Rethinking MSX Selection: Optimizing Glutamine Synthetase Expression for Enhanced Bioproduction

By Yusuf Johari, Ph.D.
Senior Principal Scientist, R&D

As therapeutic protein formats continue to diversify, the industry is re-evaluating not only how we design expression vectors, but also how we select and stabilize producing cells. While methionine sulfoximine (MSX) has historically played a central role in glutamine synthetase-based CHO expression system, our data show that optimizing glutamine synthetase selection stringency, in the absence of MSX, can offer distinct advantages.

Having worked extensively in expression system optimization and cell line development, including on the GS Gene Expression System® technology platform, I’ve seen how subtle differences in selection strategy can profoundly influence bioproduction outcomes. This is becoming increasingly important as molecules grow more complex, from monoclonal antibodies to bispecifics and fusion proteins. In this context, robust, predictable, and stable expression systems are no longer optional – they are essential.

Moving toward MSX-free systems

MSX is commonly used in glutamine synthetase-based expression systems to select high-producing cells (Figure 1). By inhibiting glutamine synthetase, it creates strong selection pressure, ensuring that that only cells with strong expression of the glutamine synthetase gene (along with the co-expressed product genes) survive in glutamine-deficient conditions.

Figure 1: Glutamine is required by CHO cells as an energy and nitrogen source. MSX increases selection stringency by inhibiting glutamine synthetase, thereby creating host CHO cells that require high glutamine synthetase gene expression for survival in media lacking glutamine supplementation.

While effective, this approach introduces an artificial layer of selection pressure that can sometimes lead to unintended consequences. For instance, gene expression may become less stable over time, particularly once selection pressure is reduced or removed. These effects are not universal, but they tend to become more pronounced as expression systems are pushed to accommodate increasingly complex protein formats.

MSX-free systems take a different approach. Instead of relying on chemical pressure, they depend on optimized vector design to achieve the desired level of glutamine synthetase expression. This strategy works best when combined with advanced genetic architectures, i.e. carefully tuned promoters and well-designed multi-chain constructs, to ensure that each component is expressed at the appropriate level.

A stronger foundation for the GS® expression system

As protein modalities evolve, so too must the systems used to produce them. MSX-free expression systems represent a shift toward more biologically aligned strategies – ones that emphasize stability, balance, and predictability over artificial selection.

At the core of this approach is the precise modulation of glutamine synthetase expression at the transcriptional level. Achieving this is not straightforward. The marker gene must be expressed at a level low enough to maintain sufficient selection stringency, yet high enough to preserve key performance attributes such as cell growth.

It may be tempting to simply weaken a strong promoter, for example, by removing an enhancer, but this is rarely optimal. Marker gene optimization is better understood as a “Goldilocks” challenge: glutamine synthetase expression must be finely tuned to be neither too high nor too low, but just right. This places greater emphasis on the intrinsic design of the glutamine synthetase selection cassette, including upstream and downstream promoters/enhancers, gene arrangement, and other regulatory elements.

Enhanced pool stability

One of the most compelling observations from our work is the consistency of expression stability in MSX-free bulk pools, regardless of molecular complexity. Whether producing a standard IgG or a more intricate bispecific antibody, optimized glutamine synthetase selection, without MSX, maintains stable product titers over extended culture periods (Figure 2).

Figure 2: Combining synthetic LHP-1 promoter with a novel glutamine synthetase selection cassette, enhances bulk pool stability of monoclonal (mAbs) and bispecific antibodies (bsAbs) over 30 generations in MSX-free culture. Each bar represents a bulk pool.

By contrast, MSX-selected systems can exhibit greater heterogeneity at the pool level, driven by elevated glutamine synthetase expression and external selection pressure. This instability often becomes more apparent once MSX is withdrawn. In practical terms, MSX-free expression systems can deliver more consistent yields and product quality, particularly for toxicology supply and early development material.

Simplified workflows

MSX-free approaches also enable more streamlined workflows in bioproduction system development. With optimized selection stringency, bulk pools can achieve titer performance comparable to, or better than, traditionally enriched pools. This simplification accelerates early development by enabling faster generation of stable pools and more predictable progression into clonal selection, while also reducing resource requirements (Figure 3).

Figure 3: Our optimal MSX-free process enables rapid generation of process development and tox material for fast-moving programs.

An additional advantage is the removal of MSX itself. As a potentially toxic compound, its use is not generally encouraged in commercial manufacturing. Manufacturers treat MSX as a process-related impurity that must be tightly controlled or eliminated during processing. Eliminating MSX therefore removes the risk of residual contamination in the final drug product and helps address potential regulatory concerns.

Looking ahead

The move toward MSX-free expression systems reflects a broader trend in bioproduction system development: leveraging deeper biological insight to build more robust and scalable platforms. By combining optimized vector design with MSX-free selection, it is possible to create expression systems that are not only high-performing, but also inherently stable across a wide range of therapeutic modalities.

As datasets expand, further insights will help refine best practices and guide the development of next-generation platforms capable of supporting increasingly complex biologics. Ultimately, removing dependence on artificial selection pressures may be key to achieving more consistent, scalable manufacturing outcomes in an era of growing molecular complexity.