Reducing the costs of biopharmaceuticals: a bioprocess development approach
Updated: Apr 10, 2020
In 2018 the generated revenue by biopharmaceuticals is estimated to be approximately 200 b$! Although this is an incredibly high number, this revenue will only be generated by approximately 20 metric tons of biopharmaceuticals! Or in other words: one gram of biopharmaceutics sells around 10000 €. Already in 2016 our team; members from the University Natural Resources and Life Sciences Vienna, University of Lisbon and University of Newcastle forged the idea to develop a smart solution to reduce the price of biopharmaceuticals by adapting both, process development methods and control strategies. So why exactly are biopharmaceuticals so expensive? What made the Austrian Research Promotion Agency (FFG) invest 1.6 m€ into the project and how does the solution look like?
The implemented question why biopharmaceuticals are so expensive can be answered by the long development times (12-14 years), the financial effort to bring a new drug to the marked (1.8 b$ each) and the limited capacity to produce enough material, derived from the high number of batch rejections. Due to the long development times and financial effort to bring a new drug to the market, the final drug prices set huge pressure to the patients and health insurance. To reduce the cost per dose, several strategies were suggested by the Food and Drug Administration (FDA). One key element in these requests is the better understanding and control of bioprocesses.
“For each new drug entering the market 1.8 b$ are spent. Reducing this costs is of high interest for FDA, patients and health insurances”
And exactly here Novasign will step in.
So how does it evolve? Back in 2016, the former Novasign team was frustrated. After years of endless discussions about PAT (process analytical technology) QbD (quality by design) and Industry 4.0 without clear solutions, it was time to change the game. The team put a strategy together, speeding up bioprocess development to reduce development times for innovative medicine.
By combining the expertise in bioprocess development, machine learning, modeling as well as industrial software development, Novasign focuses to develop a software to reduce bioprocess development times while describing the behavior of the process.
Today in classical bioprocess development, the optimal working range for different unit operations can be found by using a design of experiments (DoE) approach. Here the impact of critical process parameters (CPPs) on critical quality attributes (CQAs) is evaluated. Such approaches are very valuable but they mostly only describe constant parameters over time, missing to address the dynamic of the process. But wouldn't it be great to describe several different CPP combination settings within a single experiment? For sure, but how can this be achieved?
We invented a disruptive method which is able to describe the dynamics of the system by changing CPPs in a defined way within a single experiment. Moritz named it intensified design of experiments (iDoE). To be able to describe the outcome of such a new experiment hybrid models are applied. These models combine the information hidden in the process data with the predictability of mechanistic equations. Does this mean the combination of iDoE and hybrid models can reduce bioprocess development times?
A certain design space can be screened 40 to 70% faster using iDoE over classical DoE approaches.
With our first experiments in a 20 L bioreactor, we will compare the cell behavior and the productivity using a classical DoE and an iDoE approach! If iDoE and the toolbox work as we expect them to do, the development times for a cell cultivation process will significantly be reduced. In our case, nine experiments for the iDoE compared to 27 experiments for the DoE.
Therefore, the screening of a certain design space in a cell cultivation process which would normally require six months of testing can be performed in only two months. The resulting reduction in materials, personal and energy costs are significant. Further, the models generated from the design screening can be used for process control because the dynamics of the system are finally deconvoluted.
Within 2019 the company will be founded and the software prototype will be finalized for commercialization. So stay tuned or directly join us on our journey.