When One Size Doesn’t Fit All

Biological systems have been used to produce therapeutic proteins since 1982, when Humulin (human insulin) was approved by the FDA as the first recombinant biopharmaceutical. Since then, biopharma has grown tremendously; in particular, monoclonal antibodies (mAbs), which continue to be the largest and fastest growing biopharmaceuticals– doubling in market-size over the past five years (1). However, peer into biopharma pipelines and you will find an abundance of more complicated, difficult-to-express molecules – next generation biologics.

The term “complex molecules” can be controversial in scientific circles. Colleagues will be quick to point out that anything beyond a simple protein, including a traditional mAb, is technically “complex.” I use the term colloquially to refer to molecules that go beyond standard antibodies. Examples would be bispecifics or other novel antibody formats or glycoproteins that combine domains from various molecules, often including engineered elements that you won’t find in nature. These are, naturally, difficult to express in existing expression platforms.

Another example of a problem manufacturers face is that these molecules will sometimes require additional genes when compared to the two required for a vanilla antibody. Here, simply getting all of the genes into a cell may be difficult, and pairing of gene products once in the system, if required, can be a real challenge. Equally, challenges exist during downstream and upstream development: purification can be tricky, especially if the protein is lacking an Fc region.

Faced with these difficulties, many in the industry have tried to shoehorn complex proteins into existing platforms, but this often results in a great deal of time and effort spent trying to find a suitable cell line and process. Trying to use existing systems can also mean there is a danger that important points are missed. For example, it can be vital to consider product quality, as well as product concentration, early in the process. We know these molecules can be difficult to express, so finding the best expressing cell line is important. But if product concentration alone is the main focus, a suitably expressing cell line could be selected, but it may not achieve the desired target product profile.

A toolbox of solutions

With monoclonal antibodies, technologies have been developed that can deal with the same molecule type again and again. Increasingly, however, expression platforms must deal with molecules of different “sizes and configurations,” so a one-size-fits-all approach doesn’t work as well.

Developing a toolbox of solutions to deal with the variety of complex proteins coming down the pipe will be essential. For example, at Lonza we have developed a multigene vector system that allows the transfection of multiple genes simultaneously on a single vector. This is far simpler and easier than two alternative approaches available today: namely co-transfection and what I would call a “mix-and-match” approach.

The mix-and-match approach involves constructing several cell lines where each makes a portion of the product. Fully purified material from each can then be mixed together and the product chemically recombined. This would be followed by a purification step to clean up and obtain the desired end product. Technically this can work, but the time and cost associated with making several cell lines can be significant.

Co-transfection, on the other hand, involves spreading your genes of interest out over multiple vectors and transfecting them all at the same time. This does avoid the need to make several cell lines, but there are several disadvantages. For example, you may need additional selection markers, which may be difficult to source. There can also be further analytical work required to identify the cell lines that have taken up the vectors and are expressing each of the genes at the required amount. Additionally, there is an increased risk of cell line instability when using multiple vectors as individual vectors can be lost. A multigene vector system, on the other hand, can be seen as the best of both worlds, since it removes the need to make several cell lines, without the additional work and risk associated with co-transfection.

Another important tool in the protein expression toolbox is Lonza’s GS piggyBac™ transposon-based technology, which is also well suited to more complex proteins. Transposons are mobile genetic elements and their mobility is mediated by transposase enzymes. The transposon, as part of a DNA vector, contains the genes you want to insert into your host cell, surrounded by inverted terminal repeat sequences. Once the vector is introduced into the host cell along with the enzyme, the transposase recognizes the inverted terminal repeat sequences at each end of the genes you want to move, at which point the transposase cleaves the DNA. The target gene is then pasted into specific sites within the genome associated with stable, high expression. These sites have a specific sequence (TTAA) and are found within regions of open chromatin. The combination of GS System® and piggyBac™ technologies therefore allows you to select cell lines where vectors have been inserted into highly transcriptionally active sites that are associated with stable, high expression.

This system can deal with the large gene cargos that are often associated with complex proteins (it has a cargo capacity of over 200 kb). GS piggyBac™ can preferentially target genetically stable parts of the genome at high efficiencies, which means that, it also has the potential to improve cell line stability.

The key benefit of GS piggyBac™ is that it works very well for low-expressing proteins, which complex proteins often are. For example, in 2016, a group at Eli Lilly tried combining piggyBac™ with GS CHO technology with four different antibodies, including a bispecific antibody (2). A two- to twelve-fold increase in product concentration was observed with piggyBac™ compared to control CHO pools – and in follow on work (2017) this group demonstrated that product quality was similar between piggyBac™ and control pools. They concluded in this follow on work that the higher product concentration could be explained by a combination of increased average gene copy number, significantly higher messenger RNA levels and the homogeneity of the piggyBac™ pools, relative to the control (3).

Lonza has carried out proof-of-concept studies with GS piggyBac™. Using a difficult-to-express antibody – one expressing less than one gram per liter (pretty poor for an antibody), we observed a >200 percent increase in product concentration with GS piggyBac™ compared to the control. Lonza’s commitment to ongoing technology advancements of our expression systems has resulted in us evaluating a number of different technologies alongside GS. Increases in product concentration of the scale observed with these studies are rare.

The ripple effect

The benefits of improving expression levels are clear when one thinks about cost and efficiency upstream, but they also have further benefits. For example, if you have low expression levels with a complex protein then you might have to extend your cell line construction to try and find a good expresser, or you may have to perform multiple rounds of process optimization to improve product concentration. This can be time consuming and can extend development timelines. This is a problem for the industry as a whole, but especially for small biotechs who are under intense pressure to get ahead of the pack.

It’s far too early to write off mAbs – they will continue to play a prominent role in biologic medicine for the foreseeable future. But as developers look to move beyond the low-hanging fruit by tinkering with nature to create new, responsive drug systems, complexity and variety will eventually become the norm. But these new therapies will never fulfill their therapeutic potential unless manufacturers are able to achieve appropriate expression levels whilst meeting the desired target product profile. And for that, a toolbox of solutions, tailored to each molecule will be key.