- Email: [email protected]
Viral antigen production in cell cultures on microcarriers
Available online at www.sciencedirect.com
Vaccine 25 (2007) 7785–7795
Viral antigen production in cell cultures on microcarriers Bovine parainfluenza 3 virus and MDBK cells M.M. Conceic¸a˜ o a , A. Tonso b , C.B. Freitas c , C.A. Pereira a,∗ a b
Laborat´orio de Imunologia Viral, Instituto Butantan, Av. Vital Brasil 1500, 05503-900 S˜ao Paulo, Brazil Departamento de Engenharia Qu´ımica, Escola Polit´ecnica, Universidade de S˜ao Paulo, S˜ao Paulo, Brazil c Vall´ ee SA, Montes Claros, Minas Gerais, Brazil Received 6 March 2007; received in revised form 7 August 2007; accepted 26 August 2007 Available online 14 September 2007
Abstract Viral antigens can be obtained from infected mammalian cells cultivated on microcarriers. We have worked out parameters for the production of bovine parainfluenza 3 (PI-3) virus by Mandin–Darby Bovine Kidney (MDBK) cells cultivated on Cytodex 1 microcarriers (MCs) in spinners flasks and bioreactor using fetal bovine serum (FBS) supplemented Eagle minimal essential medium (Eagle-MEM). Medium renewal during the cell culture was shown to be crucial for optimal MCs loading (>90% MCs with confluent cell monolayers) and cell growth (2.5 × 106 cells/mL and a μx (h−1 ) 0.05). Since cell cultures performed with lower amount of MCs (1 g/L), showed good performances in terms of cell loading, we designed batch experiments with a lower concentration of MCs in view of optimizing the cell growth and virus production. Studies of cell growth with lower concentrations of MCs (0.85 g/L) showed that an increase in the initial cell seeding (from 7 to 40 cells/MC) led to a different kinetic of initial cell growth but to comparable final cell concentrations ((8–10) × 105 cells/mL at 120 h) and cell loading (210–270 cells/MC). Upon infection with PI-3 virus, cultures showed a decrease in cell growth and MC loading directly related to the multiplicity of infection (moi) used for virus infection. Infected cultures showed also a higher consumption of glucose and production of lactate. The PI-3 virus and PI-3 antigen production among the cultures was not significantly different and attained values ranging from, respectively, 7–9 log10 TCID50 /mL and 1.5–2.2 OD. The kinetics of PI-3 virus production showed a sharp increase during the first 24 h and those of PI-3 antigen increased after 24 h. The differential kinetics of PI-3 virus and PI-3 antigen can be explained by the virus sensitivity to temperature. In view of establishing a protocol of virus production and based on the previous experiments, MDBK cell cultures performed under medium perfusion in a bioreactor of 1.2 L were infected and the PI-3 virus production in 12 L attained 12 log10 TCID50 . Other than establishing a protocol for PI-3 production in MDBK cell cultures on Cytodex 1, the experiments are proposed as a basis for approaching the development of a virus production protocol in mammalian cells cultivated on microcarriers in bioreactors. © 2007 Elsevier Ltd. All rights reserved. Keywords: MDBK cells; PI-3 virus; Microcarriers
1. Introduction The development of animal cell culture bioprocesses for the production of viral antigens is today a solid basis for industries to provide vaccines for human and veterinarian use. The technology of animal cell cultivation constitutes the essence of the production based in fermentation and differs ∗
Corresponding author. Tel.: +55 11 37267222; fax: +55 11 37261505. E-mail address: [email protected] (C.A. Pereira).
0264-410X/$ – see front matter © 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.vaccine.2007.08.048
significantly from that using bacteria or yeast. Animal cells usually show a slower cell growing, anchorage dependency, higher hydrodynamic stress, metabolic complexity and need strict aseptic working conditions [1]. In view of the increasing demand of animal cell bioprocess for biological products, research & development in this area is subject of great scientific and economical interest [2,3]. Since most of the cell lines suitable for virus replication show an anchorage dependence, and that virus infection is often more effective in attached and spread out cells, dif-
7786
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
Fig. 1. Kinetics of MDBK cell growth on microcarriers. Medium renewal. 105 MDBK cells/mL were seeded on 2 g/L of Cytodex 1 MCs (12 cells/MC). Cultures were performed in 100 mL spinner flasks in batch (A) or in repeated batch with a half volume (50 mL) medium renewal at 24, 48, 60, 72 and 84 h (B). Cultures were also performed in 1.2 L bioreactor with a half volume (600 mL) medium renewal at 24, 48, 60, 72 and 84 h (C). The concentrations (g/L) of glucose, glutamine and lactate were periodically measured in culture supernatants and are indicated. The % of totally covered MCs (D) and the specific cell growth, μx (h−1 ) (E) in the cultures were measured. Cultures were performed at 37 ◦ C and 60 rpm. Dissolved oxygen (DO) in bioreactor cell cultures was controlled at 40% and pH at 7.3. Arrows indicate medium change. Data are the mean value of at least three separated experiments with standard deviations.
ferent cell substrates and particular bioprocesses have been developed, the scaling-up being always dependent on the surface available for cell attachment, spreading and multiplication [1,2,4,5]. A great step towards the establishment of high-density cell cultures and industrial bioprocesses was done by the cell cultivation on suspended beads, named microcarriers [6]. Cell cultures on microcarriers, at different concentrations, can be performed in bioreactors with fine control of cell environment (nutrients, pH, O2 and agitation) and using different modes of operation (batch, fed-batch and perfusion) [1]. Batches of 1000 L VERO cell cultures on microcarriers can be performed in bioreactors and are used for the production of vaccines as rabies and poliomyelitis [7,8]. Although the approach of viral antigen production in cell cultures on microcarriers is well known and consists in cell attachment and multiplication followed by virus infection, supernatant or cell fraction harvesting, concentration, purification and formulation, the bioprocess parameters may vary considerably from a cell/virus system to another. Parameters such as, the microcarrier type and concentration, cell culture medium composition, pH, dissolved O2 , multiplicity and time of virus infection, nutrients consumption and
supply, need to be worked out and rigorously determined [3–5,9,10]. An animal cell-based bioprocess for virus production has the particularity of constituting a system where the cell physiology is often sharply modified and redirected by the virus infection. It differs considerably from a cell-based recombinant protein synthesis that constitutes ultimately a consequence of a stable cell physiological state. As a consequence, parameters for a virus production bioprocess establishment have to be worked out by studying both, the cell growth and the virus replication steps [4,5,11–17]. The Mandin & Darby Bovine Kidney (MDBK) cell line represents an important cell line for the replication and consequent identification and production of several viruses of veterinarian interest [18–21]. The bovine parainfluenza 3 (PI-3) virus is an important agent of worldwide respiratory disease in cattle [22–25]. It induces cytopathic effect (CPE) in several cell lines including the MDBK and can be inactivated by heat. Vaccines containing inactivated or attenuated PI-3 virus have shown to be antigenic and protected animals from disease [26–29] In addition, the bovine PI-3 virus has been proposed as a human vaccine [30] and as virus vaccine vector to express foreign viral antigens for human immuniza-
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
7787
guideline for a bioprocess establishment involving other cell/virus/microcarrier systems.
2. Materials and methods 2.1. Cell line and virus The Mandin & Darby Bovine Kidney (MDBK) cell and the Bovine Parainfluenza 3 (PI-3) virus were obtained from de Laborat´orio de Viroses Bovinas do Instituto Biol´ogico (S˜ao Paulo, Brazil) and originated from the American Type Culture Collection (ATCC). MDBK cells were cultivated following established procedures with Eagle’s minimal essential medium (MEM) (Cultilab) supplemented with 10% fetal bovine serum (FBS) (Cultilab), 1.5 g/L of NaHCO3 (Labsynth) and 0.11 g/L of sodium pyruvate (Gibco-BRL). A cell bank was constituted at 132nd passage and stored in liquid nitrogen. The PI-3 virus was already adapted to MDBK cells. A virus stock in MEM + 10% FBS constituted at 4th passage with a titer of 8.16 log10 TCID50 /ml, frozen at −80 ◦ C, was used in all the experiments. Viable cell concentration was determined by cell counting in hemocytometer with trypan blue exclusion. 2.2. MDBK cell cultures and infection
Fig. 2. Light microscopy of MCs cell covering during MDBK cell cultures. 105 MDBK cells/mL were seeded on 2 g/L of Cytodex 1 MCs (12 cells/MC). Cultures were performed in 100 mL spinner flasks with a half volume (50 mL) medium renewal at 24, 48, 60, 72 and 84 h. Samples of MCs were taken at 36 and 96 h of culture, stained with trypan blue and observed at microscopy for evaluation of MC-cell covering degree. Dark areas are regions of the MC without cells and stained by trypan blue. This procedure was used throughout the study for observation of the MC cell covering and estimation of the % of totally covered MCs. In (A) and (B) we have, respectively, 13.8% and 95.2% of MCs totally covered by MDBK cells.
tion [31,32]. In spite of all these evidences, the literature is very scarce concerning studies of MDBK cell multiplication and/or PI-3 replication on microcarriers, in view of establishing a protocol for scaling up bioprocess of virus antigen production [21]. Besides establishing a protocol for production of PI3 virus antigen in MDBK cell cultures on Cytodex 1 in bioreactor, the approach in this study is proposed as a
For cell culture experiments, the MEM + 10% FBS was supplemented with 1 UI/mL of penicillin (Cultilab), 1 g/mL of streptomycin (Cultilab) and 2.5 g/mL of amphotericin B (Sigma). Cells were maintained in 75 or 150 cm2 T-flasks incubated at 37 ◦ C with 5% CO2 and passages were performed after trypsinization (tripsin solution at 0.2%). Cytodex 1 microcarriers (MCs) (GE Health Care) were used to cultivate the MDBK cells in spinners (Bellco) or in a bioreactor (celligen, New Brunswick) and had the following characteristics: density of 1.03 g/mL, average size of 190 m, surface of 4.400 cm2 /g and a number of MCs/g of 4.3 × 106 . They were PBS hydrated, following the manufacturer recommendation, and “in situ” sterilized. Spinners (Bellco) and bioreactor vessels were previously treated with dimetildiclorosilane (Merck). For cell culture, 100 mL spinners containing cells and MCs were placed on a magnetic stirrer (Bellco) inside an incubator at 37 ◦ C and 5% CO2 . For cell cultures in the 1.2 L working volume bioreactor, medium and cells were added by pressuring a bottle connected to the reactor. In both cases initial cell seeding is indicated as cells/mL and cells/MC, as calculated by knowing the number of cells/mL, the g/L of MCs and the number of MCs/g. The bioreactor had an automatically controlled fourgas using a double-screen cell-lift impeller for low shear, avoiding sparging and foam formation and providing the culture with O2 , CO2 , N2 and air for pH and dissolved oxygen (DO) control. It operated with adjustable temperatures (37 ◦ C) and agitation speed (60 rpm). Samples were collected via a pipe in the culture fluid and transferred
7788
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
Fig. 3. Kinetics of MDBK cell growth on microcarriers. MCs concentration. In (A), (B) and (C), 105 MDBK cells/mL were seeded on, respectively, 1–3 g/L of Cytodex 1 MCs (resulting in, respectively, 24, 12 and 8 cells/MC). Cultures were performed in 100 mL spinner flasks with a half volume (50 mL) medium renewal at 24, 48, 60, 72 and 84 h. The concentrations (g/L) of glucose, glutamine and lactate were periodically measured in culture supernatants and are indicated, the specific cell growth, μx (h−1 ) (D), the average number of cells per MC (E), and the % of totally covered MCs (F) in the cultures were measured. Cultures were performed at 37 ◦ C and 60 rpm. Arrows indicate medium change. Data are the mean value of at least 3 separated experiments with standard deviations.
to sample flasks by aspiration. Medium renewal was performed by using peristaltic pumps (Watson–Marlon) and a MC decanter to avoid the outflow of MCs. PI-3 viral infection of cell cultures was performed by adding virus suspensions at the indicated multiplicity of infection (moi). Medium containing virus during perfusion was collect in recipients placed at +4 ◦ C to avoid virus inactivation or degradation.
2.3. Analytical procedures Kinetics of cell multiplication on MCs, expressed by the number of cells per mL (cells/mL), was obtained by counting the cells attached to the MCs, at the indicated intervals throughout the culture, as follows: cell loaded-MCs in 1 mL of culture sample were allowed to settle, then cell lysis and nucleus staining were carried out in 1 mL of a hypotonic solution (0.1 M citric acid and 0.1% crystal violet) at 37 ◦ C for 10 min, and the nuclei counted in a Neubauer chamber [5].
The specific cell growth, expressed as μx (h−1 ), was determined as already described [33,34]. Kinetics of cell distribution on MCs, expressed by the % of MCs showing a cell confluent monolayer (% of totally covered MCs), was obtaining by periodically collecting cell samples from the cultures, staining the MCs with blue trypan (0.025%) for 5 min at room temperature, and counting the number of MCs totally or partially covered with cells or uncovered (Fig. 2). Kinetics of cell loading of MCs, expressed by the number of cells per MC (cells/MC), was obtained from the values of cell counting (cells/mL) and number of MCs per mL used for settling the culture. Kinetics of glucose (GLC) and glutamine (GLN) consumption or lactate (LAC) production, expressed by their concentrations (g/L) in the culture supernatants, were determined by enzymatic reaction using the YSI 2700 analyzer (Yellow Springs). Kinetics of PI-3 virus replication, expressed by the log10 TCID50 /mL, was measured as following. Dilutions
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
7789
Fig. 4. Kinetics of MDBK cell growth on microcarriers. Batch cultures, low MCs concentration and variable cell seeding. Cell seeding of 0.25 × 105 , 0.75 × 105 , 1.5 × 105 cells/mL or, respectively, 7, 20 or 40 cells/MC were used to initiate MDBK cell cultures on 0.85 g/L of Cytodex 1 MCs. Batch cultures were performed in 100 mL spinner flasks at 37 ◦ C and 60 rpm. Kinetics data show the cell growth and the number cells/MC (A), the specific cell growth, μx (h−1 ), (B), the % of totally covered MCs (C) and the concentrations (g/L) of glucose and lactate. Data are the mean value of at least three separated experiments with standard deviations.
of infected cell culture supernatants were added to 96well microplates containing a MDBK cell monolayer. The microplates were incubated at 37 ◦ C for 72 h and then observed under the microscope for the presence of cytopathic effect (CPE). Based on the highest sample dilution capable of inducing CPE in 50% of cell, the viral titer was calculated using the Reed and Muench [35] test. Kinetics of PI-3 viral antigen production, expressed by the optical density (OD), was evaluated in cell culture supernatants by a commercially available enzyme-linked immunosorbent assay (ELISA Pulmotest Parainfluenza 3, Bio-X diagnostic). Briefly, dilutions of infected cell culture supernatants were added to 96-well microplates containing polyclonal antibodies against the PI-3 virus. The microplates were incubated for 1 h at 37 ◦ C, washed and then peroxidaselabeled antibodies against PI-3 virus were added. The microplates were incubated again for 1 h at 37 ◦ C, washed and the reaction developed by the addition of hydrogen peroxidase and TMB (tetrametilbenzidine). OD was measured at 540 nm in ELISA reader.
volume change every 12 h after 48 h of cell seeding. Under these conditions, the level of glucose, glutamine and lactate were maintained and the spinner flasks cultures could reach up to 2.4 × 106 cells/mL at 96 h with 95% of totally covered MCs and a μx (h−1 ) of 0.035 at 36 h, in contrast to batch cultures that reached only 106 cells/mL at 120 h with 7% of totally covered MCs and a μx (h−1 ) of 0.030 at 24 h. Bioreactor cultures with medium renewal, although showing also good performance (2 × 106 cells/mL at 96 h with 85% of totally covered MCs and a μx (h−1 ) of 0.032 at 36 h), partially reproduced the data obtained with spinner flasks. As shown in Fig. 2, MCs without cells attached to the surface or partially covered by cells, could be easily identified and quantified upon trypan blue staining which does not stain the MCs bearing confluent cells or MC areas covered by cells. The trypan blue staining procedure was used throughout the study to monitor cell cultivation in terms of MC cell loading and distribution.
3. Results
The above-mentioned conditions of medium renewal were used in experiments designed to evaluate the influence of the MC concentration (1–3 g/L) in cell culture performance (Fig. 3). The procedures of the medium renewal established in the previous experiments (Fig. 1) were shown to provide all the cultures with suitable levels of glucose, glutamine and lactate. Cultures performed with 2 or 3 g/L of MCs reached, at 96 h, higher cells concentra-
3.1. MDBK cell cultures in spinners and bioreactor—medium renewal As shown in Fig. 1, the medium renewal during MDBK cell culture was shown to be determinant for cell growth and gave best results when started at 24 h and performed at a half
3.2. MDBK cell cultures in spinner flasks—MC concentration
7790
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
Fig. 5. Kinetics of MDBK cell growth on microcarriers. PI-3 virus infection. Cell seeding of 0.25 × 105 , 0.75 × 105 , 1.5 × 105 cells/mL or, respectively 7, 20 or 40 cells/MC were used to initiate MDBK cell cultures on 0.85 g/L of Cytodex 1 MCs. Batch cultures were performed in 100 mL spinner flasks at 37 ◦ C and 60 rpm. At 60 h the cultures were infected with 0.001, 0.1 or 10 moi of PI-3 virus. Data show the kinetics of cell growth (A), % of totally covered MCs (B) and the concentrations (g/L) of glucose and lactate (C) after infection. Data are the mean value of at least three separated experiments with standard deviations.
tions (∼2.3 × 106 cells/mL) than those performed with 1 g/L (1.5 × 106 cells/mL). The cell cultures with 1, 2 and 3 g/L of MCs showed, respectively, a μx (h−1 ) of 0.031, 0.035 and 0.036 at 36 h, a cell loading of 340, 280 and 190 cells/MC at 108 h and a % of totally covered MCs of 85, 95 and 80 at 96 h. 3.3. MDBK cell cultures in spinner flasks—batch cultures, low MCs concentration, variable cell seeding and virus infection In view of establishing conditions for PI-3 virus infection production, batch cell cultures were performed with 0.85 g/L of MCs and seeded with 7, 20 and 40 cells/MC. These cultures although showing different kinetics of cell growth, reached comparable final cell concentrations (8.5 × 105 , 8 × 105 and 9 × 105 cells/mL at 120, 96 and 84 h for cultures with, respectively, 7, 20 and 40 cells/MC). The higher was the cell seeding, the sooner a plateau of cell concentration was
reached and MCs loaded and faster glucose was consumed and lactate produced. Glucose attained lower levels at 96 h, and lactate higher levels at 120 h. The MCs cell loading and distribution values in the cultures ranged, respectively, from 230 to 280 cells/MC and 83–90% of totally covered MCs. The μx (h−1 ) values ranged from 0.030 to 0.031 (Fig. 4). Paralleled cultures were mock or PI-3 virus infected with moi of 0.001, 0.1 and 10 and cell replication, MC cell distribution and glucose/lactate metabolism evaluated. As compared to mock-infected cultures, the virus infected ones showed a decreased cell growth and MCs cell distribution directly related to the virus moi used, which was clearly observed after 24 h of infection. Glucose was consumed faster by infected cultures as compared to controls, reaching low levels (<0.2 g/L) at 48 h of infection. Differences in lactate production among mock and virus infected cultures were less evident (Fig. 5). Cytopathic effect with cell detachment from the MCs in these cultures could be clearly observed by microscopic examination of MCs (Fig. 6).
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
7791
moi 0.001 the virus titers plateau was reached only at 48 h of infection (8–9 log10 TCID50 /mL). As expected, a different kinetic was observed for antigen production. Cultures infected with the different moi showed comparable kinetics and values. In cultures infected with 0.1 or 0.001 moi the PI-3 antigen production sharply increased from 24 to 72 h attaining higher values in the first one. In cultures infected with 10 moi, remaining PI-3 virus used for inoculation was detected as antigen together with antigen produced by the cell cultures during infection. No significant differences were observed among cultures with different cell seeding (Fig. 7). Differences of kinetics of PI-3 virus and antigen observed are in a great part due to the virus sensitivity to the temperature. In assays of virus inactivation, as measured by the virus titer detected in samples stored for different times at different temperatures, the PI-3 virus titer decrease was directly related to the time and temperature of inactivation. At 37 ◦ C, the incubation temperature during cell culture and infection, the PI-3 virus titer sharply decreases attaining already <20% of its original value after 18 h of incubation (Fig. 8). 3.4. MDBK cell cultures and virus infection in bioreactor
Fig. 6. Light microscopy of MDBK cell cultures on MCs upon infection with PI-3 virus. 0.75 × 105 MDBK cells/mL were seeded on 0.85 g/L of Cytodex 1 MCs (20 cells/MC). Batch cultures were performed in 100 mL spinner flasks. They were infected with PI-3 virus at 60 h with a multiplicity of infection (moi) of 0.1. Cytopathic effect can be observed after 24 h (A), 48 h (B) and 72 h (C).
The kinetics of PI-3 virus production was measured by “in vitro” cell infection assay (log10 TCID50 /mL) (Fig. 7A) and the PI-3 antigen by ELISA (OD) (Fig. 7B). In cultures infected with moi 10, there was no significant increase in the virus titers, the kinetics showing a plateau from 0 to 72 h (7–8 log10 TCID50 /mL). In cultures infected with moi 0.1, the virus titers increased sharply during the first 24 h of infection, reaching then a plateau, where they remain until 72 h (7.5–8.5 log10 TCID50 /mL). In cultures infected with
MDBK cell cultures were performed in bioreactor for PI3 virus infection with medium perfusion before and after virus infection. Data show the kinetics of cell growth and virus production (Fig. 9). Cell concentration attained values of ∼2.8 × 106 cells/mL at 108 h of culture, when they were infected with 0.1 moi of PI-3 virus. At this time the cell culture showed ∼75% of MCs with confluent cell monolayers and the concentration of glucose and lactate were controlled at, respectively, ∼0.2 and ∼0.7 g/L. To attain these values the medium (MEM + 10% FBS) perfusion rate was increased from 0.8 to 1.7 volumes/24 h. After infection, the medium (MEM + 2% FBS) perfusion rate was settled in 1 volume/24 h allowing the cell concentration and MC loading to remain constant for the first 36 h and the concentration of glucose and lactate to be controlled at, respectively, ∼0.1 and ∼0.7 g/L. In view of the perfusion procedure, the virus produced by the cultures could be collected both in the supernatant of the culture as well as in a reservoir placed at +4 ◦ C with the perfused medium. A total of 3 L of perfused medium and 1 L of supernatant were collected after 72 h of infection. The PI-3 virus production showed virus titers of ∼8 log10 TCID50 /mL in culture supernatant and from 6 to 7.5 log10 TCID50 /mL in perfused medium at different times after infection. A virus titer of 12 log10 TCID50 was obtained in 4 L of total collect medium. The PI-3 antigen production increased with the time of culture reaching high values.
4. Discussion Animal cell culture bioprocesses are today imperative for the preparation of cell-derived products. The main
7792
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
Fig. 7. Kinetics of PI-3 virus (A) and antigen (B) synthesis in MDBK cell cultures. Cell seeding of 0.25 × 105 , 0.75 × 105 , 1.5 × 105 cells/mL or, respectively 7, 20 or 40 cells/MC were used to initiate MDBK cell cultures on 0.85 g/L of Cytodex 1 MCs. Batch cultures were performed in 100 mL spinner flasks at 37 ◦ C and 60 rpm. At 60 h cultures were infected with 0.001, 0.1 or 10 moi of PI-3 virus. Data show the kinetics of PI-3 virus synthesis, as evaluated by the infectious virus titers (A) and the antigen quantification by ELISA (B). Data are the mean value of at least three separated experiments with standard deviations.
difficulty encountered when one attempts to establish a protocol for large-scale production is to provide high-density cell cultures with the best environment possible allowing them to synthesize the large amounts of product. When virus particles are the envisaged product, the establishment of a production protocol has to be approached in two steps, the first one consisting in the cell multiplication to high densities and the second one consisting in the virus infection, replication and synthesis. The main reason for this particular characteristic is that the virus infection often redirects the cell metabolism leading to cell culture conditions that may differ considerably from the one observed for non-infected cells. Upon infection, cells frequently show a modified nutrients consumption, metabolites production, oxygen requirement and hydrodynamic sensitivity. The present study was carried out with the aim of establishing bioprocess conditions for large-scale production of bovine parainfluenza 3 (PI-3) virus antigen in MDBK cell cultures on MCs, a veterinarian vaccine with commer-
Fig. 8. Loss of infectivity of PI-3 virus suspensions. Temperature sensitivity. A stock of PI-3 virus was exposed for 24, 48 and 72 h to temperatures of 10, 20, 30 and 37 ◦ C in a water bath. The viral suspensions were than frozen and titrated. Data, expressed as log10 TCID50 /mL, are the mean value of at least 3 separated experiments with standard deviations.
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
7793
Fig. 9. Kinetics of cell growth and virus synthesis in medium perfused and PI-3 infected MDBK cell cultures on microcarriers. MDBK cell cultures were performed on 3.4 g/L of Cytodex 1 MCs with a cell seeding of 1.5 × 105 cells/mL (10 cells/MC) in a 1.2 L bioreactor. Cultures were performed at 37 ◦ C, 60 rpm with dissolved oxygen (DO) controlled at 40% and pH at 7.3. PI-3 virus infection was done with 0.1 moi at 108 h of culture. Data show the perfusion rates employed (volume/24 h), the kinetics of cell growth (A), % of totally covered MCs (B), the concentrations (g/L) of glucose and lactate (C), the infectious PI-3 virus (D) and antigen produced (E). Data are the mean value of at least three separated experiments.
cial, environmental and animal health importance [23,25]. Our approach for establishing these bioprocess conditions is proposed as a guideline for the establishment other cell/virus/microcarrier systems. Basic conditions of MDBK cell multiplication on MCs were first investigated in 100 mL spinner cultures with or without medium renewal. Kinetics of cell multiplication and metabolism as well as specific cell growth and MC cell loading/distribution were studied. Data show clearly that specific cell growth, cell density and MCs cell loading/distribution were dependent on medium renewal. Among several protocols of time and medium renewal (data not shown) the one leading to best performances indicated half volume of medium renewal initiated 24 h after cell seeding. Preliminary cultures on 1.2 L bioreactor using parameters established for 100 mL spinners roughly reproduced the cell growth performance (Fig. 1).
In view of studying the influence of MCs concentration on cell culture performance, the established conditions of medium renewal were used in experiments with different MCs concentrations (1–3 g/L). In all the cultures, the specific cell growth and % of fully covered MCs were comparable, indicating good environmental culture conditions. Although showing a lower final cell density, the cultures performed with 1 g/L of MCs showed a higher MC cell loading. At this stage and considering the limitations of spinner cell cultures, these observations allowed us to conclude that the surface available for cell multiplication limited the cell growth and that the cell loading, in cultures with higher MCs concentrations (2 and 3 g/L), could be further increased by providing the cultures with even better environment (Fig. 3). In order to investigate the cell growth and virus infection in batch conditions with high MCs loading, experiments with lower MCs concentration and different cell seeding were
7794
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
designed. As an indication that this approach was suitable for studies of virus infection, all the cultures, regardless the cell seeding, were shown to attain relatively high levels of specific cell growth, cell density (cell/mL) and MC cell loading/distribution. It is important to take into consideration that they were performed without medium renewal and as a consequence low concentrations of glucose in cultures were rapidly attained (Fig. 4). Upon PI-3 virus infection, the cell cultures showed a decreased cell growth and MCs cell loading directly related to the virus moi used and paralleling the induced cytopathic effect observed (Figs. 5 and 6). Higher glucose consumption was also detected in infected cultures. Infected culture supernatants were assayed for the presence of PI-3 virus and PI-3 antigen. As expected, and explained by the virus temperature sensitivity, the virus titers increased during the first 24 h attaining a plateau and the virus antigen increased progressively during the infection. Best results were obtained with virus infection of 0.1 moi, regardless the cell seeding. The ELISA data suggested that PI-3 antigen was not undergoing important degradation during the culture period studied (72 h). The virus titer and antigen measures are of importance and express actually different viral product characteristics. The virus titer express the ability of the cultures to produce infectious virus that in turn are capable of further promote secondary infections, and the virus antigen expressing, ultimately, the virus protein relevant as a product for immunization (Figs. 7 and 8). Once the conditions of cell growth with medium renewal (Figs. 1 and 3) and virus infection (Figs. 4–8) were established, we could assay a suitable protocol for MDBK cell culture and PI-3 virus infection in a 1.2 L bioreactor. Conditions for MDBK cell cultures under medium perfusion in bioreactor were preliminary assayed (data not shown) and then established with a MC concentration of 3.4 g/L (10 cells/MC) and rates of perfusion varying from 0.8 to 1.7 volumes/24 h. They were infected with 0.1 moi and the kinetics of virus and antigen production evaluated. The cell cultures attained, at the moment of infection (108 h), a cell density of ∼2.8 × 106 cells/mL with ∼75% of MCs with confluent cell monolayers and the concentration of glucose and lactate controlled in, respectively, ∼0.2 and ∼0.7 g/L. These conditions were maintained after virus infection and allowed the harvesting, after 72 h, of 4 L of virus suspension with an infectious titer of 12 log10 TCID50 . The PI-3 virus antigen production increased with the time of culture reaching high values at 72 h. Our development study showed here led to the establishment of a pilot protocol for the production of PI-3 viral antigen in MDBK cell cultures on Cytodex 1 MCs performed in a 1.2 L bioreactor, that can be employed for the preparation of PI-3 antigen batches for immune protection investigations and easily scaled-up for production in larger volumes. Although the establishment of a given production protocol may vary depending on the virus, cell substrate, medium and vessel culture, common features exist and a general technological approach is helpful as a guideline for optimization.
So, experiments showed here are proposed as a flowchart basis for approaching the development of a virus production protocol in mammalian cells cultivated on microcarriers in bioreactors.
Acknowledgements We thank Maryvonne Kientz for reviewing the manuscript. This work was supported in part by grants from Vall´ee S.A and Fundac¸a˜ o Butantan. Carlos A. Pereira is recipient of CNPq—IA senior research fellowship. Mateus M. Conceic¸a˜ o had a scholarship from Vall´ee S.A.
References [1] Freshney RI. Culture of animal cells. A manual of basic technique. 4th ed. New York: John Wiley & Sons Inc.; 2000. [2] Hu WS, Aunins JG. Large-scale mammalian cell culture. Curr Opin Biotechnol 1997;27:148–53. [3] Merten O-W. Introduction to animal cell culture technology–past, present and future. Cytotechnology 2006;50:1–7. [4] Mendonc¸a RZ, Arr´ozio SJ, Antoniazzi MM, Ferreira Jr JMC, Pereira CA. Metabolic active-high density VERO cell cultures on microcarriers following apoptosis prevention by galactose/glutamine feeding. J Biotechnol 2002;97:13–22. [5] Yokomizo AY, Antoniazzi MM, Galdino PL, Azambuja Jr N, Jorge SAC, Pereira CA. Rabies virus production in high Vero cell density cultures on macroporous microcarriers. Biotechnol Bioeng 2004;85:506–15. [6] Van Wezel AL. Growth of cell strains and primary cells on microcarriers in homogeneous culture. Nature 1967;216:64–5. [7] Montagnon B, Vincent-Falquet JC, Fanget B. Thousand litre scale microcarrier culture of VERO cell for killed poliovirus vaccine: promising results. Dev Biol Stand 1984;55:37–8. [8] Pullen K, Johnston MD, Philips AW, Ball GD, Finter WB. Very large scale suspension cultures of mammalian cells. Dev Biol Stand 1984;60:175–7. [9] Mendonc¸a RZ, Prado JCM, Pereira CA. Attachment, spreading and growth of VERO cells on microcarriers for the optimization of large scale cultures. Bioprocess Eng 1999;20:565–71. [10] Ng YC, Berry JM, Butler M. Optimization of physical parameters for cell attachment and growth on macroporous microcarriers. Biotechnol Bioeng 1996;50:627–35. [11] Berry JM, Barnab´e N, Coombs KM, Butler M. Production of reovirus type-1 and type-3 from VERO cells grown on solid and macroporous microcarriers. Biotechnol Bioeng 1999;62:9–12. [12] Bleckwenn NA, Golding H, Bentley WE, Shiloach J. Production of recombinant proteins by vaccinia virus in a microcarrier based mammalian cell perfusion bioreactor. Biotechnol Bioeng 2005;90:663–74. [13] Genzel Y, Fischer M, Reichl U. Serum-free influenza virus production avoiding washing steps and medium exchange in large-scale microcarrier culture. Vaccine 2006;24:3261–72. [14] Ghendon YZ, Markushin SG, Akopova II, Koptiaeva IB, Nechaeva EA, Mazurkova LA, et al. Development of cell culture (MDCK) live coldadapted (CA) attenuated influenza vaccine. Vaccine 2005;23:4678–84. [15] Mohler L, Flockerzi D, Sann H, Reichl U. Mathematical model of influenza A virus production in large-scale microcarrier culture. Biotechnol Bioeng 2005;90:46–58. [16] Sun MB, Jiang YJ, Li WD, Li PZ, Li GL, Jiang SD, et al. A novel process for production of hepatitis A virus in Vero cells grown on microcarriers in bioreactor. World J Gastroenterol 2004;10:2571–3.
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795 [17] Wu SC, Liu CC, Lian WC. Optimization of microcarrier cell culture process for the inactivated enterovirus type 71 vaccine development. Vaccine 2004;22:3858–64. [18] Gogorza LM, Moran PE, Larghi JL, Segui R, Lissarrague C, Saracco M, et al. Detection of bovine viral diarrhea virus (BVDV) in seropositive cattle. Prev Vet Med 2005;72:49–54. [19] Takiuchi E, Medici KC, Alfieri AF, Alfieri AA. Bovine herpesvirus type 1 abortions detected by a semi-nested PCR in Brazilian cattle herds. Res Vet Sci 2005;79:85–8. [20] Townsend J, Duffus WP, Williams DL. Immune production of interferon by cultured peripheral blood mononuclear cells from calves infected with BHV1 and PI3 viruses. Res Vet Sci 1988;45:198–205. [21] Varani J, Bendelow MJ, Chun JH, Hillegas WA. Cell growth on microcarriers: comparison of proliferation on and recovery from various substrates. J Biol Stand 1986;14:331–6. [22] Murphy FA, Gibbs EPJ, Horzinek MC, Studdert MJ. Veterinary virology. London: Academic Press; 1999. p. 411–28. [23] Hagglund S, Hjort M, Graham DA, Ohagen P, Tornquist M, Alenius S. A six-year study on respiratory viral infections in a bull testing facility. Vet J 2006 [Epub ahead of print]. [24] Bryson DG. Parainfluenza-3 virus in cattle. In: Dinter Z, Morein B, editors. Virus infections in ruminants. Amsterdam: Elsevier Publishers; 1990. p. 319–33. [25] Wellenberg GJ, van der Poel WH, Van Oirschot JT. Viral infections and bovine mastitis: a review. Vet Microbiol 2002;88:27–45. [26] Platt R, Burdett W, Roth JA. Induction of antigen-specific T-cell subset activation to bovine respiratory disease viruses by a modified-live virus vaccine. Am J Vet Res 2006;67:1179–84. [27] Toth TE, Jankura D. Analysis of bovine parainfluenza-3 virus replication in bovine embryonic lung cells by indirect fluorescent antibody and hemadsorption assays. J Virol Methods 1990;27:113–9.
7795
[28] van Wyke Coelingh KL, Winter CC, Tierney EL, London WT, Murphy BR. Attenuation of bovine parainfluenza virus type 3 in nonhuman primates and its ability to confer immunity to human parainfluenza virus type 3 challenge. J Infect Dis 1988;15:655– 62. [29] Frank GH. Serial passage of two plaque types of parainfluenza-3 virus: changes in hemagglutinating properties and cytopathic effect. Am J Vet Res 1970;31:1085–91. [30] Karron RA, Wright PF, Hall SL, Makhene M, Thompson J, Burns BA, et al. A live attenuated bovine parainfluenza virus type 3 vaccine is safe, infectious, immunogenic, and phenotypically stable in infants and children. J Infect Dis 1995;171:1107– 14. [31] Pennathur S, Haller AA, MacPhail M, Rizzi T, Kaderi S, Fernandes F, et al. Evaluation of attenuation, immunogenicity and efficacy of a bovine parainfluenza virus type 3 (PIV-3) vaccine and a recombinant chimeric bovine/human PIV-3 vaccine vector in rhesus monkeys. J Gen Virol 2003;84:3253–61. [32] Schmidt AC, McAuliffe JM, Murphy BR, Collins PL. Recombinant bovine/human parainfluenza virus type 3 (B/HPIV3) expressing the respiratory syncytial virus (RSV) G and F proteins can be used to achieve simultaneous mucosal immunization against RSV and HPIV3. J Virol 2001;75:4594–603. [33] Le Duy A, Zajic JF. A geometrical approach for differentiation of an experimental function at a point: applied to growth and formation. Biotechnol Bioeng 1973;15:805–10. [34] Schmidell W, Lima UA, Aquarone E, Borzani W. Biotecnologia industrial. S˜ao Paulo: Edgard Bluecher; 2001. [35] Reed LJ, Muench H. A simple method of estimating 50% endpoints. Am J Hyg 1938;27:493–7.
Vaccine 25 (2007) 7785–7795
Viral antigen production in cell cultures on microcarriers Bovine parainfluenza 3 virus and MDBK cells M.M. Conceic¸a˜ o a , A. Tonso b , C.B. Freitas c , C.A. Pereira a,∗ a b
Laborat´orio de Imunologia Viral, Instituto Butantan, Av. Vital Brasil 1500, 05503-900 S˜ao Paulo, Brazil Departamento de Engenharia Qu´ımica, Escola Polit´ecnica, Universidade de S˜ao Paulo, S˜ao Paulo, Brazil c Vall´ ee SA, Montes Claros, Minas Gerais, Brazil Received 6 March 2007; received in revised form 7 August 2007; accepted 26 August 2007 Available online 14 September 2007
Abstract Viral antigens can be obtained from infected mammalian cells cultivated on microcarriers. We have worked out parameters for the production of bovine parainfluenza 3 (PI-3) virus by Mandin–Darby Bovine Kidney (MDBK) cells cultivated on Cytodex 1 microcarriers (MCs) in spinners flasks and bioreactor using fetal bovine serum (FBS) supplemented Eagle minimal essential medium (Eagle-MEM). Medium renewal during the cell culture was shown to be crucial for optimal MCs loading (>90% MCs with confluent cell monolayers) and cell growth (2.5 × 106 cells/mL and a μx (h−1 ) 0.05). Since cell cultures performed with lower amount of MCs (1 g/L), showed good performances in terms of cell loading, we designed batch experiments with a lower concentration of MCs in view of optimizing the cell growth and virus production. Studies of cell growth with lower concentrations of MCs (0.85 g/L) showed that an increase in the initial cell seeding (from 7 to 40 cells/MC) led to a different kinetic of initial cell growth but to comparable final cell concentrations ((8–10) × 105 cells/mL at 120 h) and cell loading (210–270 cells/MC). Upon infection with PI-3 virus, cultures showed a decrease in cell growth and MC loading directly related to the multiplicity of infection (moi) used for virus infection. Infected cultures showed also a higher consumption of glucose and production of lactate. The PI-3 virus and PI-3 antigen production among the cultures was not significantly different and attained values ranging from, respectively, 7–9 log10 TCID50 /mL and 1.5–2.2 OD. The kinetics of PI-3 virus production showed a sharp increase during the first 24 h and those of PI-3 antigen increased after 24 h. The differential kinetics of PI-3 virus and PI-3 antigen can be explained by the virus sensitivity to temperature. In view of establishing a protocol of virus production and based on the previous experiments, MDBK cell cultures performed under medium perfusion in a bioreactor of 1.2 L were infected and the PI-3 virus production in 12 L attained 12 log10 TCID50 . Other than establishing a protocol for PI-3 production in MDBK cell cultures on Cytodex 1, the experiments are proposed as a basis for approaching the development of a virus production protocol in mammalian cells cultivated on microcarriers in bioreactors. © 2007 Elsevier Ltd. All rights reserved. Keywords: MDBK cells; PI-3 virus; Microcarriers
1. Introduction The development of animal cell culture bioprocesses for the production of viral antigens is today a solid basis for industries to provide vaccines for human and veterinarian use. The technology of animal cell cultivation constitutes the essence of the production based in fermentation and differs ∗
Corresponding author. Tel.: +55 11 37267222; fax: +55 11 37261505. E-mail address: [email protected] (C.A. Pereira).
0264-410X/$ – see front matter © 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.vaccine.2007.08.048
significantly from that using bacteria or yeast. Animal cells usually show a slower cell growing, anchorage dependency, higher hydrodynamic stress, metabolic complexity and need strict aseptic working conditions [1]. In view of the increasing demand of animal cell bioprocess for biological products, research & development in this area is subject of great scientific and economical interest [2,3]. Since most of the cell lines suitable for virus replication show an anchorage dependence, and that virus infection is often more effective in attached and spread out cells, dif-
7786
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
Fig. 1. Kinetics of MDBK cell growth on microcarriers. Medium renewal. 105 MDBK cells/mL were seeded on 2 g/L of Cytodex 1 MCs (12 cells/MC). Cultures were performed in 100 mL spinner flasks in batch (A) or in repeated batch with a half volume (50 mL) medium renewal at 24, 48, 60, 72 and 84 h (B). Cultures were also performed in 1.2 L bioreactor with a half volume (600 mL) medium renewal at 24, 48, 60, 72 and 84 h (C). The concentrations (g/L) of glucose, glutamine and lactate were periodically measured in culture supernatants and are indicated. The % of totally covered MCs (D) and the specific cell growth, μx (h−1 ) (E) in the cultures were measured. Cultures were performed at 37 ◦ C and 60 rpm. Dissolved oxygen (DO) in bioreactor cell cultures was controlled at 40% and pH at 7.3. Arrows indicate medium change. Data are the mean value of at least three separated experiments with standard deviations.
ferent cell substrates and particular bioprocesses have been developed, the scaling-up being always dependent on the surface available for cell attachment, spreading and multiplication [1,2,4,5]. A great step towards the establishment of high-density cell cultures and industrial bioprocesses was done by the cell cultivation on suspended beads, named microcarriers [6]. Cell cultures on microcarriers, at different concentrations, can be performed in bioreactors with fine control of cell environment (nutrients, pH, O2 and agitation) and using different modes of operation (batch, fed-batch and perfusion) [1]. Batches of 1000 L VERO cell cultures on microcarriers can be performed in bioreactors and are used for the production of vaccines as rabies and poliomyelitis [7,8]. Although the approach of viral antigen production in cell cultures on microcarriers is well known and consists in cell attachment and multiplication followed by virus infection, supernatant or cell fraction harvesting, concentration, purification and formulation, the bioprocess parameters may vary considerably from a cell/virus system to another. Parameters such as, the microcarrier type and concentration, cell culture medium composition, pH, dissolved O2 , multiplicity and time of virus infection, nutrients consumption and
supply, need to be worked out and rigorously determined [3–5,9,10]. An animal cell-based bioprocess for virus production has the particularity of constituting a system where the cell physiology is often sharply modified and redirected by the virus infection. It differs considerably from a cell-based recombinant protein synthesis that constitutes ultimately a consequence of a stable cell physiological state. As a consequence, parameters for a virus production bioprocess establishment have to be worked out by studying both, the cell growth and the virus replication steps [4,5,11–17]. The Mandin & Darby Bovine Kidney (MDBK) cell line represents an important cell line for the replication and consequent identification and production of several viruses of veterinarian interest [18–21]. The bovine parainfluenza 3 (PI-3) virus is an important agent of worldwide respiratory disease in cattle [22–25]. It induces cytopathic effect (CPE) in several cell lines including the MDBK and can be inactivated by heat. Vaccines containing inactivated or attenuated PI-3 virus have shown to be antigenic and protected animals from disease [26–29] In addition, the bovine PI-3 virus has been proposed as a human vaccine [30] and as virus vaccine vector to express foreign viral antigens for human immuniza-
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
7787
guideline for a bioprocess establishment involving other cell/virus/microcarrier systems.
2. Materials and methods 2.1. Cell line and virus The Mandin & Darby Bovine Kidney (MDBK) cell and the Bovine Parainfluenza 3 (PI-3) virus were obtained from de Laborat´orio de Viroses Bovinas do Instituto Biol´ogico (S˜ao Paulo, Brazil) and originated from the American Type Culture Collection (ATCC). MDBK cells were cultivated following established procedures with Eagle’s minimal essential medium (MEM) (Cultilab) supplemented with 10% fetal bovine serum (FBS) (Cultilab), 1.5 g/L of NaHCO3 (Labsynth) and 0.11 g/L of sodium pyruvate (Gibco-BRL). A cell bank was constituted at 132nd passage and stored in liquid nitrogen. The PI-3 virus was already adapted to MDBK cells. A virus stock in MEM + 10% FBS constituted at 4th passage with a titer of 8.16 log10 TCID50 /ml, frozen at −80 ◦ C, was used in all the experiments. Viable cell concentration was determined by cell counting in hemocytometer with trypan blue exclusion. 2.2. MDBK cell cultures and infection
Fig. 2. Light microscopy of MCs cell covering during MDBK cell cultures. 105 MDBK cells/mL were seeded on 2 g/L of Cytodex 1 MCs (12 cells/MC). Cultures were performed in 100 mL spinner flasks with a half volume (50 mL) medium renewal at 24, 48, 60, 72 and 84 h. Samples of MCs were taken at 36 and 96 h of culture, stained with trypan blue and observed at microscopy for evaluation of MC-cell covering degree. Dark areas are regions of the MC without cells and stained by trypan blue. This procedure was used throughout the study for observation of the MC cell covering and estimation of the % of totally covered MCs. In (A) and (B) we have, respectively, 13.8% and 95.2% of MCs totally covered by MDBK cells.
tion [31,32]. In spite of all these evidences, the literature is very scarce concerning studies of MDBK cell multiplication and/or PI-3 replication on microcarriers, in view of establishing a protocol for scaling up bioprocess of virus antigen production [21]. Besides establishing a protocol for production of PI3 virus antigen in MDBK cell cultures on Cytodex 1 in bioreactor, the approach in this study is proposed as a
For cell culture experiments, the MEM + 10% FBS was supplemented with 1 UI/mL of penicillin (Cultilab), 1 g/mL of streptomycin (Cultilab) and 2.5 g/mL of amphotericin B (Sigma). Cells were maintained in 75 or 150 cm2 T-flasks incubated at 37 ◦ C with 5% CO2 and passages were performed after trypsinization (tripsin solution at 0.2%). Cytodex 1 microcarriers (MCs) (GE Health Care) were used to cultivate the MDBK cells in spinners (Bellco) or in a bioreactor (celligen, New Brunswick) and had the following characteristics: density of 1.03 g/mL, average size of 190 m, surface of 4.400 cm2 /g and a number of MCs/g of 4.3 × 106 . They were PBS hydrated, following the manufacturer recommendation, and “in situ” sterilized. Spinners (Bellco) and bioreactor vessels were previously treated with dimetildiclorosilane (Merck). For cell culture, 100 mL spinners containing cells and MCs were placed on a magnetic stirrer (Bellco) inside an incubator at 37 ◦ C and 5% CO2 . For cell cultures in the 1.2 L working volume bioreactor, medium and cells were added by pressuring a bottle connected to the reactor. In both cases initial cell seeding is indicated as cells/mL and cells/MC, as calculated by knowing the number of cells/mL, the g/L of MCs and the number of MCs/g. The bioreactor had an automatically controlled fourgas using a double-screen cell-lift impeller for low shear, avoiding sparging and foam formation and providing the culture with O2 , CO2 , N2 and air for pH and dissolved oxygen (DO) control. It operated with adjustable temperatures (37 ◦ C) and agitation speed (60 rpm). Samples were collected via a pipe in the culture fluid and transferred
7788
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
Fig. 3. Kinetics of MDBK cell growth on microcarriers. MCs concentration. In (A), (B) and (C), 105 MDBK cells/mL were seeded on, respectively, 1–3 g/L of Cytodex 1 MCs (resulting in, respectively, 24, 12 and 8 cells/MC). Cultures were performed in 100 mL spinner flasks with a half volume (50 mL) medium renewal at 24, 48, 60, 72 and 84 h. The concentrations (g/L) of glucose, glutamine and lactate were periodically measured in culture supernatants and are indicated, the specific cell growth, μx (h−1 ) (D), the average number of cells per MC (E), and the % of totally covered MCs (F) in the cultures were measured. Cultures were performed at 37 ◦ C and 60 rpm. Arrows indicate medium change. Data are the mean value of at least 3 separated experiments with standard deviations.
to sample flasks by aspiration. Medium renewal was performed by using peristaltic pumps (Watson–Marlon) and a MC decanter to avoid the outflow of MCs. PI-3 viral infection of cell cultures was performed by adding virus suspensions at the indicated multiplicity of infection (moi). Medium containing virus during perfusion was collect in recipients placed at +4 ◦ C to avoid virus inactivation or degradation.
2.3. Analytical procedures Kinetics of cell multiplication on MCs, expressed by the number of cells per mL (cells/mL), was obtained by counting the cells attached to the MCs, at the indicated intervals throughout the culture, as follows: cell loaded-MCs in 1 mL of culture sample were allowed to settle, then cell lysis and nucleus staining were carried out in 1 mL of a hypotonic solution (0.1 M citric acid and 0.1% crystal violet) at 37 ◦ C for 10 min, and the nuclei counted in a Neubauer chamber [5].
The specific cell growth, expressed as μx (h−1 ), was determined as already described [33,34]. Kinetics of cell distribution on MCs, expressed by the % of MCs showing a cell confluent monolayer (% of totally covered MCs), was obtaining by periodically collecting cell samples from the cultures, staining the MCs with blue trypan (0.025%) for 5 min at room temperature, and counting the number of MCs totally or partially covered with cells or uncovered (Fig. 2). Kinetics of cell loading of MCs, expressed by the number of cells per MC (cells/MC), was obtained from the values of cell counting (cells/mL) and number of MCs per mL used for settling the culture. Kinetics of glucose (GLC) and glutamine (GLN) consumption or lactate (LAC) production, expressed by their concentrations (g/L) in the culture supernatants, were determined by enzymatic reaction using the YSI 2700 analyzer (Yellow Springs). Kinetics of PI-3 virus replication, expressed by the log10 TCID50 /mL, was measured as following. Dilutions
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
7789
Fig. 4. Kinetics of MDBK cell growth on microcarriers. Batch cultures, low MCs concentration and variable cell seeding. Cell seeding of 0.25 × 105 , 0.75 × 105 , 1.5 × 105 cells/mL or, respectively, 7, 20 or 40 cells/MC were used to initiate MDBK cell cultures on 0.85 g/L of Cytodex 1 MCs. Batch cultures were performed in 100 mL spinner flasks at 37 ◦ C and 60 rpm. Kinetics data show the cell growth and the number cells/MC (A), the specific cell growth, μx (h−1 ), (B), the % of totally covered MCs (C) and the concentrations (g/L) of glucose and lactate. Data are the mean value of at least three separated experiments with standard deviations.
of infected cell culture supernatants were added to 96well microplates containing a MDBK cell monolayer. The microplates were incubated at 37 ◦ C for 72 h and then observed under the microscope for the presence of cytopathic effect (CPE). Based on the highest sample dilution capable of inducing CPE in 50% of cell, the viral titer was calculated using the Reed and Muench [35] test. Kinetics of PI-3 viral antigen production, expressed by the optical density (OD), was evaluated in cell culture supernatants by a commercially available enzyme-linked immunosorbent assay (ELISA Pulmotest Parainfluenza 3, Bio-X diagnostic). Briefly, dilutions of infected cell culture supernatants were added to 96-well microplates containing polyclonal antibodies against the PI-3 virus. The microplates were incubated for 1 h at 37 ◦ C, washed and then peroxidaselabeled antibodies against PI-3 virus were added. The microplates were incubated again for 1 h at 37 ◦ C, washed and the reaction developed by the addition of hydrogen peroxidase and TMB (tetrametilbenzidine). OD was measured at 540 nm in ELISA reader.
volume change every 12 h after 48 h of cell seeding. Under these conditions, the level of glucose, glutamine and lactate were maintained and the spinner flasks cultures could reach up to 2.4 × 106 cells/mL at 96 h with 95% of totally covered MCs and a μx (h−1 ) of 0.035 at 36 h, in contrast to batch cultures that reached only 106 cells/mL at 120 h with 7% of totally covered MCs and a μx (h−1 ) of 0.030 at 24 h. Bioreactor cultures with medium renewal, although showing also good performance (2 × 106 cells/mL at 96 h with 85% of totally covered MCs and a μx (h−1 ) of 0.032 at 36 h), partially reproduced the data obtained with spinner flasks. As shown in Fig. 2, MCs without cells attached to the surface or partially covered by cells, could be easily identified and quantified upon trypan blue staining which does not stain the MCs bearing confluent cells or MC areas covered by cells. The trypan blue staining procedure was used throughout the study to monitor cell cultivation in terms of MC cell loading and distribution.
3. Results
The above-mentioned conditions of medium renewal were used in experiments designed to evaluate the influence of the MC concentration (1–3 g/L) in cell culture performance (Fig. 3). The procedures of the medium renewal established in the previous experiments (Fig. 1) were shown to provide all the cultures with suitable levels of glucose, glutamine and lactate. Cultures performed with 2 or 3 g/L of MCs reached, at 96 h, higher cells concentra-
3.1. MDBK cell cultures in spinners and bioreactor—medium renewal As shown in Fig. 1, the medium renewal during MDBK cell culture was shown to be determinant for cell growth and gave best results when started at 24 h and performed at a half
3.2. MDBK cell cultures in spinner flasks—MC concentration
7790
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
Fig. 5. Kinetics of MDBK cell growth on microcarriers. PI-3 virus infection. Cell seeding of 0.25 × 105 , 0.75 × 105 , 1.5 × 105 cells/mL or, respectively 7, 20 or 40 cells/MC were used to initiate MDBK cell cultures on 0.85 g/L of Cytodex 1 MCs. Batch cultures were performed in 100 mL spinner flasks at 37 ◦ C and 60 rpm. At 60 h the cultures were infected with 0.001, 0.1 or 10 moi of PI-3 virus. Data show the kinetics of cell growth (A), % of totally covered MCs (B) and the concentrations (g/L) of glucose and lactate (C) after infection. Data are the mean value of at least three separated experiments with standard deviations.
tions (∼2.3 × 106 cells/mL) than those performed with 1 g/L (1.5 × 106 cells/mL). The cell cultures with 1, 2 and 3 g/L of MCs showed, respectively, a μx (h−1 ) of 0.031, 0.035 and 0.036 at 36 h, a cell loading of 340, 280 and 190 cells/MC at 108 h and a % of totally covered MCs of 85, 95 and 80 at 96 h. 3.3. MDBK cell cultures in spinner flasks—batch cultures, low MCs concentration, variable cell seeding and virus infection In view of establishing conditions for PI-3 virus infection production, batch cell cultures were performed with 0.85 g/L of MCs and seeded with 7, 20 and 40 cells/MC. These cultures although showing different kinetics of cell growth, reached comparable final cell concentrations (8.5 × 105 , 8 × 105 and 9 × 105 cells/mL at 120, 96 and 84 h for cultures with, respectively, 7, 20 and 40 cells/MC). The higher was the cell seeding, the sooner a plateau of cell concentration was
reached and MCs loaded and faster glucose was consumed and lactate produced. Glucose attained lower levels at 96 h, and lactate higher levels at 120 h. The MCs cell loading and distribution values in the cultures ranged, respectively, from 230 to 280 cells/MC and 83–90% of totally covered MCs. The μx (h−1 ) values ranged from 0.030 to 0.031 (Fig. 4). Paralleled cultures were mock or PI-3 virus infected with moi of 0.001, 0.1 and 10 and cell replication, MC cell distribution and glucose/lactate metabolism evaluated. As compared to mock-infected cultures, the virus infected ones showed a decreased cell growth and MCs cell distribution directly related to the virus moi used, which was clearly observed after 24 h of infection. Glucose was consumed faster by infected cultures as compared to controls, reaching low levels (<0.2 g/L) at 48 h of infection. Differences in lactate production among mock and virus infected cultures were less evident (Fig. 5). Cytopathic effect with cell detachment from the MCs in these cultures could be clearly observed by microscopic examination of MCs (Fig. 6).
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
7791
moi 0.001 the virus titers plateau was reached only at 48 h of infection (8–9 log10 TCID50 /mL). As expected, a different kinetic was observed for antigen production. Cultures infected with the different moi showed comparable kinetics and values. In cultures infected with 0.1 or 0.001 moi the PI-3 antigen production sharply increased from 24 to 72 h attaining higher values in the first one. In cultures infected with 10 moi, remaining PI-3 virus used for inoculation was detected as antigen together with antigen produced by the cell cultures during infection. No significant differences were observed among cultures with different cell seeding (Fig. 7). Differences of kinetics of PI-3 virus and antigen observed are in a great part due to the virus sensitivity to the temperature. In assays of virus inactivation, as measured by the virus titer detected in samples stored for different times at different temperatures, the PI-3 virus titer decrease was directly related to the time and temperature of inactivation. At 37 ◦ C, the incubation temperature during cell culture and infection, the PI-3 virus titer sharply decreases attaining already <20% of its original value after 18 h of incubation (Fig. 8). 3.4. MDBK cell cultures and virus infection in bioreactor
Fig. 6. Light microscopy of MDBK cell cultures on MCs upon infection with PI-3 virus. 0.75 × 105 MDBK cells/mL were seeded on 0.85 g/L of Cytodex 1 MCs (20 cells/MC). Batch cultures were performed in 100 mL spinner flasks. They were infected with PI-3 virus at 60 h with a multiplicity of infection (moi) of 0.1. Cytopathic effect can be observed after 24 h (A), 48 h (B) and 72 h (C).
The kinetics of PI-3 virus production was measured by “in vitro” cell infection assay (log10 TCID50 /mL) (Fig. 7A) and the PI-3 antigen by ELISA (OD) (Fig. 7B). In cultures infected with moi 10, there was no significant increase in the virus titers, the kinetics showing a plateau from 0 to 72 h (7–8 log10 TCID50 /mL). In cultures infected with moi 0.1, the virus titers increased sharply during the first 24 h of infection, reaching then a plateau, where they remain until 72 h (7.5–8.5 log10 TCID50 /mL). In cultures infected with
MDBK cell cultures were performed in bioreactor for PI3 virus infection with medium perfusion before and after virus infection. Data show the kinetics of cell growth and virus production (Fig. 9). Cell concentration attained values of ∼2.8 × 106 cells/mL at 108 h of culture, when they were infected with 0.1 moi of PI-3 virus. At this time the cell culture showed ∼75% of MCs with confluent cell monolayers and the concentration of glucose and lactate were controlled at, respectively, ∼0.2 and ∼0.7 g/L. To attain these values the medium (MEM + 10% FBS) perfusion rate was increased from 0.8 to 1.7 volumes/24 h. After infection, the medium (MEM + 2% FBS) perfusion rate was settled in 1 volume/24 h allowing the cell concentration and MC loading to remain constant for the first 36 h and the concentration of glucose and lactate to be controlled at, respectively, ∼0.1 and ∼0.7 g/L. In view of the perfusion procedure, the virus produced by the cultures could be collected both in the supernatant of the culture as well as in a reservoir placed at +4 ◦ C with the perfused medium. A total of 3 L of perfused medium and 1 L of supernatant were collected after 72 h of infection. The PI-3 virus production showed virus titers of ∼8 log10 TCID50 /mL in culture supernatant and from 6 to 7.5 log10 TCID50 /mL in perfused medium at different times after infection. A virus titer of 12 log10 TCID50 was obtained in 4 L of total collect medium. The PI-3 antigen production increased with the time of culture reaching high values.
4. Discussion Animal cell culture bioprocesses are today imperative for the preparation of cell-derived products. The main
7792
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
Fig. 7. Kinetics of PI-3 virus (A) and antigen (B) synthesis in MDBK cell cultures. Cell seeding of 0.25 × 105 , 0.75 × 105 , 1.5 × 105 cells/mL or, respectively 7, 20 or 40 cells/MC were used to initiate MDBK cell cultures on 0.85 g/L of Cytodex 1 MCs. Batch cultures were performed in 100 mL spinner flasks at 37 ◦ C and 60 rpm. At 60 h cultures were infected with 0.001, 0.1 or 10 moi of PI-3 virus. Data show the kinetics of PI-3 virus synthesis, as evaluated by the infectious virus titers (A) and the antigen quantification by ELISA (B). Data are the mean value of at least three separated experiments with standard deviations.
difficulty encountered when one attempts to establish a protocol for large-scale production is to provide high-density cell cultures with the best environment possible allowing them to synthesize the large amounts of product. When virus particles are the envisaged product, the establishment of a production protocol has to be approached in two steps, the first one consisting in the cell multiplication to high densities and the second one consisting in the virus infection, replication and synthesis. The main reason for this particular characteristic is that the virus infection often redirects the cell metabolism leading to cell culture conditions that may differ considerably from the one observed for non-infected cells. Upon infection, cells frequently show a modified nutrients consumption, metabolites production, oxygen requirement and hydrodynamic sensitivity. The present study was carried out with the aim of establishing bioprocess conditions for large-scale production of bovine parainfluenza 3 (PI-3) virus antigen in MDBK cell cultures on MCs, a veterinarian vaccine with commer-
Fig. 8. Loss of infectivity of PI-3 virus suspensions. Temperature sensitivity. A stock of PI-3 virus was exposed for 24, 48 and 72 h to temperatures of 10, 20, 30 and 37 ◦ C in a water bath. The viral suspensions were than frozen and titrated. Data, expressed as log10 TCID50 /mL, are the mean value of at least 3 separated experiments with standard deviations.
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
7793
Fig. 9. Kinetics of cell growth and virus synthesis in medium perfused and PI-3 infected MDBK cell cultures on microcarriers. MDBK cell cultures were performed on 3.4 g/L of Cytodex 1 MCs with a cell seeding of 1.5 × 105 cells/mL (10 cells/MC) in a 1.2 L bioreactor. Cultures were performed at 37 ◦ C, 60 rpm with dissolved oxygen (DO) controlled at 40% and pH at 7.3. PI-3 virus infection was done with 0.1 moi at 108 h of culture. Data show the perfusion rates employed (volume/24 h), the kinetics of cell growth (A), % of totally covered MCs (B), the concentrations (g/L) of glucose and lactate (C), the infectious PI-3 virus (D) and antigen produced (E). Data are the mean value of at least three separated experiments.
cial, environmental and animal health importance [23,25]. Our approach for establishing these bioprocess conditions is proposed as a guideline for the establishment other cell/virus/microcarrier systems. Basic conditions of MDBK cell multiplication on MCs were first investigated in 100 mL spinner cultures with or without medium renewal. Kinetics of cell multiplication and metabolism as well as specific cell growth and MC cell loading/distribution were studied. Data show clearly that specific cell growth, cell density and MCs cell loading/distribution were dependent on medium renewal. Among several protocols of time and medium renewal (data not shown) the one leading to best performances indicated half volume of medium renewal initiated 24 h after cell seeding. Preliminary cultures on 1.2 L bioreactor using parameters established for 100 mL spinners roughly reproduced the cell growth performance (Fig. 1).
In view of studying the influence of MCs concentration on cell culture performance, the established conditions of medium renewal were used in experiments with different MCs concentrations (1–3 g/L). In all the cultures, the specific cell growth and % of fully covered MCs were comparable, indicating good environmental culture conditions. Although showing a lower final cell density, the cultures performed with 1 g/L of MCs showed a higher MC cell loading. At this stage and considering the limitations of spinner cell cultures, these observations allowed us to conclude that the surface available for cell multiplication limited the cell growth and that the cell loading, in cultures with higher MCs concentrations (2 and 3 g/L), could be further increased by providing the cultures with even better environment (Fig. 3). In order to investigate the cell growth and virus infection in batch conditions with high MCs loading, experiments with lower MCs concentration and different cell seeding were
7794
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795
designed. As an indication that this approach was suitable for studies of virus infection, all the cultures, regardless the cell seeding, were shown to attain relatively high levels of specific cell growth, cell density (cell/mL) and MC cell loading/distribution. It is important to take into consideration that they were performed without medium renewal and as a consequence low concentrations of glucose in cultures were rapidly attained (Fig. 4). Upon PI-3 virus infection, the cell cultures showed a decreased cell growth and MCs cell loading directly related to the virus moi used and paralleling the induced cytopathic effect observed (Figs. 5 and 6). Higher glucose consumption was also detected in infected cultures. Infected culture supernatants were assayed for the presence of PI-3 virus and PI-3 antigen. As expected, and explained by the virus temperature sensitivity, the virus titers increased during the first 24 h attaining a plateau and the virus antigen increased progressively during the infection. Best results were obtained with virus infection of 0.1 moi, regardless the cell seeding. The ELISA data suggested that PI-3 antigen was not undergoing important degradation during the culture period studied (72 h). The virus titer and antigen measures are of importance and express actually different viral product characteristics. The virus titer express the ability of the cultures to produce infectious virus that in turn are capable of further promote secondary infections, and the virus antigen expressing, ultimately, the virus protein relevant as a product for immunization (Figs. 7 and 8). Once the conditions of cell growth with medium renewal (Figs. 1 and 3) and virus infection (Figs. 4–8) were established, we could assay a suitable protocol for MDBK cell culture and PI-3 virus infection in a 1.2 L bioreactor. Conditions for MDBK cell cultures under medium perfusion in bioreactor were preliminary assayed (data not shown) and then established with a MC concentration of 3.4 g/L (10 cells/MC) and rates of perfusion varying from 0.8 to 1.7 volumes/24 h. They were infected with 0.1 moi and the kinetics of virus and antigen production evaluated. The cell cultures attained, at the moment of infection (108 h), a cell density of ∼2.8 × 106 cells/mL with ∼75% of MCs with confluent cell monolayers and the concentration of glucose and lactate controlled in, respectively, ∼0.2 and ∼0.7 g/L. These conditions were maintained after virus infection and allowed the harvesting, after 72 h, of 4 L of virus suspension with an infectious titer of 12 log10 TCID50 . The PI-3 virus antigen production increased with the time of culture reaching high values at 72 h. Our development study showed here led to the establishment of a pilot protocol for the production of PI-3 viral antigen in MDBK cell cultures on Cytodex 1 MCs performed in a 1.2 L bioreactor, that can be employed for the preparation of PI-3 antigen batches for immune protection investigations and easily scaled-up for production in larger volumes. Although the establishment of a given production protocol may vary depending on the virus, cell substrate, medium and vessel culture, common features exist and a general technological approach is helpful as a guideline for optimization.
So, experiments showed here are proposed as a flowchart basis for approaching the development of a virus production protocol in mammalian cells cultivated on microcarriers in bioreactors.
Acknowledgements We thank Maryvonne Kientz for reviewing the manuscript. This work was supported in part by grants from Vall´ee S.A and Fundac¸a˜ o Butantan. Carlos A. Pereira is recipient of CNPq—IA senior research fellowship. Mateus M. Conceic¸a˜ o had a scholarship from Vall´ee S.A.
References [1] Freshney RI. Culture of animal cells. A manual of basic technique. 4th ed. New York: John Wiley & Sons Inc.; 2000. [2] Hu WS, Aunins JG. Large-scale mammalian cell culture. Curr Opin Biotechnol 1997;27:148–53. [3] Merten O-W. Introduction to animal cell culture technology–past, present and future. Cytotechnology 2006;50:1–7. [4] Mendonc¸a RZ, Arr´ozio SJ, Antoniazzi MM, Ferreira Jr JMC, Pereira CA. Metabolic active-high density VERO cell cultures on microcarriers following apoptosis prevention by galactose/glutamine feeding. J Biotechnol 2002;97:13–22. [5] Yokomizo AY, Antoniazzi MM, Galdino PL, Azambuja Jr N, Jorge SAC, Pereira CA. Rabies virus production in high Vero cell density cultures on macroporous microcarriers. Biotechnol Bioeng 2004;85:506–15. [6] Van Wezel AL. Growth of cell strains and primary cells on microcarriers in homogeneous culture. Nature 1967;216:64–5. [7] Montagnon B, Vincent-Falquet JC, Fanget B. Thousand litre scale microcarrier culture of VERO cell for killed poliovirus vaccine: promising results. Dev Biol Stand 1984;55:37–8. [8] Pullen K, Johnston MD, Philips AW, Ball GD, Finter WB. Very large scale suspension cultures of mammalian cells. Dev Biol Stand 1984;60:175–7. [9] Mendonc¸a RZ, Prado JCM, Pereira CA. Attachment, spreading and growth of VERO cells on microcarriers for the optimization of large scale cultures. Bioprocess Eng 1999;20:565–71. [10] Ng YC, Berry JM, Butler M. Optimization of physical parameters for cell attachment and growth on macroporous microcarriers. Biotechnol Bioeng 1996;50:627–35. [11] Berry JM, Barnab´e N, Coombs KM, Butler M. Production of reovirus type-1 and type-3 from VERO cells grown on solid and macroporous microcarriers. Biotechnol Bioeng 1999;62:9–12. [12] Bleckwenn NA, Golding H, Bentley WE, Shiloach J. Production of recombinant proteins by vaccinia virus in a microcarrier based mammalian cell perfusion bioreactor. Biotechnol Bioeng 2005;90:663–74. [13] Genzel Y, Fischer M, Reichl U. Serum-free influenza virus production avoiding washing steps and medium exchange in large-scale microcarrier culture. Vaccine 2006;24:3261–72. [14] Ghendon YZ, Markushin SG, Akopova II, Koptiaeva IB, Nechaeva EA, Mazurkova LA, et al. Development of cell culture (MDCK) live coldadapted (CA) attenuated influenza vaccine. Vaccine 2005;23:4678–84. [15] Mohler L, Flockerzi D, Sann H, Reichl U. Mathematical model of influenza A virus production in large-scale microcarrier culture. Biotechnol Bioeng 2005;90:46–58. [16] Sun MB, Jiang YJ, Li WD, Li PZ, Li GL, Jiang SD, et al. A novel process for production of hepatitis A virus in Vero cells grown on microcarriers in bioreactor. World J Gastroenterol 2004;10:2571–3.
M.M. Concei¸ca˜ o et al. / Vaccine 25 (2007) 7785–7795 [17] Wu SC, Liu CC, Lian WC. Optimization of microcarrier cell culture process for the inactivated enterovirus type 71 vaccine development. Vaccine 2004;22:3858–64. [18] Gogorza LM, Moran PE, Larghi JL, Segui R, Lissarrague C, Saracco M, et al. Detection of bovine viral diarrhea virus (BVDV) in seropositive cattle. Prev Vet Med 2005;72:49–54. [19] Takiuchi E, Medici KC, Alfieri AF, Alfieri AA. Bovine herpesvirus type 1 abortions detected by a semi-nested PCR in Brazilian cattle herds. Res Vet Sci 2005;79:85–8. [20] Townsend J, Duffus WP, Williams DL. Immune production of interferon by cultured peripheral blood mononuclear cells from calves infected with BHV1 and PI3 viruses. Res Vet Sci 1988;45:198–205. [21] Varani J, Bendelow MJ, Chun JH, Hillegas WA. Cell growth on microcarriers: comparison of proliferation on and recovery from various substrates. J Biol Stand 1986;14:331–6. [22] Murphy FA, Gibbs EPJ, Horzinek MC, Studdert MJ. Veterinary virology. London: Academic Press; 1999. p. 411–28. [23] Hagglund S, Hjort M, Graham DA, Ohagen P, Tornquist M, Alenius S. A six-year study on respiratory viral infections in a bull testing facility. Vet J 2006 [Epub ahead of print]. [24] Bryson DG. Parainfluenza-3 virus in cattle. In: Dinter Z, Morein B, editors. Virus infections in ruminants. Amsterdam: Elsevier Publishers; 1990. p. 319–33. [25] Wellenberg GJ, van der Poel WH, Van Oirschot JT. Viral infections and bovine mastitis: a review. Vet Microbiol 2002;88:27–45. [26] Platt R, Burdett W, Roth JA. Induction of antigen-specific T-cell subset activation to bovine respiratory disease viruses by a modified-live virus vaccine. Am J Vet Res 2006;67:1179–84. [27] Toth TE, Jankura D. Analysis of bovine parainfluenza-3 virus replication in bovine embryonic lung cells by indirect fluorescent antibody and hemadsorption assays. J Virol Methods 1990;27:113–9.
7795
[28] van Wyke Coelingh KL, Winter CC, Tierney EL, London WT, Murphy BR. Attenuation of bovine parainfluenza virus type 3 in nonhuman primates and its ability to confer immunity to human parainfluenza virus type 3 challenge. J Infect Dis 1988;15:655– 62. [29] Frank GH. Serial passage of two plaque types of parainfluenza-3 virus: changes in hemagglutinating properties and cytopathic effect. Am J Vet Res 1970;31:1085–91. [30] Karron RA, Wright PF, Hall SL, Makhene M, Thompson J, Burns BA, et al. A live attenuated bovine parainfluenza virus type 3 vaccine is safe, infectious, immunogenic, and phenotypically stable in infants and children. J Infect Dis 1995;171:1107– 14. [31] Pennathur S, Haller AA, MacPhail M, Rizzi T, Kaderi S, Fernandes F, et al. Evaluation of attenuation, immunogenicity and efficacy of a bovine parainfluenza virus type 3 (PIV-3) vaccine and a recombinant chimeric bovine/human PIV-3 vaccine vector in rhesus monkeys. J Gen Virol 2003;84:3253–61. [32] Schmidt AC, McAuliffe JM, Murphy BR, Collins PL. Recombinant bovine/human parainfluenza virus type 3 (B/HPIV3) expressing the respiratory syncytial virus (RSV) G and F proteins can be used to achieve simultaneous mucosal immunization against RSV and HPIV3. J Virol 2001;75:4594–603. [33] Le Duy A, Zajic JF. A geometrical approach for differentiation of an experimental function at a point: applied to growth and formation. Biotechnol Bioeng 1973;15:805–10. [34] Schmidell W, Lima UA, Aquarone E, Borzani W. Biotecnologia industrial. S˜ao Paulo: Edgard Bluecher; 2001. [35] Reed LJ, Muench H. A simple method of estimating 50% endpoints. Am J Hyg 1938;27:493–7.