Optimizing Protein Aggregate Analysis by SEC

Monoclonal antibodies (mAbs) are a dominant class of protein biotherapeutic that have achieved outstanding success in treating many life-threatening and chronic diseases. Indeed, more than 20 mAb targeted therapy products reached ‘blockbuster’ status in 2015 (1). With many commercially successful biologic patents now expired, or nearing expiry over the next few years, there exists a great opportunity for mAb-based biosimilars to enter the global market.

Due to the structural and compositional complexity of mAbs, there are a number of critical quality attributes that require monitoring to meet standards for human use. Protein aggregation, which may occur throughout the manufacturing process and during product storage, may be indicative of partial denaturation or other perturbations of protein structure, which can seriously affect product safety and efficacy. Robust analytical methods capable of quantifying the extent of protein aggregation are therefore essential to maintain product performance and patient safety.

Size-exclusion chromatography (SEC) is commonly used for this purpose. Without proper optimization of instrument set-up and conditions, however, the technique may experience several problems that can deleteriously affect analytical performance. Firstly, as SEC is one of the few chromatography methods that does not exhibit on-column focusing, optimizing pre-column dispersion is important to achieve suitable peak resolution. Additionally, proteins can undergo non-specific hydrophobic binding to the column, resulting in poor quality results.

This article presents a robust ultra-high-performance liquid chromatography (UHPLC) SEC approach for protein aggregate analysis of five structurally diverse biotherapeutic mAbs (bevacizumab, cetuximab, infliximab, rituximab, and trastuzumab) that employs a silica column with a covalently-bonded diol hydrophilic layer designed to prevent secondary interactions. It is also shown that pre-column dispersion can be minimized through optimization of instrument tubing and use of appropriate injection 
volumes.

Quantifying protein aggregation by SEC

Monoclonal antibodies are known to form aggregates in the course of product expression during fermentation, product purification in downstream processing, or in storage or through mishandling of the product prior to patient administration.

Aggregation of mAb monomers to dimers, trimers, and other higher order structures is undesirable for two key reasons. Firstly, aggregates may cause a decrease in product efficiency by lowering the effective concentration of the product. Secondly, aggregation can expose normally unexposed epitopes, leading to increased immunogenicity. To demonstrate the safety and efficacy of the mAb and gain regulatory approval, it is essential to monitor the formation of aggregation products throughout the production process.

SEC is the standard method for protein aggregate analysis. The technique involves passing molecules through a column containing porous polymer or silica beads. The choice of pore size is related to the size of the molecule to be separated. For the separation of mAbs and their aggregates, this is around 300Å.  Molecules are separated based on their hydrodynamic volume. Smaller molecules can penetrate fully into the pores of the stationary phase, while larger molecules cannot get totally inside the porous bead. Therefore, the larger molecules have less distance to travel and elute through the column more quickly.

One requirement of the technique is that the analyte does not interact with the surface of the stationary phase. Ideally, differences in elution time are based solely on a protein’s hydrodynamic volume, rather than its chemical or electrostatic interactions with the stationary phase.

However, mAbs are structurally diverse and can exhibit unwanted secondary interactions with residual groups on the column during analysis, affecting analytical data quality. Non-specific hydrophobic binding of proteins to the columns can, for example, lead to retention time shifts, peak tailing, or even a complete loss of protein peaks (2,3). A general SEC method applicable to a wide range of mAbs is therefore highly sought after.

Materials and methods

A UHPLC method, which employed a silica column (MAbPac SEC-1, Thermo Scientific) covalently modified with a diol-based hydrophilic layer, was used for the aggregate analysis of five therapeutically-important and structurally diverse mAbs (bevacizumab, cetuximab, infliximab, rituximab, and trastuzumab). These mAbs were chosen as they covered a range of isoelectric points (7.6-8.7) and possessed different glycosylation patterns, from very simple (bevacizumab) to highly complex (cetuximab).

Chromatography experiments were performed on a UHPLC system (Vanquish Flex Quaternary, Thermo Scientific), with UV detection at 214 nm. For aggregation determinations, a 7.8 x 300 mm (MAbPac SEC-1, Thermo Scientific) silica column was employed. This column was selected as the 300 Å pore size was well matched to give good separation of the molecular weight of proteins studied. Separation was performed at 30 °C with a flow rate of 1.0 mL/min. Pre-column dispersion experiments were also performed at 30 °C and used a 4 x 300 mm version of this column with a flow rate of 0.3 mL/min. All experiments used 0.2 M NaCl in 100 mM phosphate buffer (pH 6.8) in HPLC-grade water as the eluent and sample buffer. Monoclonal antibodies were solubilized in water (5-25 mg/mL) and injected directly into the instrument. An injection volume of 1 μL was used unless otherwise stated.

Determination of extent of mAb aggregation

Using the silica column, suitable resolution of aggregates and fragments from the monomer was achieved for all five mAbs, permitting determination of the percentage aggregation in each case. Table I shows the percentage aggregation for each mAb.

Table I: Percentage aggregation in each monoclonal antibody (mAb).
[ALL FIGURES ARE COURTESY OF THE AUTHORS.] 

Cetuximab and infliximab showed the lowest levels of aggregation and fragmentation. Bevacizumab exhibited a higher level of aggregation and a complex aggregation pattern, but had a lower level of fragmentation.

Given the narrow range of molecular weights of the proteins studied, the similar retention times obtained for the samples, shown in Table II, indicates a lack of secondary interactions with the column. This conclusion is further supported by the good symmetry of the chromatogram peaks. The addition of solvent did not improve the symmetry of the infliximab peak, which exhibited the worst asymmetry. 

Table II: Comparison of monoclonal antibody (mAb) retention times and peak asymmetries. 

In general, the expected relationship between retention time and size was observed, with the largest mAb, cetuximab, eluting first, and the smaller mAbs eluting in reverse molecular weight order. The exception to this rule is bevacizumab, which elutes last, although it is not the smallest protein. This may be a result of protein folding or differences in the N-glycan structures present on this molecule relative to the other mAbs studied. These subtle effects are magnified due to the narrow molecular weight range of the analytes.

These results demonstrate the applicability of this method for accurate aggregate analysis of this broad range of commercially important mAbs. The data show that the silica column eliminates non-specific interactions with the column often encountered during SEC analysis.
 

Effect of pre-column dispersion on analytical performance

SEC is one of the few chromatography methods that exhibits no on-column focusing. Therefore, careful control of pre-column dispersion is essential to achieve optimum separation results, especially at reduced flow rates on smaller identity (i.d.) columns as broad peak volumes are not focused at the column head. To investigate the effect of pre-column dispersion on analytical performance, pre-column tubing of various diameters were used in combination with 4 and 7.8 mm i.d. columns for the analysis of bevacizumab.

Using the 7.8 mm i.d. column at the higher flow rate of 1.0 mL/min, there was essentially no difference in the produced chromatograms when the pre-column tubing was changed from the standard i.d. of 100 μm to 75 μm. Peak width, asymmetry, and resolution were the same in both analytical runs, as shown in Figure 1.

Figure 1: Effect of tubing on 7.8mm size-exclusion chromatography (SEC).

In contrast, using the 4 mm i.d. column at the lower flow rate of 0.3 mL/min, pre-column dispersion had a significant impact on analytical performance. Using 180 µm i.d. tubing (typical of a standard HPLC system) produced a marked reduction in peak shape quality (Figure 2). While the relative area of the main peak remained consistent, notable dispersion and reduced sensitivity were observed, with an increase in peak width at half height of more than 40%. In reality, this effect would be compounded by the dispersion associated with the injection valves of older HPLC systems. Reducing tubing i.d. resulted in a significant improvement in peak resolution, with an optimum tubing diameter of 75 µm. These findings highlight the importance of correct instrument set-up when using UHPLC systems, operating at lower flow rates.

Figure 2: Bevacizumab monoclonal antibody (mAb) with a 4mm size-exclusion chromatography (SEC) column. UHPLC is ulta-high-performance liquid chromatography.

As sample injection is an inherent source of dispersion in HPLC methods, the effect of injection volume on analytical performance was also studied. The chromatogram obtained for a 1 μL injection of a solution of bevacizumab was compared against that of a 10 μL injection of a 10-fold dilution of the same sample (Figure 3). The 10 μL injection volume resulted in greater peak width and a loss in resolution and definition of the smaller aggregate peak, highlighting the importance of maintaining small injection volumes for SEC.

Figure 3. Changes in injection volume on 4mm size-exclusion chromatography (SEC).

Conclusion

Robust and accurate protein aggregation analysis was achieved for five structurally diverse mAbs using a SEC method that employed a silica-based column. The column’s hydrophilic diol layer eliminated the non-specific protein-column interactions often encountered with this technique, producing high resolution chromatographic peaks that allowed the percentage protein aggregation to be determined in each case. The poor peak resolution associated with pre-column dispersion at lower flow rates could be minimized through the use of pre-column transfer tubing of narrower i.d. and smaller injection volumes. Through correct instrument optimization and the appropriate choice of chromatography columns, protein aggregation analysis by SEC can be a useful technique to ensure the safety and efficacy of mAb-based biotherapeutic products.

References

1. Blockbuster Biologics 2015: Sales of Recombinant Therapeutic Antibodies & Proteins, 2015
2. P. Hong, S. Koza, E. S. P. Bouvier,  J. Liq. Chromatogr. Relat. Technol. 2012; 35(20): 2923-2950. 
3. T. Arakawa et al., BioProcess International, 2006, 4(10), 42-43. 

Article Details

BioPharm International
Volume 30, Number 10
October 2017
Pages: 46-47, 50-51

Citation

When referring to this article, please cite it as A. Farrell, J. Bones, K. Cook “Optimizing Protein Aggregate Analysis by SEC," BioPharm International 30 (10) 2017.

About the Authors

Amy Farrell, PhD, is an applications development team leader at NIBRT; Jonathan Bones, PhD, is principal investigator at NIBRT; and Ken Cook, PhD, is a bio-separations support expert at Thermo Fisher Scientific.