CRISPR technology and a computer model can enhance the production and purity of drugs

Disease and treatment 11. jul 2020 4 min Associate Professor of Pediatrics and Bioengineering Nathan E. Lewis, Associate Professor of Pediatrics and Bioengineering Nathan E. Lewis Written by Kristian Sjøgren

A new model can show pharmaceutical manufacturers the genes they need to knock out in their drug-producing cell cultures to improve product purity and productivity. The model can also advance researchers’ knowledge of diseases such as cancer and nonalcoholic fatty liver disease.

Downstream processing to remove impurities accounts for up to 80% of the production costs of making certain types of biopharmaceuticals.

However, this process can be much improved now that researchers from the Technical University of Denmark and the University of California, San Diego and colleagues have developed a computer model that can identify the impurity-producing genes in their drug-producing cell cultures.

The researchers can then use the CRISPR gene-editing tool to knock out these genes before they cause trouble.

When the researchers knock out the right genes, the cells produce fewer impurities, and this can increase the quality and affordability of high-value biopharmaceuticals.

By understanding this system better, drug manufacturers can also increase their production of biopharmaceuticals.

“This does not mean testing gene by gene but requires systematically examining the overall picture of what the cells are doing and what can be deleted while still retaining viable cells. We can do this with our model, and it can also be used in research on diseases, such as advancing knowledge on cancer,” says a researcher behind the new studies, Nathan Lewis, Associate Professor, Systems Biology and Cell Engineering and Novo Nordisk Foundation Center for Biosustainability, University of California, San Diego, USA.

The research has been published in Nature Communications. Read the articles here and here.

Modelling the secretory pathway in a cell

Nathan Lewis and his colleagues have developed a computer model of the protein secretion pathway in a mammalian cell.

Cells in the body coexist with both neighbouring cells and those further afield, requiring all cells to constantly communicate with their surroundings.

One way they achieve this is by sending many hormones and other signalling molecules into their surroundings, and Nathan Lewis’ research has shown that a substantial amount of a cell’s resources are dedicated to producing membrane and secreted proteins, most of which are processed through the secretory pathway.

In addition, the cells use one third of their protein-coding genes on communicating and forming molecules that are unwanted by-products from a pharmaceutical perspective.

“So far, most research has focused on the individual signalling pathways and interactions between the individual substances, but in this study we created a model that provides a much better overview of the whole system of the secretory pathway and how it affects signalling,” explains Nathan Lewis.

Integrating thousands of metabolites and proteins

Nathan Lewis and his colleagues have created a model that keeps track of the diverse processes going on in the secretory pathway of a mammalian cell based on data from previous scientific studies and other data.

Overall, the model includes more than 250 proteins present in many mammals, including humans, mice and Chinese hamsters, which are used to make more than 2000 secreted proteins and signalling molecules and more than 5000 membrane proteins and receptors that decode the signalling molecules in the surroundings.

In addition, the model contains the several thousand metabolites that help fuel the cells and keep the secretory and signalling pathways running.

“The model shows how the whole system is connected, and this enables us to determine that, if a cell needs to produce a specific type of signal molecule, it must use some specific resources and in given quantities to keep producing them. We can therefore keep track of the cellular activity level,” says Nathan Lewis.

Researchers can advance understanding on cancer

The new model can be used for various purposes, all related to understanding how cells allocate resources.

In research on disease, the model enables researchers to study what happens in, for example, nonalcoholic fatty liver disease when the signalling pathways in the liver are altered to produce different secretory hormones, which ultimately affects metabolism in the entire body.

Another example is cancer, in which the cancer cells also change expression and begin to produce more or less of some secretory substances that influence the surroundings, enabling the cancer cells to proliferate and metastasize.

“The model enables researchers to investigate which signal pathways are involved and how they are supplied. You can determine the substances a cancer cell needs to make the signalling molecules or enzymes the cancer uses to affect the surroundings and then perhaps find a drug that can remove these substances from the cells so that they cannot produce the signalling molecules,” says Nathan Lewis.

Major potential for pharmaceutical companies

Nathan Lewis envisions the pharmaceutical industry being interested in the benefits of using the new computer model.

Mammalian cells are widely used in producing many drugs such as cancer drugs and vaccines.

Pharmaceutical companies want the cells to produce as much of the desired drug as possible with as few by-products as possible.

Nathan Lewis and his colleagues used the model to investigate which parts of the secretory and signalling pathways account for the largest production of by-products in a specific cell line that makes drugs in the form of antiviral monoclonal antibodies.

The researchers also examined how this cell line generally uses its resources.

The results from the model showed that a single gene was responsible for taking up resources in drug production, and that knocking out this one gene could theoretically increase the production of the pure product by almost 25%.

The researchers tested the model and knocked out this specific gene and achieved almost exactly what the model predicted, increasing production by 18%.

“We can also flip the question and ask the model what genes we need to knock out to enable the cell to use 50% of its resources on making biopharmaceutical products. In our experiments, we found that we needed to knock out 30 genes. It is rather fascinating to think that these cells make thousands of proteins, and we just have to knock out 30 genes to get 50% of their resources reallocated to make the desired drugs,” explains Nathan Lewis.

Eliminated 50% of impurities

In a subsequent experiment, the researchers tried to eliminate the impurities in biopharmaceutical manufacturing.

The researchers used the model and found that they could substantially affect cell performance and eliminate impurities in a cell line by knocking out 14 proteins.

Then they used CRISPR to knock out the necessary genes one by one.

The genes related to proteins that could be divided into three groups.

• Abundant secretory proteins that probably do not play an essential role in cell survival, since they are nevertheless released into the surroundings.

• Proteases that cut other proteins into fragments. The researchers chose to remove the proteases because they did not want molecular scissors mixed with their drug.

• Proteins that bind to the desired drug and other things. The researchers removed these, since they substantially increase cost of purifying drugs.

“When we knocked out 11 of the 14 genes, the cells actually grew better than the wild type, probably because they could focus more resources on growing. However, when we knocked out the last three genes, the production of the drug was reduced, probably because the proteins play some role in the production of our desired drug. Nevertheless, we showed that our approach was feasible and that there are ways to increase the production of a given drug and obtain a purer product from the start,” says Nathan Lewis.

Nathan Lewis explains that both researchers and pharmaceutical companies have expressed interest in the model, which is freely available for everyone to use.

Multiplex secretome engineering enhances recombinant protein production and purity” and “Genome-scale reconstructions of the mammalian secretory pathway predict metabolic costs and limitations of protein secretion” have been published in Nature Communications. Nathan Lewis is Associate Professor, Systems Biology and Cell Engineering and Novo Nordisk Foundation Center for Biosustainability, University of California, San Diego, USA.

Dr. Nathan E. Lewis is an Associate Professor of Pediatrics and Bioengineering at the University of California, San Diego, where his lab develops and...

Dr. Nathan E. Lewis is an Associate Professor of Pediatrics and Bioengineering at the University of California, San Diego, where his lab develops and...

© All rights reserved, Sciencenews 2020