Vaccine28 (2010) 309–316
Contents lists available at ScienceDirect
Vaccine
j o u r n a l h o m e p a g e:w w w.e l s e v i e r.c o m/l o c a t e/v a c c i n
e
Development of a simple and high-yielding fed-batch process for the production of influenza vaccines
Jamal Meghrous a,b,Wafaa Mahmoud a,Danielle Jacob b,Rick Chubet a,Manon Cox a,Amine A.Kamen b,∗
a Protein Sciences Corporation,1000Rearch Parkway,Meriden,CT06450,United States
b Biotechnology Rearch Institute,National Council of Canada,6100Royalmount Avenue,Montreal,Quebec,Canada H4P2R2
a r t i c l e i n f o
Article history:
Received10August2009
Received in revid form8October2009 Accepted12October2009
雅思英语培训机构
Available online 30 October 2009
Keywords:
英语学习辅导Influenza vaccine
Fed-batch cultures
Hemagglutinin
Baculovirus
expres SF+cells
FluBlok®a b s t r a c t
A robust and reliable GMP-compatible fed-batch process was successfully developed for the production of recombinant hemagglutinin(rHA)proteins by expres SF®cells.The feeding solution,feeding strategy as well as the cell density at infection were optimized to maximize thefinal rHA production yields with-out affecting the existing rHA recovery protocol and downstream process.
A simple and stable feeding solution was formulated and a rational feeding regimen designed to yield,depending on the rHA bac-ulovirus ud,between2-and3-fold enhancements in volumetric rHA production with incread specific productivity compared to the batch culture.Recombinant HA from fed-batch cultures could be simply recovered following cell lysis and purified through chromatographic steps.Overall,the incread rHA yield was maintained throughout the whole process.The performance,reproducibility and scalability of the fed-batch process was successfully demonstrated in12bioreactor runs of2-and10-L working volume usingfive different rHA encoding baculovirus.
Crown Copyright © 2009 Published by Elvier Ltd. All rights rerved.
1.Introduction
Seasonal influenza is a highly contagious respiratory dia causing about300,000–500,000deaths annually worldwide(World Health Organization,WHO)[1].The emergence and rapid spread of possibly highly pathogenic influenza virus variants,through antigenic drift or recombination with other strains,has prompted pharmaceutical companies to make strong efforts to counter the imminent possibility of a pandemic[2,3].The risk of a pandemic exists and the most effective strategy to control
such a threat of emerging virus is preventive vaccination[4].To date,most of the commercially availableflu vaccines are produced using embryonated chicken eggs[2].The current annual vaccine con-tains hemagglutinin antigen(HA)from three influenza virus,two strains of the type A virus and one strain of the type B virus[4,5]. There are two egg-derived influenza vaccines that are currently licend for commercial u in the US:(1)trivalent inactivated vaccine(TIV),which in turn could be split(subvirion)or a sub-unit vaccine,and(2)live-attenuated influenza vaccine(LAIV).The main drawbacks of using the egg-bad technology for influenza vaccine production is that it is labor-intensive and difficult to scale-up in the event of a pandemic and takes veral months following the identification of new strains as demonstrated by the respon ∗Corresponding author.Tel.:+15144962264;fax:+15144966785.
E-mail address:Amine.a(A.A.Kamen).to the emergence of the novel H1N1influenza virus and vaccine shortages.Cell culture-bad technologies are promising and safe systems for mass and rapid production of candidate vaccines.Sev-eral pharmaceutical companies have paid increasing attention to the technologies over the last decade.Madin Darby Canine Kid-ney(MDCK)and African green monkey kidney Vero cells(Vero)are two continuous mammalian cell lines that have been frequently ud or are currently being considered by veral companies for the
零七八碎production of influenza vaccines.However,the cell lines are adherent cell lines making them inherently difficult to scale-up becau of their requirement for an attachment surface.Another cell line growing in suspension,the human retina-derived cell line PER.C6,is currently being evaluated by Sanofi-Pasteur for manufac-turing of influenza vaccines,however no data have been relead to date.Efforts have been dedicated to develop new cell lines such as the avian embryonic derived stem cell line[6]or the duck retina cell line[7]for large-scale production of influenza vaccines.In gen-eral,mammalian cell culture-bad technologies might require the production of a high-yielding re-assorted virus with other high-yielding strains in cell cultures instead of chicken eggs,which could introduce host specific mutations in the viral genome.Using mammalian systems would involve steps of inactivation and/or attenuation of live virus that needs to be produced under various levels of biocontainment.Substantial progress has been made with the technologies for influenza vaccine manufacturing and some mammalian cell-bad influenza vaccines have obtained licen for commercialization;however,low production yields are often the
尺子英文
0264-410X/$–e front matter.Crown Copyright © 2009 Published by Elvier Ltd. All rights rerved. doi:10.1016/j.vaccine.2009.10.048
310J.Meghrous et al./Vaccine28 (2010) 309–316
major limitations of using this technology.In the event of pandemic, high-level expression is an esntial requirement for an economic and effective influenza vaccine.Inct cell technology is an alter-native cell culture-bad technology for manufacturing candidate influenza vaccines.It takes advantage of the progress in recombi-nant DNA technology and the safety profile of inct cell cultures to successfully produce sub-unit influenza vaccines such as Viral Like Particle(VLP)or recombinant viral proteins[8,9].Protein Sciences Corporation has developed an influenza vaccine,FluBlok®,using the inct cell/baculovirus system[5,10,11].FluBlok®has been shown to be safe and efficacious in human clinical trials[11–13]. FluBlok®vaccine was granted fast track status and priority review by FDA and a Biological Licen Application was submitted to FDA in April2008.
Inct cell technology is a promising platform for mass-production of influenza vaccine and might be a practical alternative to overcome the limitations and drawbacks prented by the egg-bad manufacturing technology.An inct cell-derived influenza vaccine has been shown to provide equivalent or better immuno-genicity,an established surrogate marker for protection in clinical trials than egg-derived vaccine[11,14–16].Potentially,it has great advantages compared to conventional systems with regards to safety,and ea and rapidity of scaling to large-scale production in the event of worldwide pandemic.The major advantage of the inct cell/baculovirus expression system(BEVS)
-derived sub-unit vaccine is that the manufacturing of the HA proteins does not require the cultivation and subquent inactivation of live influenza virus as required for embryonated eggs or mammalian cells pro-duction systems[5].
BEVS has been ud extensively for the production of viral anti-gen and vaccine products[17,18].BEVS has recently been ud in the manufacture of both veterinary and human vaccines and more than10BEVS-derived products are now either commercially avail-able or in clinical trials.For instance,the baculovirus system has been ud for the production of animal vaccines such as Porcilis®Pesti,Advasure®and Bayovac,vaccines against the classical swine fever,produced by Intervet International(Boxmeer,The Nether-lands,now Intervet/Schering-Plough),Pfizer(New York,NY)and Bayer(Liverkun,Germany),respectively;as well as Circumvent®PCV vaccine and Ingelvac®CircoFLEX TM,against Porcine Cir-covirus Associated Dia produced by Intervet/Schering-Plough and Boehringer Ingelheim(St.Joph,MO),respectively.Also,the BEVS has been ud by GlaxoSmithKline(GSK,Rixensart,Belgium) for manufacturing the human vaccine Cervarix TM,a bivalent human papillomavirus(HPV16/18)vaccine against cervical cancer.Other potential human vaccines are currently in the late-stage clinical development such as Provenge,an immuno-therapeut
18 and lifeic vaccine for prostate cancer,produced by Dendreon Inc.(Seatlle,WA)and Chimigen vaccines,against chronic hepatitis B and C,developed by Virexx Medical Corp.(Calgary,Canada).
To obtain high yields in a scalable and a cost-effective sys-tem,the choice of the production process is of great importance. Fed-batch technology has become a platform technology for the large-scale production of therapeutic and recombinant proteins due to its simplicity of operation,flexibility to be implemented and a highfinal product concentration[19,20].Fed-batch culture is a frequently ud technology for the industrial manufacturing of cell culture-bad recombinant and therapeutic proteins.The productivity of the cell culture process has incread dramatically by optimizing the fed-batch technology and particularly improv-ing the feed formulation and the feeding strategy.An important conquence of this increa in productivity is the decrea in prod-uct manufacturing cost.For instance,the antibody production was 40-fold improved using the fed-batch process[19].
The most challenging task is to maximizefinal yields of hemag-glutinin through the development of a simple fed-batch process while limiting changes to the downstream protein purification pro-cess.Here,we describe the development of a simple fed-batch process for high yielding of rHA proteins using SF+cells and rum-free Protein Sciences Formulary Medium.Robustness and reproducibility of this developed fed-batch are evaluated by asss-ing productions offive different rH
As in two different core facilities, at the Biotechnology Rearch Institute(National Rearch Council Canada,Montreal,Canada)and Protein Sciences Corporation(PSC, Meriden,CT,USA).
2.Materials and methods
2.1.Cell maintenance and viral stocks preparation
expres SF®+cells(Sf9-derived cell line)(US Patent#6,103,526) and Protein Sciences Formulary Media(PSFM)are proprietary to Protein Sciences Corporation(PSC,Meriden,CT).Cells were rou-tinely grown and maintained in250-mL(50-mL working volume) shake-flasks(Corning)in rum-free PSFM at28◦C and110rpm. Baculovirus vectors expressing human hemagglutinin proteins under the control of polyhedrin promoter were constructed at Pro-tein Sciences Corporation(PSC,CT).The recombinant baculovirus were amplified by infecting expres SF®+cells(SF+cells)and har-vested3days post-infection as previously described by Wang et al.[10].The virus stock solution titers were determined by plaque assay as previously described by O’Reilly et al.[21].Low pas-sage viral stocks(passage#4)were prepared in shake-flask SF+ cultures and stored at4◦C.The rHA derived from the follow-ing strains were ud in this study:A/New Caledonia/20/1999 (H1N1),B/Malaysia/2506/2004,H5/Vietnam/1203/2004(H5N1), A/Wisconsin/67/2005(H3N2),and A/S
olomon Islands/03/2006 (H1N1).The rHA designation refers to the baculovirus construct carrying a specific hemagglutinin gene encoding for the expression of one of thefive rHA proteins.
2.2.Feeding solution preparation
All concentrates and solution components of the feed cocktails were prepared with Milli-Q water and cell culture grade bio-chemical andfilter-sterilized through a0.2-mfilter(Millipore). The concentrated nutrients were stored either at4or−20◦C and are combined just before adding to the cell cultures.The orig-inal feeding solution is referred to as the generic cocktail and contained concentrated non-animal derived hydrolysates,carbo-hydrates,lipids,amino acids and vitamins and trace elements. Overall,this feeding solution consisted of44ingredients that were combined just prior to addition to the cell cultures as detailed by Bédard et al.[22]and summarized in Table1.
2.3.Development of the feeding solution for SF+cells
The composition of the feeding solution was simplified by re-asssing the effect of each component one by one to support cell growth and sustain rHA productivity at a targeted cell density.The cells were eded at a viable cell density(VCD)of1.0×106cells/mL and fed with the different solutions at VCD of2.5×106cells/mL. Cells were infected using a multiplicity of infection(MOI)of1.
2.4.Determination of cell density at infection in fed-batch
cultures
Unless otherwi stated,all small-scale experiments were per-formed in Erlenmeyer shake-flasks(Corning,NY)in duplicate.
A fed-batch experiment was initiated by inoculating one300-mL culture at 1.0×106cells/mL in PSFM media(PSC).At VCD of2.5×106cells/mL,cells were distributed in250-mL disposable
J.Meghrous et al./Vaccine28 (2010) 309–316311
Table1
List of the components ud for the preparation of the generic cocktail.
Ingredients of the cocktail Composition of the solutions ud in the cocktail
Amino acids l-Arginine,l-asparagine,l-aspartic acid,l-glutamic acid,l-glutamine,l-glycine,l-histine,l-isoleucine,l-leucine,l-lysine,
l-methionine,l-phenylalanine,l-proline,dl-rine,l-threonine,l-tryptophan,l-valine,l-cysteine·HCl·H2O,tyrosine
Vitamins solution Thiamine·HCl·1/2H2O,riboflavin,d-calcium pantothenate,pyridoxine·HCl,para-aminobenzoic acid,nicotinic acid,i-inositol,
biotin,choline chloride,vitamin B12,folic acid
Trace elements Molybdic acid ammonium salt,cobalt chloride hexahydrate,cupric chloride·2H2O,mangane chloride·4H2O,zinc chloride,
ferrous sulfate·7H2O
Lipid emulsion preparation a Cholesterol,␣-tocopherol acetate,cold liver oil fatty acid methyl esters,Tween80and Pluronic F-68
Yeastolate solution50×Yeastolate solution(Invitrogen)
Gluco solution Gluco concentrate
a Lipid emulsion solution was freshly prepared by combining1000×lipid/ethanol mixture and10%of Pluronic F-68(SAFC BioSciences).eyes wide shut
shake-flasks(working volume of40mL)and fed with the cocktail solution.Feeding was carried out24h before the viral infection. Fed-batch cell cultures were infected during the mid-exponential growth pha,at a targeted cell density of infection(CDI)of2.5,3.0,
4.0,
5.0and8.0×106cells/mL with recombinant baculovirus using
函授大学本科a MOI of1.Cells were eventually fed again at the time of infection depending on the targeted CDI.A batch culture of infected SF+cells at a VCD of2.5×106cells/mL was run in parallel to the fed-batch experiments as a control.
2.5.Validation of the fed-batch cultures in bioreactors
The fed-batch cultures were performed in a controlled3.5L water jacket Chemap CF-3000bioreactor(Mannedorf,Switzerland) equipped with three surface baffles and pitched blade impellers as previously described[23].pH was monitored using a gel electrode (Mettler-Toledo GmbH,
Switzerland).Dissolved oxygen(DO)was measured with an Ingold polarographic oxygen probe,and con-trolled with a mixture of oxygen and nitrogen at a constant level of 40%of air saturation byflushing the headspace environment with a total gasflow of0.3L/min.Oxygen was also sparged into the culture, using a polypropylene sparger,with250m porosity atflow rate from0.01L/min up to0.06/min of the total(0.3L/min)gasflow.Cul-tivation was controlled and monitored on-line with the interface FIX MMI software(Intellution,Norwood,MA,USA)for data acqui-sition.Fed-batch bioreactor cultures were started by inoculating the cells at VCD of1.0×106cells/mL.Cells were initially grown in batch mode for2days with a specific growth rate between0.026 and0.034h−1.At VCD of2.5×106cells/mL,170mL of feeding solu-tion was added to the cell cultures.Cells were then allowed to grow to a targeted cell density of4.0×106cells/mL,and subquently infected with60mL of baculoviral stock(passage4).
2.6.Reproducibility and performance of the fed-batch process
86届奥斯卡颁奖典礼The fed-batch process was validated by asssing four rHA encoding baculovirus in duplicate in 3.5-L biore-actors(working volume of2L):A/New Caledonia/20/1999 (H1N1),B/Malaysia/2506/2004,H5/Vietnam/1203/2004(H5N1), A/Wisconsin/67/2005(H3N2).In addition,fed-batch cultures at10-L scale were performed using A/New Caledonia/20/1999 (H1N1)and
A/Solomon Islands/03/2006(H1N1)encoding bac-ulovirus in15-L glass Applikon bioreactor(Applikon,Foster City, CA).Culture temperature,dissolved oxygen(DO)were automat-ically monitored and controlled as standard t-points using an Applikon controller ADI1012(Applikon).pH and gasflow rates were monitored by the same system.
2.7.Analytical methods
Infected cultures were sampled daily(4×1mL).Cell concen-tration,cell viability and cell diameter were determined using a Cedex(Innovatis,Germany)from the same sample.The three remaining1-mL samples were centrifuged at3000rpm for5min and the supernatants were carefully removed and stored at−80◦C for subquent gluco,lactate,ammonia and amino acid analy-sis.The concentrations of lactate and ammonia produced as well as the residual gluco were measured using IBI Biolyzer Rapid Analysis System(Kodak,New Haven,CT).Amino acids were quan-tified by rever pha high performance liquid chromatography (HPLC)using AccQ Tag method following manufacturer’s instruc-tions(Waters,Milford,MA).Cell pellets were incubated with rHA extraction buffer for30min(each rHA strain has a specific extrac-tion buffer).Cell extracts were then centrifuged at3000rpm for 20min at4◦C and the supernatant containing extracted recom-binant rHA proteins were collected and analyzed by single radial immunodiffusion assay(SRID)as previously described[2
4,25].The SRID reference standard reagents as well as anti-hemagglutinin ra or rHA strains were obtained from Food&Drug Administration (FDA).SRID assay was optimized and validated at PSC(Meriden,CT). The identity of extracted rHA recombinant proteins was confirmed by running western blot.
2.8.Purification of hemagglutinin
A protocol for hemagglutinin purification was previously described by Wang et al.[10].A new protocol has been developed. Briefly,the cell pellets were extracted with an appropriate buffer for30min at room temperature.The cell lysate was centrifuged at 12,000rpm for10min at4◦C and the supernatants were clarified and concentrated through a depthfiltration.The lysate was quen-tially purified using an ion-exchange SP-Sepharo Big Beads chromatography(SP-Sepharo BB,GE Healthcare,Uppsala,Swe-den)and a hydrophobic interaction chromatography(Phenyl High Performance,Phenyl HP)(HIC,GE Healthcare).The eluted fraction was then applied to an Ion-Exchange Membrane Chromatography (Q chromatography,Pall Corporation,NY)to remove the resid-ual DNA and theflow-through fraction was concentrated using ultrafiltration(Ultrafiltration casttes,Sartorius Stedim Biotech, Gottingen,Germany).
3.Results
The main objective of this study was to develop a streamlined and versatile fed-batch process that could be implemented at PSC core production facility.
3.1.Asssment of the original feeding solution and fed-batch strategy to sustain high SF+cell density and rHA production in shake-flasks experiments
The preliminary experiments in shake-flasks showed that amino acids were depleted from the culture medium,suggesting that the cells stopped producing due to nutrient limitation(data not
312J.Meghrous et al./Vaccine
28 (2010) 309–316
Fig.1.Cell growth profile and hemagglutinin production by the infected SF+cells in fed-batch and control batch cultures in shake-flask experiments.For the fed-batch cultures SF+cells were infected by the baculovirus harboring the H1N1type (A/New Caledonia/20/1999)at approximately8×106cells/mL at an MOI of1.The feed was added24h prior to infection and at the time of infection as shown by the solid vertical arrows.Clod and open symbols reprent the fed-batch and batch cultures,respectively;total cells(squares),cell viability(triangles).The inrt shows the volumetric HA production expresd in mg/L for batch and fed-batch cultures after72hpi.Data reprent the average of two independent experiments.Error bars reprent standard deviation.
shown).The original feeding solution described in Section2was evaluated in shake-flask experiments to improve both cell growth and rHA product yields using A/New Caledonia/20/1999(H1N1) strain as a model.Twice,the feeding solution was added to the culture as single pul-batch;thefirst feed was added24h before infection and the cond at the time of infection to avoid any potential nutrients limitation in high cell density cultures.The feed additions sustained a successful growth and infection at a density of infected SF+cells as high as8.0×106cells/mL(Fig.1).The viable cells consistently maintained a higher concentration in fed-batch cultures.This result was not surprising since this feedi
ng solution was originally designed to support very high cell density cultures and very likely contained an excess of the necessary nutrients.The maximum cell density after infection remained unchanged which suggested synchronous infection process of the infected cells. More importantly,the culture showed an extended lifespan of the infected SF+cells from2days in batch culture to3days in fed-batch culture.At72h post-infection(hpi),the cell viability in the fed-batch cultures was maintained at70%,which is significantly higher than the30%obtained for the batch cultures.No significant difference in cell size was obrved between the infected SF+cells in batch and fed-batch cultures;the average cell size incread by more than30%post-infection for both culture mode.Cell size is a reliable and early indicator for the cell infection process.The data showed that the feeding solutions enhanced the cell density at infection and extended the cell viability as compared to the batch cultures.
More importantly,the feed additions sustained a successful pro-duction at density of infected SF+cells as high as8.0×106cells/mL, which is more than three-times higher than the density obtained in the batch cultures(Fig.1).Batch culture lasts approximately3 days and achieves an average concentration of20mg of rHA per liter of cultures[10,this study].Improvement of the cell viabil-ity and cell density at infection resulted also in an increa in rHA yields in fed-batch cultures.r
HA yields,as measured by SRID assay, achieved a value of62mg/L in the fed-batch versus20mg/L in batch cultures(e inrt of Fig.1),corresponding to more than3-fold increa over the batch control.Additional experiments with fur-ther additions of the cocktail at similar cell density did not improve the volumetric yields beyond62mg/L.Table2
Comparison of feed solutions for rHA production in fed-batch cultures using SF+ cells in shake-flasks.
Feeding solution rHA yields
admiraltyVolumetric(mg/L)Specific(mg/109cells)
First-generation48.7±3.07.2
Second-generation46.4±4.07.7
Generic feeding solution61.6±4.17.2
Batch cultures20.0±2.7 6.8
Data reprent the average and standard of two independent experiments.
igt
Two feeding additions were carried out(24h prior cell infection and at time of infection).rHA yields are determined by SRID assay.
3.2.Streamlining andfine-tuning the feed for the fed-batch
process
Although the original feeding solution was successfully ud to maximize rHA productivity in SF+cell cultures,its addition involved a complex procedure using multiple solutions as a con-quence of the low solubility of some ingredients.A cond drawback is the increa in the culture osmolality following the feed addition that might be solved through a continuous feeding but this will add to the complexity of operations.Hence,we decided to further simplify the formulation of the generic cocktail to facil-itate its preparation and better align the fed-batch process with the current manufacturing process at PSC.Eight different formu-lations were prepared according to the metabolic requirements of SF+cells to promote high cell densities and maintain specific production yield of rHA.Among the feeding solutions,one sim-plified cocktail met PSC requirements with regard to the yields and cell density;it showed comparable profiles of cell viability and slightly higher specific rHA production as compared to the orig-inally designed feed solution.Furthermore,the osmolality of the cell growth medi
um at the time of feeding infected cells fed with the simplified feed was comparable to that of the PSFM growth media and be added to the culture in a single pul-batch instead of a mi-continuous mode of addition as previously done with the generic cocktail.Although the formulation of the cocktail was reasonably improved and yielded comparable values of rHA pro-duction,the precipitation of some trace elements was still one of its drawbacks.This required further reducing the formulation of the simplified feeding solution while maintaining a high productivity. Table2summarizes the results obtained with the feeding solutions tested for their performance to sustain high rHA yields.Bothfirst-and cond-generation feeds yielded comparable rHA values while maintaining constantly the specific rHA yields around7.2mg/109 cells,comparable or even slightly higher to what was obtained with the generic feed.
3.3.Determination of the peak cell density for rHA production in shake-flask experiments
The cond-modified feeding solution was lected for its sim-plicity and ud in the rest of the study to furtherfine-tune the fed-batch process by optimizing the cell density at infection.Sev-eral CDI ranging from2.5to8.0×106cells/mL were investigated. Two pul-batch feeds were added quentially to the culture to avoid potential nutrient limitations:thefirst feed was added24h before infection and the cond at the time of infection.
At a CDI of2.5×106cells/mL,a similar cell density ud in batch production,the feed addition resulted in an increa of the volumetric yield indicating possible limitations in current batch operations.rHA production incread in proportion to the cell den-sity at infection and leveled off at around4.0×106cells/mL and 55mg/L of rHA concentration(Fig.2).Doubling the CDI up to
J.Meghrous et al./Vaccine28 (2010) 309–316
313
Fig.2.Effect of cell density at infection(CDI)on hemagglutinin production in fed-batch cultures.One feed was added to the cells infected by the baculovirus harboring the H1N1type(A/New Caledonia/20/1999)at CDI of2.5and3.0×106cells/mL whereas two subquent feeds were added to the cultures for CDIs of 4.0, 5.0and8.0×106cells/mL.In the batch control cells were infected at a CDI of 2.5×106cells/mL.Values are an average of at least two independent experiments. Error bars reprent standard deviation.
8.0×106cells/mL had a marginal effect on rHA production and the yields remained around59mg/L.The data indicated that a CDI of 4.0×106cells/mL can be ud for fed-batch operation with the sim-plified feed.Furthermore,minor scale-up modifications to the DSP protocol would be needed to process cultures at4.0×106cells/mL versus2.5×106cells/mL in the current batch process.
A similar trend was also obrved with the specific produc-tivity,which was consistently maintained at high values for CDIs ranging between2.5and5.0×106cells/mL before dropping below 8.0mg/109cells at a CDI of8.0×106cells/mL.The highest specific rHA yield was achieved at CDI of4.0×106cells/mL(1
2mg/109 cells),which is higher than the6.8mg rHA proteins per109cells obtained in batch control.
The cond step in simplifying the fed-batch process con-sisted in evaluating the feeding strategy by using a single addition of the feed instead of two,given that the peak cell density is around4.0×106cells/mL,and assuming that this feed concentrate could sustain this cell density.Interestingly,a single feed addition appeared to be sufficient to sustain cell viability and rHA produc-tivity to comparable levels to what was obtained with two feeds as previously described,indicating that the cond feed was probably not esntial at a CDI of4.0×106cells/mL.The operation simplifi-cations offer the advantage of reducing the number of intervention and thereby contribute to streamlining the fed-batch process.
3.4.Validation of fed-batch process in3-L bioreactor scale
The next step was to validate the simplified feeding strategy in bioreactor operations and to demonstrate scale-up before imple-mentation of the fed-batch process for rHA production in pilot-scale bioreactors.The simplified fed-batch protocol as described in the last ction was operated in two bioreactor runs and the results are shown in Fig.3A.Both cell growth and cell viability profiles are comparable for the two bioreactor runs showing a similar trend to previously obtained results in shak
e-flasks experiments.The aver-age cell viability in fed-batch cultures was around80%after51hpi. Recombinant HA production kinetics has been also evaluated indi-cating an optimal harvest time for maximal rHA yield around50h post-infection.Overall,the results confirm the potential scalabil-ity of the fed-batch process.
Operations in bioreactors allowed frequent sampling of large volumes to analyze the basic metabolites in SF+fed-batch pro-duction under controlled conditions.The gluco consumption and lactate and ammonia accumulation during cell infection are shown in Fig.3B.The gluco concentrations gradually decread during cell infection period and about60%of its initial concentration in the medium was consumed by SF+cells.Accumulation of ammo-nia incread from2.8to6.4mM over the entire cultures whereas lactate remained at1.0mM during growth and declined to nil after 44hpi.Accumulation of lactate resumed after51h post-infection indicating a shift in metabolism or relea in the media becau of decline of culture viability.None of the19measured amino acids appeared exhausted in the fed-batch cultures.For example,Fig.4 shows ratios of the most consumed amino acids in the cultures as measured at92hpi.Both glutamine and rine were preferably con-sumed as compared to other amino acids.Therefore we concluded that none of the measured amino acids was limiting in fed-batch production with SF+cells at4.0×106cells/mL.
Fig.5summarizes rHA production in3-L fed-batch and batch cultures and compares to the results from shake-flask experiments. An average of44mg of rHA per liter of cultures was obtained in3-L bioreactor fed-batch cultures,significantly higher than the20mg/L routinely achieved in batch cultures.This improvement reprents almost2.5-fold increa over the batch culture,while the spe-cific productivity was slightly higher in fed-batch cultures(9and 7mg/109cells for fed-batch and batch cultures,respectively).This data shows clearly that the simplified feeding solution was appro-priate for the fed-batch production process at4.0×106
cells/mL
Fig.3.Production of baculovirus H1N1type(A/New Caledonia/20/1999)infected SF+cells in3-L bioreactor fed-batch cultures with simplified cond-generation feed. Cell cultures were infected at4.0×106cells/mL.One single feed was added24h prior to infection at a target cell density of2.5×106cells/mL as indicated by the vertical arrow.Data reprent the average of two bioreactor runs.Error bars repre-nt standard deviation.(A)Cell growth profile and hemagglutinin total cells(open squares),cell viability(clod triangles).Histogram reprents the kinetic of HA production and numbers indicate the exact time of sampling.(B)Gluco(clod squares)consumption and lactate(open circles)and ammonia(clod triangles) fed-batch productions.
