1 Introduction

Public–private partnerships (PPPs) in drug discovery are becoming more common. Many of them focus on accessing novel biology from academia, because this is currently a significant bottleneck for pharmaceutical companies who have historically focused the largest part of their resources on chemistry, biochemical screening technologies, and clinical trials. The unacceptably high rates of failure in late-stage clinical trials due to lack of drug efficacy (clinical Phases II and III) is evidence for a lack of biological wisdom in the sector. Conversely, academia has amassed significant knowledge and expertise into complex biological models and pathways that are pertinent to innovation but not necessarily accessible or useful to the pharmaceutical industry. Consequently, many pharmaceutical companies are currently accessing biological innovation, expertise, knowledge, and education from academia by setting up bilateral collaborations as an integral part of their early-stage drug discovery strategy. However, valuable knowledge from academia can be difficult to translate into meaningful assets, due to a difference in experimental ethos and a misalignment of purpose, thus resulting in a translational gap.

By engaging in larger consortia, rather than in one-to-one collaborations, pharma can benefit from complementary inputs from partners that have different abilities and perspectives. These grander schemes that often include multisector stakeholders have the advantage of creating a common ethos and purpose, which can help bridge the translational gap. These coalitions often provide a broader platform from which to internalize novel assets, offer avenues to break down silo mentality, and reduce the chances of the collaboration being driven purely by academic motivators.

This article aims to present insights into biology-focused PPPs between academia and industry that are at least part funded by the pharma industry. After exposing what drives industry and academia to engage in PPPs, we present three different exemplars of PPPs that we have been engaged in, which had very different aims, partners, and budgets, namely the MSD-Scottish Life Sciences Fund (MSD-SLSF), the Phenomics Discovery Initiative (PDi), and the European Lead Factory (ELF).

The exemplars are based in Europe, although the pharmaceutical companies involved are multinationals, and all partners have, in any case, an international outlook. SMEs mentioned in this article (who are also industrial partners) are part of the PPP ecosystem but are only considered as part of consortia involving large pharmaceutical companies, not in any bilateral collaborations with academia.

By offering our experience from an academic perspective, we hope to educate and advise pharmaceutical professionals who are considering building or participating in a biology-focused PPP. Our insights include an understanding of what motivates the different stakeholders, and some lessons learnt, often the hard way, on what can enhance or degrade a collaboration. Hopefully, this practical approach will help professionals get a head start when engaging with the exciting and productive collaborative adventures that are PPPs.

2 Why Should Industry Seek New Biology from Academia?

Between its instigation in 1827 and 2018, the FDA has approved around 1666 New Molecular Entities (NMEs) as therapeutics (89). The annual rate of NME approvals has varied greatly over the years. After what was thought of as the heydays of drug approval in the late 1990s (on average >36 drugs approved/year) (90), the 2000s saw a sharp reduction in approvals (on average ∼22 drugs approved/year), despite ever-increasing levels of investment from the industry. In fact, from the 1990s to the 2010s, the cost of bringing a drug to market increased from $320 million to over $2 billion (91). Also, the total global R&D spend by the industry (pharma and biotech) increased from $12 billion in the 1990s to $108 billion in 2006 and continuing to rise to $141 billion in 2015 (92). The major upheavals in the pharma industry including mergers, corporate restructuring, and pipeline streamlining were rife in the 2000s, and the disruption caused to the industry could be considered a major factor in the drop of drug approval rates around that time.

Although still hugely variable on a year-by-year basis, since 2010, the drug approval figures have slowly improved and attained, in 2018, the all-time record of 59 drugs. This encouraging observation is also supported by a trend toward improved clinical success rates, which had been in steady decline between 1997 and 2010 (93). Although clinical success rates have remained at low levels in Phase I (change of 6–7% success rate from Phase I to launch from 2010 to 2017), they are coupled with slightly increased success rates in Phase II (increase from 11% to 15% success rate from Phase II to launch from 2010 to 2017) and Phase III (increase from 49% to 62% success rate from Phase III to launch from 2010 to 2017), suggesting that pharma companies, through their business improvement strategies, may have successfully focused on improving the quality of their portfolio and are managing to fail drugs early before the cost of clinical trials becomes too exorbitant. Higher success in the clinic, coupled with more refined portfolios, has predictably increased drug approval rates.

However, realistically, success rates are still very poor, considering that fewer than 1 in 10 000 early translational programs successfully conclude in the launch of an FDA-approved drug, a process that still costs in excess of $2 billion and takes over 11 years to get to market (94). Even when a candidate enters the clinical pipeline (only 0.14% of translational programs achieve this), the probability that it will make it to an approved drug is at around 7% (just below 1 in 10, or put another way, a 93% failure rate). From a business perspective, these are unacceptably high failure rates, considering the extreme levels of investment required to be a player in this field. In fact, failures in the later stages of clinical trials have the greatest impact on the expected returns of any investment in pharma. Also, 7 out of 10 drug products that go to market do not recover the average cost of development. This means that the risk versus returns profile of the pharma industry makes it look like a very poor business model, and with this in mind, a pharma company's value can only come from a carefully considered portfolio that spreads the investment risk (95).

Until recently, the industry's business model has been predicated on launching blockbuster drugs such as Atorvastatin. Sold as Lipitor, Atorvastatin reduces de novo synthesis of cholesterol, is a competitive inhibitor of HMG-CoA reductase, and was launched by Pfizer in 1996. Atorvastatin was the world's best-selling drug, realizing $125 billion in sales over 15 years, and came off patent in the United States in 2011. Similarly, Adalimumab (an antibody which targets TNF alpha and is sold as Humira) was launched in 2003 by Abbvie to treat rheumatoid arthritis and was the best-selling drug in 2018 with sales of $19.9 billion. US exclusivity extends until 2023.

It is these phenomenally profitable blockbuster drugs that have been supporting the high-failure rates and poor profitability in the rest of pharma's R&D portfolios. Unfortunately, modern-day chronic diseases such as dementia and diabetes are more complex to address therapeutically than was reducing cholesterol synthesis with Lipitor. For complex, multifactor diseases, patient variability has a far greater impact on drug discovery outcomes and implies that achieving a blockbuster such as Lipitor is becoming less likely.

Historically, drug companies were very much built around the knowledge, skills, and logistics of creating large diverse chemical compound collections, and the medicinal chemistry expertise required to morph them into bioactive drugs.

Including biology in early-stage drug discovery is clearly necessary, but due to the adoption of reductionist high-throughput approaches in screening, it has mostly been reduced to biochemical rather than biological assays (reductionist and target-based approaches) (96, 97). This means that the more complex biological pathways that are linked to drug efficacy have been pushed out to later stages of the discovery process.

As a consequence of the spate of pharmaceutical sector mergers and acquisitions in the 2000s, several thousand industry R&D jobs were axed between 2005 and 2012 in some of the major pharmaceutical companies such as Pfizer, Roche, GSK, MSD, Johnson & Johnson, AstraZeneca, and Sanofi (98). This has resulted in reduced in-house R&D activities and an increased focus of company resources on in-licensing and supporting clinical trials. Pharmaceutical companies have in effect been revising their strategies away from trying to achieve end-to-end drug discovery and development, to focus on the parts of the process they have historically been good at, and as a result collaborating on other parts of the process they are not so good at, such as innovation. In-licensing, open innovation activities, and outsourcing the more standard R&D processes to CROs are now much more common.

Pharmaceutical companies have focused their expertise on the ability to manage and fund new therapeutic chemical entities through expensive and lengthy clinical trials to achieve regulatory approval. However, at the same time, regulatory hurdles have become more onerous, meaning that proof of clinical efficacy and added clinical value are harder to demonstrate, making approval harder.

By confounding all the changes that have occurred in the pharma industry (summarized in Table 3), it is becoming apparent that innovation, especially from biological sciences, is required to redress the pharma innovation and productivity gap and to deal with the challenges of medicine in an aging population. Over the past 15 years or so, industry has noticed this gap and is working on opportunities to plug it to be more competitive.

One of the opportunities is to slowly revise methods of interaction with academia by promoting more productive exchanges, collaborations, and partnerships, including PPPs. Ultimately, industry views these partnerships from the standpoint of how they are able to influence the bottom line of the company.

3 Why Would Academia Want To Collaborate With Industry?

Before trying to engage with academia, it is important to understand what would motivate them to engage.

Academia is good at educating, innovating, and generating new knowledge, regardless of its utility or tractability for translation into industry applications and products. This is because success in academia is measured differently to that in industry. Quality of university education is monitored by an independent panel of experts, including academics, students, and employer representatives. Assessments are based on a set of metrics using continuation rates, student satisfaction, and employment outcomes for undergraduate students. On the other hand, university research and innovation quality are measured according to gained grant income and quantity and quality of peer-reviewed journal publications. None of these metrics are particularly conducive to creating an environment that generates solutions to industry's immediate drug discovery needs with regard to biological innovation. This is made worse, because at the outset there has been no real incentive for academics to generate results that are either specifically relevant to drug discovery's R&D safety and efficacy challenges or reproducible enough to be seamlessly translated into an industrial environment. The pharmaceutical industry often suffers from an inability to reproduce published research methods and findings. It is rather alarming that according to studies by pharmaceutical companies Bayer and Amgen, when industry labs try to reproduce academic results they are unsuccessful approximately 80% of the time (99).

Viewed by the average faculty member, a pharmaceutical corporation does not necessarily have the hallmarks of a trusted partner. There is concern from many academics that collaborating companies may try and seize any intellectual property (IP) and profits that are issued from any collaborative efforts. There is also a worry that industry collaborators will encourage a delay in publications, which are critical to academic careers. It has been argued that collaborating with industry actually reduces publication rates because industry is so keen to disclose findings in patents or the keep critical know-how secret (100). Corporate pressures could also divert scientific efforts from fundamental studies, toward applied research, which has less publication value.

However, recently, a cultural shift has occurred in Europe, resulting in universities being more engaged in impact-driven outputs, because these are now factored in as an additional measure of research quality. Impact from interactions with industry can be defined as a collaboration, knowledge-exchange activity, or technology transfer with industry that contributes toward a product or business process that measurably benefits the public or the economy. For biology-focused university departments, this has resulted in a will to translate knowledge from the bench to the bedside, as an obvious demonstrator of human and financial impact. Achieving impact, which can be described in case studies as well as publications, now comes with financial rewards for universities when they apply for core funding from the state.

Beyond the impact agenda, industry collaborations may actually benefit academic publication records. A recent study at the London School of Business has postulated that academics with industry collaborators actually published more follow-on papers than those without the relationships (101), and that publications with a corporate coauthor were more widely publicized when measured by altmetric attention score (the altmetric attention score tracks the number of discussions around a published paper, which almost doubled for life sciences papers with a corporate coauthor), increased the academics' scholarly impact when measured by h-factor, and was associated with a 26% increase in citation rates (100).

There are also other incentives for academics to collaborate with industry, such as generating new and untapped sources of income, accessing uncommercialized technologies, expanding researcher networks, generating opportunities for staff and knowledge exchanges, and using collaboration outputs as a base for commercially valuable IP generation (patent filing and licensing) and commercialization (spinout generation).

Through these relationships, faculty can keep up to date with trends in industry, students are made ready for the workplace, and university resources are often updated. The staff that is engaged can also expand horizons and develop new career skills that are relevant to commercialization and industry perspectives.

So, in summary, there are newly found reasons for university biologists to collaborate more with industry. However, it is important to note that the mechanisms, knowledge, and guidelines to do it well are not yet fully established in academic institutions. It is still very much an activity that is carried out by motivated individuals who work out how to do it themselves rather than a strategy that departments adopt seriously and comprehensively to boost their impact statements and funding goals. Tech transfer offices do help, but they generally respond to a strategy developed by the scientist and only deal with the administrative aspects of setting up the collaboration rather than providing real guidance on strategy. It is worth being aware of this situation when negotiating a collaboration with academia.

4 What Can Academia Contribute To Industry?

The majority of innovation occurs outside of the pharmaceutical industry. The total budgets spent by academic institutions on research and innovation are higher than that spent by pharmaceutical institutions by a long shot, making them one of the main engines of innovation (Table 4). Academic freedom also promotes innovation and creativity, due to the lack of boundaries put on research goals.

One of academia's missions is to acquire and disseminate knowledge as an end in itself. Faculties tend to emphasize theory over practice, which is in direct contradiction with industry. This kind of curiosity-driven biological research has the potential to deliver novel drug targets and important enabling chemical tools.

University life scientists are defined by the specific and detailed knowledge they hold in a very narrow field of biology. They will be world-class experts on the pathway, molecule, or biological model they have spent years developing. This potentially makes each scientist unique, with a potential huge value to a pharma wanting to work in their field. The cellular assays, animal models, diagnostic tools, and know-how around the pathways that define them can be of immense value in an early-stage drug discovery program. Finding the right Key Opinion Leader (KOL) can be the defining feature of a successful early-stage drug discovery program.

University clinicians also provide industry with access to clinical samples and patients, which are useful in in vitro experiments, but also later on in clinical trials. Sourcing relevant patient material in an efficient and ethical way can be difficult, expensive, and time consuming. Pairing up with the right clinician can ease all of these bottlenecks.

Ultimately, the greatest benefit for both industry and academia is the ability to engage in academic research that benefits both the society and the economy.

5 Has Academia Historically Contributed to Industry Success in Drug Discovery?

This recent contraction of in-house pharmaceutical R&D activities, combined with an amplified need for innovation, has meant that collaboration between academia and industry in the field of early-stage drug discovery should be a win–win situation. On paper, both parties are almost perfectly compatible, providing complementary skills and expertise.

Looking at the motivators and current status in academia and industry, we believe it is becoming increasingly clear that there is a real need from the pharma industry to collaborate more systematically with academia, to access biological knowledge and innovation that is useful in the early stages of the drug discovery process. However, organizations that are innovating, be they academic or industrial, only choose to do so if they feel they have to. Realistically, for IP ownership reasons, they prefer to keep as much in-house as they can.

So is there any hard evidence that academic involvement in early-stage drug discovery confers any advantages to industry with regard to increased rates of drug approval?

Pharma has in fact always benefitted from knowledge and innovation coming from academia. Bilateral – vertical one-to-one – collaborations have been quite common, and the rate of collaboration creation has been on the rise in the past 10–15 years. This is evidenced by an increased percentage of research projects supported by industry and an increased financial investment of university research by pharmaceutical companies. According to Nature Index 2017, the number of academic–industry collaborations more than doubled from 12 672 in 2012 to 25 962 in 2016, and half of those 2016 collaborations were in the life sciences (100). In 2018, industry funded about 7% of university research, which is about double that of 20 years ago (1998).

It has been proposed that increased pharmaceutical productivity can be achieved by transforming the way R&D is conducted from a fully integrated approach to one that was more highly networked, partnered, and leveraged (94). This networked approach has been predicted to improve productivity by improving the quality of the discovery pipeline in a cost-effective manner. Academia is part of this networking equation and can therefore play a role in improving productivity in the pharmaceutical industry.

Since the 2000s pharma has actually changed the nature of its overall engagement with academia, in the form of longer term and more profound collaborations such as PPPs. This shift, toward deeper and more complex relationships, has correlated with an increase in the number of FDA-approved drugs, suggesting that the collaborations have had a positive impact on drug discovery (102).

A general observation is that at every stage of the drug discovery process, projects that have benefitted from partnering (or any form: academia, biotech, etc.) deliver improved clinical success rates (93, 103), further supporting the notion that collaboration is an advantage in drug discovery.

Between 1998 and 2007, 58% of the 252 FDA drug approvals were first reported by a pharmaceutical company alone, 24% came from a university, and 18% came from biotechnology companies. In 2015, 55% of all 1453 approved drugs were first reported by academia, and finally in 2018, of the 210 NMEs approved between 2010 and 2016, all of them had an initial academic input (102). These figures clearly indicate that academic innovation is increasing its impact on drug approvals.

It is also important to note that many disruptive technologies that over the years have improved research productivity, such as PCR amplification, CRISPR/Cas9 gene editing, Western blotting, phage display libraries, and CAR-T immunotherapy, to name but a few, all originated in universities.

6 Different Types of Collaborations between Industry and Academia for Early-Stage Drug Discovery

There are many ways in which academia typically interacts with industry. The list in Table 5 attempts to detail different forms of engagement along with associated activities, funding, and outputs. Currently, the most common focus of industry–academia collaborations is to achieve impacts on student education (funding PhDs, for example) or develop applied research and technology transfer activities. Other more complex forms of collaboration, such as PPPs and joint research labs, are slowly gaining in popularity but are still in the minority.

What is evident is that each form of collaboration presents a different model, with different modes of funding, different governance structures, different levels of engagement, different goals, and different outputs.

Even collaborations that are set up meticulously well, with all five dimensions defined and agreed in advance, are likely to develop sources of tension, regardless of the nature of the collaboration. Partners may clash over IP or decisions about how much a company will pay to license a patent from a university lab. Liabilities are also a source of tension: who will be legally responsible for what, and under what circumstances will partners be sued. Each side is interested in protecting its interests from damages resulting from lawsuits. Such tensions can significantly delay a project's progress and need to be dealt with seriously as and when they emerge in the partnership.

7 Biology-Focused PPPs That Aid Education and Innovation in Early-Stage Drug Discovery

Drug discovery PPPs integrate subject matter experts, drug development experts, and other relevant stakeholders under a common goal to deliver novel findings that are relevant to developing new drugs. To meet their objectives, they should also be supplied with enough resources to translate these novel findings far enough down the pipeline to be seen as potential starting points for fully-fledged drug discovery programs. Underfunding these partnerships or not providing them with the correct kind of support will only result in outputs that are too early stage to be of interest to industry and therefore not ultimately translated.

The collective power and motivation of PPPs reside in the shared vision of improving human health by realizing the potential of the collaborative research. However, success requires the building of trust, open-minded debate, and incentives appropriate to each situation. Most issues in PPPs are human and come from poor communication, preparation, and expectations. It is important to have an insight into your partners' vulnerabilities to anticipate what challenges they may have. Something that is easy for one partner may be challenging for another.

The next sections of this article will aim to describe three different drug discovery PPPs that we were involved in over the past seven years, including defining their vision, establishing their governance, and administrating their delivery. Data in Table 6 summarize the partners, high-level aims, budgets, and timelines for each consortium. For each project we will also aim to describe the PPP requirements, how and why they were set up, which partners were required, how to justify the budgets and resources, and what lessons were learnt through each partnership.

7.2 Phenomics Discovery Initiative (PDi)

7.2.1 Background

PDi was a PPP between industrial pharmaceutical companies and the National Phenotypic Screening Centre (NPSC) that launched operations in 2016. NPSC is an academic phenomics center with laboratories in the Universities of Dundee, Edinburgh, and Oxford.

PDi was a PPP addressing industry's biological research priorities for phenotypic drug discovery. Each industry partner paid an annual fee (£650K) to NPSC to access novel disease-relevant phenotypic assays that were recruited globally from academia, developed to industry standards for high-content screening and then screened against two compound libraries (annotated and diverse). Assays were crowdsourced globally from academics (assay owners) who put their assays forward for selection by a panel of experts.

Using the data from the screens that were selected to be in the portfolio, PDi aimed to deliver its partners with validated assays that were automatically licensed to the main consortium partners who could exploit them privately and validated hits and hit structures that are owned by the academic but formed part of the shared consortium data that could be used to inform downstream strategies.

7.2.1.1 Gap Addressed by the Consortium

The PDi addressed several gaps that are pertinent to the current lack of productivity in the pharmaceutical industry. Through smart crowdsourcing the consortium had a sophisticated mechanism for efficiently accessing near-patient physiologically relevant biology, which is one of the bottlenecks in the industry. PDi also allowed this biology to be developed in industry standard assays, addressing the reproducibility crisis in translational efforts. Focusing on phenotypic rather than target-based approaches means that the hits that were obtained from screening were more likely to be successful in the later stages of the drug discovery process due to the chances of efficacy being higher.

7.2.2 Scientific Focus

PDi sought to identify, develop, screen, and validate innovative phenotypic assays that were relevant to human disease. PDi phenotypic assays were crowdsourced globally from academia and small businesses in the form of project proposals made by academics, clinicians, and SME employees. These were submitted to the NPSC online recruitment platform (Figure 1). Best in class-predictive assays were sought out using top-down and bottom-up approaches, accompanied by a ruthless selection mechanism, which took into account the quality of the model (right cell, right stimulus, and right readout), scientific excellence, disease relevance, market relevance, and need from industry (Figure 1).

All applications were made online and submitted to a tailored recruitment platform, allowing them to be accessed and tracked globally. Projects that were selected were fully funded through assay development and phenotypic screening at NPSC against a ∼1000-compound annotated small-molecule collection and an ∼80 000-compound industry small-molecule collection (Janssen's Jumpstarter library) as shown in Figure 2.

The portfolio of projects was selected by a scientific committee made up of experts in the field, who at each yearly project recruitment round iteratively select projects to best meet industry needs. The portfolio contained a range of assays from different disease areas: Oncology (6), Neurology (2), Viral respiratory (2), Immunology and Inflammation (3), and Metabolic (2).

7.2.3 Budget

The consortium was initially funded with a £2.275M cash contribution from Janssen as well as supported with significant in-kind and leveraged assets such as:

• Access to a ∼1000 cpd annotated compound set from Janssen
• Access to Janssen's 80 000 Jumpstarter diversity set (both libraries combining to an estimated value of around £8M)
• Leverage of SFC's £8M Capital Expenditure investment, which founded the NPSC phenotypic screening facilities in 2015 (Dundee and Oxford laboratories).

7.2.4 Governance

PDi's governance structures were at three levels:

• At a board level – The PDi Board aimed to meet industry needs by setting the consortium strategy and approving and funding a phenotypic screening portfolio put forward by the PDi Scientific Committee. The Board was made up of industry and academia experts.
• At a scientific level – a scientific committee, made up of representatives from industry and academia, reviewed incoming crowdsourced projects and recommends a portfolio of screening projects to the board. Projects were selected on a set of defined criteria such as scientific excellence, disease relevance, tractability, potential for translation, and quality of the research team. They also appointed suitable project teams to manage the research projects and scrutinize results and outputs on a day-to-day basis.
• On a project-by-project basis – a project team made up of a team leader from academia and one from industry who established and managed research plans that were specific to the project.

NPSC's executive team managed the screening operations at the three university sites carrying out the assay development and screening operations.

7.2.5 Outputs and Portfolio

Multiparametric, near-patient phenotypic assays were generally complex and challenging to develop into high-throughput format for screening. Many of the cellular models contain human-derived cell types that are hard to source, expensive, and difficult to culture in vitro. Furthermore, the assays often involve building multicellular models or organoids, which require patience and skill to miniaturize for screening formats. These pathophysiology-relevant assays that were painstakingly developed at NPSC were one of the main assets of the PDi. Combined with the novel research aspect that comes from academia, their disease relevance made them attractive to industry. The assays represent novel biology that is hard to come by and is often not reproducible enough if taken straight form the academic lab into industry.

PDi's portfolio, shown in Figure 3, was made up of 15 programs, 4 of which have been completed by the NPSC and are moving into downstream screening cascades. Other assays are at various stages of assay development and screening.

For academia, PDi provides a translational platform for knowledge and education, because all the innovative outputs are validated to industry standards, making them more valuable to industry. For industry, reproducible assays that explore new biology can easily be turned into proprietary material (Table 8).

Table 8. Outputs of the PDi to Academia, Industry, and the NPSC

PDi consortium activities and assets	Outputs for academia	Outputs for industry
Validated assays with comprehensive SOP	Validated translational asset: An assay that is built to industry standards, which can be used as a primary assay in screening cascades	Complex near-patient assay that can be screened on private collections to generate private IP
Validated annotated hit list with 10pt dose–response curves	Insight into the biology at play in the near-patient model. Hits that inform on possible deconvolution strategies	Insight into the biology at play in the near-patient model. Hits that inform on possible deconvolution strategies.
Validated hits from diversity set with 10pt dose–response curves	Validated hits that can be used as staring points for a drug discovery project	Ideas for further drug discovery strategies
Access to raw and processed data	Research material	Research material that can be cross-referenced against other programs
Staff exchanges	Training of staff on high-throughput robotics systems	n/a

7.2.6 Intellectual Property

The IP model for this program is of particular interest, as it is simple and promotes the creation of ongoing relationships and collaborations to take projects forward into development.

The industry-standard data that is generated by PDi (validated assay, SOP, and structures of up to 80 phenotypic hits) is owned by the academic/SME applicants who are free to publish these results and use the hits as starting points for novel drug discovery programs. This is possible because the compound collection used for screening is made up of commercially available compounds.

Assay development and screening results data are also shared precompetitively between all consortium members. Industry participants and the NPSC are free to use the validated assays in-house on proprietary compound sets in the so-called private screens, and this is where the industry partners generate IP and benefit from the consortium. This means there is a clear distinction between the consortium-owned results, which are automatically licensed for use to all partners, and the potential for privately generated IP which is private to the party that generates it.

7.2.7 Lessons Learnt from PDi

7.2.7.1 Governance

PDi taught us that having an internal industry champion who truly owns the program is really important. Having an industry coordinator who is simply assigned to be in charge of the program is not sufficient.

7.2.7.2 Administration

Dealing with corporate restructuring in pharma companies is detrimental to PPPs, as the change of staff can destabilize the consortium, especially when trying to grow it with the entry of new partners. Accelerating the decision cycles and making them more discrete and iterative can mitigate the effects of this problem by setting shorter term objectives that can be met within the committee lifetime of staff members or can easily be revised without too much upheaval on the joining of a new member.

7.2.7.3 Mutuality

It is critical to anticipate and mitigate the issues partners may have to preserve the synergistic values of the collaboration. For example, when crowdsourcing programs from the public, having very simple and beneficial T&Cs makes academic uptake easier. With PDi, the clear-cut split between the public assets and the potential for private screens by industry partners meant that academics were happy to engage with the program because they could easily understand the T&Cs. Perception of hurdles is often more important than the hurdles themselves.

Helping academics with assay development is really important in making transitional projects a success. Most academics neither have the skills nor the equipment to make assays robust or reproducible enough for high-throughput operations. By seeking the best scientific ideas and then putting resources into making them fit for industrial grade screening is a very productive way to translate science, rather than trying to select projects that are already robust enough for industrial applications.

In PDi, it was important to keep individual industrial participants sharp with their needs, interests, and expectations from the consortium, as this is directly linked to achieving the success anticipated by the company as a whole. This was difficult to do because for each individual contributor, the PPP probably only represented a small fraction of their workload, which may not be critical to their career. This means that it is important to make sure they understand the synergistic value of the consortium and what they have to do to make it work, regardless of whether it is any benefit to them. Hopefully, it is possible to find good reasons for them to be involved which are indeed beneficial to them.

7.2.7.4 Norms

As with most collaborations, communication is key to achieving success. Regular small updates at every management level of the project are more productive than adopting a top-down managerial approach. This habit will establish trust early on and build an environment where partners are keen to engage and deal with issues as they appear and before they become a treat to the project. However, with project partners dotted around the globe in different time zones, this can be difficult to achieve in a smooth and fruitful manner.

7.2.7.5 Organizational Autonomy

From a biology perspective, the pharmaceutical industry is mostly organized in therapy areas (TAs). Each TA has a director (or VP) who is a very important decision-maker and budget holder. PDi was initially set up and mostly funded through the discovery arm of the pharma partners (screening technologies and chemistry) who are not therapy focused. This meant that defining the disease focus and getting buy-in was difficult. If collaborating on biological assets, involving the Heads of TAs is critical to success, as it is their organizational interests that drive many critical aspects of the project for decision-making and financial support. As the project evolved from being funded through discovery to the TAs, this lack of engagement became a real issue.

7.3 The European Lead Factory (ELF)

7.3.1 Background

The ELF was launched in 2013 and entered its second round of funding in May 2019. It is funded by the EU commission's Framework 7 funding scheme via the Innovative Medicines Initiative (IMI).

The ELF consortium is currently made up of eight EFPIA pharmaceutical companies, nine SMEs, and four academic partners. It is an EU-wide platform for scientists from universities and SMEs to translate novel disease-relevant biochemical (target-based) and phenotypic assays and deliver high-quality, relevant chemical assets, with a view to discovering investable starting points for new therapies. Researchers from academia or biotech would not be able to achieve this rapid progression without ELF's industry standard facilities, financial support, staff expertise, and access to small molecule libraries. The contributing scientists are called program owners.

The program uses an extensive and unique compound library (called the Joint European Compound Library – JECL), which is made up of compounds from different industry and public sources, which are contributed to the program by the compound owners, including:

• 200 000 highly novel small molecules that were crowdsourced and developed in the early stages of ELF from academic and SME partners
• Around 350 000 in-house compounds from the eight EFPIA partners
• Different selections from the library are available for phenotypic screens, including a small collection of annotated compounds and chemical probes.

7.3.1.1 Gap Addressed by the Consortium

ELF bridges an important gap between basic research and drug development. It facilitates the translation of fundamental biological insights and chemical novelty into innovative drug starting points. The setup allows innovators from EU academics and SMEs to have access to state-of-the-art industry-grade facilities, drug discovery expertise, and a top-quality, curated library with over 550 000 unique compounds suitable for screening potential drug targets. These are amenities they would never normally have access to. The JECL compound library in particular is one of a kind, as it includes proprietary compounds from eight different EFPIA companies as well as compounds with truly creative chemistries that were sourced from academia. Screening this unique library against novel targets and novel biological models from academia and industry has a great potential for addressing the lack of productivity in drug discovery due to a lack of innovation. The size and scope of the consortium also bridges an investment gap by providing funders with an attractive proposition for exploitation of the consortium findings. Ultimately, this will result in novel medicines bridging a therapy gap for patients.

7.3.2 Scientific Focus

The eight EFPIA partners that are engaged in ELF will screen around 130 programs using the unique ELF library over five years on programs that are proposed and selected through a program selection committee, made up of members of industry, SMEs, and academia. The program selection committee ensures that the program portfolio is novel, balanced, has therapeutic and commercial potential, is tractable, and has no duplication. As in the first iteration of the program, ELF will crowdsource novel screening ideas such as high-value biological targets and disease-linked phenotypic assays from the European research community (public partners from academia and SMEs). The aim is to source around 50 public screening programs in this way, which are selected like the EFPIA programs according to a strict set of criteria set by a program selection committee made up of members of industry, SME, and academic partners. MMV will also propose five screens, relevant to their mission, to be prosecuted by the public partners, and additional screens (around 20) will be prosecuted and externally funded by other charities and private organizations. Program recruitment efforts are supported by therapeutic area champions and aimed at a range of different disease areas, with a special focus on neglected tropical diseases (Figure 4).

7.3.3 Budget

The ELF has had significant funding delivered over two funding cycles:

• In the first cycle (2013–2018), the consortium was funded with €196M, including an €80M cash contribution from the IMI, €91M provided as in-kind contributions from the participating EFPIA companies, and €25M in a variety of contributions from non-EFPIA participants.
• In the second cycle (2019–2024), the consortium is funded with €36.5M, around €18M is a cash contribution from the EU commission, €0.75M is contributed by the Medicines for Malaria Venture (MMV – a charity focused on finding novel antimalarials), and the rest (around €17.75M) is contributed as in-kind contributions by the eight EFPIA companies.

7.3.4 Governance

The size of the consortium means that the governance structure is extensive and comprises the following main structures (Figure 5):

• A coordinator from the public partners and project leader from the EFPIAs are the key intermediaries between the IMI and the other consortium members and actors
• A project management board supervises the overall progress and strategy of the project
• A project executive – responsible for the operational governance of the consortium activities and overall coordination of the partnership
• Three committees dealing with legal matters and IP, ethics, and program review
• Three theme-led teams dealing with portfolio management, project operations, and ELF sustainability
• A working group that manages the crowdsourced programs
• A general assembly involving representatives from all companies and institutions, to present and review results and strategies on an annual basis.

Most governance structures and bodies have a mix of partners from industrial, SME, and public partners, allowing a balanced approach to manage strategy and operations.

7.3.5 Intellectual Property

The following description of the IP conditions for ELF is at a very high level and is just a rough distillation of a far more complex agreement. IP in ELF mostly resides in the structures of the qualified hit lists (QHLs) that are delivered after screening and triaging the JECL library against a public or EFPIA assay, and also to a lesser extent in the screening results, and assays themselves. Qualified hits have gone through an extensive triaging process, giving them significant value for downstream drug discovery activities. Program assets (JECL compounds) and results generated under the project (assay results, screening results, QHL results, and other results) initially belong to the party that generated them. However, the driving principle behind the rights and obligations of the consortium is that program owners should be granted access to, and be in control of, any results from which they can generate value from their program. This means that result ownership, depending on its importance for generating program value, can be transferred to and/or shared between the party generating the data and the program owner. To facilitate this, compound owners (EFPIA or public) grant program owners with royalty-free research use to relevant program results and any relevant background IP required to execute the project. If an asset is directly exploited by the program owner, there are predefined milestones to compensate the compound owner for direct exploitation and also an obligation to purchase newly synthesized compounds from one of the chemistry SMEs that are engaged in ELF. If program owners decide to partner the program, the ELF EFPIA partners have first right of refusal. If an EFPIA directly exploits a public target or assay, the downstream compensation scheme is negotiable with the program owner (Figure 6).

7.3.6 Current Outputs from ELF

ELF is an ambitious long-term project, which in time is aiming to move from the earliest stages of drug discovery (screening and delivery of hits) into later preclinical stages such as lead and candidate generation. The project has so far accepted 88 screening programs, for which 58 QHLs have been delivered.

The ELF will accumulate many valuable assets as a result of the longevity and scope of the consortium activities including novel chemistry, novel biology, world-class facilities and translational platform, shared knowledge, and IP. These include assets that were delivered in the first round of funding and those that will be delivered in the second. Put together, these outputs provide significant benefits to the consortium partners and correspond to a level of screening activity that might exist in a reasonably sized pharma company. Data in Table 9 shows the activities since ELF's inception in 2013; it has had a number of significant outputs as a result of activities downstream of the screening, which are listed in Table 9.

Table 9. ELF Assets and Benefits to ELF Partners

Type of activity/asset	What has ELF delivered so far?	New goals and benefits to partners
Access to world-class drug discovery infrastructure and expertise	88 accepted target programs, of which 58 have been screened and delivered QHL	Delivery of the facilities, logistics, and expertise for up to 185 new drug discovery programs 50 assays will be sourced from the public and screened by the ELF public screening centers 135 assays will be screened by the EFPIA companies Screening efforts will include phenotypic screening as well as target-based screening
A translational platform delivering QHLs	58 QHLs that are ready for translation	Translation of up to 50 novel biological targets and phenotypes from the European Research community (academia and SMEs)
Novel chemistry	∼200 000 novel chemical compounds sourced from public partners	EFPIA partners and public program owners will have access to a unique compound library made up of a mix of ∼550 000 proprietary industry and academic compounds. The large-scale screening activity will generate a significant portfolio of new innovative and biorelevant hits that have the potential to be progressed up the value chain
New biology	Development of >250 bespoke assays	ELF will source 185 novel targets from industry and EU scientists. Novel insights are expected to emerge from the implementation of phenotypic and high-content screening. These insights can be exploited for drug discovery activities or to further life sciences knowledge
	The delivery of 58 public QHLs
	Resolution of >40 crystal structures of target–compound complexes
New businesses	The generation of two spinout companies: ScandiCure AB and Keapstone Therapeutics, who have both secured funding for their ventures through different sources	ELF is expected to facilitate the creation of more start-ups and business opportunities
New collaborations	Translation of two programs into either preclinical or in vivo trials	The ELF partners will actively build new partnerships that will benefit the public and will strengthen the EU life sciences community
New intellectual property	Ongoing generation from round 1 funding	ELF will continue to support the generation of patents around chemical matter and cellular models
	Filing of one patent to date
Leveraged funding	Ongoing	To expand and extend its activities and services, ELF partners will continuously and actively seek additional financing from a range of different sources
Economic and health impact	n/a as too early	Taking academic and biotech research programs to the next stage will create more value, more jobs, and result in new therapies, an improvement in global health, and a direct effect on the economy
Shared knowledge	>60 peer-reviewed journal papers	ELF will actively pursue publications and dissemination of novel results in journals, at conferences, and by digital means that enhance the available knowledge and provide tool compounds to the scientific community
Education and training	>90 academic postdoctoral fellows trained in industry methods and approaches ELF data has contributed to three PhD theses	ELF will continue to support education and training opportunities
New IT platforms	Two custom-built data management platforms	ELF will continue to implement bespoke IT platforms to address issues of partner confidentiality, rights, and obligations

7.3.7 Lessons Learnt from the ELF

7.3.7.1 Governance

ELF's governance structures generally work well. The size and scope of the consortium mean that the setup is a little large and complex, but this seems inevitable.

7.3.7.2 Administration

When the ELF grant was applied for to the IMI, it was in two halves in the first round of selection: chemistry and biology. In the second round, the two halves of the consortium had to come together into a single project. This caused some issues with the administration, because both parties had processes and structures that were already in place to manage the project, including management entities. These management entities were sometimes at odds with each other. This caused some confusion at the start of the project. With hindsight it may have been more efficient to have a single step in which to form the consortium, which would have included a single management entity.

7.3.7.3 Mutuality

The ELF's vision was to recruit novel biology from the academic and industry communities, and screen it against a unique compound library, in the hope that this kind of shake-up would produce some true innovation. This approach means taking risks, which does not sit well with the pharmaceutical industry. However, with a clear communication of the PPP's visions, and the fact that it was part funded by public funds, we have a position where around 1/3 of the screening programs are indeed really novel and do not fit into more common categories of targets that are screened by the pharmaceutical partners, such as enzymes, receptors, and ion channels. It was only by forcing through the obvious synergies and repackaging them into values that could be understood by each of the partners to account for their different perspectives and objectives that brave decisions could be achieved. It was therefore important to adapt the message to each stakeholder to get the best outcome for the PPP as a whole.

The vision for ELF is to be a platform that is available to scientists from all EU countries. However, some member states, with greater research capacities, are benefitting more from the platform.

7.3.7.4 Norms

With so many partners involved in ELF, establishing trust was a lengthy process. The project had two natural splits: pharmaceutical industry versus public and chemistry versus biology. At the start of the program, there was an “us-and-them” attitude between the four camps, which resulted in trust issues and disagreements on liabilities. It took a massive collaborative effort between a large group of talented lawyers from all the different ELF entities to establish a suitable legal framework that would allow trust to flourish. It is surprising what a good legal team can do! Their detachment from the scientific substrate of the operations allows them to be much clearer on the legal and administrative processes that need to be put in place to have an operational PPP. The complex legal framework actually catalyzed some of the initial innovation in the program by creating one of the IT management platforms that allows confidentiality, rights, and obligations to be respected and monitored in the consortium. This is a tool that everyone relies on and trusts to facilitate the correct flow of data within the partnership.

7.3.7.5 Organizational Autonomy

Some aspects of the partnership were initially too focused on the way pharmaceutical companies regard biological research. This comes from the fact that biochemical reductionist approaches dominate in this sector, and reproducibility is of paramount importance to succeeding in high-throughput screening. This led to very high barriers to entry for project owners when submitting programs for review, as the assay requirements in terms of reproducibility, signal to noise, and coefficients of variation were too high. Most academics are neither trained nor equipped to deliver this kind of data, meaning that they were unlikely to submit a project to ELF. This led to a gap in assay development, where either the ELF screening centers or regional screening hubs had to put more resources into this activity. We believe this is an example where the interests and motivations of one type of partner (pharmaceutical partners) took precedence on that of the consortium as a whole. Fortunately, this issue was addressed early on in the project, and academics are now supported by local screening centers and with local funding agencies to get their assays developed enough to meet ELF recruitment requirements.

8 Discussion and Conclusion

Through the experience of three PPPs, we can extract some common guidelines and points of discussion that are useful to bear in mind when setting up this type of collaboration.

Good governance comes from choosing the right team. Having team members who are seniors is also an advantage, as they will most likely be natural decision-makers and strategists. Giving every partner a voice that is truly heard requires effort from all parties and has to break through different cultural, sector, and personal barriers. As in most ambitious projects, recruiting the best people is always a winning ticket.

In terms of administration, it is critical to have an internal champion in each participating organization who truly owns the program and also has a very good understanding of the admin procedures and chain of command in their organization. Most operational issues in collaborations are due to human failures, so ensuring good communication and resolving conflicts as they arise will enhance the quality of the partnership. Nothing is more important than clarity, coordination, and agreement on the PPPs goals.

The mutuality of the collaboration and its benefits needs to be rational and understood by all parties. There also needs to be an agreement that partner benefits may evolve during the course of the partnership, and any changes should be discussed and agreed by all in a timely manner. The synergistic values of the partnership are often what motivate the human and emotional aspects of partner involvement, so are a strong component for engagement and action. If the individual benefits of each partner are matched with a larger positive societal benefit, then the stars are truly aligned for a successful PPP.

Gaining trust comes with time and is in general poor at the start of the partnership. Establishing norms by being open minded about different cultural and professional approaches to a problem is important at the start of a PPP. With patience and understanding, partners can learn to appreciate the strengths and weaknesses of their colleagues and eventually avoid “us-versus-them” politics.

Each partner suffers from the issues of organizational autonomy, but these can be mitigated if the partners are well chosen to be compatible, meaning that their involvement in the project leads to outputs that they can easily defend with their line manager or board. They should also ideally be allowed to make decisions on behalf of their organization, yielding smooth and efficient engagements.

1 The State of the Transfer of Japanese University Research Findings to Industry

Amidst discussions of pharmaceutical industry challenges such as expiring patents for compounds on the market and depleting novel drug targets, major pharmaceutical companies continue to appeal for research collaboration and are actively working to partner with external organizations, or “open innovation,” including academia.

The Japanese government equally continues to promote its project aimed at strengthening the links between the latest findings in life science and pharmaceutical industries. The Ministry of Education, Culture, Sports, Science, and Technology (MEXT); Ministry of Health, Labour and Welfare; and Ministry of Economy, Trade and Industry have begun initiating unique programs to this cause. However, as a result of overlapping policies, the Japanese Agency for Medical Research and Development (AMED) was established in 2015 to intensify the implementation of policies for medical-related research and development and to partner on increasing their efficiency.

There are many notable successes documenting academia-initiated drug discovery. Pharmaceutical development cannot be achieved solely depending on academia; hence, universities have been cooperating with pharmaceutical companies. However, university governance has recently been called into debate. In the past, a typical Japanese university's administration did not strategically direct collaborations with the pharmaceutical industry. In terms of research, universities have long been an aggregate of the so-called private shops, which are managed by each professor. Alternatively, universities only act as shopping malls, where each shopkeeper (professor) talks to familiar companies, calls the relevant people, and markets his/her groups' discoveries based on personal judgment. Subsequently, a good number of the drug discovery leads introduced to these companies tend to be courteously declined, or distant level affiliations are built with the professor, which eventually fade away.

Although there have been successful cases where a handful of university pioneers collaborated with industry and flawlessly developed their research into a pharmaceutical product, the vast majority of university researchers are only amateurs in the field of drug discovery research. Basically, it is a general trend for companies to reject academia-based researchers mostly because the research fails to align with the company's needs and sometimes due to concerns of insufficient or irreproducible data.

After incorporation in 2004, Japanese national universities modified their approach to intellectual property rights, establishing the Technology Licensing Organization (TLO) as a point of contact for academic–industrial collaboration. Despite the recent effort by universities to accumulate knowledge, the necessary support to exploit this innovation are either insufficient or completely absent (in some universities). Consequently, the relationships between individual faculty members and companies remain relevant.

2 Barriers Faced by Basic Researchers Getting into Drug Discovery Research

Excluding coincidental discovery, basic researchers need to conduct compound screening in order to discover small molecules that control biological functions. Having a high degree of originality is of utmost importance in academic research. Hence, academic projects themselves are not so competitive. However, “drug discovery” researchers while drafting their research plan must review the previous research through the lens of existing drugs and disease treatments, including alternate treatment methods not involving the target molecule of interest. If such research exists, the researcher is then tasked with clarifying the superiority of the drug they aim to develop. Basic researchers need to have a sufficient understanding of the different facets of drug discovery outside science, including clinical and business aspects. The required compound screening skills cannot be acquired from university lectures or practical laboratory courses, making basic researchers inexperienced in these skills. Multisample screening is thought of as an easy, repetitive task consisting of investigating whether a small number of compounds have effects. However, the skills and knowledge vary, so guidance from experienced individuals is ideal. Forcefully conducting multisample screening with inadequate preparation may result in large variations, poor reproducibility, and data which cannot be analyzed/interpreted, leading to a waste of time and resources.

Daily compound screening is conducted within pharmaceutical companies, and it is not uncommon for a company to screen hundreds of thousands of samples in a day. High-throughput screening (HTS) technology, in which multiple samples are tested at a high speed, was hailed as a dream technology for discovering drug “hits.” However, with the rate of “hit” discoveries failing to meet up with the resources invested in HTS, many companies are drifting toward areas with higher development efficiencies such as antibody drugs. Despite this drift, the cheap prices of small molecules that control biological target molecules compared to biopharmaceuticals, as well as their applicability to highly versatile oral drugs, make them difficult to replace. Thus, screening cannot simply be stopped, as it is an irreplaceable method for the discovery of small molecules with pharmacological activity.

3 Chemical Screening Support of DDI

The University of Tokyo founded the Chemical Biology Research Initiative (CBRI) in 2006; it was the organization responsible for building a large-scale compound library and compound screening facilities in order to utilize the findings of a large-scale structural biology project started by MEXT in 2002. Switching government grants, the CBRI was then renamed twice: first, the Open Innovation Center for Drug Discovery (OCDD) in 2011 and then the Drug Discovery Initiative (DDI) in 2015 (Figure 7). The organization continues to develop its support activities for researchers based on the research base. Currently, DDI is a member of the Basis for Supporting Innovative Drug Discovery and Life Science Research (BINDS) program of AMED (105).

The DDI provides the same drug discovery research support to academic and industrial researchers. The drug discovery targets may be medicines, agrochemicals, or research reagents. The main condition is that confidential information such as the intended use and assay data be disclosed to the DDI. Chemical samples are equally provided free of charge, although charges covering the costs of sample plates (container fee) and delivery (delivery fee) are paid by the user. The DDI only uses the reported assay data to improve the quality of the compound library by discarding compounds with nonspecific functions, for instance. Recently, assay ready plates, which allow for direct reagent and culture media addition and assay measurements without creating intermediate dilutions of chemical samples, make up a big portion of the total usage. The Echo® liquid handlers, which can dispense solution of chemical samples at the nanoliter level, are used for preparing the assay ready plates (Figure 8).

The DDI's compound library currently maintains and manages approximately 280 000 samples (Figure 9). In 2007, when the plan to build the compound library in Japan was made, only major pharmaceutical companies had HTS compound libraries with over 100 000 samples, and no large-scale library was available to academic researchers in Japan. Such libraries already existed in countries such as the United States and South Korea, with the United States standing out, as it is not uncommon for individual universities to have them. We observed compound libraries at such institutions and at pharmaceutical companies as reference.

Several people have the opinion that “drug discovery cannot be accomplished within academia” or question if anything can be accomplished with such a scale of foundation. Even with 280 000 samples, the DDI's compound library is insignificant compared to the libraries of the major pharmaceutical companies with over 1 million samples. However, considering the time when no academia-targeted drug discovery research foundation existed, we believe that the difference between 280 000 and 1 million is not high. Rather, an increased number of samples in turn increase the management and assay costs, which would make the library less accessible to academia.

The compound library consists of several commercial compounds made by foreign suppliers, selected for their diversity in chemical structure, predicted affinity with biomolecules, and “druglikeness.” In addition, the library contains approximately 10 000 noncommercial compounds from university laboratories with unique chemical structures and approximately 60 000 compounds deposited by pharmaceutical companies. Compounds from pharmaceutical companies typically require a link between the providing company and the research, so the conditions for use of these compounds are maintained in accordance with the intentions of the providing company. If academic researchers agree to those conditions, they may use these samples, but the sample amount and number of chemical structures disclosed are considerably limited. Hence, these samples are only used in research assay systems that have previously been cleared from any issues associated with using generally provided samples.

All samples are prepared in solutions of dimethyl sulfoxide (DMSO) at either 10 or 2 mM, which is miscible with water in all proportions. The quantity needed is dispensed into microplates, and the plates are quickly supplied. Outside pharmaceutical company samples, solid samples are stored for repreparation of solution samples. Generally, solid samples are not provided for bioassays. For research such as animal studies, the supplier information will be disclosed to the researchers to enable self-procurement of the required samples. Academic researchers are often forced to keep their screening to a small scale due to limited funding and manpower. If there is any information such as the three-dimensional structural information of target proteins or information on known active compounds, a computer may narrow down the candidate compounds, but when searching for changes in phenotypes before and after the administration of a compound with no clear target molecule, researchers are forced to use a standard random search. To that end, a Core Library of 9600 samples selected for pilot screening or a validated compound library containing existing medications/samples reported to have pharmacological activity is commonly provided to researchers. Furthermore, for cell-based assays testing, the Whole Core Library is normally not feasible. Thus, we have constructed an intracellular target core library which consists of a subset of 2400 Core-Library samples predicted to relate to membrane permeability. We are also devising alternative diverse subset collections such as the Advanced Core Library (contains 22 400 samples), to enable efficient screen where conducting assays on the full library is impractical.

We also provide support for building assay systems that meet the needs of users. It is necessary to build individual systems suited to various targets for assay methods. Trial and error and a certain amount of experience are required for this. It often takes at least a year to build a suitable assay system, and commonly takes longer than the assay itself. Commercial luminescence kits for measuring kinase activity are expensive, so it is unfeasible to conduct large-scale kinases screening within academia. However, a new fluorescence assay method developed by the DDI has made it extremely low cost to run these analyses (106). A laboratory (Figure 10) is open to users who wish to use it. The laboratory has staff with extensive experience in exploratory research at pharmaceutical companies who are called in to guide academia-linked researchers with no experience in building large-scale compound libraries or conducting HTS. Advice alone is useful, but they equally provide practical guidance which allows beginners and students to ease into their research. Semi-hands-on compound screening technical courses are also offered a few times a year, which hundreds of people attended.

The DDI is a system which provides chemical samples to all researchers. Another popular system is screening centers, where samples are not provided to common researchers, but assays are contracted to screening professionals. This system is common, owing to its high efficiency and yield of good-quality data. However, successful drug discovery research requires perseverance, hard work, and enthusiasm; hence, having researchers conducting their respective goal-oriented research facilitates research progression. Moreover, above anything, this is connected to popularizing know-how in screening and drug discovery research in academia, and training drug discovery research personnel and instructors, which we believe can contribute to a bottom-up approach to drug discovery research in Japan.

Promising drug targets are said to be buried in academia, and pharmaceutical company representatives are eagerly searching for such existing targets; multiple companies publicly appeal for proposal-based targets. Furthermore, AMED conducts target search surveys, and mostly existing targets may have been played out. However, because drug discovery research continues to evolve, it is therefore crucial to continually release new drug discovery ideas, improve the ability to validate them, and cultivate an environment where target development is easier.

Even if screening is contracted to the DDI, from a resource perspective, research themes must be highly scrutinized and prioritized; thus, there are concerns that this may narrow the door to opportunities in drug discovery research. We believe that amidst the fierce global competition in drug discovery research, our limited resources and a small library with commercially available compounds cannot compete with pharmaceutical companies even when conducting similar research on a small scale. Thus, we need to cherish the unrestricted approach of academia with undefined drug discovery targets. Meanwhile, speed is extremely important in drug discovery research. Researchers should consider applying to the iD3 Booster project of AMED, in which research funding is given to promising research themes relating to drug discovery (107), in order to accelerate research through collaborations with companies seeking such opportunities. Given that “transferring to pharmaceutical companies” is one of the goals of academic drug discovery, academia should work to partner innovation with companies that have a strong strategic interest in the target indication.

Based on these points, we believe that a system which provides necessary support to individuals in need, specifically recipients who propose novel drug discovery ideas, while moving forward with screening research, is ideal.

4 Hit-to-Lead Synthesis and ADMET/Physical Property Evaluation Support of DDI

In cases where compound screening results in “hits,” those compounds do not necessarily have satisfactory characteristics, owing to poor activity, nonideal targets, nonselective subtypes, or poor oral absorbability/pharmacokinetics. In other cases, the compound may be metabolized in the body and converted into a different compound or easily excreted. In such cases, improvements need to be planned using hints from the structures of the “hits” from screening, followed by chemical modification and substitutions. It is not recommended to transition to animal studies without evaluating the pharmacokinetic properties of a compound; hence, if disregarded, its results (favorable or not), chemical structure stability, and effective concentration will remain unclear, creating controversies and decreasing the persuasive power of the data.

The DDI's lead exploration unit, which has a team of researchers presently working at pharmaceutical companies, was set up in 2016 with the aim of consulting with biologists or medical researchers not specialized in “hit-to-lead” generation or pharmacokinetics and taking charge of the kinetic parameter measurements. Requests for this support can be made through the BINDS program.

5 Regarding Intellectual Property Rights

When no elements of the DDI's collaborative research are involved in research conducted with DDI-provided general compound samples, the intellectual property of the results belongs to the user. Synthesis support entails collaborative research with academia; hence, the intellectual property rights will be handled accordingly. Findings obtained from the use of public compound samples should be published. However, the nature of the research field may require that the information be retained until the publication of a patent application or a paper with the findings. Users working for companies make up approximately 10% of all users, but company-affiliated users generally seem to have no challenges with following this rule.

The possession of findings that are considered inventions in drug discovery research leads to patent application issues. It usually takes 10 years or more from the discovery in basic research to the actual sale of the drug, including the clinical trials. If a patent expires, other companies could sell the generic versions of the drug, making the patent term of utmost importance in the drug discovery business. It is therefore obvious that patent applications must be submitted as rapidly as possible, given the competitive environment. However, given that the term of research and development after the date of application is subtracted from the effective term of the patent, filing early applications with immature data packages may not always be the best approach to secure exclusivity. Moreover, application details are published a year and a half after the application is filed; hence, competing companies will be notified of the novel technology. The application timing for the technology in question is based on a “strategy” in anticipation of commercial development. Conversely, in academia, publishing papers on research findings and attending conferences are of utmost importance, as opposed to commercializing a pharmaceutical product. This is especially true in cases where graduate students contributed to an invention and have to fulfill their requirements to publish their research thesis to obtain the respective degree. It is in the nature of academic researchers to publish their research findings as soon as possible, to avoid getting scooped if other researchers end up publishing first. The department handling intellectual property at their university requests that the patent application is considered before the research findings are published. Hence, there is no so-called strategy in patent applications based on research findings from universities, leading to substandard patent applications (patent omissions are acceptable), failing to meet company demands, causing a barrier to cooperation between the two parties. To ensure university rights, filling out opportunities for collaborative research with companies after applications is virtually unavoidable, although the patent is initially meant for company activities. If the contents of the patent application cannot be applied in the pharmaceutical industry, then there is no point to it. In this field, it is ideal for inventors and candidate companies for future invention development to be matched in many cases, so that collaborative applications can be based on a business strategy that allows for the progress of collaborative research. It is ideal to understand that the most important things to each party are different and find amicable compromises such as allowing the publication of academic papers only on the “hits” that are not being developed, on a case-to-case basis.

6 Examples of Past Findings

The DDI has provided a total of approximately 25 million samples for over 600 research themes as of March 2019. The target diseases of about half of the research themes were cancers or infectious diseases. At the drug discovery stage, projects are often in their early development, with more than half being in the idea stage.

Major findings obtained using the DDI's compound library have been involved in the discovery of AdipoRon (prospective hypoglycemic agent) (108), the invention of RK-20449 (prospective leukemia medication) (109), the repositioning of a biliary tract carcinoma drug for use in organoids (110), and the discovery of RK-287107 (tankyrase inhibitor and prospective colorectal cancer drug) (111). Additionally, an inhibitor of autotaxin, which is a chronic renal disease target, discovered in collaboration with Dr. Junken Aoki of Tohoku University and Dr. Osamu Nureki of the University of Tokyo, was successfully used as a lead compound by Shionogi & Co., Ltd.

The number of yearly requests for support remains fairly unchanged. In the future, the support of the DDI holds promise in contributing to the creation of new “hits.”

1 Introduction

When transitioning from a 20-year career in the pharmaceutical industry more than a dozen years ago, one of us (SVF) had the opportunity to create an academic drug discovery center at the University of North Carolina at Chapel Hill – the Center for Integrative Chemical Biology and Drug Discovery (CICBDD, http://cicbdd.web.unc.edu). During this transition, collaboration between the CICBDD and the Structural Genomics Consortium (SGC, https://www.thesgc.org, led by AME) contributed greatly to our discovery of many small-molecule chemical probes (112-114). Over the years, this collaboration and our practical experiences in the academic drug discovery ecosystem have informed our thinking about the future of efforts in this area. This article aims to describe the initial impetus for creation of the CICBDD, its successes and challenges, and why we believe that the future role of academic institutions in drug discovery should embrace a fully open model of drug discovery (115).

It should be noted that our discussion will focus on small-molecule drugs not because other approaches, such as antibodies, cell, and gene therapies are not important, or subject to some of the same challenges, but simply because this is the area of our expertise.

2 Problem Statement

Spurred by access to human genome sequence information, tremendous technological advances, and massive increases in funding by the public and philanthropic sectors, scientific progress in the biomedical sciences has accelerated enormously over the past two to three decades. The cellular signal transduction processes and biochemical pathways that enable life are increasingly being defined at the molecular level, and the aberrations that result in disease can be addressed within this rational framework. The sequencing and manipulation of model organism genomes has further explicated the identity and roles of specific proteins in pathophysiologic processes, increasing our ability to preclinically assess pharmacologic targets. We are clearly approaching the “end of the beginning” in our efforts to comprehend the molecular basis of disease.

The pharmaceutical industry has responded to these exciting advances in basic science and technology by increasing their own investment in R&D, with the aim of translating these advances into a revolution in the availability of potent, selective, and safe drugs. However, the rates of approval of completely new medicines have not kept pace with these massive increases in investment. Consequently, pharma is rethinking how to organize its research, particularly in the light of additional business challenges, such as pricing pressures and litigation (116). This reimagined pharma has increased emphasis on external research and also relied on frequent mergers, reorganizations, and reductions in scientific staff across the industry to maintain profits. Indeed, given the rate of organizational change in the industry, and the long timelines associated with drug discovery, it is increasingly difficult for a research project or strategy to bear fruit before it is abandoned (117).

The travails of the drug discovery sector, coupled with pressures on health care systems, are now causing some to rethink the fundaments of the business model. It is increasingly being voiced that the current economic incentives that society has put in place to support the discovery of new medicines, which are based on protection of intellectual property (IP) and prolonged periods of data secrecy, are inherently inefficient; they result in many organizations simultaneously exploring the same molecular target hypotheses, in secret, with failure as the most likely outcome. To support this process, pricing of new medicines must be set at levels needed to recover the costs of these failures, to drive profit expectations, and to turn every new drug, even ones that treat only a few hundred patients (https://thehill.com/policy/healthcare/445451-fda-approves-worlds-most-expensive-drug-at-%242.1M), into blockbusters (annual sales >$1 billion). In our view, this scenario presents an opportunity for public sector drug discovery efforts to contribute, especially to pilot alternative business models.

3 Initial CICBDD Organizational Model

Part of the rationale for pursuing these dual objectives was that they provided the center flexibility in controlling its portfolio, staff requirements, training culture, stakeholder support, and funding options: increasing the resiliency of the center while also providing differing advantages and limitations. Drug discovery projects were responsive collaborations in which the center assessed targets proposed by disease-centered PIs and then engaged to cocreate projects where there was a perceived opportunity to make a unique contribution. To establish an area of scientific excellence, the CICBDD focused on the chemical biology of chromatin regulation and creation of open-access chemical probes, frequently in collaboration with the SGC. For these projects, the center selected its own scientific problems and proactively aligned with the best collaborators to enable success. It should be noted that the objectives of our drug discovery efforts and basic science were often complementary; both led to publications, funding, and training opportunities.

Although our drug discovery and scientific objectives were complementary and often synergistic, they differed significantly in how IP was handled. In order for drug discovery projects to advance into human studies, we anticipated needing to raise significant (∼$5–10 million) funds to enable lead optimization (LO) and preclinical development (toxicology (tox) and chemical manufacturing and controls work (CMC)). Should LO, tox, and CMC all be achieved, we would require even more funding (∼$5–10 million) to pursue initial Phase 1 studies to establish human pharmacokinetics (PK) and tolerability and, only occasionally, evidence of efficacy. The cost for each subsequent stage of clinical development is in turn even higher – perhaps $20–50 million to achieve evidence of efficacy in Phase II, the typical stage for a first readout of a compound's potential as a medicine. This need for capital, especially when focused on high-risk endeavors and when adopting the traditional business model, means that we felt we needed to protect the IP around any new chemical entities in order to attract private investors and offer them a potential for return, should the project lead to a useful therapeutic. We appreciated though that this approach was not without limitations. For example, patent costs can also be a considerable burden as projects progress and national phase filings are required. And we also realized that if any of our discoveries led to a new medicine, the price would be set by downstream investors to maximize profits, not to ensure access for the taxpayers who funded our initial discovery work.

Our chemical probes program adopted a different IP strategy. Since chemical probes are most impactful when shared, we decided they require no IP protection. This strategy eliminates concerns about publications or presentations creating prior art, facilitates transfer of material among scientists, and has been shown to accelerate subsequent scientific discoveries (112).

To be successful in both our missions, the CICBDD recruited staff with industry expertise in protein expression and purification, assay development, computational drug design, and medicinal chemistry. This breadth of expertise means the center has all the resources and knowledge required to optimize small molecules for potency and selectivity in vitro and in cells. We rely on collaborators and outsourced activities for data in animal models and PK/tox. Because the core staff and research faculty of the center provide effort and input to drug discovery and chemical probe projects equally, they increase their opportunities for scientific growth and have greater connection to the overall mission of the center. In addition to research faculty and staff, the CICBDD is home to a number of tenured/tenure-track faculties with research programs in medicinal chemistry, natural products, and chromatin regulation. These faculties enrich the scientific life of the center and benefit from ready availability of expertise and shared equipment that support the center's drug discovery efforts.

A major design principle behind the CICBDD's two-part mission is the capacity it provides to optimize the training experience for students. Our drug discovery projects are well staffed, involve several years of focused effort, and necessarily have to tackle multiobjective optimization of a compound series for potency, selectivity, PK, pharmacology, and tolerability. These projects also face rigorous “go, no-go” decision points as they progress. This team-based science is an ideal training ground for postdoctoral fellows who are aiming for careers as medicinal chemists or other careers that involve multidisciplinary activities. Our typical postdocs in these roles arrive with a Ph.D. in organic synthesis and are able to learn drug design for structure–activity relationship development, biochemistry of their target, PK assay parameter interpretation, and some elements of drug development. In contrast, development of a chemical probe for a novel protein–protein interaction (PPI) involved in chromatin regulation, for example, is an excellent thesis material for graduate students. Students can participate in selecting their probe target with an emphasis more on the biological knowledge to be gained from the use of a probe than on established or potential disease relevance. New technologies are frequently in need of development (119), and because no IP is created on probes, students are able to collaborate with anyone in the world, sharing compounds freely (no MTAs) throughout the discovery process. We learned that avoiding IP also eliminates the potential for conflicts of interest in trying to balance a student's need to publish with timing of patent filing and alleviates this important concern about drug discovery in academics. However, our postdoctoral fellows engaged in traditional drug discovery projects publish or speak externally about their work only coincident with filing or publication of a related patent. Team meetings and a journal club in the center enable students, postdocs, faculty, and staff to learn about the issues specific to drug discovery programs as well as chemical biology-focused probe development strategies, reinforcing our educational mission. Journal club is also open to collaborators, where they frequently give disease-focused seminars, and because this is an internal UNC meeting, we can openly discuss our drug discovery projects and avoid creating a veil of secrecy that conflicts with academic culture.

Funding for the CICBDD, which has an annual operating cost of ∼$3 million, derives from competitive grants and contracts to center principal investigators (PIs) or collaborators (∼70% of total funding) and from institutional support from the University Cancer Research Fund (UCRF ∼30% of total funding, administered by the Lineberger Comprehensive Cancer Center). These funding sources have been supplemented by a gift from a donor and a small amount of indirect cost returns over the years. An active decision was made when creating the center not to carry out any research on a fee-for-service basis. The thinking behind this was twofold: fee-for-service implies a routine and predictable activity with minimal intellectual input, which is certainly not the case for drug discovery, and there might be few customers – very few academic labs have funds to cover the real costs of drug discovery through fees. To engage academic labs, the center decided to operate on a collaborative basis, utilizing UCRF funding to cover some of the initial costs of projects to create preliminary data for joint grant applications.

While scientific and technical skills are obviously important to the CICBDD's success, the training in leadership, organizational design, and strategic planning from the time spent in industry (SVF) were the basis for the purposeful design of the center described above. Explicit training in these areas has recently become more accepted in academic medical centers and is a welcomed addition to faculty skills. “An investment in knowledge pays the best interest.” – Benjamin Franklin.

4 What Works Well

All academic centers struggle with developing meaningful quantitative metrics for drug discovery efforts. The ultimate, and perhaps best, measure of impact would be the number of safe, effective medicines approved for significant unmet medical needs, delivered at lower cost than would have been spent in industry. Unfortunately, the long timelines associated with drug discovery make the use of this metric of success impractical. And thus, academic centers must work to imperfect proxy metrics, such as the derivation of clinical candidates, the novelty of the drug target, and the knowledge gained by drug discovery efforts. It should be noted that these long timelines create the same, and perhaps even greater, challenges in industrial drug discovery. As long as the half-life of senior leadership is much less than the cycle time for the organizational change they impose on pharma to meaningfully impact end-to-end productivity in discovery and development of approved medicines, the “tail” of mergers and acquisitions to feed profits will always wag the “dog” of the early discovery engine (117).

With these caveats about metrics, the CICBDD has identified two clinical candidates for previously unprecedented drug targets. One is now progressing through Phase 1 under the auspices of Meryx Pharmaceuticals (http://www.meryxpharma.com) and targets the Mer tyrosine kinase (MerTK) (120), and the other is under consideration for out-licensing. The center also contributed to the creation of IP that helped launch four additional start-up companies from UNC that are actively pursuing drug discovery programs.

Drug discovery projects for unprecedented drug targets usually involve both “validation risks” (likelihood that modulation of the target will have a favorable outcome in patients) and “technical risks” (likelihood that a tolerable molecule can be discovered that modulates the target in patients) (121). Our projects bear risks in each of these categories, and tellingly our most successful project, MerTK, was of lower technical risk due to the accumulated expertise in drug discovery versus the kinase target class and the ability to use structure-guided approaches (121). Over the course of these drug discovery efforts, more than 20 postdoctoral fellows received training in medicinal chemistry, and more than 30 publications resulted. These projects were funded through the CICBDD's participation in the NCI's Chemical Biology Consortium (https://next.cancer.gov/discoveryResources/cbc.htm) and using grants from the National Institutes of Health (NIH).

Our chemical biology program is based on targeting methyl-lysine-binding domains (Kme readers) and methyltransferases. The scientific challenges in probe discovery for Kme readers provided excellent fodder for students and fellows because the probes target a PPI modulated by the eponymous posttranslational modification (PTM). Methyltransferase enzymes were considered inherently more technically feasible than PPIs, but the knowledge base for this class was minimal at the initiation of our program, and we contributed significantly to the progress in the field (122). The methyltransferase project was initiated by Prof. Jian Jin, now at Mount Sinai, as a close collaboration with the SGC, and this and Kme readers program resulted in six high-quality chemical probes for these target classes (114, 121). These probes have been purchased more than 3000 times from commercial vendors and requested via the SGC probe library more than 6000 times. Five students will have soon completed their Ph.D. training based on targeting Kme readers, and more than 10 postdocs also received training in chemical biology as part of this program. Forty papers have been published describing this research, and our chemical probes have been shared freely with many labs around the world, both through the SGC and directly from the CICBDD. Perhaps, the most gratifying aspect of our chemical probe program has been our ability to share probes freely and then see them used as tools in other labs to increase scientific understanding of the proteins they target – an IP-free approach is truly liberating.

Indeed, the creation of chemical probes is ideally suited to academic models of funding, training, and collaboration. Once an initial knowledge base and strategy is in place for a target class, such as Kme readers (121), the efforts of one to two students/postdocs can progress a target to the chemical probe stage during their training. The absence of IP constraints enables sharing of probes with multiple biological collaborators who can test many hypotheses in differing model systems and uncover unanticipated connections between a target and biological consequences of its modulation – often with implications for treatment of disease (https://news.oicr.on.ca/2019/01/new-potential-treatment-for-leukemia-discovered-by-oicr-scientists-draws-major-industry-investment%EF%BB%BF/). While NIH funding is always competitive, the more exploratory nature of probe discovery better fits the expectations of study sections for innovation and creation of foundational knowledge; accordingly, we have been successful in maintaining funding for our Kme reader probe program for more than a decade, and the majority of our NIH funding has come from our chemical probe research. While open-science chemical probe efforts may not have the immediate allure associated with drug discovery, they are in fact one of the best opportunities for chemists, structural biologists, and biochemists to create de-risked drug discovery opportunities. Dozens of proprietary drug discovery programs have been launched from the data generated from open chemical probes created by our center and the SGC.

5 Challenges

An ongoing challenge in our drug discovery efforts, and in most other academic drug discovery efforts, is the limited number of target proposals that are novel, have some promise of validity, or are even remotely technically feasible. Of course, risk is always high prior to some investment of effort, and it also takes time; technical de-risking through assay development, hit discovery, and hit-to-lead chemistry is minimally a 2-year endeavor, and biological de-risking of target validity, best achieved using a chemical probe/lead in preclinical models, only adds to these timelines. In other words, good projects do not come knocking on your door; they are almost always created through expenditure of resources, by cultivating internal expertise in a system, and through partnerships with disease experts. One must also distinguish drug discovery, which requires these significant investments in time and intellectual resources, from hit discovery, which is a more generic, platform-based endeavor, and can be achieved with any of a number of approaches, such as high-throughput screening. Our thesis is that to be impactful, academic drug discovery centers should specialize in a target class (121) or a unique technology and thus acquire and apply distinct insights.

Another limitation of academic drug discovery settings, indeed of all drug discovery settings, is that of funding. Grant funding can be used to diminish technical and validation risks, to discover a high-quality chemical probe, and to generate an interesting therapeutic hypothesis, but grant funding is less likely to support a 2–3-year LO effort whose next major scientific payoff – testing the hypothesis in humans – is many years and many millions of dollars down the road. It is understandable that public funders are reluctant to finance such efforts (and are often unqualified to review them), but starting a new company or partnering at this early stage is often equally difficult. For these reasons, initially competitive academic projects can move at such a slow pace that they are no longer in a leading position when a candidate is identified. Alternative funding mechanisms are clearly needed if the traditional, IP-driven model of academic drug discovery is to flourish, and this issue is being partly addressed by firms or funders that aim to fill this gap for LO. These include pharmaceutical companies (https://www.tritdi.org/), contract research organizations (https://www.evotec.com/en/innovate/bridges), and pools of private capital (such as Deerfield's investment into UNC (https://uncnews.unc.edu/2018/10/22/unc-chapel-hill-and-deerfield-management-announce-the-creation-of-pinnacle-hill-to-accelerate-the-discovery-of-new-medicines/). These funding mechanisms may be the solution to enable support for efforts across the spectrum of activities required in drug discovery, activities that cannot be supported by traditional grants.

6 New Approaches

Whether early-stage drug discovery is carried out in industry or in academia, the core scientific problem is the same – our understanding of disease is too poor to adequately predict the efficacy of a new treatment targeting an unprecedented mechanism, and many exciting therapeutic hypotheses will prove wrong. And whether early-stage drug discovery is carried out in industry or in academia, if the traditional drug discovery business model is pursued, the process will be kept confidential until after patents are filed, the licensor will negotiate for the most lucrative terms, and any resulting drug will be priced as high as the market can bear. In practical terms, there is little to distinguish an academic drug discovery operation from a biotech company – except that academics are using students and fellows as their labor source.

However, we believe that academic drug discovery efforts have one distinct advantage: the public good mandate of these institutions and their associated faculty's commitment to service provide university-based drug discovery efforts the freedom to explore completely new business models – models that might aim for impact, affordability, and accessibility as opposed to maximizing returns on investment. We have decided to explore one such model, in which the data and reagents are shared openly and frequently as the project goes along, and marketing rights of any developed product will be protected by regulatory data exclusivity and not by patents (115). By eschewing patenting, the model both promotes partnerships and reduces the many conflicts of interest that emerge when one tries to maximize financial gain within a publicly funded institution.

The genesis of our open science business arose from two observations. First, we noted that there were massive amounts of funding available in the public sphere. The sum total of global nonprivate investment in biomedical research totals well over $150B per annum. This is an immense pool of funds compared with the ∼$10B that venture capital spends annually in biotech. Second, we were convinced that most academic scientists and institutions would be willing, even delighted, to contribute and align their intellectual capacity and laboratory funding to the discovery of a new medicine under the following conditions: (i) the science must be impactful and have the potential to lead to high-quality publications; (ii) there must not be any restriction on publication of the scientific outputs; (iii) the science must be shared in real time, as the project progresses; (iv) all materials (i.e. chemical probes and pre- and clinical candidates small molecule or antibodies) must be shared with the community as the project progresses, and under terms that restrict actions that may encumber any resulting research (such as patenting); (v) the pricing of the new medicine must be as low as is possible; and (vi) any profits from drug sales must be given away to charity – no participating institution or scientist should receive any compensation (equity or any form of income) from the sales of an approved drug.

This approach might sound overly optimistic of the willingness of institutions and individuals to forego financial rewards, and even naïve, but it has already been implemented in a company called M4K Pharma, and the two-year experience in operating the company has only increased our confidence as to its potential. Our commitment to openness and affordable pricing has attracted dozens of collaborators and resources from both the public and private sectors. We have also observed that scientists in pharmaceutical companies are more than willing to donate their time to consult for the project – a phenomenon that has also been observed by malaria and neglected disease drug discovery researchers. We have seen that every single academic institution with whom we have wished to collaborate, including those in Canada, the United States, Spain, and the United Kingdom, have been willing to forego IP rights in the M4K Pharma projects. Clearly, this model provides a vehicle for many academics to fulfill their desire to make a difference and participate in the development of a new medicine, yet not forsake their commitment to open knowledge generation for the public good.

7 Conclusions

A dozen years into our collaboration and the life of the CICBDD, we have learned a great deal about the challenges and opportunities in academic drug discovery. While it has proven possible to balance the traditional business model for drug discovery with the training and knowledge-generation mission of an academic center, there are inherent conflicts and the time seems ripe for a new approach. Our open-science chemical probe effort has created new small-molecule tools for unprecedented targets. Importantly, these widely shared tools have proven highly effective to explore new biology and to identify new therapeutic hypotheses, as well as to seed new drug discovery programs – notably without the incentive, the cost, or the encumbrances of patents nor with any hint of sacrifice of our public good mission. Indeed, with the impact achieved by inventing and sharing open science chemical probes, it is increasingly difficult to justify allocating public funding to proprietary drug discovery.

It may be time for academic drug discovery centers to reconsider their mission. In many ways, these centers are structured ineffectively. They are given the aims of a biotechnology company but not the resources. They are embedded within public good institutions, yet are forced to adopt both the proprietary concerns of industry and the financial models of an early-stage corporate investor. Given that there may be a viable path to achieve public good through open science drug discovery, and a mechanism to license assets under terms that promote affordability and access, should not academic drug discovery centers abandon the traditional model for one that seeks to maximize public good and embrace the academic ideals of knowledge creation and sharing? Of course, there is an element of risk. Given that the discovery of completely new medicines often takes decades, there is no evidence now that the open model will be effective. But what are the downsides? At worst, we will generate openly available knowledge and research tools about unprecedented targets and mechanisms, including data about the therapeutic relevance of the target or mechanism in patients, and these will spur others to discover anew or to abandon unproductive paths. At best, our new drugs will achieve clinical proof-of-concept in one or more diseases, and we will then develop the drugs ourselves, or with an industry partner aligned with our values, and provide society with a safe, and affordable, new medicine.

Conflicts

S.V.F holds stock in Meryx Inc., which is developing MerTK inhibitors discovered in the CICBDD. Aled Edwards is a chief executive of the SGC and chair of the Board of M4K Pharma.

1 Establishment of the ITDD

The University of Minnesota (UMN) has a long history of biomedical innovation, including drug discovery and development. Perhaps most notably, in the 1980s, a family of carbocyclic nucleosides known as carbovirs were designed by Professor of Medicinal Chemistry Robert Vince. In 1992, UMN exclusively licensed a series of patents for these antiviral agents to Burroughs Wellcome (later Glaxo Wellcome, later Glaxo Smith Kline or GSK). In 1998, Glaxo started marketing the new anti-HIV drug abacavir (Ziagen). After a legal dispute over whether abacavir was covered by UMN's patents was settled, royalties from the sales of this life-saving drug eventually brought in over $600 million to the University. This showed that academic drug discovery is not only scientifically rewarding but can also be financially rewarding. In the early 2000s, partially inspired by the success of Ziagen, the UMN made a strategic pivot toward developing the infrastructure to support translational sciences. The general scientific funding environment was favorable as well, thanks to the creation of the new NIH Roadmap program supporting academic translational science, including drug discovery and development. The University decided to create a comprehensive drug discovery and development center as part of the Medicinal Chemistry Department within the College of Pharmacy. This initiative was spearheaded by Dr. Marilyn Speedie, the dean of the College at the time, and Dr. Frank Cerra, who was the head of the Academic Health Center. Professor Gunda Georg was recruited from the University of Kansas to lead the new center named the Institute for Therapeutics Discovery and Development, or ITDD.

The ITDD opened its doors on 8 January 2007 in a newly renovated facility with an initial staff of two dozen scientists and 20 000 sf of premium scientific space. In the following months, the institute director, Dr. Georg, and the associate director, Dr. Vadim Gurvich, who oversaw the facility renovation effort, established four scientific cores and recruited industrially trained scientists as the core leaders. The structure was intended to mirror the typical industrial drug discovery approach. It included a highly automated High-Throughput Screening and Assay Development Core, a Lead and Probe Discovery Core, a Medicinal Chemistry Core, and a Therapeutics Process Development (TPD) Core with a GMP facility. Several years later, a Preclinical Pharmacology Core was added. With its collective expertise among the five cores, the Institute's structure allows the development of a drug candidate starting from a biological target and moving it all the way through preclinical development. The ITDD was also designed to leverage UMN's outstanding basic biomedical research infrastructure and excellent clinical capabilities. Through its highly collaborative network, the ITDD established access to various complementary capabilities, such as X-ray crystallography, computational chemistry, pharmaceutical formulations, bioanalytical chemistry, toxicology, translational medicine, cancer biology, and many other specialties. Within the first few years, the Institute became involved in dozens of internal and external collaborations, including several funded by the NIH. The Institute's flexible business model was designed to accommodate the diverse funding needs of its collaborators and available opportunities from funders. The Institute's strategic goal is to discover, design, and develop drugs that can reach patients. With this in mind, intellectual property protection is at the heart of its activities. Projects in the ITDD are not focused on a single therapeutic area, but rather span a wide range of areas that include cancer, infectious diseases, contraception, glaucoma, epilepsy, and Alzheimer's disease, among others.

2 Organization of the ITDD

The ITDD is housed within the Department of Medicinal Chemistry at the UMN. One of the primary benefits to this organizational structure is that graduate students in the Department have exposure to modern methods of drug discovery that they might not experience elsewhere in their training. Additionally, undergraduate and graduate students from other departments often work in the ITDD laboratories, thus gaining valuable experience. Each of the core directors in the ITDD has an unpaid research faculty appointment in the Department of Medicinal Chemistry. In general, the directors are not part of the graduate faculty of the Department but teach in the graduate and professional programs in the College of Pharmacy, participate in departmental and collegiate committees, and informally mentor students.

The ITDD is organized into core groups, much like those that exist in biotechnology companies. There are five cores in the ITDD: (i) High-Throughput Screening and Assay Development, (ii) Lead and Probe Discovery (LaPD), (iii) Medicinal Chemistry, (iv) TPD Core, and (v) Preclinical Pharmacology. The cores allow the ITDD to work on a variety of projects ranging from assay development and optimization to scale-up and GMP manufacturing for Phase I and II clinical trials. The cores are staffed by permanent scientific personnel and also by graduate students and postdocs. Projects that enter the ITDD can involve either a single or multiple cores.

3 Project Examples

3.1 Minnelide: A Prodrug of the Natural Product Triptolide with Strong In Vivo Anti-Cancer Activity

Triptolide is a natural product isolated from the plant Tripterygium wilfordii used in traditional Chinese medicine to treat rheumatoid arthritis and inflammation. It has also been reported to have potent anticancer activity in vitro against a number of cancers. Our collaborator, Dr. Ashok Saluja, then with the Department of Surgery at UMN, was particularly interested in Triptolide's strong activity against pancreatic cancer. However, Triptolide's poor aqueous solubility, narrow therapeutic window, and poor patentability as a natural product made clinical development of triptolide extremely challenging.

We were able to overcome these difficulties by designing and developing the prodrug minnelide (Figure 11), which incorporates a phosphonooxymethyl group that greatly increases its water solubility (125). This promoiety allows for rapid and reliable cleavage in vivo, releasing Triptolide and formaldehyde after activation with alkaline phosphatase. Minnelide showed very good efficacy at low doses against multiple animal models of cancer, including human colon adenocarcinoma (Figure 12). Further studies showed similar results in models of pancreatic cancer (126).

The discovery and development of minnelide at the ITDD was supported primarily through internal UMN grants, including funds from the Masonic Cancer Center, the College of Pharmacy, and the Academic Health Center. Collaborations between the ITDD and other UMN translational groups, such as the CTM, enabled the internal development of minnelide through the IND-preparation stage. At that point, minnelide was licensed to Minneamrita Therapeutics. It is currently in Phase I and II clinical trials.

3.2 CKLP1: A Water-Soluble Prodrug with In Vivo Ocular Hypotensive Activity

Elevated intraocular pressure (IOP) is a major risk factor for glaucoma, and most treatments designed to slow the progression of glaucoma focus on reducing IOP. Our collaborator, Dr. Michael Fautsch of the Mayo Clinic, identified adenosine triphosphate-sensitive potassium (K_ATP) channels as key cellular effectors of IOP. He found that administration of K_ATP channel openers such as diazoxide or cromakalim led to significant decreases in IOP in a normotensive mouse model. These drugs also lowered IOP in cultured human anterior segments. Utilizing pharmacologic K_ATP channel openers with selective subunit specificity, and a combination of immunohistochemistry and treatment of Kir6.2^(−/−) mice, Dr. Fautsch determined that the IOP-lowering effect of K_ATP channel openers was due to their action on K_ATP channels that comprised SUR2B and K_ir6.2 subunits.

However, cromakalim and diazoxide have limited aqueous solubility, and in his initial in vivo studies Dr. Fautsch used DMSO and Cremophor EL to solubilize the K_ATP channel openers and permeabilize the cornea so that the drugs could be administered topically. Because the toxicity of DMSO makes its clinical use in eye drops undesirable, we prepared a series of water-soluble derivatives of K_ATP channel openers and worked with Dr. Fautsch to evaluate whether they could be administered topically to lower IOP (127).

The compound that we identified with the best mix of efficacy, solubility, stability, and synthetic scalability was CKLP1 (Figure 13), a prodrug in which a phosphate group is directly attached to the alcohol moiety of cromakalim. CKLP1 was highly water soluble (>10 mg/mL), allowing it to be administered in a simple phosphate-buffered saline (PBS) vehicle. Once daily, CKLP1 showed good IOP-lowering efficacy in normotensive mice (Figure 14) as well as in rabbits, monkeys, and dogs. It also worked to lower IOP in an additive manner with existing glaucoma therapeutics (128).

Funding for this project was initially obtained via grants from the Mayo Clinic and the UMN Office of Discovery and Translation, followed by a larger grant from the Minnesota Partnership for Biotechnology and Medical Genomics. The Partnership is a program funded by the State of Minnesota designed to encourage collaborations between UMN and the Mayo Clinic. Once a potential lead compound was identified, we secured an additional grant from the Partnership to fund development work on the compound. CKLP1 was licensed in early 2019 to a start-up company that plans to begin human clinical trials with the compound in 2021.

3.3 Development of Gut-Restricted Bile Acid Analogs Designed to Inhibit Clostridium difficile Spore Germination

Clostridium difficile is a spore-forming bacterium that can cause a potentially lethal infection of the colon. Clostridium difficile infection (CDI) normally occurs after a course of antibiotics disrupts a patient's intestinal flora, allowing spores of C. difficile room to germinate and propagate in the colon. The typical treatment for CDI is an additional round of antibiotics. However, this causes further disruption to the patient's indigenous microbiota, and approximately 20–40% of patients relapse.

Rather than finding new ways to eradicate vegetative cells, our approach to solving the problem of recurrent CDI is to focus on preventing C. difficile spores from germinating in the colon and reestablishing the infection after the cessation of antibiotics. Spores of C. difficile use cues from their environment to determine if they are in the right location to germinate. In particular, the bile acid taurocholic acid (TCA) promotes spore germination, while derivatives in the chenodeoxycholic acid (CDCA) and ursodeoxycholic acid (UDCA) families typically inhibit germination. We have synthesized a large set of CDCA and UDCA derivatives and evaluated their ability to prevent TCA-promoted spore germination in vitro. We have identified several synthetic compounds that inhibit germination significantly more potently than naturally occurring compounds (129). We then focused on incorporating features into these molecules that reduce their uptake from the digestive tract into enterohepatic circulation by working to block both the active and passive transport mechanisms (Figure 15). We are currently evaluating our compounds in animal models of CDI to determine their efficacy.

To advance this project we have brought together a project team that includes Dr. Alex Khoruts, a gastroenterologist who regularly treats CDI patients and has access to clinically relevant strains of C. difficile, and Dr. Michael Sadowsky, an experienced microbiologist who has the facilities necessary to work with the infectious disease. The program was initially funded by a grant from the UMN Academic Health Center (AHC) that was specifically designed to encourage collaborations between investigators in different colleges of the AHC. Our continuing research in this area is funded through a pair of grants from the US Department of Defense.

3.4 A Prodrug/Activating Enzyme Delivery System to Treat Epileptic Seizure Emergencies

A key advantage of academic drug discovery is that the university environment can sometimes bring together project teams that would not naturally occur in an industrial setting. This can lead to solutions to problems that would not be arrived at by a single researcher working alone or even a team working in a corporate environment. One example of this is our project to develop diazepam derivatives that can be delivered nasally during epileptic seizures. The project team members include Drs. Gunda Georg (a medicinal chemist), James Cloyd (an expert in epilepsy and orphan diseases), Ronald Siegel (Pharmaceutics), and Edward Paterson (Veterinary Medicine).

Benzodiazepines derivatives such as diazepam are used clinically for the treatment of seizure emergencies. In many cases these drugs are administered intravenously, which can lead to delays in treatment, especially when seizure emergencies occur outside of a hospital setting. Rectal administration can also be effective but is hindered by poor patient compliance. Nasal delivery is an attractive alternative to these other routes of administration, but the limited volume of the nasal cavity requires dosing with highly concentrated solutions. Diazepam is not water soluble enough to be administered nasally in aqueous solution, and the use of organic cosolvents causes nasal irritation and pain. One solution to this problem is to administer a water-soluble prodrug of diazepam, such as avizafone (Figure 16). When administered by injection, avizafone is cleaved enzymatically to intermediate 1, which is then chemically converted to diazepam. Unfortunately, the nasal cavity has insufficient amounts of enzymes available capable of rapidly activating avizafone. To overcome this problem, the project team developed a dual-delivery system that administered avizafone along with an enzyme to cleave the promoiety. The resulting supersaturated solution of diazepam was highly permeable (130), and excellent PK properties were observed upon nasal application in rats, providing in vivo proof of principle. This project was funded through a mix of UMN internal grants and philanthropic and foundation funds. Efforts to commercialize the discovery are underway.

3.5 Discovery of Nonhormonal Male and Female Contraceptive Agents

A major focus of the ITDD is the discovery of nonhormonal contraceptive agents to regulate male fertility by pharmacological means. Sixty years after the female birth control pill came to market, a similar option still does not exist for males, who are restricted to condom use and vasectomy. To provide couples and men with alternative methods for birth control, the ITDD has been engaged in targeting multiple validated contraceptive targets in collaboration with reproductive biologists. One such genetically validated target is the testis-specific plasma membrane alpha 4 isoform of Na,K-ATPase. In collaboration with Professor Gustavo Blanco from the University of Kansas Medical Center, we have developed a picomolar inhibitor of the alpha 4 Na,K-ATPase that is a completely selective and orally bioavailable analog of the natural product ouabain (Figure 17). The analog reduces both total and progressive sperm motility in rats to levels that are predictive of male infertility according to WHO standards (131).

The ITDD is currently (2017–2021) the recipient of a U54 Center grant from the NICHD Contraception Research Branch to discover and develop male contraceptive agents. We are developing selective antagonists for the retinoic acid receptor alpha in collaboration with Dr. Debra Wolgemuth from Columbia University; inhibitors of the testis-specific bromodomain BRDT with Dr. Jun Qi from the Dana-Faber Cancer Institute; and inhibitors of testis-specific serine/threonine protein kinases (TSSKs) with Dr. Pablo Visconti from the University of Massachusetts, Amherst. Protein crystallography is supported by Dr. Ernst Schönbrunn from the H. Lee Moffitt Cancer & Research Institute. The U54 Center also involves a behavioral research component led by Dr. Jennifer Barber from the University of Michigan.

4 Lessons Learned

Over the past 12 years, many lessons have been learned regarding the best practices for the operation of a drug discovery and development institute at a public university in the Upper Midwest.

5 Features of Successful Projects

1 Mission

Chemical biology is an evolving, multidisciplinary field at the interface of chemistry and biology. Detailed definitions of chemical biology research are still being discussed (132) but in general involve manipulation and interrogation of biology using chemical entities. One subcategory of chemical biology concerns the generation and validation of new small-molecule-based tool compounds. These serve as excellent tools to modulate biological systems and possess unique characteristics that nicely complement some of the shortcomings with modern precise genome editing. Needless to say, high-quality basic research in this field is a critical foundation for life science innovations, with emphasis on medicine and health, and with more unexplored potential in plant sciences and environmental research. However, access to well-characterized molecules with appropriate biological selectivity is limited to a small portion of the proteome, such that significant investments are motivated to broaden the repertoire of accessible biologies.

Chemical Biology Consortium Sweden (CBCS, http://www.cbcs.se, (133)) was inaugurated in 2010 as a distributed national infrastructure for chemical biology in Sweden. The mission of CBCS is to perform collaborative research with research groups in Sweden, aiming to discover, optimize, and use new chemical tools for exploration of complex biology. CBCS aims to enable high-quality basic research in this field, generate open access publications, and foster cross-fertilization between universities and industry in the life sciences arena. Although CBCS' mission is primarily focusing on enabling basic life science research, the operations have after 10 years led to an excess of 130 co-publications and 11 patent applications and has served as the basis of the formation of six start-up companies.

2 Description

2.1 How CBCS was Formed and Organization Today

CBCS was formed in 2010 after an initiative by employees of the pharmaceutical company Biovitrum (former Pharmacia and Upjohn). This led to the establishment of a chemical biology research node at Karolinska Institutet, including a consolidation of already existing chemical biology infrastructures in Sweden at Umeå University and Uppsala University (134). In 2013, CBCS became part of SciLifeLab, a national center for molecular biosciences (135), and is now part of a combined platform for chemical and genetic perturbations of biological systems. Today, CBCS comprises a multidisciplinary team of 16 staff scientists in areas spanning from biochemistry and cell biology, organic/medicinal/enabling/computational chemistry to compound logistics and management (see Figure 18 for a schematic illustration of CBCS operations).

As a complement to CBCS, SciLifeLab established a dedicated platform for drug discovery and development (DDD) in 2014 (136) with the primary goal to support academic research projects with improved translational potential through offering of industry standard services, both in small-molecule projects and also using other drug modalities. The combined chemical biology and drug discovery platform provides a fantastic opportunity for Swedish academics to initiate a chemical biology-related research project, which later can extend into the drug discovery value chain and mature into a lead therapeutic for further development. In fact, several projects that were initiated and validated at CBCS have later continued their drug discovery journey within the SciLifeLab DDD platform (137).

3 CBCS Services

Most CBCS projects are characterized by a combined basic research and human health application focus, and a large portion of the users have a strong strive to understand disease biology and a wish to advance their projects toward drug discovery. CBCS provides services in assay development – both biochemical and cell-based assays, access to and screening of the SciLifeLab compound collection and medicinal and computational chemistry for validation, structure–activity relationship (SAR) exploration and optimization of hit compounds, and chemistry for mechanism of action elucidation efforts (chemical proteomics) (Figure 18). The goal is to generate validated compounds and targets – assess “ligandability” of pathways and targets – and create a foundation toward initiation of drug discovery projects.

4 CBCS Operational Model

4.1 Operational Model

The most successful projects are based on a tight collaboration with the academic research group. Research groups interested in comprehensive CBCS services comprising assay development, small-molecule screening, and synthetic chemistry activities are defined as a “large collaborative project,” and principal investigators (PIs) have to send in an application to a CBCS Project Review Committee (PRC). In addition, CBCS assists in small collaborative projects that require limited resources, e.g. use of instrumentation/compound libraries or intellectual input. CBCS projects are categorized according to the type of project: isolated protein, targeted cell based, phenotypic cell based, plant biology, technology development, diagnostic tool, and others (i.e. ADME profiling and chemistry analysis); see project distribution in Figure 19.

The PRC committee evaluates applications based on the following criteria:

• Overall biological rational and potential scientific impact
• Plans and availability of assays and models for validation of research tool compounds
• Technical feasibility for CBCS and the user
• CBCS resource requirements
• The importance of CBCS efforts
• Publication strategy

4.7 Infrastructure Accessibility and Publication Output

A main objective for CBCS is to be available in a transparent way to all users in the Swedish research community. Accordingly, CBCS has engaged in collaborative projects resulting in a project portfolio of >400 projects in various stages from ∼236 individual users between 2010 and 2018 with users from all major universities with a national spread, see Figure 20, but focus on the location of CBCS nodes. User interest in CBCS services has grown steadily during this period with approximately one new project discussion meeting, often with a new PI, per week. In these meetings, including CBCS experts with different expertise, all potential users are provided with useful input to their projects. Thereof, 25% results in full PRC applications, whereas other projects get CBCS limited support through a small project. CBCS publications normally have a lag time of a couple of years after project completion and have currently reached a steady state of about 20 publications annually. Besides the publications and educational impact, CBCS operations have generated more than 11 patent applications and the start-up of six companies.

5 CBCS Projects – Examples

We have chosen to exemplify two projects representing the different CBCS services: (A) This is a collaborative project that started with a phenotypic screening campaign in primary cells and altogether required significant assay development, screening, and chemistry support for elucidation of the mechanism of action of low nanomolar compounds (Figure 21). This project has resulted in a first publication (140) and awaits disclosure of a more complete study. The project has also generated patent applications, and the project is now being pursued by a biotech company. (B) This represents a major technology development around the cellular thermal shift assay (CETSA), which started as a user-driven project and lead onto internal technology development within CBCS (Figure 22). The overall project has so far generated six CBCS publications and have significantly impacted the business of a Swedish biotech company Pelago Bioscience. Further developments are still ongoing in this active project.

A. Exploring the druggability of Diamond-Blackfan anemia: Johan Flygare at Lund University has an interest in a rare hereditary disease called Diamond-Blackfan anemia (DBA). The mechanism of this bone marrow failure syndrome, caused by a mutation in ribosomal proteins, is poorly understood but involves reduced proliferation of erythroid progenitor cells. Johan Flygare contacted CBCS in 2012 and asked for guidance in a screen for small molecule that could rescue the viability of erythroid precursors and use such compounds and mechanisms to explore DBA biology.

Together with CBCS, a phenotypic assay was developed using fetal liver erythroid progenitors from a doxycycline-inducible DBA mouse model (see Figure 21). The addition of doxycycline to the culture medium induces the growth defect phenotype and reduces proliferation to <10% of normal such that rescue of proliferation could be used as a simple readout for screening. The primary screening of the CBCS collection, which at that time amounted to 10 500 cmpds, was followed by hit validation activities and generated several hit series of which one was particularly promising showing a crude SAR. Based on these data, the PI could apply for a follow-up project aiming at elucidating the mechanism of action of this hit series, and the project was transferred from a biology-oriented to a chemistry-focused program.

Substantial “detective work” was carried out by the chemists including in-depth literature and database mining, synthesis of hit series analogues and known reference compounds, and multitarget assay panel testing. Making a long story short, this resulted in a proposed target responsible for the rescuing effect of proliferation of the mutant mice erythroid progenitors. The rescue effect for this and other chemotypes for the same target was later validated using patient-derived cells. Based on these findings the researcher applied to the SciLifeLab DDD platform, which supported the filing of a patent and initiation of a collaboration with a biotech company that is presently pursuing this project toward the clinic.

B. Method development around the cellular thermal shift assay (CETSA): The ability to directly verify the interaction of small molecules with their putative protein targets in native cellular environments is a critical step in elucidating and validating pharmacological concepts. The CETSA was developed as a novel biophysical approach to investigate ligand binding in intact cells (141). CETSA is based on the principle of increased thermal stability of proteins upon (concentration-dependent) ligand binding and its detainments in the soluble fraction of cell lysis fractions after heat challenge. In contrast to most other target engagement techniques, no modification of the ligand or cloning work to include reporter groups on the target protein is required. Instead, the methodology is based on ligand-induced resistance of target proteins from thermal denaturation when the cells are heated, i.e. while ligand-stabilized protein remains in solution, unbound protein denatures and aggregates.

Already in the original CETSA paper, CBCS' intellectual input was acknowledged (Figure 22a) (141). This formed the basis of a collaboration in which CBCS, together with Professor Pär Nordlund's group and Pelago Biosciences, has demonstrated that CETSA is amenable to automation and screening, thus making it possible to apply large chemical libraries for screening purposes (Figure 22b) (142, 143). CBCS has now successfully developed several homogeneous AlphaScreen®-based assays (142-145) to monitor target engagement and also performed a first 384-well format screening campaign in live nonengineered cells expressing thymidylate synthase (TS) (143). The screen successfully identified all drugs within the test library that are known to modulate TS as well as novel compounds capable of binding and inhibiting this enzyme (Figure 22b) (143).

This work nicely showed the opportunity to identify cell-permeable compounds that show direct target engagement already at early-screening stages of a project. In a follow-up internal technology development project, CBCS scientists developed a method for quantitative interpretation of intracellular drug binding and kinetics using CETSA (144). This work describes the experimental path that is required for making proper comparisons of CETSA data with functional cellular responses at 37 °C and is believed to have broad implications in the appropriate use of CETSA for target and compound validation. A follow-up technology development project developed a CETSA protocol for detecting intracellular drug binding to p38α (MAPK14) in live A431 cells in situ, using high-content imaging (Figure 22d) (146, 147). The assay concept was further validated using a number of known p38α inhibitors, and the potential of this technology for chemical probe and drug discovery purposes was demonstrated performing a small pilot screen. Importantly, this protocol creates a workflow that is amenable to adherent cells in their native state and yields spatially resolved target engagement information measurable at the single-cell level. Finally, CBCS has translated this expertise into additional projects, including, for example, the first example of the use of CETSA as the main SAR-driving cellular assay in a high-impact publication on NUDT5 (148).

CETSA is today broadly used within academia and the pharmaceutical industry, which have also adopted the CBCS microplate format (e.g. AstraZeneca and GSK) (149). In addition, Perkin Elmer recently launched a number of Alpha SureFire® CETSA® Target Engagement kits for use in combination with Pelago Bioscience's CETSA® technology assessing target engagement in live cells (https://www.perkinelmer.com/, (150)).

CBCS has established itself as a world-leading authority for developing modifications of this technology for high-throughput screening and target-identifications purposes. Investigators collaborating with CBCS have access to our full arsenal of expertise and technology, allowing for powerful, conclusive scientific deduction of molecular mechanisms of action.

6 Lessons Learned and Future Directions

CBCS have during these 10 years successfully built an efficient and sustainable national platform for chemical biology open to all researchers in Sweden with qualified projects. Several lessons were learned on the way which has impacted the future directions of which some are mentioned below.

7 Concluding Remarks

Small molecules are excellent tools to unravel complex biology and “enlightening the (un)targeted proteome.” The understanding of gene and protein function in disease biology is of utmost importance for drug discovery and explains the majority of small-molecule application failures in the clinic. In order to maximize this development, there are lots of opportunities ahead of which some are excellently exploited by the SGC (https://www.thesgc.org/, (157)) and their collaborators (158). New and already explored high-quality chemical probes need to be made available for the entire research community (159). Expert scientists in drug discovery and chemical biology national infrastructures widely present in many countries have a great role to enable this development by expert guidance in selecting and using these tools. The development will further be enabled by cross-disciplinary collaborations and new method development together with related scientific research areas such as single and spatial omics and imaging, genome editing, advanced integrative data analysis, and artificial intelligence. CBCS is really excited and has great expectations for the future within the field of chemical biology.

1 Introduction

Food and Drug Administration (FDA) approvals of new therapeutic drugs (NTDs) have rebounded from a decade low in 2016. The FDA has recently presented a concerted effort to grant market authorization more quickly through a number of recent procedural changes. The approval of 59 drugs in 2018 is a record for the agency and complements the stunning curative breakthroughs in recent years (160). Representative approvals have included new immunotherapeutic treatments that have added years, not the traditional months, to response rates for some cancers, gene therapies for spinal muscular atrophy and sickle cell disease, and actual cures for a number of afflictions including hepatitis C and certain forms of AML. Importantly, the mean number of NTD approved annually has increased significantly (30 in 2000–2011 to >40 since 2012), creating a clear trend toward increased innovation. On the other hand, only 16 (26%) new molecular entities (NME), a significant decrease, were associated with companies classically described as Big Pharma. Thus, one might conclude that either scale is no longer as useful as it once was or smaller biopharma companies have obtained the capacity to shepherd innovation through to commercialization likely out of necessity. Accompanying the recent increased pace of approvals is a widening breadth of their impact to address clinical needs. While Big Pharma has remained somewhat focused on more common diseases, start-ups have been increasingly targeting lesser known diseases to many members of the public or even much of the biomedical research community. In fact, a majority of the 2018 approvals (31) were given FDA orphan designation and target rare diseases. Rare diseases are defined as disorders that affect fewer than 200 000 persons in the United States, and orphan status is provided to effective treatments that are not expected to recover the cost of development and marketing. Thus, not surprisingly, the commercial potential of the 2018 class of approvals is considered solemn. Only two of these approved treatments are expected to achieve greater than $2 billion in sales, and the average projected peak sales per NTD have continued to drop over the past several years.

FDA approvals of new therapeutic drugs and aggregate peak global annual sales.

Source: Data from Baedeker, M.; Ringel, M.; Schulze, U. “2018 FDA approvals hit all-time high – but average value slips again” Nat. Rev. Drug Disc. 2019, 18, 90.

2 Academic Drug Discovery Centers

If the revenue from new medicines cannot support their own research and development programs, the business model is no longer sustainable. While the role of academic research in the drug development process has been a subject of some, and often including high-profile, debates, there is agreement that it is an outstanding source of a broad range of fundamental scientific discoveries. Moreover, it is clear that these discoveries can be the foundation that ultimately leads to innovative therapies with proper shepherding through the development pipeline. According to the latest industry research, about 90% of drugs that reach the clinical stage development never make it to FDA approval and commercialization with most of these candidates has shown to have limited efficacy (163). This suggests that even the recent industry trend toward “quick-kill” decision-making strategies, which aim to reduce late-stage attrition, suffers from cognitive biases and inadequate predictors of success (164). Contemporaneously, pharmaceutical companies have also drastically cut early-stage research budgets and reduced research staff as a measure to de-risk their business models and focus on short-term gains. Thus, academic and not-for-profit research sectors have been given the opportunity and are taking an increasing proportion of responsibility in early-stage drug discovery (165). In response to these opportunities, there has been an increase in the number of academic drug discovery (ADD) centers, where often these programs are a part of a university's basic research support structure.

Internal support for translational research has a number of benefits beyond their potential to impact on societal health including but not limited to strengthening education and training programs, promoting faculty and institutional reputation, and enabling potential revenue through licensing and commercialization. Drug discovery centers at some universities are focused on single research groups and small collaboration between complementary researcher programs within specific focus areas, such as the Vanderbilt Center for Neuroscience Drug Discovery (VCNDD). The VCNDD is an outstanding example of a collaborative model for ADD based on the complementary expertise in neuroscience biology and industry experience in drug discovery and development. VCNDD facilities, which have been staffed by scientists with significant pharmaceutical industry experience, replicate the infrastructure typically found in the industrial sector. As a testament to their translational capabilities, this Center recently received approval from the FDA to advance one of their compounds into clinical trials for Alzheimer's disease.

From an alternative perspective, many other ADD centers play the role of a general support structure for research programs with overlapping interests more broadly defined. In 2012, the Academic Drug Discovery Consortium (ADDC) was created to connect the growing numbers of these centers (http://www.addconsortium.org/), and today it has over 150 ADD centers and close to 2000 individuals in its membership. A majority of the centers report research interests in small molecule and biologics. While 80% of the centers state a research focus in oncology, most are involved in several additional areas including infectious disease, neurological disorders, immunology, metabolic disorders, cardiology, and pain. From a unique perspective, Johns Hopkins Drug Discovery (JHDD) has partnered with Eisai Inc. to establish a hybrid industry/academic development team. While initially focused on the disorders of the central nervous system, the program expanded to provide services to all departments and therapeutic areas. In this collaboration, Eisai screens its internal proprietary compound library using assays developed and targets identified by biological researchers within Johns Hopkins University. Within this model, Eisai is motivated by the opportunity to participate in further development of the resulting hits through pharmacology and drug metabolism/pharmacokinetics studies and/or license the intellectual property (IP) at this early-development stage. The Center for Integrative Chemical Biology and Drug Discovery (CICBDD) at the University of North Carolina-Chapel Hill was created as a collaborative model in support of UNC researchers. The Center faculty and staff provide expertise in high-throughput screening and medicinal chemistry to help translate biological target discoveries into drug candidates. The Center maintains a focus on oncology with a strong partnership with the UNC Lineberger Comprehensive Cancer Center and announced their first clinical candidate for solid tumors and leukemia in 2018.

At the University of Notre Dame, we established the Warren Family Research Center for Drug Discovery and Development in 2014 through philanthropic support. As is typical of many academic research programs, the Warren Center was built upon a strong foundation of faculty colleagues with a history of drug discovery research interests. In this case, the Department of Chemistry & Biochemistry at the University of Notre Dame had long been known, in part, for excellence in antibiotic research including bacterial cell wall biochemistry and the discovery of related and novel small-molecule inhibitors. Over the years, this has been well complemented by a diverse group of faculties with expertise overlapping synthetic organic chemistry, medicinal chemistry, natural product research, and diagnostic tool development. Prior to the establishment of the Warren Center, faculty researchers had two avenues for translating their early-stage biomedical discoveries: (i) Develop expertise within their individual labs or (ii) identify expertise outside the university and develop personal relationships, collaborations, and partnerships which secured complementary capabilities and experience.

2.3 Intellectual Property Protection and Enterprise Development

Beyond the potential impact on human health and subsequent benefit to the academic reputation of the institution, further justification for the internal support for biomedical research has included the promise of potential financial gain. Several universities have benefited significantly from royalty payments associated with commercialization of their pharmaceutical technologies including Florida State University, the University of Minnesota, Emory University, Princeton University, UCLA, and Northwestern University. The Bayh-Dole Act, signed into law in 1980, gave academic institutions IP rights on the products of federally funded research. A recent evaluation of Bayh-Dole has shown that the legislation is working as intended helping American universities contribute to the American life science ecosystems and maintain economic leadership in the life sciences (177). Thus, universities are incentivized to patent research discoveries and inventions and license the IP to private-sector entities which then become responsible for the sometimes lengthy, difficult, and risky development effort. Unfortunately, as previously discussed, the age of the blockbuster drug has passed, and newer pharmaceutical business models are seeking to spread the risk of the development process shifting licensing decisions to later in the development process.

Traditional academic technology transfer offices are responsible for the protection of university IP through a process that first promotes faculty submissions of invention disclosures. The invention disclosure describes a potentially new idea, a current understanding of its commercial potential, and a brief evaluation of the current state of the societal problem the idea seeks to address. This information is supplemented, to varying degrees, by a market analysis and patent search, in an effort to determine whether or not there is adequate justification for proceeding with patent protection. From the author's experience, the process is highly reliant on the experience of the tech transfer staff, the promotional ability of the inventor, and the willingness for decision-makers to seek specific outside expertise. When it comes to drug discovery, patent type, timing, and jurisdiction are all critical. Protecting states of matter, their observed biological activity, and their therapeutic potential prior to their readiness for clinical development can eliminate interest from potential licensees. It is important to remember that the drug discovery development process is long and expensive. Shortening the lifetime of the patent, 20 years from filing, by protecting IP too early can negatively effect its commercial potential. Thus, a process that enables the delay of first public disclosure is preferable through the use of confidentiality agreements and carefully controlled dissemination of technical information.

Despite the opportunity to obtain revenue from IP licensing, in any given year, it is well documented that most universities fail to recover the costs of running their tech transfer offices. In fact, only a small number of universities consistently secure lucrative licensing deals that produce significant revenue. While blockbuster discovery can alter a university's fortunes, the vast majority of licensing deals yield little or no money, and for most universities the royalty returns are low. The problem is not only related to the high-risk nature of scientific research. Only a fraction of fundamental discoveries ultimately relate to and provide a solution to real-world problems. Much has been made of the so-called Valley of Death relating to the gap in external support for academic research that is found between early-stage, federal funding for basic research and proof-of-concept, where routine but critical supporting data is needed to attract pharmaceutical companies and biomedical investors (178). To the PI, this chasm not only exists in the paucity of external funding opportunities but also in access to internal university support structures; i.e., seed programs and research cores support fundamental discoveries and intellectual property offices that are most excited by projects with clear commercialization potential.

Notre Dame, like many other R1 universities, has begun to expand their technology transfer services beyond classic IP protection and into the realm of enterprise development. The lack of early-stage capital and support structures for innovation development is a major challenge in advancing promising university technology from the lab to commercialization. Research institutions are implementing gap funding and accelerator programs to provide support for proof- of-concept prototype development and start-up ventures. These programs have evolved into sophisticated evaluation, development, investing, and commercialization support structures that can provide the necessary space, services, and expertise for idea development, commercialization, business formation, prototyping, and addressing the need for entrepreneurial education for both students and faculty.

Academic incubator spaces are particularly attractive as they not only support translational activity arising from the university environment and moving into the private sector but also open up new sources of research funding including STTR (small business technology transfer grants) and SBIR (small business innovation research) grants from federal agencies as well as venture capital (VC) funds. State and local government are often partners in university-sponsored tech parks as new businesses can return investments to the economic development of the region in the form of job creation and tax revenue. STTR and SBIR grants are awards to the new businesses, with SBIRs offered by a dozen federal agencies and STTRs a subset of five. Specific institutes within the National Institutes of Health are the most likely source of these types of grants for drug development-related businesses. In contrast to typical basic research grant applications, SBIR and STTR applications must include a commercialization plan and thus have very different review criteria creating an initial barrier for inexperienced faculty applicants. The PI on an SBIR grant application must be employed at least 51% of the time at the small business, at the time of the award, and for duration of the research period. Thus, it is complicated for the PI to retain their academic appointment without securing an extended and at least partial leave of absence. University faculties already struggle to comprehend the intricacies of percent effort compliance. Thus, the appointment of an alternative representative as CEO for the business entity is valuable, and university enterprise development centers can be an important resource for identifying experienced and interested individuals for this role. Alternatively, an STTR application can be an easier path as this program allows a PI to come from either the small business or the academic institution requiring a 10% effort commitment which is in the range of typical academic grant expectations. Therefore, faculties who serve as the PIs for an STTR would not be required to take a leave of absence from a research institution.

In some cases, universities have created high-risk venture funds as a potential source of financial support for early-stage companies associated with their IP in return for partial ownership of the entity. They may also mediate networking for entrepreneurs to philanthropic alumni, alumni angel investors, and VC firms. While financial support is critical to ensuring continued progress down the development pipeline, success is not possible without experienced mentorship to advise on the best use for these funds. Overall, universities must provide new entrepreneurs training in securing funds for commercial entities along with access to serial founders whose experience maximize the likelihood for a successful transition.The Indiana Center for Biomedical Innovation was established in 2016 to support biomedical entrepreneurship and technology commercialization associated with Indiana University School of Medicine (IUSM). The Center provides access to office and laboratory spaces as well as IUSM instrumentation cores including animal facilities. Participants are provided a range of support structures including seed funding through the Indiana Clinical and Translational Sciences Institute (vide infra) and mentorship from a large, diverse advisory council. Since 2016, the ICBI has launched 13 commercial entities. Of particular note is the ICBI's record in generating SBIR and STTR applications and awards by their affiliated start-ups. For example, HB Therapeutics, Inc., which recently received a Phase I SBIR grant, is seeking to develop the first small-molecule degraders of small ubiquitin-like modifier-1 protein (SUMO1). In addition, Allinaire Therapeutics, which aims to develop novel therapeutics for chronic obstructive pulmonary disease (COPD) and other pulmonary diseases, was awarded a $1.6M Phase II SBIR from the National Heart, Lung, and Blood Institute within the NIH. Their technology is based on the discovery of a novel therapeutic target that plays a central role in the development and progression of lung disease. The director of the Indiana Center for Biomedical Innovation, Dr. Jaipal Singh, attributed their success in securing SBIR and STTR awards to their partnership with Parmelee Consulting Group (http://www.parmeleeconsulting.com/). Kris Parmelee has almost 20 years of experience in project management and fundraising and now almost exclusively focuses on SBIR/STTR grant application consultation. Parmelee commented that her biggest challenge in preparing a strong application is ensuring that the PI, who typically has exclusively academic grant writing experience, now views their technological discovery as a business. “While the proposed research goals may be creative and well thought out, it is critical that they are complemented by key benchmarks that demonstrate a path towards commercialization,” she stressed.

Indianapolis-based Parmelee consulting is also providing services to start-ups associated with the University of Notre Dame's IDEA Center. Standing for Innovation, De-Risking and Enterprise Acceleration, the IDEA Center is the exclusive home for commercialization and entrepreneurial activities at the University of Notre Dame. In addition, to overseeing entrepreneurship-related educational programs at the undergraduate, graduate, and professional levels, the IDEA Center is a commercialization and enterprise development program that seeks to advance new technologies and innovations from the Notre Dame community to the market. In its first two years, the IDEA Center founded more than 40 start-ups, which exceeds the number of ND-related commercial entities in its history. As with typical university technology transfer offices, the IDEA Center provides IP-related services from invention disclosure through patent protection and licensing agreement negotiation. In addition, projects are now managed through several stages of lean-canvass and lean-start-up process, which often includes financial support for prototype development and supplemental data collection through a de-risking fund. In the drug discovery space, this support can enable the hiring of external consultants and obtaining important pharmacological and toxicology data to complement fundamental biological and efficacy data already in hand. For example, our current effort to develop a natural product lead compound for AML that is being supported by the IDEA Center's commercial engine has proceeded from their risk assessment stage, called the two-stroke, to the four-cylinder de-risking phase. Funding has been provided to recruit the oncology-focused, contract research organization (CRO) TD2, Inc. (Translational Drug Development; Scottsdale, AZ) to provide additional preclinical data necessary to support an investigational new drug (IND) filing to the FDA for permission to support human clinical trials. Technologies that have been sufficiently de-risked are eligible to pitch for funding from the IDEA Center's Venture Fund as a complement to external funding sources such as STTR and SBIR grants. The IDEA Center has recently supported a drug discovery-based venture generated out of the Notre Dame's Department of Chemistry & Biochemistry. Extending on basic and translational research conducted with diabetic animals and human diabetic wound samples to elucidate the dysregulated mechanisms of wound healing, Professors Chang and Mobashery have developed a topical drug that addresses the underlying condition in pathology of the recalcitrant healing of diabetic wounds. The only currently approved drug has modest efficacy and comes with an increased risk of cancer and mortality. The new lead compound is currently undergoing IND evaluation through a $5 million Therapeutic Development Award from the Department of the Army with the goal of starting Phase I clinical trials in 2021 through the start-up, SalvePeds.

A recent $50 million dollar gift allowed the creation of a similar support structure at Harvard University. The Blavatnik Biomedical Accelerator provides strategic, monetary, and advisory support for the translating applied life-science research discoveries building upon their Office of Technology Development. The Accelerator has provided over $20 million in direct research support of more than 100 projects from across Harvard targeting most major disease areas. Supported projects have included the development of therapeutic strategies, new diagnostic tools and biomarkers, and other biomedical technology and instrument development. This effort has already demonstrated tangible results as nearly half of all completed projects have developed licenses with pharmaceutical and biotech industry or other major industry-sponsored research agreements as well as new venture start-ups. One particularly strong example demonstrates the value of providing support for the advancement of potential lead compounds to a relatively late stage of preclinical development. A faculty researcher in Harvard University's Department of Chemistry & Chemical Biology discovered a potentially novel therapeutic strategy for AML through the inhibition of kinases associated with transcription regulation. The technology was licensed to Merck in 2016 for an upfront fee of $20M along with future milestone payments. In return, Merck will shepherd the further development of these new cancer lead compounds including clinical development and commercialization.

3 Cross-Institutional Models for Promoting Academic Translational Research

3.2 Indiana Clinical and Translational Sciences Institute

The progress of new ideas through the development process, either to the market as therapeutic products or as implementations into health and disease management, faces many barriers. The NIH established the National Center for Accelerating Translational Sciences (NCATS) and Clinical and Translational Science Awards (CTSAs) with the goal to “transform the translational science process so that new treatments and cures for disease can be delivered to patients faster.” As the national platform for achieving this goal, the CTSA funding opportunity charged applicants to “create local ‘Hubs’ which support innovations in methodology, training and career development, as well as high quality clinical and translational research locally and nationally.” The Indiana Clinical and Translational Sciences Institute (Indiana CTSI) was created in 2008 as a statewide colaboratory to accelerate clinical and translational research across the three state research universities, Indiana University, Purdue University, and the University of Notre Dame along with partners within the corporate sector, the regional health care systems, and local foundations. The mission of the Indiana CTSI is to serve as the statewide catalyst for translational research, improve human health across Indiana, and promote the application of new paradigms that are capable of being implanted nationally. The Indiana CTSI is structured with the specific aims to promote collaborative research, provide resources and services to support clinical and translational research, offer education and training development programs for next-generation translational researchers, partner with the local community, and integrate with the national CTSA network.

ADD programs have objectives that typically include but are not limited to contributions to fundamental biology, biochemistry, and chemistry, the development of tool compounds or potential therapeutic leads, and the education and professional development. These wide-ranging goals can create conflicts of interest associated with partitioning of resources within a single institution or administrative structure. With hopes of addressing this issue, the Indiana CTSI created a Molecular Therapeutics Program (MTP) and charged it with the mission of building upon and coordinating with internal ADD support mechanisms and affiliated research groups with the partner institutions. The MTP provides supplementary specialized expertise and complementary resources critical to translational research but potentially difficult to find or organize within a single university. The MTP facilitates collaborative translational research across the state institutions, through identification of complementary expertise and support for team building across diverse disciplines. The MTP achieves its mission by leveraging the combined resources and expertise of the academic and industrial partners throughout the State of Indiana through a cross-state drug discovery consortium, the Indiana Drug Discovery Alliance (IDDA). The IDDA promotes and supports promising early-stage drug discovery research and provided expertise with a multi-institutional advisory committee which includes drug discovery expertise from Eli Lilly & Co. and the Indiana Biosciences Research Institute, along with the Indiana CTSI academic partners: Purdue University, Indiana University School of Medicine, Indiana University-Bloomington, and the University of Notre Dame. The Indiana Drug Discovery Alliance Advisory Committee is loosely based on a similar support structure used at Purdue University.

The Purdue University Institute for Drug Discovery has expanded a program originally offered through the University's Center for Cancer Research to provide specific advisement to projects with translational potential in therapeutics development. On a semiannual basis, a drug evaluation committee (DEC) is assembled with diverse team of experts including current and former large pharmaceutical company executives, research clinicians, serial biotechnology company entrepreneurs, and experienced academicians. The DEC provides expert reviews of projects presented by members of Purdue's faculty. In addition to scientific and technical advice, these discussions often include drug development topics not typically considered by academics including potential market size, current market competition, clinical trial indications, potential clinical trial recruitment challenges, chemistry manufacturing controls, and other manufacturing challenges. The Institute for Drug Discovery considers the activities of the DEC, an essential part of their successful effort to advance several drug candidates in their drug pipeline. It not only provides the PIs valuable advice enabling them to focus on critical milestones, but this type of expert analysis can also be used to justify internal funding, develop a proper IP strategy, and hone SBIR/STTR applications.

By including representatives from multiple institutions, the IDDA Advisory Committee can be a clearinghouse for drug discovery and development resources available through Indiana-CTSI partners and, when necessary, identify and expand current capabilities by initiating new, external partnerships. The MTP has connected with drug discovery research programs and their PIs across the state primarily through two annual seed grant programs: the annual IDDA Support Program and the MTP Pre-Clinical Service Award. The applications are simple to prepare, two pages in length, and focus on background, current status, and commercialization potential. By not requiring an upfront proposed budget, the program provides an opportunity for the review process to help shape specific aims. In the author's experience, the success of seed grant programs to support academic research is frequently underwhelming. Typically, the amount of funding is less than what is truly needed to complete the project, and reasonable participation rates require active recruitment of applicants. More consequentially, the funding is typically allocated at the discretion of the PI and thus may not be ultimately utilized in the productive way. This is particularly true for translational biomedical research projects where key milestones and data collection efforts that are needed for the applied aspects of a project are at times at odds with support for additional basic research goals. Results needed for publication and basic research grant applications are different than data that would help support a successful SBIR/STTR application. Thus, new MTP projects are first assessed for clinical, biological target, and chemical relevance by the IDDA advisory committee. The diverse experiences of the committee, including representatives from Eli Lilly with significant drug discovery and development experience, are extremely helpful for this critical first step and play an important role in identifying immediate needs and the development of a project budget. Whether or not selected for a second stage discussion, each project is assigned an IDDA project manager to directly assist the PI(s), provide valuable feedback, track the project for key milestones, and routinely assess future needs. A second round of evaluation is carried out through short presentations and interactive discussion. This provides an opportunity for the review committee to ask questions, evaluate the current plans, and offer guidance in generating the most appropriate specific aims. A strong subset of projects is provided funding with the expectations of directing the support toward key milestones recommended by the review committee.

Despite similarity in structure, the goals of the two seed grant programs are distinct and complementary in their effort to support translational research. The IDDA Support Program provides small grants to facilitate collaborative translational research and partnerships related to drug discovery through support for team building across diverse disciplines, the identification of complementary expertise across the Indiana CTSI, and, when necessary, the creation of new, external partnerships. The IDDA advisory committee is motivated by a desire to connect complementary biological and chemical expertise. From a drug discovery perspective, this includes the design and development of new assays based on the identification of a novel biological target, the identification of potential hits from an in silico screen or actual hits from a high-throughput screen, or the validation of target within a particular disease or the confirmation of the activity from a screening hit through follow-up experimentation. When possible, funding is provided to research service cores within the Indiana CTSI consortium and not directly to the PIs associated with the project.

This indirect support mechanism is an approach offered by SPARK Translational Research Program at Stanford University School of Medicine. This program provides technical expertise and $50 000 per year for up to two years to support the translation of Stanford University discoveries into new therapeutics and diagnostics to addressing unmet biomedical needs. As discussed previously, funding is not provided to the academic lab but directed toward partners that would provide valuable data necessary for continued progress. Grantees must participate in weekly seminars and present quarterly updates. These gatherings provide an opportunity for teams to learn from one another's experiences and interact with industry experts who also attend. Participants benefit significantly from their location in the tech world of Silicon Valley which provides routine access to biotech industry experts, serial founders, and venture capitalists. The SPARK program was created by a professor of Stanford Chemical and Systems Biology, Daria Mochly-Rosen, and is currently codirected by Kevin Grimes. The two recently coauthored an excellent primer on drug development that seeks to fill in the gaps between an academic view of drug discovery and the actual industry experience. It not only provides a fairly detailed overview of the steps in the development pathway but also provides experiential anecdotes from their SPARK experience (180). There is a strong argument that ADD centers should provide a copy of “A Practical Guide to Drug Development in Academia” as an essential part of their support efforts.ADD projects often include the development of a structure–activity relationships (SAR) based on a particular hit obtained from a screen or other biological assay. An SAR is developed through an iterative process of chemical modifications and biological testing seeking to identify critical structural features necessary for the compound's biological activity. In striking contrast, early-stage drug discovery programs in the pharmaceutical industry include optimization of pharmacological properties such as pharmacokinetics/ADME (absorption, distribution, metabolism, and excretion) and preliminary toxicological profiling. Funding for these types of experiments is often difficult to obtain despite the fact that the data provides critical information necessary for the identification of a lead or development candidate for animal studies and provides justification for more rigorous IND-enabling preclinical studies (165). Therefore, the advisory committee often advises PIs about the structures obtained from preliminary screening helping to ensure that researchers are focused on classes of molecules that are amenable to optimization and not pan-assay interference compounds (PAINS) (181) or classes of compounds with known cytotoxicity. From complementary perspective, we created the MTP Service Grant Program. This opportunity provides support for hit-to-lead development studies through pharmaceutical development services provided by local CROs. The Indiana-CTSI negotiated contracts with two CROs, frequently used by Eli Lilly for early-stage hit-to-lead chemical assessment in particular to provide Tier 1 ADME services including but limited to solubility, metabolic stability, and protein-binding evaluation as well as Caco-2 drug transport assays. In addition, in vivo pharmacokinetic evaluation and brain–plasma ratios are also valuable pieces of data. Finally, mutagenicity, hERG cardiotoxicity screening, and general toxicity profiles can support or eliminate further studies of potential leads. A number of ADD centers offer a selection of these services through internal research core facilities. Unfortunately, these services can be expensive to maintain in part due to the infrequent use and do not always provide the data quality essential to justify go-no-go decision-making for PIs or reviewers of external financial resources such as venture capitalists and funding agencies.

The MTP contains a number of programmatic innovations that complement the broad translational goals of the other programs within the Indiana CTSI. These benefits include routine feedback from the advisory committee's pharmaceutical experts, implementation of a project management paradigm, and financial support for CRO-generated preclinical data. Drug discovery projects supported by the MTP benefit from combined intellectual resources and alternative perspectives provided by the IDDA advisory committee which complement the PI who is often inspired, ultimately supported, by objectives rooted in more basic research discovery. Drug discovery is very much an iterative process where a number of steps must be accomplished sequentially, and each step is informative to future decision-making. A project management approach tracks the progress of the project and helps to ensure the project remains on a trajectory through the translational pipeline toward a potential clinical benefit (182). Over the past seven years, the MTP has been able to connect and provide mentoring advice to researchers and research teams associated with more than 70 projects across the Indiana-CTSI-associated institutions. More than 25 projects have received some financial assistance, but these and several more have been placed into our project management portfolio for routine oversight. Three projects have received a second award for significant progress made along the drug discovery pathway and maintain momentum.

3.2.1 Novel Protein–Protein Interaction (PPI) Inhibitors that Target Cancer Signaling Pathways

Indiana-CTSI researchers utilized in silico screening to identify potential classes of compounds that could affect the transcription factor TEAD interactions with YAP, a coactivator that has been implicated as an oncogene and is amplified in a number of human cancers. The IDDA support program provided funds to screen hundreds of compounds in an in vitro assay. The results of this screen provided four compounds that bound to TEAD and inhibited the interaction. Based on these results, a second round of funding was considered to synthesize a small library of compounds related to these hits and characterize their activity in biochemical and cellular assays of TEAD activity. The IDDA advisory also suggested that the library be screened for selectivity with a number of known PPIs. The results of these studies serve as an important basis for not only an NIH R01 application but also for the next stage of the drug discovery pathway preliminary preclinical data collection and animal model studies (183).

3.2.2 Development of Small-Molecule Inhibitors for Novel Cancer Therapeutic Target PRMT5

Indiana-CTSI researchers discovered protein arginine methyltransferase 5 as an activator of NF-kB, a factor shown to be constitutively activated in colon cancer. In 2014, the IDDA support program provided funds for the development of a high-throughput screen of PRMT5-specific AlphaLISA technique and identified a number of hits (184). Biochemical studies validated a few lead compounds, and additional funds were provided in 2015 to generate a small library of analogues based on the most active lead. Current efforts have begun to explore the activity of the prepared derivatives in colon cancer cellular assays. The results of these studies serve as an important basis for not only an NIH R01 application but also for the next stage of the drug discovery pathway preliminary preclinical data collection and animal model studies.

3.2.3 New Therapeutic Lead Compounds for Neurodegenerative Diseases Associated with Aberrant Proteostasis

Indiana-CTSI researchers have shown that a bacterial-derived, natural product corrects the aberrant cholesterol-storage phenotype associated with the lysosomal storage disease, Niemman-Pick Type C, and specific forms of AML. The researchers access gram quantities of the material through fermentation of Streptomyces chromofuscus. Through an award from the MTP Service Grant Program, the lead compound has been evaluated for solubility, in vitro metabolic stability, P-gp assessment, and murine pharmacokinetic profiling. These results suggest that the natural product is a strong lead compound and worthy of testing in animal models of both diseases and helped secure an NIH grant to study the therapeutic potential of the molecule (173).

3.2.4 Novel Lead Compound for Diabetic Foot Ulcers

Indiana-CTSI researchers have developed potent and highly selective inhibitors of matrix metalloproteinase-9 (MMP-9) which are water soluble. Inhibition of MMP-9 has been shown to lower inflammation and enhance angiogenesis. Topical application of the lead compound has been shown to accelerate diabetic wound healing in mouse models of the disease. Diabetic foot ulcers (DFUs) are a significant health problem. An award from the MTP Service Grant Program was leveraged to obtain additional funding from the University of Notre Dame and support IND-enabling toxicological studies (185).

4 Additional Collaborative Opportunities

Despite the advent of combinatorial chemistry and high-throughput screening in the 1990s, which enabled the preparation and screening of large, diverse compound libraries, the number of drug candidates in clinical trials did not scale proportionally (186). Major factors contributing to this fact were that screening libraries were limited in quality and chemical diversity. Over the past 20 years, the process has evolved from simply a number game to specific advances in the quality of biological assays and the creation of mutually beneficial research partnerships even between pharmaceutical companies. Compound sharing is becoming routine in drug discovery as an efficient way to rapidly couple increased chemical diversity to biological screening opportunities (187). Recent examples include agreements between Sanofi, Syngenta, and AstraZeneca (AZ) to exchange 240 000 compounds, and AZ has also partnered with Bayer to provide access to each other's entire compound library. Molecular contributions for a European consortium of seven companies as well as compounds from several academic centers have created the Joint European Compound Library (JECL). Unfortunately, academic institutions in the United States are more reticent about compound sharing due to, somewhat unfounded, concerns about IP. Thus, efforts to create a statewide compound resource never got traction. More recently, however, Notre Dame's Warren Center for Drug Discovery established a new relationship with the Boston University – Center for Molecular Diversity (BU-CMD) as a part of a larger consortium of molecular providers. Under the arrangement, the BU-CMD will curate unique chemical entities from chemistry sources within six academic institutions. The compounds would be made available to a complementary network of screening facilities and expand their own compound collection. Intellectual property concerns are alleviated through the use of material transfer agreements and consideration of the relationship as one-on-one collaborations despite the larger programmatic organization. While some of our core facilities have purchased commercial libraries, a majority of the unique compounds currently sit idle on shelves or in freezers of a number of laboratories and represent a largely, untapped resource for drug discovery. Compound sharing across institutions can be mutually beneficial and benefit from the utilization of common protocols for compound and chemoinformatics data storage.

5 Summary

Due to its simultaneous contributions to fundamental biological discoveries and accompanying training next-generation researchers, ADD should remain an important contributor to the field. With the dramatic increase in cost and time to develop new therapies, the pharmaceutical industry has contracted particularly in the support for early-stage biomedical research. Academic scientists are well positioned to fill this gap, but unfortunately, funding for academic research has waned, until very recently, over the past two decades. However, universities have responded by tapping into alternative sources of financial support including foundations and philanthropy in order for ADD to expand their role to a larger portion of the development pathway. In order to ensure the generation of a sustainable platform and establish a more reliable return on investment, support structures from ADD must complement the skills and technology from what is typical to individual research laboratories. Biological experts can provide critical evaluation of proposed screening assays and provide insight into target validation for specific indications. Chemical expertise is useful in ensuring the proper selection of screening libraries and the optimization of compound structures to provide leads with drug-like properties. While computational tools have long been utilized for structure-based design, the application of newer artificial intelligence and machine-learning approaches to compound property prediction can significantly help academic researchers predict PAINS (188), toxicological profiles (189), and ADME profiles (190). Expertise in both of these areas need to be complemented by pharmaceutical industry experience. Industry experts can focus projects toward key translational goals and away from conflicting academic priorities. Universities can access these needed skills by creating policies that promote mutually beneficial partnerships with the pharmaceutical industry. Seed grants should be used to engage researchers and garner information about the status of current projects which would assist in developing a needs assessment strategy. When funding is provided, support should be focused on obtaining key pharmacological data that helps shepherd the project from academic research lab into early-stage start-up entities. There is no doubt that the current environment provides expanding opportunities for academic contributions to the development of new therapies despite the increased risk placed on early-stage research.

Finding quality chemical matter from academic screening sets can pose some unique challenges. These sets are usually not proprietary but are obtained from commercial vendors. The proportion of promiscuous, unstable, and reactive compounds is higher than the typical pharma proprietary library, whose collections have been designed for physicochemical properties and stabilities which are more druglike. Even in relatively clean libraries, these undesirable compounds rise to the top due to activity across multiple assays. Complex solution behavior of these collections, including very low aqueous solubilities and aggregation, must also be dealt with to select reliable leads for optimization. We have come to these conclusions after conducting target-based high-throughput screening for over a decade on scores of targets. More recently, we conducted an accelerated screening program of about two dozen novel targets over a three-year period using a library of around 200 000 diverse compounds from commercial vendors. This experience has resulted in a workflow to address the liabilities of poor screening matter, avoid wasting effort on risky compounds, and focus on the most promising leads. As more academic laboratories engage in industry collaborations around drug discovery, it is important that the academic laboratory has thoroughly vetted the quality of the chemical matter they bring to the negotiating table. The disclosure of well-researched discovery leads from academia will go a long way to encourage these collaborations and mend the somewhat sullied reputation of academic discovery lead matter (191).

The building and selection of a new screening library should include the computational filtering for undesirable functionality and properties. Compounds which are unstable in DMSO, aqueous buffer, or air should be flagged with suitably constructed SMARTS queries (192). Examples of the sort of functionality in the design of these SMARTS queries include compounds such as nonalkali metals, extended fused aromatics, and functionality with extremely high or low pK_a values which preclude their cell permeability. The exact nature of these filters would depend on the investigators' needs and biases, and there are published SMARTS filters available (193). In total, we have designed over 300 SMARTS queries which can reduce vendors screening offerings by half. These SMARTS filters include published lists of pan-assay interference compounds (PAINS) (194, 195). These queries for undesirable chemical functionality can be followed by filtering for custom physicochemical properties such as molecular mass, topological polar surface area, and calculated octanol/water partition coefficient as examples. Vendors are aware of drug-like physicochemical properties and PAINS and may offer collections prefiltered for these properties. However, we have found that after full SMARTS and properties filtering, a vendor's collection can be reduced to as low as 20% of the initial offering. We typically conduct this filtering in MOE (196), but most computational software suites can either directly filter using SMARTS queries or one can develop a KNIME protocol (197) to filter and calculate the physicochemical properties.

In the case of the existing libraries which have not been prefiltered, this analysis occurs in the hit follow-up or triage phase of the screening program where questionable chemical matter is flagged for a critical review by the medicinal chemist (198). Hits meeting a predetermined activity threshold are computationally filtered using PAINS and SMARTS queries. We assign a three-tiered ranking system to all hits consisting of green, yellow, and red categories, corresponding to no known issues with green compounds, problematic but fixable functionality with yellow compounds, and red compounds regarded as unsuitable for follow-up due to reactivity, stability, or known promiscuous behavior. This “stoplight” analysis can be an effective method to communicate the complexity of high-throughput screening to faculty collaborators unfamiliar with the complexities of selecting quality lead matter. A final assessment of potential promiscuity is conducted by searching individual hits and their substructures in online bioactivity databases such as SciFinder (199), ChEMBL (200), UniProt (201), and Binding DB (202). This final check of reported activities is important since actives in our assays often show up as frequent hitters in these databases. In many instances, the mechanism of promiscuous activity is not obvious, and an online check of these databases can alert one to a potential new PAINS. Once this triage analysis is complete, we can focus on the best chemical matter and begin assessing the next critical property which is solubility.

The assessment of complex solubility behavior in screening is difficult to predict computationally, and we therefore rely on physical measurements. Compounds must be soluble throughout the dose–response titration. Behaviors such as failure to reach 100% inhibition or steep Hill slopes (203) are warnings that activity may be solubility limited, or aggregation may be responsible for activity. In addition to these readouts, we consider any potential lead compound having a calculated logP octanol/water of greater than 3 to be suspect, and the aqueous solubility and tendency to form aggregates should be determined.

Our method to determine problematic solubility and aggregation behavior starts with a kinetic solubility determination. Plate-based nephelometers are commercially available for high-throughput determinations, but we find a simple low-tech single-cell nephelometer or turbidimeter to be superior in sensitivity and ease of operation. In this method, a stock solution of the compound of interest in DMSO is added incrementally to a large stirred cell containing PBS or assay buffer. We typically prepare these stock solutions at 10–20 mM. The addition of the compound DMSO solution to the buffer spans a range of 1–100 μM solubilities, and the final DMSO concentration is kept below 5%, or whichever value reproduces the assay conditions. The reflectance at 90° to the incident light is monitored by a nephelometer or turbidimeter, of which several are commercially available and inexpensive. A plot of reflectance vs. concentration gives a straight line while the compound concentration is soluble but rapidly curves upward as precipitating or aggregating particles causes increasing reflectance. Figure 23 shows three types of solubility behaviors and their characteristic plots. The kinetic solubility of beta-estradiol illustrates the near ideal behavior for a well-behaved solute approaching its aqueous solubility limit at a literature value of approximately 13 μM. We could have confidence in any measured activity higher than this solubility threshold. By contrast, Compound A does not achieve a level baseline which indicates solubility, and the reflectance rises from the first addition of DMSO solution. We cannot have confidence that either a single-point activity value or a dose–response activity would be meaningful. We cannot determine the nature of the particles, whether they are aggregates or a solid precipitate, however. We would not have confidence that any activity values are reliable for this compound and would normally choose to flag this compound as low priority or disqualify it for follow-up altogether. Aggregation itself does not disqualify a compound as a potential drug; in fact, several marked compounds have been shown to form aggregates (204). However, in an academic discovery lead finding campaign where promising leads must be selected based on screen activity from among several hit classes, the presence of aggregation or solubility limited behavior, which would consume resources in the search for a “fix” which may never materialize.

Aggregation is characterized by a critical aggregation concentration, CAC, which defines the phase change from solute to the inhibitory aggregation state. The presence of a defined CAC as determined by DLS is a strong evidence of colloidal aggregation (205). Finally, Triton X-100 displays a typical behavior of surfactants which show maximum reflectance just below its critical micelle concentration (CMC). Reflectance then drops following the completed phase transition to micelles at concentrations above the CMC. The Triton X-100 stock solution was from a 10% aqueous stock rather than a DMSO solution. Since surfactants such as Triton X-100 are recommended to disrupt aggregation in a biochemical assay, they are typically used at concentrations higher than their CMC. The loss of compound inhibitory activity on addition of surfactants is another hallmark of aggregation-based inhibition (206).

The white light source from these instruments may be filtered to pass only the red light (610 nm) to avoid fluorescent interference. The use of a stirred cell, as opposed to a plate reader for these measurements, allows higher accuracy, longer path lengths for greater sensitivity, and more reproducible values. Obviously, the design of the nephelometer must allow the inclusion of a magnetic stir bar in the cell with no obstruction of the light path. Some instrument models do not allow this. However, the design of a nephelometer is straightforward, and an instrument could be easily constructed in collaboration with an analytical or engineering department in most universities.

The determination of solubility via nephelometry is not without its caveats. The compound should be relatively free of insoluble matter such as silica gel. Aggregates may be forming, which are beyond the sensitivity of the nephelometer. Compounds which can form micelles give a characteristic initial upward inflection followed by return to baseline at the CMC. In addition, in our hands, some compounds will aggregate at submicromolar concentrations forming sub-micron-sized particles while simultaneously precipitating larger particles, suggesting these are two independent phase transitions from solute. The nanoparticle formation and aggregation may be further characterized after initial nephelometry using dynamic light scattering. Compound A (Figure 24) was analyzed with DLS starting at the low 4 μM concentrations where the nephelometry reflectance was weak. DLS scattering showed a population centering around 80 nm and much larger micron-sized particles at 1–10 μm. Population analysis shows these larger particles increasing dramatically with higher concentrations (not shown).

The workflow we have presented addresses the two most pressing issues we have experienced in assessing high-throughput screening campaigns, unsuitable chemical functionality and problematic solubility. These problems can be effectively addressed with a minimum of effort and cost in an academic setting where resources are often limiting.

Historically, pharmaceutical companies were known to be quite generous with supporting academic research. Companies often awarded unrestricted grants that served to support, either directly or indirectly, postdocs in top-tier academic laboratories. The hope was that those trained scientists would eventually find their way into highly competitive roles within their own industrial laboratories. For the interested reader, a number of editorials and case studies provide historical snapshots of academic collaboration trends over time (207-215). Government-based sources of funding for academic research have always been competitive, but historic funding levels by national agencies such as the NIH and NSF continue to decrease, which has led to significantly increased pressure on academics to find alternative sources for funding their research (216). Despite long-standing concerns over reproducibility of data generated in academic laboratories (217, 218), many scientists across the pharmaceutical and biotech industry are routinely collaborating with academic colleagues when it comes to developing enabling tools and knowledge such as in vitro and in vivo models and new technologies for screening and imaging. Some of this partnering is being conducted under the banner of open innovation (219). There is a good deal of partnering conducted in a relatively informal manner at no to minimal cost, but most is conducted only after having established formal material transfer agreements and/or research agreements. In an effort to minimize transaction costs and perhaps create more of a free flow of information and materials, some companies have worked to establish formal institutional relationships with academic institutions and/or research institutes, taking advantage of natural geographical relationships or with the strategic intent of expanding their geographic footprint (219).

1 Conception of LRAP

As the business model within the pharmaceutical industry has continued to evolve, limited resources have been increasingly diverted to support portfolio-enabling priority work, and unrestricted grants appear to be rarely offered in the present environment. In 2010, a group of senior scientists at Eli Lilly and Company sought to reverse this trend in part. They successfully recommended to Dr. Jan Lundberg, the then president of Lilly Research Laboratories, that limited financial resources be set aside in reserve to promote academic partnerships and scientific development of Lilly's best scientists. One of the work streams led by Drs. Jim Stevens and Stephen Burley (distinguished Lilly fellows – investigative toxicology and structural biology, respectively, at the time) worked to propose a new approach to partnering for Lilly called the Lilly Research Award Program (LRAP). The goal of the LRAP program was to support excellent precompetitive science independent of Lilly's direct portfolio-facing work. The hope was that funded research would focus on discovery, development, and technology/methodology-related areas that might serve to augment innovative science at Lilly. Peer-reviewed publications and external scientific presentations of the collaborative research were the main expected outcomes, but if the work were to lead to patentable inventions, a free nonexclusive license to leverage the work was anticipated to be included in the overall agreement.

A critical first step was to work with Lilly's legal team to help define an agreement that would serve to be seen by most academic institutions as an equitable and easily executable research agreement. It was also important to limit the resources required by Lilly legal to get the agreements executed in a reasonably rapid period of time with a very limited amount of negotiation. Lilly had precedent for building contract efficiencies after establishing alignment with the Association of University Tech Transfer Managers (AUTM) to craft an equitable special Material Transfer Agreement that proved to be foundational for Lilly's premier open innovation initiative. That collaborative program was Lilly's phenotypic drug discovery (PD²) program, which evolved and expanded to include other opportunities for crowdsourcing and collaboration. It was subsequently renamed the open innovation drug discovery (OIDD) program (220, 221).

2 LRAP Process

To efficiently initiate the LRAP process with a potential principal investigator, a concise letter of intent (LOI) was crafted, and a process was established for LRAP to have the LOI signed by the appropriate institutional representative. The LOI informs the target institution and academic principal investigator of the key details of the LRAP process and eventual agreement if awarded. This includes (i) a deadline for reaching an agreement once the award is granted, (ii) granting of a nonexclusive license to Lilly of any intellectual property generated as a result of the collaboration, and (iii) appropriate recognition of Lilly scientists' contributions to the research in any publications.

After an internal call for proposals is issued, scientists at Lilly work with their selected academic principal investigator(s) to submit a brief research summary (∼1 page) proposal to the LRAP team. The primary purpose of the research summary document is to ensure that the scope of the submission will adhere to the precompetitive guidelines of research under LRAP. Feedback is provided at that time to help guide the scientists on scope (e.g. is it exciting, and will it be doable?), and any potential legal questions are addressed early. A signed LOI and full proposal is then submitted to the LRAP team. Three internal scientific peers are selected to review the proposal, and one is identified as the primary reviewer. Notably, the authors are blinded to the expert reviewers selected. The reviewers are asked to make a recommendation to advance the proposal (or not) for a full committee review. If a proposal is advanced for review, the primary reviewer is asked to present the proposal's strengths and weaknesses (15 min including Q&A) to the Scientific Committee, which is composed of senior scientists (not Senior Management) from across Lilly. The committee review is conducted similar to the approach used in an NIH study section. Scores from the committee review are used to inform the final decisions for funding. For proposals that are not funded, by rule the authors can elect one time to resubmit a reworked proposal that sufficiently takes into account written feedback from the review process.

Once an agreement is signed, an initial payment (25%) is made to get the project going. At the end of year 1 of an agreement, the Lilly investigator and collaborator prepare a short summary progress report that serves to inform a decision on authorizing a second payment (50%). A final summary report is required prior to issuing the final agreed payment (remaining 25%). Some projects have been cancelled due to insufficient progress on the science, and others have been rescoped based on breaking science. Sometimes, given delays in project progression or just interest in continuing the collaboration, no-cost extensions to the original agreement have been executed.

Since the initiation of the program in 2011, nearly 600 proposals have been reviewed in approximately 15 funding rounds, and at the time of writing about 140 have been fully funded, constituting a commitment that totals >$35M to date. Proposals have been initiated from all of Lilly's key R&D sites, all therapeutic areas, and functional areas ranging from discovery to development, including regulatory and global health outcomes. More than 80 institutions have been engaged in LRAP agreements. Many of the agreements have been signed in <90 days and with minimal negotiation on the details of the agreements. The program is managed by a small group of volunteers who make up the LRAP operations team, and their combined effort represents approximately 0.5 full-time employee commitment. At the time of writing, we have logged more than 70 peer-reviewed publications (just from the first 11 LRAP rounds) that have issued across a broad range of disciplines citing the collaborative work conducted under the banner of the LRAP program. A select set of seven example publications are cited below that demonstrates the type of precompetitive work being conducted through this unique approach to industry–academia collaboration. The first example (222, 223) was a collaboration to characterize the binding modes of various ligands to several GPCR targets (M1, M2, and M5 muscarinic acetylcholine receptors) via X-ray crystallography. The second example (224) highlights some of the outcomes of a project seeking to develop new in silico-based approaches to improving process chemistry efficiency as part of Lilly's investment in Green Chemistry production approaches. The final example highlights a collaborative effort to better predict, via in silico methods, the potential existence of novel (e.g. more stable) polymorphic crystal forms of drug candidates that might prove to complicate pharmaceutical development and commercialization (225-228).