EP0808361A1

EP0808361A1 - Novel replication process

Info

Publication number: EP0808361A1
Application number: EP95939952A
Authority: EP
Inventors: Robert G. Webster; Nicolai V. Kaverin
Original assignee: St Jude Childrens Research Hospital
Current assignee: St Jude Childrens Research Hospital
Priority date: 1994-11-16
Filing date: 1995-11-13
Publication date: 1997-11-26
Also published as: NO972239L; JPH11509081A; NO972239D0; AU694592B2; WO1996015232A1; AU4158996A; EP0808361A4; NZ296861A

Abstract

The proposed process permits replication of human influenza virus at a low multiplicity of infection in a Vero cell line by maintaining a trypsin concentration of at least 0.05 νg/ml of the culture medium throughout the growth cycle.

Description

DESCRIPTION

NOVEL REPLICATION PROCESS

Technical Field

This invention relates to a process for viral replication in mammalian cells and particularly viral replication of human influenza virus in Vero cell culture.

Background Art

The major influenza virus glycoprotein, he agglutinin (HA) , is synthesized in infected cells as a single polypeptide. Post- translational cleavage of the HA forms two subunits, HA1 and HA2, joined by a disulfide bond. Cleavage is essential to the production of infectious virus; virions containing uncleaved HA are non-infectious. The process can occur intracellularly or extracellularly. The HAS of human, swine, and most avian influenza virus strains cannot be cleaved by ubiquitous intracellular proteases. Therefore, replication of these viruses in cell culture requires the addition of trypsin to the maintenance medium to ensure HA cleavage thereby permitting activation of the progeny virus so that the infection can proceed.

For the past several decades, fertilized chicken eggs have been used to produce influenza virus in large quantities. Killed influenza vaccines are purified from virus-containing chick embryo allantoic fluid. However, a large body of data now suggests that this is not an ideal system. Even a single passage of a human influenza virus isolate in eggs can lead to the selection of variants that differ in their antigenic specificity from the original virus. By contrast, viruses isolated and passaged exclusively in mammalian cell cultures fully retain their antigenic characteristics, a feature that would prove highly advantageous in vaccine production. However,' the cell lines routinely used in laboratory studies, including the favored line, Madin-Darby canine kidney (MDCK) cells, have not been certified for virus vaccine production.

In contrast to influenza A and B viruses grown in eggs, those isolated in maπunalian host cells possess structurally homogenous hemagglutinin molecules (Has) that are identical to the predominant Has of the original clinical isolate [Katz et al, Virology, Vol. 165 (1988), pp. 446-456; Robertson et al, Virology, Vol. 179 (1990), pp. 35-40]. Moreover, influenza viruses grown in mammalian cells elicit neutralizing and hemagglutinin inhibition (HI) antibodies in human sera more readily and at higher titers than do their egg-grown counterparts. An experimentally inactivated influenza virus grown in Madin-Darbin canine kidney (MDCK) cells introduced higher HI in neutralizing antibody titers than did egg- grown counterpart virus, and provided superior protection of ferrets against subsequent challenge with infectious virus grown in either MDCK cells or embryonated eggs [Katz et al, J. Infect. Dis., Vol. 160 (1989), pp. 191-198; Wood et al, Virology, Vol. 171 (1989), pp. 214-221]. These observations underscore the need for a mammalian cell line that could be used to replace chicken eggs in the production of influenza virus vaccines and diagnostic reagents. Mammalian cell grown virus may also have advantages for easier virus purification.

Influenza viruses can be propagated in several types of primary cell cultures including chick embryo kidney, chick embryo lungs, monkey kidney, canine kidney, bovine kidney, chick kidney, guinea pig kidney, and chick embryo fibroblasts. However, primary tissue cultures are unlikely to be useful as a substrate for vaccine production for several reasons, including contamination by various endogenous agents, the variable quality of the cells, different sensitivities to variants of the same virus, and, of course, high cost and difficulties in obtaining and preparing the tissue cultures. Diploid tissue cultures, such as WI-38, have been used to produce vaccines against poliomyelitis, adenovirus types 4 and 7, rubella, measles, and rabies viruses. Although human diploid (MRC-5) cells can support the growth of influenza viruses, such systems have stringent growth media requirements and are expensive to maintain, making them suboptimal for large-scale production of vaccines. Disclosure of the Invention

This invention of a process for ensuring replication of human influenza virus at a low multiplicity of infection in a mammalian cell line involves maintaining a consistent minimum concentration of trypsin (about 0.05 g/ml) in the culture medium.

Vero cells are sensitive to a spectrum of viruses, including: enteroviruses, measles and parainfluenza viruses, herpes viruses, andenoviruses , rhabdoviruses and some arboviruses. Low-passage- number Vero cells lack tumorigenicity, do not contain adventitious viruses and can support efficient proliferation of many types of viruses. This cell line has been used successfully for the production of vaccines against poliomyelitis and rabies. The Vero cell line is suitable for cultivation of infectious influenza A viruses and for primary isolation of currently circulating influenza A (H3N2) strains. The data in the examples on the adaptation of influenza A/England/lZ53 (H1N1) [HG] strain to Vero cell culture, its growth characteristics and antigenic stability, and the likelihood of obtaining high yields of viral proteins with Vero is compared to MDCK cells.

The first attempts to obtain high yields at low multiplicities of infection in mammalian cell line certified for vaccine production [Vero (WHO) , a subline of African green-monkey kidney cells] were unsuccessful. Subsequent studies to identify the cause(s) of this failure implicated loss of trypsin from the cell maintenance medium. When infected with influenza A virus at a multiplicity of at least 0.005 TCID_j0 per cell, Vero (WHO) cells produced yields of virus comparable to those produced in Madin- Darby canine kidney (MDCK) cells. However, at lower multiplicities of infection, multicycle growth was blocked early in the course of infection, the progress of the cytopathic effect was stopped, and the final virus yields were low.

To test the possibility that loss of trypsin activity was responsible for the observed effect, we used a sensitive fluorogenic substrate to measure this activity in the culture fluid of Vero (WHO) cells. The results indicated a rapid decrease of trypsin activity. Similar findings were made with MDCK, rhesus monkey kidney LLC-MK_2/ and swine kidney cells although the rates of decrease were much slower than in Vero (WHO) cells. The causative role of trypsin was verified in experiments in which repeated addition of the enzyme to cell cultures restored multicycle virus growth and permitted high virus yields to be obtained at a low multiplicity of infection.

Tests showed that trypsin concentrations of at least 0.05 μq/ml in the cell culture were essential for securing high virus yields and that a concentration of about 0.1 μgZml was optimal when the multiplicity of infection ranged from about 1 x 10"⁵ and 1 x 10"⁶ TCID₅₀ per cell; satisfactory results were obtained at about 5 x 10"⁷ TCID_J0 P^er cell. Thus, trypsin had to be maintained from about 0.05 to 0.5 μq/T&l to secure adequate yields of virus.

A higher volume of maintenance medium per area of cell monolayer also required slight improvement of multicycle virus growth at low input doses, most likely because of a lower concentration of trypsin-inactivating factor in the medium. It seems that efficient replication of virus in MDCk cells is possible because of a relatively slow rate of trypsin inactivation. The level of trypsin activity necessary for efficient HA cleavage, to the extent of ensuring multicycle virus growth, is much lower than the initial trypsin concentration in the maintenance medium, so that the infection proceeds even though the trypsin activity decreases, provided the decrease in not so rapid as in Vero (WHO) cells.

It might prove useful, however, to supplement even MDCk cell cultures with trypsin, particularly in situations where small volumes of medium are used with large amounts of cells, such as roller cultures, etc. At the very least, trypsin activity in culture fluid should be monitored routinely.

The nature of the factor(s) responsible for trypsin inactivation in cell cultures is not known. Once it is secreted into the medium and collects there, it rapidly inactivates trypsin. The kinetics of trypsin inactivation in cell cultures appears to reflect the accumulation of the inhibitory factor rather than its interaction with trypsin.

Passage of the trypsin-inactivating factor through a series of graded filters indicated a molecular mass very close to 100 kilodaltons (kDa) . Alternately, the protein may consist of two fractions, one larger than 100 Kda and the other between about 50 and about 100 (Kda) . Of the many inhibitors of serine proteinases that block the cleavage of low molecular weight substrates by trypsin only, very few have a molecular mass as high as the one estimated for our factor. Inter-α-trypsin inhibitor (ITI) of human plasma is represented by a native molecule of 180 Kda as well as a species of lower molecular mass that retains inhibitory activity. Related inhibitors were detected in baboon plasma. If, as suggested by our molecular weight estimates, the trypsin- inactivating factor in cell cultures belongs to the class of proteins that inhibit proteinases, our findings of a putative inhibitor of trypsin activity may be of value in the studies of serine proteinase inhibitors. The inhibitors of this family, although numerous and extensively studied, are mostly derived from such substances as plants, bovine pancreas, human and animal plasma, tissues of invertebrate species, etc. For this reason, their structure in enzymological properties are far better known than their biosynthesis, intracellular transport and secretion mechanism. Thus, inhibitors produced by cultured cells may prove to be a valuable asset. In the best mode for carrying out the invention, there are described several preferred embodiments to illustrate the invention. However, it is to be understood that the invention is not intended to be limited to the specific embodiments contained therein.

Best Mode for Carrying Out the Invention

The following materials and methods were used.

Example 1 Cells:

The Vero (WHO) cell line, deposit no. 1297, was obtained from the American Type Tissue Collection at the level of the 134th passage. The cells were cultivated as monolayers in Falcon Labware 250 cm³ flasks at 37°C and 5% Co₂ in a growth medium of Eagles minimal essential medium (MEM) supplemented with 10% unheated fetal calf serum. For the growth of Madin-Darby canine kidney (MDCK) cells and rhesus monkey kidney (LLC-MK₂) cells, the medium used was MEM with 5% fetal calf serum heated 30 minutes at 56°C. For the cultivation of swine kidney cell line (SwK) , RPMI 1640 medium was used with 5% heated fetal calf serum. For the experiments involving infection or mock-infection, the cells were grown either in 50 cm³ flasks or in 6-well, 24-well, and 96-well plates (Falcon Labware) . Cell monolayers were washed three times with PBS and overlaid with maintenance medium. The latter had the same composition as the growth medium for each cell line, the serum being omitted and 0.3% bovine serum albumin (BSA) added. Unless otherwise stated, the maintenance medium contained TPCK-trypsin (Worthington) at 1.0 μgZml. Plaque assays were performed with TPCK-treated trypsin (2.5 μgZml) .

Viruses:

Vero (WHO) -adapted influenza AZEnglandZlZ53 (HlNl) [HG] , A/FW/1Z50 (HlNl), and A Aichi 2 68-PR 8 34 (H3N2) [X-31] viruses were used. The viruses were passaged 5 times in Vero (WHO) cell cultures, and the final stock virus preparations contained about 10⁷³ to 10⁸-²⁵ TCID₅₀/°-2 ml and about 32 to about 128 HAU. In the preliminary experiments, the Vero (WHO)-adapted A Rome 4 (HlNl) strain was used (10⁶⁷ TCID_J0Z0.2 ml, about 15 to about 32 HAU). HA and infectivity titration were performed essentially as described in "Advanced Laboratory Techniques for Influenza Diagnosis" [Immunol. Ser. 6, pp. 51-57 (1975)]. HA titrations were done in mirotiter plates. Infectivity was measured by an end point titration technique in MDCK cells grown in 96-well plates with CPE evaluation at 72h postinfection.

Assessment of Trypsin Activity:

A highly sensitive assay of trypsin activity based on application of a fluorogenic substrate, BAAMC (Na-benzoyl-L- arginine-7-amido-4-methylcoumarin-hydro-chloride; Sigma) was used. The substrate was dissolved to a final concentration of about 0.2 Mm in a buffer containing 50 Mm of Tris-Hcl, Ph about 8.0, 10 Mm CaCl₂ and 1% DMSO. A 0.1 ml sample of trypsin-containing cell culture fluid was added to about 0.9 ml of BAAMC solution and incubated at 37°C for 1 hour. The samples were placed on ice and assayed in a Perkin-Elmer MPF-44B fluorescence spectrophotometer at activation and emission wavelengths of about 380 and 460 nm, respectively.

RESULTS

Inefficient Multicycle Replication of Influenza in Vero (WHO) Cells:

In the attempts to passage influenza A virus in Vero (WHO) cells, difficulties were encountered in the production of high virus yields using low multiplicities of infection (m.o.i.). When the cells were grown in 50 cm³ flasks, the m.o.i. had to be at least 0.005 TCID₅₀Zcell to produce maximal yields—a concentration that would be impractical for use in vaccine production. At lower input doses, the tiers were low or the virus failed to accumulate at all. In most instances, the accumulation of virus in the cultures infected at the low m.o.i. stopped after 48 hours postinfection. The progress of the cytopathic effect also ceased. This pattern, however, occurred to a different extent in different kinds of plasticware; it was strongly expressed in 50 cm³ flasks, less strongly in 6-well plates, and even less in 24-well plates and practically not at all in 96-well plates. The mode of infection and incubation in the experiments with all kinds of plasticware was identical. The only difference was the volume of maintenance medium per unit of monolayer, that is, the amount of culture fluid per cell. This dependence of the final yields on the input dose was not observed in MDCK cells, irrespective of the plasticware used. An example of the multiplicity dependence of the final yields is presented in an experiment with Vero (WHO)-adapted influenza A/Rome/49 (HlNl) virus strain (Table 1) . Restoration of Multicycle Virus Growth by Repeated Addition of Trypsin:

To verify the abrogation of influenza virus accumulation in Vero cell cultures was due to the loss of trypsin activity in the culture medium, several experiments were performed in which trypsin concentration was restored in the course of infection by repeated additions of trypsin to the culture medium. This procedure led to an increase of the virus production in the cultures infected with low input doses, thus ensuring high final yields irrespective of the multiplicity of infection. The effect was especially evident in 50 cm³ flasks (Table 2) and 6-well plates with dense confluent monolayers (Table 3) , that is, in the conditions favoring a rapid loss of trypsin activity. In 6-well plates with non-confluent monolayers, as well as in 24-well plates, the effect was much less dramatic because in this case the multicycle growth of the virus was fairly efficient under standard conditions (Table 3) .

Example 2

Viruses:

Seventy-two influenza A virus strains, obtained from the repository of St. Jude Children's Research Hospital, were investigated for their growth characteristics in Vero cells.

Vero cells were infected with the AZEnglandZlZ53 (HlNl) [High Growth, HG] strain of influenza virus, a reassortant containing the gene segments coding for the two surface glycoproteins (HA and NA) from A/EnglandZlZ53 (HlNl) and the remaining six genes from A PRZ8 4 (HlNl) . For the first four passages, the virus was left to adsorb for 1 hour at 37°C, after which the monolayer was washed twice with warm phosphate buffered saline (PBS) solution to remove the unadsorbed viruses. Serum-free MEM with 0.3% bovine serum albumin (BSA) was then added; the maintenance medium contained

TPCK-treated trypsin at about 1.0 μgZml. The input dose of virus was 10"²-10^"3 PFU cell. The material for further passage was collected 72 hours postinfection (p.i.), with trypsin (final concentration, about 1.0 μg/ml) added at 48 hours p.i.. Cells were infected with serial lθ-fold dilutions of virus, which were added to the washed cell monolayer without previous adsorption. Virus accumulation was estimated by visual determination of the cytopathic effect (CPE) and HA titration of culture fluid at different times p.i. (24, 48 and 72 hours). Infectivity titrations were performed in 96-well plates. Tissue culture infectious doses (TCID_joZml and egg infectious doses (EID)₅₀/ml values were calculated by the formula of Karber [Arch, Exp. Path. Pharmak. , Vol. 162, pp. 480-483 (1931).

Virus-containing culture fluids were concentrated in an Amicon system and purified by differential sedimentation through 25-70% sucrose gradients. Whole virus protein estimates were made by the method of Bradford (1976) . To determine the yield of HA protein in virus grown in Vero and MDCK cells, the virus proteins were separated by gradient (4-20%) SDS-PAGE and intensity of Coomassie blue-stained protein bands was quantified by densitometry.

Virus Isolation From Clinical Material:

Influenza A viruses were isolated from the throat washings of patients with clinical signs of influenza and collected in PBS to which 0.7% BSA was added. Cell culture (both Vero and MDCK) or embryonated chicken eggs were infected directly with freshly collected (not frozen) throat washings. Chicken eggs were inoculated amniotically and allantoically. Clinical samples used for isolation were inoculated undiluted and at 10^"' and 10^"2 dilutions and incubated for 72-96 hours. Trypsin was added at 0 and 48 hours p.i. (about 1.0 /.g/ml) and tested for virus replication with chicken and guinea pig erythrocytes. Each sample was given at least two passages in chicken eggs or cell cultures before being considered negative.

Immunoloσical Tests:

Monolayer antibodies to the A/Baylor/5700/82 (HlNl) and A/Baylor 11515/82 (HlNl) strains were prepared by the method of Kδhler and Milstein (1976) . Polyclonal antisera to influenza A/England/1/53 virus (20 passages in Vero cells) were prepared in chickens by intravenous injection of virus-containing culture fluid. HA and HI reactions were performed in microtiter plates with about 0.5% (v/v) chicken erythrocytes. Guinea pig erythrocytes (about 0.4% v v) were used to analyze primary influenza A isolates from the 1993-1994 winter epidemic season.

Gene Amplification:

RNA was isolated by treating virus-containing allantoic or culture fluids with proteinase K and sodium dodecyl sulfate and then extracting the product with phenol-chloroform (1:1) and ethanol precipitation as previously described (Bean et al, 1980) . Viral RNA was converted to cDNA with the use of U12 (5ΑGCGAAAGCAGG3') and AMV reverse transcriptase. The sequences of the oligonucleotide primers used in this study for molecular characterization of internal genes (PB2, PB1, PA, NS and M) are available on request.

Amplification proceeded through a total of 35 cycles of denaturation at 95°C (1 min) , annealing at 50°C (1 min) , and primer extension at 74°C (3 min). Amplified DNAs were analyzed by electrophoresis, visualized with ethidium bromide and then purified with either the Magic™ PCR Preps DNA purification system (Promega, Madison, WI) or the Geneclean® kit (BIO 101, La Jolla, CA) according to the manufacturers' instructions.

Nucleotide Sequence Determination:

Nucleotide sequencing was performed dideoxynucleotide chain termination method with the fmol™ DNA sequencing system (Promega) . The reaction products were separated on 6% polyacrylamide-7M urea gels, 0.4 mm thick.

Morphological Observations:

For electron microscopic detection of virus particles on the cell surface and for comparison of cytopathological changes, Vero and MDCk cell monolayers were infected with the Vero-adapted influenza virus strain A/England/1/53 [HG] at 10^"3 PFU/cell multiplicity of infection, trypsin (about 1.0 μg/ml) was included in the medium. Infected and control cell monolayers were fixed at 48 hours postinoculation in cacodylate-buffered 2.5% glutaraldehyde, post-fixed in 1% osmium tetroxide, dehydrated in graded series of alcohols and embedded in Spurr low-viscosity embedding medium (Ladd Research Industries, Burlington, VT) . Ultrathin sections of cells were cut with a diamond knife on a Sorvall MT 6000 ultramicrotome, and the sections were examined in a Philips EM 301 electron microscope operated at 80 kV. Immunohistochemical assay for detection of apoptotic changes in Vero and MDCk virus-infected cells was performed with ApopTag™ In Situ Apoptosis Detection Kit-Fluorescein (ONCOR®) according to the manufacturer's instructions.

RESULTS

Screening of Influenza A Viruses in Vero Cells

Influenza viruses can replicate to high titers in a limited number of mammalian cells, provided that trypsin is present for cleavage of the HA molecule. To determine whether Vero cells are a suitable alternative system for replication of influenza A viruses, a virus repository was screened and a master strain was selected that would replicate sufficiently in the mammalian epithelial-like cell line. MDCK cells, which are widely used to isolate and culture viruses, were included in the study as a reference.

The influenza A virus strains that we examined had been isolated from a wide range of human and avian hosts, and represent 12 of the 14 HA (not H5 and H7) and 9 NA subtypes. Viruses were passaged three times in Vero and MDCk cells, and the virus yield was estimated from HA and infectivity titers. Of the 72 strains investigated, 65 (90.3%) replicated to the level that can be detected by HA titration in Vero cells after the first passage and 37 (51.4%) after the second. By comparison, all strains could replicate in MDCK cells during the first and second passages. Six humans and four avian influenza A viruses were selected as strains with the highest growth potential (Table 4) , among which A/England/1/53 (HlNl) [HG] virus was chosen for further adaptation to Vero cells.

If the A/England/1/53 (HlNl) [HG] virus is to be used as a master strain for generation of high growth reassortants, it is necessary to establish the genotype of this virus. We, therefore, partially sequenced the genes encoding the internal proteins and compared their nucleotide sequence with the prototype influenza strain, A/PR/8/34 (HlNl) . As shown in Table 5, the A/England/1/53

[HG] strain selected for adaptation to growth in Vero cells is a reassortant between the original A/England/1/53 strain and

A/PR/8/34. Six genes of the reassortant encoded internal proteins of A/PR/8/34 and two surface glycoproteins of A/England/1/53.

Infectivity of A/England/1/53 rHG] after Serial Passaging:

To enhance the yield of virus in Vero cells, we performed 20 serial passages of A/England/1/53 [HG] at limiting dilutions, comparing the results with those for the parental strain (Table 6) . Although the infectivity of the parent was lower in Vero cells than in either MDCK or chicken embryos, the progeny showed increased activity in Vero cells by the 10th passage, exceeding that in both reference systems. By the 20th passage, the infectivity of the virus was superior in Vero cells, but the HA titers remained comparable (64-128) . The infectivity titer (TCID₅₀) was 26 times higher than that of the parental strain. By contrast, adaptation of replication in Vero cells resulted in a slight attenuation of the virus in chicken embryos, as indicated by a reproducible decrease in EID₅₀ titer from about 8.2 to about 7.7 log₁₀. The plaques formed by the Vero-adapted A/England/1/53 [HG] influenza strain were not as clear in Vero as in MDCK cells, and the efficiency of the production was 10-fold lower. Plaque-forming capacity in Vero cells increased during serial passages of the virus but not in direct relation to the TCID₅₀ titers. Thus, after 20 serial passages in Vero cells, the yield of infectious virus was high by comparison with that in MDCk cells and embryonated chicken eggs. Viral Protein Yield of Influenza A/England/1/53 rHGI :

Viral protein yield is an important feature of any system used to produce influenza virus vaccines. To establish the amount of virus-specific proteins that can be obtained from Vero cells, we compared the protein yields of A/England/1/53 [HG] (20-passage) virus after replication in Vero and MDCK cells (Table 7) . Determination of the HA protein yield was done using SDS-PAGE separated virus proteins and was quantitated by densitometry. Tests of culture fluids indicated that approximately 6 x 10⁸ of infected cells could produce 4.38 mg of virus protein in Vero and 4.13 mg in MDCk cells. It was also possible to obtain viral proteins from disrupted virus-infected cells of either type; the protein yields were lower than in the supernatant but there was no significant difference between the cell types in the amount of virus protein.

Antigenic Stability of Vero-Adapted Influenza A/England/1/53 fHGI Virus:

Because repeated passage of influenza viruses in mammalian cells could lead to changes in antigenicity, it was thought that it was important to access the influence of Vero cell culture on this property. In HI tests with polyclonal chicken, rabbit and goat antisera with monoclonal antibodies to cross-reacting influenza A (HlNl) viruses, there were no appreciable differences in HA reactivity between the parental strain of A/England/1/53 [HG] and its serially passaged variants (Table 8) . This finding, which extends to antibodies specific to HlNl strains other than A/England/53 [HG] , indicates that serial passage of the virus in Vero cells did not modify its HA antigenic properties.

Primary Isolation of Influenza A Viruses:

Currently, MDCK cells provide the most sensitive host cell system for the primary isolation of influenza viruses. Vero cells have been successfully used to isolate parainfluenza and mumps viruses, but they were judged unsuitable for the isolation of influenza viruses. To reassess this issue, we tested nine clinical specimens collected during the 1993-1994 epidemic season in three culture systems (Vero and MDCK cells and embryonated chicken eggs) . Six influenza A (H3N2) strains were isolated in Vero cells, seven in MDCK cells and only two in embryonated chicken eggs (Table 9) . Two samples failed to yield virus in any host system. During the first passage in Vero cells, CPE observed 48-72 hours after inoculation was the only evidence of virus reproduction. HA activity was detectable on the second passage, and, by the third passage, the positive samples produced both CPEs and HA titers that ranged from 2-32. In all three culture systems, it was necessary to use guinea pig erythrocytes to determine HA titers, chicken erythrocytes failed to be agglutinated as was first described by Burnet and Bull. To examine whether replication of influenza A (H2N2) viruses in Vero cell lines could select antigenic variants, we analyzed viruses that had been passaged three times in this system. The reactivity patterns of the HA with polyclonal antisera to reference A (H3N2) influenza strains and monoclonal anti-HA antibodies did not indicate differences between the strains isolated in Vero cells (results not shown) . These results indicate that Vero cells would provide a useful and nearly as sensitive a culture system as MDCK cells for primary isolation of influenza A (H3N2) viruses.

Ultrastructural Features of Virus-Infected Vero Cells:

To determine (i) if influenza virus infected Vero cells, (ii) if virus is released from the apical surface of Vero cells as in other epithelial cells, and (iii) if Vero cells undergo apoptosis as reported for other epithelial cells, we studied ultrastructural features of this system as compared to MDCK cells, following infection with the A/England/1/53 [HG] influenza virus (20 passages). At the m.o.i. used, both types of cells showed nuclear and cytoplasmic inclusions typical of influenza virus-infected cells, as well as numerous budding virions. As in MDCk cells, virions were released from the apical surface of Vero cells, a feature typical of epithelial cells infected with influenza virus. The budding virions in MDCK and Vero cells appeared filamentous. A fraction of infected cells in both systems showed cytopathological changes indicative of apoptosis. The nuclear changes included fragmentation and condensation of chromatin, margination of chromatin to the nuclear envelope, and blebbing of the nuclear envelope. The cytoplasmic changes consisted of condensation, extensive vacuolation, and blebbing and vesiculation of the plasma membrane to form "apoptotic bodies."

To confirm the apoptotic character of these electron microscopic alterations, the DNA fragmentation in Vero and MDCK cells was assayed. The histochemical assay consisted of addition of digoxigenin-labeled nucleotides to the 3'-OH ends of broken DNA with use of terminal deoxynucleotidyl transferase and detection of the added nucleotides by reactivity with fluorescein-labeled antidigoxigenin antibodies. The infected Vero and MDCK cells showed 20% and 30% positive cells, respectively, by this assay, whereas the uninfected cells were negative. The reported percentage of apoptotic cells in both Vero and MDCk cells may be underestimates because some of the apoptotic cells appeared to detach from the substratum during the extensive washing required by these procedures. The label in certain cells is clearly seen over spherical masses with the nucleus, which may represent condensation of chromatin. These results suggest that a fraction of infected Vero and MDCK cells undergo endonucleolytic cleavage of DNA—a typical feature of apoptosis.

TABLE 1

Accumulation of influenza A/Roruc/_'19 (HIND Yirui in MUCK cells and in Vero (WHO) cells in different plajli vαrc

'Hour. po-L-infcction

"Reciprocal, of HA litor in culture fluid

ND - not done

TAΠ K 2

Effect of Lhc rcβtorαllou uf Lrj-pslπ CQDCEntratWH on X-31 virus grown in Vero (WHO) cell, in 50 cm³ flask-

Input doβo 1.5 x 10' 1.5 λ 10¹ 1.5 x 10' (TCI jo/IliisJ

48^* 72 48 72 48 72

Trypsin added at 0 h p.i.

Trypsin added at 0, 12, 24, 64 128 32 128 128 and 48 lι p.i.

^"Hours post-infection

"KccipnjwLb of HA Utcrs

TPCK - irypsin added to final concentration of 1.0 g/nil

18 Table 4. Highest Yields of Influenza A Viruses in Vero Cells

'Titers were determined for the first passage and measured at 72 hours postinoculation. ND = not done, HG = High Growth.

Table 5. Genotyping of A/England/1/53 (HlNl) Influenza Virus

P = 'genes from A/PR 8/34, E = genes from original A/England/1/53, HG = High Growth.

The origin of HA and NA was determined by HI and NI tests.

Table 6. Growth Characteristics of Influenza A/England/1/53

(HlNl) [HG] Virus After 10 and 20 Passages in Vero Cells

HG = High Growth: P = passages in Vero cells ^"Mcasuicd at 72 hours postinoculation

Table 7. Viral Protein Yield of Influenza A/England/1/53 (HlNl) [HG] in Vero and MDCK Cells

'The concentrated virus was derived from 1 liter of culture fluid or sonicated infected cells

6.1 x 10⁸ Vero and 5.8 x 10⁸ MDCK cells were initially infected.

"Quantitation of HA protein was determined by dcnsitomctry after polyacrvlamidc gel clcctrophorcsis.

TABLE 8 Antigenic Analysis of Influenza A/England/1/53 [HG] Virus Passaged in Vero Cells

r

Table 9. Primary Isolation of Influenza A (H3N2) Viruses in Vero Cells During 1993-1994 Epidemic Season

< = HA tiler less than 1 :2, + = virus isolated and typed as influenza A (H3N2), - = virus not isolated

"HA titration was performed with 0.4% (V/V) guinea pig erythrocytes

'Virus was detected after the first passage in MDCK cells ^zVirus was detected on the second amniotic passage Various modifications of the process of the invention may be made without departing from the spirit thereof and it is to be understood that the invention is intended to be limited only as defined in the appended claims.

Claims

WHAT IS CLAIMED IS:

1. A process of replication of human influenza virus in Vero cell culture comprising infecting the Vero cells with the influenza virus in the presence of a minimum concentration of trypsin of about 0.05 μg/ml in the culture medium throughout the influenza virus growth cycle.

2. The process of Claim 1 wherein the Vero cells are infected with influenza virus at a multiplicity of infection between about 1 × 10'⁵ and about 1 × 10^-6 TCID₅₀ per cell.

3. The process of Claim 1 wherein the multiplicity of infection is between about 1 × 10^-5 and about 1 × 10^-6 TCID 5₀ per cell.

4. The process of Claim 1 wherein the trypsin is regularly added during the replication of the Vero cells to the culture medium to maintain the trypsin concentration greater than 0.05 μg/ml.

5. The process of Claim 1 wherein the concentration of trypsin is maintained between about 0.05 and about 0.5 μg/ml throughout the growth cycle.