September 15, 2015 (Vol. 35, No. 16)

Appropriate Post-Translational Modification Techniques Must Be Critically Employed

Therapeutic proteins are subject to deamidation, oxidation, glycosylation and other post-translational modifications (PTMs). These finishing touches are not merely ornamental, they can alter essential factors such as structure (including higher-order structure), binding, stability, and activity—factors that can influence a biotherapeutic drug’s clinical effectiveness and safety.

Accordingly, “PTMs become critical quality attributes,” advises Asish Chakraborty, Ph.D., business development manager at Waters. “Developers need to determine which modifications are present and their implications.”

Because of their sensitivity and selectivity, liquid chromatography–mass spectrometry (LC-MS) methods constitute the foundation of modern PTM analysis. LC-MS identifies PTMs, specifies their locations, and determines their relative abundances. This kind of information helps define product quality. Low-level occurrence at a critical location—say, the complementarity-determining region of a monoclonal antibody—may be more critical than higher preponderance at a less-critical site.

Data handling becomes the bottleneck in LC-MS identification and quantitation of PTMs. Making sense of PTM occurrence and trends involves comparing data from many batches and large numbers of samples over time. “Data mining, generation, interpretation, and communication are what slow the process down,” Dr. Chakraborty explains.

Waters has developed an analytical platform solution that provides the critical tools needed for biotherapeutic drug development. It is called the Biopharmaceutical Platform Solution with UNIFI® (BPS/UNIFI).

UNIFI software enables the deployment of high-resolution analytics, such as LC/MS, across a biopharmaceutical organization, with workflows for specific PTMs. For analyzing deamidation and oxidation, the package applies a peptide mapping workflow that generates precursor and fragment ions in a single run to identify, quantify, and compare the PTMs from different lots and batches. This workflow approach, which is called an LC/MSE, uses simultaneous acquisition of exact mass at high and low collision energy.

BPS/UNIFI integrates robust ultra-performance LC/MS characterization with comprehensive software suitable for applications in bioseparations, intact protein mass analysis, peptide mapping, and released glycan analysis. “UNIFI is the first commercial software combining LC and MS data in a single package that supports data acquisition, data processing, visualization, reporting, and compliance,” asserts Dr. Chakraborty.

SAFC’s recent innovations in media development have enabled clients to identify critical raw materials at early stages of cell culture development. Besides significantly expanding process knowledge, this next-generation media development approach has made robust manufacturing processes more predictable.

MS—The Way to Go

Since PTMs result in proteins that are chemically nearly identical with the original molecules, analytical methods must deliver maximum resolution and sensitivity. “Multidimensional chromatographic and MS/MS platforms enable researchers to break down complex molecules to look for specific PTMs,” says Scott Kuzdzal, Ph.D., life science business unit manager at Shimadzu Scientific Instruments. “Ion traps are especially powerful as they enable MSn detection by holding analytes in a trap and sequentially ejecting the fragmented ions.” (The n in MSn refers to the number of product ion stages in multistage MS experiments.)

For example, researchers employ ion traps to characterize glycan composition and sequencing information without labeling or glycosidase treatment.

Phosphoproteomics and glycomics continue to be the principal methods for analyzing PTMs. “Nearly half of all proteins are glycosylated,” Dr. Kuzdzal notes. Phosphoproteomics remains a significant challenge because of issues such as the low stoichiometry of phosphorylation, lowered ionization efficiencies of phosphopeptides, reduced digestion efficiencies, and phosphopeptide losses during sample chromatographic separations.

For glycan work with highly branched structures, MSn analysis is the way to go. An alternative is collision-induced dissociation, which provides valuable information such as simultaneous sequencing of the peptide and determination of glycan structures. Obtaining such information at lower collision energies is not possible.

Current PTM detection has limitations. Whereas modifications occur once per position, many different PTM types may occur simultaneously. “Biological systems are not restrained to single—or even a handful—of modifications,” Dr. Kuzdzal explains. “No single technology available can interrogate all potential modifications simultaneously, and we have a long way to go technically and computationally to even begin to look at the synergistic effects of multiple, tandem modifications.”

Multi-Attribute Assays

Much has been written on the significance of PTMs occurring in therapeutic proteins. According to St. John Skilton, Ph.D., global market manager at Sciex, PTMs remain topical for issues related to patient benefits, bioproduction, and meeting quality control requirements.

“Deeply understanding PTMs strengthens manufacturers’ ability to correlate them with safety and efficacy,” Dr. Skilton says. “Furthermore, monitoring PTMs as efficiently as other bioprocess parameters will boost manufacturing efficiency.”

Fully understanding PTMs helps biomanufacturers meet internal quality goals, monitor stability, and convince regulators of batch-to-batch comparability because the variations can be clearly understood.

These benefits are not lost on Sciex, whose MS instruments form the cornerstone of the company’s PTM analysis platforms. “The more information we can provide in terms of hard analytical, statistical information, with measures of abundance, the better off our customers will be,” Dr. Skilton remarks.

There is a massive trend toward multi-attribute PTM assays, which Sciex offers by combining traditional separations like capillary electrophoresis or liquid chromatography with mass spectrometry: CESI (a type of electrospray MS ionization with a capillary electrophoresis front end is suitable for peptide mapping, intact mass analysis, and glycan analysis) plus MS, MS/MS, and Sciex’s BioPharmaView™ software. The company also offers its own SelexION™ differential ion mobility instrument as an additional separation mode.“This set of capabilities enables us to construct multi-attribute assays from one analysis,” asserts Dr. Skilton. “From one peptide map, I can look at all the PTMs flagged as important, quantify them, and determine their variability across a batch. Having separations that are complementary and orthogonal makes this comprehensive.”

These analysis platforms enable scientists to probe deeply into PTM structure so that they can determine, for example, the degree of branching of specific sugar residues in glycans. This is one strength of capillary electrophoresis coupled with MS. “We can get to crazy levels of finding very low abundance glycosylations that people were not aware of,” Dr. Skilton explains. This will be a boon to developers of biosimilars who must demonstrate both physicochemical and therapeutic similarity.

Dr. Skilton likens the ability to exploit deep, broad glycan analysis to an arms race: “Innovator companies can use this information to create market entry barriers, but biosimilars firms can also use it to demonstrate similarity with analytical tools and measure the heterogeneity.”

Subtle Distinction

One challenge with characterizing a complex mixture, and in particular PTMs in a complex mixture, is reconciling an aggregate measurement and a higher-resolution measurement. “The question is whether existing analytical methodology for, say, total degree of deamidation, is representative of the molecule’s characteristics,” says Brian Collins, Ph.D., director of research at Momenta Pharmaceuticals. Aggregate numbers do not distinguish between a single site that is heavily modified and 10 sites that are changed to a much lower degree. “One of the first challenges,” Dr. Collins adds, “is to develop probes that allow us to move away from aggregate determinants.”

Biosimilars are the ideal example. Where they may be similar to the originator molecule through aggregate characterization, the molecules may be dramatically different in terms of distribution of specific PTMs.

Dr. Collins also warns about analytical methods that introduce PTMs during sample preparation. Elevated temperatures, exposure to ambient atmosphere, and proteolysis are in some cases sufficient to introduce modifications that were absent in the starting analyte. “We need to build in assurance that we are not introducing additional PTMs derived from the sometimes hostile environment of sample preparation,” he continues.

The third component of Momenta’s strategy is related to the last one: reliance on multiple, orthogonal analytical techniques. Every methodology has a unique bias, particularly for quantitative, high-resolution analyses. “Comparing measurements of the same PTM obtained through multiple techniques will further build confidence that your measurement is accurate, and you have not introduced additional modifications during analysis,” Dr. Collins advises.

In summary, Momenta’s approach involves breaking down proteins, characterizing their components, and assembling a comprehensive view. The challenge is balancing the cost and time of orthogonal, non-aggregate methods against conventional analysis. Aggregate methods are, after all, cheap and fast.

“The great challenge of characterizing biologics is that from the beginning they’re not a single molecular entity, like small molecule drugs,” Dr. Collins concludes. “You are by definition analyzing a complex mixture.”

Control Strategy

Over the last four years SAFC has put significant effort into understanding how cell-line development and culture media affect PTMs, particularly glycosylation and charge variants. The glycan epitope alpha-gal, for example, is expressed on glycoproteins of mammalian cell lines that are widely used in biomanufacturing. The gene for alpha-gal was muted millions of years ago in primates, which now carry alpha-gal antibodies responsible for immune responses.

“During early upstream development, we can attack immunogenic moieties with zincfinger nuclease technology,” says Steven Richardson, technical manager at SAFC.

Media and supplement development constitute the flip side of the company’s strategy to control undesirable PTMs. SAFC has systematically correlated critical raw materials of its cell-culture media to levels of PTMs associated with charge variance, glycosylation, and immunogenicity. “Removing those ingredients or altering their concentrations enables us to control PTMs downstream of cell-line development,” Richardson explains.

The success of this strategy depends on development stage. It is more appropriate, for example, for noncommercial or next-generation processes.

Biosimilars and PTM profiles are a recurring theme for which the media/supplement strategy is particularly useful. An SAFC customer may have developed a biosimilar process through pilot manufacturing scale and at acceptable titer, but the product may lack an acceptable PTM similarity profile. “If they need to match a target product profile, we can help them by characterizing and optimizing the raw materials,” Richardson explains.

SAFC is also working on a new class of supplements that can correct PTM differences. Like the media/ingredient strategy, this technique works best before the process is set in stone.

“Four years ago, we were still asking whether raw materials affect critical quality attributes, but we very quickly discovered they did,” states Richardson. “Where the effects of pH and osmolality on PTMs were well understood, we lacked an appreciation for leveraging media development to control quality attributes. Now we have the technology to fully understand and control quality-related PTMs. The relationship between media and critical quality attributes is no longer a black box.”

Trends in PTM Analysis

Scott Kuzdzal, Ph.D., life science business unit manager at Shimadzu Scientific Instruments, identifies the following trends in PTM analysis:

  • Labeled quantitative proteomics has been losing momentum in favor of label-free techniques.
  • Collision-induced dissociation is the dominant fragmentation method, but other fragmentation methods, such as higher-energy collisional dissociation or electron-transfer dissociation, are growing.
  • Bead-based isolation approaches continue to be developed, despite the long wash times and nonspecific adsorption.
  • Online digestion approaches, such as those developed by Shimadzu and by Perfinity Biosciences (including Perfinity Integrated Digestion Platform and Perfinity Workstation) have greatly improved protein digestion reproducibility and enabled powerful immuno-MS assays.
  • Researchers are moving toward more multiplexed strategies, but appear to be struggling and sacrificing data quality for higher throughput.
  • More targeted studies, such as exosomic studies, which concerns the nano- and micro-sized membrane-bound vesicles that are shed from live intact cells, offer refreshing, new perspectives on PTM analyses.
  • Some PTMs, such as phosphorylation, are unstable and therefore require novel sample preparation and analysis methods.

PTM Challenges and Best Methods

GEN:  What are the key challenges in testing for correct post-translational modifications (PTMs) after protein production?

Dr. Stokes: PTMs affect nearly all cellular processes in some way. Study of these critical modifications can be challenging due to the low abundance of some post-translationally modified proteins in a particular cell line or tissue. Even for more highly abundant proteins, the percent of that protein carrying a particular PTM may also be low, making identification and quantification of the modified pool of protein difficult without some type of enrichment. The specificity of the tools used for this enrichment is key, as some PTMs are highly related (such as symmetric and asymmetric dimethyl arginine).

Ms. Wake: The heterogeneity of PTMs certainly presents many challenges in terms of the resolution and, in some cases, the sensitivity of the analytical method. The potential change in the PTM profile and potential batch-to-batch distribution of PTMs equally give rise to a continual challenge for the analytical toolkit and methodologies.

Aside from this, one factor not related to the actual protein is the effect of formulation excipients, such as polysorbate, on the analytics themselves. Often these excipients can interfere with the analysis required to facilitate characterization, meaning the only option, in some cases, is complex sample workup, which can complicate the overall analytical program.

Dr. Somiari: In spite of technological advances, accurate and consistent detection of specific PTMs is still a challenge due in part to the transient nature of many PTMs and sample preparation steps that must be followed during the analysis. Since in many cases the PTM is almost always the less abundant population, compared to the non-PTM component, the use of a sensitive instrument and/or the introduction of a PTM enrichment step is required prior to mass spectrometric analysis. The limited availability of mass spec-compatible reagents that can stabilize specific types of PTMs makes accurate and reproducible testing for PTMs challenging.

Dr. Herren: A common issue is overcoming false negatives that result when operating below the detection limit of your chosen technique. PTMs are susceptible to detection issues because they are often: substoichiometric (that is, only a fraction of the total protein is modified), buried by protein structure (making them inaccessible), and only present under defined conditions (may need to be induced directly by chemical stimulation or indirectly through specific signaling pathways).

Furthermore, sample processing may result in alteration or loss of a PTM; thus, preservation of PTMs during processing is critical. Lastly, precise localization of a PTM to a particular amino acid site can be challenging, especially when multiple similar PTMs occur tightly clustered together.

GEN: What particular technique or techniques do you employ as a best method for detecting and analyzing PTMs of a protein? Why do you prefer this method?

Dr. Stokes: Cell Signaling Technology (CST) offers a wide variety of validated antibodies, which can be used in applications such as immunohistochemistry, flow cytometry, immunofluorescence, and western blotting to study PTMs. CST developed a proprietary technology to generate antibodies highly specific for a particular phosphorylation motif or PTM (ubiquitination, acetylation, methylation, etc.). These antibodies are used to enrich even the lowest abundance post-translationally modified peptides from cell or tissue samples, thus simplifying the mixture of peptides.

The enriched peptides are then identified and quantified using liquid chromatography-tandem mass spec (LC-MS/MS). This method, called PTMScan, allows researchers to quantitatively compare thousands sites of PTM in a single LC-MS/MS run. PTMScan is widely used for the study of PTMs in biomarker discovery efforts, such as the discovery of novel ALK fusions in ovarian cancer.

Ms. Wake: The choice of technique depends on the output required and in some ways the PTM or PTMs of relevance to the molecule studied. Mass spectrometry is in many ways the most sensitive and all-encompassing technique for defining the position of a PTM as well as establishing the heterogeneity of the modification.

For example, with glycosylation, you can use the qualitative and quantitative functions of mass spec not only to establish that the modification is on the same amino acid, but also to demonstrate that the added glycan unit is equivalent. This, however, may not be a practical technique for application in general quality-control testing when comparison of gross profiles between a test and well-characterized reference is often the underlying requirement. Capillary electrophoresis or hydrophilic interaction liquid chromatography approaches are best in these instances.

Dr. Somiari: We rely on mass spec for PTM analysis because it is fast and sensitive. But because of the nature of mass spec, all systems—including sample preparation, LC, MS/MS, and database search—must be optimized and in synergy. Sample prep is critical for consistent and reproducible results. The type of enzyme used in digestion, the extent of enzymatic digestion, and how close the modified amino acid of interest is to the site of cleavage are variables that could have profound effects on the accuracy and consistency of detecting PTMs.

To help ensure successful PTM testing, we attend to the following tasks: 1. Perform in silico digestion of the protein to determine the best enzyme to use. (The predicted number of cleavages and the proximity of the cleavage site or sites to target PTM are important considerations.) 2. Control the extent of digestion to achieve about 50% total digestion. 3. Preselect modified peptides when possible.

Dr. Herren: High-resolution LC-MS/MS in a bottom-up proteomics workflow. This strategy offers unsurpassed precision in the comprehensive detection of PTMs. The presence of many PTMs can be inferred from the specific masses they add to amino acids within a protein.

Because the mass spectrometer is detecting the masses of fragmented peptides associated with a protein and not the PTMs specifically, it allows simultaneous identification and localization of PTMs across any of the primary amino acid sequence of the protein(s) that fall within the instrument’s detection limit. To overcome detection limits, biochemical strategies have been devised to enrich PTMs in the sample processing steps before LC-MS/MS. LC-MS/MS is scalable, time efficient, cost effective, and suitable for analyzing protein samples of modest (few purified proteins) to high complexity (protein lysates). 

Previous articleNew Insights Into “The Mind’s Eye”
Next articleBroad Institute, MD Anderson Named NCI Genome Characterization Centers