A perfusion culture system for virus vaccine manufacture in diploid cell cultures

(1)

A PERFUSION CULTURE

SYSTEM FOR VIRUS VACCINE

MANUFACTURE

IN DIPLOID CELL CULTIIIUS’

G. F. Mann’ and Julio de Mucha Ma&s9

A perfusion culture system for large-scale virus vaccine production is described and comsared with the stationary bottle @ocedure. The perfusion system gave much greater cell production while still maintaining satisfactory titer. This

find-

ing could have important application in the large-scale manufacture of vaccines.

Iutroduction

A new perfusion

culture system designed to meet the requirements of large-scale virus vaccine manufacture has previously been reported (I).

The culture vessels in this system consist of glass plates separated by Teflon strips and sealed within an open-sided box fitted with inlet and outlet ports. Vessels of any desired surface growth area in the range of <O.Ol to

>lO mz

can be constructed, giving uniform operational characteristics. When a

surface

area of more than 10 m2 is required, multiples of 5- or 10-m” vessels can be operated in parallel.

The vessels are connected through a pump, to separate the medium reservoirs, and the entire assembly is operated as a closed system, thus minimizing the risk of contamination.

It was expected that this system would permit greater densities of cell production than those obtained with the normal

‘Also appearing in Bol Of Sunit Panam 80(5):388- 396, 1976.

“Scientist, Area II Office, Pan American Health Organization, Mexico City, Mexico.

3Director, Poliomyelitis Vaccine Production, Na- tional Institute of Virology, Ministry of Health and Welfare, Mexico City, Mexico.

monolayer, since Kruse and Mediema (2) and others had shown that cell density is directly related to the volume of medium used. Thus, considering the advantages that this would offer for the large-scale manufacture of vaccines, the Pan American Health Organization and the Mexican Ministry of Health and Welfare undertook a project to develop the system at the National Institute of Virology in Mexico City. The results of this work, carried out since

1973,

are the

subject

of the present report.

Experimental Work

Perfusion culture vessels with a surface growth area of 200 cmz were used to study the many parameters of cell growth and virus production. The production of poliomyelitis virus (Lsc 2ab) in human diploid cells was selected as the test system. The main findings from the experimental work were then tested on production scale using vessels with a surface area of 2.5 m2.

Assessment of Culture Vessel Design

No significant difficulty was encountered in the manufacture, cleaning, assembly, or sterilization of culture vessels. All compo-

(2)

206 PAHO BULLETIN . vol. X, no. 3, 1976

nents of the vessels and operating systems were shown to be nontoxic by stringent tests using human diploid cell cultures. The assembled, sterilized vessels were shown to retain 5 lb/in2 air pressure indefinitely and to miintain {he sterility of recirculated bacteriologic and tissue-culture media. It was demonstrated that microscopic exami- nation of cells on the outer plates on both sides of the vessel can be satisfactorily carried out. With long-working distance objectives, several cell layers on each side of the vessel can be examined. Cells grown on both sides of the glass plates in multiplate vessels were shown to be evenly distributed over the plates and from plate to plate by projection counts of nuclei stained in situ.

Culture Vessel Ofieration

Simple operating systems were devised using standard experimental and industrial equipment according to the flow pattern illustrated in Figure 1. For both the cell growth and the poliovirus production studies, the perfused medium was recycled through medium reservoirs. The operational methods followed are outlined in the draft vaccine production procedure (Ap- pendix).

Growth of Di#loid Cells

The preliminary studies were carried out using the WI-38 strain of human diploid cells, received from L. Hayflick. All studies reported here, however, were carried out using the MRC-5 strain, obtained from P. Jacobs, Medical Research Council, London.

Eagle’s basal medium supplemented with fetal calf serum was used for cell growth in all experiments. Cell suspension derived from bottle cultures at the 27th or 28th population doubling was used as inoculum for both perfused cultures and stationary bottle controls. Cell counts were carried out by staining fixed cells in situ and counting the nuclei by projection microscopy. They

Figure 1. Basic syntem for perfusion culture vessel operation in ccllgmwth ad virun

pmduction atudks.

RC

n

r

f

1

-cY--Ll

R r Rescmir P =plmrp cv = cultureveMe1

were done routinely at 18 hours (inoculum count) and at the end of the growth period; additional counts were done when cell growth was under study (Figure 2).

Cells inoculated into the perfusion vessels showed complete attachment at 30 minutes and by one hour had begun to spread out on the glass surface.

The lag phase of cells in the perfused cultures was demonstrated to be of the same order as that for equivalent cell inocula in bottles.

The generation time of the cells in the perfused cultures (26 to 28 hours) was constant, even at high cell density, and was similar to that observed in parallel low-density bottle cultures.

(3)

Mann and Much Macias l PERFUSION CULTURE SYSTEM FOR VACCINES 207

Figure 2. Growth of diploid cells (MRC-5) in perfused vessels and stationary bottles (nuclear counts by

projection microscopy after fixing and staining cells in situ).

No. of Fold increase generations in normal

observed saturation 6.0 -j density 5.8 -

5.6 - 5.4 - mg 5.2 -

3 aI 0 5.0 -

2 s” 4.8 -

4.6 -

4.4 - Perfused vessel Perfused vessel O---O Stationary bottles O---O Stationary bottles 4.2

1 4.0 1

1 2 3 4 5 6 1 8 Incubation time (days at 36oC)

‘Perfused cultures, 6 ml medium/cm2. 2Bottle cultures, 0.5 ml medium/cm2. 3Count imprecise due to high cell density.

physical dimensions of the growth vessel, but rather by the reservoir capacity, which is virtually unlimited. With large volumes of medium, cell growth in the perfused cultures continued at a logarithmic rate and reached or exceeded the level at which multilayered cultures, containing about 1 x 106 cells per cmz, are observed (Figure 2).

Multilayered cultures are considered un- suitable for human vaccine production, since the population doubling level may not be the same in all horizontal layers.

In the culture preparation for the virus production studies, the cell density was

1.5 - 2.0 x 105 cells per cm’2, equivalent to twice the saturation density observed in bottle cultures. The volume of medium required, the inoculum, and the incubation time were defined. Such cultures are dense but not multilayered, since, as in bottle cultures, the overlapping cells are observed only at the cross-striations in the cell sheet. Typical diploid-cell orientation was observed in all cultures.

Cells grown in perfusion by the above procedure then subcultured back into bottles showed no karyologic changes.

Profiagation of Poliomyelitis Virus (IStrain

Lsc Zab)

The propagation of virus in perfused MRC-5 cells was compared with parallel stationary bottle cultures handled by tradi- tional methods. Medium 199 was used as the seed diluent and virus production medium in all the experiments. A virus preparation derived from patas monkey kidney cell cultures at the SO + 4 level was used as seed virus at final virus concentrations of 10 *.* or 10 ~3 TCID so/ml. Virus titrations were carried out using a Vero cell/microtiter system having similar sensitivity to the HEp 2 C system recommended by A. Sabin. The methods followed for the perfused cultures were generally those set forth in the draft vaccine production procedure (Appendix). All infected cultures were incubated at

34.5%.

(4)

208 PAHO BULLETIN l vol. X, no. 3, 1976

Table 1. Production of poliomyelitis virus (bc 2ab) in perfused and stationary bottle cultures of MRC-5 cells-Some typical experimental results.

Experi- Culture ment vessel no. _type

Perfused

6

Falcon bottles

Perfused

8

Falcon bottles

Cell cultures _I Harvest Virus yield (loglo TCID~O)~

Surface growth

area (cm2)

200

75

200

Cell density

(cells /ml

/cm21

2 100 8.04 3 100 7.48 1.64~ 10’ 4 100 6.60

2-4 300 (7.69)

2 20 6.98

3 20 7.35

0.79x10” 4 20 6.98

2-4 60 (7.14)

I 2 100 8.25 3 100 7.50 100 7.38

300 7.89

-i

I

t

1

/vessel

I I

/cell /cm2

10.04 2.57 7.74 9.48 2.01 7.18 8.60 / 1.13 6.30

*

8.28 1.55 6.40 8.65 1.92 6.77 8.28 1.55 6.40

8.92 ’ 2.19 7.04

aAll titrations for each experiment in parallel against a standard. Pool titers (days 2-4) were by calcula- tion only in experiment 6 but were confirmed by titration in experiment 8.

Table 2. Production of poliomyelitis virus (Lsc 2ab) in perfused and stationary bottle cultures of MRCd cells-Summary of results.

Culture vessel

Perfused

Falcon bottles

- Virus yield (log, o TCID~O)~

.__- Operating No.

of re- per cm2 of surface per cell procedure

sults

tados Mean Range Mean Mean Range

Mean

difference difference

Recirculating

medium, 12 8.02 7.75-8.35 +0.73 - - -

dally

harvest ₍₈₎ - - - 2.83 2.62-3.10 +0.56

Stationary

medium, 7 7.29 7.04-7.61 -0.73 - - -

dally

harvest (5) - , - , - , 2.27 , 2.19-2.32, -0.56

(5)

Mann and Mucha Macias l PERFUSION CULTURE SYSTEM FOR VACCINES 209

Table 3. Potential of production-scale perfusion vessels for vaccine production (kc 2ab/MRC cells).

Perfused Potential yield from production vessels

vessel Experimental yield (doses in millions)

surface 5 (log1 o TCID, &m2) ~__-

area (cm2)

1O5*y5 TCID50/dosea 105*’ TCID5,,/doseb

Mean Range Mean Range i Mean Range

12,500 1.48 0.79-3.16 2.63 1.41-5.62

25,000 I

8.02 7-75-8.35 2.96 1.5X-6.32 5.26 2.82-l 1.24

50,000 I 5.92 3.16-12.64 10.52 5.64-22.48

100,000 11.84 6.32-25.28 21.04 11.28-44.96

aAssessed as trivalent vaccine containing 5 x 105, 3 x lb5 and 1 x IO5 TCID

bAssessed as monovalent vaccine of types 1, 2, and 3, respectively. containing 5 x 105 TCID50 of type 1 only. 50

experiments, the last harvest being taken within 96 hours.

Some of the typical results are presented in Table 1, and the summarized findings from several of the experiments are given in Table 2. It will be noted that the mean yield from the perfused vessels in 10 e.83 TCID 50 per cell, as compared with a value of 10 2.27 TCIDso per cell in the parallel control groups. This difference of 0.56 loglo may represent an inherent advantage of the perfusion system, or it may reflect intrinsic differences between the two methods.

The yield per square centimeter reflects not only the difference in yield per cell but also the increased cell density. In this case, the perfusion system shows a 0.73 loglo advantage over the bottle system, for which the mean values are lOa.oz and 10 7.&g TCID ~0 /cm”, respectively.

In this work emphasis has been placed on efficiency of virus production (i.e. yield per cell), rather than on high titers. Never- theless, the mean yield of 108.0” TCIDso,/

cmz represents a titer of 1O7*s4 TCID,o/

ml,

which is satisfactory for vaccine manufacture.

The number of vaccine doses per production batch will, of course, depend on the surface growth area used, the yield, and the vaccine formulation. The numbers of doses have been calculated for two vaccine formulations over a range of surface growth area per batch on the basis of the experimental yields shown in Table 2. This information is presented in Table 3.

Production-ScaEe Trials

In the first technically successful production-scale trial, which used vessels with a surface growth area of 2.5 cmz, the yield obtained was within the experimental range (107~s8 TcIDs~,/cm~). This result is considered extremely important, since it confirms the unified design and operational concepts of the system over a 125-fold difference in scale.

Complete testing of material from the production trials is now in progress.

SUMMARY

Development of a new perfusion culture The growth characteristics of the diploid cells

system for the production of attenuated polio- (MRC-5) were found to be normal in the

(6)

210 PAHO BULLETIN l vol. X, no. 3, I976 density of stationary bottle cultures were field was confirmed to be within the experimen- established, tal range when small production-scale vessels

The yield of virus (Lsc Zab) per cell and per were used. unit of surface growth area were observed to be

significantly higher in the perfused system than Subject to satisfactory results in the testing of in parallel stationary bottle culture controls-by the virus product, this system could be highly factors of three- and five-fold, respectively. The economic in large-scale vaccine manufacture.

REFERFiNCES

(I) Kruse, P. F., and E. Mediema. Production (2) Mann, G. F. A new system for cell and characterization of multiple-layered popu- cultivation in perfused layered cultures. I. lations of animal cells. J Cell Biol 27:273-279, Design and operating principles. Bull PAHO

1965. 6(S):%-36, 1972.

APPENDIX

Draft Procedure for Poliovirus Vaccine Manufacture in Perfused Cultures of MRC-5 Cells

Stage Operations in clean tissue-culture area Control samples Notes

Prepare cell inoculum at P25 by serial passage of seed cells in bottle cultures using conven- tional methods

Inoculate cells at P25 into a perfusion culture vessel at a I:4 split ratio; adsorb cells on both faces of internal glass plates; incubate at 36.50C for 6 days, recirculating a defined volume of growth medium to give P28

Sterility test at Inoculum density each passage 1.88-2.5 x 104/cm2

Saturation density 0.75-1.0 x lOs/cm2 No. of generations = 2

Inoculum cell Inoculum density counts 1.88-2.5 x 104/cm2

Saturation density 1.5-2 x lOs/cm2 No. of generations = 3

3 Remove growth medium; resuspend cells by Sterility tests trypsin/versene treatment; prepare a single Spent medium pool of cells in growth medium (P28) tests

Viable cell tests Nonviable cell

tests

Karyologic mo- nitoring Control cultures

in bottles or perfusion

(7)

Mann and Mucha Macias l PERFUSION CULTURE SYSTEM FOR VACCINES 211

Stage Operations in clean tissue-culture area Control samples Notes

5 Remove growth medium; wash cultures by re- Spent medium circulating serum-free medium to give a calcu- tests lated dilution factor for bovine serum of 1: 1 Sterility test million in the first harvest

6 Remove wash medium and fill vessel with virus Seed virus titra- In experimental seed diluted in 199 medium; adsorb 1 hour at tion studies the seed virus 34.5oC, redistributing at intervals to bring all Sterility tests was used at 101.7 seed virus into contact with cells; commence TCID 5 o/ceil and the recirculation of 199 medium to give a total harvest contained a

volume of 0.5 ml/cm2 mean of 102.83

TCID, o/cell, giving an increase in virus by a factor of 103.13 or x 1350

7 Harvest virus by replacing reservoir of 199 Virus titrations Harvest at 24 hours medium at 24 hour intervals starting at 22 or Sterility tests is known to contain 46 hours and continuing through to 94 hours; insignificant amounts remove harvest samples and store at 4oC up of virus

to 4 days or -45oC indefinitely

8 Prepare a pool of the three harvests to give combined harvest which also constitutes bulk suspension; sample

9 Immediately after sampling at stage 8, filter the bulk suspension; sample; distribute into bottles and store at -45oC pending test results

All virus harvest tests

All bulk suspension tests

All filtered bulk Consistency to be suspension established on tests batches presented