Across the world, more than 15,000 chemical compounds are concocted every day, most of them in university laboratories. Until recently, the majority ended up in deep freeze — or on the rubbish heap.

Credit: Bratislav Milenkovic

“When I was a PhD student I made 300 molecules,” says Matt Cooper, a chemist at the University of Queensland in Brisbane, Australia. “Three years' work, completely new molecules, all weird, all handcrafted. I remember having each of them labelled and bottled, but basically they went in the bin.”

More than 20 years later, Cooper is director of one of the world's largest molecular-compound-screening programmes: the Community for Open Antimicrobial Drug Discovery (CO-ADD). Frustrated by the waste of novel molecules, any one of which might hold the key to life-saving treatment for antibiotic-resistant bacterial infection, malaria or tuberculosis, Cooper wants to put these under-appreciated creations to use.

CO-ADD is collating and screening vast quantities of small molecules in an effort to help spur the development of powerful new antibiotics. Funded by the University of Queensland and Britain's Wellcome Trust, the programme has already received 40,000 compounds for screening since its launch in early 2015. Of these, around 26,000 have undergone high-throughput screening, an automated procedure that can test the biological activity of thousands of compounds a day and an established part of the early drug-discovery process.

The compounds are tested for antimicrobial activity against the most dangerous hospital-acquired, antibiotic-resistant infections Escherichia coli, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa and methicillin-resistant Staphylococcus aureus (MRSA), as well as the common causes of fungal infections Cryptococcus neoformans and Candida albicans. The programme is working with 88 groups from 26 countries and has a further 300,000 compounds on the way, including those from what will be the first nationwide antimicrobial screening project to be coordinated by France's National Centre for Scientific Research.

It is an initiative born of necessity at a time when profit-driven antibiotic research is floundering, says Cooper. “We think one of the reasons the industry is not productive any more is because we have been looking in the wrong chemical space.” He thinks that new drugs, including antibiotics, are likely to come from outside the conventional chemical structures held in the vast collections of pharmaceutical companies.

To see whether the molecules created in academic settings are more fruitful than the commercial compound libraries, Cooper and his colleagues began to screen compounds put forward by an academic group for efficacy against the selected bacteria and fungal pathogens. Compared with the biological activity of commercial libraries, the academics' compounds had a 20–30 times higher hit rate. “Academics make compounds for a whole range of reasons,” says Cooper. “They are more eclectic and more diverse.”

CO-ADD is now leading the way in tapping this diverse pool of potential drug leads by making it as easy as possible for chemists to share their creations and uncover potentially valuable medical uses. “We make no claims on intellectual property, so all the results and all of the intellectual property rests with the provider of the compounds. There are no strings attached at all,” says Cooper. “We have become the first group to take open access to the nth degree and really reach out to people all around the world — crowdsourcing our molecules.”

In an era of low productivity in early drug discovery, medicinal chemists in the corporate realm have reached a similar conclusion to Cooper — the lack of chemical diversity is affecting early drug development. In late 2015, pharmaceutical giants AstraZeneca and Sanofi took the unusual step of exchanging 210,000 proprietary compounds to boost the variety of compounds in their collections. And, like CO-ADD, some big pharma are inviting external researchers to put forward their compounds to determine whether they may have some medical application. In a world in which knowledge is fiercely protected, the new approach is prompting ongoing adjustments by everyone involved. Academic chemists are becoming more open about their research, and companies are now more willing to share intellectual property to encourage researchers to participate. This open-innovation approach is young, and drug development is notoriously slow — a decade at best in most cases — so it may be years before a blockbuster drug results from this process. But the platforms are showing promise, which is sparking excitement among the enthusiastic pharmaceutical executives who are driving them forward.

Chemical diversity

A lot of the academic chemistry is quite unusual and quite specialized.

Screening external compounds is a ready-made source of innovation for big pharma and requires little investment, given that the companies already have the sophisticated screening technology in place. “It gives us access to compounds that are really different to the types of compounds that we have in our collection,” says Garry Pairaudeau, head of external sciences at AstraZeneca in Cambridge, UK. “A lot of the academic chemistry that is going on is quite unusual and quite specialized and would be difficult for us to reproduce in our lab.”

Merck KGaA, based in Darmstadt, Germany, Leo Pharma and Eli Lilly are also hosting compound-screening platforms that are designed to engage external researchers (see 'Search history'). The corporate platforms ask academic scientists or those associated with biotechnology start-ups to first submit their compounds for computer analysis. The structure of the compound is not shared with the drug giants at this stage, ensuring that the original researchers' intellectual property is protected. This analysis is carried out to determine whether the molecule meets the company's threshold for further examination. Compounds can be rejected because they are too reactive or unstable, or too similar to previously screened compounds. In the programmes run by AstraZeneca and Leo, this initial analysis is performed by a third party as an added precaution to protect confidentiality.

If compounds meet the various thresholds, the external researchers then provide a sample (typically 1–5 milligrams) for further investigation. It is at this point that the molecular structure is shared, and agreements that outline the obligations and expectations of the various parties, as well as details of where the intellectual property rights lie, are signed. With the exception of Merck, the intellectual property remains with the external researcher.

The results of the screening are then either shared with or transferred to the external researcher — again, except for the Merck programme, which provides no biological information unless an agreement to collaborate further is signed.

Each company then has a different procedure if it decides that it wants to develop the compound. Eli Lilly asks only for the right of first approach if a compound shows promise. But the company has no legal recourse if a researcher takes a molecule elsewhere following the screening. “We hope that when somebody sends a compound to us, that we have the first right of approach,” says Marta Piñeiro-Núñez, head of Lilly's Open Innovation Drug Discovery (OIDD) platform, based in Indianapolis, Indiana. “We hope that by working with them and engaging in the research with them that we become a preferred partner.” It's a similar story for Leo and AstraZeneca.

Merck's approach is more proprietorial. The standard agreement, signed when a compound is accepted for screening as part of its Open Compound Sourcing initiative, sets out that the arrangement must remain confidential and gives the company an exclusive three-year licence to develop the compound. Merck pays €200 (US$228) for the privilege, with a promise to pay a further €5,000 if the company files a patent application that includes the compound within 5 years. In all cases, the agreements at this point can become complex and, unlike the simpler no-intellectual-property approach of CO-ADD, require significant legal insight. “These things can take a long time,” says Niclas Nilsson, head of open innovation at Leo Pharma in Ballerup, Denmark, “which only further highlights the need for speedier and less-difficult processes to explore new collaborations”. These legal hurdles may be one reason why the response to some of the corporate screening platforms has been relatively weak compared with the CO-ADD programme. Merck has signed only around 30 contracts to screen small groups of compounds since the initiative's 2011 launch. In the seven years that the less prescriptive Lilly OIDD platform has been operating, the company has screened over 40,000 molecules selected from more than 400,000 submissions, and at the start of 2015 had 345 groups participating. The screenings have been done at a consistent rate of about 500 compounds per month, but according to Piñeiro-Núñez, the company has the capacity for thousands more. Meanwhile, Leo has screened just 150 compounds submitted by 15 partners in the first 12 months of its Open Innovation programme — a long way from the roughly 26,000 that have been assayed by Cooper and his CO-ADD colleagues during that time. Cooper attributes the variation to concern about the motivation behind the programmes. “People are probably more likely to engage with a not-for-profit,” he says. “They feel more comfortable that we're doing it for the right reasons.” Aware of the challenge of fostering participation, Nilsson set out to devise a screening programme that would be attractive to both small industry and academia. Nilsson says, “Our approach was based on interviews with biotech and university researchers to determine how low the bar would have to be set in order to persuade them to take part and share their creations”. He then negotiated with the company's legal and business development arms to determine how barriers to submission could be reduced. Researcher feedback drove Nilsson's team to further adapt the conditions of the programme. Initially, they arranged to share the data with participants, giving both parties rights of access and ownership. But it quickly became clear that this would be difficult for small biotechs to agree to — if they don't own the data, they can't use them as leverage when negotiating with others. “So now we transfer the ownership of the data back to the external partner. It means that they can increase the value of their assets by using our services and then they can walk away if they want to. That is, of course, our risk,” says Nilsson. Although the numbers are small, Nilsson says that the fledgling programme has already generated positive responses and the company is in the process of negotiating with one contributor on a future collaboration in relation to a treatment for the skin condition psoriasis. “They submitted compounds that are targeting new proteins that we didn't even know existed. So, not only are they providing new chemistry, but they also open up a new science that we didn't know could develop,” Nilsson says. “That is exactly what we're looking for with the open innovation — something new that we couldn't predict ourselves.” Promising progress It is too early to see the fruits of this open-innovation approach in the marketplace. But there are signs that this increased engagement with researchers and their molecules is having an effect. Doug Frantz, a chemist at the University of Texas at San Antonio, has been contributing to the Eli Lilly compound-screening programme for five years. His laboratory has submitted around 500 compounds and recently became one of just two OIDD contributors in the programme's history to reach an agreement to progress a molecule to trials in animals. In the context of drug development, this is more than one-third of the way towards clinical use. The progress of the potential chronic-pain drug (an antagonist for a form of sodium channel called NaV 1.7) is a milestone for him as well as for Lilly. “It's just a fantastic opportunity,” Frantz says, “to be able to explore the biological activity of my molecules, rather than just making them and sticking them in the freezer and publishing a paper. I can take that data and I can take the compounds and publish or put it into a grant application if I want.” Ana Cristina Parra Rivera, a student of Doug Frantz at the University of Texas, works remotely for Eli Lilly. Credit: Robynne Neff Frantz and his students can keep tabs on the assays that their compounds are undergoing and any results through a secure online account. Many of their compounds have shown activity in these assays, and there are a few particularly promising compounds that they are considering adapting to meet Lilly's threshold for further collaboration. As well as the NaV 1.7 antagonist, they have a second contract with Lilly that relates to a possible treatment for schizophrenia (an antagonist of the nervous-system receptor mGluR2). This has involved one of Frantz's students working over the Internet to create compound analogues in Lilly's automated-synthesis lab. The University of Texas at San Antonio is one of 255 universities that are contributing compounds for screening with Lilly. Piñeiro-Núñez says that, initially, “people would get up on their seats and say that we were crazy”, but now many in the academic community as well as industry are embracing external compound screening. The results of the screening of compounds produced by external researchers by Lilly have been comparable with, and sometimes better than, the 1–5% hit rate of internally produced molecules. “Even if the molecules don't advance, they introduce starting points and they contribute to each project,” Piñeiro-Núñez says. The screening platforms are also a way to save money — AstraZeneca has put the cost of the internal production of the 210,000 compounds that it received from Sanofi last year at around$25 million.

After about a year of screening, the CO-ADD project has uncovered 128 hitherto unknown potential antibiotics that have passed tests for cell toxicity. On the basis of these results, the group is already talking to health-care-research funders, including the US-based Bill & Melinda Gates Foundation and the Wellcome Trust, about extending the programme. By the end of 2020, Cooper hopes to have screened 1 million compounds. “When we do these things collectively and collegially, they're of incredible value to society,” he says. “This is like a molecule bank — we want to use this to screen for new compounds against malaria, against tuberculosis, against other pathogens and threats to world health.”

CO-ADD seems to be well on its way to sourcing those 1 million molecules, but it remains very early days for this kind of open-innovation approach. Already, there is new engagement between big pharma and academia, and more diversity in compound collections and novel research leads. But only time will tell if drug developers can capitalize on this early promise. With only one compound thought to have made it to animal trials, these programmes still have a long way to go if society is to feel their true effect.Footnote 1