Assessing the impact of viscosity lowering excipient on liquid-liquid phase separation for high concentration monoclonal antibody solutions

Abstract
With continued interest in high concentration monoclonal antibody drug products to meet subcutaneous administration requirements, there is heightened attention on balancing protein-protein interactions, solution properties and overcoming instabilities such as increased in viscosity, particle formation, loss in potency, and aggregation of drug products. L-arginine hydrochloride is a commonly used viscosity reducing excipient used to influence protein-protein interactions of high concentration of mAbs. Contrary to literature, we observed that slight modifications to L-arginine hydrochloride concentrations in model drug product formulations can result in liquid-liquid phase separation if excipient and pH conditions are not well tightly controlled. We utilized a biophysical toolkit to assess the potentials of liquid-liquid phase separation (LLPS) that informs the limits of excipient and pH levels using structural- and molecular interaction-based assessments. While liquid-liquid phase separation observed in this study is reversible and does not impact inherent protein folding and structure, we demonstrated that increased ionic content in the formulations can significantly alter the balance of osmolarity toward the occurrence of LLPS. The aim of this work is to demonstrate the diversity of the toolbox used to evaluate the observed LLPS and the decision-making for optimization formulation development.
Introduction
Delivering monoclonal antibodies (mAbs) via the subcutaneous route of administration is increasingly desired during drug product development as it yields to better patient-centric, making new therapies more accessible and easing the burden on the healthcare system1,2. As mAbs are highly selective therapeutics, larger subcutaneous doses are typically employed to achieve maximum therapeutic levels for optimal pharmacokinetic exposure compared to an IV based regimen therapy3. This consideration is under the assumption that doses between IV and SC are different based on clinical pharmacology attributes, blood pharmacokinetic parameters such as Cmax, Cmin, Tmax, AUC, and concentration within the therapeutic window. For subcutaneous delivery, the dose can be altered depending on the injection volume or the drug product (DP) concentration, while the delivery can be managed with a host of device options4,5. Historically, for subcutaneous injections, volumes within the ranges of 0.5 – 2.25 mL are used without causing substantial injection site related pain, however up to 5 mL may be tolerated in the abdomen due to the limiting size of the subcutaneous space5. More recently, work has focused on trying to increase subcutaneous injection volume to greater than 3 mL (i.e. up to 25 mL has been tried in the clinic6) and this area of drug product delivery is growing1. Another approach for high dose mAb product development is to push the drug product concentrations beyond the 100 mg/mL paradigm and evaluate its potential in aiding for a robust subcutaneous delivery. The aim is to achieve the high concentration while understanding key structural and analytical focal areas enabling stable higher concentration DPs. Alternative technologies have looked at atypical approaches to managing high to ultra-high concentrations of drug products that can be delivered either in small volumes or moderately large volumes managed via viable device optionality.
With increasing mAb concentrations, the propensity of protein-protein interactions increases substantially and the need to control these interactions by varying different excipient levels takes priority in drug substance/drug product formulation and process development. The excipient levels in the formulation require modulation for favorable solution properties that enables ease of manufacturing, aiding stable products over its shelf life, and patient acceptability and useability7. Often there is a fine balance between achieving optimal solution properties and colloidal stability in a high concentration mAb formulation that is indicative of a stable DP. A big watch-out for high concentration mAb formulations is the propensity to form highly viscous solutions primarily due to intermolecular interactions and changes to associated rheology, changes to intramolecular-excipient interactions, and sheer sensitiveness8.
A typical formulation strategy (primarily driven to reduce viscosity and improve manufacturability) is by modulating protein-protein interactions and behavior of physiochemical properties in different buffer systems. Increased ionic strength is typically known to decrease viscosity through shielding protein charges on either hydrophobically or solvent exposed surfaces and weakening the long-range attractive interactions between protein molecules9. While increased ionic strength decreases repulsive interactions, it may also increase protein-protein attractive forces and hydrophobic interaction between the heterogenous distribution of charge and hydrophobic residue of mAbs10. This can in turn result in an alteration of balance of osmolarity and self-association interaction causing either greater reversible or non-reversible protein self-associations such as opalescence, and liquid-liquid phase separation (LLPS)11. While opalescence is more of an aesthetic problem, LLPS can inhibit dose administration and have a detrimental effect on colloidal stability impacting overall drug product stability, ability to meet batch release criteria’s, and use-case under different administration settings. It can also lead to rejection of manufactured clinical or commercial DP batches, and is shown to be associated with long-term stability impacts.12. LLPS is usually a thermodynamically reversible phenomenon in which mixtures of different components separate into two liquid phases13. During LLPS there is nucleation of droplets or protein domains (in mAb formulations), which is dependent on temperature and composition, explaining an interplay between thermodynamics and kinetics14. As lower temperatures reduce the activation energy of molecules, movement is restricted leading to the nucleation that may otherwise be prevented at higher temperatures with higher energy reserves15. The temperature reversibility can be limiting for the DPs stability, storage and use at various temperature conditions. While it is favorable to increase number of excipients or the excipient concentrations to improve solution properties such as viscosity and modulate protein-protein interactions, care and optimization must be taken to ensure excipients do not shift the balance of attractive-repulsive forces between proteins, leading to potential LLPS. For the production of highly concentrated monoclonal antibodies (mAbs) during drug substance downstream processing, understanding the Donnan effect via DLVO theory on electrostatic interactions, protein charge density, and excipient levels is vital for ultra-diafiltration of buffer systems. Chemistry manufacturing controls (CMC) for drug substances and drug products considers DLVO theory (Derjaguin–Landau–Verwey–Overbeek theory) to optimize high concentration mAb formulations for yields, purity, and process efficiencies. Strategies in formulation and processing must balance both protein repulsion and self-association like phenomenon’s.
Modifying excipients, whether by adding new ones or adjusting quantities, can impact mAb interactions, solution properties, and stability. These changes depend on alterations to surface charge, hydrophobic/hydrophilic exposure, water molecule exclusion, and dynamic behavior of self-association, protein structure dynamics, folding-unfolding kinetics of the solution. Previously, it has been demonstrated that L-arginine hydrochloride (HCl) is an effective excipient at reducing viscosity and LLPS for majority of the IgG formulations at either 50 or 100 mg/mL and marketed products in the market16,17. It was summarized that the LLPS phenomena was reduced likely due to decreases in attractive forces between proteins by shielding protein charges and reducing key intermolecular interactions between either Fab-Fc or Fc-Fc. Due to the ability to shield protein charges, L-arginine HCl has also been used as a viscosity reducing agent. However, as explained earlier, it must be noted that while increasing excipients can shield charge, it may have other downstream effects, leading to other intermolecular interactions with the protein and excipients and may further contribute to LLPS.
The regulation of excipient incorporation, along with the maintenance of rheologically pertinent solution properties, balanced protein-protein interactions, and optimal colloidal stability, presents a significant challenge18. During the formulation development of an IgG1 monoclonal antibody, L-arginine HCl was used primarily to reduce viscosity. Although the formulation showed favorable results in molecular and biophysical assessments, an increase of 10% in excipient concentration and a 0.8-unit increase in pH resulted in reversible liquid-liquid phase separation (LLPS). The occurrence of LLPS has prompted additional investigation into its effects on the IgG isotype formulation and the contributing factors involved. Biophysical characterization techniques such as diffusion interaction parameter and thermal melting temperature are commonly used to predict the stability of mAb formulations19. However, there are limitations to these tools for high concentration mAb formulations as they are typically conducted at low concentration and predict stability at accelerated conditions20,21.
To investigate the impact of varying formulation conditions on LLPS formation, we utilized a biophysical toolbox to evaluate the potential occurrence of LLPS across different formulations with variations in excipients and pH levels. Our assessment involves implementing molecular assessments of protein species via size-exclusion chromatography, isoelectric focusing capillary electropheris (iCE), and microfluidic modulation FTIR infrared (MM-IR) spectroscopy, in addition to biophysical characterization by dynamic light scattering, and differential scanning fluorimetry. We used a plate-based method with poly-ethylene glycol (PEG) to study forced precipitation and assess the apparent solubility of the mAb formulation. Lower solubility in PEG-6000 indicated liquid-liquid phase separation (LLPS). An inverse linear relationship was found between osmolality and solubility, suggesting that adjusting ionic strength/osmolality could reduce LLPS risk during processing or storage. While changes in excipients may not affect viscosity, thermal stability, or protein interactions, they can alter protein solubility and increase LLPS risk, potentially leading to failure to meeting drug product specifications.
Read more here
Chelsea R. Thorn, Deep Bhattacharya, Lindsey Crawford, Vicky Lin, Advait Badkar, Parag Kolhe, Assessing the impact of viscosity lowering excipient on liquid-liquid phase separation for high concentration monoclonal antibody solutions., Journal of Pharmaceutical Sciences, 2025, 103804, ISSN 0022-3549, https://doi.org/10.1016/j.xphs.2025.103804.