Space Age Drug Development
Professor Lee Cronin, Regius Chair of Chemistry at Glasgow University, UK, has interests ranging from the origins of life to making drugs on Mars. Here, he shares his vision for the future of drug manufacturing.
The drug industry problem that DARPA wishes to address is very similar to the challenge I set for myself, and which I articulated in a 2011 TED talk on 3D printing for drugs (1). The problem is analogous to books going out of print in the publishing industry. Prior to the digitization of publishing, you would print a book by setting up a printing press and doing a printing run, but once the stock of printed copies sold out, the book would no longer be available. Similarly, drug manufacture requires constructing a complex and expensive production facility, but the know-how and infrastructure for making the drug is easily lost if, for example, the facility were to be adapted for a different product. This is because the facility is often bespoke. In the laboratory, the fact that some discoveries are done under bespoke conditions often means that it can be hard to understand how to reproduce them. This is part of the reproducibility crisis in chemistry – it is not often discussed , but it can be very hard for laboratories to reproduce each other’s work. This issue of reproducibility, not just in chemistry but in all of science is now actively being discussed. It is a frustrating problem, but then I realized that “digitizing” chemistry could help not only solve the problem, but aid collaboration and further discoveries. This is because the process of digitizing chemistry combines knowledge of both the chemical instructions, the hardware for doing the reaction, and the precise way of executing the instructions complete with analytical data and observation, such that the entire process can be replicated without fail time and time again. I then realized that digitization meant that you wouldn’t need to be a chemist to synthesize chemicals; and then I realized that you wouldn’t even need a human present – the process could be fully automated as long as the system had the required software, hardware and wetware.
Protein Synthesis – But Not as We Know It
In a previous issue of The Medicine Maker, we reported on the work of James Collins, a faculty member at the Wyss Institute at Harvard University and the Henri Termeer professor of medical engineering and science at MIT (1). Collins and his colleagues are not working with DARPA or focusing on synthesizing drugs on demand, but they have developed a method for producing therapeutic molecules on-demand with freeze-dried synthetic gene networks (2). The technique could be used to produce complex biopharmaceuticals that do not require refrigeration – making them ideal for use in the developing world. “The lyophilized format negates the need for a cold chain, and is very simple to use – it requires only the addition of water to synthesize the protein of interest,” explains Collins.
The work of the Collins Lab focuses on engineered gene networks using synthetic biology and systems. “Our work brings together engineers and molecular biologists to model, design and construct synthetic gene circuits, and to use these to reprogram living organisms for specific applications,” he says. “The work stems from the Human Genome Project in the 1990s – the project produced large ‘parts lists’ for different organisms. We want to explore engineering these ‘parts’ into new and useful combinations.”
To create the freeze-dried synthetic gene networks, a mixture of DNA, RNA, ribosomes and enzymes is removed from the cell and adsorbed to a solid support, such as paper. The preparation is freeze-dried and stored at room temperature – and protein synthesis takes place as normal once water is added.
“We have shown that these preparations can be the basis for rapid and inexpensive point-of-care diagnostics such as for Ebola and Zika (3). Now, we are investigating the use of similar cell-free extracts, but non-adsorbed, to make therapeutic proteins on demand (4),” says Collins. “These could be beneficial for providing biotherapeutics in remote locations, such as in emergency relief efforts, or in space.”
Looking ahead, Collins and his colleagues are investigating the advantages of embedding the dried systems into clothing, for example, to serve as sensors to warn of exposure to an infectious agent, or as components of educational kits for students.
Coding chemistry
In my lab, we have developed an “app the industry” approach. Basically, we are pursuing the digitization of chemical space. The idea is to go from molecules to code, and code to molecules; once a manufacturing process is reduced to code, we can use the code to duplicate that process anywhere in the world. This will make drug manufacture very portable and easily distributable. In fact, it could disrupt the pharma industry in the same way that internet file sharing disrupted the music industry, but it’s my job to disrupt. (I hope, however, that our chemical digitization will be enabling rather than destructive!)
In one approach, we have developed a device that can not only 3D-print reaction vessels, but also add chemicals to the vessels, run a reaction and purify the end-product. Essentially, this system can make a drug from nothing more than code and simple ingredients. What are the benefits of this? Remote chemical manufacturing is one key application. To demonstrate this, a few weeks ago we put our simplified version of the system on a nanosatellite, making drug manufacture possible (here we selected a reaction that makes a drug like molecule and the molecule is purified by crystallization) using a remotely operated device 500 km up, traveling at 8 km/s!
This has obvious applications in manned space exploration, where weight constraints limit the number of drugs we can transport. With our system, we can take a pallet of chemicals and a synthesizer, and make any drug, as required. So perhaps Elon Musk will take one of our drug synthesizers to Mars! At present, you’d need maybe fifty or a hundred different chemicals in order to synthesize any known small-molecule drug, but that is only because the technology is so new. Image recognition software needed thousands of lines of bespoke code at the beginning of its development, but now requires much less as the standard libraries are available to be applied to many different situations as a module. Similarly, as we get better at digitizing drug manufacture, we will learn how to perhaps reduce the number of chemicals needed. We may also be able to add in more steps, or develop new approaches to design chemicals that can be used to expand the number of accessible drugs. For example, the chemical outputs from one process could be used as inputs for another process, so you wouldn’t need two sets of chemicals.
The internet of chemistry
The idea is to digitize the synthesis of every known molecule using a common chemistry set. Just as Google Maps records every street, so we will digitize molecules, such that every year the number of digitized, downloadable synthesis code for the molecules will increase, until the doubling time comes down to just years or months. And just as the increasing processing power described by Moore’s Law revolutionized computing – think of the capabilities of your smartphone – so the digitization of chemistry will transform the speed and efficiency of drug discovery.
This chemical digitization – the internet of chemistry – will have huge implications. It will take the manual labour out of R&D, allowing chemists to focus on discovery. Maybe chemists will not need to be spending such long hours in the lab. Rather, they’ll be able to design incredibly complex molecules, and validate the synthetic pathways using software alone before implementing the correct practical solution in the laboratory. Once this occurs, I can imagine that advances in automation will dramatically increase the productive throughput of interesting molecules, as a result of digitization. In this way perhaps drug discovery and manufacturing will be conducted from the computer interface. And for industry, chemical digitization will offer practical, low-cost ways to manufacture drugs. Centralized facilities may be replaced by massively distributed manufacturing, such that drug manufacture directly responds to individual prescriptions. This would reduce costs and eliminate logistical difficulties in drug manufacture; for example, these systems could be easily deployed in remote regions where drug access traditionally is inadequate. It also would enable cost-effective manufacture at low volume, thus supporting personalized medicine. Furthermore, it would reduce drug counterfeiting – if you can access the real drug at reasonable cost, why buy a potentially dangerous fake? Hopefully, big pharma will recognize the advantages, club together and use this digitization approach to streamline manufacturing, reduce costs, increase flexibility and simplify logistics. Distributed manufacturing sites, say a small unit in every city (not a massive refinery, just a small industrial unit), would make the industry more resilient and drug shortages should become a thing of the past. However, I don’t see it ever getting to the point where every home has its own drug synthesizer; the health risks would be too great, and there would be no incentive to do-it-yourself because getting it made up at the pharmacy would be so cheap.
Nevertheless, developing the technology will require much investment and collaboration. In particular, it may be tricky to ensure that drugs produced from distributed facilities comply with regulatory and safety requirements. We’ll have to address this issue at some point, but looking ahead, I am confident that the digitization of chemistry, and the development of new synthetic methods using networks and robots, will generate a vast number of new markets. We may end up doing chemistry in the Cloud; certainly, research will be transformed and many more molecules than are available now will be discovered. Low-cost drugs of improved efficacy will become available to everyone on the planet, just as low-cost mobile phones are now almost ubiquitous. And when Elon Musk gets a headache on Mars, he’ll be able to synthesize his preferred analgesic!
<link issues protein-synthesis-but-not-as-we-know-it>Protein Synthesis – But Not as We Know It
Lee Cronin is Regius Chair of Chemistry at Glasgow University, UK.
