote held on March 3 that went in favor of venBio. As a result, it looks like a new plan may be in place for IMMU-132.
Service firm’s government-approved assets are right for developing a drug based on psychoactive found in pot
Investors are hot on cannabinoids. Shares in GW Pharmaceuticals, marketer of the world’s first plant-derived cannabinoid prescription drug, have more than tripled in the past year. The stock of Zynerba Pharmaceuticals, which is developing transdermal synthetic cannabinoids, has more than doubled.
In contrast, shares in Nemus Bioscience, the newest publicly traded cannabinoid company, have not caught on yet. And unlike its high-visibility competitors, which trade on the NASDAQ stock exchange, Costa Mesa, Calif.-based Nemus trades quietly on the less-prestigious over-the-counter market.
But Brian Murphy, Nemus’s CEO, is biding his time. Murphy sees his firm as an up-and-coming player in second-generation cannabinoids. Nemus has been raising modest amounts of money from investors in preferred stock offerings. And importantly, it has signed up a respected contract research and manufacturing organization, Albany Molecular Research Inc. (AMRI), to develop and produce the active ingredient in its lead product.
Nemus got its start in 2012 after one of its founders, Cosmas Lykos, was approached by investors looking to capitalize on growing interest in cannabinoids. It went public two years later via a reverse stock merger with a trucking company called Load Guard Logistics.
“As is apropos for the area of cannabinoids, we are a little unconventional in how we were founded,” Murphy acknowledges. Murphy himself has a pretty conventional background as chief medical officer for pharmaceutical firms including Valeant International, InterMune, and Roche.
Nemus’s ace in the hole is a partnership with the University of Mississippi (UM), which since 1968 has held the sole federal contract to cultivate cannabis for research purposes. One of Nemus’s scientific advisers is Mahmoud A. ElSohly, director of UM’s marijuana project. “He literally wrote the book on cannabinoid chemistry,” Murphy says.
In 2014, UM licensed Nemus the rights to tetrahydrocannabinol-valine-hemisuccinate (THC-VHS), a prodrug of THC, which is the main psychoactive component of the cannabis plant. Nemus renamed the compound NB1111 and is now developing it as a treatment for glaucoma. The company has several other cannabinoids in its pipeline as well.
Researchers have known since the 1970s that smoking marijuana reduces intraocular pressure in patients with glaucoma. “It’s the first time in 20 years of drug discovery that I’m working with a drug I already know works before human trials have started,” Murphy says.
But the smoking effect is short-lived, he notes, meaning patients have to stay almost constantly high for lasting glaucoma relief. THC eye drops would seem to be an option, but the molecule is lipid soluble and doesn’t easily cross the cornea and get absorbed into the eye.
UM scientists discovered that attaching VHS to THC improves THC’s water solubility. And once THC-VHS traverses the cornea, Murphy explains, esterases in the body cleave the amide-ester bond that links the molecule’s two halves, freeing up THC.
UM followed up with rabbit tests showing that the effect of THC-VHS eye drops on intraocular pressure is better than that of THC alone or of existing treatments such as pilocarpine. However, researchers found no THC in the animals’ circulatory system, indicating that a psychotropic effect was unlikely.
The next step for Nemus was finding a contract services firm that could further develop and produce THC-VHS to the standards that FDA demands before agreeing to human clinical trials. And because THC is considered a controlled substance, Nemus needed a partner with the right Drug Enforcement Agency (DEA) license.
“Right away that filtered a lot of companies out of the mix,” Murphy says. With the help of consultants who work with Nemus, he settled on AMRI.
AMRI, Murphy soon learned, isn’t just licensed by the DEA to produce controlled substances. It has become something of an expert at cannabinoid chemistry. Pete Michels, a senior director and biocatalysis expert at AMRI, says the company has worked with close to a dozen partners on cannabinoid-based therapeutics.
As part of that push, Michels notes, the company is becoming a provider of unique cannabinoid molecules to the research community. Last year, for example, it licensed biosynthesis technology from Teewinot Life Sciences allowing it to produce the first commercial analytical standard for cannabichromenic acid, a nonpsychotropic cannabinoid being investigated as antiviral, antifungal, and anti-inflammatory.
Teewinot’s technology uses biocatalysis to create cannabichromenic acid and other cannabinoids. Alternatively, cannabinoids including THC can be recovered through extraction from the cannabis plant.
But for Nemus’s THC-VHS, AMRI is using a purely synthetic process that it developed in Rensselaer, N.Y., and scaled up for manufacturing at its Grafton, Wis., site. As with many natural products, AMRI and Nemus had some complex stereochemistry to solve, Michels says. AMRI also streamlined the synthesis with several process improvements, such as more-selective chemistry and the elimination of a chromatographic separation step.
Michels says AMRI and Nemus chose synthesis because it is reproducible, scalable, and unencumbered by the related cannabinoids that can come along with an extraction. “You want to ensure that no potentially psychoactive components are generated,” he says. “This can be a significant advantage of synthesis versus botanical generation, which starts from a mix of compounds.”
Others have noted that FDA and DEA are more comfortable with synthetic cannabinoids than with plant-derived ones. Tellingly, GW Pharmaceuticals’ product Sativex, a multiple sclerosis treatment based on a mix of cannabinoids extracted from the cannabis plant, has yet to win approval in the U.S.
Yet AMRI also uses extraction and fermentation in its broader pursuit of cannabinoid chemistry. “There are about 100 cannabinoids out there,” Michels says. “Clearly, each of the three approaches can have advantages for specific molecules.” By drawing on all three, he adds, AMRI is able to help customers investigate common cannabinoids such as THC and cannabichromenic acid, as well as lesser-known molecules.
Through acquisitions and internal growth, AMRI has become one of the largest U.S.-based providers of chemistry services to pharmaceutical companies. Its customers include some of the biggest names in the global drug industry. Nemus, on the other hand, is the smallest of start-ups.
But Christopher Conway, AMRI’s senior vice president of discovery and development services, sees no disconnect. Serving small and virtual biotechs, he says, is core to what the company does. “Many of our largest commercial products started with a small biotech,” he points out. “Part of our strategy is to create these ‘sticky’ customers.”
And it doesn’t hurt when that biotech firm is pursuing a high-profile area of chemistry like cannabinoids. “The market is very hot,” Conway observes. “There’s so much untapped potential that demand will continue to grow.”
CORRECTION: This story was updated on March 31, 2017, and Sept. 11, 2019, to correct the structure of tetrahydrocannabinol-valine-hemisuccinate.