Aggregates in MAbs and Recombinant Therapeutic Proteins: A Regulatory Perspective

November 1, 2008

BioPharm International

Volume 21, Issue 11

FDA perspectives on specs and effective control strategies.


The dynamics of a protein aggregate mixture are complex and require multiple analytical methods for detection, evaluation, and monitoring. A successful assessment of aggregates in protein-based pharmaceuticals rests on a science-driven, risk-based program that evaluates the protein stability profile and determines the impact that a given aggregate mixture will have on the safety and efficacy of the drug product during its lifecycle. This article provides a regulatory perspective on aggregates in protein-based pharmaceuticals, including their characterization, detection methods, and various control strategies that have been implemented by manufacturers.

Aggregates in biologic products differ in origin, type, and size, and are caused by multiple factors. Of particular interest to regulatory agencies are aggregates that have the potential to enhance immune responses causing adverse clinical effects, or aggregates that may compromise the safety and efficacy of the drug product. Enhanced immune responses to protein aggregates have been reported in animal and clinical studies.1–4 Although classical immune responses to proteinaceous material foreign to humans are expected, the immune system may mount a strong response to aggregated protein preparations that have endogenous counterparts through a tolerance breakdown mechanism. In the mechanism of tolerance breakdown, protein aggregates may serve as facilitators in the formation of protein complexes that trigger B cell production of antibodies against the protein, independent of T-helper cell recruitment. The basis for this type of response comes from the immunon concept, in which antigen presented as a polymeric structure with more than 10 haptens, spaced 5–10 nm apart in a viral-like particle organization, and sized above 100 kDa, can overcome immune tolerance.5,6 Such a scenario may account for the unexpected neutralization of the endogenous protein, and have a profound clinical effect. High molecular weight (HMW) aggregates that conserve most of the native configuration of their monomer counterpart, and that can nucleate haptens in this way, are of most concern for this type of mechanism.7 Alternatively, aggregates displaying non-native protein conformations may be seen by the immune system as neoantigens, which could trigger antibody formation. 7 This article provides a regulatory perspective on aggregates in protein-based pharmaceuticals, their characterization, detection methods, and various controls that have been implemented by drug manufacturers.


The classification of aggregates is a difficult task and currently, there is no comprehensive classification available. The difficulty resides in the multiple categories within which aggregates can be grouped. Table 1 lists the most common aggregate categories found in biopharmaceuticals. Dimers, reversible aggregates, covalent aggregates, and particulates are all terms used and, for lack of a better classification, help in understanding the type of entity being discussed, but also add to the confusion in the literature.

Table 1 An alternative classification of aggregates could be based on their sizes, as this may have a direct correlation to potential adverse clinical outcomes. Aggregate size ranges from the soluble dimers and other multimers (approximately 5–10 nm in apparent globular diameter) including high molecular weight (HMW) aggregates through nucleated aggregates that can be either soluble or insoluble, to larger, insoluble species identified as subvisible and visible particulates (approximately 20–50 µm in apparent globular diameter). From the soluble aggregates group, the larger ones such as the HMW species may be more capable of eliciting immunogenic responses that could have an adverse clinical outcome.7 With respect to their molecular weight, aggregates of sizes above 102 kDa may deserve more careful evaluation based on their potential for undesirable immunogenic responses.5,6

In addition to aggregate size, the rate of aggregate change in size over time, where time corresponds to the protein product lifecycle, is a useful parameter that can provide a functional characterization of aggregates. An aggregate can initially exist as a small dimer or fragment, and progress toward larger structures, such as subvisible or visible particles, if such a transition becomes thermodynamically favorable. At any given moment, a protein may be transitioning between a thermodynamic state that favors the monomer or native configuration of the protein, and an intermediate state that favors an unfolded native-like protein configuration. Under the right set of conditions, the unfolded protein may form a complex with other native and non-native forms, gaining enough free energy to transition to an aggregated state that may become the most stable state of the new proteinaceous entity. The basis for these transitions was studied by Lumry and Eyring in the 1960s, who proposed the kinetic transitions of protein aggregates in solution, and was further modified to first-order transition reaction kinetics in a case of interferon-γ aggregation.8,9 Evaluating the rate of aggregate growth in a drug product is complex. A single drug product typically has a non-homogeneous aggregate mixture (Table 1). Stable protein preparations will have aggregates present in a solution of a heterogeneous nature, but the rate of growth is minimal compared with more unstable preparations. Aggregates that have increased rates of growth are more worrisome and more aggressive strategies of control and minimization may be needed.


Drug product aggregates may originate from multiple sources and various types of aggregates may be present in a given drug product vial. The potential for protein aggregate formation exists at all levels of protein-based pharmaceutical manufacturing. Starting with the sequence and characterization of the protein, each protein will have a physicochemical signature that can render it more or less stable. This is the case of a CHO-derived monoclonal antibody (MAb) found to aggregate in culture because of elevated levels of free thiol, which could be prevented if copper sulfate were added to the culture.10 Protein heterogeneity also can be a contributing factor for protein aggregation, as the probability of multiple protein forms interacting with their environment is increased. In the case of epratuzumab, disulfide bond scrambling favored covalent aggregate formation.11

Therapeutic MAbs are typically formulated at high concentrations that favor the increased incidence of molecular interaction, and therefore, the potential for aggregate formation. Manufacturers therefore dedicate much time and effort to developing a formulation that will keep the protein drug product stable during its lifecycle, whether in solution or lyophilized. For example, adding sucrose to interleukin-1 receptor agonist or to a lyophilized MAb inhibited or reduced aggregate formation respectively. 12,13

Bulk freezing presents a challenge to stable protein preparations because of the solute concentration effect that occurs during the freezing of solutions. An ideal strategy is to freeze the entire solution at the same time and rapidly at –80°C where thermal transitions (such as eutectic melting) and glass transitions are minimized. This strategy is not practical for large-volume solutions. Changes in solute concentration and pH during bulkfreezing can promote protein aggregation.14,15 Bulk thawing also presents challenges, primarily related to surface adsorption at the ice–liquid interface. The appropriate formulation becomes critical when bulk freezing and thawing is required, and excipients may serve as protein cryoprotectants.

Fill-and-finish operations may use pumps that can mechanically denature the protein because of shear stress or introduce impurities that serve as nucleation sources of protein aggregates. Some piston-displacement pumps, for example, can interact with protein drug product in a similar way that a car motor engine piston interacts with lubricant oil. The intimate contact between protein drug product and a piston rod can disrupt an otherwise stable drug product. This was the case for an antibody drug product that was found to have increased levels of aggregate particles with more pump passes, as determined by absorbance and light obscuration methods.16 In the case of piston shedding, stainless-steel passivation should reduce the risk of introducing stainless-steel deposits to the final drug product, which has the potential to induce protein heteronucleation.17

New delivery systems have increased the complexity of container compatibility, along with the potential for protein aggregate formation. Glass from vials, rubber from stoppers, silicone from stoppers and syringes, and tungsten from syringes are some of the foreign particles that may find their way into a protein drug product. Many of these foreign particles are electrostatically charged, and therefore, have the potential to interact with proteins, protein aggregates, and protein aggregate precursors, to form heteronuclei. Such was the case with prefilled syringes containing tungsten particles shed during syringe barrel manufacturing, which served as nuclei for aggregate formation.18


Based on aggregate kinetic models, protein aggregates may be more hydrophobic than their monomeric counterparts. This is because protein aggregation may occur when the protein transitions to a native, partially unfolded state, exposing its hydrophobic residues.8,9 Studies revealed that aggregates are precipitated better than monomers by ammonium sulfate, and that they bind more strongly to polyvinylidene fluoride (PDVF) membranes.19

Commonly, aggregates are studied by exposing the protein solution to extreme conditions of temperature, pH, humidity, and photon incidence, in what are known as forced and photodegradation studies. The rationale is based on the expectation that protein degraded in this manner reflects the degradation pathway(s) experienced during the protein lifecyle. These parameters may also be of value when establishing a stress stability program.20

Another important part of protein aggregation studies is evaluating the biological activity of the aggregate. Differences in biological activity of the aggregates compared to the activity of the monomeric protein can profoundly influence the potency of a protein-based drug. In such cases, product efficacy may be compromised. In general, a risk-based assessment of aggregates may warrant specific studies that may help elucidate which types of aggregates are more worrisome. A thorough investigation of the different environments to which the drug product is exposed during its lifecycle, including manufacture, storage, shipping freeze and thaw cycles, oxygen exposure, light, and physical stress, can provide a rationale for special degradation studies such as seeding or spiking of a specific aggregate or impurity into a protein solution to observe the potential for further aggregation.


The biopharmaceutical industry has experienced an increase in the number of analytical methods available to detect, characterize, quantify, and monitor aggregates in biopharmaceutical protein products. A list of the most common methods used to detect, monitor, and study aggregates appears in Table 2. Although not meant to be comprehensive, Table 2 offers a general concept of the analytical methods and their main advantages and limitations. Test methods available can be clustered into two groups: those that detect small aggregates such as dimers, LMW, HMW, soluble aggregates, and protein fragments (first section of Table 2), and those that detect large aggregates, such as insoluble subvisible and visible particles (second section of Table 2). Within the group of test methods used to detect small aggregates, size exclusion chromatography (SEC) is commonly used for routine detection and monitoring of aggregates during lot release. Methods that do not lend themselves to lot release can be used for additional characterization or as confirmatory tests.

Table 2

Size exclusion high-pressure liquid chromatography is the most widely used analytical method for protein aggregates such as dimers and LMW, and HMW species. The method is amenable to validation for high throughput analytical testing with good sensitivity, precision, resolution, and accuracy. However, as a chromatographic method, it can also induce aggregation, cause existing aggregates to be removed during sample preparation, or underestimate the presence of aggregates when HMW species cannot be discriminated or recovered. Perhaps the main limitation of SEC is the need to inject samples at low concentrations (e.g., 1 mg/mL). For therapeutic proteins, this often means a 100-fold dilution causing reversible soluble aggregates to disaggregate. This concern leads to more general questions, such as, How relevant is the aggregate profile obtained? Does it represent the aggregates present in the final drug product? As noted elsewhere, there is not a single analytical method capable of evaluating all aggregates present in a given protein solution.21 A combination of methods is usually needed to cover the microscopic and visual range of aggregates that may be present. In addition, because of the limitations of each method, orthogonal confirmatory methods may be needed. For example, SEC results may require confirmation with other orthogonal methods, such as analytical ultracentrifugation (AUC). AUC can be used as a confirmatory method because it provides a good separation of aggregates, and it does not require sample dilution or sample preparation.

Among the methods used for aggregate characterization, field flow fractionation (FFF), and dynamic light scattering (DLS) are capable of evaluating aggregates directly in solution (without dilution). FFF has a detection range broader than that of SEC, but data interpretation of concentrated samples may be difficult. DLS is a good quantitative method, but the read-out is proportional to surface area so large aggregates can mask detection of small ones if differences in size are not enough. Calorimetry is very useful to evaluate the stability of a protein solution because it can detect protein unfolding. Some methods may be combined to have a more powerful tool, such as mass spectrometry coupled to chromatography, which provides information on chemical and physical degradation. In addition, DLS in combination with SEC can overcome the limitations that may be encountered using either method.

Microscopy and light obscuration methods are routinely used for detecting and counting subvisible particles according to United States Pharmacopeia (USP) chapter <788>. Both tests apply to small and large volumes, but in general, multiple drug product vial samples are used during this kind of testing. Normally, samples are first tested by the light obscuration method. If the sample fails the specified limits, the microscopic assay method can then be used. However, the microscopic method can be the sole test if there is a documented technical reason or interference from the product undergoing testing that would make the light obscuration method unsuitable or the results invalid. Turbidimetry measures the opalescence or clarity of a solution against a reference standard. A combination of light obscuration, turbidimetry, and DLS has been used to evaluate the time course of particle formation in protein solutions.22

Figure 1 When these methods are correlated to the apparent globular size of the aggregate detected (Figure 1), there is a size gap in the capability of methods used to detect and monitor small aggregates that range between a few nanometers and about 50 nm in diameter, and the methods used to detect and monitor large aggregates that range between a few micrometers and about 50 µm in diameter. This gap may pose a problem because the ability to detect, measure, and evaluate the fate of some small aggregates as well as the precursors of larger aggregates may not be implemented in protein aggregate control strategies. In fact, it has been pointed out that aggregates with apparent globular diameters around 0.5 µm are not routinely tracked and analyzed.23


Various approaches are taken by manufacturers to control aggregates in protein-based pharmaceuticals, which depend on the nature and levels of the aggregate, and their potential impact on the safety, quality, and stability of the specific protein product. Although a protein drug product may contain aggregates of different natures and sizes than can be controlled by proper monitoring, some aggregates may require aggressive control strategies depending on their risk assessment and evaluation. In some cases, controls are geared toward minimizing or inhibiting aggregate formation. Sometimes mutational changes can improve the stability of a protein product susceptible to aggregation, such as during freeze–thaw.24 Control strategies may also include the addition of excipients. Such is the case for a recombinant human platelet activation factor that showed aggregate formation by heteronucleation with silica particles on storage. The addition of surfactant pluronic acid F68, or a change in the pH of the formulated drug product, reduced heteronucleation formation.25 In other cases, control strategies are oriented toward increased monitoring and re-evaluation of the limits and specifications of the aggregate levels. This is the case with liquid Remicade, which showed higher turbidity than its lyophilized counterpart. The measures taken were to add a turbidity specification limit, and to tighten monomer specification by gel filtration–HPLC.26


A robust program that studies product aggregates will have the capability to establish limits for product aggregate levels. Whether aggregate levels are a stability indicator for the particular drug product or whether the aggregate mixture present in the drug product is assigned as low risk, it is important to accumulate adequate data to support the conclusions. The rationale for this is that the data accumulated during preclinical and clinical phases of product developmen, together with the study of the degradation pathways of the protein, will contribute to a better understanding of the impact that aggregates will have on drug product safety and quality.

There is no consensus on the maximum allowable limit for protein-based pharmaceutical aggregates because some proteins may be largely stable and safe despite certain levels of aggregates, while for other proteins very small changes in aggregate levels may profoundly affect protein stability, and even safety. By the time of license application, the manufacturer will have collected sufficient data to justify a control strategy. These data include results from clinical lots, information from in vivo and in vitro studies, the product stability profile, the type of aggregates, and their potential impact on product safety and quality. Usually this background allows a formal analysis that can predict the expected level of aggregates during the complete lifecycle of the product.

The only group of aggregates that has maximum allowable limits based on USP <788> (also European Pharmacopeia [Ph. Eur.] 2.9.19 and Ph.Eur. 2.9.20), is the group subvisible particles. Some manufacturers have initiated efforts to optimize alternative methods for evaluating subvisible particles to reduce the volume of sample required, and limitations that USP <788> may pose when analyzing high-concentration, highly viscous protein solutions using the recommended light obscuration and microscopy methods.27 However, the USP testing for sub-visible particles is designed to mitigate the risk associated with the presence of extraneous particles in injectable solutions that might lead to blood-vessel occlusion and were not intended to address safety issues potentially associated with large protein aggregates. Visible particles have only qualitative specifications according to USP <788>, which indicates that the injectable sterile solution should be essentially free from particulate matter that can be observed with visual inspection. Because of the subjective nature of visual inspection, care must be taken by manufacturers when routine analysis of visible particles is performed.


A specific regulatory guidance documents on biopharmaceutical protein aggregates is currently not available. The International Conference on Harmonization (ICH) has put forth guidelines that deal with drug product and drug substance impurities (ICH Q5C, Q6B, Q1AR). Stability guidances deal with the evaluation of impurities as part of the stability program during long-term, stress, accelerated, and photonic exposure (ICH Q1 series), and comparability guidances consider impurities during manufacturing changes (ICH Q5e). These guidances offer general guidelines for the evaluation of impurities, but do not specifically address aggregates as a separate issue in protein-based pharmaceutical manufacturing.

In general, the manufacturer relies strongly on its research and development group for a comprehensive program that assesses its drug product aggregates during characterization, manufacturing, and storage of the drug product. In addition, regulatory agencies encourage implementing aggregate specifications, and have seen a surge in relevant information provided by manufacturers that is based on sound analysis of their product aggregates. By the time of license application, a manufacturer should be capable of providing a data-supported justification for its specifications for aggregate levels.


There are important considerations to be made when assessing aggregates. Aggregates in protein-based pharmaceuticals can be viewed as an evolving mixture of various types, large and small, that actively undergo transitional equilibrium states. At any given time, the population of such a mixture may reach a new equilibrium and advance toward even larger aggregate structures. The dynamics of a protein aggregate mixture are complex and require multiple analytical methods for detection, evaluation, and monitoring.

Formulation remains a critical aspect of protein stabilization, and an important step in protein aggregate minimization. In the same way that chaperones assist endogenous proteins to maintain their folded state, the right combination of excipients, pH, and temperature keeps the protein-based pharmaceutical from assuming aggregate-permissive states.

A successful assessment of aggregates in protein-based pharmaceuticals rests on a science-driven, risk-based program that evaluates the protein stability profile and determines the impact that a given aggregate mixture will have on the safety and efficacy of the drug product during its lifecycle. Knowledge of the potential degradation pathways that a protein will encounter in every environmental exposure during its shelf-life sets the basis for a well-characterized protein product. This knowledge should be linked to preclinical, clinical, and manufacturing experience to have a better understanding of what types and quantities of aggregates should be allowable.


The views offered in this article represent the author's personal perspective on the issue and should not be construed as regulatory policy or guidance.


The author would like to thank Steve Kozlowski, Kathleen Clouse, Patrick Swann, Barry Cherney, and Chana Fuchs from the Office of Biotechnology Products at the Center for Drug Evaluation and Research, FDA, for their insightful comments.

Ruth V. Cordoba-Rodriguez, PhD, is a biologist in the division of monoclonal antibodies, Office of Biotechnology Products, Center for Drug Evaluation and Research, US Food and Drug Administration, Rockville, MD, 301.827.0454,


1. Dresser DW. Specific inhibition of antibody production: II. Paralysis induced in adult mice by small quantities of protein antigen. Immunol. 1962;5(3):378–388.

2. Braun A, Kwee L, Labow MA, Alsenz J. Protein aggregates seem to play a key role among the parameters influencing the antigenicity of interferon alpha in normal and transgenic mice. Pharm Res. 1997;14(10):1472–1478.

3. Moore WV, Leppert P. Role of aggregated human growth hormone (hGH) in development of antibodies to hGH. J Clin Endocrinol Metab. 1980;51:691–697.

4. Prümmer O. Treatment-induced antibodies to interleukin-2. Biotherapy. 1997;10(1):15–24.

5. Dintzis RZ, Okajima M, Middleton MH, Greene G, Dintzis HM. The immunogenicity of soluble haptenated polymers is determined by molecular mass and hapten valence. J Immun. 1989;143(4):1239–1244.

6. Bachmann M, Zinkernagel R. Neutralizing antiviral B-cell responses. Annu Rev Immunol. 1997;15:235–270.

7. Rosenberg A. Effects of protein aggregates: an immunological perspective. AAPS J. 2006;8(3):E501–E507.

8. Lumry R, Eyring H. Conformation changes of proteins. J Phys Chem. 1954;58:110–120.

9. Kendrick BS, Carpenter JF, Cleland JL, Randolph TW. A transient expansion of the native state precedes aggregation of recombinant human interferon-g. Proc Natl Acad Sci USA. 1998;95:14142–14146.

10. Chaderjian WB, Chin ET, Harris RJ, Etcheverry TM. Effect of copper sulfate on performance of a serum-free CHO cell culture process and the level of free thiol in the recombinant antibody expressed. Biotech. Prog. 2005;21:550–553.

11. Remmele JR. RL, Callahan WJ, Krishnan S, Zhou L, Bondarenko PV, Nichols AC, et al. Active dimer of epratuzumab provides insight into the complex nature of antibody aggregate. J Pharm Sci. 2006;95(1):126–145.

12. Kendrick BS, Chang BS, Arakawa T, Peterson B, Randolph TW, Manning MC, Carpenter JF. Preferential exclusion of sucrose from recombinant interleukin-1 receptor antagonist: Role in restricted conformational mobility and compaction of native state. Proc Natl Acad Sci USA. 1997;94:11917–11922.

13. Andya JD, Hsu CC, Shire SJ. Mechanisms of aggregate formation and carbohydrate excipient stabilization of lyophilized humanized monoclonal antibody formulations. AAPS J. 2003;5(2):1–11

14. Webb SD, Sesin DF, Kincaid AC, Webb JN, Hughs TG. Freezing bulk-scale biopharmaceuticals using common techniques and the magnitude of freeze concentration. BioPharm Int. 2002;5:22–34.

15. Pikal-Cleland KA, Carpenter JF. Lyophilization-induced protein denaturation in phosphate buffer systems: monomeric and tetrameric b‚ galactoside. J Pharm Sci. 2001;90(9):1255–1268.

16. Cromwell MEM, Hilario E, Jacobson F. Protein aggregation and bioprocessing. AAPS J. 2006;8(3):E572–E579.

17. Carpenter JF, Nikolai T. Pumping-induced aggregation of monoclonal antibodies. AAPS Workshop on protein aggregation, Oral presentation. Breckenridge, CO. 2006 Sep.

18. Narhi L. Effects of manufacturing processes on a therapeutic protein. AAPS Workshop on protein aggregation, Oral presentation. Breckenridge, CO. 2006 Sep.

19. Wang L, Hale G, Ghosh R. Non-size-based membrane chromatographic separation and analysis of monoclonal antibody aggregates. Anal Chem. 2006;78:6863–6867.

20. International Conference on Harmonization. Q1A(R2), Stability testing of new drug substances and products. Geneva, Switzerland;2003.

21. Philo JS. Is any measurement method optimal for all aggregate sizes and types? AAPS J. 2006;8(3):E564–E571.

22. Kiese S, Pappenberger A, Friess W, Mahler H-C. Shaken, not stirred: mechanical stress testing of an IgG1 antibody. J Pharm Sci. 2008;Oct;97(10):4347–66.

23. Mahler H-C. Break-out session on analytical methods. AAPS workshop on protein aggregation. 2006 Sep 26–27; Breckenridge, Colorado.

24. Lu Y, Harding SE, Rowe AJ, Davis KG, Fish B, Varley P, Gee C, Mulot S. The effect of a point mutation on the stability of IgG4 as monitored by analytical ultracentrifugation. J Pharm Sci. 2008;97(2):960–969.

25. Arakawa T, Kita Y. Protection of bovine serum albumin from aggregation by Tween 80. J Pharm Sci. 2000;89:646–651.

26. European Commission (Enterprise Directorate General). EMEA. Remicade Scientific Discussion. 2004.

27. Ives CM, Soderquist R, Stoner MR, Kendrick BS. Light obscuration particulate analysis for protein solutions: challenges and limitations. Colorado Protein Stability Conference. Poster presentation by Amgen. 2007.