sign in sign up
<<
Home , Diarylheptanoid, Epigallocatechin gallate

Article Screening Naturally Occurring Phenolic Antioxidants for Their Suitability as Additives to CHO Cell Culture Media Used to Produce Monoclonal Antibodies

Luis Toronjo Urquiza 1 , David C. James 1, Tibor Nagy 2 and Robert J. Falconer 3,* 1 Department of Chemical & Biological Engineering, ChELSI Institute, University of Sheffield, Sheffield S1 3JD, UK; ltoronjourquiza1@sheffield.ac.uk (L.T.U.); d.c.james@sheffield.ac.uk (D.C.J.) 2 Fujifilm Diosynth Biotechnologies, Belasis Ave, Stockton-on-Tees, Billingham TS23 1LH, UK; tibor.nagy@fujifilm.com 3 Department of Chemical Engineering, University of Adelaide, Adelaide, SA 5005, Australia * Correspondence: [email protected]

 Received: 4 April 2019; Accepted: 30 May 2019; Published: 3 June 2019 

Abstract: This study identified several antioxidants that could be used in Chinese hamster ovary (CHO)cell culture media and benefit monoclonal antibody production. The flavan-3-ols, , epicatechin, and all had no detrimental effect on cell viability at the concentrations tested, and they reduced the final viable cell count with a resulting rise in the cell specific productivity. The flavone, behave similarly to the flavan-3-ols. at 50 µM concentration resulted in the most pronounced reduction in viable cell density with minimal decrease in IgG synthesis and the largest increase in cell specific productivity. Low concentrations of α- (35 µM) reduced viable cell density and raised cell specific productivity, but at higher concentration it had little additional effect. As high concentrations of α-tocopherol are not toxic to CHO cells, its addition as an anti-oxidant has great potential. up to 50 µM, up to 20 µM and piceid up to 100 µM showed little effect on growth or IgG synthesis and could be useful as antioxidants. Caffeic acid phenethyl ester was toxic to CHO cell and of no interest. Seven of the phenolic compounds tested are potential cell cycle inhibitors as well as having intrinsic properties.

Keywords: flavan-3-ol; catechin; epicatechin; epigallocatechin gallate (EGCG); resveratrol; curcumin; monoclonal antibodies (mAb); immunoglobulin G (IgG); immunoglobulin; chemical chaperone

1. Introduction Antioxidants are of potential benefit in the production of monoclonal antibodies (mAb) by Chinese hamster ovary (CHO) cells due to their capacity to reduce the damage by free radicals to the mAb and to the CHO cells in cell culture media. A range of small molecules of no direct nutritional value have been added to cell culture media to improve productivity or to protect the mAb being produced. It is the aim of this paper to assess whether a range of naturally occurring polyphenolic antioxidants can be added to the beneficial ingredients used in CHO cell culture media. The productivity of mammalian cell culture has been the focus of considerable research and has resulted in marked improvements in reactor design and media composition. Early advances resulted in improved reactor design, enhanced feed strategies and improved media composition [1]. Researchers have also used mild hypothermia, hyperosmotic pressure and the addition of small molecules with no direct nutritional value to improve the cell specific productivity (qp) or reduce protein aggregation [2]. Temperature is a variable that impacts both the cell growth rate and the protein expression rate. By modulating temperature conditions can be provided to favour cell growth then protein synthesis

Antioxidants 2019, 8, 159; doi:10.3390/antiox8060159 www.mdpi.com/journal/antioxidants Antioxidants 2019, 8, 159 2 of 18 optimising fed-batch or perfusion systems. Mild hypothermia (reduced temperature) has been used to slow cellular growth while allowing protein synthesis to continue, improving the qp values [3–6]. Temperature reduction can also reduce protein aggregation in a cell culture reactor [2]. Another method for modulating cell growth rates is by altering the osmotic pressure of the media. The negative effect caused by the hyperosmotic pressure conditions is counteracted by the addition of the osmolyte glycine betaine which achieves slower cellular growth and can in some circumstances improve the qp value [7,8]. The presence of low concentrations of glycine betaine does not interfere with cell growth or protein synthesis, its effect is solely due to modulating the osmotic pressure. Some small molecules (often referred to as chemical chaperones) are capable of enhancing protein folding, improving protein and slowing protein aggregation. The fatty acids, butyric acid, valeric acid, valproic acid and 4-phenylbutyric acid have been added to cell culture media [9–13]. Fatty acids have a hydrophobic end and a negatively charged end, so they are capable of hydrophobic interaction with a proteins hydrophobic surfaces displaying the charged end to the liquid-phase or an electrostatic interaction with positively charged side chains on the protein. The former interaction may enhance protein solubility while the second interaction could reduce solubility, so the effect is likely to be protein specific. These interactions are non-specific and will occur with the recombinant protein, other proteins and with the cells and their constituents. The fatty acids may also affect cell function. Sub-toxic doses might be cytostatic (which could increase the qp value) or induce apoptosis. Sodium butyrate slowed CHO cell growth and improved qp of recombinant tissue plasminogen activator production [9]. Valeric acid (pentanoic acid) produced higher product titres and increased qp for an unnamed recombinant protein in CHO cells [10]. Valproic acid (2-propylpentanoic acid) addition to cell culture media has also been reported to improve antibody expression [11,13]. 4-phenylbutyrate has also been used to minimise aggregation in protein expressed in CHO cells [12]. One family of chemicals that has received extensive research due to potential benefits to human health are the phenolic compounds, but research in their use as potential additives in cell culture media is limited [14–16]. Phenolic compounds are common in nature and play a range of roles, especially in vascular plants where they have been associated with antimicrobial activity, UV protection, antioxidant activity [17], and allelopathy [18]. Mammals have evolved in parallel with plants, and in many species plants form a major part in their diet. Herbivores and omnivores ingest significant quantities of phenolic compounds usually with little negative impact. In some cases, plant-derived phenolic compounds have been credited with a diverse range of benefits for the feeding animal. have had diverse medical attributes allocated to them, including the treatment and prevention of cancer and cardiovascular disease. They have also been suggested to be antiulcer, antithrombotic, anti-inflammatory, antiallergenic, anticoagulant, immune modulating, antimicrobial, vasodilatory, and analgesic [19]. Epigallocatechin gallate (EGCG), in particular, has been widely studied and it is claimed to regulate signal transduction pathways, factors, DNA methylation, mitochondrial function, and autophagy [20]. The phenolic compound with obvious importance in human health is α-tocopherol, which is the form of found in dietary supplements. Recent investigations into the beneficial properties of green , red wine, and leaf vegetables has resulted in extensive studies focusing on the flavan-3-ols, the flavonol kaempferol, the diarylheptanoid curcumin and the stilbene resveratrol. In this study, we screened a range of naturally occurring phenolic compounds found in the human diet for their effect on Chinese hamster ovary cells and the synthesis of IgG to detect whether they were potential candidates for cell cycle arrest, chemical chaperones or to provide antioxidant capabilities in cell culture media. The candidates included the flavan-3-ols, catechin, epicatechin, epigallocatechin gallate and gallocatechin gallate (Figure1). These compounds are ubiquitous in vascular plants, and is found in products like cocoa, wine and tea. Kaempferol is a flavonol and is found in a wide range of vegetables and fruit. Apart from being an antioxidant it has also been linked to a wide range of beneficial health outcomes in humans. Luteolin is a flavone, is found in a range of leaf vegetables and is an antioxidant (Figure1). Resveratrol is a and has two phenol groups connected by two Antioxidants 2019, 8, x FOR PEER REVIEW 3 of 18 Antioxidants 2019, 8, 159 3 of 18 to a wide range of beneficial health outcomes in humans. Luteolin is a flavone, is found in a range of leaf vegetables and is an antioxidant (Figure 1). Resveratrol is a stilbenoid and has two phenol groups double bonded carbons and is found in wine. It is an effective antioxidant and its potential health connected by two double bonded carbons and is found in wine. It is an effective antioxidant and its benefits have been widely studied. Piceid (also known as polydatin) is the form of resveratrol. potential health benefits have been widely studied. Piceid (also known as polydatin) is the glycoside Like resveratrol, it is found in a range of plants including grapes. Curcumin is a diarylheptanoid, it form of resveratrol. Like resveratrol, it is found in a range of plants including grapes. Curcumin is a has two phenolic groups linked by a seven carbon chain. It is found in turmeric, where it imparts the diarylheptanoid, it has two phenolic groups linked by a seven carbon chain. It is found in turmeric, yellow color and has antioxidant properties. Caffeic acid phenethyl ester (CAPE) is the ester of caffeic where it imparts the yellow color and has antioxidant properties. phenethyl ester (CAPE) acid and phenethyl and is found in honey. α-tocopherol is one form of vitamin E and it is a is the ester of caffeic acid and phenethyl alcohol and is found in honey. α-tocopherol is one form of fat soluble antioxidant found in a range of vegetable oils. The structures of CAPE, piceid, curcumin, vitamin E and it is a fat soluble antioxidant found in a range of vegetable oils. The structures of CAPE, resveratrol and α-tocopherol are shown in Figure1. In screening candidates that are causing cell cycle piceid, curcumin, resveratrol and α-tocopherol are shown in Figure 1. In screening candidates that arrest, it would be expected to reduce the viable cell count in the culture while having a less dramatic are causing cell cycle arrest, it would be expected to reduce the viable cell count in the culture while reduction in IgG synthesis. Chemical chaperones on the other hand, may increase the apparent IgG having a less dramatic reduction in IgG synthesis. Chemical chaperones on the other hand, may concentration by preventing aggregation. All candidates are capable of providing antioxidant capacity increase the apparent IgG concentration by preventing aggregation. All candidates are capable of in cell culture media. providing antioxidant capacity in cell culture media.

Figure 1. Cont. Antioxidants 2019, 8, 159 4 of 18 Antioxidants 2019, 8, x FOR PEER REVIEW 4 of 18

Figure 1.1. TheThe structure structure of of the flavan-3-ols;-3-ols; catechin catechin,, epicatechin, gallocatechin gallate and epigallocatechin gallate. gallate. The The flavonol flavonol kaempferol kaempferol and and flavone flavone luteolin. luteolin. Caffeic Ca ffacideic phenethyl acid phenethyl ester ester(CAPE), (CAPE), piceid, piceid, curcum curcumin,in, resveratrol resveratrol and α and-tocopherol.α-tocopherol. 2. Materials and Methods 2. Materials and Methods The following reagents were used during the process: dimethyl sulfoxide (DMSO), phosphate The following reagents were used during the process: dimethyl sulfoxide (DMSO), phosphate buffered saline (PBS), caffeine, 4-phenylbutyrate, glycine betaine, ( ) catechin, ( ) epicatechin, ( ) buffered saline (PBS), , 4-phenylbutyrate, glycine betaine, (±)± catechin, (−−) epicatechin, (−−) gallocatechin gallate, epigallocatechin gallate, luteolin, kaempferol, resveratrol, piceid, caffeic acid gallocatechin gallate, epigallocatechin gallate, luteolin, kaempferol, resveratrol, piceid, caffeic acid phenethyl ester (CAPE), curcumin and α-tocopherol were supplied by Sigma-Aldrich (St. Louis, MO, phenethyl ester (CAPE), curcumin and α-tocopherol were supplied by Sigma-Aldrich (St. Louis, USA). OptiCHO free serum media and l-glutamine were supplied by Thermo Fisher Scientific (Ulm, Missouri, USA). OptiCHO free serum media and L-glutamine were supplied by Thermo Fisher Germany). Further well known chaperon chemicals were also assessed as reference (AppendixB). Scientific (Ulm, Germany). Further well known chaperon chemicals were also assessed as reference The cell line used was a stable recombinant DG44 CHO cell line that expressed an IgG monoclonal (Appendix B). antibody, provided by Fujifilm Diosynth. This cell line was then left to grow in suspension in an The cell line used was a stable recombinant DG44 CHO cell line that expressed an IgG OptiCHO™ free serum media in a 125 mL Erlenmeyer Flask (Corning, NY, USA) containing 33 mL monoclonal antibody, provided by Fujifilm Diosynth. This cell line was then left to grow in of culture media, supplemented with 8 mM l-glutamine and 100 mg/L hygromycin B. Culture was suspension in an OptiCHOTM free serum media in a 125 mL Erlenmeyer Flask (Corning, NY, USA) maintained in a Infors HT Multitron cell shaking incubator (Bottmingen, Switzerland) set at 130 rpm containing 33 mL of culture media, supplemented with 8 mM L-glutamine and 100 mg/L hygromycin with 5% CO modified atmosphere at 37 C. Cells were seeded into new media at 0.2 106 viable B. Culture was2 maintained in a Infors HT Multitron◦ cell shaking incubator (Bottmingen, Switzerland)× cells/mL every three to four days for a maximum of 16 passages prior to use. set at 130 rpm with 5% CO2 modified atmosphere at 37oC. Cells were seeded into new media at 0.2 × 2.1.106 viable 24-Well cells/mL CHO Cell every Growth three and to IgGfour Expression days for a maximum of 16 passages prior to use.

2.1 24-WellThe screening CHO Cell test Growth was designed and IgG Expression to estimate the concentration at which these chemicals become toxic on CHO cell lines and to determine whether the treatments affects IgG expression. Chemical treatmentsThe screening were carried test was in a designed Greiner bio-one to estimate Cellstar the® concentration24-well-plate at (Stonehouse, which these UK), chemicals where become 1 mL of exponentiallytoxic on CHO growingcell linescells and wereto determine seeded at whether an initial the viable treatments cell density affects (VCD) IgG expression. of 0.2 106 Chemicalcells/mL. ® × Wellstreatments were thenwere treatedcarried atintime a Greiner 0 using bio-one a different Cellstar range 24-well-plate of concentrations (Stonehouse, for each UK), chemical where diluted 1 mL 6 inof 5exponentiallyµL of DMSO. growing The control cells wellswere wereseeded treated at an initial with the viable same cell volume density of (VCD) DMSO. of Cells0.2 × 10 were cells/mL. grown Wells were then treated at time 0 using a different range of concentrations for each chemical diluted for a period of three days in a static incubator (Infors HT) at 5% CO2 modified atmosphere and at a in 5 μL of DMSO. The control wells were treated with the same volume of DMSO. Cells were grown temperature of 37 ◦C. for a period of three days in a static incubator (Infors HT) at 5% CO2 modified atmosphere and at a 2.2.temperature Viable Cell of Density, 37 °C. Viability and IgG Concentration Analysis 500 µL were used to measure viable cell density (VCD) and viability by cytometry using a 2.2. Viable Cell Density, Viability and IgG Concentration Analysis Beckman-Coulter Vi-cell™ XR cell viability analyser (Indianapolis, IN, USA). 300 µL from the same 500 μL were used to measure viable cell density (VCD) and viability by cytometry using a Beckman-Coulter Vi-cellTM XR cell viability analyser (Indianapolis, IN, USA). 300 μL from the same Antioxidants 2019, 8, 159 5 of 18 samples were centrifuged for 10 min at 400 g, the supernatant collected and the IgG concentration measured using bio-layer interferometry using a Pall FortéBio Octet® QKe (Fremont, CA, USA). Specific protein production (qp) was determined by dividing the total protein produced after three days incubation by the final VCD [21].

Final Recombinant titer Concentration (ug/mL) qp(pg/cell day) = · VCD (106 viable cells/mL) days · 2.3. Statistical Analysis A permutation test was performed to compare the differences between the control and the treatments for all the variables studied. p-values below 0.05 were considered significant. Data was analysed using R programming tool and the coin package [22].

3. Results Catechin and epicatechin are isomers and only differ in the direction of the hydroxyl group on the carbon 3 on the dihydropyran (C) ring (see Figure1). The e ffect of catechin on CHO cell viability, viable cell density (VCD) and IgG synthesis was tested at 10, 25, 50, 75 and 100 µM (Figure2). Cell viability remained high over the whole concentration range. CHO cell growth was affected, the VCD dropped from 1.29 0.18 106 cells per mL at 0 µM catechin to 0.73 0.18 106 cells per mL for ± × ± × 25 µM and 0.64 0.10 106 cells per mL for 50 µM. Higher concentrations of catechin did not result in ± × any further drop in VCD. IgG synthesis was reduced, but not to the same extent as the cell growth, final IgG concentration dropped from 8.2 0.6 mg/L in the 0 µM catechin control to 6.3 0.9 for 10 µM ± ± and 5.9 0.8 for 50 µM, higher concentrations of catechin did not result in any further drop in IgG ± concentration. The calculated qp rose from 2.1 0.1 pg/cell/day at 0 µM to 3.1 0.2 pg/cell/day at ± ± 50 µM catechin. The effect of epicatechin on CHO cell viability, viable cell density (VCD) and IgG synthesis was tested at 1, 5, 10, 15 and 20 µM (Figure A1). Cell viability remained high over the whole concentration range. VCD dropped to 0.79 0.04 106 cells per mL for 0 µM and 0.56 0.03 106 cells ± × ± × per mL for 20 µM epicatechin. IgG synthesised remained stable over the concentration range tested. The calculated qp rose from 5.2 0.1 pg/cell/day at 0 µM to 7.3 0.2 pg/cell/day at 20 µM epicatechin. ± ± Unsurprisingly, the two isomers had a similar effect CHO cells. Gallocatechin gallate (GCG) and epigallocatechin gallate (EGCG) are also isomers. They are esters of gallocatechin and epigallocatechin, respectively. They differ in the direction of the ester bond on the carbon 3 on the dihydropyran (C) ring. GCG and EGCG are significantly larger than catechin and epicatechin, and have greater capacity to hydrogen bond than catechin and epicatechin. The effect of GCG on CHO cell viability, viable cell density (VCD) and IgG synthesis was tested at 5, 25, 50, 75 and 100 µM (Figure A2). Cell viability remained high over the whole concentration range. VCD dropped to 0.80 0.06 106 cells per mL for 0 µM and 0.44 0.09 106 cells per mL for 100 µM ± × ± × GCG. IgG synthesised remained stable over the concentration range tested. The calculated qp rose from 3.3 0.1 pg/cell/day at 0 µM to 5.7 0.4 pg/cell/day at 100 µM GCG. The effect of EGCG on CHO cell ± ± viability, viable cell density (VCD) and IgG synthesis was tested at 5, 25, 50, 75 and 100 µM (Figure3). Cell viability remained high over the whole concentration range. VCD dropped from 0.94 0.02 106 ± × cells per mL for 0 µM to 0.39 0.03 106 cells per mL for 5 µM, higher concentrations of EGCG did ± × not result in any further drop in VCD. IgG synthesised dropped slightly over the concentration range in the experiment from 8.2 0.7 mg/L in the 0 µM catechin control to 6.2 0.4 mg/L for 100 µM EGCG. ± ± The calculated qp rose from 3.3 0.1 pg/cell/day at 0 µM to 4.8 0.3 pg/cell/day at 100 µM EGCG. ± ± Both GCG and EGCG had a marked impact on CHO VCD at the lowest concentration tested but no additional drop in VCD was observed at higher GCG or EGCG concentrations. Antioxidants 2019, 8, x FOR PEER REVIEW 6 of 18 AntioxidantsAntioxidants2019 2019, ,8 8,, 159 x FOR PEER REVIEW 66 of of 18 18

Figure 2. A: Effect of different catechin (Cat) concentrations on viable cell density (VCD) (blue) and Figure 2. (A)Effect of different catechin (Cat) concentrations on viable cell density (VCD) (blue) and viabilityFigure 2. (yellow) A: Effect B of: effect different of different catechin catechin (Cat) concentrat concentrationsions on (10–100 viable μcellM) densityIgG concentration (VCD) (blue) (grey) and viability (yellow) (B) effect of different catechin concentrations (10–100 µM) IgG concentration (grey) andviability relative (yellow) IgG productionB: effect of different(qP) (orange) catechin for CHOconcentrations cells grown (10–100 in 24-well μM) IgG plates. concentration Each data (grey)point and relative IgG production (qP) (orange) for CHO cells grown in 24-well plates. Each data point representsand relative the IgG average production from a (qP) quadruplicate, (orange) for and CHO the cellserro rgrown bars correspond in 24-well toplates. ±SD. Each* p-value data

Figure 3. (A)Effect of different epigallocatechin gallate (EGCG) concentrations on viable cell density (VCD)Figure (blue)3. A: Effect and viabilityof different (yellow) epigallocatechin (B) effect ofgallate different (EGCG) EGCG concentrations concentrations on viable (10–100 cellµM) density IgG Figure 3. A: Effect of different epigallocatechin gallate (EGCG) concentrations on viable cell density concentration(VCD) (blue) (grey)and viability and relative (yellow) IgG productionB: effect of (qP) different (orange) EGCG for CHO concentrations cells grown (10–100 in 24-well μM) plates. IgG (VCD) (blue) and viability (yellow) B: effect of different EGCG concentrations (10–100 μM) IgG Eachconcentration data point (grey) represents and relative the average IgG production from a quadruplicate, (qP) (orange) andfor CHO the error cells bars grown correspond in 24-well to plates.SD. concentration (grey) and relative IgG production (qP) (orange) for CHO cells grown in 24-well plates.± *Eachp-value data< point0.05 for represents the treatment the average compared from to a thequad control.ruplicate, and the error bars correspond to ±SD. * pEach-value data < 0.05 point for represents the treatment the average compared from to thea quad control.ruplicate, and the error bars correspond to ±SD. * p-value < 0.05 for the treatment compared to the control. Antioxidants 2019, 8, x FOR PEER REVIEW 7 of 18

All of the flavan-3-ols had a similar impact on CHO cells and the synthesis of IgG being responsible for a reduction in VCD but a smaller effect on IgG synthesis resulting in a rise in 𝑞. This suggests that the flavan-3-ols are all of potential use as a cell culture additive where a reduction cell growth is desirable and IgG synthesis is relatively unaffected. The flavonol, kaempferol’s effect on CHO cell growth and IgG synthesis was quite different to the flavan-3-ols. It was tested at 10, 50, 100, 150, and 200 μM kaempferol (Figure 4). Cell viability was Antioxidants 2019, 8, 159 7 of 18 unaffected between 0 and 50 μM, though there was a drop in VCD from 1.16 ± 0.12 × 106 cells per mL at 0 μM to 0.80 ± 0.08 × 106 cells per mL at 50 μM and a drop in IgG synthesis from 13.6 ± 0.4 mg/L at

0 μMAll to of 8.7 the ± flavan-3-ols0.5 mg/L at had50 μ aM. similar The 𝑞 impact remained on CHO stable cells between and the synthesis 0 and 50 ofμM IgG kaempferol. being responsible At 100 forμM a kaempferol reduction in viability VCD but and a smallerVCD dropped effect on drastically, IgG synthesis suggesting resulting this in dose a rise was in q toxicp. This to suggeststhe cells. thatThe theevidence flavan-3-ols from this are experiment all of potential suggest use askaempferol a cell culture could additive be used where as a cell a reduction culture media cell growth additive is desirableup to a concentration and IgG synthesis of 50 μ isM, relatively but there una is ffnoected. productivity benefit from its addition and could only be incorporatedThe flavonol, for kaempferol’s its anti-oxidant effect properties. on CHO cell growth and IgG synthesis was quite different to the flavan-3-ols.The flavone, It was luteolin tested was at 10, tested 50, 100, at 2, 150, 5, 10, and 15, 200andµ 20M μ kaempferolM (Figure 5). (Figure Cell viability4). Cell remained viability washigh unaoverff ectedthe whole between concentration 0 and 50 µ M,range. though VCD there dropped was a to drop 0.80 in ± VCD0.06 × from 106 cells 1.16 per0.12 mL for10 06 μcellsM and per 0.49 mL ± × at± 00.01µM × to10 0.806 cells 0.08per mL10 6forcells 20 perμM. mL IgG at synthesised 50 µM and aremained drop in IgG stab synthesisle over the from concentration 13.6 0.4 mg range/L at ± × ± 0tested.µM to The 8.7 calculated0.5 mg/L at𝑞 50 roseµM. from The q5.0p remained ± 0.1 pg/cell/day stable between at 0 μ 0M and to 7.3 50 µ±M 0.2 kaempferol. pg/cell/day At at 100 20 µμMM ± kaempferolluteolin. This viability suggests and that VCD luteolin dropped is of potential drastically, use suggesting as a cell culture this dose additive was toxicwhere to a thereduction cells. The cell evidencegrowth is from desirable this experiment and IgG synthesis suggest is kaempferol relatively unaffected. could be used as a cell culture media additive up to a concentrationCAPE was toxic of 50 toµ M,CHO but cells there even is no at productivity the lowest benefitconcentration from its (2 addition μM) and and was could considered only be incorporatedunsuitable as for a cell its anti-oxidantculture media properties. additive (Figure A3).

Figure 4. (A)Effect of different kaempferol (Kmp) concentrations on viable cell density (VCD) (blue) andFigure viability 4. A: (yellow)Effect of (differentB) effect ofkaempferol different kaempferol(Kmp) concentr concentrationsations on viable (10–100 cellµM) density IgG concentration (VCD) (blue) (grey)and viability and relative (yellow) IgG productionB: effect of different (qP) (orange) kaempferol for CHO concentrations cells grown in (10–100 24-well plates.μM) IgG Each concentration data point represents(grey) and the relative average IgG from production a quadruplicate, (qP) (orange) and the for error CHO bars cells correspond grown in to 24-wellSD. * pplates.-value Each< 0.05 data for ± thepoint treatment represents compared the average to the fr control.om a quadruplicate, and the error bars correspond to ±SD. * p-value < 0.05 for the treatment compared to the control. The flavone, luteolin was tested at 2, 5, 10, 15, and 20 µM (Figure5). Cell viability remained high overCurcumin the whole was tested concentration at 5, 10, 20, range. 40 and VCD 80 μ droppedM (Figure to A4). 0.80 Cell0.06 viability106 remainedcells per mLhigh for between 0 µM ± × and0 and 0.49 20 μ0.01M, but10 lost6 cells viability per mL at for40 μ 20M.µ M.Between IgG synthesised 0 and 20 μ remainedM curcumin stable there over was the a concentrationslight drop in ± × rangeVCD tested.and IgG The but calculated the calculatedqp rose 𝑞 from remained 5.0 0.1 stable. pg/cell Curcumin/day at 0 µ Mcould to 7.3 be added0.2 pg /cellto media/day at up 20 µtoM a ± ± luteolin.concentration This suggests of 20 μM that if its luteolin anti-oxidant is of potential properties use aswere a cell desired culture but additive are unlikely where ato reduction provide cellany growthbenefit isin desirableraising the and 𝑞 IgG value. synthesis is relatively unaffected. Antioxidants 2019, 8, 159 8 of 18 Antioxidants 2019, 8, x FOR PEER REVIEW 8 of 18

FigureFigure 5.5. (A:)E Effectffect of didifferentfferent luteolinluteolin (Ltl)(Ltl) concentrationsconcentrations onon viableviable cellcell densitydensity (VCD)(VCD) (blue)(blue) andand viabilityviability (yellow)(yellow) (B:) effect effect of didifferentfferent luteolinluteolin concentrationsconcentrations (10–100(10–100 µμM)M) IgGIgG concentrationconcentration (grey)(grey) andand relativerelative IgGIgG productionproduction (qP)(qP) (orange)(orange) forfor CHOCHO cellscells growngrown in in 24-well 24-well plates. plates. EachEach datadata pointpoint represents the average from a quadruplicate, and the error bars correspond to SD. * p-value < 0.05 for represents the average from a quadruplicate, and the error bars correspond± to ±SD. * p-value < 0.05 thefor treatmentthe treatment compared compared to the to control.the control. CAPE was toxic to CHO cells even at the lowest concentration (2 µM) and was considered The stilbenoid, resveratrol had a considerable effect on CHO cells and the synthesis of IgG. It unsuitable as a cell culture media additive (Figure A3). was tested at 10, 25, 50, 75, and 100 μM (Figure 6). Cell viability remained high between 0 and 50 μM Curcumin was tested at 5, 10, 20, 40 and 80 µM (Figure A4). Cell viability remained high between but lost viability at 75 μM. There was a drastic reduction in VCD from 1.22 ± .10 × 106 cells per mL at 0 and 20 µM, but lost viability at 40 µM. Between 0 and 20 µM curcumin there was a slight drop 0 μM to 0.28 ± 0.01 × 106 cells per mL at 50 μM. IgG synthesis also had a slight decline and the in VCD and IgG but the calculated qp remained stable. Curcumin could be added to media up to calculated 𝑞 rose from 3.2 ± 0.1 pg/cell/day at 0 μM to 5.1 ± 0.2 pg/cell/day at 100 μM. This suggests a concentration of 20 µM if its anti-oxidant properties were desired but are unlikely to provide any that resveratrol is of potential use as a cell culture additive where a reduction cell growth is desirable, benefit in raising the qp value. IgG synthesis is relatively unaffected and an improvement in 𝑞 is desired. Resveratrol has been The stilbenoid, resveratrol had a considerable effect on CHO cells and the synthesis of IgG. It was added to cell culture media before, as a potential inhibitor of autophagy that could improve tested at 10, 25, 50, 75, and 100 µM (Figure6). Cell viability remained high between 0 and 50 µM but recombinant production. The work used a single dose of 10 μM resveratrol [23] and observed little lost viability at 75 µM. There was a drastic reduction in VCD from 1.22 0.10 106 cells per mL at 0 µM effect. Our research6 indicates a higher concentration of resveratrol± is required× for a measurable to 0.28 0.01 10 cells per mL at 50 µM. IgG synthesis also had a slight decline and the calculated qp impact± on CHO× cell growth and IgG synthesis. Resveratrol is also an attractive candidate as an anti- rose from 3.2 0.1 pg/cell/day at 0 µM to 5.1 0.2 pg/cell/day at 100 µM. This suggests that resveratrol oxidant. Resveratrol± can be radicalised at each± of the three hydroxyl groups, the 4′ being the most is of potential use as a cell culture additive where a reduction cell growth is desirable, IgG synthesis is stable. The structure of resveratrol enables the electron from the free radical to freely move around relatively unaffected and an improvement in q is desired. Resveratrol has been added to cell culture the molecule and it is able to spread the energyp across entire molecule providing good stabilisation media before, as a potential inhibitor of autophagy that could improve recombinant production. The [24]. work used a single dose of 10 µM resveratrol [23] and observed little effect. Our research indicates a Piceid (also known as polydatin) is the glycoside form of resveratrol. It was tested at 10, 25, 50, higher concentration of resveratrol is required for a measurable impact on CHO cell growth and IgG 75, and 100 μM (Figure A5). Piceid had little effect in viability or VCD, but did reduce IgG synthesis synthesis. Resveratrol is also an attractive candidate as an anti-oxidant. Resveratrol can be radicalised and the resulting low 𝑞 value suggesting picead is a poor candidate for a cell culture additive. at each of the three hydroxyl groups, the 4 being the most stable. The structure of resveratrol enables α-tocopherol is the form of vitamin 0E used in vitamin tablets. It was tested at 35, 175, 350, 525, the electron from the free radical to freely move around the molecule and it is able to spread the energy and 700 μM (Figure 7). These concentrations were selected to stay under the critical micelle across entire molecule providing good stabilisation [24]. concentration. Cell viability remained high over the whole concentration range. The VCD dropped from 0.58 ± 0.01 × 106 cells per mL at 0 μM to 0.36 ± 0.03 × 106 cells per mL at 35 μM. The IgG synthesis remained stable so the calculated 𝑞 rose from 8.1 ± 0.1 pg/cell/day at 0 μM to 12.1 ± 0.3 pg/cell/day at 35 μM. Higher concentrations of α-tocopherol had no further effect on IgG synthesis or the calculated 𝑞value. This suggests that α-tocopherol is of potential use as a cell culture additive where a reduction cell growth is desirable, IgG synthesis is relatively unaffected and an improvement in 𝑞 is desired. It is also of added value if its anti-oxidant properties are required as relatively high Antioxidants 2019, 8, x FOR PEER REVIEW 9 of 18 concentrations can be added to cell culture media (700 μM) without a detrimental effect on cell Antioxidants 2019, 8, 159 9 of 18 viability.

Figure 6. (A)Effect of different resveratrol (Rsv) concentrations on viable cell density (VCD) (blue) andFigure viability 6. A: Effect (yellow) of different (B) effect resveratrol of different (Rsv) resveratrol concentrat concentrationsions on viable (10-100 cell densityµM) IgG (VCD) concentration (blue) and (grey)viability and (yellow) relative B IgG: effect production of different (qP) resveratrol (orange) for concentrations CHO cells grown (10-100 in 24-well μM) IgG plates. concentration Each data point(grey) representsand relative the IgG average production from a quadruplicate,(qP) (orange) andfor CHO the error cells bars grown correspond in 24-well to plates.SD. * p -valueEach data< 0.05 point for ± therepresents treatment the compared average from to the a control. quadruplicate, and the error bars correspond to ±SD. * p-value < 0.05 for the treatment compared to the control. Piceid (also known as polydatin) is the glycoside form of resveratrol. It was tested at 10, 25, 50, 75, and 1004-phenylbutyrate,µM (Figure A5). and Piceid glycine had little betaine effect also in viability were added or VCD, to butcell didculture reduce media IgG synthesisto provide and a thecomparison resulting lowto theqp value polyphenolic suggesting screening picead is results a poor candidate(see Figures for aA2 cell and culture A3 additive.in the Appendix). Interestingly,α-tocopherol none is of the these form “chemical of vitamin chaperones” E used in sh vitaminowed potential tablets. as It a was cell testedculture at additive 35, 175, in 350, this 525,study. and Glycine 700 µM betaine (Figure resulted7). These in a concentrations slight rise in VCD, were but selected no corresponding to stay under rise the in IgG critical synthesis. micelle 4- concentration.phenylbutyrate Cell was viability toxic above remained 2 mM. high At sub-toxi over thec wholeconcentrations, concentration it decreased range. Theboth VCD VCD dropped and IgG fromsynthesis, 0.58 but0.01 with10 6nocells corresponding per mL at 0 µriseM toin 0.36 𝑞 . The0.03 behaviour106 cells of per 4-phenylbu mL at 35 µtyrateM. The observed IgG synthesis here is ± × ± × remained stable so the calculated qp rose from 8.1 0.1 pg/cell/day at 0 µM to 12.1 0.3 pg/cell/day at in agreement with previous experimentation where± it was added to act as a “chemical± chaperone” 35andµM. reduce Higher aggregation concentrations of a of misfoldedα-tocopherol protein had no[12]. further Caffeine effect was on IgG also synthesis tested as or theit has calculated known qphysiologicalp value. This suggestsproperties that butα-tocopherol demonstrated is of no potential beneficial use effects as a cell on culture CHO cell additive growth where or IgG a reduction synthesis cell(see growth Figure isA4 desirable, in the Appendix). IgG synthesis is relatively unaffected and an improvement in qp is desired. It is alsoThe of added screen value was repeated if its anti-oxidant and despite properties variation are requiredin the final as relativelyVCD and highIgG concentration concentrations and can the be µ addedresulting to cell calculated culture media𝑞 values (700 inM) the without wells containing a detrimental the e ffuntreatedect on cell controls, viability. the trends seen in differences4-phenylbutyrate, in VCD, IgG and and glycine the 𝑞values betaine were also consistent were added between to cell the culturetwo sets media of plates, to provideindicating a comparisonthat the effect to of the the polyphenolic additives on screeningCHO cell resultsgrowth (seeand FiguresIgG synthesis A2 and were A3 consistent in the Appendix and thatA the). Interestingly,findings were none valid. of these “chemical chaperones” showed potential as a cell culture additive in this study. Glycine betaine resulted in a slight rise in VCD, but no corresponding rise in IgG synthesis. 4-phenylbutyrate was toxic above 2 mM. At sub-toxic concentrations, it decreased both VCD and IgG synthesis, but with no corresponding rise in qp. The behaviour of 4-phenylbutyrate observed here is in agreement with previous experimentation where it was added to act as a “chemical chaperone” and reduce aggregation of a misfolded protein [12]. Caffeine was also tested as it has known physiological properties but demonstrated no beneficial effects on CHO cell growth or IgG synthesis (see Figure A4 in the AppendixA). AntioxidantsAntioxidants2019 2019, ,8 8,, 159 x FOR PEER REVIEW 1010 of of 18 18

Figure 7. (A)Effect of different α-tocopherol (Tcp) concentrations on viable cell density (VCD) (blue) Figure 7. A: Effect of different α-tocopherol (Tcp) concentrations on viable cell density (VCD) (blue) and viability (yellow) (B) effect of different α-tocopherol concentrations (10–100 µM) IgG concentration and viability (yellow) B: effect of different α-tocopherol concentrations (10–100 μM) IgG (grey) and relative IgG production (qP) (orange) for CHO cells grown in 24-well plates. Each data point concentration (grey) and relative IgG production (qP) (orange) for CHO cells grown in 24-well plates. represents the average from a quadruplicate, and the error bars correspond to SD. * p-value < 0.05 for Each data point represents the average from a quadruplicate, and the error bars± correspond to ±SD. * the treatment compared to the control. p-value < 0.05 for the treatment compared to the control. The screen was repeated and despite variation in the final VCD and IgG concentration and 4. Discussion the resulting calculated qp values in the wells containing the untreated controls, the trends seen in differencesThe early in VCD, use of IgG antioxidants and the qp valuesin cell wereculture consistent media aimed between at reduction the two sets of mortality of plates, indicatingof Chinese thathamster the e ovaryffect of cells the by additives reducing on the CHO intracellular cell growth concentration and IgG synthesis of reactive were oxygen consistent species and [25,26]. that theThe findingsspecies include were valid. superoxide anions, hydrogen peroxide, hydroxyl radicals, and singlet oxygen all of which can react directly with the cells and with the recombinant protein product. In one study, the 4.non-phenolic Discussion antioxidants reduced and L-ascorbic acid 2-phosphate were shown to suppressThe early apoptosis use of and antioxidants consequently in cellincrease culture synthe mediasis aimedof tissue at plasminogen reduction of activator mortality [25]. of Chinese A range hamsterof phenolic ovary antioxidants cells by reducing have also the been intracellular tested for their concentration capacity to of reduce reactive reactive oxygen oxygen species species [25,26 in]. TheCHO species cells [15]. include The superoxide phenolic antioxidants anions, hydrogen included peroxide, , hydroxyl catechin, radicals, chlorogenic and singlet acid, oxygen , all ofgenistein, which can hesperetin, react directly , with the , cells and withpelargonidin, the recombinant , protein resveratrol, product. rosmarinic In one study, acid theand non-phenolic . Interestingly, antioxidants only reduced two of glutathionethe phenolics and tested,l-ascorbic daidzein acid and 2-phosphate , were were toxic shown to the to suppressCHO cells. apoptosis The experimental and consequently work indicated increase synthesis catechin ofand tissue rosmarinic plasminogen acid performed activator [ 25best]. A in range their ofcultures phenolic reducing antioxidants build-up have of alsolactate been and tested ammoni for theirum, capacityand improving to reduce the reactiverecombinant oxygen protein species titre. in CHOThe impact cells [15 of]. the The antioxidants phenolic antioxidants on chemical includedmodification apigenin, of the catechin,recombinant chlorogenic product acid,was not daidzein, tested. genistein,The phenolic hesperetin, antioxidants, melatonin, EGCG naringenin, and rutin, pelargonidin, have been used quercetin, to reduce resveratrol, acidic species rosmarinic charge acid variants and silibinin.of recombinant Interestingly, therapeutic only two proteins of the produced phenolics by tested, CHO daidzein cells [14]. and Experiments genistein, were using toxic rosmarinic to the CHO acid cells.in CHO The experimentalcell culture media work indicatedreduced catechinoxidative and stress rosmarinic which acidalso performedreduced protein best in glycation their cultures and reducingformation build-up of acidic of charge lactate variants and ammonium, [16]. and improving the recombinant protein titre. The impact of theThe antioxidants results of on this chemical paper modification and the work of theof recombinantothers [15,27–29] product suggest was not th tested.e pallet The of phenolicphenolic antioxidants,antioxidants EGCGthat could and rutin,be used have to beenreduce used reactive to reduce oxygen acidic species species in chargecell cultures variants is very of recombinant broad and therapeuticcould include, proteins but not produced be confined, by CHO to cells the [four14]. Experimentsflavan-3-ols, usingluteolin, rosmarinic resveratrol, acid α in-tocopherol, CHO cell culturechlorogenic media acid, reduced silibinin oxidative and rosmarinic stress which acid. also Th reducede effect of protein the polyphenols glycation and on formationthe CHO cells of acidic may chargebe to reduce variants oxidative [16]. stress but they also slowed the cell growth (observed as a drop in VCD and an increase in the 𝑞value). The cause of the slowing of CHO cell growth is not apparent. Polyphenols interact with other molecules by a combination of hydrophobic interaction, hydrogen bonding [30,31] Antioxidants 2019, 8, 159 11 of 18

The results of this paper and the work of others [15,27–29] suggest the pallet of phenolic antioxidants that could be used to reduce reactive oxygen species in cell cultures is very broad and could include, but not be confined, to the four flavan-3-ols, luteolin, resveratrol, α-tocopherol, , silibinin and rosmarinic acid. The effect of the polyphenols on the CHO cells may be to reduce oxidative stress but they also slowed the cell growth (observed as a drop in VCD and an increase in the qp value). The cause of the slowing of CHO cell growth is not apparent. Polyphenols interact with other molecules by a combination of hydrophobic interaction, hydrogen bonding [30,31] and potential pi-cation interaction. Polymers of phenolics such as (polymer of catechin, epicatechin and their gallic acid esters) can cross link and aggregate proteins [32–34]. The monomers reversibly interact with proteins with minimal negative impact on solubility. The effect of phenolics on CHO cells in this screen is likely to be due to a range of non-specific interactions with the proteins and phospholipids on the CHO cell surface. The net effect of the interaction of the phenolics (the flavan-3-ols, luteolin and α-tocopherol) was to slow cell growth, retain high cell viability and retain the capability to synthesise IgG. The options for cell culture strategies that are currently available are diverse. One strategy is to reduce the growth of the cells while maintaining protein synthesis. Reduction of the bioreactor temperature, increasing the osmotic pressure and the addition of chemicals that interfere with the cell cycles can all slow cell growth and increase qp values. Cell culture additives that reduce reactive oxygen species include both non-phenolic and phenolic antioxidants. The reduction of reactive oxygen species can benefit the cells directly by reducing oxidative stress on the cells but also protect the recombinant protein product from chemical damage caused by free radicals. The use of fatty acid chemical chaperones (butyric acid, valeric acid, valproic acid and 4-phenylbutyric acid) capacity for reducing aggregation of the recombinant protein product, whether in the cell or in the cell culture media is a separate strategy for improving cell culture operations. The beneficial attributes of phenolic antioxidants is their interaction with CHO cells, and the recombinant product is likely to be a combination of their ability to reduce reactive oxygen species and slow cell growth.

5. Conclusions The unpredicted result from this work was that out of the 11 phenolics tested only one was toxic at low doses (CAPE) and only three had no effect on specific cell productivity (kaempferol, curcumin and piceid). The remaining seven phenolics (the four flavan-3-ols, luteolin, resveratrol and α-tocopherol) are potential cell culture media additives with the capacity to raise specific cell productivity as well as the potential of providing some protection from free radical damage. α-tocopherol is the strongest potential candidate for an additive with anti-oxidant properties, due to its low toxicity even at high concentrations (700 µM), which is close to α-tocopherol’s critical micelle concentration. Resveratrol was the most potent cell growth inhibitor whilst enabling IgG synthesis to continue. Resveratrol also is an effective anti-oxidant making it an attractive candidate for further work. The flavan-3-ols and luteolin are potential candidates for causing cell growth inhibition whilst not supressing IgG synthesis and also providing a degree of anti-oxidant protection.

Author Contributions: L.T.U. conducted the research work and wrote the early draft of the manuscript. D.C.J. advised on the project strategy and supervised the cell culture work conducted by L.T.U. at the University of Sheffield. T.N. supervised the analytical work at Fujifilm Diosynth Biotechnologies, co-applicant on the ICASE proposal that funded the work and provided input from a biotechnology industry perspective. R.J.F., L.T.U.’s primary supervisor, and wrote the later drafts of the manuscript. Funding: The research was funded by the BBSRC CASE studentship (BB/K021036/1) and with contributions by Fujifilm Diosynth Biotechnologies. Conflicts of Interest: The authors declare no conflict of interest. Antioxidants 2019, 8, 159 12 of 18 Antioxidants 2019, 8, x FOR PEER REVIEW 12 of 18

AppendixAntioxidants 2019 A , 8, x FOR PEER REVIEW 12 of 18

Figure A1. A: Effect of different epicatechin (EC) concentrations on viable cell density (VCD) (blue) Figureand viability A1. (A (yellow).)Effect of B di: effectfferent of epicatechin different epicatechin (EC) concentrations concentrations on viable (10–100 cell μ densityM) IgG concentration (VCD) (blue) andFigure(grey) viability andA1. relativeA (yellow).: Effect IgG of (B different)production effect of epicatechin di ff(qP)erent (orange) epicatechin (EC) concentrationsfor CHO concentrations cells grownon viable (10–100 in 24-wellcellµM) density IgG plates. concentration (VCD) Each (blue) data (grey)andpoint viability andrepresents relative (yellow). the IgG average productionB: effect fr omof (qP)different a quadruplicate, (orange) epicatechin for CHO and concentrations cellsthe e grownrror bars in (10–100 24-wellcorrespond plates.μM) toIgG Each±SD. concentration data* p-value point < represents(grey) and the relative average IgG from production a quadruplicate, (qP) (orange) and the for error CHO bars cells correspond grown in to 24-wellSD. * pplates.-value Each< 0.05 data for 0.05 for the treatment compared to the control. ± thepoint treatment represents compared the average to the fr control.om a quadruplicate, and the error bars correspond to ±SD. * p-value < 0.05 for the treatment compared to the control.

Figure A2. (A)Effect of different gallocatechin gallate (GCG) concentrations on viable cell density (VCD) Figure A2. A: Effect of different gallocatechin gallate (GCG) concentrations on viable cell density (blue)(VCD) and (blue) viability and (yellow).viability ((yellow).B) effect ofB di: effectfferent of GCG different concentrations GCG concentrations (10–100 µM) IgG(10–100 concentration μM) IgG (grey)concentrationFigure and A2. relative A: (grey)Effect IgG and of production different relative IgG (qP)gallocatechin production (orange) for gallate (qP) CHO (orange) (GCG) cells grown forconcentrations CHO in 24-well cells grown plates. on viable in Each 24-well cell data density plates. point represents(VCD) (blue) the averageand viability from a (yellow). quadruplicate, B: effect and theof different error bars GCG correspond concentrations to SD. * (10–100p-value <μ0.05M) IgG for Each data point represents the average from a quadruplicate, and the error bars± correspond to ±SD. * thepconcentration-value treatment < 0.05 compared (grey)for the and treatment to relative the control. compared IgG production to the control.(qP) (orange) for CHO cells grown in 24-well plates. Each data point represents the average from a quadruplicate, and the error bars correspond to ±SD. * p-value < 0.05 for the treatment compared to the control. Antioxidants 2019, 8, x FOR PEER REVIEW 13 of 18 Antioxidants 2019, 8, 159 13 of 18 Antioxidants 2019, 8, x FOR PEER REVIEW 13 of 18

Figure A3. A: Effect of different curcumin (Crc) concentrations on viable cell density (VCD) (blue) and viability (yellow). B: effect of different curcumin concentrations (10–100 μM) IgG concentration FigureFigure A3.A3. (A:)E Effectffect of didifferentfferent curcumincurcumin (Crc)(Crc) concentrationsconcentrations onon viableviable cellcell densitydensity (VCD)(VCD) (blue)(blue) (grey) and relative IgG production (qP) (orange) for CHO cells grown in 24-well plates. Each data andand viabilityviability (yellow).(yellow). (B:) effect effect of didifferentfferent curcumincurcumin concentrationsconcentrations (10–100(10–100 µμM)M) IgGIgG concentrationconcentration point represents the average from a quadruplicate, and the error bars correspond to ±SD. * p-value < (grey)(grey) and and relative relative IgG IgG production production (qP) (qP) (orange) (orange) for CHOfor CHO cells cells grown grown in 24-well in 24-well plates. plates. Each dataEach point data represents0.05 for the the treatment average fromcompared a quadruplicate, to the control. and the error bars correspond to SD. * p-value < 0.05 for point represents the average from a quadruplicate, and the error bars correspond± to ±SD. * p-value < the0.05 treatment for the treatment compared compared to the control. to the control.

Figure A4. (A)Effect of different caffeic acid phenethyl ester (CAPE) concentrations on viable cell densityFigure (VCD)A4. A: (blue)Effect andof different viability caffeic (yellow). acid (B phenethyl) effect of di esterfferent (CAPE) CAPE concentrations concentrations on (10–100 viableµ M)cell density (VCD) (blue) and viability (yellow). B: effect of different CAPE concentrations (10–100 μM) IgGFigure concentration A4. A: Effect (grey) of different and relative caffeic IgG acid production phenethyl (qP) ester (orange) (CAPE) for concentrations CHO cells grown on inviable 24-well cell IgG concentration (grey) and relative IgG production (qP) (orange) for CHO cells grown in 24-well plates.density Each (VCD) data (blue) point and represents viability the (yellow). average B from: effect a quadruplicate, of different CAPE and theconcentrations error bars correspond (10–100 μM) to plates.SD. * p Each-value data< 0.05 point for represents the treatment the average compared from to thea quadruplicate, control. and the error bars correspond to ±IgG concentration (grey) and relative IgG production (qP) (orange) for CHO cells grown in 24-well ±SD. * p-value < 0.05 for the treatment compared to the control. plates. Each data point represents the average from a quadruplicate, and the error bars correspond to ±SD. * p-value < 0.05 for the treatment compared to the control. AntioxidantsAntioxidants 20192019, 8,, x8 ,FOR 159 PEER REVIEW 14 of14 18 of 18 Antioxidants 2019, 8, x FOR PEER REVIEW 14 of 18

FigureFigure A5.A5. (A:)E Effectffect of didifferentfferent piceidpiceid (Pcd)(Pcd) concentrationsconcentrations onon viableviable cellcell densitydensity (VCD)(VCD) (blue)(blue) andand Figureviabilityviability A5. (yellow).A(yellow).: Effect (ofBB:) differenteffect effect of different di piceidfferent (Pcd) piceid concentrat concentrationsconcentrationsions on (10–100 viable µμM) M)cell IgGIgG density concentrationconcentration (VCD) (blue) (grey)(grey) and and relative IgG production (qP) (orange) for CHO cells grown in 24-well plates. Each data point viabilityand relative (yellow). IgG B production: effect of (qP)different (orange) piceid for concentrations CHO cells grown (10–100 in 24-well μM) plates. IgG concentration Each data point (grey) represents the average from a quadruplicate, and the error bars correspond to ±SD. * p-value < 0.05 andrepresents relative theIgG average production from a(qP) quadruplicate, (orange) for and CHO the error cells bars grown correspond in 24-well to SD. plates. * p-value Each< 0.05data for point for the treatment compared to the control. ± representsthe treatment the average compared from to the a quadruplicate, control. and the error bars correspond to ±SD. * p-value < 0.05 for the treatment compared to the control. AppendixAppendix BB

Appendix B

Figure A6. The structures of the “chemical chaperones” 4-phenylbutyrate, the osmolyte glycine betaine Figure B1. The structures of the “chemical chaperones” 4-phenylbutyrate, the osmolyte glycine and the stimulant caffeine. betaine and the stimulant caffeine.

Figure B1. The structures of the “chemical chaperones” 4-phenylbutyrate, the osmolyte glycine betaine and the stimulant caffeine. Antioxidants 2019, 8, x FOR PEER REVIEW 15 of 18

AntioxidantsAntioxidants2019 2019, ,8 8,, 159 x FOR PEER REVIEW 1515 ofof 1818

Figure B2. A: Effect of different 4-phenylbutyrate (4-PBA) concentrations on viable cell density (VCD) (blue) and viability (yellow). B: effect of different α-tocopherol concentrations (10–100 μM) IgG Figureconcentration A7. A (grey) and relative IgG production (qP) (orange) for CHO cells grown in 24-well plates. Figure B2. (A: )EEffectffect of of different different 4-phenylbutyrate 4-phenylbutyrate (4-PBA) concentrations on viable cellcell density (VCD)(VCD) (blue)Each data and point viability represents (yellow). the (averageB) effect from of di aff quaderentruplicate,α-tocopherol and the concentrations error bars correspond (10–100 µ toM) ±SD. IgG * (blue) and viability (yellow). B: effect of different α-tocopherol concentrations (10–100 μM) IgG concentrationp-value < 0.05 (grey)for the and treatment relative compared IgG production to the (qP)control. (orange) for CHO cells grown in 24-well plates. concentration (grey) and relative IgG production (qP) (orange) for CHO cells grown in 24-well plates. Each data point represents the average from a quadruplicate, and the error bars correspond to SD. Each data point represents the average from a quadruplicate, and the error bars correspond to ±SD.± * * p-value < 0.05 for the treatment compared to the control. p-value < 0.05 for the treatment compared to the control.

Figure A8. (A)Effect of different glycine betaine (BTN) concentrations on viable cell density (VCD) Figure B3. A: Effect of different glycine betaine (BTN) concentrations on viable cell density (VCD) (blue) and viability (yellow). (B) effect of different α-tocopherol concentrations (10–100 µM) IgG (blue) and viability (yellow). B: effect of different α-tocopherol concentrations (10–100 μM) IgG concentration (grey) and relative IgG production (qP) (orange) for CHO cells grown in 24-well plates. concentration (grey) and relative IgG production (qP) (orange) for CHO cells grown in 24-well plates. EachFigure data B3. point A: Effect represents of different the average glycine from betaine a quadruplicate, (BTN) concentrations and the error on viable bars correspond cell density to (VCD)SD. ± *(blue)p-value and< 0.05 viability for the (yellow). treatment B: compared effect of todifferent the control. α-tocopherol concentrations (10–100 μM) IgG concentration (grey) and relative IgG production (qP) (orange) for CHO cells grown in 24-well plates. Antioxidants 2019, 8, x FOR PEER REVIEW 16 of 18

AntioxidantsEach 2019data, 8point, 159 represents the average from a quadruplicate, and the error bars correspond to ±SD.16 * of 18 p-value < 0.05 for the treatment compared to the control.

Figure A9. A Figure B4. (A:)E Effectffect of of different different caffeine caffeine (Cff) (Cff) concentrationsconcentrations on viableviable cellcell densitydensity (VCD)(VCD) (blue)(blue) andand viability (yellow). (B) effect of different α-tocopherol concentrations (10–100 µM) IgG concentration viability (yellow). B: effect of different α-tocopherol concentrations (10–100 μM) IgG concentration (grey) and relative IgG production (qP) (orange) for CHO cells grown in 24-well plates. Each data point (grey) and relative IgG production (qP) (orange) for CHO cells grown in 24-well plates. Each data represents the average from a quadruplicate, and the error bars correspond to SD. * p-value < 0.05 for point represents the average from a quadruplicate, and the error bars correspond± to ±SD. * p-value < the treatment compared to the control. 0.05 for the treatment compared to the control. References References 1. Keeley, B. Industrialization of mAb production technology. The bioprocessing industry at a crossroads. mAbs 1. Keeley, B. Industrialization of mAb production technology. The bioprocessing industry at a crossroads. 2009, 1, 443–452. [CrossRef] mAbs 2009, 1, 443–452. https://doi.org/10.4161/mabs.1.5.9448 2. Johari, Y.B.; Estes, S.D.; Alves, C.S.; Sinacore, M.S.; James, D.C. Integrated cell and process engineering for 2. Johari, Y.B.; Estes, S.D.; Alves, C.S.; Sinacore, M.S.; James, D.C., Integrated cell and process engineering for improved transient production of a “difficult-to-express“ fusion protein by CHO cells. Biotechnol. Bioeng. improved transient production of a ‘‘difficult-to-express‘‘ fusion protein by CHO cells. Biotechnol. Bioeng. 2015, 112, 2527–2542. [CrossRef] 2015, 112, 2527–2542. https://doi.org/10.1002/bit.25687 3. Trummer, E.; Fauland, K.; Seidinger, S.; Schriebl, K.; Lattenmayer, C.; Kunert, R.; Vorauer-Uhl, K.; Weik, R.; 3. Trummer, E.; Fauland, K.; Seidinger, S.; Schriebl, K.; Lattenmayer, C.; Kunert, R.; Vorauer-Uhl, K.; Weik, Borth, N.; Katinger, H.; et al. Process parameter shifting: Part I. Effect of DOT, pH, and temperature on the R.; Borth, N.; Katinger, H.; Müller, D. Process parameter shifting: Part I. Effect of DOT, pH, and temperature performance of Epo-Fc expressing CHO cells cultivated in controlled batch bioreactors. Biotechnol. Bioeng. on the performance of Epo-Fc expressing CHO cells cultivated in controlled batch bioreactors. Biotechnol. 2006, 94, 1033–1044. [CrossRef] Bioeng. 2006, 94, 1033–1044. https://doi.org/10.1002/bit.21013 4. Kou, T.C.; Fan, L.; Zhou, Y.; Ye, Z.Y.; Liu, X.P.; Zhao, L.; Tan, W.S. Detailed understanding of enhanced 4. Kou, T.C.; Fan, L.; Zhou, Y.; Ye, Z.Y.; Liu, X.P.; Zhao, L.; Tan, W.S. Detailed understanding of enhanced specific productivity in Chinese hamster ovary cells at low culture temperature. J. Biosci. Bioeng. 2011, 111, specific productivity in Chinese hamster ovary cells at low culture temperature. J. Biosci. Bioeng. 2011, 111, 365–369. [CrossRef][PubMed] 365–369. https://doi.org/10.1016/j.jbiosc.2010.11.016 5. Gomez, N.; Subramanian, J.; Ouyang, J.; Nguyen, M.D.H.; Hutchinson, M.; Sharma, V.K.; Lin, A.A.; Yuk, I.H. 5. Gomez, N.; Subramanian, J.; Ouyang, J.; Nguyen, M.D.H.; Hutchinson, M.; Sharma, V.K.; Lin, A.A.; Yuk, Culture temperature modulates aggregation of recombinant antibody in CHO cells. Biotechnol. Bioeng. 2012, I.H. Culture temperature modulates aggregation of recombinant antibody in CHO cells. Biotechnol. Bioeng. 109, 125–136. [CrossRef] 2012, 109, 125–136. https://doi.org/10.1002/bit.23288 6. Chen, Z.L.; Wu, B.C.; Liu, H.; Liu, X.M.; Huang, P.T. Temperature shift as a process optimization step for the 6. Chen, Z.L.; Wu, B.C.; Liu, H.; Liu, X.M.; Huang, P.T. Temperature shift as a process optimization step for production of pro-urokinase by a recombinant Chinese hamster ovary cell line in high-density perfusion the production of pro-urokinase by a recombinant Chinese hamster ovary cell line in high-density culture. J. Biosci. Bioeng. 2004, 97, 239–243. [CrossRef] perfusion culture. J. Biosci. Bioeng. 2004, 97, 239–243. https://doi.org/10.1016/S1389-1723(04)70198-X 7. Kim, N.S.; Lee, G.M. Response of recombinant Chinese hamster ovary cells to hyperosmotic pressure: Effect 7. Kim, N.S.; Lee, G.M. Response of recombinant Chinese hamster ovary cells to hyperosmotic pressure: effect of Bcl-2 overexpression. J. Biotechnol. 2002, 95, 237–248. [CrossRef] of Bcl-2 overexpression. J. Biotechnol. 2002, 95, 237–248. https://doi.org/10.1016/S0168-1656(02)00011-1 8. Ryu, J.S.; Kim, T.K.; Chung, J.Y.; Lee, G.M. Osmoprotective effect of glycine betaine on foreign protein

8. productionRyu, J.S.; Kim, in hyperosmotic T.K.; Chung, recombinantJ.Y.; Lee, G.M. Chinese Osmoprotective hamster ovaryeffect cellof glycine cultures betaine differs on among foreign cell protein lines. Biotechnol.production Bioeng. in hyperosmotic2000, 70, 167–175. recombinant [CrossRef Chinese] hamster ovary cell cultures differs among cell lines. Antioxidants 2019, 8, 159 17 of 18

9. Avello, V.; Tapia, B.; Vergara, M.; Acevedo, C.; Berrios, J.; Reyes, J.G.; Altamirano, C. Impact of sodium butyrate and mild hypothermia on metabolic and physiological behaviour of CHO TF 70R cells. Electron. J. Biotechnol. 2017, 27, 55–62. [CrossRef] 10. Coronel, J.; Klausing, S.; Heinrich, C.; Noll, T.; Figueredo-Cardero, A.; Castilho, L.R. Valeric acid supplementation combined to mild hypothermia increases productivity in CHO cell cultivations. Biochem. Eng. J. 2016, 114, 101–109. [CrossRef] 11. Paul, A.J.; Handrick, R.; Ebert, S.; Hesse, F. Identification of process conditions influencing protein aggregation in Chinese hamster ovary cell culture. Biotechnol. Bioeng. 2018, 115, 1173–1185. [CrossRef][PubMed] 12. Yam, G.H.-F.; Gaplovska-Kysela, K.; Zuber, C.; Roth, J. Sodium 4-phenylbutyrate acts as a chemical chaperone on misfolded myocilin to rescue cells from endoplasmic reticulum stress and apoptosis. Investig. Ophth. Vis. Sci. 2007, 48, 1683–1690. [CrossRef] 13. Yang, W.C.; Lu, J.; Nguyen, N.B.; Zhang, A.; Healy, N.V.; Kshirsagar, R.; Ryll, T.; Huang, Y.-M. Addition of valproic acid to CHO cell fed-batch cultures improves monoclonal antibody titers. Mol. Biotechnol. 2014, 56, 421–428. [CrossRef] 14. Hossler, P.; Wang, M.; Mcdermott, S.; Racicot, C.; Chemfe, K.; Zhang, Y.; Chumsae, C.; Manuilov, A. Cell culture media supplementation of bioflavonoids for the targeted reduction of acidic species charge variants on recombinant therapeutic proteins. Biotechnol. Prog. 2015, 31, 1039–1052. [CrossRef] 15. Tian, J. Use of Phenolic Antioxidants in Cell Culture for the Production of Proteins. U.S. Patent WO 2016130734, 2016. 16. Chung, S.; Tian, J.; Tan, Z.; Chen, J.; Zhang, N.; Huang, Y.; Vandermark, E.; Lee, J.; Borys, M.; Li, Z.J. Modulating cell culture oxidative stress reduces protein glycation and acidic charge variant formation. mAbs 2019, 11, 205–216. [CrossRef] 17. Jovanovic, S.V.; Steenken, S.; Tosic, M.; Marjanovic, B.; Simic, M.G. as antioxidants. J. Am. Chem. Soc. 1994, 116, 4846–4851. [CrossRef] 18. Thorpe, A.S.; Thelen, G.C.; Diaconu, A.; Callaway, R.M. Root exudate is allelopathic in invaded community but not in native community: Field evidence for the novel weapons hypothesis. J. Ecol. 2009, 97, 641–645. [CrossRef] 19. Ozdal, T.; Capanoglu, E.; Altay, F. A review on protein–phenolic interactions and associated changes. Food Res. Int. 2013, 51, 954–970. [CrossRef] 20. Kim, H.-S.; Quon, M.J.; Kim, J. New insights into the mechanisms of polyphenols beyond antioxidant properties; lessons from the , epigallocatechin 3-gallate. Redox Biol. 2014, 2, 187–195. [CrossRef] 21. Renard, J.M.; Spagnoli, R.; Mazier, C.; Salles, M.F.; Mandine, E. Evidence that monoclonal antibody production kinetics is related to the integral of the viable cells curve in batch systems. Biotechnol. Lett. 1988, 10, 91–96. [CrossRef] 22. Hothorn, T.; Hornik, K.; van de Wiel, M.A.; Zeileis, A. Implementing a class of permutation tests: The coin package. J. Stat. Softw. 2008, 28, 1–23. [CrossRef] 23. Baek, E.; Kim, C.L.; Kim, M.G.; Lee, J.S.; Lee, G.M. Chemical inhibition of autophagy: Examining its potential to increase the specific productivity of recombinant CHO cell lines. Biotechnol. Bioeng. 2016, 113, 1953–1961. [CrossRef] 24. Leopoldini, M.; Russo, N.; Toscano, M. The molecular basis of working mechanism of natural polyphenolic antioxidants. Food Chem. 2011, 125, 288–306. [CrossRef] 25. Yun, Z.-Y.; Takagi, M.; Yoshida, T. Effect of antioxidants on the apoptosis of CHO cells and production of tissue plasminogen activator in suspension culture. J. Biosci. Bioeng. 2001, 91, 581–585. [CrossRef] 26. Yun, Z.-Y.; Takagi, M.; Yoshida, T. Combined addition of glutathione and iron chelators for decrease of intracellular level of reactive oxygen species and death of Chinese hamster ovary cells. J. Biosci. Bioeng. 2003, 95, 124–127. [CrossRef] 27. Lombardo, E.; Sabellico, C.; Hájek, J.; Staˇnková, V.; Filipský, T.; Balducci, V.; De Vito, P.; Leone, S.; Bavavea, E.I.; Proietti Silvestri, I.; et al. Protection of cells against oxidative stress by nanomolar levels of hydroxyflavones indicates a new type of intracellular antioxidant mechanism. PLoS ONE 2013, 8, e60796. [CrossRef][PubMed] 28. Barontini, M.; Bernini, R.; Carastro, I.; Gentilib, P.; Romani, A. Synthesis and DPPH radical scavenging activity of novel compounds obtained from and derivatives. New J. Chem. 2014, 38, 809–816. [CrossRef] Antioxidants 2019, 8, 159 18 of 18

29. Bernini, R.; Carastro, I.; Palmini, G.; Tanini, A.; Zonefrati, R.; Pinelli, P.; Brandi, M.L.; Romani, A.J. Lipophilization of -Enriched Fractions from Olea europaea L. Byproducts and Evaluation of the in Vitro Effects on a Model of Colorectal Cancer Cells. Agric. Food Chem. 2017, 65, 6506–6512. [CrossRef] 30. McRae, J.M.; Falconer, R.J.; Kennedy, J.A. Thermodynamics of grape and wine interaction with polyproline: implications for red wine astringency. J. Agric. Food Chem. 2010, 58, 12510–12518. [CrossRef] 31. Scollary, G.R.; Pasti, G.; Kallay, M.; Blackman, J.; Clark, A.C. Astringency response of red wines: Potential role of molecular assembly. Trends Food Sci. Tech. 2012, 27, 25–36. [CrossRef] 32. Poncet-Legrand, C.; Gautier, C.; Cheynier, V.;Imberty,A. Interactions between flavan-3-ols and poly(L-proline) studied by isothermal titration calorimetry: Effect of the tannin structure. J. Agric. Food Chem. 2007, 55, 9235–9240. [CrossRef][PubMed] 33. Harbertson, J.F.; Kilmister, R.L.; Kelm, M.A.; Downey, M.O. Impact of size as individual and mixed polymers on bovine serum albumin precipitation. Food Chem. 2014, 160, 16–21. [CrossRef] [PubMed] 34. Kilmister, R.L.; Faulkner, P.; Downey, M.O.; Darby, S.J.; Falconer, R.J. The complexity of condensed tannin binding to bovine serum albumin—An isothermal titration calorimetry study. Food Chem. 2016, 190, 173–178. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).