Going with the Flow
Three researchers at the Massachusetts Institute of Technology – chemical engineer Klavs Jensen, chemist Tim Jamison, and crystallization expert Allan Myerson – decided to collaborate and bid for a DARPA grant, resulting in a pharmacy on demand system based on continuous chemical flow processing.
Having previously collaborated in the MIT-Novartis Center for Continuous Manufacturing on the development of a scaled-down, end-to-end flow chemistry process to manufacture tablets from simple chemical inputs, Jensen, Jamison and Myerson were well-positioned to respond to DARPA’s call for PoD proposals. However, the original system, although much more compact than a normal pharmaceutical process, was the size of a shipping container – hardly the portable device that DARPA was looking for.
“Since then, however, advances in flow chemistry have expanded the chemist’s toolbox, allowing for faster reactions in smaller vessels,” says Jensen. “We have now developed a fridge-sized, continuous flow system that can be reconfigured to produce a variety of different small-molecule drugs – with different chemical structures and synthesis routes – to US Pharmacopeia standards.”
Considering that a normal pharmaceutical process operates in large batches requiring big vessels, squeezing it into something portable isn’t straightforward. Part of the solution lies in the geometry of the upstream reaction tubes. According to Jensen, the right design means that the chemical synthesis reaction can be heated and cooled more quickly, achieving higher temperatures and completing the reaction in less time than a traditional batch process.
The downstream process required devices capable of crystallizing and purifying the API output from the upstream process. Initially, the team focused on liquid formulations. “The original project brief specified that the drugs would be used within 14 days, which means that solutions or suspensions are acceptable – these don’t have a long shelf life, but if you’re making medicines on demand and using them quickly then this isn’t a problem,” says Myerson. “Subsequently, DARPA has funded an additional project focusing on solid formulations. We’ve now built a device that blends the API with excipients, and forms the mixture into tablets. We’re testing this now, but dealing with powders on such a small scale has been difficult.”
“Building the whole system has been a significant challenge, especially in terms of making it relevant to the needs of DARPA. Counter-intuitively, it has resulted in fundamental research leading to new chemistry and other new technology,” says Jamison. “For example, because we were unable to source commercial equipment suitable for the scale on which we needed it to operate, we had to develop many mechanical components ourselves. One challenge was designing a pump that worked reliably over extended periods with many different chemicals.”
Initially, the team experimented with simple pharmaceuticals, such as diphenhydramine, lidocaine, fluoxetine and benzodiazepine, but they are now working to broaden the range of molecules that the system can manufacture. Recently they’ve been looking at drugs with more complicated structures: ciprofloxacin and doxycycline.
Jamison adds, “Keeping the device relevant to real-world needs has been a fundamental requirement of DARPA from the very beginning, and a longer-term goal is for the system to be useable in the field by non-experts. DARPA also had very specific requirements regarding the number of doses that the equipment would be required to make. It wasn’t enough to just run the process for 30 minutes and declare victory!” The system has been designed using a “plug and play” philosophy that allows components and units to be easily changed. For example, if it’s not convenient to clean the system by flushing through a solvent then the contaminated tubing can be easily replaced and discarded after use.
“This type of device isn’t just useful for the battlefield. For example, there’s an industry trend towards drugs that target genetically defined populations – and manufacturers of personalized medicines would certainly benefit from flexible, fast production technologies,” says Myerson. “Some companies are also interested in the potential of the technology for the cost-effective manufacture of clinical supplies in low volume.”
“There are also benefits associated with the uniquely mobile nature of the system. It can be put in the back of a truck or on a plane, and it doesn’t require much power, so it’s ideal for remote locations,” adds Jensen. “Others have raised the possibility of pharmacy-on-demand devices in drugstores and hospitals, so that organizations can make some drugs as required rather than keeping large, limited shelf-life stocks.”
As Tyler McQuade mentioned, however, getting a distributed manufacturing system – particularly one that proposes to manufacture multiple drugs from a single device – to comply with regulatory requirements will be a challenge. “Essentially, our PoD system is no different from a pharmaceutical plant that makes several different drugs at one site,” says Jensen. “In traditional manufacture, a site and process approval would be required for each product. In our case, as well as the device itself, each flow process would need approval.”
Despite the challenges that lie ahead, the team are confident that portable systems will, in time, lead to important benefits. Jamison says, “Our post-docs and students put a huge amount of effort into this, and when it worked it was like a moon shot; we all felt that something new and important had been achieved.”
Klavs Jensen is Warren K. Lewis Professor of Chemical Engineering, Timothy Jamison is Professor of Chemistry, and Allan Myerson is Professor of the Practice of Chemical Engineering, all at MIT.
<link issues protein-synthesis-but-not-as-we-know-it>Protein Synthesis – But Not as We Know It
