Unspent Media

Current biopharmaceutical process development requires the use of living cell lines with highly specific nutritional and environmental needs, which poses a number of complex challenges – least of all finding the optimal cell culture media formulation. But getting the media formulation right is crucial; the nutritional composition has a direct effect on cell growth – as well as the yield and quality of the biotherapeutic molecules they produce. By fully optimizing media, biopharmaceutical developers can dramatically improve productivity and cost-efficiency. 

As with all scientific processes, the key to successful optimization is understanding data – in this case, it’s about understanding how each medium component influences the cells. Traditionally, these data have been collected using a technique known as “spent media analysis.” This iterative, empirical approach compares levels of different components in media samples before, during, and after cell growth to provide insight into component utilization over time, which feeds into optimization of media.

However, despite its longstanding use, the level of detail that can be obtained using spent media analysis is fundamentally restricted. And that’s true for both understanding the components themselves – as the technique only permits analysis of major metabolites such as vitamins and amino acids – and how they are being used. The latter issue arises because spent media analysis can identify only a limited number of molecules that are taken up or secreted by the cells, rather than identifying global molecular changes, such as signaling, and which metabolic pathways the components are involved in.

The solution? Multi-omics analysis. Specifically, in the case of media optimization, the application of proteomics and metabolomics, which refers to the molecular characterization of proteins and metabolites, respectively. Much like spent media analysis, these techniques rely on an iterative approach to identify how media components are being used by cells, and then use this insight to optimize the media. Unlike spent media analysis, the level of detail that these two techniques can obtain is unparalleled.

By enabling precise identification and quantification of the proteins expressed by the cells, proteomics enables identification of the intracellular pathways that are being activated or inactivated. This information can then be layered upon metabolomics data to establish how individual metabolites are flowing through these pathways. As a result, potential pathway bottlenecks, which could be impacting cell growth or product quality and yield, can be discovered. The combined knowledge can then inform the design of additional experiments to further optimize the media formulation.

For example, consider a process where the amino acid serine is rapidly depleted despite relatively low consumption by the synthesized therapeutic protein. In this scenario, a hypothesis for where the serine is used could be developed and tested using spent media analysis, but this would be a time-consuming process. By using proteomics and metabolomics instead, the actual intracellular pathways can be followed and the specific component that the cells are synthesizing using serine can be identified. Knowing this, the developer can then undertake further investigations to determine whether to add the missing component to their media, rather than more (potentially unnecessary) serine.

This example illustrates the considerable impact of using multi-omics rather than spent media analysis in the design of experiments undertaken during media optimization. In particular, it highlights how the extra level of granular intracellular detail that multi-omics provides can enable developers to either gain actionable results through less experimental iterations or – with an equal or higher number of iterations – gain increasingly more information.

Notably, I’d be the first to agree that spent media analysis has been an invaluable tool over the years, playing a pivotal role in the development of numerous life-saving biotherapeutics. However, as interest in biologics continues to grow, it is clear that we need to render their development and manufacture even more efficient and cost-effective. In my view, the only way to accelerate the development of next-generation biopharmaceuticals is to leverage next-generation analytical solutions.

To make full use of advanced process development analytics, there is an onus upon the entire industry to think big in terms of potential applications. For example, the combined use of proteomics and metabolomics is not restricted to new media optimization projects; it could also be applied to existing processes to enable efficient and reliable media troubleshooting in the event of unexplained product and process variations. 

Beyond media optimization, the use of further omics analyses – such as genomics and transcriptomics during cell line development – holds even more promise. By applying these techniques collaboratively at different workflow stages, biopharmaceutical developers could not only benefit from a significantly improved process, but also from an expedited development timeline and, in turn, an accelerated speed to market.

Another area where collaboration (albeit between more diverse scientific disciplines) has the power to further advance process development is the management and use of the data collected during these analyses. By working with computer scientists to implement AI processes using machine learning, we can create models based on data collected from thousands of experiments. And as more and more data are collected, these metabolic pathway models will become a vital part of the multi-omics toolbox by allowing process developers to escape the traditional limitations of so-called local “tribal” knowledge. Instead, they will have direct access to detailed company-wide – or even industry-wide – global knowledge, which can be used to support new optimization processes. 

I’ll admit I have good reason to be biased – but I truly believe multi-omics analysis should be considered an essential part of a modern cell culture media optimization process. And if we spread our wings further to consider the full spectrum of its applications across the entire development process – not to mention how it could be enhanced by cutting-edge data science – the introduction of multi-omics analysis could even contribute to a tipping point in overall biopharmaceutical development.

This article was originally published in The Medicine Maker