A monoclonal antibody formulated at 150 mg/mL should, in principle, be simple to deliver. The protein is well-characterized, the excipients are standard, and the container closure system is proven. When the formulation team runs injectability testing, however, the force required to push through a 27-gauge needle exceeds what patients can comfortably self-administer. The viscosity at high shear is too high. Reformulating with a viscosity-reducing excipient helps the flow problem but introduces aggregation over storage. The project timeline slips by months.

Scenarios like this play out regularly in biopharmaceutical development, reflecting a broader shift in how drugs are made and delivered. The pharmaceutical industry is moving away from simple small molecules and toward complex biological systems: monoclonal antibodies, bispecific antibodies, antibody-drug conjugates, mRNA therapeutics, viral vectors, lipid nanoparticles, and cell therapies. These modalities are not just chemically different from traditional drugs, they are physically different. They are colloidal systems, suspensions, emulsions, and nanoparticle dispersions whose behavior depends on physicochemical properties that were once secondary concerns.

The way drugs reach patients is changing as well. Biologics that once required IV infusion in clinical settings are being reformulated for subcutaneous self-injection at home. Higher protein concentrations in smaller volumes concentrate every formulation challenge. Proteins interact with each other, with interfaces, and with the shear forces encountered during injection. What worked at 10 mg/mL may fail completely at 150 mg/mL.

The biologics market now exceeds $400 billion globally, with subcutaneous delivery growing faster than IV administration. Success of mRNA vaccines during the COVID-19 pandemic accelerated interest in lipid nanoparticle delivery systems for a range of therapeutic applications. GLP-1 receptor agonists have become one of the largest drug classes in history, driving demand for advanced injectable delivery systems. Across all these areas, product success increasingly depends on controlling a small set of physico-chemical properties: colloidal stability, viscosity as a function of shear rate, surface tension at the air-water interface, and foam behavior.

Trends Reshaping Drug Delivery

Several developments are converging to make formulation science central to biopharma innovation.

The push toward subcutaneous self-administration is perhaps the most significant. Hospital-based IV infusion is expensive, inconvenient, and limits patient access. Payers and patients both prefer treatments that can be administered at home, and this preference is reshaping how biologics are developed. A subcutaneous injection typically requires delivering a therapeutic dose in 1-2 mL of fluid, which means antibody concentrations often need to reach 100-200 mg/mL. At these concentrations, protein-protein interactions dominate the rheological behavior, and viscosity can increase exponentially with small changes in concentration. The formulation challenge shifts from ensuring stability at low concentration to managing a concentrated colloidal system under mechanical stress.

GLP-1 receptor agonists illustrate how quickly the landscape can change. Clinical success of drugs for diabetes and obesity has created enormous demand for injectable peptide delivery systems, including multi-dose pens, autoinjectors, and long-acting depot formulations. These products must meet demanding specifications for injection force, dose accuracy, and stability across temperature excursions. Rheological behavior directly affects whether a device can deliver a consistent dose, while surface tension affects whether air bubbles form during cartridge filling or injection.

Cell and gene therapies represent an even more complex frontier. CAR-T cells, viral vectors for gene therapy, and mRNA therapeutics delivered via lipid nanoparticles are all highly sensitive colloidal systems. A viral vector suspension can lose infectivity if exposed to excessive shear during pumping or filtration. Lipid nanoparticles can aggregate or fuse if ionic strength, temperature, or interfacial conditions change even slightly. These systems are manufactured in small batches at enormous cost, and a formulation failure late in development can set a program back significantly.

Long-acting injectable depots add another dimension. Sustained-release formulations based on biodegradable polymers, in situ forming gels, or microsphere suspensions are expanding in areas ranging from psychiatric medications to oncology. Rheology determines whether these systems can be injected, how they form depots in tissue, and how they release drug over weeks or months. The interplay between polymer concentration, solvent composition, and shear behavior is complex and sensitive to small changes.

Pulmonary and nasal delivery are gaining traction for biologics and vaccines. Delivering drugs through the respiratory tract avoids first-pass metabolism and can provide rapid systemic or local effects. Aerosolization, however, requires precise control of droplet size, which depends on surface tension and dynamic interfacial behavior. Nebulization subjects formulations to high shear and creates large interfacial areas where proteins can adsorb and denature.

Wearable injectors and connected drug delivery devices sit at the intersection of formulation science and device engineering. Large-volume on-body injectors can deliver 3-10 mL of a biologic over several minutes, enabling higher doses without multiple injections. These devices have strict requirements for viscosity, since the formulation must flow reliably through narrow tubing driven by spring mechanisms or small motors. Bubble formation, which depends on surface tension and foam behavior, can interrupt dosing or cause delivery failures.

Next-generation biologics like bispecific antibodies and antibody-drug conjugates are structurally more complex than conventional monoclonal antibodies. They have additional sites for aggregation, are often more sensitive to interfacial stress, and may exhibit unusual rheological behavior. Formulation development for these modalities requires particularly careful attention to physico-chemical properties.

The Properties That Matter

Four physicochemical properties have emerged as central to biopharmaceutical formulation: colloidal stability, viscosity versus shear rate, surface tension at the air-water interface, and foam generation and decay. These properties are interconnected, and understanding that interconnection is essential for effective formulation design.

Colloidal Stability

Biologics are colloidal systems. Proteins, lipid nanoparticles, viral vectors, and cell suspensions all exist as dispersed phases that can aggregate, phase separate, or undergo structural changes over time. Colloidal stability refers to the ability of these dispersed systems to remain uniformly distributed without unwanted interactions.

For protein therapeutics, aggregation is the primary stability concern. Aggregates can form through multiple pathways: physical association driven by protein-protein interactions, chemical degradation that exposes hydrophobic regions, or interfacial adsorption that unfolds proteins and promotes clustering. The consequences are serious. Aggregates reduce drug efficacy by decreasing the concentration of active monomers. More concerning, they can trigger immunogenic responses in patients, potentially generating anti-drug antibodies that neutralize therapeutic effect or cause adverse reactions.

Lipid nanoparticles face different stability challenges. LNPs can fuse with each other, leading to particle growth and changes in biodistribution. They can lose their encapsulated payload through leakage. The lipid bilayer itself can undergo phase transitions or oxidation. Maintaining the narrow particle size distribution required for consistent pharmacokinetics demands careful control of formulation conditions and storage.

Viral vectors are perhaps the most sensitive of all. Structural integrity of the viral capsid or envelope is essential for infectivity, and even mild stresses can reduce potency. Shear during processing, freeze-thaw cycles, ionic strength changes, and interfacial exposure can all damage vectors irreversibly.

Stability testing in biopharma is time-consuming and expensive, often requiring months of real-time storage studies at multiple temperatures. Accelerated stability studies can help predict long-term behavior, but the complexity of degradation pathways in biological systems makes extrapolation uncertain. In many cases, subvisible aggregates emerge, much before visible effects take place. Early detection of subvisible particles is still a characterization challenge as it relates to understanding early-stage instability in these systems. By the time instability is detected, substantial resources may have been invested in a formulation that cannot succeed.

Viscosity Versus Shear Rate

Measuring viscosity at a single shear rate is not sufficient for biopharmaceutical formulations. These systems typically exhibit non-Newtonian behavior, meaning their viscosity changes depending on how fast they are flowing. Understanding the full viscosity-shear rate relationship is essential for predicting behavior during manufacturing, filling, and injection.

High-concentration protein solutions often show shear-thinning behavior: viscosity decreases as shear rate increases. Shear-thinning is actually beneficial for injectability, since the formulation flows more easily during the high-shear conditions of injection through a fine needle. At rest, higher viscosity can help prevent sedimentation or phase separation. The degree of shear-thinning depends on protein concentration, ionic strength, pH, and the presence of excipients, and these dependencies are difficult to predict from first principles.

At very high protein concentrations, protein-protein interactions become increasingly important. Attractive interactions can dramatically increase viscosity, while repulsive interactions can mitigate the concentration effect. Small changes in formulation conditions can shift the balance between attraction and repulsion, causing large changes in rheological behavior. Such sensitivity makes formulation development iterative and time-consuming.

Rheological behavior also matters during manufacturing. Pumping, filtration, and filling operations expose formulations to a range of shear rates. If viscosity at those shear rates is too high, processing becomes difficult or impossible. If shear sensitivity is too great, the act of processing can alter the formulation’s properties or damage sensitive components like viral vectors.

For depot formulations and polymer-based systems, rheology determines even more fundamental aspects of product behavior. The viscosity profile affects how the formulation flows through a needle, how it forms a depot at the injection site, and how drug releases over time.

Surface Tension at the Air-Water Interface

Proteins are inherently surface-active. They contain both hydrophobic and hydrophilic regions, and given the opportunity, they will adsorb to interfaces between air and water or between oil and water. Adsorption at interfaces is not benign: proteins that reach the air-water boundary often partially unfold, exposing hydrophobic regions that promote aggregation.

During manufacturing and fill-finish operations, biopharmaceutical solutions encounter numerous interfaces. Pumping creates turbulence and air entrainment. Filtration generates new interfaces as liquid passes through membrane pores. Filling introduces the formulation to the headspace of vials, syringes, or cartridges. Each of these steps creates opportunities for interfacial damage.

Surfactants like polysorbate 20 and polysorbate 80 are added to protein formulations specifically to compete for interfacial sites. The surfactant reaches the interface faster than the protein, occupying the available area and preventing protein adsorption. Protective efficacy depends on surfactant concentration, the relative surface activities of the surfactant and protein, and dynamic factors like how quickly new interfaces are created.

Surfactant concentration must be carefully optimized. Too little surfactant leaves the protein vulnerable to interfacial damage. Too much surfactant can cause other problems, including potential effects on stability, interactions with container closure systems, or regulatory concerns about excipient levels. Polysorbates themselves can degrade during storage, losing their protective function and generating degradation products that may affect the formulation.

For lipid nanoparticles, surface and interfacial behavior affects particle formation, stability, and biological function. LNPs are assembled through processes that involve rapid mixing of lipid and aqueous streams, with particle characteristics depending on the interfacial properties of the lipid components. Changes in surface behavior can alter particle size, encapsulation efficiency, and ultimately therapeutic performance.

Foam Generation and Decay

Foam might seem like a minor issue, but it creates real problems in biopharmaceutical manufacturing. Foam forms when air becomes incorporated into liquid and stabilized by surface-active components. In protein solutions, the proteins themselves can stabilize foam by forming films at bubble surfaces. These films trap protein at air-water interfaces precisely where unfolding and aggregation can occur.

Foam generation during mixing, pumping, or filling can increase particulate levels in the final product. Foam can interfere with filtration by blocking filter pores, cause dose variability if air bubbles end up in filled containers, and in some cases create equipment problems or slow down filling operations.

Foam behavior over time matters as well. Some foams collapse quickly as liquid drains from bubble films, while others persist for extended periods. Stability of foam depends on the viscoelastic properties of the interfacial film, which in turn depends on the surface-active components present. Formulations that generate persistent foam are particularly challenging to process.

Foam is also relevant for delivery devices. Air bubbles in prefilled syringes or autoinjector cartridges can affect dose accuracy and potentially cause discomfort during injection. Wearable injectors that deliver large volumes over extended periods are particularly sensitive to bubble formation, since air entrainment can interrupt drug delivery.

Why These Properties Are Interconnected

The challenge in biopharmaceutical formulation is that these four properties do not vary independently. Changing one aspect of a formulation typically affects multiple properties simultaneously, and optimizing for one target can move another property out of specification.

Consider the surfactants used to protect proteins from interfacial damage. Adding polysorbate reduces protein adsorption at the air-water interface and can decrease foam stability by altering the interfacial film. Surfactants can also affect protein-protein interactions and potentially influence colloidal stability. At high surfactant concentrations, micelles form that may interact with proteins in complex ways. The surfactant that solves one problem may create another.

Protein concentration is an even more connected variable. Increasing concentration to reduce injection volume increases viscosity, potentially to levels that impair injectability. Higher concentration also increases the frequency of protein-protein encounters, raising aggregation risk. More protein available to adsorb to interfaces may affect surface tension and foam behavior as well.

Ionic strength, pH, and buffer composition all affect multiple properties. Changing ionic strength can modulate protein-protein interactions, shifting viscosity and stability simultaneously. pH affects protein charge and conformation, which influences aggregation propensity, viscosity, and surface activity.

Such interconnection makes formulation development inherently iterative. Improving one property often requires re-checking others. The design space is constrained from multiple directions simultaneously, and finding a formulation that meets all specifications can require extensive experimentation.

Where Predictive Tools Fit

The traditional approach to biopharmaceutical formulation is empirical: propose a formulation, measure its properties, identify shortcomings, and iterate. This approach eventually works, but it is slow and resource intensive. Measuring viscosity versus shear rate requires rheometer time. Stability testing takes months. Foam and surface tension measurements require specialized equipment and expertise. Each iteration through the development cycle consumes expensive protein and delays timelines.

Complexity of the design space compounds the problem. With multiple interacting variables and multiple property targets, the number of possible formulations grows rapidly. Systematic exploration through design of experiments can map response surfaces, but even efficient experimental designs require substantial investment when each measurement is time-consuming and materials are costly.

Predictive modeling offers a different approach. If the relationships between formulation composition and physicochemical properties could be captured in computational models, formulators could explore the design space virtually before committing to experiments. They could identify regions likely to meet specifications and avoid regions likely to fail. Instead of testing exhaustively, they could test strategically.

Building such models is difficult because the underlying physics is complex and context dependent. Protein-protein interactions depend on sequence, structure, and solution conditions in ways that are not easily reduced to simple rules. Interfacial behavior involves dynamic processes that resist first-principles prediction. Standard machine learning approaches struggle with the small datasets and chemical complexity typical of formulation development.

Chemistry-aware models trained on physicochemical data offer a path forward. By learning from experimental measurements across diverse formulations, these models can capture relationships that are too complex to derive analytically. They can predict colloidal stability trends, viscosity profiles, and interfacial behavior for new formulations before synthesis.

FastFormulator has developed this capability through a combination of proprietary data generation and purpose-built machine learning architectures. The platform includes models for predicting viscosity versus shear rate behavior, colloidal stability indicators, and surface tension characteristics, with the capability to create custom models for other relevant properties. These predictions allow formulation scientists to prioritize experiments on candidates most likely to succeed and to understand how changes in composition will affect multiple properties simultaneously.

Predictive tools do not replace experimental validation. Biologics are too complex and too important for unsupported predictions. They can, however, substantially reduce the number of experiments required to find a viable formulation and accelerate the path from early development to clinical supply.

Takeaways

The shift toward complex biological modalities and patient-centric delivery has made physicochemical properties central to biopharmaceutical success. Colloidal stability, viscosity behavior, surface tension, and foam characteristics are no longer secondary considerations to be addressed late in development. They are design variables that determine whether a promising molecule can become a manufacturable, deliverable product. Formulation failures in these areas delay programs, increase costs, and in some cases prevent good therapies from reaching patients.

The difficulty is that these properties are interconnected through underlying physics and chemistry. Protein concentration affects viscosity, aggregation propensity, and surface activity simultaneously. Surfactants that protect against interfacial damage influence rheology and stability. Buffer composition modulates protein interactions that manifest across multiple measured properties. Optimizing a formulation requires navigating this coupled design space, and traditional empirical approaches are resource-intensive and slow.

Predictive tools built on chemistry-aware modeling can help by providing early insight into how formulation choices affect multiple properties. By identifying promising regions of the design space before extensive experimentation, these tools allow development teams to focus resources on candidates most likely to succeed. The result is faster development, fewer failed iterations, and a more reliable path from discovery to delivery.

References

• Shire, S.J., Shahrokh, Z., and Liu, J. (2004). Challenges in the development of high protein concentration formulations. Journal of Pharmaceutical Sciences, 93(6), 1390-1402.
• Wang, W. (2005). Protein aggregation and its inhibition in biopharmaceutics. International Journal of Pharmaceutics, 289(1-2), 1-30.
• Carpenter, J.F., et al. (2009). Overlooking subvisible particles in therapeutic protein products: gaps that may compromise product quality. Journal of Pharmaceutical Sciences, 98(4), 1201-1205.
• Garidel, P., et al. (2017). High-concentration protein formulations: How high is high? European Journal of Pharmaceutics and Biopharmaceutics, 119, 353-360.
• Hou, J.J.C., et al. (2022). Lipid nanoparticles for mRNA delivery. Nature Reviews Materials, 6, 1078-1094.
• Bee, J.S., et al. (2009). Response of a concentrated monoclonal antibody formulation to high shear. Biotechnology and Bioengineering, 103(5), 936-943.
• Allmendinger, A., et al. (2014). Rheological characterization and injection forces of concentrated protein formulations. European Journal of Pharmaceutics and Biopharmaceutics, 87(2), 318-328.
• Patapoff, T.W. and Esue, O. (2009). Polysorbate 20 prevents the precipitation of a monoclonal antibody during shear. Pharmaceutical Development and Technology, 14(6), 659-664.
• Grand View Research (2024). Biologics Market Size, Share & Trends Analysis Report, 2024-2030.
• Schoenmaker, L., et al. (2021). mRNA-lipid nanoparticle COVID-19 vaccines: Structure and stability. International Journal of Pharmaceutics, 601, 120586.