Analytics Playing Increasing Role in Formulation Development

Formulation and analytical sciences are separate disciplines by name, yet according to Sachin Dubey, senior director, Analytical, Formulation and Drug Product Development at KBI Biopharma, in practice these two areas are inseparable parts of the same process.

The connection is particularly evident, Dubey explains, as biologics and complex modalities come to dominate pipelines, gaining even more significance as the preservation of protein integrity becomes a priority.

In a landscape defined by rapid change, including the ongoing evolution of automation, digitalization, and data processing, Dubey says that the ability to align formulation strategies with analytical rigor is emerging as a defining feature of successful drug development, helping to advance molecules from discovery through commercialization.

A synergistic relationship

“For me, formulation and analytics go hand in hand,” Dubey says, adding that there are a few particularities about formulation that must be noted.

“For example, unlike process development, formulation development is highly analytical-intensive,” Dubey adds. “The studies in formulation development are small, which means that you need much more analytical power there.” By small, he means that except for long-term stability studies, most formulation development studies are designed to be conducted over a period of 10 days to a month—just mere weeks long.

“That means that you will have to analyze a lot of samples before intervention, after intervention, in those short periods,” Dubey says, adding that because of this, there is a necessity for the studies to be fit for purpose.

Also, according to Dubey, because early-stage formulation may look different than late-stage, accompanying analytics should account for those differences. As formulation is selected for later stages, much more robust and accurate analysis is needed.

“There can be different types of analytical means,” Dubey explains. “At one extreme, we need high-concentration formulations. And for high-concentration formulations, analytics like viscosity and injectability become extremely important. At the other side of the extreme, if you look into developing low-concentration formulations, then challenges like adsorption become much more important.”

Simply put, analytics as a discipline must cater to a wide variety of requirements from particular formulations. Particularity is also crucial for analysis of finished products in a way that it may not be for drug substances, according to Dubey—for example, when it comes to subvisible and/or visible particulate analysis.

Keys to protein analysis

Dubey addresses how analytical capabilities can specifically assist the preservation of protein integrity in the formulation process. “Protein integrity is absolutely critical to be preserved,” he says. “From an analytical perspective, there are several options, and they all need to come together and play at the same time.”

Dubey explains that routine analytical processes need to be established throughout each of the four levels of protein structures; in other words, confirmation or characterization of primary and secondary structures is often necessary for proper analysis of tertiary and quaternary structures.

“‘Triple-A analysis’—amino acid analysis—and mass spectrometry analysis are often performed to confirm the primary structure,” Dubey notes. “Secondary analysis is often done using circular dichroism or differential scanning calorimetry.”

While this category of analyses does not need to be carried out for every batch or stability point, tertiary and quaternary structures need to be studied and followed up on regularly. “Those can be achieved by more routine analysis,” Dubey explains. “Tertiary analysis can be done by routine size-exclusion chromatography or capillary electrophoresis. Quaternary structure can be followed by functionality.” In theory, he adds, this approach also confirms that primary and secondary structures should be intact at the time of analysis. “These analyses are equally important for monitoring and guiding processing as well,” he adds.

Stages of scalability

In terms of taking a product from preclinical to commercial use, analytics has a hand in scalability. Similar to the progression of stages of formulation, Dubey notes that as a product is developed from early to late stage, needs and purposes are different, so the utilized analytics will be as well.

Dubey points to International Council for Harmonisation (ICH) guideline Q14, Analytical Procedure Development, as a guidepost (1). “For very early-stage discovery or cell line development-type work, we focus our analytics more on screening power,” he says, with goals that include discovering which among 10 hypothetical candidates is the best performing or most stable.

As a molecule moves to the clinical phase, starting with a platform approach is typical. “At this point, the molecule is still a molecule,” Dubey states. “It has not become a drug; it’s still exploratory, and the analytics is also within an appropriate fit for purpose efforts.”

After optimizing the molecule or molecules and performing phase-appropriate qualification, analytics must be at their most robust for the late and commercial stages. “This is now further strengthened from a regulatory perspective as well,” Dubey adds. “A few years back, ICH Q14 became effective, which covers the lifecycle management of analytics—how the quality target product profile needs to drive the analytical target profile, meaning, ‘What is the purpose of that analysis?’”

To put it another way, Dubey asks,“How can a risk-based approach be implemented, what is expected of that, and what are the boundaries of those analytics?” The answer can be complex, with sometimes 20-plus analytical attributes to be followed, or as simple as following just four or five main parameters, which would include such factors as the pH of the mobile phase or the number of injections onto the column.

“If we start building those files, it helps build a wealth of information,” Dubey explains, “and it helps build a database based on which we can realistically make data-based decisions and accordingly adopt the later-stage development of analytics.”

Automation evolution

Dubey says that as we reach the end of 2025, new and emerging digital tools, along with automation, are helping advance formulation strategies.

“For quite a significant time, simpler automation has existed in the formulation and development area,” Dubey says, using buffer exchange as an example that has been deployed for 10 to 15 years. “But now, as we are advancing, automation is becoming more and more complex and has more and more advantages.”

According to Dubey, things like pH measurements, physical appearance, color, clarity, and particulate analysis can now all be concentrated into one system, which can also compile the data gathered into a cohesive package, making the lives of scientists and research associates easier. “That’s a move coming to digitalization, which has more to do with the data or the information flow,” Dubey adds. “This can be how the data [are] communicated, for example, how to establish a workflow, and that’s becoming easier with more connected platforms.”

On the other hand, Dubey says, data processing is also becoming more and more powerful and advanced. For instance, he says, a fairly new evolution in microfluid imaging has occurred. “There’s a high-speed camera taking several hundreds of pictures of your sample as it flows through, and those several hundred pictures are then analyzed,” Dubey explains. “And until about five or eight years ago, the capabilities of these softwares were not strong or efficient enough to have very good differentiation of different particles. But now with higher powers and better datasets, as well as more information available, the systems and the data processing are capable of differentiating between types of particles in a very, very robust manner."

Reliance on the operator is still needed to a certain extent for such simple analyses as comparing color or clarity to a reference sample, but Dubey argues that some level of subjectivity is still present. “I feel there is a scope here to have a more automated, more advanced system in which the subjectivity can at least be reduced,” he says.

Tips for troubleshooting

As with any process, formulation development has its share of pitfalls and trouble spots.

When troubleshooting, there are hard and soft aspects to consider in the relationship between analytics and different formulation strategies, Dubey says. “The soft aspect could be as simple as the communication,” he continues. “There has to be fluid communication in the sense that they are interdependent; the formulation can impact analytics in terms of whether the formulation has a different type of buffer, and the analytics can be influenced by the matrix effect, or the other way around.”

Be it for formulation development, process development, or something else, Dubey says, an analytical team with the right mindset about problem solving is very important. “They always bring in the sample with a question; they are looking for an answer from that data,” he says.

In talking about the harder aspects, Dubey says the aforementioned matrix effect is something that must be taken into consideration while developing analytics from a formulation perspective—but also for different purifications. “There has to be sufficient reproducibility of analytics, because you’re selecting a composition that is ultimately going to be administered into the patient, and we do not want to have additional variability coming out of the analytics,” he says.

Reversible aggregation is not a big threat from Dubey’s perspective, because as the term suggests, it is reversible in nature. Still, a measurement made at the right concentration is of utmost importance, as is the understanding necessary to ensure the accuracy of that step.

Analytical instrumentation limitation is another consideration. “You need to respect that kind of correction or limitation or use other orthogonal techniques,” Dubey says.

Finally, analytical teams should understand that any changes within an analytical setup will impact other teams, especially formulation teams that can also be responsible for stability studies, explains Dubey. A change to the process could make a big difference to a study that may need to run for four or five years.

Ahead of the curve

Dubey says his passion for formulation development has caused him to keep a watchful eye on new modalities that are coming into prominence, each of which requires a new and different approach.

“Antibody-drug conjugates (ADCs) are getting a lot of traction, and ADCs have, on one side, a chemical moiety that is apolar in nature, and on the other side, a protein,” Dubey says. “So, having these molecules, from a formulating perspective as well as from a stability perspective, becomes important.”

Dubey adds that the era of bi-specifics is quickly becoming one of tri-specifics and other very complex, more influencing types of molecules, designed in ways that affect the biology of first-in-human type doses. “Just by virtue of their foreignness into our body, they often have a very low dose to start with, and those kinds of challenges need to be looked at very, very carefully,” he says.

Peptides are also coming back to center stage, and as they are different from small molecules, they must be dealt with in a specific manner. Overall, formulation scientists should continue to be vigilant, Dubey suggests. “Listening, from an industry perspective, and preparing and staying ahead of the curve is really the key,” he concludes.

Reference

1. ICH, Q14 Analytical Procedure Development, Step 5 version (2024).