Concentration and Purification of Influenza Virus for the Preparation of Vaccines

Abstract

Discussion and Summary In the present report, apparatus and procedures are described for the culture, harvest and concentration of the influenza virus, for the purpose of preparing vaccines. With this technic chorio-allantoic fluid infected with influenza viruses A (PR8 strain) and B (Lee strain) can be collected and the virus concentrated to a high degree with but slight contamination with bacteria. A team of 4 workers can candle, grind, inoculate and handle about 400 eggs an hour. To harvest the chorio-allantoic fluid at the same rate, a team of 5 is necessary. The maintenance of sterility or low degree of contamination of the final virus concentrate is dependent on a number of factors; the initial choice of vigorous embryos by careful preliminary candling, inoculation of the eggs under ultraviolet light away from dust, careful candling and choice of embryos after incubation, the harvest of the fluid under ultraviolet light, and, finally, the operation of the Sharples centrifuge in an atmosphere of low content of viable bacteria. The healthy chick embryo is relatively resistant to infection with bacteria. This resistance is lowered but little, apparently, by the presence of influenza viruses A (PR8 strain) and B (Lee strain). In contrast, infection of the embryo with the swine influenza virus appears to increase greatly the susceptibility of the embryo to infection with bacteria, with the result that thus far chorio-allantoic fluid containing the swine virus has frequently been heavily contaminated with bacteria. The virus concentrates obtained in a single sedimentation with the Sharples centrifuge are of a high degree of purity, as shown by electron micrographs and sedimentation diagrams in the analytical ultracentrifuge (23). Little material extraneous to the virus is present. With the handling of large quantities of the concentrates, however, thorough dispersion of the sedimented virus is time-consuming, and there is evident a greater degree of aggregation than when smaller amounts are handled. This would not be expected to lessen the antigenicity of the virus in vaccines. The chief factor limiting the production of concentrated influenza viruses is the capacity of the Sharples centrifuge. With the standard 160-ml separator bowl, the rate of flow for practical purposes is limited to about 1 l per hour, or about 20 l per day, allowing for removal of the concentrate and setting the machine up again. On this basis a single machine could handle the output of about 2,000 to 2,500 eggs per day. The efficiency of the machine can be increased approximately 2-fold by the use of a modified batch bowl of 50-ml capacity. By this means the yield of 4,000 to 5,000 eggs can be handled by a single machine in a day. It would appear unlikely that a Sharples machine of a type much larger than the laboratory model could be used to increase the output. The larger machines now available have a rating of about 15,000 g at full speed. A field of this low magnitude would necessitate a very low rate of flow as indicated by the sedimentation experiments described above. The use of the centrifugal process provides a number of advantages in the concentration of virus. The procedures for the maintenance of asepsis can be controlled throughout. Once the chorio-allantoic fluid has been obtained, a single operator can attend several machines and carry out the entire concentration process. The final virus product is practically pure virus and can be standardized simply by chemical analysis for nitrogen or carbohydrate. The yield of virus can be controlled and kept at a level of about 90 per cent or more of the virus initially present in the chorio-allantoic fluid. The degree of concentration of the virus would be dependent on the quantity of fluid run through the machine in a single sedimentation, since the final volume is fixed at 50 ml. In a run of 20 hours, 40 l of fluid passed through the machine would result in a concentration of 800 times by volume. The yield of virus in the instance of influenza virus A (PR8 strain) would be approximately 2.0 to 4.0 gm per 40 l, or 50 to 100 mg of virus per liter, and 2.0 gm per 40 l, or 50 mg per liter in the instance of influenza virus B (Lee strain). The number of doses of vaccine per machine day would be determined by the amount of virus required for each dose.