key: cord-0010457-ytnv0map authors: Bahnemann, Hans G. title: Inactivation of viral antigens for vaccine preparation with particular reference to the application of binary ethylenimine date: 2002-11-12 journal: Vaccine DOI: 10.1016/0264-410x(90)90083-x sha: 77193a568623a2585c6f13f8ba545a79ee6cfddc doc_id: 10457 cord_uid: ytnv0map Viral antigens for human and veterinary vaccines are still inactivated with formaldehyde. This is not an ideal inactivant and the problems of formaldehyde inactivation of vaccines are discussed. Vaccines inactivated with aziridines are superior in safety and antigenicity. Aziridines inactivate viruses in a first-order reaction and the inactivation rate and endpoint can be determined. The preparation and application of the aziridine compound binary ethylenimine (BEI) and the necessary conditions for and controls of the inactivation process are described and discussed. A computer program has been written for assistance in the use of BEI for controlled inactivation of viral antigens. A perusal of recent and current literature on the preparation of viral vaccines with inactivated antigen and, in particular, of experimental vaccines of this type shows that very often the inactivation is still obtained with formaldehyde and without the necessary controls. In the preparation of an inactivated viral vaccine the inactivation process is a very important step. The innocuity of the vaccine must be assured before the question of potency can be addressed. It seems worthwhile to discuss the inactivation of viral antigens with reference to a few older (and perhaps forgotten) publications as well as some more recent studies. The procedures and process controls which must be applied in order to assure a safely inactivated vaccine will be described. The example of foot-and-mouth disease (FMD) vaccine preparation will be used as this vaccine is by volume the largest viral vaccine produced at present. For many years most of the viral vaccines with inactivated antigen were prepared with formaldehyde as inactivating agent. The work of Sven Gard and his collaborators t'2 with poliovirus during the period of 1956 to 1958 demonstrated that the inactivation of this virus with formaldehyde was not a linear or first-order reaction. Similar results were obtained for the formaldehyde inactivation of FMD virus by Wesslrn and Dinter in 19573 and by Graves 4 in 1963. A recent publication stating the linearity of formaldehyde inactivation of FMD virus 5 comes to this erroneous conclusion because the infectivity titration for the inactivation slope was based on final readings of plaque forming units (p.f.u.) at 2 days. This is far too The extended incubation period for formaldehyde treated virus also means that an innocuity test in animals is inappropriate for detecting small amounts of residual infectious virus. When this virus begins to reiglicate with a delay of between several days and 2 weeks, the animal is already beginning to produce antibodies. The new virus will then be neutralized by antibody and the animal has an abortive or subclinical infection. This subclinical infection can be detected by testing the animal for the -virus infection associated antigen (VIAA), the viral RNA polymerase 8. In 1975 Alonso et al. 9 reported a study of cattle exposed to FMD and found that at 21 days postvaccination, of 18 animals vaccinated with formaldehyde inactivated vaccine five animals were positive for VIAA. Of 16 animals vaccinated with N-acetylethylenimine (AEI) inactivated vaccine no animal was positive for VIAA. Pinto and Garland 1° later found VIAA positive cases also in animals revaccinated with AEI inactivated vaccine, but emphasized that the response in these animals was much weaker and only transient. Alonso et al. ~ were able to confirm this transient and weak response and also found, with binary ethylenimine (BEI) inactivated AI(OH)3 and oil adjuvant vaccines, a response to VIAA only after revaccination. The weak response to VIAA is caused by the presence of this antigen in the inactivated virus suspension used for vaccine preparation. However VIAA antibodies induced by safely inactivated vaccines are only detected after revaccination and the presence of such antibodies after primovaccination still indicates recent virus replication in the animal. Lucam et al. analysed in 1958 the FMD vaccination campaigns in France ~ 2 and suspected residual infectivity in some formaldehyde inactivated FMD vaccines, which had passed official innocuity tests. Recently Beck and Strohmaier ~ 3 studied viruses from field outbreaks of FMD in Europe by determination of their nucleotide sequences. They found that most of these isolates were related to virus strains in (formaldehyde inactivated) vaccines. This led Strohmaier 14 to the opinion that most of the FMD outbreaks in Europe in the last 20 years were 'homemade', i.e. were caused by vaccination. He made the recommendation that inactivation of FMD vaccine antigens should be changed from formaldehyde to first-order inactivants or that vaccination should be stopped altogether. The FMD vaccine production regulations in several South American countries have for several years now permitted only the use of first-order inactivants. At least one European country has recently also adopted this position. Most of the FMD vaccine production laboratories in these countries apply the aziridine compound EI in the form of BEI. One group of laboratories uses diluted EI. The first report of a (bacterial) virus inactivation by ethylenimine, the basic aziridine substance, was published in 1955 by Raettig and Uecker Is. Hurst in 1957 x6 was of the opinion that vaccines prepared with AEI as inactivant were antigenically superior to vaccines inactivated with formaldehyde and would guarantee inactivation of the virus. The antigenic superiority was later confirmed also for vaccines inactivated with BE117-19. ICI patented the use of AEI for inactivation of microorganisms in 19592°. The first report on the inactivation of FMD virus by AEI was published by Brown and Crick also in 195921 . This compound was subsequently used by a leading FMD vaccine production laboratory for many years in the preparation of inactivated antigens. However it did not come into general use because of its patent protection. In 1961 Uecker 22 reported the linearity of inactivation of bacterial viruses by ethylenimine derivatives and Graves and Arlinghaus described in 1967 the linearity of AEI inactivation of foot-and-mouth disease virus 23. At ambient temperatures AEI is not stable and it therefore has to be kept at 4°C or preferably at -20°C. Fellowes remarked in 196524, that AEI has a low boiling point and very little is left in a biologic preparation at reaction temperatures of 20°C or above. This observation probably made him use an inactivation temperature of 23°C instead of the usual 37°C. The problem of the stability of AEI was perhaps the reason for introducing the double dosing regimen for inactivation at 37°C, i.e. the application of two doses for 24 h each as described by Pay et al. 25 . This procedure is still being used with BEI by the same laboratory 26 although EI is much more stable 27. An extended incubation of the antigen at 37°C damages the antigen, as was shown in comparative inactivations of FMD virus at 26°C and at 37°C 2a. This damage is not due to the inactivant but is probably caused by the action of proteolytic enzymes present in the virus suspension. Other laboratories continued to work on FMD virus inactivation by aziridines and in the early 1970s reports on ethylethylenimine, EEI zg, ethylenimine, E127 and binary ethylenimine, BEI 3° were published. Both EEI as well as E1 are difficult to obtain in quantity. For this reason, as well as the ease of preparation and handling, BEI is now the preferred inactivating agent for FMD and other veterinary vaccines. The viruses which have been reported as inactivated with BEI are given in Table I . They belong to a variety of families of viruses with either RNA or DNA, which makes it very likely that most known viruses would be inactivated by an aziridine. Inactivation in vaccine preparation transforms an infectious antigen into a non-infectious one. This transformation step should therefore be done in a well identified intermediate area between the virus-containing and the virus-free area. Access to the intermediate area should be limited, and only be possible from the virus-containing side. The antigen must be held in the intermediate area until completion of the necessary control tests (inactivation endpoint and innocuity). The facilities for holding of the inactivated antigen in this area, cold room or cooled storage tanks, should be able to accommodate a volume of at least 2-3 weeks production of virus in order to allow termination and if necessary a repeat of the control tests. Only after confirmation of the innocuity can the antigen be transferred to the virus-free area and be used for vaccine preparation. The inactivation process should be done under slow agitation in two different vessels, with perhaps one quarter or a third of the time in the first vessel, and transfer of the virus suspension under inactivation in a closed system to the second vessel for the remainder of the time. This procedure is indicated in order to avoid pockets of the virus suspension into which the inactivant did not enter or reinfection from non-inactivated virus on the tank wall above the liquid level at the end of the inactivation period and after hydrolization of the inactivant. The virus suspension should be checked to determine that it is at the desired temperature and is at a pH of ~7.4 before the inactivant is added. It is also advisable to control the osmolarity of the virus suspension, which for cell culture produced FMD virus is 320-340 mOsM. The pH and osmolarity of the virus suspension affect the velocity of inactivation. Figures 1 and 2 show the results of a study some years ago of the effect of pH and osmolarity on the inactivation rate of EI. The experimental conditions were the same as given previously 27. It can be seen from Figure 1 that an increasing alkalinity reduces the inactivation rate. This is in contrast to formalin inactivation, where the velocity of inactivation was found to increase with increasing alkalinity 4°. Figure 2 shows that increasing osmolarity also slows down the inactivation rate. The pH and osmolarity effects on the inactivation rate are perhaps due to conformational changes in the viral capsid proteins which affect the permeability for the inactivant. The inactivant was called binary ethylenimine or BEI because it is prepared from two substances, 2-bromoethylamine HBr (BEA) and NaOH, and also to distinguish this preparation from pure EI. BEA converts in an alkaline solution to 'binary' ethylenimine. The active substance is the ethylenimine ring as in all aziridine compounds. The BEA solution is 0.1M or 20.5g1-1 of a 0.175N NaOH solution. The conversion to BEI is completed in 30-60min at 37°C and is accompanied by a pH drop from = 12.5 to =8.5. The formation of BEI is indirectly controlled by visualizing this change with the pH indicator fl-naphthol violet (BNV). From a 1% aqueous stock solution of BNV, 0.5 ml is added per litre of NaOH Inactivation of viral antigens: H. G. Bahnemann solution. The colour of the BEA solution changes at 37°C in = 15 min from violet to orange upon formation of BEI. The BEI preparation should not be used for inactivation if the colour has not changed to orange. The 0.1 ra BEI preparation contains only 0.5% EI and therefore is much easier to manage than concentrated EI or AEI. However it should still be handled with care and prepared in a closed vessel which allows transfer of the BEI in a closed system to the inactivation vessel. Higher concentrations of BEI, such as 1 or 2M, can be prepared. But the preparation and handling of such BEI solutions requires much more attention and precautions as the El concentration is much higher. It is recommended to use the 0. Na-thiosulphate hydrolyses BEI. The BEI inactivation is stopped by the addition ofa 1 ta sterile Na-thiosulphate solution at 10% of the volume of the BEI solution used. It should also be used for hydrolysationt of BEI in any spills or for cleaning of the vessel in which the BEI solution was prepared. The inactivation process must be accompanied by appropriate in-process controls. This begins with the BEI preparation. As described above, the formation of the ethylenimine ring is monitored in a simplified manner. If necessary, a colorimetric determination of the produced BEI can be performed using the method described by Epstein et al. 4x , and has been used by Czelleng et al. 42 for BEI and Preaud et al. 43 for EI. It should not be necessary to do this colorimetric determination with each BEI preparation. However it is advisable to do so at intervals, as it is also a control for the BEA salt. No information is available on the shelf life of BEA. Some inactivation failures are suspected to have been caused by old BEA. The inactivation of viruses with BEI is a first order or linear reaction. For each inactivation, the rate has to be determined. This is done by taking samples for infectivity titration during the early part of the inactivation process, at 0, 1, 2, 3 and 4h for example. The inactivation rate is obtained from the infectivity titres by calculation of the regression coefficient and is used to calculate the inactivation endpoint. The endpoint is the most important parameter to be determined and is a function of the inactivation rate, the reaction time and the volume of the virus suspension. The minimum endpoint is defined as being one log(10) lower than the titre which gives one infectious unit in the total volume under inactivation. For example: One infectious unit in a volume of one litre has a titre of Vaccine, Vol. 8, August 1990 301 Inactivation of viral antigens: H. G. Bahnemann log 10 -3 (for the usually expressed titre in ml). The minimum endpoint for this volume therefore is log 10 -4. For a successful inactivation the calculated endpoint has to be lower than the minimum endpoint. The difference between the calculated and the minimum endpoint, or the DIM (Difference of Inactivation endpoint to Minimum) value has to be positive. It is important to determine the DIM value for each inactivation process for an assessment of a successful inactivation. After termination of the inactivation each virus suspension has to be tested for innocuity on cell cultures. This can be done by inoculation of at least two roller bottles or similar cell culture vessels and two subsequent blind passages at 48 h intervals. The cell cultures for this test have to be prepared in the virus-free area and the test should be done in the intermediate area. Only after determination of the inactivation endpoint and completion of the innocuity test can the virus suspension be considered to be properly inactivated. The major application of BEI will be in the preparation of inactivated antigen for vaccines. But there are other areas in which BEI can be used. Since BEI does not react with proteins it can be used for inactivation of adventitious viruses in biological preparations from animal or human tissues or fluids. The BEI treatment of bovine serum used for cell cultures has been reported in 197634 . The usefulness of this method was later confirmed in 1984 by Heuschele 36, who applied it over a 6-year period for the inactivation of adventitious bovine viral diarrhoea virus in calf serum used in primary or secondary animal cell cultures. BEI can also be used for inactivation of viruses in enzyme preparations of animal origin. A commercial trypsin preparation was treated with BEI without any loss of activity (unpublished results). It is very likely that other biologicals, like Factor VIII, could be treated with BEI for inactivation of adventitious viruses. Pure aziridines are highly toxic and have to be handled with special precautions and extreme care. This high toxicity is the reason for the 0.1 M preparation of BEI. At this molarity the BEI preparation contains only 0.5% El and the vapour pressure at this concentration is low enough that for temperatures under 50°C no EI will escape into the atmosphere (the boiling point for EI is 57°C). On the basis of experiments in laboratory animals, aziridines are considered to be carcinogenic substances. According to Dermer and Ham 45, however, no cases of human cancer caused by E1 have ever been reported. Fellowes z4 cites Hurst who states that an injection of 0.5 mg of AEI into rats did not produce any tumours during an observation period of 515 days. The total annual production of FMD vaccine worldwide is probably between 700 and 800 million doses, of which nearly 500 million doses are produced in South America alone. Furthermore between 70% and 80% of all FMD vaccines are inactivated with BEI, which means that about 500 million doses of FMD vaccine are inactivated with BEI and applied in cattle annually. The majority of these cattle are revaccinated many times over the years. No increase in the incidence of cancer in cattle after vaccination with BEI inactivated vaccines has ever been reported from any country. Neither is there any reason to expect any increase, as the residual BEI after inactivation is hydrolized with Na-thiosulphate. A computer program has been written to assist with the calculations needed for a controlled inactivation of viruses with BEI or any other first-order inactivant. The screen menu for the program is given in Figure 3 . Inactivation of poliomyelitis virus by formaldehyde Inactivation of poliovirus by formaldehyde The inactivation of foot-and-mouth disease virus by formalin Formaldehyde inactivation of foot-and-mouth disease virus as applied to vaccine preparation Formaldehyde inactivation of foot-and-mouth disease virus. Conditions for the preparation of safe vaccine partially inactivated poliomyelitis virus: initiation of infection in tissue culture Inactivation of poliomyelitis virus by formaldehyde. Incubation time in tissue culture of formalin treated virus Inhibition of cell-free foot-and-mouth disease virus ribonucleic acid synthesis by antibody The use of virus-infection-associated antigen (VIA) in the detection of cattle exposed to foot-and-mouth disease The induction of antibodies against VIAA in cattle vaccinated and revaccinated with inactivated FMD vaccine La vaccination antiaphteuse en France Subtyping of European foot-and-mouth disease strains by nucleotide sequence determination Die Maul-und Klauenseuche IlL Erfahrungen und Folgerungen aus 20 Jahren Bek~mpfung der Maul-und Klauenseuche Inaktivierung yon Bakteriophagen durch Athyleniminderivate Approaches to the chemotherapy of virus diseases Inactivation of O~ FMD virus by binary ethylenimine (BEI) Binary ethylenimine and foot-and-mouth disease vaccine production in Thailand A preliminary vaccine potency trial of a Newcastle disease virus inactivated with binary ethylenimine Application of agar-