Journal of Medical Virology 4 1 : 6 1 4 (1993)

Inactivation of Hepatitis A Virus by Heat Treatment in Aqueous Solution

Paula Murphy, Thomas Nowak, Stanley M. Lemon, and Joachim Hilfenhaus Department of Medicine, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina (P.M., S.M.L.); Behringwerke A G Research Laboratories, Marburg, Germany (T.N., J.H.)

Hepatitis A virus infections have been reported recently among hemophilic patients in Italy and Germany, leading to speculation that infectious hepatitis A virus (HAW might have been present in some factor Vlll concentrates. In both cases, the implicated factor concentrates had been treated by a solvent/detergent method, which in- activates enveloped viruses but which would not be expected to inactivate HAV, a nonenveloped picornavirus. To determine whether HAV would be inactivated during pasteurization of factor Vlll concentrate, an alternative method employed for virus inactivation, we determined the extent to which the infectivity of cell culture-adapted HAV, suspended either in cell culture medium or in a proprietary stabilizing buffer, was reduced by heat treatment at 60°C for 10 hr. The titer of infec- tious HAV declined rapidly at 60"C, but the stabi- lizer considerably delayed HAV inactivation. In cell culture medium, HAV was inactivated by >3.6 log,, within 30 min, but 3.6 log,, inactiva- tion of HAV was reached only after 6 hr in the presence of the stabilizer. Residual infectious HAV was present after even 10 hr of heat treat- ment in the stabilizer, indicating that <5.2 log,, infectious HAV particles are inactivated under these conditions. In the presence of the stabi- lizer, HAV was significantly more stable than po- liovirus type l , which has been used to validate virus inactivation by pasteurization. We con- clude that pasteurized factor Vlll concentrate should pose little if any risk for transmission of HAV if pooled plasma used for its manufacture contained low levels of the virus. Q 1993 Wiley-Liss, Inc.

KEY WORDS: factor Vlll concentrate, pasteur- ization, HAV, poliovirus

INTRODUCTION Hepatitis A virus (HAV) is almost always transmit-

ted by the fecal-oral route. However, transmission of HAV by transfusion of blood or blood products does occur, albeit at a low frequency compared to the blood- 0 1993 WILEY-LISS, INC.

borne transmission of hepatitis B or C viruses. Cases of hepatitis A have been described that appear to have been caused by transfusion of whole blood donations, erythrocyte concentrates, or plasma [Hollinger e t al., 1983; Sheretz et al., 1984; Weisfuse et al., 19901. The possible transmission of HAV by a purified human plasma derivative, however, was not reported until re- cently. An apparent outbreak of hepatitis A involving 41 patients with hemophilia A, diagnosed in Italy be- tween 1989 and 1992, was suspected to be related to virus present in a high-purity factor VIII concentrate manufactured by an Italian company [Mannucci, 19921. In addition, the possible transmission of HAV by factor VIII concentrates administered to patients with hemo- philia A has also been reported recently from Germany [Gerritzen et al., 19921. The implicated factor VlII con- centrates were manufactured using a solventldetergent method to inactivate blood-borne viruses commonly present in large human plasma pools, including hepati- tis B virus, hepatitis C virus, and the human immuno- deficiency virus (HIV-1). This inactivation method would not be expected to inactivate nonenveloped vi- ruses, however, such as HAV. The possibility that HAV was transmitted via these plasma protein preparations is supported by a recent report that HAV RNA could be detected by a reverse transcriptiodpolymerase chain reaction method in a t least some of the factor VIII con- centrates involved [Normann et al., 19921.

The heat treatment of factor VIII concentrates in sta- bilized aqueous solution at 60°C for 10 hr (pasteuriza- tion) represents a n alternative approach to inactivation of viral contaminants. Previous studies have shown that poliovirus, a picornavirus like HAV, is inactivated by pasteurization [Hilfenhaus et al., 1986; Nowak et al., 19921. However, HAV is known to have signifi- cantly greater thermostability than poliovirus [Siegl et al., 19841, and its inactivation has never been studied under these conditions. Because of the reported out- break of hepatitis A among Italian hemophiliacs, we

Accepted for publication January 20,1993. Address reprint requests to Stanley M. Lemon, Department of

Medicine, The University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7030.

62 Murphy et al.

each flask. After incubation of the cell cultures for 14 days at 35”C, the supernatant fluids were decanted and the cell sheets were lysed by the addition of 5 ml me- dium containing 0.1% sodium dodecyl sulfate (SDS). Lysates were clarified by low-speed centrifugation and the supernatant was stored a t -20°C until testing for hepatitis A antigen by a microplate-based, solid-phase radioimmunoassay [Lemon et al., 19821. As a positive control, similar cultures were carried out with the 6 hr heat inactivation sample, which contained virus a t the limits of detection by radioimmunofocus assay (see Re- sults).

Poliovirus was titrated in a standard microplate as- say, starting with a 1:lO dilution of the original sample. The detection limit of this assay was 40 ID,,/ml. If no infectious virus was detected in the microplate assay, four replicate 1 ml aliquots of the inactivated material were diluted 1:lO with medium and inoculated onto confluent Hep-2 cells in 25 cm2 flasks. If all four flasks did not show any cytopathic effect after a 7 day incuba- tion at 35”C, the virus titer was assumed to be < 1 ID5,/ml.

RESULTS The inactivation of HAV and poliovirus type 1 by

pasteurization is shown in Figure 1. Both viruses were rapidly inactivated at 60°C when suspended in cell cul- ture medium. However, both these picornaviruses were considerably more resistant to heat inactivation when suspended in the stabilizer solution, and HAV was much more stable under these conditions than poliovi- rus type 1. Infectious HAV was still detectable by radio- immunofocus assay after 6 h r a t 60”C, indicating that only 3.6 log,, infectious HAV particles were inacti- vated under these conditions. Moreover, infectious HAV was identified in each of the five replicate cell cultures inoculated with 1 ml aliquots of virus samples, which had been heated at 60°C for 6 hr, 8 hr, or 10 hr. The final HAV titer of these samples was thus between 1 and 40 rfdml, indicating that 6 . 2 log,, infectious HAV particles were inactivated after 10 hr a t 60°C in the stabilizer solution. In contrast to HAV, the inacti- vation of poliovirus type 1 was complete after 4 hr of heat inactivation, even in the stabilizer (Fig. 1). No infectious poliovirus was detected in 1 ml aliquots of virus heated for 4 h r or more.

DISCUSSION The results presented here demonstrate that HAV is

substantially, but incompletely, inactivated by a pas- teurization process used for the manufacturing of hu- man plasma derivatives. HAV was significantly more heat stable than poliovirus when suspended in a solu- tion used to stabilize factor VIII during pasteurization. This stabilizer contains high concentrations of sucrose, which is used commercially to stabilize some oral polio- virus vaccines. The greater thermal stability of HAV, compared with poliovirus, confirms earlier studies [Siegl et al., 19841. The structural features of the HAV

investigated the inactivation of HAV by pasteuriza- tion. These studies included a comparison of the heat stabilities of HAV and poliovirus type 1, because polio- virus has routinely been included in virus safety stud- ies carried out according to the CPMP guidelines for “Validation of Virus Removal and Inactivation Proce- dures” [Committee for Proprietary Medicinal Products: Ad Hoc Working Party on Biotechnology/Pharmacy and Working Party in Safety Medicines, 19911.

MATERIALS AND METHODS Viruses and Cells

We studied the thermal inactivation of a rapidly rep- licating, cell culture-adapted variant of HAV (HM175/ 18f virus) [Lemon et al., 19911 and the Sabin vaccine strain of poliovirus type 1. The infectivity titer of HAV was determined in BS-C-1 cells and that of poliovirus in Hep-2 cells. Virus samples either were assayed imme- diately after heat treatment, or they were diluted 1 : l O with cell culture medium and stored frozen at -20°C prior to inoculation of cell cultures.

Heat Inactivation of Virus Separate suspensions of HAV and poliovirus were

prepared. One volume of each virus stock was added either to nine volumes of cell culture medium or to nine volumes of a protein solution used to stabilize factor VIII concentrate during pasteurization (20 mg/ml pro- tein, 588 mg/ml sucrose, 88 mg/ml glycine, pH 7.3). The final titer of HAV in the suspension was lo5 radioim- munofocus-forming units (rfu)/ml, and that of polio- virus was --lo7 plaque-forming units (pfu)/ml. After thorough mixing, the suspensions were divided into multiple aliquots (-5 ml each), which were submerged in a conventional water bath maintained at 60°C. Tem- perature was monitored throughout a 10 hr inactiva- tion period, with individual aliquots removed at, peri- odic intervals and placed at 0°C (or frozen a t -20°C) until virus titration.

Virus Infectivity Assays Because the undiluted stabilizer solution was toxic to

cell sheets, virus titrations were carried out with ten- fold or greater dilutions. Tenfold serial dilutions (be- tween 10-1 and of HAV samples were tested for infectious virus by radioimmunofocus assay [Lemon et al., 19911. F’etri dish cultures of BS-C-1 cells wert, Inoc- . ulated with 0.25 ml volumes of these dilutions, result- ing in a detection limit of 40 rfu/ml in inactivated virus suspensions. Samples that had been heated in the stabi- lizer solution for 8 or 10 h r and that were negative in this assay (see Results) were further investigated for residual infectious HAV as follows. Five replicate 75 cm2 flasks containing nearly confluent BS-C-1 cells were inoculated with 10 ml of a 1:lO dilution of each sample. Two control flasks were inoculated with an equal volume of nonheated stabilizer solution diluted 1:lO in cell culture medium. The flasks were incubated for 2 h r a t 35”C, and 20 ml fresh medium was added to

Thermal Inactivation of HAV 63

t

P \

; 3 E lo4

I I I I I I I I I I

0 2 4 6 8 10 0 2 4 6 8 10

Hrs at 6OoC Hrs at 6OoC

Fig. 1. Inactivation of HAV (left) or the Sabin vaccine strain of poliovirus type 1 (right) by pasteurization at 60°C for up to 10 hr. Circles represent virus suspended in cell culture medium; squares represent virus suspended in the factor VIII stabilizer solution. The horizontal dashed line indicates the detection limits of the quantita-

tive assay, while values below this line represent the semiquantita- tive detection of virus by inoculation of cell cultures with large vol- umes of heat-treated materials. Solid symbols represent samples shown to contain infectious virus by either type of assay.

particle that account for this remarkable thermal sta- bility are not known. Moreover, it is not clear to what extent aggregation of the virus might contribute to its thermostability profile. However, our observations demonstrate the value of using “relevant” viruses such as HAV, rather than model viruses such as poliovirus, in studies aimed at validating virus inactivation proce- dures. The HAV strain used in these studies (HM175/ 180 is highly adapted to growth in cell culture but has only three amino acid substitutions in the capsid poly- peptides. Previous studies have demonstrated that its thermostability is indistinguishable from that of low- passage virus possessing the wild-type capsid amino acid sequence [Lemon et al., 19911.

The fact that low titers of infectious HAV remained detectable in a factor VIII stabilizer that had been spiked with HAV prior to pasteurization does not indi- cate that this virus poses any special risk to recipients of heat-treated factor VIII concentrates. Such a risk would depend, in part, on the probability of a plasma pool being contaminated with HAV and the titer of HAV present in contaminated pools. This is difficult to estimate and certainly is dependent on the source of the plasma. HAV does not cause chronic infections, but viremia is probably always present during the incuba- tion period as well as the acute phase of hepatitis A

[Krugman et al., 1959, 19621. Viremias in monkeys infected experimentally with a cell culture-adapted vi- rus reached titers in excess of lo4 rfdml serum and had a mean duration of -3 weeks [Lemon et al., 19901. Thus it would not be surprising to find HAV present in pools of plasma donations of the size used for commercial factor VIII manufacture (4,000 or more donors). At present, no practical means is available for screening donor units for the presence of HAV. Plasma collected exclusively from donors with antibody to HAV would have little if any risk of containing virus, but such a policy would almost certainly result in an unacceptably high rate of donor exclusion.

Because the prevalence of HAV viremia must be very low, even in populations at high risk for hepatitis A, it is likely that the titer of HAV present in contaminated plasma pools would also be low. Large plasma pools would be expected to contain substantial quantities of neutralizing antibodies to HAV, which would signifi- cantly reduce the infectious titer of HAV present in any plasma pool prior to processing. In addition, the manu- facturing process used for factor VIII production is ca- pable of eliminating viruses to a considerable extent. Previous studies demonstrated an elimination factor of lo5 for poliovirus during manufacture of high-purity factor VIII concentrate (T. Nowak and J. Hilfenhaus,

64

unpublished results). Although such studies have not been carried out with HAV, these data provide further assurance that HAV contaminating a plasma pool would be reduced to a level that would be effectively eliminated by the pasteurization process.

We emphasize that the risk of transmission of HAV by processed human plasma proteins remains poorly defined. Although the reported cases of HAV infection among Italian hemophilics may have been related to administration of solvent-detergent-treated factor VIII concentrate, it is not clear a t present whether the plasma pool or fresh water used for manufacturing the product was contaminated with HAV [Mannucci, 19921. Similarly, the epidemiology of the infections re- ported recently from Germany [Gerritzen et al., 19921 requires further investigation. Additional studies are clearly warranted to determine the extent to which HAV infection might be associated with administration of processed human plasma products such as factor VIII concentrate.

Murphy et al.

mophilic plasma protein preparations by pasteurization. Vox San- guis 50:20%211.

Hollinger FB, Khan NC, Oefinger PE, Yawn DH, Schmulen AC, Dreesman GR, Melnick J L (1983): Posttransfusion hepatitis type A. JAMA 250:2313-2317.

Krugman S, Ward R, Giles JP (1962): The natural history of infectious hepatitis. American Journal of Medicine 32:717-728.

Krugman S, Ward R, Giles JP, Bodansky 0, Jacobs AM (1959): Infec- tious heoatitis: Detection of virus during the incubation oeriod and

ACKNOWLEDGMENTS We thank R. Weber for expert technical assistance

and A. Baumbach and S. Hildebrants for preparation of the manuscript. Work in the laboratory of S.M.L. was supported in part by a gift from Behringwerke AG Re- search Laboratories.

REFERENCES Committee for Proprietary Medicinal Products: Ad Hoc Working

Party on 13iotechnology/Pharmacy and Working Party on Safety Medicines (1991): EEC Regulatory Document, Note for Guidance: Validation of virus removal and inactivation procedures. Biologi- cals 19:247-251.

Gerritzen A, Schneweis KE, Brackmann HH, Oldenburg J, Hanfland P, Gerlich WH, Caspari G (1992): Acute hepatitis A in hemophili- acs. Lancet 340:1231-1232.

Hilfenhaus J, Herrmann A, Mauler R, Prince AM (1986): Inactivation of the AIDS-causing retrovirus and other human viruses in antihe-

in clinically inapparent infection New-England Journal of Medi- cine 261 729-734

Lemon SM, Binn LN, Marchwicki R, Murphy PC, Ping LH, Jansen RW, Asher LVS, Stapleton JT, Taylor DG, LeDuc JW (1990): In vivo replication and reversion to wild type of a neutralization- resistant antigenic variant of hepatitis A virus. Journal of Infec- tious Diseases 161:7-13.

Lemon SM, LeDuc JW, Binn LN, Escajadillo A, Ishak KG (1982): Transmission of hepatitis A virus among recently captured Pana- manian owl monkeys. Journal of Medical Virology 10:25-36.

Lemon SM, Murphy PC, Shields PA, Ping LH, Feinstone SM, Cromeans T, Jansen RW (1991): Antigenic and genetic variation of cytopathic hepatitis A virus variants arising during persistent infection: Evidence for genetic recombination. Journal of Virology 65:20562065.

Mannucci PM (1992): Outbreak of hepatitis A among Italian patients with haemophilia. Lancet 339:819.

Normann A, Graff J, Gerritzen A, Brackmann H-H, Flehmig B (1992): Detection of hepatitis A virus RNA in commercially available fac- tor VIII preparation. Lancet 340:1232.

Nowak T, Gregersen JP , Klockmann U, Cummins LB, Hilfenhaus J (1992): Virus safety of human immunoglobulins: Efficient inacti- vation of hepatitis C and other human pathogenic viruses by the manufacturing procedure. Journal of Medical Virology 36:209- 216.

Sheretz RJ, Russel BA, Reuman PD (1984): Transmission of hepatitis A by transfusion of blood products. Archives of Internal Medicine 144:1579.

Siegl G, Weitz M, Kronauer G (1984): Stability of hepatitis A virus. Intervirology 22:218-226.

Siegl G, Weitz M, Kronauer G (1984): Stability of hepatitis A virus. Intervirology 22:218-226.

Weisfuse IB, Graham DJ, Will M, Parkinson D, Snydman DR, Atkins M, Karron RA, Feinstone S , Rayner AA, Fisher RI, Mills BJ, Dutcher JP , Weiss GR, Glover A, Kuritsky JN, Hadler SC (1990): An outbreak of hepatitis A among cancer patients treated with interleukin-2 and lymphokine activated killer cells. Journal of Infectious Diseases 161547-652.