j ourna l ho me page: www.elsev ier .com/ locate / jb io tec
ovel human renal proximal tubular cell line for the production ofomplex proteins
ukas Fliedla, Gabriele Manharta, Florian Kasta, Hermann Katingerb, Renate Kunertb,ohannes Grillarib, Matthias Wiesera, Regina Grillari-Voglauera,b,∗
ACIB, Muthgasse 11, A-1190 Vienna, AustriaDepartment of Biotechnology, University of Natural Resources and Life Sciences Vienna, Muthgasse 18, A-1190 Vienna, Austria
r t i c l e i n f o
rticle history:eceived 19 November 2013eceived in revised form 31 January 2014ccepted 6 February 2014vailable online 16 February 2014
eywords:PTECrythropoietinlycosylationialylation
a b s t r a c t
Human host cell lines for the production of biopharmaceutical proteins are of interest due to differencesin the glycosylation patterns of human and animal cell lines. Specifically, sialylation, which has a majorimpact on half-life and immunogenicity of recombinant biopharmaceuticals, differs markedly. Here, weestablished and characterized an immortalized well documented and serum-free host cell line, RS, fromprimary human renal proximal tubular epithelial cells (RPTEC). In order to test its capacity to producecomplex glycosylated proteins, stable recombinant human erythropoietin (rhEpo) producing clones weregenerated. The clone with highest productivity, RS-1C9 was further characterized and showed stableproductivity. Biological activity was observed in in vitro assays and 28% of rhEpo glyco-isoforms producedby RS-1C9 were in range and distribution of the biological reference standard (BRP) isoform, as comparedto 11.5% of a CHO based rhEpo. Additionally, cellular �-2,6 sialylation, Galactose-alpha-1,3-galactose
soelectric focussing (alpha-Gal) and N-glycolylneuraminic acid (NeuGc) patterns compare favourably to CHO cells. Whileproductivity of RS still needs optimization, its amenability to upscaling in bioreactors, its productionof glyco-isoforms that will increase yields after down-stream processing of about 2.5 fold, presenceof sialylation and lack of Neu5Gc recommend RS as alternative human host cell line for production ofbiopharmaceuticals.
Today the recombinant production processes of complex pro-eins mainly rely on Chinese Hamster Ovary (CHO) cell lines (Wurm,004). To be fully active when applied, such complex proteinsust meet characteristics like complete and correct amino acid
ackbone, proper folding and a ‘human-like’ glycosylation patternKwaks and Otte, 2006). Although one of the benefits of mammalianell lines is that they have the ability to add glycan-structures to
protein, the repertoire of glycosidases and glycosyltransferasess species specific and even cell type specific (reviewed in (Bergert al., 2012)). The resulting incomplete and non-authentic glyco-
ylation patterns can subsequently result in reduced stability andalf-life of the glycoprotein or it might trigger an immune response
n patients (Cai et al., 2012 and reviewed in Berger et al., 2012).
∗ Corresponding author at: Department of Biotechnology, University of Naturalesources and Life Sciences Vienna, Muthgasse 18, A-1190 Vienna, Austria.el.: +43 1476546231.
Moreover, activity as well as signal transduction and cell adhesionof complex proteins (reviewed in Berger et al., 2012) depend onthe glycosylation pattern of the protein. Especially, the last step ofglycosylation, the terminal linkage of sialic acids on glycostructureshas major effects on the in vivo half-life (Ashwell and Harford,1982; Fukusa et al., 1989). Although CHO cells efficiently addcarbohydrate chains to the proteins backbone (Sasaki et al., 1987),differences were found in the glycan structures, which correlate tothe proteins’ biological activity (Takeuchi et al., 1989). One differ-ence in terms of sialylation is that proteins produced by CHO cells,only contain �-2,3 linkages between galactose and the terminalsialic acid residue of N-linked oligosaccharides due to their lack of�-2,6 sialyltransferase. In order to bypass this limitation, BHK andCHO cells were engineered to express the human �-2,6 sialyltrans-ferase (Grabenhhorst et al., 1997; Schlenke et al., 1997; Zhang et al.,1998). In addition, CHO cells in general have a reduced potentialof sialylation compared to humans (Brooks, 2004). In order to
improve this phenomenon as well as productivity, the 30Kc19 geneof the silkworm (Bombyx mori) has been introduced into recom-binant human erythropoietin (rhEpo) producing CHO cells (Parket al., 2012; Wang et al., 2011). Furthermore, most non human
ammalians produce Galactose-alpha-1,3-galactose (alpha-Gal)nd N-glycolylneuraminic acid (NeuGc) (Padler-Karavani andarki, 2011). These differences in glycosylation can lead to
mmunogenicity and increased clearance of the recombinantroduct (Ghaderi et al., 2010; Padler-Karavani and Varki, 2011).
On the contrary, recombinant proteins produced in human cellsre expected to perfectly match essential characteristics of the cor-esponding endogenous proteins like folding and post-translationalodifications (Swiech et al., 2012). Thus, in vitro cultivated, highly
ifferentiated normal human cells with secretory capacity repre-ent optimal expression systems for complex proteins produced byhe same cell type in vivo. However, the use of such cells for biotech-ological applications is limited by the fact that these cells can onlye propagated for a limited number of population doublings (PD)efore entering the phase of replicative senescence (Hayflick andoorhead, 1961). Therefore, the cellular life span of the cells has to
e extended as well as the maintenance of differentiated features,hich are essential for proper glycosylation. Erythropoietin (Epo),
major humoral regulator of erythropoiesis (Krantz and Jacobson,970 and reviewed in Jelkmann, 1992) represents one of the lead-
ng biopharmaceutical protein products with applications in thereatment of hypoxia and anaemia and was the first biopharma-eutical to reach blockbuster status (Wurm, 2004). Epos 165 aminocid backbone is linked to 3 N- and one O-glycan chain that makep about 40% of the proteins mass (Jelkmann, 1992). Additionally,ialic acid residues terminate glycan chains and thereby prolong theirculation time of the glycoprotein. Desialylated Epo is recognizedy galactose binding receptors of the liver, namely asialoglycopro-ein receptors, and rapidly cleared from the blood stream (Ashwellnd Harford, 1982; Fukusa et al., 1989). In vivo, Epo is producednd secreted in the kidney by renal proximal tubular epithelial cellsRPTEC) (Shanks et al., 1996). Since different glycosylation patternsre due to specialized cells, human kidney cell line endowed withharacteristics of the normal counterpart in vivo might be an idealost system to produce this complex protein with high quality.
In order to proof this hypothesis we isolated human RPTECs fromidney tissue biopsies. The cells were maintained in vitro underulture conditions lacking animal derived products under full doc-mentation and immortalization was initiated using SV40 earlyegion. After characterization of the cellular phenotype, we usedhe cell line for production of recombinant human Epo in small andarge scale and the produced protein was characterized in termsf its glycosylation pattern. Thereby, we have shown that immor-alized human RPTECs have a high potential as novel host for theroduction of complex biopharmaceuticals.
. Material and methods
.1. Cell lines
The local ethic commission approved the study and the patientsave their informed consent (Biobank Graz). Thus, the study waserformed in accordance with the Declaration of Helsinki.
nRPTEC: Normal human renal proximal tubular epithelial cellsnRPTEC) were isolated as described in detail previously (Wiesert al., 2008) and cultivated in ProxUp-Pri medium (Evercyte GmbH).
RS cells: nRPTEC were transfected with a plasmid encodingV40 early region (Voglauer et al., 2005) using Lipofectamine 2000Invitrogen) according to the manufacturers instructions and pos-tively transfected cells were selected by outgrowth of SV40 earlyegion overexpressing cell clones. These cells were cultivated asass culture designated RS cells and grown in OptiPro serum-free
edium (OptiPro SFM, Gibco) supplemented with 4 millimolar
mM) L-glutamine (Sigma Aldrich) in culture flasks (Greiner Bione) pre-coated with human collagen (Sigma Aldrich) for at least0 min prior to passaging the cells.
nology 176 (2014) 29–39
2.2. Plasmids and transfection
The plasmid pDEPT carrying the early region of SV40 (largeT/small t), including the SV40 promoter/enhancer sequences, isdescribed in detail by (Banerji et al., 1983). The plasmid pEPO wasconstructed using the pCI-neo (Promega) backbone habouring aCMV enhancer/promoter and SV40 late poly(A). The human codon-optimized human erythropoietin sequence was integrated usingthe cloning sites MluI and XbaI. Codon optimization and vectorcloning was performed by Geneart.
All transfections were done using Lipofectamine2000 (Life Tech-nologies) according to the manufacturers’ protocols.
2.3. Culture conditions
2.3.1. Cultivation in roux flasksAll cell lines used within this study were cultivated at 37 ◦C in a
humidified atmosphere containing 7% CO2. Confluent monolayerswere passaged with a split ratio of 1:2–1:4 depending on the cellline. Therefore, the cells were detached using 0.1% trypsin-0.02%EDTA (Sigma), which was inactivated by addition of soybean trypsininhibitor (1 mg/ml, Sigma), followed by a centrifugation step (5 min,170 g) and transfer of the resuspended cells to new culture vessels.
2.3.2. Cultivation in spinner flasksSpinner flask cultivation was performed as described previously
(Fliedl and Kaisermayer, 2011). Briefly, prior to cultivation glassvessel (Techne, Abington, England) were siliconised using Sigma-cote (Sigma Aldrich) to prevent adhesion of microcarrier to the glasssurface. Microcarriers (Cytotex 3; Sigma Aldrich) were washed incalcium and magnesium free-phosphate buffered saline (PBS) (PAA,Pasching, Austria) and autoclaved.
The spinner flasks were prepared 24 h before inoculation with50% of medium and the appropriate amount of carriers to reach aconcentration of 3 g/l in the final working volume of 60 ml. Flaskswere incubated at 7% CO2 for pH adjustment.
After detachment of cells, spinner flasks were inoculated with2 × 105 cells/ml. Cultivation was performed at 37 ◦C and 50 rota-tions per minute. Samples were taken daily to determine cellconcentration, cell morphology and metabolite concentrations.Cultivation medium was replaced as necessary to maintain a resid-ual glucose concentration higher than 1 g/l.
2.4. Cell counting and calculation of population doublings (PD)
For T-flask cultures cell number and viability were deter-mined using a ViCell (Beckman Coulter). To establish a growthcurve, cells were passaged twice a week and population dou-blings (PD) were calculated using the following equation: PD = 3.32(log X2 − log X1) + B. (X1 is the cell number at the beginning of theincubation time, X2 is the cell number at the end of the incubationtime, B PD at the beginning of the incubation time).
In microcarrier cultures 1 ml of suspension was used to deter-mine the total cell concentration. The carriers were left to settle,the supernatant was removed and the carriers were resuspendedin 1 ml of 0.1% crystal violet in 0.1 M citric acid. After a minimumincubation period of 1.5 h the released nuclei were counted in ahaemocytometer.
2.5. Microscopy and photography
Cell morphology was documented by photographs taken at100-fold magnification using an Olympus microscope (IMT-2) andan Olympus digital camera (XC 50). Cells on microcarriers werestained using haematoxylin solution (haematoxylin 0.9 g/l, sodium
For staining of Occludin, E-cadherin and SV40 large T/small tntigens cells were grown on chamber slides (IBIDI) to conflu-nce, fixed with 4% paraformaldehyde (Sigma Aldrich) for 10 min atoom temperature (RT) and permeabilized with 0.3% Triton X-100Sigma Aldrich) in PBS. Then, the cells were incubated over nightt 4 ◦C in the presence of the primary antibodies against OccludinZO-1) 1:100 (rabbit anti-ZO-1, Zymed 61-7300), E-Cadherin (goatnti E-Cadherin, R&D Systems AF648), or SV40 large T/small t anti-en (mouse anti-SV40 t/T antigen, Oncogen DP01-100UG) diluted:100 in PBS with 10% FCS. Secondary antibodies goat anti-rabbitgG, DyLight 488 (Jackson Immuno Research, Newmarket, UK 111-85-144), donkey anti-goat IgG, Alexa Flour 488 (Life technologies,-11055) or goat anti-mouse IgG, DyLight 488 (Jackson Immunoesearch, Newmarket, UK 115-485-146) were applied for 1 h at RT
n a 1:1000 dilution. Nuclei were counterstained with 0.1 �g/mlAPI solution (Sigma Aldrich) and slides were analyzed on a LeicaMI6000 CS confocal microscope.
For detection of Aminopeptidase N (APN) using flow cytometry,ells were harvested, blocked for 20 min in 10% FCS in PBS, resus-ended in a 1:100 dilution of mouse anti-APN antibody (Southerniotech 9556-01) followed by a washing step and incubation withhe secondary antibody (goat anti-mouse DyLight 488; Jacksonmmuno Research, Newmarket, UK 115-485-146). Cells were ana-yzed on a Gallios Flow Cytometer (Beckman Coulter) and theercentages of positive cells were evaluated using the KALUZABeckman Coulter) software.
.7. Telomeric repeat amplification protocol
Telomerase activity was performed according to already pub-ished protocols (Kim 1994, Falchetti 1998). In brief, a PCR from cellysates (TRAPeze lysis buffer, Millipore) was performed using theollowing primer pair TSmod IV 5′AATCCGTCGAGAACAGTT 3ıandxa 5′GTGTAACCCTAACCCTAACCC 3. The PCR product was loadednto a 10% denaturing polyacryamid gel (AA/BAA 19:1) and bandsere stained with SYBR-Green I (Molecular Probes) and analyzedsing a Lumi-Imager F1 (Roche).
.8. Gamma-glutamyl transferase activity
The protocol for determination of GGT activity was adapted fromMeister et al., 1981). In brief, cells were cultured in 12-well platesnd incubated with substrate (1 mM G-glutamyl-para-nitroanilide,0 mM Tris-HCl pH 8.0, and 20 mM glycyl-glycine) for 20 min atT. The enzymatic reaction was stopped by addition of 10% aceticcid (1:2). Para-nitroanilide release was determined with a Tecannfinite M200 at 405 nm. Values were normalized to cell numberss determined by ViCell. All tests were performed with biologicalnd technical triplicates. For biological replicates cells were grownnd treated separately and technical replicates were three paralleleasurements. While technical replicates were only used to eval-
ate stability of measurement, biological replicates were used tovaluate standard deviation.
.9. Hormonal assay
Cells were treated with parathyroid hormone (PTH) and vaso-
ressin (AVP) according to Wieser et al. (2008). After hormonalreatment the medium was removed and cell extracts were pre-ared using 0.5% trichloroacetic acid. The levels of cAMP weressayed with cAMP Direct BiotrakTM EIA RPN2251 (GE-Healthcare)
nology 176 (2014) 29–39 31
according to the manufacturers instructions. Values were nor-malized to cell numbers as determined by ViCell. All tests wereperformed with biological triplicates, which were used to evaluatestandard deviation.
2.10. Generation of hEpo producing clones
RS cells, grown in 6-well plates to 80-95% confluence, weretransfected with plasmid DNA containing a codon optimizedhuman Epo gene and neomycin phosphotransferase as selectionmarker using Lipofectamine 2000 (Life technologies) according tothe manufacturers instruction. In order to improve transfectionefficiency the plasmid DNA was linearized with Fast Digest restric-tion enzyme XmnI (Fermentas). 24 h post transfection, cells wereexposed to medium containing 100 �g/ml G418 (Geneticin, Invivo-gen). Clones were selected and passaged to higher cell numbersaccording to growth and productivity.
2.11. ELISA
Secreted recombinant human erythropoietin (rhEpo) was quan-tified via sandwich ELISA. Therefore, mouse anti-hEpo antibody(500 �g/ml, MAB287; R&D Systems) was used in a 1:250 dilution forcoating a 96-well Maxisorp plate (Nunc) and blocked with Tween20 (Sigma). Standard (500 ng/ml, Erythropoietin BRP, HVS0008,Lot.-no.: BP907-F67029) in duplicates and samples were diluted(1:1.5 dilution row) and applied onto the plate. For detection, apolyclonal IgG anti-hEpo antibody (MAB286; R&D systems) thatwas biotinylated previously (GE Healthcare Amersham ECL ProteinBiotinylation Module) was applied in a 1:200 dilution, followedby the addition of Streptavidin-Horseradish Peroxidase Conjugate(RPN1231-2ML, GE) in a 1:2000 dilution. The labelling reagentconsisted of 100 �l OPD (Ortho-phenylenediamine dihydrochlo-ride (Sigma)) and 6 �l of H2O2 in 10 ml dyeing buffer. 25% sulphuricacid was used as stop solution. The product titre was measuredusing an ELISA Reader (Sunrise, Tecan) at 492 nm with a referencewavelength of 620 nm and analyzed with the Magellan 4 software.
2.12. Enzymatic digestion, SDS Page and Western blots
In order to evaluate the number of N-glycan structures,supernatants of hEpo producing RS-1C9 cell line and a humanErythropoietin standard (Erythropoietin BRP, E1515000, batch 3)were treated with the enzyme PNGase F (NEB, P0704) using 0,3, 5 and 20 U/�l for 1 h. After addition of appropriate amountof 4× loading dye, enzyme treated samples were applied toa 4–15% Mini-PROTEAN TGX Precast Gel (Biorad) in a Mini-PROTEAN Tetra Cell (Biorad). 8 �l of samples and of the PAGE RulerPrestained Protein Ladder (Thermo Scientific Pierce) were loadedand run for approximately 12 min with 350 V in a 1 × LaemmliBuffer. Separated proteins within the gel were transferred ontoa PVDF membrane (Trans-Blot Turbo Mini PVDF Transfer Packs,Biorad) using the TransBlot System (Biorad) and the programme‘Mini TGX gel’ (3 min, 2.5 A, 25 V). After blocking with milk pow-der solution (1.5 g milk powder in 50 ml PBS + 0.1% Tween20),first antibody (human Erythropoietin MAb (MAB2871; R&D sys-tems) mouse IgG2A; 1:3000 in milk powder solution) andsecondary antibody anti mouse-Alexa Flour 680 (Molecular probes,1:5000 in milk powder solution) were used for staining. Detec-tion was performed with an Odysee Infrared Imager (Li-Cor) at700 nm.
2.13. Isoelectric focusing (IEF)
IEF of rhEpo was performed with self-cast gels containing animmobilized pH gradient from pH 2.5 to 7.0 cast on GelBond PAG
3 iotech
fiCpAenr2istbsbIadae
2
sstf2uB(tTubalalt(
2
UepLciorATe1i(me
2 L. Fliedl et al. / Journal of B
lm. Gels were rehydrated in a solution containing urea, DTT,HAPS and IPG buffer 3–10 for 2 h at RT. In the meantime sam-les were desalted and concentrated with spin filters (Millipore,micon-Ultra 10 K). Approximately 1 �g of rhEpo was loaded ofach sample, these were supernatant from RS-1C9 culture, super-atant from an industrially used serum-free CHO culture producinghEpo, and a rhEpo standard (Erythropoietin BRP, blending of the
rhEpo preparations, erythropoietin alpha and beta 1:1). Focus-ng was performed over night at 20 ◦C on a multiphor system (GE)tarting with 100 V up to 3000 V with 15 kV h totally. After elec-rophoresis proteins were transferred to a nitrocellulose membraney diffusion blotting over night. After blocking with milk powderolution (1.5 g milk powder in 50 ml PBS + 0.1% Tween20), first anti-ody (human Erythropoietin MAb (MAB2871; R&D systems) mouse
gG2A; 1:3000 in milk powder solution) and secondary antibodynti mouse Alexa Fluor 680 (Molecular probes, 1:5000 in milk pow-er solution) were used for staining. Detection was performed withn Odysee Infrared Imager (Li-Cor) at 700 nm. Image J was used tovaluate distribution of isoforms.
.14. Specific analysis of glycosylation
For analysis of linkage of sialyloligosaccharides on the cellurface, digoxigenin-labelled lectins from Sambucus nigra (SNA,pecific for �-2,6 sialyloligosaccharides) (DIG Glycan Differentia-ion Kit, Roche) and FITC-labelled anti-DIG Fab fragment (Roche)or detection were used. As described previously in (Genzel et al.,012). For detection of Galactose-alpha-1,3-galactose (� -Gal)sing flow cytometry, 1 × 106 cells were blocked for 20 min in 0.1%SA in PBS, resuspended in a 1:5 dilution alpha-Gal (M86) IgMEnzo Life sciences) followed by a washing step and incubation withhe 1:500 diluted Alexa Fluor 488 goat anti-mouse IgM (invitrogen).he detection of N-glycolylneuraminic acid (NeuGc), was donesing the Anti-Neu5Gc Basic Pack Kit (Sialix, Inc.). 1 × 106 cells werelocked for 20 min in the provided blocking puffer, resuspended in
1:500 dilution of the primary antibody or control antibody fol-owed by a washing step and incubation with the 1:1000 dilutednti-Chicken IgY 488A antibody from goat (Sigma Aldrich). Ana-yzes was done on a Gallios Flow Cytometer (Beckman Coulter) andhe percentages of positive cells were evaluated using the KALUZABeckman Coulter) software.
.15. In vitro assay
The in vitro biological activity of rhEpo was measured in anT-7 cell-based proliferation assay. The UT-7 cell line (Komatsut al., 1991) was maintained in RPMI 1640 (Biochrome AG) sup-lemented with 10% foetal calf serum (PAA Laboratories), 4 mM-glutamine and 5 ng/mL Epo. The cells were washed with Epo-freeulture medium and incubated for 4 h at 37 ◦C and 7% CO2. Increas-ng amounts of CHO or RS derived rhEpo were added to 100 �Lf medium containing about 104 cells in a 96-well culture plate,esulting in a final rhEpo concentration of 0.005–30 ng/mL per well.fter 4 days at 37 ◦C and 7% CO2, 10 �L of an MTT (Thiazolyl Blueetrazolium Bromide; Sigma) solution (5 mg/mL) were added toach well and the plate was incubated for 4 h as before. Finally,00 �L of 10% SDS (in 0.01 M HCl) were added to each well and
ncubated at 37 ◦C before measurement of absorbance at 570 nmreference wavelength 690 nm). The results of the fivefold deter-
ination were evaluated using Prism Graph Pad. Half maximumffective doses were calculated using 4 parameter fit-curves.
nology 176 (2014) 29–39
3. Results
3.1. Establishment of a serum-free continuously growing kidneycell line for biotechnological applications
Normal human renal proximal tubular epithelial cells grownunder well documented serum-free conditions were transfected atPD5.5 with pDEPT and arising cell clones were cultivated as massculture designated RS. Whereas the non-transduced cells couldonly be propagated for a maximum of 10 PDs before entering thephase of replicative senescence, SV40 early region overexpress-ing cells showed a remarkable increase of the replicative life span(Fig. 1A). However, the cell population showed slow growth for upto 100 days post transfection. This can be attributed to the fact thatnormal non-transfected cells are still present since positively trans-fected cells are not selected by addition of antibiotics but solelyby overgrowth of SV40 early region expressing cells due to highergrowth rates and an extended life span. As soon as normal cellshad entered replicative senescence, growth rates increased back tothe level of normal RPTECs (nRPTECs) at early PD. Indirect immu-nofluorescence stainings performed at PD24 clearly showed thatall cells homogenously express SV40 t/T antigens in the nucleus(Fig. 1B), which was not the case during selection phase (data notshown). Between PD30 and PD40 cells entered a crisis period withreduced growth rates. Thereafter, cells resumed growth again andthe cell line can be considered immortal with constant growth ratesthereafter (Fig. 1A).
Overexpression of SV40 early region influences normal cell cyclecontrol pathways by blocking essential tumour suppressor genesand thereby is able to extend the life span (Neufeld et al., 1987).However, for full immortalization, a mechanism for telomere sta-bilization such as reactivation of endogenous telomerase needs tobe activated (Bryan et al., 1995). In order to analyze telomeraseactivity in RS cell line during in vitro cultivation, lysates of nRPTECat PD6 as well as RS cells before (PD23) and after crisis (PD46 andPD61 indicated by arrows in Fig. 1A) were analyzed using TRAPassay. In accordance to literature nRPTECs do not show telome-rase activity (Fig. 1C). On the contrary, RS cells after crisis showedtelomerase activity demonstrating the spontaneous reactivation ofendogenous telomerase, which contributes to the immortal phe-notype.
3.2. Characterization of the cellular phenotype
Since it has been reported previously that the cellular pheno-type, especially the maintenance of differentiated features and theabsence of tumorigenicity strongly influences the glycosylationpattern of secreted proteins (Varki, 2010), we further aimed at adetailed analysis of cell type specific markers and functions of RScell line. As shown in Fig. 2, RS cell line showed the typical epithe-lial morphology of nRPTECs. However, immortal, SV40 early regionoverexpressing cells did not show the tight cell–cell contacts oftheir normal non-transfected counterpart, which was also mirroredby staining for E-cadherin. While normal cells showed the charac-teristic E-cadherin belt, RS cell line, albeit positive, gave a ratherdiffuse signal. On the other hand, when staining for Occludin, nor-mal as well as immortalized cells (Fig. 2, ZO-1), display a similarpositive staining pattern.
Moreover, RPTECs were tested for the activity of Gamma glu-tamyl transferase (GGT), an RPTEC specific enzyme with a majorrole in degradation and synthesis of gluthatione in RPTEC (Tateand Meister, 1981). The observed data shows that RS cells express
a 9-fold lower GGT activity compared to nRPTECs (0.02 versus0.18 nmol/min/105 cells) (Fig. 3A). Furthermore, response to hor-monal treatment of cells was tested. Response to parathyroidhormone (PTH) is characteristic for cells of proximal origin, while
L. Fliedl et al. / Journal of Biotechnology 176 (2014) 29–39 33
Fig. 1. Immortalization. Growth of nRPTEC and RS was followed (A). Whereas the nRPTEC stop growth after 10 population doublings, RS show an extended life span thatwas followed up to PD 71. Expression of SV40-T/t was confirmed by immunofluorescence, and localised to cell nuclei visualized by DAPI at PD24 (B). To confirm successfulimmortalization of RS TRAP assay was performed, revealing telomerase activity (C). Time points for sampling of cell lysates for TRAP assay are indicated by arrows in thegrowth curves (A). No telomerase activity was detected in nRPTEC at PD6 and RS at PD23, while RS at PD46 and PD61 were positive for telomerase activity. RNAse(R) treatedcell lysates were included as controls.
Fig. 2. Morphology und cellular junctions of RS and nRPTEC as assessed by microscopy. Upper panels show the morphology of the two cell lines, both showing epithelialcobblestone shape. Middle panels are immunofluorescence pictures of cells stained for E-cadherin, nRPTEC show distinct signal at cell membranes whereas RS cell show lowblurry signal. Lower panels shows images of both cells lines stained with ZO-1, showing a clear positive signal for both cell lines.
34 L. Fliedl et al. / Journal of Biotechnology 176 (2014) 29–39
Fig. 3. Enzymes and hormone response as markers of RPTECs. Gamma glutamyl transferase activity was tested, revealing that RS cells have significantly lower GGT activityc show
c over 9p cell b
iinintno
daFprAarino
3
bpaft1sciiw
adherence and morphology during the cultivation on microcar-riers (Fig. 5B) as the cells homogenously attached to the carrierand resumed growth instantly. Importantly, the comparison of
ompared to nRPTEC (A). RS show no response to PTH and AVP, whereas nRPTEC
hange when treated with AVP (B). In flow cytometric analysis all cell lines were
assages after isolation and for RS at 20 passages after establishment of the master
ncrease of cAMP levels after vasopressin (AVP) treatment is typ-cal for distal background (Toutain et al., 1991). RS cells showedo response to treatment with PTH. While a strong increase of the
ntracellular cAMP level was detected in nRPTEC treated with PTH,o cAMP was measured in RS cells (Fig. 3B). Cells did not respondo AVP treatment, which is in accordance with nRPTECs being alsoon-responsive to AVP treatment and suggesting that the cells aref proximal tubular origin.
Additionally, cells were tested for expression of Aminopepti-ase N (APN) an enzyme involved in metabolism of peptides, again
specific marker of RPTECs (Riemann et al., 1999). As shown inig. 3C, over 98% of normal as well as SV40 early region overex-ressing RS cells homogenously expressed APN. Finally, growthates of RS and nRPTECs were calculated from their growth curves.
mean growth rate was calculated over 6 passages for nRPTECsnd over 20 passages for RS. Both cell lines have a mean growthate of 0.26 per day (Fig. 3D). Thus, SV40 early region overexpress-ng cells, although having reduced some features of their normalon-transfected counterparts, are still clearly of proximal tubularrigin and express cell type specific markers.
.3. Expression of rhEpo
In order to test if the RS cell line has a potential as a recom-inant protein production host, it was transfected with pEPO, alasmid coding for human erythropoietin. rhEpo was selected as
model protein since it is highly glycosylated, and does originaterom the adult kidney in vivo. For selection of positive transfec-ants, cells were seeded into 96 well plates in medium containing00 �g/ml G418 24 h after transfection. Wells were screened foringle clones visually using a bright field microscope with phase
ontrast. After 14 days supernatants of wells identified as contain-ng single clones were analyzed for secreted rhEpo. Best performersn terms of growth and production were propagated and 7 clones
ere chosen for further characterization (data not shown). Finally,
the typical response of RPTEC, raised cAMP levels when treated with PTH and no8% positive for APN (C). The mean growth rates were calculated for nRPTEC at 6
ank (D).
RS-1C9 clone showed the highest productivity whereas growth ratewas similar to that of all other selected clones (data not shown).Therefore, RS-1C9 was used for all subsequent experiments.
To evaluate the stability of rhEpo production, RS-1C9 was pas-saged continuously and productivity was analyzed over 9 passages.Thereby we have shown that the productivity of RS-1C9 was stableand constant with a mean productivity of 1.29 pg/cell/day (Fig. 4).
Furthermore, to test cellular growth and productivity in astirred production system, RS-1C9 was grown in spinner flaskson microcarriers (Cytodex3), a stirred lab-scale process wellaccepted as model for large scale production processes in stirredbioreactors (Fliedl and Kaisermayer, 2011). The growth rate ofRS-1C9 in the stirred microcarrier cultivation during exponen-tial phase reached a maximum of 0.3 per day and a final cellconcentration of over 1.5 × 106 cells/ml was reached. The prod-uct accumulated constantly over the cultivation time reaching afinal volumetric productivity of 15.8 �g/ml and a specific produc-tivity of 2.3 pg/cell/day on day 14 (Fig. 5A). RS-1C9 showed good
Fig. 4. Stability of RS-1C9 productivity. The clone (1C9) was passaged continuouslyand productivity was analyzed over 9 passages. The productivity of RS-1C9 wasstable and constant with a mean productivity of 1.29 pg/cell/day.
L. Fliedl et al. / Journal of Biotechnology 176 (2014) 29–39 35
Fig. 5. Upscale of RS-1C9. RS-1C9 were cultivated on microcarriers in spinner flask. Cell densities of up to 1.8 × 106 c/ml were reached (A). The growth during exponentialp . RS-1o and sp( he spi
gsewTd
3
fisfrpdafpc(optBsiswctci
distribution of isoforms as defined by the European pharmacopoeia(“Erythropoietin concentrated solution,” 2007).
Finally, when combining the calculations of the percentage ofsufficient rhEpo by range and the distribution of isoforms a total
hase reached a maxiumum rate of 0.3 d–1 . Final qp on day 13 was 0.47 pg/cell/dayn carriers on day 3 of cultivation (B). The comparision of growth rates from T-flaskC). When comparing the specific productivity, RS-1C9 was significantly higher in t
rowth rates from T-flask and spinner flask cultivation showed noignificant differences between the two systems (Fig. 5C). How-ver, as shown in Fig. 5D, the specific productivity of RS-1C9as significantly improved in the stirred system, compared to the
-flask cultivation, showing the large potential of increasing pro-uctivity by optimizing the culture system used for production.
.4. Analysis of glycosylation
In order to analyze the presence of the 3 N-glycans and to getrst insights into the quality of the secreted rhEpo, we treated theupernatant of 1C9 cells with PNGase F. The purified and finallyormulated BRP standard served as positive control. A stepwiseemoval of carbohydrates was clearly visible for both the rhEporoduced with RS-1C9 (Fig. 6A) and the BRP standard (Fig. 6B) inependence on time/dose of PNGase F. To elaborate the presencend distribution of terminal sialic acids on the rhEpo isoelectricocusing was performed. rhEpo from RS-1C9 and CHO were com-ared to the BRP standard. While with the rhEpo produced in CHOells only 50% were found to be in the range of the BRP standardFig. 6C), more than 70% of the RS-1C9 Epo was within the rangef the isoforms present in the BRP standard (Fig. 6D). The mostrominent isoforms in BRP standard are 4–7, as also published inhe “Collaborative Study for the Establishment of ErythropoietinRP Batch 3” (Behr-Gross et al., 2007). This distribution is slightlyhifted for the rhEpo produced in RS-1C9 where most prominentsoforms are 2–6. Even more difference is present for the CHOupernatant, where isoforms 1–5 are most pronounced. Isoform 7as proportionally the least expressed form, in both host systems
ompared to the BRP standard and therefore was used to calculateheoretical yields. While only 23% of Epo produced in CHO cellsan be used due to isoform distribution, this yield is significantlyncreased for RPTECs, where up to 40% of the product meet the
C9 showed good attachment and morphology on microcarriers. Photo shows cellsinner flask cultivation showed no significant differences between the two systems
nner, compared to the T-flask cultivation (D).
Fig. 6. The RS-1C9 derived hEPO was digested with PNGase F to prove the presenceof all 3 N-glycans (A) and compared to BRP standard (B). Isoelectric focusing of hEPOfor analysis of terminal sialylation. hEPO was compared to the BRP standard. 50% ofthe CHO derived rhEPO was with in the range of the BRP standard (C). Over 70% ofthe RS-1C9 derived rhEPO was in the zone of the isoformes of the BRP standard (D).
36 L. Fliedl et al. / Journal of Biotechnology 176 (2014) 29–39
F ells were negative for �-2,6 sialyloligosaccharides, while over 99% RS cells were detecteda lls. N-glycolylneuraminic acid (NeuGc) was negative for RS cells, while 90% of CHO cellsw
yR2i
3m
ts(sto�M(ann(b
3
ieofJoSt1pb
Fig. 8. In vitro activity assay of rhEPO derived from either RS-1C9 or CHO cells.
ig. 7. Analysis of specific glycopatterns using specific lectins and antibodies. CHO cs positive. Galactose-alpha-1,3-galactose (�–Gal) was negative for RS and CHO ceere detected positive for Neu5Gc.
ield of only 11.5% was calculated for CHO derived product whileS-1C9 Epo reached a total value of 28%. Thus, an approximately.5 fold difference in percent of total was observed in the yield of
soforms meeting the European pharmacopoeia regulations.
.5. Analysis of specific glycosylation patterns of the host cells asirror of the rhEPO glycosylation
Since proper human like glycosylation and subsequent func-ionality as well as in vivo half life involves the presence of �-2,6ialyloligosaccharides, absence of Galactose-alpha-1,3-galactose�–Gal) as well as N-glycolylneuraminic acid (NeuGc) have to behown, RS cells as well as CHO cells were tested for presence ofhese sugar residues. As shown in Fig. 7 flow cytometric analysisf linkage of sialylation revealed that CHO cells were negative for-2,6 sialyloligosaccharides, while over 99% RS cells were positive.oreover, �–Gal was not expressed in RS as well as in CHO cells
Fig. 7). On the contrary, cells derived from pig, which were useds positive control, showed a homogenous positive staining (dataot shown). Immunostaining of NeuGc showed that RS cells wereegative, while 90% of CHO cells were detected positive for Neu5GcFig. 7). Therefore, RS cells represent a host cell system being capa-le of adding sugar residues in their genuine human form.
.6. In vitro biological activity of rhEPO
In order to test the in vitro biological activity of rhEpo producedn RS-1C9 and CHO cells, we treated UT-7 cells with rhEpo fromither CHO cells or RS-1C9 cells. Proliferation of UT-7 cells dependsn the presence of receptor binding Epo, and thus these cells arerequently used as indicator for product quality (Cohan et al., 2011;ez et al., 2013). As shown in Fig. 8, CHO derived rhEPO had an ED50f 0.017 pmol/ml, whereas RS-1C9-EPO reached 0.131 pmol/ml.ince, it was shown earlier that removal of N-glycosylation leads
o higher in vitro activity (Tsuda et al., 1990; Yamaguchi et al.,991), one possible explanation of this lower activity is that RS cellsroduced rhEpo might have a higher level of sialylation. This idea isased on the fact that the surface protein staining of RS-1C9 shows
Half maximum effective doses (ED50) were calculated using 4 parameter fitcurves.RS-1C9 reached an ED50 of 0.131 pmol/ml and CHO 0.0165 pmol/ml.
homogenous sialylation (see above), as well as on data showingthat plant derived rhEpo without sialylation had a higher activitycompared to its sialylated counterpart (Jez et al., 2013). Eventhough lack of glycostructures is favourable for receptor bindingin vitro, it is detrimental for in vivo half-life and accordinglybioavailability, as it was shown that a higher degree of sialylationof rhEpo leads to higher in vivo efficacy (Egrie and Browne, 2001).
4. Discussion
Currently almost 70% of recombinant proteins are produced by
CHO cells (Jayapal et al., 2007). This predominance is due to highgrowth rates and productivity and the ability to form “human like”glycosylation patterns. Nevertheless, the remaining differences in
iotech
tg
hSbupepcfafi
c(dOsmr(gaf
r(rreiunadataoti
cdwoosEccpombb
twscag
L. Fliedl et al. / Journal of B
he glycopattern, such as the degree of sialylation, and non-humanlycopatterns are still a major drawback of CHO cells.
Therefore, human cells are of ever increasing importance asosts for production of complex proteins, as recently reviewed bywiech et al. (2012). Human embryonic kidney cells transformedy fragments of adenovirus type 5 (HEK293) have been extensivelysed in transient processes. To improve performance in transientroduction derivative cell lines of HEK293 were developed enablingpisomal replication of appropriate plasmids. Therefore, high yieldroviding HEK293 and derivates are probably the most dominantell lines for the production of research grade material. However,or the stable production of pharmaceutical proteins HEK293 cellsre rarely used. This might be related to their high degree of trans-ormation (Graham et al., 1977), which on the one hand has majormpact on glycosylation (Varki et al., 2009) and raises safety issues.
Another human cell line used for the production of pharma-eutical proteins is HT-1080, a cell line derived from fibrosarcomaRasheed et al., 1974). HT-1080 cells have been used to produce theesired proteins using gene-activation technology (Moran, 2010).ne product emerging from this technology is Dynepo, which
hows significant differences compared to the CHO derived com-ercial products. While Dynepo does not contain Neu5Gc, the
hEpos from CHO cells show 1.1%–1.4% of this sialic acid variantShahrokh et al., 2011). The presence of Neu5Gc leads to immuno-enicity, increased clearance of the product (Ghaderi et al., 2010)nd may also lead to rhEpo resistance (Shahrokh et al., 2011). Thisurther demonstrates the high potential of human host cell lines.
The non-tumour derived cell lines Per.C6 (human embryonicetina cells), AGE1.HN (human neuronal precursors) and CAP cellsamniocytes) have also been evaluated for the production ofecombinant proteins and have already proven the technologicalelevance of human cell lines (Blanchard et al., 2011; Kuczewskit al., 2011; Schiedner et al., 2008). Thus, Per.C6 at the moments the most promising alternative to CHO having 14 related prod-cts in clinical trails phase I/II. Nevertheless, due to the limitedumber of available non-transformed human production cell linesnd the previously noted differences in glycosylation patterns pro-uced by cells from different tissues (Berger et al., 2012), there isn urgent need for alternative human host systems. In order to testhe hypothesis that human cells producing a specific protein in vivolso are best producers of high quality proteins, we here focussedn the establishment of a rhEpo expression human renal proximalubular epithelial cell line and the analysis of the resulting proteinn terms of glycosylation, especially sialylation.
Therefore, all steps including isolation, immortalization andultivation were done under serum free conditions and are wellocumented. Immortalization initiated by the early region of SV40as confirmed by immunofluorescence of SV40 t/T, analysis of tel-
merase activity using TRAP assay and the continuous monitoringf elongated life span. All data meet the previously described twotep model of SV40 antigen immortalization (Wright et al., 1989).xamination of the cell line and comparison with normal parentalells revealed the maintenance of several characteristics of theseells, although some specific functions seem diminished in com-arison to their parental cells. Still, RS cells fulfil important criteriaf a high quality human host cell line (i) complete absence of ani-al derived products during all stages of cultivation and (ii) balance
etween growth and differentiation in order to fulfil the needs foriotechnological application.
In order to test its recombinant protein production capabili-ies, Epo was overexpressed in RS cells and the highest produceras selected in small scale and further tested in a stirred lab-scale
ystem. Thereby, we were able to double the specific productivityompared to the static system. Thus, RS-1C9 clone can be up-scalednd there seems to be much room for improvement in terms of bothrowth and productivity by optimization of the production process.
nology 176 (2014) 29–39 37
The analysis of glycan structure of rhEpo produced in RS-1C9cells by enzymatic digestion using PNGase resulted in a staircasepattern clearly showing the presents of all 3 N-glycans. Surpris-ingly, Western blot analysis of RS-1C9 produced rhEpo revealed anadditional smaller molecular weight band when all three N-glycanshave been cleaved off (Fig. 6B). We assume that this double bandmight derive from a small fraction lacking O-glycosylation, whichis of minor importance for the biological activity of Epo (Higuchiet al., 1992). Moreover, serum was reported to contain Epo withand without O-glycan structure (Skibeli, 2001) and also marketedEpo is not fully O-glycosylated ((Llop et al., 2008) and reviewed in(Deicher and Hörl, 2004)).
Using IEF we further analyzed rhEpo produced with RS cells andcompared it to the finally formulated BRP standard, which is a com-bination of two commercially available rHEpo products producedin CHO cells. When comparing the IEF distribution of rhEpo iso-forms, we observed that a higher proportion of RS-1C9 productisoforms are in the range of the BRP standard, proving that moresialic acids are present on the hEpo produced with the human cellline as compared to CHO-rhEpo. When combining isoform rangeand distribution we calculated that only 11.5% of CHO product ismeeting the European pharmacopoeia regulations (EPR) while RS-1C9 Epo reached a total value of 28%. Thus, we estimate that anapproximately 2.5 fold higher yield will be available after down-stream processing, where the isoforms that do not meet the EPRcriteria have to be discarded.
Although we cannot fully exclude changes in the degree of sialy-lation caused by media differences, pH or productivity, human cellsas host systems were already shown to be preferential in terms ofsialylation (Shahrokh et al., 2011 and reviewed in Swiech et al.,2012).
The analysis of specific glycopatterns, namely �-2,6 linkage ofterminal oligosaccharide, �-Gal and Neu5Gc, revealed differencesbetween RS and CHO cells. Although, the impact of absence of ter-minal �-2,6 oligosaccharide linkage is not fully understood up tonow (Swiech et al., 2012), we could show that CHO do not formthis linkage while RS cells indeed are homogenously positive. Fur-ther we have shown that Neu5Gc is not present on RS cells, whilemeasured on CHO cells. The analysis of �-Gal was negative for bothCHO and RS, but this glycopattern may increase for individual CHOclones and was shown to be present on commercial product (Oren-cia) derived from CHO (Bosques et al., 2010). Finally, we showedthat RS-1C9 derived rhEpo is biologically active in vitro. Higheramounts of rhEpo produced in the human host cell line are requiredfor growth stimulation of UT-7 cells when compared to the CHOproduct. This fact might be seen as an additional indicator thatthe product carries a high degree of sialylation and therefore hashigh in vitro activity, as the here used in vitro assay is negativelycorrelated with in vivo bioactivity (Tsuda et al., 1990). Taking ourdata together we here suggest that the novel, well documentedand serum-free human production platform of RS cells might rep-resent an alternative for the production of complex glycosylatedproteins, especially if a more human-like glycosylation pattern is ofinterest.
Notes
RG and JG are cofounders of Evercyte GmbH and declare Conflictof Interest.
Acknowledgments
This work has been supported by the Federal Ministry ofEconomy, Family and Youth (BMWFJ), the Federal Ministry ofTraffic, Innovation and Technology (bmvit), the Styrian Busi-ness Promotion Agency SFG, the Standortagentur Tirol and
3 iotech
ZCPHcs
R
A
B
B
B
B
B
B
B
C
C
D
E
EF
F
G
G
G
G
H
H
J
J
J
8 L. Fliedl et al. / Journal of B
IT – Technology Agency of the City of Vienna through theOMET-Funding Programme managed by the Austrian Researchromotion Agency FFG. Additionally, RG was supported by FWF anderzfeldersche Familienstiftung. Moreover we thank the Imagingentre of BOKU and especially Monika Debreczeny for excellentupport.
eferences
shwell, G., Harford, J., 1982. Carbohydrate-specific receptors of the liver. AnnualReview of Biochemistry 51, 531–554.
anerji, J., Olson, L., Schaffner, W., 1983. A lymphocte-specific cellular enhancer islocated downstream of the joining region in immunoglobulin heavy chain genes.Cell 33, 729–740.
ehr-Gross, M.-E., Daas, A., Burns, C., Bristow, A.F., 2007. Collaborative study for theestablishment of erythropoietin BRP batch 3. Pharmeuropa Bio, 49–67.
erger, M., Kaup, M., Blanchard, V., 2012. Protein glycosylation and its impacton biotechnology. Advances in Biochemical Engineering/Biotechnology 127,165–185.
lanchard, Liu, X., Eigel, S., Kaup, M., Rieck, S., Janciauskiene, S., Sandig, V., Marx, U.,Walden, P., Tauber, R., Berger, M., 2011. N-glycosylation and biological activityof recombinant human alpha1-antitrypsin expressed in a novel human neuronalcell line. Biotechnology and Bioengineering 108, 2118–2128.
rooks, S.A., 2004. Appropriate glycosylation of recombinant proteins for humanuse: implications of choice of expression system. Molecular Biotechnology 28(3), 241–255.
ryan, T.M., Englezou, a., Gupta, J., Bacchetti, S., Reddel, R.R., 1995. Telomere elon-gation in immortal human cells without detectable telomerase activity. EMBOJournal 14, 4240–4248.
ai, H., Huang, Z.-H., Shi, L., Sun, Z.-Y., Zhao, Y.-F., Kunz, H., Li, Y.-M., 2012. Variation ofthe glycosylation pattern in MUC1 glycopeptide BSA vaccines and its influenceon the immune response. Angewandte Chemie (International ed. in English) 51,1719–1723.
ohan, R.A., Madadkar-Sobhani, A., Khanahmad, H., Roohvand, F., Aghasadeghi, M.R.,Hedayati, M.H., Barghi, Z., Ardestani, M.S., Inanlou, D.N., Norouzian, D., 2011.Design, modeling, expression, and chemoselective PEGylation of a new nano-size cysteine analog of erythropoietin. International Journal of Nanomedicine 6,1217–1227.
eicher, R., Hörl, W.H., 2004. Erythropoiesis-Stimulating Agents Kidney Disease, vol.64., pp. 499–509.
grie, J.C., Browne, J.K., 2001. Development and characterization of novel ery-thropoiesis stimulating protein (NESP). Nephrology, Dialysis, Transplantation:Official Publication of the European Dialysis and Transplant Association – Euro-pean Renal Association 16 (Suppl. 3), 3–13.
rythropoietin concentrated solution, 2007. Ph. Eur. Monograph 1316, 6th ed.liedl, L., Kaisermayer, C., 2011. Transient gene expression in HEK293 and vero cells
immobilised on microcarriers. Journal of Biotechnology 153, 15–21.ukusa, M.N., Sasaki, H., Lopez, L., Fukuda, M., 1989. Survival of recombi-
nant erythropoietin in the circulation: the role of carbohydrates. Blood 73,84–89.
enzel, Y., Behrendt, I., Rödig, J., Rapp, E., Kueppers, C., Kochanek, S., Schiedner,G., Reichl, U., 2012. CAP, a new human suspension cell line for influenza virusproduction. Applied Microbiology and Biotechnology.
haderi, D., Taylor, R.E., Padler-Karavani, V., Diaz, S., Varki, A., 2010. Implicationsof the presence of N-glycolylneuraminic acid in recombinant therapeutic glyco-proteins. Nature Biotechnology 28, 863–867.
rabenhhorst, E., Costa, J., Conradt, H.S., 1997. Construction of novel BHK-21 cell lines coexpressing Golgi resident or soluble forms of human�2,6-sialyltransferase and �1,3/4-fucosyltransferases together with secretoryglycoproteins. In: Carrondo, M.J.T., Griffiths, B., Moreira, J.L. (Eds.), AnimalCell Technology: From Vaccines to Genetic Medicine. Kluwer Acad. Publ, theNetherlands, pp. 481–487.
raham, F.L., Smiley, J., Russell, W.C., Nairn, R., 1977. Characteristics of a humancell line transformed by DNA from human adenovirus type 5. Journal of GeneralVirology 36, 59–74.
ayflick, L., Moorhead, P.S., 1961. The serial cultivation of human diploid cell strains.Experimental Cell Research 25, 585–621.
iguchi, M., Oh-eda, M., Kuboniwa, H., Tomonoh, K., Shimonaka, Y., Ochi, N., 1992.Role of sugar chains in the expression of the biological activity of human eryth-ropoietin. Journal of Biological Chemistry 267, 7703–7709.
ayapal, K.P., Wlaschin, K.F., Hu, W.-S., Yap, M.G.S., 2007. No TitleRecombinant pro-tein therapeutics from CHO cells—20 years and counting. Chemical EngineeringProgress, 40–47.
elkmann, W., 1992. Erythropoietin: structure, control of production, and function.Physiological Reviews 72, 449–489.
ez, J., Castilho, A., Grass, J., Vorauer-Uhl, K., Sterovsky, Thomas, Altmann, F.,Steinkellner, H., 2013. Expression of functionally active sialylated human eryth-ropoietin in plants. Biotechnology Journal 8, 371–382.
nology 176 (2014) 29–39
Komatsu, N., Nakauchi, H., Miwa, A., Ishihara, T., Eguchi, M., Moroi, M., Okada,M., Sato, Y., Wada, H., Yawata, Y., 1991. Establishment and characterizationof a human leukemic cell line with megakaryocytic features: dependency ongranulocyte-macrophage colony-stimulating factor, interleukin 3, or erythro-poietin for growth and survival. Cancer Research 51, 341–348.
Krantz, S., Jacobson, L., 1970. Erythropoietin and the Regulation of Erythropoiesis.University of Chicago Press, Chicago.
Kuczewski, M., Schirmer, E., Lain, B., Zarbis-Papastoitsis, G., 2011. A single-use purifi-cation process for the production of a monoclonal antibody produced in a PER.C6human cell line. Biotechnology Journal 6, 56–65.
Kwaks, T.H.J., Otte, A.P., 2006. Employing epigenetics to augment the expres-sion of therapeutic proteins in mammalian cells. Trends in Biotechnology 24,137–142.
Llop, E., Gutiérrez-Gallego, R., Segura, J., Mallorquí, J., Pascual, J.a., 2008. Structuralanalysis of the glycosylation of gene-activated erythropoietin (epoetin delta,Dynepo). Analytical Biochemistry 383, 243–254.
Meister, A., Tate, S.S., Griffith, O.W., 1981. Gamma-glutamyl transpeptidase. Meth-ods in Enzymology 77, 237–253.
Padler-Karavani, V., Varki, A., 2011. Potential impact of the non-human sialic acidN-glycolylneuraminic acid on transplant rejection risk. Xenotransplantation 18,1–5.
Park, J.H., Wang, Z., Jeong, H.-J., Park, H.H., Kim, B.-G., Tan, W.-S., Choi, S.S., Park,T.H., 2012. Enhancement of recombinant human EPO production and glycosyl-ation in serum-free suspension culture of CHO cells through expression andsupplementation of 30Kc19. Applied Microbiology and Biotechnology 96, 671–683.
Rasheed, S., Nelson-Rees, W.A., Toth, E.M., Arnstein, P., Gardner, M.B., 1974. Char-acterization of a newly derived human sarcoma cell line (HT-1080). Cancer 33,1027–1033.
Riemann, D., Kehlen, A., Langner, J., 1999. CD13—not just a marker in leukemiatyping. Immunology Today 20, 83–88.
Sasaki, H., Bothner, B., Dell, a., Fukuda, M., 1987. Carbohydrate structure of eryth-ropoietin expressed in Chinese hamster ovary cells by a human erythropoietincDNA. Journal of biological chemistry 262, 12059–12076.
Schiedner, G., Hertel, S., Bialek, C., Kewes, H., Waschütza, G., Volpers, C., 2008. Effi-cient and reproducible generation of high-expressing, stable human cell lineswithout need for antibiotic selection. BMC Biotechnology 8, 13.
Schlenke, P., Grabenhhorst, E., Wagner, R., Nimtz, M., Conradt, H.S., 1997. Expressionof a human �2,6-sialyltransferase in BHK-21A cells increases the sialylation ofcoexpressed human erythropoietin: NeuAc-transfer onto GalNAc(�1-4)GlcNAc-R motives. In: Carrondo, M.J.T., Griffiths, B., Moreira, J.L. (Eds.), Animal CellTechnology: From Vaccines to Genetic Medicine. Kluwer Acad. Publ, theNetherlands, pp. 475–480.
Shahrokh, Z., Royle, L., Saldova, R., Bones, J., Abrahams, J.L., Artemenko, N.V., Flatman,S., Davies, M., Baycroft, A., Sehgal, S., Heartlein, M.W., Harvey, D.J., Rudd, P.M.,2011. Erythropoietin produced in a human cell line (Dynepo) has significantdifferences in glycosylation compared with erythropoietins produced in CHOcell lines. Molecular Pharmaceutics, 395.
Shanks, J.H., Hill, C.M., Lappin, T.R., Maxwell, a.P., 1996. Localization of erythropoietingene expression in proximal renal tubular cells detected by digoxigenin-labelledoligonucleotide probes. Journal of Pathology 179, 283–287.
Skibeli, V., 2001. Sugar profiling proves that human serum erythropoietin differsfrom recombinant human erythropoietin. Blood 98, 3626–3634.
Swiech, K., Picanc o-Castro, V., Covas, D.T., 2012. Human cells: new platform forrecombinant therapeutic protein production. Protein Expression and Purifica-tion 84, 147–153.
Takeuchi, M., Inoue, N., Strickland, T.W., Kubota, M., Wada, M., Shimizu, R., Hoshi, S.,Kozutsumi, H., Takasaki, S., Kobata, a., 1989. Relationship between sugar chainstructure and biological activity of recombinant human erythropoietin producedin Chinese hamster ovary cells. Proceedings of the National Academy of Sciencesof the United States of America 86, 7819–7822.
Tate, S.S., Meister, A., 1981. gamma-Glutamyl transpeptidase: catalytic, structuraland functional aspects. Molecular and Cellular Biochemistry 39, 357–368.
Toutain, H., Vauclin-Jacques, N., Fillastre, J.-P., Morin, J.-P., 1991. Biochemical,functional, and morphological characterization of a primary culture of rabbitproximal tubule cells. Experimental Cell Research 194, 9–18.
Tsuda, E., Kawanishi, G., Ueda, M., Masuda, S., Sasaki, R., 1990. The role ofcarbohydrate in recombinant human erythropoietin. European Journal of Bio-chemistry/FEBS 188, 405–411.
Varki, A., 2010. Uniquely human evolution of sialic acid genetics and biology.Proceedings of the National Academy of Sciences of the United States of America107 (Suppl), 8939–8946.
Varki, A., Kannagi, R., Toole, B.P., 2009. Glycosylation Changes in Cancer, in: Essentialsof Glycobiology, 2nd ed. Cold Spring Harbor Laboratory Press, NY.
Voglauer, R., Grillari, J., Fortschegger, K., Wieser, M., Sterovsky, T., Gunsberg, P.,Katinger, H., Pfragner, R., 2005. Establishment of human fibroma cell lines from aMEN1 patient by introduction of either hTERT or SV40 early region. International
Journal of Oncology 26, 961–970.
Wang, Z., Park, J.H., Park, H.H., Tan, W., Park, T.H., 2011. Enhancement of recombinanthuman EPO production and sialylation in chinese hamster ovary cells throughBombyx mori 30Kc19 gene expression. Biotechnology and Bioengineering 108,1634–1642.
20434–20439.Zhang, X., Lok, S.H., Kon, O.L., 1998. Stable expression of human alpha-2,6-
L. Fliedl et al. / Journal of B
ieser, M., Stadler, G., Jennings, P., Streubel, B., Pfaller, W., Ambros, P., Riedl, C.,Katinger, Hermann, Grillari, Johannes, Grillari-Voglauer, R., 2008. hTERT aloneimmortalizes epithelial cells of renal proximal tubules without changing theirfunctional characteristics. American Journal of Physiology. Renal Physiology 295,F1365–F1375.
right, W.E., Pereira-Smith, O.M., Shay, J.W., 1989. Reversible cellular senescence:implications for immortalization of normal human diploid fibroblasts. Molecularand Cellular Biology 9, 3088–3092.
urm, F.M., 2004. Production of recombinant protein therapeutics in cultivatedmammalian cells. Nature Biotechnology 22, 1393–1398.
nology 176 (2014) 29–39 39
Yamaguchi, K., Akai, K., Kawanishi, G., Ueda, M., Masuda, S., Sasaki, R., 1991. Effectsof site-directed removal of N-glycosylation sites in human erythropoietin onits production and biological properties. Journal of Biological Chemistry 266,
sialyltransferase in Chinese hamster ovary cells: functional consequences forhuman erythropoietin expression and bioactivity. Biochimica et Biophysica Acta1425, 441–452.
