157

Elsevier Science Publishers

B.V. / All rights reserved / 0166-0934/92/$05.00

VIRMET 01372

Assessment of inactivation of hepatitis A vaccine by compound PCR N. Fineschi”, “Molecular

V. Pellegrinia,

Virology, I.R.I.S.

A.J. Zuckermanb

Sri.. Siena (Italy) and ‘The London (UK)

(Accepted

and F. Cavalieri”

Royal Free Hospital School of Medicine,

30 March 1992)

Summary Assuring the complete inactivation of hepatitis A virus (HAV) vaccine commonly requires prolonged tissue culture amplification, followed by detection of virus antigen in cell lysates. A reliable, but faster, alternative procedure is highly desirable, since it will permit the prescreening of experimental batches of killed HAV, prior to tissue-culture amplification. We established experimental conditions for simultaneous, polymerase chain reaction (PCR)-based amplification of viral and cellular mRNA sequences from infected cell RNA (compound PCR). Under these conditions, the presence of virus-specific amplified sequences, as detected by Southern blot, allows the identification of incompletely inactivated vaccine batches with a threshold practically identical to that of the more time-consuming subculture and ELISA. Compound PCR is, by its nature flexible enough for adaption to different requirements and it should prove useful for rapid prescreening of vaccine batches and pilot studies for improvement of inactivation protocols. HAV; Vaccine; Inactivation;

ELISA; PCR; VP1

Introduction The assessment of the level of inactivation attained is a crucial step in the production of safe, inactivated viral vaccines. The establishment of inactivation thresholds is particularly intricate and tedious in the case of viruses, such as most strains of hepatitis A virus (HAV), which grow slowly in cell cultures Correspondence to:

F. Cavalieri, I.R.I.S. srl, Via Fiorentina 1, 53100 Siena, Italy.

158

without a cytopathic effect. So far, reliable assessment of vaccine inactivation could be achieved only after prolonged attempted amplification

Materials and Methods

HAV strain LSH/S, originally isolated at the London School of Hygiene and Tropical Medicine (Garelick -et al., 1988), was propagated on MRC-5 human diploid cells. The virus grows and is mostly cell-associated, with an incubation time of three weeks and results in persistent infection without cytopathic effects. Purified HAV preparations, inactivated with formalin, where stated were used throughout this study. Conventional control of inactivation involved primary amplification of the vaccine inoculum in MRC-5 cells for 3 weeks, followed by inoculation of infected cell lysates onto fresh cell cultures and subculture for three further weeks, prior to testing by ELISA for viral antigens. The detection of viral antigens in freeze/thaw cell lysates was carried out on mouse anti-HAV Mabcoated plates, with human anti-HAV IgG and peroxidase conjugated goat antiHuman IgG (Garelick et al., 1988; Fineschi et al., 1991).

RNA was extracted from infected and control MRC-5 cell monolayers by the guanidinium thiocyanate method as described by Sambrook et al. (1989). First strand cDNA synthesis was carried out in the presence of random primers with the INVITROGEN cDNA kit, using the conditions recommended by the manufacturer. Five micrograms of total RNA were used in each cDNA

159

synthesis experiment. PCR and Southern

blot analysis

One fifth of each cDNA reaction mixture was amplified by PCR (GENEAMP, Perkins & Elmer) using the conditions recommended by the manufacturer. Amplifying oligonucleotides employed (listed 5’ to 3’) were: (a) HAV, CAAGTTGGAGATGATTCTGGAGGT (Sense, nt 2205-2228) and CTCAAATCTTTTATCTTCCTCTGA (Antisense, nt 3107-3074). These oligonucleotides encompass a 902 nt fragment, spanning most of the VP1 gene. (b) Human cytoplasmic gamma actin: ATGGAAGAAGAGATCACCATACTC (Sense, nt l-24 of the coding region) and TGTGACTCCGTCTCCAGAATCCAT (Antisense, nt 540417 of the coding region). Amplification was carried out for 30 cycles of 1 min at 94°C 4 min at 55°C and 1 min at 72°C. Aliquots of the amplified material were run on ethidium bromide agarose gels and the relative amounts of amplified actin fragments were estimated. As a rule, amplified HAV VP1 was not detectable by this method. Sample volumes containing equal amounts (estimated at about 50 ng) of amplified gamma actin fragment were electrophoresed in agarose gels, Southern blotted against radiolabelled VP1 cDNA (Fineschi et al., 199 1) and autoradiographed as described by Sambrook et al. (1989), under the following conditions: the probe was used at 500000 cpm/ml and allowed to hybridize overnight at 65°C in 5 x SSC; the filter was washed for 3 h in 1 x SSC at 65°C and exposed at room temperature for 7 h to DuPont Cronex film.

Results

ELISA-based inactivation tests of HAV vaccine candidate preparation have been carried out routinely following tissue culture amplification times ranging from as little as 10 days (Chaudary et al., 1988) to as much as ten weeks (Andre et al., 1990; Flehming et al., 1988; Lewis et al., 1990; Binn et al., 1988). Based on standards accepted for other picomaviridae (Andre et al., 1990; Duchene et al., 1990) and on previous experience, we estimated two passages to be the minimum requirement for safe inactivation screening. Our screening procedures were thus based on in-vitro propagation times of six weeks, i.e., three weeks in the original culture, followed by disruption of the cells by freezing and thawing, inoculation of a fresh cell culture with the lysate and amplification for three further weeks. As a first step towards establishing a PCR based screen, we assessed the relative sensitivities of PCR amplification/ Southern blot and tissue culture propagation followed by ELISA. Serial, ten-fold dilutions of HAV (10 J TCIDsO/ml) stock were used to infect human MRC-5 diploid cells (1 ml/25 cm2 flasks containing 1.5 ’ lo5 cells/cm2). Infected cultures were harvested for PCR 1,2, and 3 weeks post-infection (p.i.). Duplicate cultures were harvested in parallel, lysed by two cycles of freezing

160

and thawing and tested by quantitative ELISA for HAV antigens (ELISA 1 and 2 weeks (p.i.). Three-week-old cultures were either tested directly (ELISA 3 weeks p.i.) or used as inocula for fresh cultures, which were incubated for three further weeks before assay for viral antigens by the same method (ELISA 6 weeks p.i.). Mock infected cultures served as negative controls. Total RNA was extracted from the PCR series, first strand cDNA synthesis was carried out with random primers and HAV VP1 gene fragment (nt 2205-3107) was amplified from the heteroduplexes using VP1 specific primers. Detection of amplified sequences was carried out by Southern blot hybridization against radiolabelled VP1 cDNA. Anti-HAV ELISA was carried out on freeze/thaw cell lysates (see Materials and Methods). As expected, the PCR-based screen proved to be far more sensitive than the ELISA-based test, with the lowest viral inocula detected by each method being: PCR, 2 TCIDsa at 1 and 2 weeks and 0.002 TCIDsc at 3 weeks; ELISA, 1800 TCIDsO at 2 weeks; 18 TCIDso at 3 weeks and 1.5 TCIDSO at 6 weeks (data not shown). These results confirmed the value of the test, but, at the same time, indicated a need to reduce the level of detection to values closer to those obtained by ELISA,. Based on theoretical considerations on the nature of PCR amplification (Mullis, 1991) we assumed that contemporary amplification of two different fragments should leave fewer enzyme molecules available for the amplification of each fragment, in the later rounds, than if each were amplified separately under identical conditions. To avoid introducing further variables, a target sequence was chosen which is both ubiquitous and is expressed at fairly constant levels under our experimental conditions, and designed oligonucleotides for the amplification of a 540 nt fragment of cytoplasmic y actin. Coamplification of y-actin, which we termed ‘compound PCR’, offers two more crucial advantages: (a) it constitutes an internal control of the efficiency of each reaction; (b) it permits rough comparison of yields. Since all cultures should contain comparable levels of actin mRNA; in the absence of a visible VP1 band, lower 540 nt fragment yields can only stem from experimental variability and they can be corrected by loading proportionately more sample onto the analytical gel. Fig. 1 shows an ethidium bromide-stained agarose (1%) gel, on which equal aliquots of compound PCR reaction mixtures are compared. Each reaction was carried out on cDNA/RNA heteroduplexes synthesised on total RNA from infected or control cultures. The only visible band in most samples, is the 540 bp actin band, present at comparable levels in all reactions but for those carried out in the absence of RNA or in the absence of actin-specific primers. The VP1 specific band shows up only in the sample obtained from cells infected with live HAV (18 TCIDso, 2 weeks p.i.) Fig. 2 shows Southern blot analysis of PCR products obtained as in Fig. 1 and loaded to comparable y-actin amplimer levels, carried out with an HAV VP1 probe. Three conclusions can be drawn from these results: (a)

161

ABCDEFGH

Fig. 1. Ethidium bromide stain of PCR products after agarose gel (1%) electrophoresis. Infected and control cultures were harvested and processed for compound PCR, as described in Materials and Methods, at 1, 2, and 3 weeks p.i. Lanes A-C: compound PCR products from cultures infected with inactivated HAV batch A harvested 1,2, and 3 weeks p.i., respectively. Lane D: compound PCR products from mock-infected cultures harvested 3 weeks p.i. Lane E: compound PCR product of a reaction carried out in the absence of cDNA. Lane G: void. Lanes F, H: PCR products of reactions carried out in the presence of VP1 primers only on the same sample as lane B, and on cDNA from cultures infected with 18 TCIDss, harvested 2 weeks p.i., respectively.

amplification of the VP1 fragment from the same heteroduplex mix is more efficient when carried out in the presence of VP1 primers alone, as opposed to VP1 and actin primers together (Fig. 2, lanes H and J), as we had assumed; (b) cell cultures infected with incompletely inactivated vaccine, i.e., cell cultures ABCDEFGHIJKL

Fig. 2. Southern blot analysis of compound PCR products. Comparable amounts of amplified y actin fragments (see Fig. 1) were electrophoresed in 1% agarose gels and Southern blotted against a radiolabelled HAV VP1 specific probe. After stringent washing the filter was exposed to a Du Pont Cronex film for 7 h at room temperature. Lanes AC: compound PCR products of reactions performed on cDNA from cells infected with inactivated vaccine batch A, harvested 1, 2, and 3 weeks p.i., respectively. ELISA 6 weeks p.i. of batch A was negative. Lane D-F: compound PCR ptoducts of reactions carried out on cDNA from cells infected with inactivated vaccine batch B, harvested 1, 2, and 3 weeks p.i., respectively. ELISA 3 weeks p.i. of batch B was negative, ELISA 6 weeks pi. of the same batch was positive. Lanes G and H: compound PCR products of reactions performed on cDNA from cells infected with inactivated vaccine batch C, harvested 1 and 2 weeks p.i., respectively. ELISA 3 weeks p.i. of batch c was negative, ELISA 6 weeks p.i. of the same batch was positive. Lane I: negative control 1: compound PCR products of a reaction carried out on cDNA from mock-infected cells, harvested 3 weeks p.i. ELISA 6 weeks p.i. was negative. Lane J: products of PCR reaction performed on same cDNA as in lane H (batch C, 2 weeks p.i.) in the presence of HAV specific primers only, all other conditions unchanged. Lane K: negative control 2: compound PCR products of reaction carried out in the absence of cDNA. Lane 12: positive control: compound PCR products of reaction performed on cDNA from cells infected with 18 TCIDss of virulent HAV, harvested 2 weeks p.i. ELISA 3 weeks p.i. was weakly positive.

162

which were found borderline positive by ELISA 6 weeks p.i., are detected as positive by a clear PCR/blot hybridization signal in l-2 weeks (Fig. 2, lanes D to H) while cultures infected with inactivated vaccine, as established by ELISA 6 weeks p.i., were found negative (lanes A to C). In conclusion, compound PCR actually lowers the yield of amplified VP1 and brings the threshold of detection very close to that of the accepted ELBA as a test of inactivation. Further tests of inactivated virus stocks showed a perfect correlation between positive score by a compound PCR at 2 weeks and positive ELISA score at 6 weeks, thereby providing a net gain of 4 weeks.

Discussion The concept of HAV vaccine inactivation is completely dependent on the screening technique employed. The same viral inoculum may not be able to replicate in the animal while giving rise to viable progeny in tissue culture, thus scoring as inactivated in one test but not in the other. If we restrict examination to tissue culture screening, the duration of viral amplification becomes crucial, whereby samples which are found negative at 3 weeks may convert to weak positive by 6 weeks (V.P., unpublished observation). We have established conditions for a PCR-based test for vaccine inactivation capable of not increasing the sensitivity of tissue culture amplification followed by ELISA, but to reach similar results in a shorter period of time. Due to marked sensitivity, PCR based screens are best suited for ‘all or none’ situations, e.g., diagnosis of the presence of integrated viral genomes in host DNA. This is not necessarily the case when dealing with HAV vaccine inactivation. In fact, the extremely low threshold of detection of PCR-based screens, may be expected to reveal the presence of minute amounts of replicating virus, which would escape detection by conventional methods, unless incubation times were protracted for impractically long periods. This may in turn lead to a new definition of a ‘TCIDSePCR. Moreover, little is known about the half-life of HAV genomic RNA in inactivated vaccines. Thus the risk exists, at least theoretically, of amplifying fragments of carry-over dead virus, and noting it as positive. In a PCR-based test of inactivation, the negative samples are crucial. Any more likely to complex, multi-step procedure is, barring contaminations, generate false negative results than false positive results. Thus, suitable and quantifiable controls must be designed internally for each experimental sample. This was achieved in our protocol by coamplification of a 540 nt fragment of cytoplasmic y actin coding sequences by the same PCR which amplifies HAV VP1 sequences. Screening PCR products on ethidium bromide agarose gels makes both qualitative and quantitative comparisons possible. Finally, Southern blot hybridization with specific HAV probes was included in the

163

procedure. As a result, the fragment detected is positively identified by three criteria: (a) primers specificity and PCR conditions; (b) correspondence of the expected with the observed size and; (c) recognition by a specific probe under stringent conditions. This protocol has several advantages: it introduces an internal control into each experimental stage, which, in our opinion, is vital for a correct assessment of negative results (i.e., successful inactivation) and its parameters, such as oligonucleotide composition and concentration, and buffer strength lend themselves to tailor the scope of the methods to a vast spectrum of particular needs. References Andre, F.E., Hepburn, A. and D’Hondt, E. (1990) Inactivated candidate vaccines for hepatitis A. In: J.L. Melnick (Ed), Progress in medical virology, 37, Karger, Basel, pp. 72-95. Binn, L.N., Bancroft, W.H., Eckels, K.H., Marchwicki, R.H., Dubois, D.R., Asher, L.V.S., Le Due, J.W., Trahan, C.J. and Burke, D.S.(1988) Inactivated hepatitis A vaccine produced in human diploid MRC-5 cells. In: A.J. Zuckerman (Ed), Viral Hepatitis and liver disease, Alan R. Liss Inc., New York, pp. 91-93. Chandary, R.K., Parker, C. and MO, T. (1988) Tissue culture grown hepatitis A virus vaccine. In: A.J. Zuckerman (Ed), Viral hepatitis and liver disease, Alan R. Liss, Inc., New York, pp. 97-99. Duchene, M., Peetermans, E., D’Hondt, E., Harford, N., Fabry, L. and Stephenne, J. (1990) Production of Poliovirus vaccines: present, past and future. Vir. Immunol. 4, 243-272. Fineschi, N., Cavalieri, F., Garelick, H., Prugnola, A., Pellegrini, V. and Zuckerman, A.J. (1991) Characterization of a hepatitis A virus strain suitable for vaccine production. J. Hepatol. Sup 4, Sl-6. Flehming, B., Heinricy, U. and Ptisterer, M. (1989) Immunogenicity of a killed hepatitis A vaccine in seronegative volunteers. Lancet 1, 1039-1041. Garelick, H., Mann, G.F., Harrison, T.J. and Zuckerman, A.J. (1988) Defective interfering particles in hepatitis A. In: A.J. Zuckerman (Ed), Viral hepatitis and liver disease, Alan R. Liss Inc., New York, pp. 12-15. Lewis, J.A., Armstrong, M.E., Larson, V.M., Emini, E.A., Midthun, K., Ellerbeck, E., Nalin, D., Provost, P.J. and Calandra, G.B. (1990) Use of a live, attenuated hepatitis A vaccine to prepare a highly purified, formalin inactivated hepatitis A vaccine In: F.B. Hollinger, SM. Lemon, H.S. Margolis (Eds) viral hepatitis and liver disease, Williams and Williams, Baltimore, pp. 94-97. Mullis, K.B. (1991) The Polymerase chain reaction in an anemic mode: how to avoid cold oligodeoxyribonuclear fusion PCR 1, l-5. Sambrook, J., Fritsch E.F. and Maniatis, T. (1989) Molecular cloning, a laboratory manual. Cold Spring Harbour Laboratory Press, Cold Spring Harbour.

Assessment of inactivation of hepatitis A vaccine by compound PCR.

Assuring the complete inactivation of hepatitis A virus (HAV) vaccine commonly requires prolonged tissue culture amplification, followed by detection ...
567KB Sizes 0 Downloads 0 Views