Development and validation of a (RP)UHPLC-UV-(HESI/Orbitrap)MS method for the identification and quantification of mixtures of intact therapeutical monoclonal antibodies using a monolithic column

https://doi.org/10.1016/j.jpba.2018.07.013Get rights and content

Highlights

  • Mixtures of intact mAbs can be separated by (RP)LC with divinylbenzene-based monolithic column.

  • (RP)UHPLC-UV-(HESI/Orbitrap) MS method validated for intact mAb quantification.

  • MS isoform profile depends on the ion pairing agent used.

  • New heteroscedasticity function to predict standard deviation associated with mAb concentration.

  • Bevazizumab, cetuximab, infliximab, rituximab and trastuzumab were quantified in mixtures.

Abstract

Monoclonal antibodies (mAbs) are one of the most important types of biopharmaceutics and have proved enormously successful in the treatment of cancers and autoimmune diseases. In this paper, we present a fast, straightforward reversed phase (RP)UHPLC-UV-(HESI/Orbitrap) MS method for the separation and identification of five of the most commonly used mAbs, i.e. bevazizumab (BEV), cetuximab (CTX), infliximab (INF), rituximab (RTX) and trastuzumab (TTZ) in mixtures. The RP mAbs separation was performed in a divinylbenzene-based monolithic column, after statistical design of the experiments with a novel approach for optimizing chromatographic conditions called the heteroscedasticity function. Results led us to split the initial mixture of five mAbs into two mixtures with four mAbs each, one containing RTX and the other TTZ. The method was validated for quantification using the signal from the UV detector and identification by (HESI-Orbitrap)MS. Direct MS characterization of the intact isoform profile of each mAb was also obtained. Advantages and disadvantages of the use of trifluoroacetic acid or formic acid as ion pairing agents for mass spectrometric analysis and chromatographic separation are discussed. Validation was performed using an internal protocol based on well-known international guidelines such as the International Conference on Harmonization (ICH) guideline, the US Food and Drugs Administration (FDA) guideline and the United Stated Pharmacopeia (USP) guideline. Performance parameters such as linearity, accuracy (precision and trueness), detection limits, quantification limits and robustness were evaluated. Robustness was established by studying the total and one-sided effects of four selected variables: column temperature, trifluoroacetic acid content in the mobile phases, initial proportion of eluent B and gradient. The results indicated the suitability of this method for quantifying these five mAbs in mixtures, as well as its robustness, reproducibility and sensitivity.

Introduction

Monoclonal antibodies (mAbs) are large glycoproteins with a molecular weight of around 150 kDa. They are currently the most important class of therapeutic proteins on the market and are used worldwide in the treatment of a broad spectrum of disorders, above all cancers [1,2]. Bevacimab (BVZ), cetuximab (CTX), infliximab (INF), rituximab (RTX) and trastuzumab (TTZ) are among the most prescribed therapeutic mAbs as they are indicated for the treatment of widespread disorders such as cancer and autoimmune diseases. They were the top-selling biopharmaceuticals in 2015 and 2016 [3,4]. BVZ (Avastin®) is indicated for the treatment of several kinds of cancer, i.e. metastatic carcinoma of the colon or rectum, breast cancer, lung cancer, etc [5]. CTX (Erbitux®) is approved for the treatment of colon, head and neck cancers [6]. INF (Remicade®) is used in the treatment of psoriasis, Crohn's disease, ankylosing spondylitis, psoriatic arthritis, rheumatoid arthritis and ulcerative colitis [7]. RTX (MabTera®) is indicated for use in the treatment of non-Hodgkin's lymphoma, rheumatoid polyarthritis and chronic lymphoid leukaemia [8]. TTZ (Herceptin®) is indicated for the treatment of patients with metastatic breast cancer with tumours that overexpress HER2 (25% of patients) [9]. All of the mAbs share the same IgG1 structure (Fig. 1), differing only in the part involved in specific interaction with the antigen (Fab). This structural similarity makes reversed phase (RP) chromatographic separation using packed bed-columns difficult [[10], [11], [12]]. The use of monolithic column could benefit separation of mAbs due to the expected more favoured mass transfer kinetics because of the lack of mesoporores in polymer monoliths [13,14].

Reversed-phase liquid chromatography ((RP)LC) is generally considered as more efficient and more sensitive for the analysis of intact biotherapeutic proteins than other types of chromatography (i.e. size exclusion or ionic exchange chromatography) [15,16]. However, this mode of chromatography has frequently been associated with low recovery due to the absorption of the proteins by the column stationary phase material and due to the low diffusion capacity towards the mobile phase, an issue that is even more complicated in large proteins such as mAbs [21]. To solve these drawbacks, high-performance reversed-phase liquid chromatography ((RP)HPLC) requires mobile phases composed of complex mixtures of organic solvents [10,17,18] combined with high temperatures in the column (up to 90 °C [16,19]) to increase the diffusion of the proteins into the mobile phase, so increasing the efficiency of the chromatographic process. Gradient elution mode must always be used. In conventional analytical chromatography, analysis is usually performed on a column packed with 3–5 μm wide-pore (300 Å) silica particles with a pressure of up to 400 bar. Nevertheless, the chromatographic separation of the mixtures of intact mAbs such as those described above has never been achieved using this approach. This could be due to the high degree of structural similarity (sharing the IgG structure), which promotes similar hydrophobicity features that prevent the separation of therapeutic mAbs by (RP)HPLC.

Ultra-high-performance liquid chromatography in reversed-phase mode ((RP)UHPLC) uses sub-2 μm particle sizes and therefore needs higher pressures in the system (around 1000 bar). This significantly enhances the efficiency of the separation of intact proteins using common organic solvent composition in the mobile phase, such as mixtures of water and acetonitrile and common ion pairing agents, such as trifluoroacetic acid (TFA) or formic acid (FA). Column temperatures of up to 80 °C are also required [20,21], as are gradient elution modes. These conditions have the advantage of being suitable to analyse intact protein when coupled with a mass spectrometer. Rigid polymer-based monolithic columns can also be used as an alternative to packed-bed columns [13,14]. In terms of protein recovery and carryover, monolithic columns have been shown to be clearly superior to particle based stationary phase formats for analysing mixtures of intact protein in the 5.7–150 kDa size range (including recombinant mAbs). This method is also very sensitive, detecting at femtomol level [22]. This means that, as indicated in [23], organic polymer-based monolithic columns are suitable for both high-resolution and high-speed biomolecule separations. High peak capacity peptide and protein separations have been reported which were at least as good, if not better, as those achieved using packed columns. An additional advantage is that the unique porous structure and chemistry of the organic polymer-based monolithic columns decreases the carryover [24].

Notwithstanding the above, the analysis of therapeutic mAbs by (RP)LC has tended to focus on the analysis of a single mAb. Normally this is done by analysing intact mAbs or by applying enzymatic strategies to analyse Fc and Fab fragments [14,16]. So far little research has been done into mixtures of mAbs. This could be due to the fact that they are mainly formulated and prescribed individually. Nevertheless, mAbs could theoretically work well in combination therapy because of their limited overlapping toxicity and the lack of pharmacokinetic interactions [25,26]. It has already been reported that a combination of mAbs can have a greater therapeutic potential for treating cancer than a single mAb [27,28]. By combining mAbs in the treatment of a particular cancer, several pathways may be targeted simultaneously, potentially creating additive or synergistic effects; preclinical and clinical studies of twenty-five different combinations of therapeutic mAbs have been reported [25]. In addition, mixtures of mAbs are using in on-going clinical trials to overcome acquired resistance to approve mAbs treatment [29,30] or to overcome resistance developed during clinical trials to new specific mAbs therapeutic agents [31]. For all this, mAb combination is therefore likely to become a promising cancer therapy in the near future. MAb mixtures can be produced either by individually manufacturing the constituent mAbs or by producing a single batch of two or more mAbs [32].

In this context, it is important to provide rigorous and reliable analytical methods that enable us to study mixtures of mAbs, and not only the individual mAb constituents [33]. It is evident that Size-Exclusion Chromatography (SEC) is not suitable for analysing intact mAb mixtures since all the mAbs have similar molecular weights of around 150 kDa [34] although it represents an excellent strategy for the analysis of individuals mAbs (and related formats) when coupling to native ion mobility mass spectrometry [35]. Cation Exchange Chromatography (CEX) has proved to be a fast, reliable method for the characterization of charge variants [36,37], also be used for the analysis of mixtures [33,38]. Very recently has been proposed the development of a universal method for charge variant characterization of single mAbs based on the use of pH gradient elution using volatile, low ionic strength buffers with direct coupling to high-resolution Orbitrap mass spectrometry [39]. Also, hydrophobic interaction chromatography (HIC) is a technique which allows separate mAbs on the basis of their hydrophobicity under native conditions [40,41]. Capillary zone electrophoresis (CZE) is another option to analyse mAbs mixtures also resolving charge variants simultaneously [42,43]. Nevertheless, as mentioned earlier, (RP)LC has greater resolving power and efficiency [44] and sharper peaks [10,11], which can be used for quantification purposes in single mAbs or in mixture solutions. (RP)LC has the additional advantage of straightforward coupling to several ionization methods, such as heat electrospray ionization (HESI), for mass spectrometry (MS) detection. Native mass spectrometry on an Orbitrap platform was recently used to characterize the composition of complex mAbs mixtures generated from single production to be used for batch to batch assessment [33]; the analysis focused on the intact mAbs without previous LC separation.

Our previous research has focused on the development and validation of (RP)HPLC/DAD methods for the rigorous analytical quantification of single therapeutic mAbs using a wide porous (300 Å) 5 μm particle size packed column [[10], [11], [12]]. We also tried to analyse mAb mixtures, but this proved impossible using this form of analytical chromatography.

In this paper, we present the development and validation of an (RP)UHPLC-UV-(HESI/Orbitrap)MS method for the quantification of mixtures of five commercially available, therapeutic mAbs (i.e. BVZ, CTX, INF, RTX and TTZ). A monolithic column based on a poly(divinylbenzene-co-ethylvinylbenzene) co-polymer was selected to perform mAb separation. The experiment was statistically designed (DoE) to optimize the liquid chromatographic conditions. Direct MS characterization of the intact isoform profile of each mAb was achieved and is described below. We also demonstrated the compatibility of the use of TFA or FA as ion pairing agents for mass spectrometric analysis and for chromatographic separation. Validation was performed according to an internal protocol based on well-known international guidelines. We also propose a new approach for obtaining a model, which we call the heteroscedasticity function, which describes the dependence of the experimental measurement standard deviation vs the mAb concentration.

Section snippets

Chemicals, reagents and mAb solutions

Reverse-osmosis-quality water was purified with a Milli-RO Milli-Q station from Merck Millipore (Darmstadt, Germany). The reagents used were LC–MS purity grade. Acetonitrile (ACN) was supplied by VWR International Eurolab, S.L. (Barcelona, Spain). Trifluoroacetic acid (TFA) was obtained from Scharlab S.L. (Barcelona, Spain). Lysozyme from chicken egg white was purchased from Sigma Aldrich (Madrid, Spain). Isotonic solution (0.9% NaCl in water for injection) was supplied by B. Braun Medical

(RP)UHPLC-UV method optimization

The experimental operational conditions of the chromatographic method were optimized by applying the methodology of statistical design of experiments (DoE). The signal from the UV detector was used for optimisation purposes as it is more sensitive than the mass detector. The latter detector was later used for isoform mAbs profile/ identification.

Before applying DoE, two ion-pairing agents (TFA and FA) were checked in order to improve the mass spectral signal. It is well known that TFA

Conclusions

This paper demonstrates that the chromatographic separation of intact commercially available therapeutic mAbs can be achieved using (RP)LC with a divinylbenzene-based monolithic column. Separation can be achieved despite the similarity in their structure –typically IgG1– which results in similar hydrophobicity features. To the best of our knowledge this is the first method to analyse mixtures of therapeutic intact mAbs focused on quantification. Up to now, (RP)LC has been used for the analysis

Acknowledgments

This work was partially funded by Projects FIS-PI10/00201 and FIS-PI17/00547 (Instituto Carlos III, Ministerio de Economía y Competitividad, Spain); therefore this project has also been partially supported by European Regional Development Funds (ERDF). The authors would like to thank the Hospital Pharmacy Unit of the University Hospital San Cecilio for kindly supplying all the medicine samples for this research.

References (54)

  • S.K. Rasmussen et al.

    Recombinant antibody mixtures: production strategies and cost considerations

    Arch. Biochem. Biophys.

    (2012)
  • A. Goyon et al.

    Evaluation of size exclusion chromatography columns packed with sub-3 μm particles for the analysis of biopharmaceutical proteins

    J. Chromatogr. A

    (2017)
  • A. Ehkirch et al.

    Hyphenation of size exclusion chromatography to native ion mobility mass spectrometry for the analytical characterization of therapeutic antibodies and related products

    J. Chromatogr. B Anal. Technol. Biomed. Life Sci.

    (2018)
  • E. Wagner-Rousset et al.

    Development of a fast workflow to screen the charge variants of therapeutic antibodies

    J. Chromatogr. A

    (2017)
  • L. Zhang et al.

    Improving pH gradient cation-exchange chromatography of monoclonal antibodies by controlling ionic strength

    J. Chromatogr. A

    (2013)
  • S. Fekete et al.

    Method development for the separation of monoclonal antibody charge variants in cation exchange chromatography, part II: PH gradient approach

    J. Pharm. Biomed. Anal.

    (2015)
  • A. Cusumano et al.

    Practical method development for the separation of monoclonal antibodies and antibody-drug-conjugate species in hydrophobic interaction chromatoraphy, part 2: optimization of the phase system

    J. Pharm. Biomed. Anal.

    (2016)
  • M. Rodriguez-Aller et al.

    Practical method development for the separation of monoclonal antibodies and antibody-drug-conjugate species in hydrophobic interaction chromatography, part 1: optimization of the mobile phase

    J. Pharm. Biomed. Anal.

    (2016)
  • S. Fekete et al.

    Ultra-high-performance liquid chromatography for the characterization of therapeutic proteins

    TrAC-Trends Anal. Chem.

    (2014)
  • L.M. Henricks et al.

    The use of combinations of monoclonal antibodies in clinical oncology

    Cancer Treat. Rev.

    (2015)
  • K. Chen et al.

    Simple NMR methods for evaluating higher order structures of monoclonal antibody therapeutics with quinary structure

    J. Pharm. Biomed. Anal.

    (2016)
  • J.M. Yang et al.

    Investigation of the correlation between charge and glycosylation of IgG1 variants by liquid chromatography-mass spectrometry

    Anal. Biochem.

    (2014)
  • P.V. Bondarenko et al.

    mass measurement and Top-down HPLC/MS analysis of intact monoclonal antibodies on a hybrid linear quadrupole ion trap-orbitrap mass spectrometer

    J. Am. Soc. Mass Spectrom.

    (2009)
  • H. Kaplon et al.

    Antibodies to watch in 2018

    MAbs

    (2018)
  • D.M. Ecker et al.

    The therapeutic monoclonal antibody market

    MAbs

    (2015)
  • C.W. Lindsley

    New 2016 data and statistics for global pharmaceutical products and projections through 2017

    ACS Chem. Neurosci.

    (2017)
  • EMA/302947/2017

    European Public Assessment Report (EPAR) Summary for Avastin, Annex I: Summary of Product Characteristics

    (2017)
  • Cited by (7)

    • Combined use of UV and MS data for ICH Stability-Indication Method: Quantification and isoforms identification of intact nivolumab

      2022, Microchemical Journal
      Citation Excerpt :

      UHPLC in reverse-phase (RP) needs higher pressures in the system (around 1000 bar) than HPLC. This significantly enhances the efficiency of the separation of intact proteins using common organic solvent composition in the mobile phase, such as mixtures of water and acetonitrile and common ion pairing agents, such as trifluoroacetic acid (TFA) or formic acid (FA) [27]. Column temperatures of up to 80 °C are also required [28,29], and a gradient must be applied during chromatographic analysis.

    • It is Never Too Late for a Cocktail - Development and Analytical Characterization of Fixed-dose Antibody Combinations

      2022, Journal of Pharmaceutical Sciences
      Citation Excerpt :

      The coupling of RP-HPLC to mass spectrometry (MS) is a well-established strategy to identify chemical changes in specific parts of the protein. For example, Perez-Robles et al. demonstrated the simultaneous identification and quantification of up to four different co-formulated mAbs by RP-HPLC coupled to MS.107 It is also possible to use peptide mapping and RP-HPLC-MS to identify the exact position of chemical changes occurring in co-formulated mAbs.42 In another example, Cao et al. developed a peptide mapping method to identify and quantify deamidation in the complementarity-determining regions of one specific mAb in a co-formulation.105

    • Freeze thaw and lyophilization induced alteration in mAb therapeutics: Trastuzumab as a case study

      2021, Journal of Pharmaceutical and Biomedical Analysis
      Citation Excerpt :

      A significant negative correlation was observed, likely because deamidation in the CDR regions results in various degrees of reduction in antigen binding affinity and potency. Deamidation in the CDR region has been known to result in various degrees of reduction in antigen binding affinity and potency of mAbs [34,35]. No significant correlation was observed between the main and basic variants for both F/T and Lyo studies.

    • Understanding the discrimination and quantification of monoclonal antibodies preparations using Raman spectroscopy

      2021, Journal of Pharmaceutical and Biomedical Analysis
      Citation Excerpt :

      Monoclonal antibodies (mAbs) are biopharmaceutics produced by bioengineering technology; they are used as targeted therapy in cancer treatment and are associated with fewer and less severe side effects compared to other chemotherapeutics [1]. In addition to their valuable role in cancer therapy, they are used in preventing or modulating viral diseases [2], for multiple sclerosis [3] and some autoimmune diseases [4]. mAbs are glycoproteins of molecular weight ∼150 KDa; their structure consists of two heavy chains, each of which contains 447 amino acids linked by 4 disulphide bonds, and two light chains, each containing 215 amino acids, linked by 2 disulphide bonds [5].

    View all citing articles on Scopus
    View full text