Enhancement of transient erythropoietin protein expression by valproic acid in CHO-K1 suspension adapted cells

Erythropoietin (EPO) is a therapeutic protein that is widely used to increase red blood cell production in chronic kidney failure. EPO protein can be produced quickly with a transient gene expression system (TGE). However, the titer produced using TGE is usually lower than the stable gene expression system (SGE). It has been known that TGE can be improved by histone deacetylase inhibitors (iHDACs) such as valproic acid (VPA). This study was conducted to examine the VPA effect on EPO protein expression in CHO‐K1 suspension adapted cells and to find the optimum concentration of VPA on transient EPO protein production. EPO proteins was quantified using the enzyme‐linked immunosorbent assay (ELISA) method. The optimization of VPA concentrations showed that VPA increased the EPO protein yield by up to 2‐fold in transient EPO production, and the optimum concentration of VPA was 4 mM. VPA optimization was very helpful to obtain the maximum increase in the transiently expressed protein. Furthermore, this study can be used as a model to produce EPO proteins or other recombinant proteins rapidly with TGE of CHO‐K1 suspension adapted cells..


Introduction
EPO is a glycoprotein that has hormone characteristic with 30,400 Dalton molecular weight (Yin and Blanchard 2000). This protein is pivotal for stimulating the produc tion of erythroid cells (Fried 2009). Many patients with kidney failure received this protein to increase red blood cell during hemodialysis. EPO treatment is also useful for patients with chemotherapy to reduce occurrence of ane mia (Ashley et al. 2002; Stein 2003; Casadevall et al. 2004; Fried 2009.
One of the main problems with cell line development for protein production using SGE system is that it is very laborious (Pham et al. 2006; Backliwal et al. 2008). On the other hand, TGE can be used as an alternative to produce protein in a more effective and convenient way. Particu larly, its production time which is relatively fast, only re quired about one to three weeks (Pham et al. 2006; Baldi et al. 2007; Carpentier et al. 2007; Geisse 2009. While the SGE technique can take up to several months. Previous studies have reported that TGE production can be improved by HDACs inhibitor, such as sodium butyrate and VPA. Those compounds were known as booster agents in TGE to produce recombinant antibody in HEK293 and CHO DG44 mammalian cells (Allen et al. 2008; Backliwal et al. 2008; Wulhfard et al. 2008. Since The Food and Drug Administration (FDA) has approved the use of VPA in human (Backliwal et al. 2008), hence the use VPA in protein production as an alternative approach can be considered, especially in TGE system. However, the magnitude of the response of VPA may be diverse in different cells. VPA increased transient antihuman RhD antibody yield about 4fold approximately in HEK293E and 1.5fold in CHO DG44 (Backliwal et al. 2008). In re lation to the ability of VPA to increase protein expression, in this current study we attempted to increase the expres sion of EPO by using VPA and optimize its concentration in TGE system in CHOK1 suspension adapted cells.

Cell culture
The CHOK1 cells were used as the mammalian host ob tained from Prof. Masashi Kawaichi, NAIST, Japan. The cells were adapted into a suspension culture by cultur ing the cells in the CHOSSFMII (Invitrogen, CA, USA), serumfree medium (Wisnuwardhani et al. 2017). CHO K1 suspension cells were routinely cultured in 125mL at 37°C and 5% CO 2 .

Transient transfection in CHO-K1 suspension cells
CHOK1 suspension cells were transfected with pJ603 EPO plasmid containing 2 additional N links (Santoso et al. 2014). The cells were transfected using lipofectamin 2000 according to the manufacturer's protocol (Invitrogen, CA, USA). One day before transfection CHOK1 suspen sion cell was passaged in 20 mL CHOSSFMII serum free medium with final density of 1x10 6 cells/ml in an 125 mL Erlenmeyer flask. DNAlipid solution with ra tio 1:3 (2.5 μg DNA: 7.5 μl lipofectamin) was prepared in 1 mL OptiMEM medium. The DNAlipid mixture was vortexed, incubated at room temperature for 20 min and added to the cells in a dropwise manner. Following in cubation at 37°C for 6 h, the cell was harvested at 3000 rpm for 5 min, resuspended in a sterile Erlenmeyer flask. The cell was then incubated on an orbital shaker at 37°C and 5% CO 2 for overnight. EPO protein was analyzed by western blotting method.

Protein characterization of EPO by N-GlycosidaseF
The presence of 5 Nlinks in EPO glycoprotein was char acterized using NglycosidaseF (Roche Mannheim, Ger many) digestion. The EPO protein was digested using N glycosidaseF according to the manufacturer's protocol. As much as 20 µl (5 μg) of EPO protein was digested with 2 units of NglycosidaseF in 50 µL total volume in sodium phosphate buffer with 20 mM final concentration, pH 7.2. Following incubation at 37°C for 120 min, using EPO pri mary antibody, the mixture was analyzed using Western blot analysis.

Optimation study of valproic acid (VPA) treatment
Optimization of VPA concentrations followed the method of (Wulhfard et al. 2010), with modifications in VPA concentrations. VPA salt (1M) (Sigma-Aldrich GbmH, Switzerland) was prepared in water and filter sterilized us ing a 0.22 μm filter (Millipore, Switzerland). VPA was added to the transfected cells at several concentrations (0 mM, 1.25 mM, 2.5 mM, 3 mM, 3.75 mM, 4 mM, 5 mM or 10 mM). The cells were incubated over night at 37°C and 5% CO 2 and medium samples were collected for quantifi cation. EPO protein was analyzed by slot blotting method and total protein was quantified by Coomassie (Bradford) Protein Assay method (Thermo Scientific, USA).

VPA treatment in transiently transfected CHO-K1 suspension cells
VPA treatment in transiently transfected cells followed the method of (Sinhadri 2009), with modifications in mam malian cell type, length of incubation time, and VPA con centration. VPA was added to the transfected cells at 4 mM VPA concentrations. The cell was then incubated on an or bital shaker at 37°C and 5% CO 2 for 12 d and supernatant was collected. The produced EPO protein was determined by EPO ELISA kit (Roche Mannheim, Germany). The control cell was untreated with VPA. Cell number and via bility were determined with the Trypan Blue (Gibco, USA) exclusion method.

Slot blotting
The protein was blotted to nitrocellulose membranes (GE Healthcare, USA) by slot blot apparatus. EPO protein was detected using antihEPO antibody (Sigma, St. Louis, USA), then was conjugated with the anti rabbit IgG (whole molecule)alkaline phosphatase anti body (Promega, Madison, WI, USA

Protein characterization of EPO by N GlycosidaseF
To prove that EPO protein has been obtained, the ex pressed EPO protein was treated with NGlycosidaseF to release the Nlinked carbohydrates. Data showed ( Figure  1) that the rhEPO protein bands was ranged around 3745 kDa which corresponds to the theoretical molecular mass for the EPO that the exact molecular weight highly de pends on the degree of glycosylation. Lane 2 it was rhEPO without any treatments has ±37 kDa molecular weight, while on lane 3 which was rhEPO with NglycosidaseF enzyme treatment has ±22 kDa molecular weight.

Optimization of VPA concentrations
To find the best concentration of VPA, in this study, sev eral VPA concentrations were performed for optimization. Culture supernatant was analyzed and the total protein was measured using Bradford method. The result showed that 4 mM of VPA gave the highest hEPO expression (Fig  ure 2a). Accordingly, the total protein concentration also showed the highest expression at 4 mM of VPA ( Figure  2b).

The effect of VPA to yield, cell density, viability cell CHOK1
Simultaneously, after six hours of transfection with pJ603 EPO plasmid, VPA was added to the cell culture with the final concentration of 4 mM. Samples were collected for every 3 d for 12 d period. Figure 3 showed that the re sult of VPA affected the protein yield, cell density and vi ability. The highest yield of EPO was obtained at day 12 which was about 15 mg/L. Meanwhile, the yield without VPA, at day 12, reached only 7.5 mg/L (Figure 3a). This data suggested that VPA can raise of EPO protein yield up to twofold compared to control. While VPA had pos itive effects on the yield, the addition of VPA to the cell culture decreased cell density and viability (Figure 3b and 3c). Cell growth of untreated cell increased strongly and reached maximum at day 4. This increase was then fol lowed by continual decrease until day 12. Profile of cell growth of VPA treated cell was almost similar to the un treated one except that the peak at day 4 was much lower (Figure 3b). Inversely proportional to the cell density pro file, cell viability of treated and untreated ones, continu ously decreases until it reaches its lowest point on day 12 which was about 20%.

Discussion
EPO protein that extracellularly expressed in CHOSFM media was successfully detected by antiEPO at a size of about 37 kDa. Furthermore, in order to find out that the EPO protein produced in CHOK1 cells is the protein of our interest, characterization was carried out using N glycosidase. EPO protein digestion with Nglycosidase F will remove most of the carbohydrates from the protein (Egrie and Browne 2001) and reduce molecular weight by around 15 kDa. The results seen in the Figure 1 showed a decrease of EPO molecular weight when treated with gly cosidase enzymes. Decreasing molecular weight of the protein to approximately 22 kDa indicates the existence of asparagine linked glycan chains. Thus, it is proven that this protein is an EPO protein that contains glycan chains. Technological approach of transient gene expression be comes more interesting to produce protein recombinant rapidly in mammalian cells, especially CHO and HEK293 cells (Pham et al. 2006; Baldi et al. 2007. Previous studies had reported that transiently trans fected CHO cell treated with VPA can enhance recom binant protein yield. However, the growth of CHO cell has also been reported to be inhibited by VPA addition (Backliwal et al. 2008; Wulhfard et al. 2010. Based on that background, the application of VPA on transient ex pression for enhancement of protein expression has to be carefully optimized. In this current optimization study, it was found that 4 mM of VPA concentration gave the high est yield of EPO (Figure 2a). This result corresponds to the profile of total protein concentration where the high est yield was also found to be at 4 mM of VPA ( Figure  2b). After reaching the optimum concentration (4 mM of VPA), the yield decreased which, in this case, presumably caused by the blocking of cell growth by VPA. Working on HEK293 cell expressing TNFR:Fc fusion protein, Sin hadri (2009) reported that the 3.75 mM VPA exhibited the highest amount of IgG. Meanwhile, Backliwal et al. (2008) have found that the best concentration of VPA to produce antiRhD antibody was 4 mM in HEK293 cell and 500 μM in CHO DG44 cell.
Overall, the data gained in this study exhibited that The effect of VPA on protein yield, cell density and viability. a. EPO protein yield was measured using ELISA. b. Cell density was revealed between VPA and non-VPA treatment on CHO-K1 suspension cell. c. Cell viability was showed between VPA and non-VPA treatment on CHO-K1 suspension cell.
VPA has an influence on TGE where the EPO protein pro duction was increased up to twofold after adding 4 mM of VPA (Figure 3a). Previous study confirms that VPA enhanced monoclonal antibody recombinant yield about 4 fold on HEK293 cell and 1.5 fold on CHO DG44 (Back liwal et al. 2008). Fan et al. (2005) stated that the im provement of this yield may have to do with the ability of VPA to block class I and II histone deasetylases (HDACs), and induce histone acetylation both in vitro and in vivo (Michaelis et al. 2004; Marchion et al. 2005. (Watson et al. 2014) reported that the removal of pos itive charge of histones by acetylation lowers the inter action of the N termini of histones with the negatively charged phosphate groups of DNA. As a result, the struc ture of chromatin becomes more relaxed and promotes greater levels of gene transcription. This relaxation sub sequently will make it easier for transcription site to be at tached by transcription initiation factors and subsequently the gene of our interest can be highly expressed to be trans lated into protein (Haberland et al. 2009).
In line with previous studies, our data shown that VPA has a strong influence of TGE, in this case, the transient expression of EPO in CHOK1 suspension adapted cells increase by up to 2 folds. This study also demonstrated that the optimization of VPA concentration was very helpful to obtain the maximum increase in the expression of our protein of interest.
Based on the research conducted by Sami et al. (2008), VPA can inhibit the cell cycle in the G1/G0 phase on HeLa cells. Inhibition of the cell cycle can affect the stability of transient genes in cells. Instability of gene on cells cause the lost of gene. According to a review by Lecharduer and Lechardeur and Lukacs (2006), the plasmid DNA will lost during the subsequent cell divisions in transient transfec tion processes. The loss of plasmid DNA in cells makes the target protein unable to be expressed. Therefore, the low expression of the EPO protein in CHOK1 cells with out VPA may be due to cell division causing the loss of EPO gene, but this still needs further study. While the CHOK1 cells were treated with VPA addition exhibited high protein expression although low viability and den sity of cells. This is probably due to the presence of VPA that can make the cell remains in the G1/G0 phase for a while. In this phase, cells do not carry out cell division processes but remain actively expressing proteins until cell death (apoptosis) occurs.

Conclusions
Valproic acid increases EPO protein titer produced tran siently in CHOK1 suspension adapted cells by up to two fold compared to the one without the addition of VPA; and 4 mM of VPA was found to be the optimum con centration. VPA addition on transient recombinant pro tein expression is highly efficient for obtaining high pro tein yield. Therefore, this study is important as a model to produce EPO protein or another recombinant protein pro duced transiently in CHOK1 suspension adapted cells.