Elsevier

Metabolic Engineering

Volume 60, July 2020, Pages 138-147
Metabolic Engineering

Enzyme capacity-based genome scale modelling of CHO cells

https://doi.org/10.1016/j.ymben.2020.04.005Get rights and content

Highlights

  • Developed a comprehensive CHO genome scale metabolic model with kinetome parameters.

  • Incorporated enzyme capacity constraints within FBA for accurate flux prediction.

  • Deciphered the clone- and media-specific lactate overflow metabolism in CHO cells.

  • Provided a profound platform to accelerate the development of bioreactor Digital Twins.

Abstract

Chinese hamster ovary (CHO) cells are most prevalently used for producing recombinant therapeutics in biomanufacturing. Recently, more rational and systems approaches have been increasingly exploited to identify key metabolic bottlenecks and engineering targets for cell line engineering and process development based on the CHO genome-scale metabolic model which mechanistically characterizes cell culture behaviours. However, it is still challenging to quantify plausible intracellular fluxes and discern metabolic pathway usages considering various clonal traits and bioprocessing conditions. Thus, we newly incorporated enzyme kinetic information into the updated CHO genome-scale model (iCHO2291) and added enzyme capacity constraints within the flux balance analysis framework (ecFBA) to significantly reduce the flux variability in biologically meaningful manner, as such improving the accuracy of intracellular flux prediction. Interestingly, ecFBA could capture the overflow metabolism under the glucose excess condition where the usage of oxidative phosphorylation is limited by the enzyme capacity. In addition, its applicability was successfully demonstrated via a case study where the clone- and media-specific lactate metabolism was deciphered, suggesting that the lactate-pyruvate cycling could be beneficial for CHO cells to efficiently utilize the mitochondrial redox capacity. In summary, iCHO2296 with ecFBA can be used to confidently elucidate cell cultures and effectively identify key engineering targets, thus guiding bioprocess optimization and cell engineering efforts as a part of digital twin model for advanced biomanufacturing in future.

Introduction

Chinese hamster Ovary (CHO) cells are the most prevalent mammalian hosts for manufacturing therapeutic proteins, accounting for more than 80% of all monoclonal antibodies in the market (Walsh, 2018). Such widespread use of CHO cells is mainly due to the capability for post-translational modifications (e.g. N-glycosylation), ease of genetic manipulation, and the potential to thrive in suspension cultures. Although the production titre of CHO cells has increased by up to a thousand times over the decades, their efficiency can be further enhanced at two different levels: 1) cell line development and engineering, and 2) bioprocess development by identifying critical process parameters (CPPs), designing cell culture media and feeding strategies for the enhanced key performance indicators (KPIs), e.g., titre and yield, of the cell culture (Kyriakopoulos et al., 2018; Hong et al., 2018a). However, industrially most of such advances to date have been achieved only through the application of empirical techniques, such as ‘design of experiments’ (DOE) method, and black-box statistical approaches which are mainly based on historical bioprocess data sets; similarly, a priori knowledge- or experience-based cell line development and engineering are also being undertaken in ad-hoc manner. Thus, it is now highly required to develop more rational approaches to better understand CHO cell physiology, thereby enhancing their KPIs (Kildegaard et al., 2013). In this regard, the recent availability of CHO cell line-specific genome sequence information facilitated the development of systematic framework whereby in silico mechanistic models can be combined with increasingly available multi-omics datasets in order to describe cell culture behaviours. Previously, a genome scale metabolic model (GEM) of mouse (Selvarasu et al., 2010) was adopted to identify limiting nutrients in the growth of CHO fed-batch culture (Selvarasu et al., 2012), and to compare metabolic differences associated with the switching of lactate production to uptake in CHO cells (Martínez et al., 2013). Then, the first-ever CHO GEM was reconstructed through world-wide community efforts (Hefzi et al., 2016). The availability of such CHO GEM has greatly aided the characterization of various CHO cell line specific traits (Hefzi et al., 2016; Lakshmanan et al., 2019; Pan et al., 2017; Yusufi et al., 2017) and several bioprocessing conditions (Calmels et al., 2019; Hong et al., 2020; Park et al., 2018; Vodopivec et al., 2019).

Various in silico approaches have been employed to describe CHO cell metabolism, which include kinetic modelling, metabolic flux analysis (MFA) and constraint-based flux balance analysis (FBA) (Galleguillos et al., 2017; Huang et al., 2017; Kyriakopoulos et al., 2018). Of them, constraint-based FBA has been widely applied for analysing cellular metabolism at genome-scale as it is amenable to large-scale model, requiring only reaction stoichiometry and the mass balance of metabolites. While the utility of FBA has been successfully demonstrated by understanding cellular behaviour of CHO cells via aforementioned studies on cell line characterization and process development, it still has a limitation in predicting intracellular fluxes and identifying potential engineering targets, due to a large degree of freedom arising from an extensive and highly-interconnected metabolic network, and often exacerbated by inadequate cell culture measurements as well as a limited knowledge on intrinsic constraints and cellular objectives in complex mammalian systems. To overcome such limitation, recently, parsimonious FBA (pFBA) has been used to describe metabolic states of CHO cells (Calmels et al., 2019); the overall enzymatic fluxes are minimized to eliminate the alternative flux solutions based on the hypothesis that cells use enzymes and carbon resources efficiently, under the pressure of natural selection (Lewis et al., 2010). While this method significantly reduces flux variability compared to basic FBA, it is still unable to replicate a few key metabolic traits and physiological behaviours observed in mammalian cell culture, such as the intrinsic restriction on growth capacity and by-product secretion. More recently, another method known as “carbon constraint FBA (ccFBA)” was proposed to improve the accuracy of flux predictions in CHO cells on the basis of carbon elemental balances (Lularevic et al., 2019). Similar to pFBA, this method was designed to reduce the intracellular flux variability. However, it simply eliminates the thermodynamically infeasible pathways without any further consideration for pathway usage and enzymatic regulatory constraints which have been argued often important for biologically relevant simulation. Therefore, it is essential to incorporate the reaction kinetic information within FBA framework for elucidating the metabolic states. Indeed, the inclusion of a physical capacity limit for metabolic enzymes has successfully reduced the potential solution space in a biologically-meaningful manner (Adadi et al., 2012; Beg et al., 2007; Sánchez et al., 2017; Shlomi et al., 2011). Hence, now we can add relevant kinetic information in CHO GEM to limit the overall metabolic enzyme usage, as such accurately identifying key bottlenecks under various cell culture conditions. By doing so, we could develop the most comprehensive and biochemically-consistent CHO GEM to-date, iCHO2291, from the previous model (Hefzi et al., 2016). Notably, this model includes the necessary kinetome parameters such as turnover number (kcat) and molecular weight of all enzymes, to enable its ready usage during the flux simulations. We then demonstrate the model utility and framework efficacy via a case study where the mechanisms underlying differential lactate overflow between two CHO clones are deciphered (Hong et al., 2018b).

Section snippets

Update of CHO genome-scale metabolic model

We significantly expanded the previous CHO genome-scale model, iCHO1766 (Hefzi et al., 2016) following a six step procedure (Materials and methods). Initially, we identified the metabolite and reaction inconsistency in iCHO1766. Here, it should be noted that iCHO1766 has been built from various human GEMs including Recon2 (Thiele et al., 2013) which was previously developed based on several resources such as Recon1 (Duarte et al., 2007) and EHMN (Ma et al., 2007). Thus, some metabolites and

Discussion

In this study, we have extensively updated the previous CHO GEM (Hefzi et al., 2016) by correcting the biochemical inconsistency and including new genes, reactions and pathways. We also incorporated kinetome parameters of enzymes such as their turnover numbers and molecular weights which are fully utilized as additional constraints within the FBA framework, i.e., ecFBA, so that intracellular flux variations can be effectively reduced to accurately characterize cell culture behaviours and

Update procedure for CHO genome-scale reconstruction

The iCHO2291 model was reconstructed by updating the previously published iCHO1766 model in a 6-step procedure: 1) identification and removal of duplicate metabolites and reactions, 2) replacement of lumped reactions into detailed steps and removal of biochemically inconsistent reactions, 3) update of GPR based on latest genome annotations (as on Nov 1, 2018), 4) correction of GPR and reaction compartment assignment based on subcellular localization, 5) inclusion of new reactions based on new

CRediT authorship contribution statement

Hock Chuan Yeo: Formal analysis, Writing - original draft, Methodology. Jongkwang Hong: Formal analysis, Writing - original draft. Meiyappan Lakshmanan: Writing - review & editing, Supervision, Methodology. Dong-Yup Lee: Writing - review & editing, Supervision.

Acknowledgements

This work was supported by Biomedical Research Council of A*STAR (Agency for Science, Technology and Research), Singapore, and the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. 2020R1A2C2007192). The authors also thank Kok Siong Ang for the assistance in collecting enzyme specific biochemical data and Lokanand Koduru for the useful discussions and technical support for implementing enzyme constrained flux balance analysis.

References (53)

  • R. Adadi et al.

    Prediction of microbial growth rate versus biomass yield by a metabolic network with kinetic parameters

    PLoS Comput. Biol.

    (2012)
  • Q.K. Beg et al.

    Intracellular crowding defines the mode and sequence of substrate uptake by Escherichia coli and constrains its metabolic activity

    Proc. Natl. Acad. Sci. U.S.A.

    (2007)
  • S. Borger et al.

    Prediction of enzyme kinetic parameters based on statistical learning

    Genome Inform

    (2006)
  • G.A. Brooks et al.

    Role of mitochondrial lactate dehydrogenase and lactate oxidation in the intracellular lactate shuttle

    Proc. Natl. Acad. Sci. U.S.A.

    (1999)
  • A. Chang et al.

    BRENDA in 2015: exciting developments in its 25th year of existence

    Nucleic Acids Res.

    (2015)
  • Y. Chen et al.

    Energy metabolism controls phenotypes by protein efficiency and allocation

    Proc. Natl. Acad. Sci. U.S.A.

    (2019)
  • Y.J. Chen et al.

    Lactate metabolism is associated with mammalian mitochondria

    Nat. Chem. Biol.

    (2016)
  • D. Davidi et al.

    Global characterization of in vivo enzyme catalytic rates and their correspondence to in vitro kcat measurements

    Proc. Natl. Acad. Sci. U.S.A.

    (2016)
  • N.C. Duarte et al.

    Global reconstruction of the human metabolic network based on genomic and bibliomic data

    Proc. Natl. Acad. Sci. U.S.A.

    (2007)
  • B. Faubert et al.

    Lactate metabolism in human lung tumors

    Cell

    (2017)
  • N.W. Freund et al.

    A simple method to reduce both lactic acid and ammonium production in industrial animal cell culture

    Int. J. Mol. Sci.

    (2018)
  • M. Glont et al.

    BioModels: expanding horizons to include more modelling approaches and formats

    Nucleic Acids Res.

    (2018)
  • D. Heckmann et al.

    Machine learning applied to enzyme turnover numbers reveals protein structural correlates and improves metabolic models

    Nat. Commun.

    (2018)
  • H. Hefzi et al.

    A consensus genome-scale reconstruction of Chinese hamster ovary cell metabolism

    Cell Syst

    (2016)
  • L. Heirendt et al.

    Creation and analysis of biochemical constraint-based models using the COBRA Toolbox v.3.0

    Nat. Protoc.

    (2019)
  • Z. Huang et al.

    Quantitative intracellular flux modeling and applications in biotherapeutic development and production using CHO cell cultures

    Biotechnol. Bioeng.

    (2017)
  • Cited by (41)

    View all citing articles on Scopus
    1

    Equally contributed to this work.

    View full text