Send to printer »

Feature Articles : Feb 1, 2013 (Vol. 33, No. 3)

Challenging the Limits of Protein Characterization

  • Josh P. Roberts

The ability to better characterize the peptides and proteins that flow through the pharmaceutical industry’s pipelines engenders a call for better characterization at the same time regulatory demands spur innovations.

Older technologies are being updated and upgraded and new ones are being developed, challenging the limits as to how small and what types of particles and degradation products, and in what kind of formulation, can be detected and differentiated one from another.

Scientists gathered at CHI’s “PepTalk” conference last month to share their insights on the tools to use for protein characterization, and reasons for choosing a particular formulation over another.

Aggregates and subvisible particles are detrimental to any protein or peptide drug formulation, and pharmaceutical companies are on the lookout for them using a variety of screening tools at various points in the process from discovery to market.

Among those especially suited to the late discovery/early formulation stage is dynamic light scattering (DLS). DLS has been around since the 1950s but was long considered a “black box technology, where the analysis was complex, and it was difficult to make the measurement,” said Kevin Mattison, Ph.D., principle scientist, bioanalytics, at Malvern Instruments.

The technology itself has become much more user friendly, and vendors have mastered data interpretation so that it’s pretty straight-forward. No sample preparation or chromatography is necessary, and it will work under very dilute or turbid conditions (the latest generation systems can handle highly concentrated samples as well). “You put your sample in, hit the go button, and about a minute later you get the size distribution across a range of about 1 nm—1 µm,” he said. “It will tell you whether or not you’ve got aggregates in there.”

While it’s easy to make the measurements, there can still be complexities of data interpretation. A lot of users get in trouble because they tend to overinterpret the data because they don’t fully understand it. That and prompting from industry is what led Dr. Mattison to co-develop and teach a short course on how to obtain reliable information from DLS by knowing what properties of the sample will influence the signal.

DLS measures the change in light scattered over a short increment of time due to Brownian motion of particles suspended in a liquid. While this is partly dependent on concentration, where DLS derives its real power is that the amount of light scattered is extraordinarily sensitive to the size of the particles doing the scattering, increasing by r6—that is, by the radius to the 6th power. A 100 nm aggregate “will dominate your DLS signal and it will stand out like a sore thumb,” Dr. Mattison noted. On the other hand, other optical techniques may not be able to pick it up (especially at low concentration), and the particles may be too big for size-exclusion chromatography.

DLS measurements can help predict shelf life and viscosity at higher concentrations, he added.

One of the common challenges Dr. Mattison faces is that people want to use DLS to count particles. He cautioned: “It can tell you how big it is, and can tell you whether or not it’s in there. But it cannot tell you to any degree of acceptable accuracy how much of it is in there.”

What Is It?

Other technologies such as light obscuration can count particles, but neither it nor DLS can identify what the particle is.

With newer techniques like resonance mass measurement and flow imaging microscopy, “you can see if it’s a protein particle, or if it’s a silicone oil droplet,” said Andrea Hawe, Ph.D., CSO for research at Coriolis Pharma, and a co-founder of the company.

The Archimedes resonance mass instrument from Affinity Biosensors will query the mass and density of a particle from about 500 nm to about 1–2 µm, while the MFI microflow imaging system from ProteinSimple can analyze properties such as particle size, shape, and transparency from 2 µm to larger micron sizes, Dr. Hawe pointed out. Software algorithms can then determine the makeup of the particles—whether protein or silicone oil, for example. Combining the two techniques allows you to characterize a wide size range of subvisible particles.

This is highly relevant for therapeutic proteins in prefilled syringes because they are siliconized on the stopper and glass barrel, and “as soon as the formulation gets into the cartridge the silicone oil droplets start to form,” she said.

In protein and peptide formulations—especially when highly concentrated—the refractive index of particles is often so similar to that of the formulation itself that there is insufficient contrast to be adequately picked up by light obscuration (which is often required by regulators for batch release).

“What you can do is switch to (or additionally do) microflow imaging where you have better contrast, or you can switch to a different measurement principle like resonance mass measurement,” Dr. Hawe noted. “You can also think about diluting your sample in a different formulation, but this is not recommended because when you change the formulation, this might also change your particles.”

Silicone oil droplets are only one example of extractables and leachables that may be picked up from process equipment, disposable systems, packaging, and delivery materials, including syringes, cartridges, tubing, filters, and gaskets. We have to better understand the interaction with the drug product, “and be careful because these have a strong impact on the product quality and on the safety profile,” said Jöel Richard, Ph.D., vp, peptides, at Ipsen.

Metallic compounds and organic compounds—such as unreacted oligomers and curing agent residues—can interact with protein molecules in the formulation and chemically degrade them, cause them to aggregate, or form composite impurities with them.

Stainless steel, cellulose, glass, and silica have all been implicated in aggregation of protein therapeutics and formation of particulates. Tungsten microparticles, formed during the boring of needle holes, have been seen to produce visible particles with proteins, and soluble tungsten polyanions in slightly acidic buffer can precipitate monoclonal antibodies within seconds.

Dr. Richard recommends first performing a careful evaluation of the manufacturing steps that can generate aggregates and particulates, including an assessment of the packaging system, and gathering and analyzing whatever data can be obtained from suppliers. Then use a quantitative risk-based approach. Derive the daily dose of leachables that would be administered and compare them to toxicity guidelines. Steps with higher risk should be monitored using independent analytical methods.

“The key is to determine what is necessary, because you can generate a lot of data,” he said. “Depending on the type of formulation you have, if you have companion packaging materials, are extraction studies necessary or not?”

Any change of formulation, container closure components, and process steps need to be analyzed and monitored as well. To drive the point home, Dr. Richard cited a famous example in which a company commercializing epoetin alfa in Europe simultaneously changed both the primary container and the formulation composition, and the new formulation extracted some low molecular weight organic compounds—residues of the vulcanizing process used to create the rubber syringe stoppers.

These acted like adjuvants, practically turning the formulation into “a vaccine” that elicited an immune response against administered and endogenous erythropoietin, leading to a depletion of the patient’s red blood cells (i.e., pure red-cell aplasia).

Guidance on exactly what constitutes a safe formulation can be somewhat lacking, Dr. Richard pointed out. “There are a lot of guidelines, a lot of regulations. They tell you basically what to do, but they don’t tell you how to do it, nor when to do it, nor propose clear acceptance criteria.”

Where Is It?

Even without the help of pump parts and rubber curing agents, proteins and peptides can degrade over time. It’s important to know how and where such degradation takes place, so as to try to minimize the impact on the formulations. Beyond the chance of adverse immunological reactions, the biological activity may be compromised as well, for example, as a deaminated or cross-linked therapeutic loses its ability to engage with a receptor.

Peter Kresten Nielsen, scientific director, protein chemistry at Novo Nordisk, examines the chemical stability of prototype formulations by incubating them at high temperatures and analyzing the degradation sites using a top-down mass spectrometry approach.

“We have to provide feedback to the chemists who then synthesize the optimized peptide based on our MS input. We don’t have time to purify the degradation products, and collect fractions, and digest the peptides by protease, and so on, so we can save time and give fast input to further peptide design,” he said.

Among the challenges he faces are detecting sites of isomerization of individual amino acids—for example, the conversion of aspartic acid (Asp) to isoAsp. “It’s really difficult to detect because there is no change in mass, so you can’t detect it by just normal MS fragmentation methods,” he lamented.

Because the isomerization reaction changes the amino acid’s backbone, using specialized methods such as electron-transfer dissociation (ETD) to fragment the peptide will yield reporter ions that are unique to isoAsp. The problem is that the peak intensities for these ions are significantly lower than those for the common fragmentation ions, and can be indistinguishable from spurious noise.

This is most often solved by first digesting with a protease that will cleave Asp peptides but not isoAsp—not an option for a top-down approach. “We’re working a lot on how to optimize this.”

Similarly, Nielsen is faced with racemized peptides (L- and D- forms), which also have the same mass, but do not yield unique reporter ions. He has been using reference peptides that contain these modifications, and experimenting with MS combined with a combination of liquid chromatography and ion mobility spectrometry. “It looks promising,” he said, “but still it requires some optimization.”

Waiting to Inhale

Once the drug is identified, it still has to be administered somehow, whether it’s sub-cutaneously, intramuscularly, intravenously, or perhaps orally or as a suppository. And there should be a reason behind the choice of delivery route. “Don’t do it just because you can—you really need to look at what is the unmet medical need, and then develop a target product profile,” said David Cipolla, senior director, pharmaceutical sciences, at Aradigm, which develops products for inhalation including macromolecular drugs.

“You need to make sure that the rationale for the product makes sense after considering all the stakeholders, to look at the competition in the marketplace that you’re competing with, and to keep your product profile updated as development progresses,” he added.

Cipolla discussed three principal scenarios for (re)positioning a drug for inhaled delivery. The first and most obvious is to treat lung-related disorders like asthma or lung cancer. The second is to find a lung-related indication for an existing drug, like using cyclosporine to prevent transplantation rejection. And the third is to use the inhalation route to treat a systemic disorder, to alter the pharmacokinetics (for pain relief, for example) or pharmacodynamics, or to avoid injection-related issues.

There are a variety of inhalation delivery systems to choose from, each with its own advantages and disadvantages that need to be weighed. Dry powder inhalers and metered dose inhalers are small, and quick and easy to use, for example. A nebulizer is bulky, inefficient, and requires long dosing times, but it’s a well-established technology that can be used off-the-shelf. Other, more advanced systems can assure more efficient and accurate delivery, with comparable or better performance for the latter versus injections.

Since biopharmaceuticals are manufactured in aqueous-based fermentation or cell culture systems anyway, Cipolla recommends going into initial safety and small proof of concept human studies using a standard jet nebulizer with a simple aqueous formulation.

“And then in parallel you can develop, if needed, another formulation that will give you the commercial specifications, such as the dose and long-term stability. And once you know that it works, that’s when you spend the time and money on whatever you want as a final delivery configuration.”