Farràs et al. BMC Molecular and Cell Biology (2020) 21:2 https://doi.org/10.1186/s12860-019-0244-x BMC Molecular and Cell Biology RESEARCH ARTICLE Open Access Heavy chain dimers stabilized by disulfide bonds are required to promote in vitro assembly of trastuzumab Mercè Farràs1* , Ramón Román2, Marc Camps1, Joan Miret2, Óscar Martínez1, Xavier Pujol1, Antoni Casablancas2 and Jordi Joan Cairó2 Abstract Background: Monoclonal antibodies (mAbs) and their derivatives have become one of the most important classes of therapeutic drugs. Their multiple applications increased the interest for understanding their complex structure. In vivo, animal cells are able to fold mAbs correctly (Song et al, J Biosci Bioeng 110:135-40, 2010), whereas previous in vitro approaches were scarce and mostly unsuccessful. Results: In this work, we compared in vitro assembly characteristics of trastuzumab, produced either by A) physical separation and refolding of its sub-units or B) direct joining of individually produced heavy and light chains. Native and denatured structures of trastuzumab were determined by SEC-HPLC, HIC-HPLC and SDS-PAGE. Conclusions: Our results demonstrate the requirement of correctly folded HC, forming disulfide-bonded dimers, in order to form a fully functional mAb. Otherwise, the unfolded HC tend to precipitate. We were able to assemble trastuzumab in this fashion by only mixing them to LC in pH-buffered conditions, while monomeric HC structure was too unstable to render a functional mAb. This approach has been used in the generation of homogeneous ADC, with results pending to be published. Keywords: mAb, Protein structure, Folding, Disulfide bonds, Renaturalization, LC, HC, mAb assembly, Urea, Anti- HER2, Non-covalent, Slow dialysis, Glutathione, 2-mercaptoethanol, Affinity chromatography, Trastuzumab, Immunoglobulin Background Mabs in vivo structure Therapeutic antibodies have risen to prominence over The complex structure of mAbs lead to further efforts in the past three decades and are the fastest growing drug order to understand their folding, denaturation and class, with currently more than 70 monoclonal anti- refolding [9, 10]. In vivo, heavy chains (HCs) and light bodies (mAbs) approved by FDA and EMA [1–3]. In chains (LCs) are co-translationally translocated into the view of the side effects and limitations of mAbs [4] sev- endoplasmic reticulum (ER) and folding begins even be- eral improvements and modifications such as conjugated fore the polypeptide chains are completely translated. mAbs fragments or bispecific antibodies [5, 6] have been Most immunoglobulins G (IgGs) assemble first as HC developed [7]. Some of its applications include the treat- dimers to which LCs are added covalently via a disulfide ment of infectious and non-infectious diseases such as bond between the CL and CH1 domains [11]. The Ig cancer, immune diseases, arthritis and other disorders fold is characterized by a greek-key β-barrel topology in resulting from organ transplantation [8]. which the barrel is not continuously hydrogen bonded, but instead composed of two sheets, forming a sandwich-like structure [11]. * Correspondence: mfcosta@fhrt.com 1 Department of Biotechnology, Farmhispania SA, Montmeló, Spain Full list of author information is available at the end of the article © The Author(s). 2020 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Farràs et al. BMC Molecular and Cell Biology (2020) 21:2 Page 2 of 9 mAbs denaturation and refolding Results Pioneering studies on antibody folding were executed on Unfolding and folding a mAb without physical separation denatured LCs, allowing in vitro refolding [12]. Cold de- Firstly, we adapted the Maeda et al. method [20] based on naturation studies using guanidine hydrochloride were slow dialysis and appropriate redox buffer in order to test performed and revealed that mAbs have a potential to the ability of the method to refold and reoxidize denatured undergo cold denaturation at storage temperatures near − and reduced trastuzumab, without physical chain separation. 20 °C (pH 6.3), and this potential needs to be evaluated in- Results obtained are shown in Fig. 1, where complete dependently for individual mAbs [13]. This previous data reduction and denaturation of anti-HER2 is achieved in shows that it is possible to refold a reversible denatured the conditions discussed (and checked by SDS-PAGE mAb without breaking disulfide bonds. and SEC-HPLC). After slow dialysis, the antibody is able The structure of antibodies and many antibody frag- to recover its disulfide bonds, showing the same profile ments is composed of multiple domains, often connected as the initial mAb in SEC-HPLC (results not shown). by inter-domain disulfide bonds, which further compli- Refolded trastuzumab effectively recognizes isolated cates the already difficult refolding processes. HER2 antigen in an ELISA test in the same levels as un- Refolding of the denatured and reduced immunoglobu- treated control (Table 1) and binds to protein A affinity lins is scarcely successful in vitro [14]. Fab fragments con- column (Fig. 1), proving that the fragment crystallizable sisting of two subunits have been vigorously employed as region (Fc) is also correctly folded. Denatured and re- model systems for studying the mechanisms of protein duced mAb showed no antigen recognition in the ELISA folding [15, 16], and many refolding technologies have test (Table 1). been used for these fragments, comprising dilution, dialy- sis, solid phase solvent exchange and size exclusion chro- Unfolding and folding anti-HER with chains physical matography [16], but several complications are observed separation during refolding, which have been overcome. The folding Anti-HER2 interchain disulfide bonds were totally re- yield of the denatured and reduced Fab fragment was low duced and mAbs chains were physically separated and by spontaneous renaturation, but in the presence of a purified by denaturing SEC (Fig. 2a) as described in GroE system (GroEL, GroES and ATP) or protein disul- methods section (Fig. 8). Protein identity of each peak phide isomerase (PDI), the folding yield of the denatured was validated through an SDS-PAGE gel (Fig. 2b) and and reduced Fab fragment was higher than that of spon- confirmed the selective chains separation, with little or taneous renaturation [17]. A similar approach using im- none cross-contamination. munoglobulin heavy chain binding protein (BiP) and PDI After a 5 times volume reduction with a 3KDa Amicon was successful in a complete mAb [18]. device, slow dialysis was performed as described before for HC and LC. In vivo and in vitro LC and HC (re)folding LC structural integrity was first checked by Capto L af- Studies regarding slow dialysis without the assistance finity chromatography and no protein could be detected of chaperone were performed to renature a denatured in the fraction corresponding to flow-through, showing and reduced IgG at a concentration of 1 mg/ml [19] that reassembled LC could interact with protein L. The with a 70% of folding yield. yield from denatured and reduced LC to the folded LC In this work, we based our In vitro refolding strategy in this was 26% (Fig. 3a). The same approach was followed with slow dialysis method but adding a physical separation step by HC by using protein A affinity chromatography to check size exclusion chromatography under denaturing conditions. proper binding. In this case, the fraction corresponding The main challenge was the physical chains separation and to the elution peak showed major aggregation in the their reversible refolding due to the mAb complex structure, non-reduced SDS-PAGE gel (Fig. 3b). formed by covalent (disulfide) and non-covalent (ionic, hydro- LC and HC were independently buffer-exchanged using gen bonds, Van der Waals, hydrophobic) interactions to main- PD Desalting G-25 column to exchange the elution buffer tain the correct conformation, which is essential to revert the for 50 mM citrate pH 6, in order to recover the original original mAb structure and functionality. mAb structure. Under these pH conditions, HC precipi- We studied the feasibility to unfold, physically separate tated almost completely and the antigen recognition of mAb chains, in vitro correctly refold them and then re- the renatured trastuzumab is reduced to the half com- assemble the original anti-HER2. This assembly ap- pared to the reference mAb (Table 1), indicating that the proach is compared with the direct reassembly of the mAb structure was not recovered completely. mAb using in vivo folded chains (independently pro- duced in HEK293 cultures). The differences between In vitro and in vivo LC folding structure comparison in vivo and in vitro folded chains are analyzed, as well as LC refolded by slow dialysis (in vitro refolding) under de- the impact on mAb assembly efficiency. naturing, non-reducing conditions (Fig. 4c LC A) shows a Farràs et al. BMC Molecular and Cell Biology (2020) 21:2 Page 3 of 9 Table 1 Isolated antigen HER2 recognition in the ELISA test to assess the mAb folding without chains physical separation Molecule Trastuzumab Denatured Renatured trastuzumab trastuzumab Isolated antigen HER2 51 ± 5 AU/ ND 22 ± 1 AU/μg recognition μg mAb refolding of independently produced LC and HC In vivo folded LC and HC were dialyzed with 50 mM cit- rate pH 6 and mixed with a proportion of 5.54 g dcHC / g LC. Under these conditions, it is possible to recover a fully functional mAb structure without the need of slow dialysis or additional chaperones. The final mAb effectively recognizes isolated HER2 antigen in an ELISA test in the same levels as original anti-HER2, while in vivo folded LC or HC alone do not. The assembled mAb forms a monomer through non-covalent interactions, but it was not stabilized through covalent disulfide bonds because under denaturing conditions only HC di- Fig. 1 SDS-PAGE of reduced and refolded anti-HER2. M: molecular mers (100 KDa), LC dimers (42 KDa) and LC weight marker; i: intact mAb; r/dn: reduced and denatured mAb; dia: monomers (21 KDa) are detected (Fig. 6b), probably mAb dialyzed by slow dialysis; FT: flow-though of the affinity because interchain thiols are capped with cellular chromatography MAb Select SURE; peak: elution peak of the affinity chromatography glutathione and/or cysteine originated during the cell culture [21]. However, mAb activity has com- pletely been recovered, as analyzed in the ELISA single 21 KDa band, corresponding to a monomer struc- test (Table 2). The HC dimers, LC and refolded ture. However, under native conditions (Fig. 4a), a molecule mAb identity was confirmed by LC-MS (data not of 42 KDa is detected, indicating that LC is folded forming shown). dimers without covalent interactions between them through disulfide bonds. Discussion LC produced by HEK293 (in vivo folding), however, ap- In this work, we have observed that a denatured and re- pears in two forms under denaturing conditions: 55% of 42 duced mAb analysed by SDS-PAGE and HPLC could be KDa band and 45% of 21 KDa band (Fig. 5c LC B). Under satisfactory refolded in vitro if HC and LC were not phys- native conditions, 100% of LC is detected as dimers (Fig. 4a). ically separated. However, if the same protocol was ap- Hydrophobic differences were also detected: mean- plied with physical separation of HC and LC followed by while in vitro refolded LC shows a single peak in HIC- the refolding of each chain independently, HC showed HPLC, in vivo folded LC appears as 2 peaks (Fig. 4b). precipitation and, therefore, the original mAb structure This result could be related to hydrophobicity differ- could not be recovered. In vivo, HC is stabilized through ences between covalently and non-covalently bonded di- disulfide bonds, to which LC are added [7, 11]. Here we mers, due to the structural differences among them. hypothesize, that in vitro HC needs LC interaction to gen- LC is not able to recognize HER2 antigen alone (Table 2). erate dimers and if it appears as a monomer form, HC is so unstable that smooth pH changes promote its precipi- tation. It is, therefore, a requisite to have a stable HC In vitro and in vivo HC folding structure comparison dimer, even if it is only stabilized by non-covalent interac- In vitro refolded HC is mostly appearing as a monomer, tions [22] as in the case of DAR 7–8 Antibody Drug Con- with little dimers forming disulfide bonds, but covalent jugates, to promote mAb assembly. aggregation is also observed in SDS-PAGE gel (Fig. 5 These observations are supported by results obtained HC A). when HC and LC were independently produced in However, in vivo folded HC is produced forming di- HEK293 cultures (in vivo). HC was produced as a cova- mers covalently linked by disulfide bonds (Fig. 5 HC B). lently bonded dimer and their mixture to dimer LC When in vitro refolded HC was dialyzed to pH 6, se- allowed obtaining a functional mAb conformation. In vere precipitation is observed. In vivo folded HC, under this approach, it has been proved that no HC or LC re- the same conditions, remained soluble. duction is required to remove glutathione from the Farràs et al. BMC Molecular and Cell Biology (2020) 21:2 Page 4 of 9 Fig. 2 mAb chains separation with exclusion size chromatography in ÄKTA Avant system under reducing and denaturing conditions. a. UNICORN Software chromatogram, where absorbance at 280 nm is shaded in blue and absorbance at 260 nm is shaded in red. b. SDS-PAGE gel corresponding with the isolated peaks collected in the chromatography capped thiol groups because interactions between LC affect the refolding kinetics. Future work will study the and HC are non-covalent. We hypothesize that the anti- optimization of the mAb refolding process. body monomer form is biologically more favorable than LC dimers, which are so unstable that the reaction is displaced to bond LC to HC instead of maintaining LC- Methods LC (even if they are covalently linked). The reduction Cell lines and maintenance and reoxidation of mAbs disulfide bonds have been Anti-HER2 antibody sequence was obtained by syn- studied in detail for manufacturing processes [23, 24] thetic nucleotide synthesis (GenScript) and cloned in and it has been demonstrated that after partially redu- a tricistronic expression vector derived from the cing the mAb, the original disulfide bonds were re- commercial pIRESpuro3 vector (Clontech) between formed because the re-oxidized state is favored [25]. NheI-AgeI (LC) BamHI-EcoRI (HC). However, the current work indicates that the antibody HEK293 cells were kindly provided by Dr. A. Kamen form is more stable than LC dimers even if the mAb (National Research Council of Canada). HEK293SF-F6 structure is not stabilized by covalent interactions. cells were then transfected with the construct and selected as described previously [26]. The new anti-HER2 antibody Conclusions producer cell lines were referred to as HEK293_T10. In this work we revealed that HC dimers are imperative for HEK293SF-3F6 producing anti-HER2 HC or anti- LC assembly in order to obtain a functional mAb structure HER2 LC were obtained by cloning the synthetic se- in vitro. If the individual subunits are not correctly or par- quences in the expression vector pIRESpuro3 (Clontech) tially folded before mAb reassembly, the functional mAb between BamHI and NotI restriction sites for HC and in cannot be obtained (Fig. 7). Also, the proportion of LC and the vector pIRESneo3 (Clontech) between NheI and HC only affects the yield of the refolding process if a subunit AgeI sites for LC. HEK293 cells were then transfected is added in excess while the protein concentration could with the different constructs and selected as described Farràs et al. BMC Molecular and Cell Biology (2020) 21:2 Page 5 of 9 Fig. 3 Gel SDS-PAGE of refolding sequence of LC and HC. a. LC refolding process. M: molecular marker; Sx: Superdex peak of denatured and reduced LC; Conc: concentrated; Dia: LC diafiltered by slow dialysis; FT: flow-through of Capto L; Peak: Capto L elution peak. b. HC refolding process. M: molecular marker; i: Superdex peak of denatured and reduced HC; dia: HC diafiltered by slow dialysis; FT: flow-through of MAb Select SURE; peak: MAb Select SURE elution peak Fig. 4 LC in vitro and in vivo structure comparison. a. SEC-HPLC of in vitro refolded LC (black) and in vivo folded LC (blue). b. HIC-HPLC of in vitro refolded LC (black) and in vivo folded LC (blue). c. Gel SDS-PAGE of in vitro refolded LC (black) and in vivo folded LC (blue) Farràs et al. BMC Molecular and Cell Biology (2020) 21:2 Page 6 of 9 Table 2 Isolated antigen HER2 recognition in the ELISA test to assess the mAb folding with chains physical separation Molecule In vivo In vivo Refolded folded LC folded HC trastuzumab Isolated antigen HER2 ND ND 54 ± 6 AU/μg recognition WAVE 20/50 EHT (GE) with a working volume of 5 L. Temperature was set at 37 °C. Agitation and angle were set at 22 rpm and 8°, respectively. Purification of products was performed after 6 days of culture starting with a clarification step by depth filtra- tion (Clarisolve 40MS, Merck) followed by bioburden re- duction filter (Millipore Express SHF, Merck). Clarified broth was concentrated by tangential flow filtration (Hydrosart 30 KDa membrane, Sartorius) and then sub- jected to protein A affinity chromatography (MAb Select Sure, GE Healthcare) for complete mAb and HC purifi- cation or protein L for LC purification (Capto L, GE Healthcare), as described by the supplier. Purified frac- tions were buffer-exchanged through PD G-25 desalting columns, (GE). The final buffer was PBS for anti-HER2 and 50 mM citrate pH 6 for LC and HC. MAb denaturation and reduction Fig. 5 HC in vitro and in vivo structure comparison. Gel SDS-PAGE In order to obtain reduced and denatured mAbs chains, of in vitro refolded HC (HC A) and in vivo folded HC (HC B) affinity purified trastuzumab at 5.75 mg/mL was buffer ex- changed with a PD Desalting G-25 columns equilibrated with 0,584 M Tris-HCl (Roche, 10,708,976,001), and 5.37 previously [27]. The new producer cell lines were referred mM EDTA (E6758, Sigma), pH 8.6 and then concentrated to as HEK293_THC and HEK293_TLC. 5 fold using 30KDa Amicon tubes (Merck). Afterwards, Cells were subcultured at 0.3 × 106 cell/mL three times 70 mM 2-mercaptoethanol (Bio-rad, 161–0710) and 8 M per week to keep them in exponential growing phase. urea (Panreac, 141,754.1211) were added as described be- Cell maintenance was performed in 125 mL polycarbon- fore [15]. The mixture was incubated for 1 h at 40 °C prior ate shake flasks (Corning), with a working volume of 12 to the size-exclusion chromatography step. mL, and maintained at 37 °C in an incubator with a 5% CO2 humidity saturated atmosphere (Steri-cult 2000 In- cubator, Forma Scientific). Flasks were continuously agi- MAb chains physical separation tated at 110 rpm on an orbital shaking platform (Stuart Trastuzumab chains were physically separated and puri- SSL110). fied by preparative SEC using an ÄKTA AVANT 150 system (GE) equipped with a 500 mL Superdex 200 XK 26/100 column (GE). The column was equilibrated with Anti-HER2, HC and LC production 0.584 M Tris-HCl, 5.37 mM EDTA, 7 mM 2- Culture media used for HEK293_T10 cell line was mercaptoethanol and 8 M urea at a flow rate of 3 mL/ SFM4Transfx-293 (HyClone, SH30382.00) supple- min. The column was kept at 4 °C in order to prevent mented with 4 mM GlutaMAX (Gibco, 35,050,061), cyanate formation due to the presence of urea. Peaks 10% (v/v) Cell Boost 5 (60 g/L solution) (HyClone, corresponding to trastuzumab HC or LC were separated SH30865.01), 2% (v/v) Kolliphor P188 (100 g/L solu- and collected using the automated fraction collector. LC tion) (Sigma, 15,759) and 0.5% (v/v) Antifoam C (10 and HC structural integrity was assessed by performing g/L solution) (Sigma, A6832). Selection pressure was a run with affinity chromatography, using HiScreen maintained during the production step by adding Capto L column (GE) for LC (binding buffer: sodium 0.2% (v/v) puromycin (1 mg/mL) (Merck, 58,582). citrate 50 mM (Sigma, S1804), pH 6.5; elution buffer: so- Bioreactor cell cultures were performed using dium citrate 50 mM, pH = 2.3) and HiScreen Mab Select Flexsafe RM 10 L bags (Sartorius, DFB010L) in a Sure (GE) for HC (binding buffer: 20 mM sodium Farràs et al. BMC Molecular and Cell Biology (2020) 21:2 Page 7 of 9 Fig. 6 mAb refolded from independently produced LC and HC. a SEC-HPLC (native structure). The identity of the peaks was checked by LC-MS (data not shown). b Gel SDS-PAGE (denatured structure) Fig. 7 Graphical conclusions. Antibody folding is a complex process which occurs in vivo and where light chains are coupled to previously assembled heavy chain dimers. In this work, denaturation and chains separation of the Trastuzumab antibody followed by their in vitro refolding through slow dialysis method has shown that heavy chain dimers stabilized by disulphide bridges are necessary in order to reassemble the whole antibody. Successful in vitro assembly of heavy and light chains has been achieved when the chains have been independently produced Farràs et al. BMC Molecular and Cell Biology (2020) 21:2 Page 8 of 9 Fig. 8 Slow dialysis scheme. The feed bottle was set in vacuum so the dialysis buffer feeds it at the same speed as the pump 1 feeds the reservoir. Pump 1 and pump 2 speed is 0,25 mL/min phosphate, 150 mM NaCl, pH 7.4; elution buffer: sodium Sepax). The mobile phase was PBS (D1408, Sigma, di- citrate 50 mM, pH 3.2). luted in ultrapure water up to 1x) at a flow rate of 0.35 ml/min. The UV absorbance was measured at a wave- MAb or chains refolding by slow dialysis length of 214 nm. Dialysis buffer: 3.6 mM 2-mercaptoethanol, 1.3 mM re- duced glutathione (Sigma, G4251) and 1 mM EDTA in HIC-HPLC 0.1 M Tris-HCl, pH 8. HIC experiments were performed using a Waters Alli- Feed initial composition: same as dialysis buffer adding ance 2695 system using Proteomix HIC Butyl-NP5 8 M urea. (4.6 × 50 mm, Sepax). The mobile phase was a gradient Slow dialysis was performed by using Slide-A-Lyzer. of 25 mM sodium phosphate pH 7 (S5136 and S5011, Dialysis 3.5KDa (ThermoFisherScientific) device sub- Sigma) with decreasing ammonium sulphate (2170, merged in a 50 mL reservoir. The device contained the Merck) concentration (1.8 M to 0 M) and increasing test sample plus 1 mL 8 M urea and 13.6 mg glutathione concentration of isopropanol (AL03151000, Scharlab) to (Fig. 8), whereas the reservoir was filled with dialysis buf- improve peaks resolution, as recommended by the col- fer and 8 M urea. Stirred feed bottle contained a constant umn manufacturer, at a flow rate of 0.8 ml/min. The UV volume of 50 mL of dialysis buffer, being constantly fed absorbance was measured at a wavelength of 214 nm. from dialysis bottle containing 400 mL of dialysis buffer. A 0.25 mL/min constant flow rate was maintained dur- ELISA assay ing a 30 h process using a Schott Iberica peristaltic pump Antibody assembly was evaluated by comparing in vitro with 2 heads, during which the reservoir volume is main- isolated anti-HER2 recognition with its in vivo assem- tained constant. Fig. 8 shows a scheme of the system. bled variant. Namely, 50 μL of 50 ng/μL HER2 in PBS All reagents used were acquired in Sigma-Aldrich un- solution (10004-HCCH-50, SinoBiological) was placed in less otherwise noted. the wells of 96-well Maxisorp plates (ThermoFisher Sci- entific) for its adsorption at 4 °C overnight. After re- Analytical methods moval of the supernatants, 100 μL of 2% skim milk SDS-PAGE gels (Oxoid, LP0031) in PBST (PBS plus 0.05% Tween 20 Protein mixture composition and relative amount of (Sigma, P2287)) were added and allowed to stand at each species in the sample was analysed by denaturing, room temperature for 1 h. The plates were washed three non-reducing SDS-PAGE (MiniProtean TGX Stain-free times with PBST and 50 μL of samples at different dilu- gels 4–20%, Biorad) following by band densitometry tions in PBS were placed in each well. After standing at using GelDoc EZ software (Biorad). room temperature for 1 h, the plates were washed three times with PBST. Then, 50 μL of a solution of 0.1 μg/mL SEC-HPLC of policlonal anti-human IgG1 conjugated to Horserad- SEC analysis was performed in Waters Alliance 2695 ish peroxidase (Genscript, A10254) was added and the (Waters) system using Zenix-C SEC-300 (4.6 × 300 mm, plates were allowed to stand at room temperature for 1 Farràs et al. BMC Molecular and Cell Biology (2020) 21:2 Page 9 of 9 h. After three washings with PBST, 50 μL of TMB 7. Weiner GJ. Building better monoclonal antibody-based therapeutics. Nat. (Pierce, 1,854,050) were added. The staining reaction Rev. Cancer. 2015;15(6):361–70. https://doi.org/10.1038/nrc3930. 8. Mahmuda A, Bande F, Kadhim Al-Zihiry KJ, Abdulhaleem N, Majid RA, was carried out at room temperature for 10–15 min and Hamat RA, Abdullah WO, Unyah Z. Monoclonal antibodies: A review of stopped by adding 50 μL of 20% sulfuric acid (Panreac, therapeutic applications and future prospects. Trop. J. Pharm Res. 2017;16:3– 141,058.1611) addition. The absorbance at 450–630 nm 713. https://doi.org/10.4314/tjpr.v16i3.29. 9. Cheng P, Wang X, Jing G, Zhao K, Zhou J, Guo Z. Monoclonal antibody, a was measured with a Labtech LT-4000 microplate novel probe for protein folding. Sci. China Ser. C Life Sci. 1998;41(2):163–8. reader (Labtech, MI, USA). 10. Goldberg ME. Investigating protein conformation, dynamics and folding with monoclonal antibodies. Trends Biochem. Sci. 1991;16(10):358–62. Abbreviations 11. Feige MJ, Hendershot LM, Buchner J. How antibodies fold, Trends. Biochem. BiP: Immunoglobulin heavy chain binding protein; CH1: Constant Heavy Sci. 2010;35(4):189–98. https://doi.org/10.1016/j.tibs.2009.11.005. chain subunit 1; CL: Constant Light chain; EDTA: Etilendiamintetraacetic acid; 12. Goto Y, Azuma T, Hamaguchi K. Refolding of the immunoglobulin light HC: Heavy Chain; HEK cells: Human Embryonic Kidney; Ig: Immunoglobulin; Chain1. J Biochem. 1979;85:1427–38. LC: Light Chain; mAb: Monoclonal antibody; ND: Not Detected; PDI: Protein 13. Lazar KL, Patapoff TW, Sharma VK. Cold denaturation of monoclonal Disulphide Isomerase; SDS-PAGE: Dodecyl sodium sulfate Polyacrilamide antibodies. MAbs. 2010;2(1):42–52. Electrophoresis Gel 14. Freund C, Gehrig P, Baici A, Holak TA, Plückthun A. Parallel pathways in the folding of a short-term denatured scFv fragment of an antibody. Folding Acknowledgements Des. 1998;3(1):39–49. https://doi.org/10.1016/S1359-0278(98)00007-8. The authors would like to thank Dr. A. Kamen (National Research Council of 15. Simpson ER, Herold EM, Buchner J. The folding pathway of the antibody V L Canada) for kindly providing the HEK293 cells which this work was domain. J Mol Biol. 2009;392:1326–38. https://doi.org/10.1016/j.jmb.2009.07.075. performed with. 16. Buchner J, Rudolph R. Renaturation, purification and characterization of recombinant Fab-fragments produced in Escherichia coli. Biotechnology. (N. Authors’ contributions Y). 1991;9(2):157–62. https://doi.org/10.1038/nbt0291-157. MC, JM, RR and MF performed all the experimental work described. AC, JC, 17. Lilie H, McLaughlin S, Freedman R, Buchner J. Influence of protein disulfide OM and XP lead the experimental strategy. MF wrote the manuscript. All isomerase (PDI) on antibody folding in vitro. J. Biol. Chem. 1994;269(19):14290–6. authors read and approved the final manuscript. 18. M. Mayer, U. Kies, R. Kammermeier, and J. Buchner, BiP and PDI cooperate in the oxidative folding of antibodies in vitro., J. Biol. Chem. 2000; vol. 275 38: Funding 29421–29425. Doi: https://doi.org/10.1074/jbc. M002655200. This work was supported by Farmhispania SA with funding from Centro de 19. Maeda Y, Ueda T, Imoto T. Effective renaturation of denatured and reduced Desarrollo Tecnológico Industrial (Spanish Government), IDI-20170546. immunoglobulin G in vitro without assistance of chaperone. Protein Eng. 1996;9:95–10. The funding bodies played no role in the design of the study and collection, 20. Maeda Y, Koga H, Yamada H, Ueda T, Imoto T. Effective renaturation of analysis, and interpretation of data and in writing the manuscript. reduced lysozyme by gentle removal of urea. Prot Eng Des Sel. 1995. https://doi.org/10.1093/protein/8.2.201. Availability of data and materials 21. Junutula JR, Raab H, Clark S, Bhakta S, Leipold DD, Weir S, Lu Y. Site-specific The datasets used and/or analysed during the current study available from conjugation of a cytotoxic drug to an antibody improves the therapeutic the corresponding author on reasonable request. index. Nat Biotechnol. 2008;26(8):925. https://doi.org/10.1038/nbt.1480. 22. Beck A, Goetsch L, Dumontet C, Corvaïa N. Strategies and challenges for the next generation of antibody-drug conjugates. Nat Rev Drug Discov. 2017 Ethics approval and consent to participate May;16(5):315–37. https://doi.org/10.1038/nrd.2016.268. No applicable. 23. O'Mara B, Gao ZH, Kuruganti M, Mallett R, Nayar G, Smith L, Meyer JD, Therriault J, Miller C, Cisney J. Fann. Impact of depth filtration on disulfide Consent for publication bond reduction during downstream processing of monoclonal antibodies No applicable. from CHO cell cultures. J. Biotechnol Bioeng. 2019 Jul;116(7):1669–83. https://doi.org/10.1002/bit.26964. Competing interests 24. Du C, Huang Y, Borwankar A, Tan Z, Cura A, Yee JC, Singh N, Ludwig R, The authors declare no competing financial interest. Borys M, Ghose S, Mussa N, Li ZJ. Using hydrogen peroxide to prevent antibody disulfide bond reduction during manufacturing process. MAbs. Author details 2018 Apr;10(3):500–10. https://doi.org/10.1080/19420862.2018.1424609. 1 Department of Biotechnology, Farmhispania SA, Montmeló, Spain. 25. Wang T, Liu YD, Cai B, Huang G, Flynn GC. Investigation of antibody disulfide 2 Department of Chemical, Biological and Environmental Engineering, reduction and re-oxidation and impact to biological activities. J Pharm Biomed Autonomous University of Barcelona, Cerdanyola del Vallès, Spain. Anal. 2015 Jan;102:519–28. https://doi.org/10.1016/j.jpba.2014.10.023. 26. R. R. J. Miret, A. Roura, C. Moreno, G. Arboix, M. Farràs, D. Cancelliere, C. Di Gesù, A. Received: 21 August 2019 Accepted: 17 December 2019 Casablancas, M., Lecina, and J. J. Cairó, PO276 comparison of bicistronic and tricistonic expression strategies for trastuzumab and trastuzumab-interferon-a2b production in CHO and HEK293 cells, in 25th ESACT MEETING, 2017. References 27. Román R, Miret J, Scalia F, Casablancas A, Lecina M, Cairó JJ. Enhancing 1. Walsh G. Biopharmaceutical benchmarks. Nat Biotechnol. 2010;28(9):917–24. heterologous protein expression and secretion in HEK293 cells by means of https://doi.org/10.1038/nbt0910-917. combination of CMV promoter and IFNα2 signal peptide. J Biotechnol. 2. Carter PJ, Lazar GA. Next generation antibody drugs: pursuit of the 'high- 2016, 239:57–60. https://doi.org/10.1016/j.jbiotec.2016.10.005. hanging fruit'. Nat Rev Drug Discov. 2018 Mar;17(3):197–223. https://doi.org/ 10.1038/nrd.2017.227. 3. Kaplon H, Reichert JM. Antibodies to watch in 2019. MAbs. 2019 Feb/Mar; Publisher’s Note 11(2):219–238. doi: https://doi.org/10.1080/19420862.2018.1556465 Springer Nature remains neutral with regard to jurisdictional claims in 4. Chames P, Van Regenmortel M, Weiss E, Baty D. Therapeutic antibodies: published maps and institutional affiliations. successes, limitations and hopes for the future. Br. J. Pharmacol. 2009;157(2): 220–33. https://doi.org/10.1111/j.1476-5381.2009.00190.x. 5. Kontermann RE. Dual targeting strategies with bispecific antibodies. mAbs. 2012;4(2):182–97. 6. Labrijn AF, Janmaat ML, Reichert JM, PWHI P. Bispecific antibodies: a mechanistic review of the pipeline. Nat Rev Drug Discov. 2019 Aug;18(8): 585–608. https://doi.org/10.1038/s41573-019-0028-1.