“When you discover something novel the medical imperative is to come up with a good use for it”
….. Paul Janssen
I was approached to report on a Reuters webinar to push forward drug discovery techniques in biologicals, my interest was immediately piqued. Well it was Reuters and who better to tackle the methodology and business models needed for this difficult discipline. When I heard it was Andy Hopkins and Chris Lipinski speaking, well it was a no-brainer I knew these guys would cover everything needed in one hour without going over old-ground which is exactly what they did.
No long drawn out lit review here, you know your stuff. You’ve read the papers, you’ve read about the failures and know that 60% of all biologicals are failing in phase III which also happens to be the most expensive phase. What we’re going to here is build on that. Let’s go outside the ecosystem and start to draw from other industries, well that’s what they do with social media, why can’t we do that with drug discovery?
We start with where we’re at. We’ve moved into poly pharmacology, promiscuous ligands and genes. We need to know which of the 23,000 human genes your protein is reacting with and what that means for the human process. We no longer talk about single targets, there’s 23,000 of them, we talk in terms of your single gene and how it’s going to affect the human physiology. We don’t even need to know the process of the disease, we need to know the phenotypical profile of the disease, that way we can predict which gene will be promiscuous with your ligand. Should we say drug discovery, no we say exploration and this is where Andy and Chris start painting outside the lines.
We go outside the ecosystem and start talking about the steel industry. The technology and machinery changes in the steel industry but the building blocks don’t. The talk references Paul Janssen, the man who invented the first chemical database. Janssens’ model was one of exploration. He got a compound with good solubility, good ADME properties and he ran with it. Janssen worked with the same basic frame adding to it to build a library of balanced compounds, producing the molecule first and then finding the indication. He knew they were efficacious and he knew the disease processes, all he did was match the drug with the disease. In his career he launched 80, yes 80, drugs to market.
Thus biological drug exploration is born, if you enter using high weight molecules you’ll fail, you’re still thinking mechanistically, bring the weight down and EXPLORE your molecule and the reaction. Build the database. It’s all about knowing your phenotype and the biological process your molecule will cause. You’ll need to be profiling and building separate human and disease phenotypical databases as well. It would be a good idea here to follow the finance industry and employ a data scientist if you don’t already have one. Data scientists are used by blue-chip firms to collect big data, analyse it and deliver the data as intelligence that is actionable. They’re the guys that deal with big data and if everything goes to plan these databases will be as big as they come.
We then move swiftly into business models, the age of the productive biotech is long gone in biologics with companies such as Pfizer, Organon, AZ, GSK, Roche and Lilly leading the way. Corporate patents have now overtaken biotech patents with a 60/40 percentile split. The big 3 patent disease areas are Cancer, Alzheimer’s, and Obesity and are completely led by the baby boom business strategy (BBBS). We listen to a description of the well-built but rigid infrastructure of big pharma, highly productive and well-funded but unlikely to react quickly to new technology. Cue the cool, biology-rich academic/biotech partners. Entrepreneurial & quick to enter the unknown technological realm, highly focused but in small groups with little funding and the global blue-chip power any drug needs to make it.
The importance of biological databases is mentioned in both models, pharma has them, biotechs & academia can’t afford them. We can talk about drug repurposing in biologics, we can talk about about finding that perfectly timed ph I or ph II B biotech/academic programme and we can talk about risk-sharing as the USA pumps more money into it’s university programmes, but we also have to talk about the fact that big pharma needs biotechs/academia to do the ground-work. So let’s start talking about using the internet as a model here, let’s start talking about crowd-sourcing.
One of the biggest complaints I hear from big pharma is finding new talent, new biotechs/academics at that perfect point, in the perfect phase of development where the risk is minimal. So more often than not the ‘tried and trusted partner’ is backed time and time again making it extremely hard for new biotech talent to surface and easy to miss workable biological compounds. Drug repurposing is also a major strategy with pharma at the moment. When we speak in terms of biologics drug exploration, repurposing takes on a new meaning. We’re not just talking about contra-indications we’re talking about going to back basics, approved biologics with good efficacy, good promiscuity and good ADME properties. We’re talking about re-mapping gene promiscuity and building on the molecules that work going back to Janssen’s model and building a biologics database from that original safe molecule.
Biological programmes without animal models are also spoken about in hushed tones within pharma circles. With only 1 in 67 biologics programmes surviving efficacy studies and an even higher attrition rate for biologicals in later phases, how useful are animal models in biologics? Would it be possible to host clinical trials without them? Your data scientist and crowd-sourced database is the answer here. Expert novel algorithms born of crowdsourced databases predicting your efficacy and datamined. We move past the old single-target DD PKPD and move into highly sophisticated modelling programs. This cannot be done without academic/biotech assistance.
The way forward here pharma is crowdsourcing, releasing all, giving free database access to verified biotechs and academia. Imagine putting together a molecule and a week later going back to your database to find a university in Japan has run with it and mapped a set of promiscuous genes for you. Imagine the physics grad who logged into the global biologics database and noted the cash reward for balancing the biologics molecule you just built. Gamification and reward schemes are a viable incentive here, an easy way to discover new talent as well as providing new algorithmic technology. This kills every bird in the park with one stone. Academia/biotechs don’t have your funding or access to your databases and you need to tap into the new technology and rich biological expertise provided by these research groups. Crowdsourcing all databases is not a suggestion it is a necessity for biologics and with 3,000 patents in cancer biological molecules expected in 2012 alone, you need someone to run with these.
You also create more partnership options here. IP-wise you have partnership rights over anything created from your molecule and this works both ways, instant partnerships formed over an open-system. Biotechs get access to these crucial but expensive databases and you have more start-up’s logging on, more talent finding you so to speak. This also allows biotechs/academia to specifically choose what to run with having more choice to correctly match their platform to the ideal molecule. No one knows their platforms better than them after all. Will we see more races to the finishing line as multiple biotechs pick the same molecule to run with? Food for a new and faster competitive model.
In conclusion we’ll see more drug approvals in smaller markets, this will be an eventuality without crowdsourced databases as the big three diseases begin to have their patent stream throttled without the said patents being further developed. Risk-sharing and academic collaborations are a must as they have more expertise in the area. Poly pharmacology, phenotypic screening, drug repurposing all lead to big data so management of multiple genomic databases are key. Expect to have databases profiling disease phenotype, patient genotype, biological molecules and gene promiscuity with sub-databases. Think exploration as opposed to discovery, this is all about profiling and targeted promiscuity. Bring your complex molecular weight down to make the reaction less complex and begin to hone and map every single reaction your molecule produces. Use the specific nature of neuroscience here, where every single link and circuit within the brain is checked and mapped by highly sensitive imaging techniques and agents in cloud-based atlases. Don’t forget to look for those new imaging techniques and optimising agents, this is how neuroscience is moving so quickly! And finally, Chris Lipinski advises that it’s better to be lucky with your molecules than smart…. Explore!
As with all Healthinnovations articles this is open to discussion and innovation.
Click on the link to listen to Thomson Reuters, The Changing Face of Drug Discovery: New Approaches, New Skills and New Technologies
Can everyone read Gary Pisano please – Pisano, Gary P. Science Business: The Promise, the Reality, and the Future of Biotech. Boston: Harvard Business School Press, 2006.