Conversion of MDCK cell line to suspension cultureby transfecting with human siat7e gene and itsapplication for influenza virus productionChia Chua,b, Vladimir Lugovtsevc, Hana Goldingc, Michael Betenbaugha, and Joseph Shiloachb,1

aDepartment of Chemical and Biomolecular Engineering, Johns Hopkins University, 3400 N. Charles Street, Baltimore, MD 21218; and bBiotechnology CoreLaboratory, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, and cCenter for Biologics Evaluation andResearch, Food and Drug Administration, 9000 Rockville Pike, Bethesda, MD 20892

Communicated by John B. Robbins, National Institutes of Health, Bethesda, MD, July 14, 2009 (received for review March 10, 2009)

MDCK cells are currently being considered as an alternative toembryonated eggs for influenza virus propagation and hemagglu-tinin (HA) production intended for vaccine manufacturing. MDCKcells were found suitable for the virus production but their inabilityto grow in suspension burdens the process of scale up and hencetheir production capability. Anchorage-dependent MDCK cellswere converted to anchorage-independent cells, capable of grow-ing in suspension as a result of transfection with the human siat7egene (ST6GalNac V). This gene was previously identified as havingan important role in cellular adhesion when the transcriptions ofgenes from anchorage-dependent and anchorage-independentHeLa cells were compared. Unlike the parental MDCK cells, thesiat7e-expressing cells were capable of growing in shake flasks assuspension cultures, achieving maximum concentration of 7 � 105

cells/mL while keeping close to 100% viability throughout thegrowth phase. In production experiments, the siat7e-expressingcells were infected with the Influenza B/Victoria/504/2000 strain. Itwas determined that the cell-derived viruses retained similarantigenic properties as those obtained from egg-derived virusesand their nucleotide sequences were identical. The specific pro-duction of hemagglutinin (expressed in hemagglutination unitsper 106 cells) from the siat7e-expressing cells was approximately 20times higher than the specific production from the parental MDCKcells. If this suspension process scales up, the production potentialof HA from 10 L of siat7e-expressing cells at a concentration of 106

cells/mL would be equivalent to the amount of HA obtained from10,000 embryonated eggs.

anchorage-independent � hemagglutinin � sialyltransferase � vaccine

Influenza-related illnesses cause an estimated 100,000 hospi-talizations and tens of thousands of deaths in the United States

annually (1). In response to rapid antigenic drift in influenzaviruses, the most effective approach taken has been the distri-bution of trivalent inactivated viral vaccines, which are tradi-tionally produced in chicken embryonated eggs (2). However, inthe event of a pandemic outbreak, this egg-based productionsystem may not be adequate to meet the surge in demand quicklyenough. The limitations associated with egg-based vaccines,which include reliable egg supplies, prolonged cultivation peri-ods, and cumbersome operations have spurred exploration ofalternatives. Among the potential alternatives for vaccine pro-duction, the use of characterized, immortalized cell lines (par-ticularly MDCK, VERO, and PER.C6) has been investigated.These cell lines have been found to produce consistently highviral titers (3–8). Nevertheless, one of the limiting aspects inscaling up the virus production in these continuous cell lines isthe fact that these cells are anchorage-dependent and thusrequire surface adhesion to proliferate (9, 10). Without surfaceattachment, these cells cannot exert their normal cyclin-dependent kinase activity through the signaling cascades initial-ized by interactions between integrins and extracellular matrix(11–15). For industrial production in bioreactors, the required

surface area can be provided using microcarrier beads (16–19).Although this approach is sufficient to obtain high virus pro-duction yield (18, 19), this propagation strategy is cumbersomecompared with propagation of cells in suspension. An MDCKcell line that can proliferate in suspension would greatly facilitatethe scale-up process of influenza virus production.

In a previous study we compared the transcription profiles ofanchorage-dependent and anchorage-independent HeLa cellsusing DNA microarrays (20). The gene siat7e (ST6GalNac V)was identified as one of the genes that play a role in controllingthe degree of cell adhesion. It was shown that higher siat7etranscription corresponded to a lower degree of adhesion bymicroscopic evaluation and by monitoring cell detachment in ashear flow chamber (20), and inhibiting siat7e transcription usingsiRNA was followed by enhanced adhesion. The human sialyl-transferase ST6GalNac V, a member of the ST6GalNac familyof sialyltransferases, is a type II Golgi membrane protein thattransfers sialic acid from the donor CMP-Neu5Ac to the GalNacresidue on the ganglioside, GM1b, forming GD1�. Tsuchida etal. (21) proposed indirect involvement of siat7e in synthesizingdisialyl Lea, a carbohydrate structure conjugated to proteins andceramides on the cell surface. In other studies, glycosphingolip-ids including gangliosides have been reported to mediate celladhesion through the sugar residue interactions in the glycosyn-apse microdomains (22, 23). These reports are consistent withour findings on the relationship between siat7e gene expressionand cell adhesion. As was indicated earlier, MDCK cells are goodproducers of several viruses including influenza A and B viruses.The conversion of these anchorage-dependent cells to cellscapable of growing in suspension will simplify the productionprocess and has the potential to supplant current productionprocedures in chicken embryonated eggs. In the present work wereport on the transfection of the anchorage-dependent MDCKcells with the human siat7e gene, on the properties of thesiat7e-expressing cells and on their capability to produce theinfluenza virus.

ResultsTransfection of MDCK Cells with Human siat7e and Its Effects onCell–Cell Adhesion and Cell Spreading. Anchorage-dependentMDCK cells exhibited changes in cell-cell adhesion and cellspreading behavior following the incorporation of the humansiat7e gene as shown in Fig. 1. Cells transfected with the siat7eshown in Fig. 1B (clone 1) and C (clone 2) appear to spread lesson the cell culture flask than the parental cells shown in Fig. 1 A;the siat7e-expressing cells also lost their ability to form a tight

Author contributions: C.C., M.B., and J.S. designed research; C.C. and V.L. performedresearch; C.C., V.L., H.G., M.B., and J.S. analyzed data; and C.C., V.L., and J.S. wrote thepaper.

The authors declare no conflict of interest.

Freely available online through the PNAS open access option.

1To whom correspondence should be addressed. E-mail: yossi@nih.gov.

14802–14807 � PNAS � September 1, 2009 � vol. 106 � no. 35 www.pnas.org�cgi�doi�10.1073�pnas.0905912106

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

26, 2

020

junctions with the neighboring cells. It was also observed thatwhen the siat7e-expressing cells undergo prolonged culture,some cells would self-detach while maintaining their viability.Assessment of transfection efficiencies with the siat7e plasmidusing the FACSCalibur machine showed that approximately 4%of MDCK cells were transfected 24 h after introducing theplasmid vector.

Gene Expression Differences Between the Parental and the siat7e-Expressing MDCK Cells. The detection of the human siat7e mRNAin the parental and the siat7e-expressing cells and the expressionof a housekeeping gene (endogenous GAPDH) are seen in Fig.2A. It is clear that there is expression of human siat7e in thetransfected cells but no expression in the parental cells, whileGADPH expression was detected in all samples. Real-time PCRwas performed to quantify the expression of siat7e and theexpression of the housekeeping gene in clones 1 and 2 (Fig. 2B).The increase in the siat7e expression was correlated with thedegree of cell-cell adhesion and cell spreading of these twotransfected clones seen in Fig. 1.

Surface Charge Differences Between the Parental and the siat7e-Expressing MDCK Cells. To assess cell surface differences betweenthe two cell lines, the cell surface charge was measured usingFITC-labeled cationized ferritin (24–26). Interactions between

cationized ferritin and negative charge sites on the cell surfacewould allow charge measurements on cell surfaces by quantifyingthe amount of bound ferritin molecules (25). The signal profilesfrom each cell line, with and without ferritin treatment, areshown in Fig. 3. Flow cytometric analysis showed a shift in theoverall signal distribution of the siat7e-expressing cells (Fig. 3B).The shift indicates higher signal intensities emitted from thefluorescein (FITC), which should correspond to higher numberof anionic sites on the membrane surface. No difference wasobserved when the ferritin was not present (Fig. 3A).

Growth Kinetics in Monolayer and Suspension of the Parental andsiat7e-Expressing MDCK Cells. Growth, viability, glucose consump-tion and lactate production of the parental and the siat7e-expressing MDCK cells grown as a monolayer in T flasks areshown in Fig. 4 A–C and as suspension culture in Fig. 4 D–F. Thesiat7e-expressing cells grew less than the parental cells in the Tflask (Fig. 4A). Their density reached 7 � 104 cells/cm2 com-pared to 2 � 105 cells/cm2 of the parental cells after 179 h ofgrowth, although the percent viability of the cells was similar(Fig. 4B). Glucose consumption and lactate production in thetwo cell lines were similar until the siat7e-expressing cellsapproached peak density in the T flasks as shown in Fig. 4C.

Fig. 1. Parental and siat7e-expressing MDCK cells grown in T flasks. (A)Parental MDCK cells. (B) Clone 1, isolated from the siat7e-expressing pool. (C)Clone 2, isolated from the siat7e-expressing pool.

Fig. 2. mRNA expression of human siat7e and endogenous GAPDH inparental MDCK and in clones 1 and 2 of the siat7e-expressing cells. (A)End-point RT-PCR. (B) Real-time PCR.

Fig. 3. FITC signal distribution obtained by FACS analysis of parental andsiat7e-expressing MDCK cells with and without ferritin. (A) Without ferritintreatment; (B) with ferritin treatment. Parental MDCK cells (gray line), siat7e-expressing cells (black line).

Chu et al. PNAS � September 1, 2009 � vol. 106 � no. 35 � 14803

APP

LIED

BIO

LOG

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

26, 2

020

Opposite growth trends were observed when the two types ofcells were propagated in shake flasks. The growth curve (Fig.4D) demonstrates that siat7e-expressing cells were able to pro-liferate in suspension culture, whereas the parental cells couldnot. The siat7e-expressing cells grew exponentially to a concen-tration of 7 � 105 cells/mL. High viabilities (Fig. 4E) of thesiat7e-expressing cells were seen throughout the 12-day growth.These cells were at least 90% viable while the viability of theparental MDCK cells declined steadily over the culture period.The glucose and lactate profiles shown in Fig. 4F indicate thatparental MDCK cells consumed more glucose and producedmore lactate than the siat7e-expressing cells especially at laterculture times when cell densities were greater in the siat7e-expressing cells. Microscopic analysis at the end of the growthshowed that the surviving parental MDCK cells were aggregatedin large clumps, while the siat7e-expressing cells, on the otherhand, appeared healthy and were suspended primarily as indi-vidual cells.

Influenza Virus Growth and HA Titer in Parental and siat7e-ExpressingMDCK Cells. The yield of influenza virus in parental and siat7e-expressing MDCK cells was evaluated by analysis of growthkinetics of a model virus B/Victoria/504/2000 per 106 cells.

Summarized in Table 1 are the highest values of both the viraland the HA titers. The values were obtained 36 to 48 h postinfection in the case of the adherent cells and 24–38 h in the caseof cells grown in suspension. The viral infectivity titers, ex-pressed as 50% egg infectious dose per mL (EID50/mL), weresimilar in three growth conditions: monolayer culture of theanchorage-dependent parental MDCK cells, monolayer cultureof the siat7e-expressing cells and the siat7e-expressing cellsgrown in suspension. However, remarkable differences wereobserved for HA titers, expressed in hemagglutinating units(HAU). When calculated per 106 cells, 2,155 HAU was obtainedfrom the parental MDCK cells, 8,606 HAU from the siat7e-expressing cells grown in monolayer, and 54,348 HAU from thesiat7e-expressing cells grown in suspension in shake flasks.Shown in Fig. 5 is the cell viability of the infected siat7e-expressing cells grown in suspension and the HA titers over thetime course of one representative kinetic experiment.

Virus Antigenic Stability During Replication in Parental MDCK Cellsand siat7e-Expressing Cells. The effect of different cell substrateson virus antigenic properties was evaluated in hemagglutinationinhibition test (HAI). The HAI titers of three ferret sera thatwere infected with egg-grown reference virus B/Victoria/504/

Fig. 4. Growth parameters of parental MDCK cells (open circles) and siat7e-expressing MDCK cells (open squares) in shake flask in suspension and in monolayerin T flasks. T flasks: (A) Growth in viable cell density (VCD). (B) Viability %. (C) Glucose consumption and lactate production (shaded) in g/L. Shake flasks: (D) Growthin viable cell density (VCD). (E) Viability %. (F) Glucose consumption and lactate production (shaded) in g/L.

Table 1. Virus titers in different cell substrates

Substrate

Viral titer Virus titer per 106 cells

HAU/mL EID50/mL, log10 HAU EID50, log10

MDCK monolayer 1,810 8.35 � 0.17 2,155 8.42siat7e-expressing cells monolayer 5,120 6.90 � 0.12 8,606 7.12siat7e-expressing cells* suspension 40,960 7.87 � 0.12 54,348 8.00

Influenza strain B/Victoria/504/2000 was used to infect the substrates between M.O.I.s of 1.0 and 2.0 TCID50.Hemagglutinin titers and infectious titers were measured using supernatant from whole cell lysate samples.*Cells were infected at 107/mL density in suspension culture and then diluted to 106/mL for propagation.

14804 � www.pnas.org�cgi�doi�10.1073�pnas.0905912106 Chu et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

26, 2

020

2000 were determined using the B/Victoria output virus from theparental MDCK cells and the siat7e-expressing MDCK cellsgrown either in monolayers or in suspension. The results areshown in Table 2. In all cases, the sera titers were within two-folddifference, demonstrating that cell-derived viruses were as an-tigenic as those obtained from the egg-derived reference virus.Direct DNA sequencing of RT-PCR products amplified fromHA and NA (neuraminidase) viral gene segments, showed thatthe cell-derived viruses and the egg-derived reference virus hadidentical nucleotide sequences. These data demonstrate thatreplication of the virus in parental or siat7e-expressing cells didnot alter the antigenic properties of the virus.

DiscussionIn a previous study, we identified two genes that have a role in celladhesion (20): siat7e, a type II membrane glycosylating sialyltrans-ferase, and lama4 which encodes laminin �4, a member of thelaminin family of glycoproteins. These two genes were identifiedfollowing a comparison of gene transcription of two phenotypicallydistinct HeLa cells, anchorage-dependent and anchorage-independent. It was demonstrated that decreased expression ofsiat7e in the anchorage-independent HeLa cells, or enhancedexpression of the lama4, resulted in greater aggregation andmorphological changes compared with the untreated anchorage-independent HeLa cells. An opposite effect was observed whenexpression of the siat7e was increased and the lama4 expression wasdecreased in the anchorage-dependent HeLa cells.

Engineering cell lines to improve biotechnology processes wasone of our major aims of the previous study. One of the importantcommercial production processes that are still in need of a well-defined cell line is the production of influenza virus. Influenza virus

is currently being produced in embryonated eggs (27). Since theproduction in eggs is quite cumbersome and time consuming,replacing the embryonated eggs process with mammalian cells,especially MDCK cells, is an appealing option (4, 6, 8). SinceMDCK cells are anchorage-dependent, the replacement of theembryonated eggs with these cells will introduce additional pro-cessing difficulty. Specifically, these processes often require inclu-sion of microcarriers, which are difficult to scale-up. Microcarriersrequire multiple additional processing steps including seeding ofcells on a surface for subsequent growth (which may be limited bysurface area), maintenance of the microcarriers in the bioreactoralong with sufficient mixing to avoid mass transfer limitation to andon the beads, and separation of the product from the microcarriers.Not surprising, current production of monoclonal antibodies andmost other biopharmaceuticals by mammalian culture utilizes celllines such as Chinese Hamster Ovary (CHO), Baby HamsterKidney (BHK), and NS0 adapted to suspension culture. Conversionof the MDCK cells to grow in suspension would considerablysimplify the production process of influenza vaccine. Although thecurrent state-of-art has already used suspension MDCK cell cul-tures to produce influenza vaccines (28), we report the targetedengineering of the MDCK cell lines to anchorage-independentculture while others relied on long serial passaging approach.

Incorporation of the human siat7e gene into the MDCK cells,as demonstrated in this study resulted in their conversion toanchorage-independent cells. The cells grew well in suspensionin shake flasks. The cultures reached a concentration of 7 � 105

cells/mL maintaining at least 90% viability throughout thegrowth period. The human gene was successfully incorporatedand transcribed, (Fig. 2 A) modifying considerably the cellphenotype. By using CF-FITC it was possible to determine thatthere is a change in the net charge on the surface of thesiat7e-expressing cells. The increased negative charge may beassociated with the increased number of sialic acids moietiesattached to the cell surface gangliosides by siat7e. Elevatednegative charge of the cell surface may contribute to a decreasedcell-to-surface adhesion and to electrostatic repulsion betweencells, and thus allowing the cells to grow in suspension.

Siat7e-expressing cells were not only able to grow in suspen-sion and to produce identical virus as the one produced inembryonated eggs, their specific production of HA was about 20times higher than the anchorage-dependent parental cells. Re-cent publications have reported on the expression of the humansiat1 gene (ST6Gal I) in MDCK cells that was associated with anincrease in �2,6-linked sialic acid on glycoproteins (29–31). Theresearchers wanted to increase the number of 6-linked sialicacids on the cell surface to enhance influenza virus sensitivity toneuraminidase inhibitor, which is the key component of antiviraldrugs for influenza. In addition, a two-log increase in theproduction of human influenza viruses in these cells was re-ported (29). Unlike ST6Gal I, ST6GalNac V expressed in thiscurrent study is responsible for adding sialic acid to the GalNacresidue located on the side position of oligosaccharide chainsinstead of the common terminal position on the Gal residue. Theinfectivity with the B/Victoria/504/2000 of the siat7e-expressingcells measured as EID50/cell was found to be slightly lower thanthat of the parental cells but the HA production was found to behigher. In a seasonal influenza vaccine, a typical dose is com-posed of 15 �g purified HA from each of the three selectedinfluenza strains (H1N1, H3N2, and B). Since the vaccine ischaracterized by the amount of the HA, it is an additional benefitof these cells. In work conducted by Tree et al., production ofinfluenza virus A strain in MDCK cells grown on various typesof microcarriers were compared to chicken eggs. They reported5.0 � 104 HAU/mL in cells grown on Cytodex I in spinner cultureand 2.0 � 105 HAU/mL in chicken eggs. Based on this infor-mation, they estimated that 1,000 L MDCK cells grown on solidmicrocarriers would be equivalent to roughly 30,000 eggs, or 1 L

Fig. 5. HA production (open squares) and cell viability following infection ofsiat7e-expressing MDCK cells with influenza B virus (filled circles). Cell viabilityof siat7e-expressing MDCK cells without infection are also shown (opencircles).

Table 2. HAI titers with viruses from different cell substrates

Substrate

HAI titers

Sera 1 Sera 2 Sera 3

Chicken eggs 128* 256 256siat7e-expressing cells†

suspension256 512 512

siat7e-expressing cells monolayer 64 128 128MDCK monolayer 64 128 128

Sera were obtained from three ferrets 3 weeks after intranasal infectionwith egg-derived reference virus B/Victoria/504/2000.*Reciprocal of the highest dilution of serum capable of completely inhibitingHA activity of the respective virus. Data are from a single representativeexperiment.

†Cells were infected at 107/mL density in suspension culture and then dilutedto 106/mL for propagation.

Chu et al. PNAS � September 1, 2009 � vol. 106 � no. 35 � 14805

APP

LIED

BIO

LOG

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

26, 2

020

would be equivalent to about 30 eggs. In another study, titer of2.5 � 103 HAU/mL was obtained in a stirred tank reactor usingCytodex I microcarriers (16) and a maximum titer of approxi-mately 4.0 � 104 HAU/mL was obtained in WAVE cellbagsusing Cytodex I microcarriers (18). It is important to note thatin these studies the Influenza A virus strain was used. For theInfluenza B strains, which was used in our study, the HA titersare commonly on the scale of 100 (32, 33). Based on the HAproduction capability, we estimate that 1 mL of culture contain-ing 106 cells can produce approximately 40,000 HAU. Takinginto consideration that the average production of HAU per eggis also around 40,000, 10-L culture of the siat7e-expressingMDCK cells can produce the equivalent amount of HAUproduced in 10,000 embryonated eggs. These data demonstratethat the established siat7e-expressing MDCK cell line has thepotential to significantly increase the efficiency of manufactureof influenza vaccines, and thus, quite possibly contributing tolower vaccine cost and wider availability to a greater number ofrecipients worldwide. In the immediate future, it is imperativethat we take on a systems approach, as demonstrated elsewhere(34), to integrate strain improvement, upstream optimization,and downstream processing to further improve our productionstrategy. A fully developed and well-characterized cell substratesystem would be advantageous not only economically but alsopresents a stronger case for approval by federal regulationagencies.

Materials and MethodsCell Line and Virus. Madin Darby Canine Kidney (MDCK) cells were acquiredfrom American Type Culture Collection (Cat. No. CCL-34). The MDCK cells weregrown in 37 °C, 5% CO2 humid incubator using Minimal Essential Mediumcontaining Earl’s salts and L-glutamine (Invitrogen) and supplemented withFetal Bovine Serum (Invitrogen) to a final concentration of 10%. Only cellsgrowing in less than 20 passages were used for this study. Influenza virus strainB/Victoria/504/2000 was obtained from the influenza virus depository of theCenter of Biologics Evaluations and Research, Food and Drugs Administration(Bethesda, MD).

Establishment of Stable MDCK Cell Line Expressing siat7e. Escherichia coli DH5�

competent cells (Invitrogen) were transformed with full-length human siat7egene expression vector (Cat. No. EX-V1581-M03, Genecopoeia). The plasmidswere purified using the QIAprep Spin Miniprep kit (Qiagen) and were used totransfect MDCK cells using Lipofectamine 2000 reagent under manufacturer’sprotocol (Invitrogen). The transfection procedure was as follows: day 1: MDCKcells were seeded at 2 � 105 cells/well in a 24-well plate; day 2: 0.8 �g plasmidDNA was mixed with 2.0 �L Lipofectamine 2000 and incubated together withthe cells in OptiMEM I medium (Invitrogen) for 4 h; the cells were than washedand suspended in growth medium; day 3: G418 was added to the growthmedium at a final concentration of 0.400 mg/mL, and the medium containingG418 (selective medium) was routinely replaced every 3 to 4 days for a periodof 3 weeks. Stably transfected pool of siat7e-expressing cells were grown andbanked. Finally, clones were isolated by limiting dilution in a 96-well plate.

Gene Expression. RNA samples were isolated from parental MDCK cells andfrom clones of the siat7e-expressing cells using RNeasy Total RNA Isolation kit(Qiagen). SuperScript One-Step RT-PCR kit (Invitrogen) was used for thereverse transcription and for PCR amplification experiments in accordance tothe manufacturer’s protocol, using the sense primer sequence 5�-ttactcgcca-caagatgctg-3� and antisense primer sequence 5�-gcaccatgccataaacattg-3�.GAPDH was selected as the endogenous control gene and was amplified usingsense primer sequence 5�-aacatcatccctgcttccac-3� and antisense primer se-quence 5�-gaccacctggtcctcagtgt-3�. Briefly: cDNA synthesis was performed at50 °C for 30 min, samples were incubated at 94 °C for 2 min to ‘‘hot-start’’ theDNA Taq polymerase. The PCR amplification cycle consisted of denaturation at94 °C for 15 s annealing at 55 °C for 30 s, and extending at 72 °C for 10 s (14s for the endogenous control). The target genes were amplified for 35 cycleswith a final extension at 72 °C for 10 min. The end products were resolved ona 1% agarose gel at 130 V for 30 min and captured on the gel imager (Bio-Rad).

Real-time PCR was performed using Power SYBR Green RNA-to-CT™ 1-StepKit (Applied Biosystems) with the same primer sequences described above.Briefly: cDNA samples were synthesized from 0.5 ng RNA sample and ampli-fied under standard thermal cycler protocol (50 °C for 2 min, 95 °C for 10 min,

and 40 cycles of 95 °C for 15 s and 60 °C for 1 min). Target Ct values wereaveraged from replicates and fold changes were calculated against the en-dogenous control, GAPDH.

Cationized Ferritin Binding Assay. Cationized ferrtin (Electron MicroscopySciences) was conjugated with FITC using the FITC Protein Labeling kit (PierceBiotechnology). Briefly: cationized ferritin was dialyzed with the suppliedborate buffer and incubated with FITC solution at room temperature for 1 h.Excess FITC dye was removed using a dialysis cassette (Pierce Biotechnology).Conjugated ferritin complex was quantified using E270 nm1% � 79.9 andMW � 750,000 for native ferritin and a correction factor of 0.3 for FITC whose�max � 494 nm. The calculated F/P ratio was approximately 12. Approximately1 � 107 cells were detached from culture flasks using Hank’s-based celldissociation buffer (Invitrogen) and washed with PBS before re-suspending in1 mL PBS containing FITC-conjugated ferritin at 50 �g/mL final concentration(24–26). The mixture was incubated on a thermomixer at 4 °C for 1 h andwashed once with PBS. Cells were spun down and suspended in 1 mL cold PBS.The cells were immediately analyzed using the FACSCalibur flow cytometer.

Growth Kinetics. For growth kinetics in anchorage-dependent manner, pa-rental and siat7e-expressing MDCK cells were seeded at a concentration of 2 �105 cells per one 25-cm2 culture flask; 21 flasks were seeded for each cell line.Glucose and lactate concentrations were measured using the YSI 2700 Selectbiochemistry analyzer (YSI Life Sciences) and cell count was measured usingCedex (Innovatis AG). Measurements were taken daily from three flasks. Forgrowth kinetics in suspension culture, cells from each line were seeded atapproximately 2 � 105 cell/mL in three 125-mL vented shake flasks containing30 mL serum-supplemented Dulbecco’s Modified Eagle’s Medium (Invitrogen)and shaken at 90 RPM. Measurements were taken at 48 h intervals.

Virus Growth Evaluations in Monolayer and Suspension Culture. Monolayerculture: Parental MDCK cells or siat7e-expressing cells were grown to conflu-ency in 25-cm2 flasks (Corning). After removal of the growth media, the cellswere washed once with serum-free medium and the virus was added to eachflask at a multiplicity of infection (MOI) of 2.0 TCID50 (50% tissue cultureinfectious dose). After adsorption for 1 h at 37 °C, the cells were washed withserum-free medium, and 10 mL of growth medium (containing 10% FBS) wereadded. The infected cells were incubated at 33 °C for the remainder of theexperiment. Cell condition (appearance of cytopathogenic effect) was con-stantly monitored and samples were collected every 8 h for virus infectivityand hemagglutination (HA) titers determination.

Suspension culture: siat7e-expressing cells grown in shake flasks wereconcentrated by centrifugation (600 rcf for 5 min) and re-suspended in aserum-free medium at a density of 107 cells/mL. After infection with theinfluenza virus at an MOI of 2.0 TCID50, the cell suspension was incubated atconstant shaking at 37 °C for 1 h. At this time, the cells were precipitated andsuspended in DMEM supplemented with 10% FBS to a density of 106 cells/mL.The infected cells were incubated at 33 °C in the same conditions for theremainder of the experiment; the controlled culture was treated in the sameway but without addition of the virus. Samples were taken every 8 h during aperiod of 4 days and stored in aliquots at �70 °C for virus infectivity titer andHA titer determination. Cell concentration, viability and metabolic parame-ters were monitored at each time point.

Determination of Virus Yield. Virus growth and concentration were deter-mined by infectivity titer in chicken embryonated eggs (EID50) and by HA titerusing standard techniques described earlier (35–37).

Determination of Virus Stability During Replication in MDCK Cells. Antigenicproperties of the progeny virus harvested from the parental or the siat7e-expressing cells (56 h post infection) were characterized by hemagglutinationinhibition test (HAI test) using a set of three homologous ferret antisera specificto strain B/Victoria/504/2000. The HAI test was performed in 96-well plates (tworeplicates for each serum sample) using 0.5% chicken red blood cells in PBS (pH7.2) (37). Two viruses were considered antigenically indistinguishable if thecorresponding HAI titers did not exceed two-fold difference. In addition thenucleotide sequences of viral gene segments encoding viral surface glycopro-teins, HA and NA, were determined by direct DNA-sequencing of the RT-PCRproducts and compared with those of the parental virus stock.

ACKNOWLEDGMENTS. We thank Dr. Bruce Raaka for his assistance with theFACS measurements and Dr. Pratik Jaluria for his communications on technicalaspects of the experiments and formulation of the manuscript. This work wassupported by the Intramural program at the National Institute of Diabetes andDigestive and Kidney Diseases, National Institutes of Health.

14806 � www.pnas.org�cgi�doi�10.1073�pnas.0905912106 Chu et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

26, 2

020

1. Simonsen L, Fukuda K, Schonberger LB, Cox NJ (2000) The impact of influenza epi-demics on hospitalizations. J Infect Dis 181:831–837.

2. Bardiya N, Bae JH (2005) Influenza vaccines: Recent advances in production technol-ogies. Appl Microbiol Biotechnol 67:299–305.

3. Kistner O, et al. (1998) Development of a mammalian cell (Vero) derived candidateinfluenza virus vaccine. Vaccine 16:960–968.

4. Govorkova EA, Kodihalli S, Alymova IV, Fanget B, Webster RG (1999) Growth andimmunogenicity of influenza viruses cultivated in Vero or MDCK cells and in embryo-nated chicken eggs. Dev Biol Stand 98:39–51, discussion 73–34.

5. Pau MG, et al. (2001) The human cell line PER.C6 provides a new manufacturing systemfor the production of influenza vaccines. Vaccine 19:2716–2721.

6. Tree JA, Richardson C, Fooks AR, Clegg JC, Looby D (2001) Comparison of large-scalemammalian cell culture systems with egg culture for the production of influenza virusA vaccine strains. Vaccine 19:3444–3450.

7. Ozaki H, et al. (2004) Generation of high-yielding influenza A viruses in African greenmonkey kidney (Vero) cells by reverse genetics. J Virol 78:1851–1857.

8. Youil R, et al. (2004) Comparative study of influenza virus replication in Vero and MDCKcell lines. J Virol Methods 120:23–31.

9. Frisch SM, Francis H (1994) Disruption of epithelial cell-matrix interactions inducesapoptosis. J Cell Biol 124:619–626.

10. Folkman J, Moscona A (1978) Role of cell shape in growth control. Nature 273:345–349.11. Guadagno TM, Ohtsubo M, Roberts JM, Assoian RK (1993) A link between cyclin A

expression and adhesion-dependent cell cycle progression. Science 262:1572–1575.12. Hansen LK, Mooney DJ, Vacanti JP, Ingber DE (1994) Integrin binding and cell spread-

ing on extracellular matrix act at different points in the cell cycle to promote hepato-cyte growth. Mol Biol Cell 5:967–975.

13. Zhu X, Ohtsubo M, Bohmer RM, Roberts JM, Assoian RK (1996) Adhesion-dependentcell cycle progression linked to the expression of cyclin D1, activation of cyclin E-cdk2,and phosphorylation of the retinoblastoma protein. J Cell Biol 133:391–403.

14. Fang F, Orend G, Watanabe N, Hunter T, Ruoslahti E (1996) Dependence of cyclinE-CDK2 kinase activity on cell anchorage. Science 271:499–502.

15. Assoian RK (1997) Anchorage-dependent cell cycle progression. J Cell Biol 136:1–4.16. Genzel Y, Behrendt I, Konig S, Sann H, Reichl U (2004) Metabolism of MDCK cells during

cell growth and influenza virus production in large-scale microcarrier culture. Vaccine22:2202–2208.

17. Genzel Y, Fischer M, Reichl U (2006) Serum-free influenza virus production avoidingwashing steps and medium exchange in large-scale microcarrier culture. Vaccine24:3261–3272.

18. Genzel Y, Olmer RM, Schafer B, Reichl U (2006) Wave microcarrier cultivation of MDCKcells for influenza virus production in serum containing and serum-free media. Vaccine24:6074–6087.

19. Hu AY, et al. (2008) Microcarrier-based MDCK cell culture system for the production ofinfluenza H5N1 vaccines. Vaccine 26:5736–5740.

20. Jaluria P, Betenbaugh M, Konstantopoulos K, Frank B, Shiloach J (2007) Application ofmicroarrays to identify and characterize genes involved in attachment dependence inHeLa cells. Metab Eng 9:241–251.

21. Tsuchida A, et al. (2003) Synthesis of disialyl Lewis a (Le (a)) structure in colon cancer celllines by a sialyltransferase, ST6GalNAc VI, responsible for the synthesis of alpha-seriesgangliosides. J Biol Chem 278:22787–22794.

22. Hakomori SI (2000) Cell adhesion/recognition and signal transduction through glyco-sphingolipid microdomain. Glycoconj J 17:143–151.

23. Regina Todeschini A, Hakomori SI (2008) Functional role of glycosphingolipids andgangliosides in control of cell adhesion, motility, and growth, through glycosynapticmicrodomains. Biochim Biophys Acta 1780:421–433.

24. Argueso P, Tisdale A, Spurr-Michaud S, Sumiyoshi M, Gipson IK (2006) Mucin charac-teristics of human corneal-limbal epithelial cells that exclude the rose bengal anionicdye. Invest Ophthalmol Vis Sci 47:113–119.

25. Danon D, Goldstein L, Marikovsky Y, Skutelsky E (1972) Use of cationized ferritin as alabel of negative charges on cell surfaces. J Ultrastruct Res 38:500–510.

26. King CA, Preston TM (1977) Fluoresceinated cationised ferritin as a membrane probefor anionic sites at the cell surface. FEBS Lett 73:59–63.

27. Palese P (2006) Making better influenza virus vaccines? Emerg Infect Dis 12:61–65.28. Audsley JM, Tannock GA (2004) The role of cell culture vaccines in the control of the

next influenza pandemic. Expert Opin Biol Ther 4:709–717.29. Hatakeyama S, et al. (2005) Enhanced expression of an alpha2,6-linked sialic acid on

MDCK cells improves isolation of human influenza viruses and evaluation of theirsensitivity to a neuraminidase inhibitor. J Clin Microbiol 43:4139–4146.

30. Matrosovich M, Matrosovich T, Carr J, Roberts NA, Klenk HD (2003) Overexpression ofthe alpha-2,6-sialyltransferase in MDCK cells increases influenza virus sensitivity toneuraminidase inhibitors. J Virol 77:8418–8425.

31. Oh DY, Barr IG, Mosse JA, Laurie KL (2008) MDCK-SIAT1 cells show improved isolationrates for recent human influenza viruses compared to conventional MDCK cells. J ClinMicrobiol 46:2189–2194.

32. Vodeiko GM, McInnis J, Chizhikov V, Levandowski RA (2003) Genetic and phenotypicanalysis of reassortants of high growth and low growth strains of influenza B virus.Vaccine 21:3867–3874.

33. Lugovtsev VY, Vodeiko GM, Strupczewski CM, Ye Z, Levandowski RA (2007) Generationof the influenza B viruses with improved growth phenotype by substitution of specificamino acids of hemagglutinin. Virology 365:315–323.

34. Lee SY, et al. (2008) From genome sequence to integrated bioprocess for succinic acidproduction by Mannheimia succiniciproducens. Appl Microbiol Biotechnol 79:11–22.

35. Lugovtsev VY, Vodeiko GM, Levandowski RA (2005) Mutational pattern of influenza Bviruses adapted to high growth replication in embryonated eggs. Virus Res 109:149–157.

36. Palmer DF, Coleman MT, Dowdle WR, Schild GC (1975) in Advanced LaboratoryTechniques for Influenza Diagnosis (Washington, DC).

37. WHO (2002) WHO manual on animal influenza diagnosis and surveillance, Geneva.

Chu et al. PNAS � September 1, 2009 � vol. 106 � no. 35 � 14807

APP

LIED

BIO

LOG

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

26, 2

020