Safety profile and immunogenicity of an inactivated vaccine derived from an attenuated strain of hepatitis A Edward F. Ellerbeck*t, John A. LewisS, David Nalinx, Kenneth Gershman* William J. Millers Marcy E. Armstrong3 Joseph P. Davide$, Audrey E. Rhoadf Brenda McGuiref, Gary Calandraj, Philip J. ProvostS and Karen Midthun*{§ To determine the safety and immunogenicity of an inactivated hepatitis A vaccine, 56 healthy adult volunteers were randomly assigned to receive an intramuscular injection of 6.3, 12.5 or 25 ng of inactivated hepatitis A vaccine or placebo at 0, 2 or 4, and 24 weeks. Adverse reactions occurred with similar frequency in vaccine and placebo recipients and consisted primarily of pain or tenderness at the injection site. By 4 weeks after a single 6.3, 12.5 or 25 ng injection, seven, nine and ten out of ten vaccinees, respectively, had antibody detectable by a HAVAB assay modified to increase its sensitivity tenfold. All vaccinees had antibodies detectable by this assay within 2 weeks of their second inoculation. Geometric mean antibody levels increased with higher doses of vaccine (p = 0.05). Neutralizing antibody was detected within 4 weeks of a single inoculation in all vaccinees. Neutralizing antibody was detected after the third inoculation at dilutions of > 1:2048 in all 12.5 and 25 ng vaccinees. All 19 vaccinees tested at 24 months still had HA V antibodies detectable by a modt$ed HAVAB assay. This inactivated hepatitis A vaccine appears to be well tolerated and immunogenic at doses of 6.3-25 ng. The choice of dose and vaccination schedule may depend on the rapidity with which seroconversion is desired. Keywords: Hepatitis A; vaccine; inactivated; tolerance; immunogenicity
INTRODUCTION Hepatitis due to infection with the hepatitis A virus (HAV) remains an important health problem worldwidel. Although immune globulin can provide shortterm protection for high risk individuals2-7, the need for repeat dosing every 6 months makes it impractical for long term prophylaxis. A vaccine that induces longlasting immunity would clearly be preferable. The propagation of HAV in cell culture in 1979* made the development of a vaccine a realistic possibility. Since then, both live-attenuated and inactivated HAV vaccines have been tested in animals and human volunteers*-23. These vaccines have generally been well tolerated, and some have been shown to protect primates against challenge with virulent HAV strains*-r4. A purified, formalin-inactivated, alum-adjuvanted vaccine prepared from HAV strain CR326 protected marmosets against challenge with virulent HAVi3. A live-attenuated vaccine *Center for Immunization Research, School of Hygiene and Public Health, and +Department of Medicine, School of Medicine, Johns Hopkins University, 624 N. Broadway, Room 125, Baltimore, MD 21206, USA. *Merck Sharp & Dohme Research Laboratories, Sunneytown Pike, West Point, PA 19486, USA. §To whom correspondence should be addressed. (Received 10 December 1991; revised 11 February 1992; accepted 14 February 1992) 0264410x/92/1-5 0 1992 Buttetworth-Heinemann 668
Ltd
Vaccine, Vol. 10, Issue 10, 1992
candidate derived from the F’ variant of HAV strain CR326F was well tolerated and immunogenic in human volunteers’2*20 and also protected marmosets against challenge with virulent HAVr2. We now report the results of a dose-escalating study of a formalin-inactivated, alum-adjuvanted vaccine prepared from the attenuated F’ variant of HAV strain CR326F and tested in healthy seronegative volunteers. Preliminary results of this study have been reported previously24. MATERIALS
AND METHODS
Hepatitis A virus vaccine strain
The HAV used to make the inactivated vaccine was derived from the master seed of the live, attenuated F’ HAV variantI by carrying out five additional serial passages of the virus in MRC-5 cells at 35°C and eight further passages of the virus in MRC-5 cells at 32°C. Vaccine was prepared from HAV contained in infected cells of the final passage. Hepatitis A antigen purification
HAV virus was purified from virus-infected MRC-5 cell pellets harvested from roller bottles by methods conventionally used in picornavirus purification, including cell lysis, organic extraction, and ion exchange and size-exclusion chromatography.
Inactivated
hepatitis
A vaccine: E.F. Ellerbeck
et al.
Hepatitis A antigen inactivation
Detection of antibodies to HAV
Purified HAV antigen in phosphate-buffered saline was inactivated by exposure to 100 pg ml-’ formalin at 37°C for 20 days with continuous stirring. At 1 and 72 h after the addition of formalin, bulk antigen was transferred from one glass vessel to another to ensure complete exposure of antigen to formalin. At 20 days, the bulk vaccine antigen was passed through a 0.2 pm filter and adsorbed to aluminium hydroxide. Thimerosal was added to 0.005%. The vaccine was stored at 4°C. A 10 pg aliquot of bulk vaccine antigen, withdrawn after 20 days of exposure to formalin, was dialysed to remove residual formalin. MRC-5 cell monolayers established in ten T25 flasks were each infected with a 1 pg aliquot of the dialysed vaccine antigen. The monolayers were refed at 7-day intervals and harvested on the 35th day by scraping. The cell harvest from each flask was lysed by freeze-thawing and sonication, and 200 ~1 aliquots of each flask were used to inoculate ten additional T25 flasks which were maintained as described above. After an additional 3 days, the cells were again harvested, lysed, and assayed for hepatitis A antigen by radioimmunoassay (RIA) I3 . The absence of detectable antigen documented complete inactivation of the virus. Further details on the preparation of this vaccine have previously been published2’.
A standard HAVAB (Abbott Laboratories, N. Chicago, IL) and HAVAB and HAVAB-M assays modified to increase their sensitivity were used to measure HAV antibody. In the modified HAVAB assay, a tenfold greater volume of serum (test or control) was tested (0.1 mlofserainO.l mlof 1251-labelled anti-HAV instead of 0.01 ml of sera in 0.2 ml of ‘251-labelled anti-HAV). Results were converted to International Units by comparing them with a reference curve generated from World Health Organization (WHO) Reference Globulin no . 12’j. The modified HAVAB-M assay tested 0.01 ml of undiluted serum instead of 0.01 ml of a 1:200 dilution. Neutralizing antibody titres were determined by the hepatitis A virus antigen-reduction neutralization assay (HAVARNA) described by Krah et a1.27.
Subjects Fifty-six healthy adults (22 males and 34 females), 18-50 years old, were recruited from the Baltimore metropolitan area. All volunteers had a normal physical examination and normal serum levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST) and gamma glutamyl transpeptidase (GGT). They were negative for hepatitis B surface antigen and HAV antibody. Persons at increased risk of HAV infection (consumption of raw shellfish or travel to HAV endemic countries) were excluded from this study. All subjects gave informed, written consent to participate in this study, which was approved by the institutional review boards of the participating institutions. Experimental design HAV vaccine, at a concentration of 12.5 ng ml-‘, was tested at doses of 6.3, 12.5 and 25 ng (see Appendix). At each dose, 14 volunteers were randomly assigned, in a single-blind fashion, to receive a single intramuscular (deltoid) injection of either the HAV vaccine (n = 10) or placebo (n = 4) at 0, 4 and 24 weeks. Fourteen volunteers received 6.3 ng of HAV vaccine (n = 10) or placebo (n = 4) at 0, 2 and 24 weeks. Volunteers recorded their temperatures, injection site reactions and other complaints for a total of 5 days following each inoculation. Serum ALT, AST and GGT were measured 2, 4, 5, 24 and 28 weeks after the initial vaccination. HAV. serology was measured in all volunteers at weeks 2-6, 8, 12, 20, 24, 28, 32 and 40. Individuals who received HAV vaccine at 0, 4 and 24 weeks were tested for neutralizing antibody at weeks 2, 4, 6 and 32. Endpoint titrations of neutralizing antibody were measured at week 32 in 12.5 ng vaccinees. After 24 months, vaccinees were asked to return to have their HAV serology measured again.
RESULTS Fifty-three (95%) of the 56 volunteers completed the study. One 6.3 ng vaccinee, one 12.5 ng vaccinee, and one placebo recipient dropped out of the study after 6 months of follow-up for reasons unrelated to vaccination. IgM antibody was detected by the modified HAVABM in all but one of the vaccinees within 4 weeks of the first vaccination. All vaccinees had detectable IgM antibody to HAV within 1 week after the second dose. By 4 weeks after a single 6.3, 12.5 or 25 ng injection given to ten vaccinees each, seven, nine and ten vaccinees, respectively, had antibody detectable by modified HAVAB. All vaccinees had seroconverted by week 6 (2 weeks after the second inoculation). Antibody levels increased approximately 20-fold in all dose groups after the 6-month boost (Figure 1). The magnitude of the antibody response was dose dependent, with higher antibody levels in recipients of higher doses of vaccine (p = 0.05, linear regression of log titres, week 32). Of the 6.3 ng vaccinees, ten received their second inoculation at week 2 and ten at week 4. All vaccinees boosted at week 2 seroconverted by week 4 as compared with seven of the 6.3 ng vaccinees who had not yet received their second inoculation. Geometric mean antibody levels from week 8 to 24 were similar in those receiving their boost at 4 weeks (GMT range: 50-l 13 mIU) or at 2 weeks (GMT range: 70-90 mIU) (Figure 2). After the third immunization (week 24), geometric mean antibody levels at week 32 were not significantly different in volunteers receiving 2-week -
‘I
7i
10000
1
z
1000
L
1
0
8
I
16 Weeks
after
24
L
32
40
vaccination
Figure 1
Geometric mean antibody responses by modified HAVAB in volunteers vaccinated with 6.3 (IJ). 12.5 (0) or 25 (0) ng of inactivated hepatitis A vaccine at 0, 4 and 24 weeks
Vaccine, Vol. 10, Issue
10, 1992
669
Inactivated 7_
hepatitis
A vaccine: E.F. Ellerbeck
et al. Table 2 Percentage of hepatitis A vaccine and placebo recipients reporting pain or tenderness at the injection site during any of 4 days after the day of vaccination
10000 c
Injection
x
I”
11
’
0
16
24
32
40
vaccination
Figure 2 Geometric mean antibody responses by modified HAVAB in volunteers vaccinated with 6.3 ng of inactivated hepatitis A vaccine at either (0) 0, 4 and 24 weeks or (a) 0, 2 and 24 weeks
Table 1 Antibody titres to hepatitis A virus by modified hepatitis A virus antibody (HAVAB) and neutralization assays at week 32 after three 12.5 ng injections of hepatitis A vaccine at 0, 4 and 24 weeks Anti-HAV titre by
134 136 137 136 139 141 143 146 147
Modified HAVAB (mlU ml-‘) 2942
1653 5296 10242 12506 6151 9365 3873 20583
Neutralization assay” 2046 4096 4096 6192 16364 6192 8192 8192 > 3276%
“Titre expressed as reciprocal of highest serum dilution with neutralizing antibody
(5464 mIU) or 4-week (3425 mIU) boosts (p = 0.18, Student’s t test). Among recipients of 6.3, 12.5 or 25 ng doses (n = 10 each), hepatitis A antibody was detected by the standard HAVAB in one, two and eight vaccinees, respectively, within 4 weeks of a single dose of vaccine. After three doses of vaccine, all HAV vaccinees had seroconverted by standard HAVAB. Individuals vaccinated at weeks 0, 4 and 24 were screened for neutralizing antibody. Neutralizing antibody, in recipients of 6.3, 12.5 or 25 ng doses (n = 10 each), was detected 2 weeks after the first inoculation in seven, nine and seven vaccinees, respectively. By 4 weeks after a single inoculation, all 30 vaccinees tested had detectable neutralizing antibody. At week 32, neutralizing antibody titres ranged from 1: 2048 to > 1: 32768 in the 12.5 and 25 ng vaccine.es and from 1:64 to > 1:4096 in the 6.3 ng vaccinees. End-point titrations of neutralizing antibody at week 32 in 12.5 ng vaccinees were highly correlated with modified HAVAB antibody levels ( Table 1; Pearson correlation coefficient (I) = 0.93, p < 0.001). Nineteen vaccinees were available for follow-up 24 months after their initial vaccination. All of these vaccinees had HAV antibody detectable by the modified HAVAB assay. HAV antibody levels ranged from 42-2756 mIU (GMT = 522 mIU) for 6.3 ng vaccinees (n = 11 ), 309-2498 mIU (GMT = 1062 mIU) for 12.5 ng
670
Vaccine,
1st
2nd
3rd
Placebo 6.3 12.5 25
19 15 0 40
13 20 10 50
33 20 11 50
(16) (20) (10) (10)
I
8
Weeks after
Subject no.
Dose” (ng) (n)
Vol. 10, Issue 10, 1992
“Volume of inoculum varied with dose: 6.3 ng (0.5 ml), 12.5 ng (1.0 ml) and 25 ng (2.0 ml). Placebo recipients received 0.5-2.0 ml depending upon the cohort in which they were enrolled
vaccinees (n = 4), and 646-7627 mIU (GMT = 1701 mIU) for 25 ng vaccinees (n = 4). No significant systemic reactions were noted after vaccination. Transient ALT elevations ( >40 IU) were noted in two of 40 HAV vaccinees (range 43-52 IU) and one of 16 placebo recipients (70 IU). Subjects reported symptoms of headache, dizziness, dyspepsia, myalgias or other mild non-localized complaints during the 5 days following three (6% ) of 47 placebo inoculations and eight (14%)of6.3ng(n=59),four(14%)of12.5ng(n=29) and two (7% ) of 25 ng (n = 30) HAV inoculations. None of these symptoms interfered with the usual daily activities of the volunteers. Injection site pain and tenderness persisted beyond the day of vaccination in O-50% of HAV vaccinees and 13-33% of placebo recipients in the different groups (Table 2). These reactions were reported more frequently in the 25 ng HAV vaccine group (40-50%) than in the 6.3 and 12.5 ng HAV vaccine groups (O-20%). The volume of inoculum, however, also varied with the dose. The three placebo recipients who received the same volume of inoculum as the 25 ng vaccine recipients reported pain or tenderness beyond the day of injection after five (62% ) of eight inoculations.
DISCUSSION This inactivated HAV vaccine was highly immunogenic at doses of 6.3 to 25 ng. The magnitude of the antibody response was dose-dependent, with higher doses of vaccine inducing higher antibody levels and earlier seroconversion. Low levels of neutralizing antibody were detected 2-4 weeks after a single 6.3, 12.5 or 25 ng dose of vaccine. Four weeks after vaccination, neutralizing antibody titres were equal to or greater than those found in immune globulin recipients 3 months after a 0.02 ml kg- ’ inoculation (unpublished data), a time when immune globulin recipients are still protected from Early seropositivity in HAV infection by HAV2-‘. vaccinees may be crucial for protecting travellers leaving on short notice to HAV endemic countries. The 6-month boost induced a 15-30-fold increase in HAV antibody levels. Neutralizing antibody titres after this boost were comparable to or higher than those reported in previous studies of inactivated HAV vaccines18,1g,21. Although antibody titres were substantially higher than those seen in a trial of the live, attenuated HAV vaccine derived from the F’ strain of CR326FZo, the persistence of the immune response to both inactivated and live attenuated HAV vaccines remains to be determined. The antibody response seen
Inactivated
with a 0, 4 and 24 week dosing schedule suggests that vaccination schedules for HAV could be coordinated with hepatitis B immunization schedules or that the HAV vaccine might be formulated in combination with hepatitis B vaccine”. Using 25 ng of vaccine, a two-dose vaccination schedule may also be possible. After a single 25 ng vaccination, antibodies can persist up to 6 months and a boost given at that time can induce antibody levels comparable to those seen after the third vaccine dose in our study (Shouval et al., unpublished results). The optimal dose and schedule of vaccination may depend upon the rapidity with which seroconversion is desired, compatibility with other vaccine schedules, and vaccine cost. This HAV vaccine provides a margin of safety not found in most other inactivated vaccines in that it is derived from an attenuated strain of HAV previously shown to be well tolerated without evidence of hepatotoxicity in human volunteers’2*20. All dose levels of the inactivated vaccine were well tolerated. The mild local reactions reported more frequently by both vaccine and placebo recipients in the 25 ng cohort were due probably to the two to four times larger volume of injection given to this cohort. The total of the different types of non-localized complaints reported was higher in HAV vaccinees than in placebo recipients, but these could not be linked to vaccination; the complaints were mild, unrelated to dose, and no particular type of symptom predominated. This inactivated HAV vaccine can induce high levels of antibodies detectable by both HAVAB and virus neutralization assays. Long-term follow-up of volunteers from this and other studies is needed to assess the duration of the immune response, and to determine whether, as in one study of natural disease, seroconversion with intact immune memory indicates long-lasting protection from disease regardless of the antibody titre 27. Additional studies are warranted to evaluate the safety of this vaccine in a larger series and to establish protective vaccine efficacy.
7
6
9
IO
11
12
13
14
15
16
17
16
19 20
21 22
23
ACKNOWLEDGEMENTS This work was supported by Merck Sharp & Dohme Research Laboratories. The authors are grateful to Janine Sherman at the Johns Hopkins University and Mary McCaughtry at Merck Sharp & Dohme Research Laboratories for assistance with this study.
24
25
26
REFERENCES 27 Lemon, SM. Type A viral hepatitis: new developments in an old disease. New Engl. J. Med. 1965, 313, 105%1067 Stokes, J. Jr and Neefe, JR. The prevention and attenuation of infectious hepatitis by gamma globulin. J. Am. Med. Assoc. 1945, 127,144-145 Hall, W.T., Madden, D.L., Mundon, F.K., Brandt, D.E.L. and Clarke, N.A. Protective effect of immune serum globulin (ISG) against hepatitis A infection in a natural epidemic. Am. J. Epidemiol. 1977. 106, 72-75 Woodson, R.D. and Clinton, J.J. Hepatitis prophylaxis abroad: effectiveness of immune serum globulin in protecting Peace Corps volunteers. J. Am. Med. Assoc. 1969, 288, 1053-1056 Havens, W.P. and Paul, J.R. Prevention of infectious hepatitis with gamma globulin. J. Am. Med. Assoc. 1945, 128, 270-272 Stokes, J. Jr, Farquhar, J.A., Drake, M.E., Capps, R.B., Ward, C.S. Jr and Kitts, A.W. Length of protection by immune serum globulin
26
hepatitis A vaccine: E.F. Ellerbeck et al.
(gamma globulin) during epidemics. J. Am. Med. Assoc. 1951,147, 714-719 Conrad, ME. and Lemon, SM. Prevention of endemic icteric viral hepatitis by administration of immune serum gamma globulin. J. Infect. Dis. 1967. 156,56-63 Provost, P.J. and Hilleman, MR. Propagation of human hepatitis A virus in cell culture in vitro. Proc. Sot. Exp. Viol. Med. 1979, 160, 213-221 Provost, P.J. and Hilleman. MR. An inactivated hepatitis A virus vaccine prepared from infected marmoset liver. Proc. Sot. Exp. No/. Med. 1976,188, 20-203 Provost, P.J., Banker, F.S., Giesa, P.A.. McAleer, W.J., Buynak, E.B. and Hilleman, M.R. Progress toward a live, attenuated human hepatitis A vaccine. Proc. Sot. Exp. No/. Med. 1962, 170, 6-14 Provost, P.J., Conti, P.A., Giesa, P.A. et a/. Studies in chimpanzees of live, attenuated hepatitis A vaccine candidates. froc. Sot. Exp. Biol. Med. 1963,172, 357-363 Provost, P.J., Bishop, R.P., Gerety, R.J. et a/. New findings in live, attenuated hepatitis A vaccine development. J. Med. Viral. 1966,28, 165-175 Provost, P.J., Hughes, J.V.. Miller, W.J., Giesa. P.A., Banker, F.S. and Emini, E.A. An inactivated hepatitis A viral vaccine of cell culture origin. J. Med. Viral. 1966, 19, 23-31 Binn, L.N., Bancroft, W.H., Lemon, S.M. et a/. Preparation of a prototype inactivated hepatitis A virus vaccine from infected cell cultures. J. Infect. Dis. 1966, 153, 749-756 Karron, R.A., Daemer, R., Ticehurst. J. et a/. Studies of prototype live hepatitis A virus vaccines in primate models. J. Infect. Dis. 1966, 187,33-345 Sjogren, M.H., Eckels, K.H., Binn, L.N. et a/. Safety and immunogenicity of an inactivated hepatitis A vaccine. In: Viral Hepatitis and Liver Disease (Ed. Zuckerman, A.J.) Alan R. Liss, New York, 1966, pp. 94-S Mao, J.S., Dong, D.X., Zhang, H.Y. et a/. Primary study of attenuated live hepatitis A vaccine (HZ strain) in humans. J. Infect. Dis. 1969, 158,621-624 Flehmig, B., Heinricy, U. and Pfisterer, M. lmmunogenicity of a killed hepatitis A vaccine in seronegative volunteers. Lancet 1969, I, 1039-1041 Flehmig, B., Heinricy, U. and Pfisterer. M. Simultaneous vaccination for hepatitis A and B. J. Infect. Dis. 1990, 181, 665-666 Midthun, K., Ellerbeck, E., Gershman, K. et a/. Safety and immunogenicity of a live, attenuated hepatitis A virus vaccine in seronegative volunteers. J. Infect Dis. 1991, 183, 735-739 Sjogren, M.H., Hoke, C.H.. Binn, L.N. et a/. lmmunogenicity of an inactivated hepatitis A vaccine. Ann. Intern. Med. 1991,114,470-471 Kusov. Y.Y.. Kazachkov, Y.A., Elbert, L.B. et a/. Characteristics of inactivated hepatitis A vaccine prepared from virus propagated in heteroploid continuous monkey cell line (letter). Vaccine 1990, 8, 513-514 Andre, F.E., Hepburn, A. and D’Hondt, E. Inactivated candidate vaccines for hepatitis A. Prog. Med. Viral. 1990, 37, 72-95 Ellerbeck, E., Lewis, J., Midthun, K. eta/. Safety and immunogenicity of an inactivated hepatitis A virus vaccine. In: Viral Hepatitis and Liver Disease (Eds Hollinger, F.B., Lemon, S.M. and Margolis, H.S.) Williams & Wilkins, Baltimore, 1991, pp. 91-93 Lewis, J.A., Armstrong, M.E., Larson, V.M. et a/. Use of a live, attenuated hepatitis A vaccine to prepare a highly purified formalin-inactivated hepatitis A vaccine. In: Viral Hepatitis and Liver Disease (Eds Hollinger, F.B., Lemon, S.M. and Margolis, H.S.) Williams & Wilkins, Baltimore, 1991, pp. 94-97 Gerety, R.J., Smallwood, L.A., Finlayson, J.S. and Tabor, E. Standardization of the antibody to hepatitis A virus (anti-HAV) content of immunoglobulin. Dev. Biol. Stand. 1963, 84, 411-416 Krah, D.L., Amin, R.A., Nalin, D.R. and Provost, P.J. A simple antigen-reduction assay for the measurement of neutralizing antibodies to hepatitis A virus. J. Infect. Dis. 1991, 163, 634-637 Villarejos, V.M., Serra, C.J., Anderson-Visona, K. and Mosley, J.W. Hepatitis A virus infection in households. Am J. Epidemiol. 1992, 115, 577-586
APPENDIX Previous trials of the Merck hepatitis A vaccine have used doses expressed as 100 to 800 ng of virus antigen, based on a hepatitis A virus antigen radioimmunoassay of the vaccine relative to a reference standard (lot 440), which was purified and characterized in 1982. Relative
Vaccine,
Vol. 10, Issue 10, 1992
671
Inactivated
hepatitis
A vaccine: E.F. Ellerbeck
et
al.
to this standard, the 6.3, 12.5 and 25 ng vaccine doses used in this study contained 100, 200 or 400 ng of virus antigen, respectively. The dosages reported in the present study are based on a new reference standard which was prepared using the current virus purification process. This new reference material has been subjected to extensive analytical testing, including quantitative amino acid analysis to determine protein concentration. Based
672
Vaccine, Vol. 10, Issue 10, 1992
on comparative analyses, including radioimmunoassay, of the purified virus bulk used to prepare the vaccine for this study and our reference standard, it was determined that the 400 ng vaccine dose contained approximately 25 ng of hepatitis A viral protein. Other doses contained proportionally analogous amounts of viral protein, e.g. 200 ng (old dosage estimate) is equivalent to 12.5 ng of hepatitis A viral protein.
INTRODUCTION Hepatitis due to infection with the hepatitis A virus (HAV) remains an important health problem worldwidel. Although immune globulin can provide shortterm protection for high risk individuals2-7, the need for repeat dosing every 6 months makes it impractical for long term prophylaxis. A vaccine that induces longlasting immunity would clearly be preferable. The propagation of HAV in cell culture in 1979* made the development of a vaccine a realistic possibility. Since then, both live-attenuated and inactivated HAV vaccines have been tested in animals and human volunteers*-23. These vaccines have generally been well tolerated, and some have been shown to protect primates against challenge with virulent HAV strains*-r4. A purified, formalin-inactivated, alum-adjuvanted vaccine prepared from HAV strain CR326 protected marmosets against challenge with virulent HAVi3. A live-attenuated vaccine *Center for Immunization Research, School of Hygiene and Public Health, and +Department of Medicine, School of Medicine, Johns Hopkins University, 624 N. Broadway, Room 125, Baltimore, MD 21206, USA. *Merck Sharp & Dohme Research Laboratories, Sunneytown Pike, West Point, PA 19486, USA. §To whom correspondence should be addressed. (Received 10 December 1991; revised 11 February 1992; accepted 14 February 1992) 0264410x/92/1-5 0 1992 Buttetworth-Heinemann 668
Ltd
Vaccine, Vol. 10, Issue 10, 1992
candidate derived from the F’ variant of HAV strain CR326F was well tolerated and immunogenic in human volunteers’2*20 and also protected marmosets against challenge with virulent HAVr2. We now report the results of a dose-escalating study of a formalin-inactivated, alum-adjuvanted vaccine prepared from the attenuated F’ variant of HAV strain CR326F and tested in healthy seronegative volunteers. Preliminary results of this study have been reported previously24. MATERIALS
AND METHODS
Hepatitis A virus vaccine strain
The HAV used to make the inactivated vaccine was derived from the master seed of the live, attenuated F’ HAV variantI by carrying out five additional serial passages of the virus in MRC-5 cells at 35°C and eight further passages of the virus in MRC-5 cells at 32°C. Vaccine was prepared from HAV contained in infected cells of the final passage. Hepatitis A antigen purification
HAV virus was purified from virus-infected MRC-5 cell pellets harvested from roller bottles by methods conventionally used in picornavirus purification, including cell lysis, organic extraction, and ion exchange and size-exclusion chromatography.
Inactivated
hepatitis
A vaccine: E.F. Ellerbeck
et al.
Hepatitis A antigen inactivation
Detection of antibodies to HAV
Purified HAV antigen in phosphate-buffered saline was inactivated by exposure to 100 pg ml-’ formalin at 37°C for 20 days with continuous stirring. At 1 and 72 h after the addition of formalin, bulk antigen was transferred from one glass vessel to another to ensure complete exposure of antigen to formalin. At 20 days, the bulk vaccine antigen was passed through a 0.2 pm filter and adsorbed to aluminium hydroxide. Thimerosal was added to 0.005%. The vaccine was stored at 4°C. A 10 pg aliquot of bulk vaccine antigen, withdrawn after 20 days of exposure to formalin, was dialysed to remove residual formalin. MRC-5 cell monolayers established in ten T25 flasks were each infected with a 1 pg aliquot of the dialysed vaccine antigen. The monolayers were refed at 7-day intervals and harvested on the 35th day by scraping. The cell harvest from each flask was lysed by freeze-thawing and sonication, and 200 ~1 aliquots of each flask were used to inoculate ten additional T25 flasks which were maintained as described above. After an additional 3 days, the cells were again harvested, lysed, and assayed for hepatitis A antigen by radioimmunoassay (RIA) I3 . The absence of detectable antigen documented complete inactivation of the virus. Further details on the preparation of this vaccine have previously been published2’.
A standard HAVAB (Abbott Laboratories, N. Chicago, IL) and HAVAB and HAVAB-M assays modified to increase their sensitivity were used to measure HAV antibody. In the modified HAVAB assay, a tenfold greater volume of serum (test or control) was tested (0.1 mlofserainO.l mlof 1251-labelled anti-HAV instead of 0.01 ml of sera in 0.2 ml of ‘251-labelled anti-HAV). Results were converted to International Units by comparing them with a reference curve generated from World Health Organization (WHO) Reference Globulin no . 12’j. The modified HAVAB-M assay tested 0.01 ml of undiluted serum instead of 0.01 ml of a 1:200 dilution. Neutralizing antibody titres were determined by the hepatitis A virus antigen-reduction neutralization assay (HAVARNA) described by Krah et a1.27.
Subjects Fifty-six healthy adults (22 males and 34 females), 18-50 years old, were recruited from the Baltimore metropolitan area. All volunteers had a normal physical examination and normal serum levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST) and gamma glutamyl transpeptidase (GGT). They were negative for hepatitis B surface antigen and HAV antibody. Persons at increased risk of HAV infection (consumption of raw shellfish or travel to HAV endemic countries) were excluded from this study. All subjects gave informed, written consent to participate in this study, which was approved by the institutional review boards of the participating institutions. Experimental design HAV vaccine, at a concentration of 12.5 ng ml-‘, was tested at doses of 6.3, 12.5 and 25 ng (see Appendix). At each dose, 14 volunteers were randomly assigned, in a single-blind fashion, to receive a single intramuscular (deltoid) injection of either the HAV vaccine (n = 10) or placebo (n = 4) at 0, 4 and 24 weeks. Fourteen volunteers received 6.3 ng of HAV vaccine (n = 10) or placebo (n = 4) at 0, 2 and 24 weeks. Volunteers recorded their temperatures, injection site reactions and other complaints for a total of 5 days following each inoculation. Serum ALT, AST and GGT were measured 2, 4, 5, 24 and 28 weeks after the initial vaccination. HAV. serology was measured in all volunteers at weeks 2-6, 8, 12, 20, 24, 28, 32 and 40. Individuals who received HAV vaccine at 0, 4 and 24 weeks were tested for neutralizing antibody at weeks 2, 4, 6 and 32. Endpoint titrations of neutralizing antibody were measured at week 32 in 12.5 ng vaccinees. After 24 months, vaccinees were asked to return to have their HAV serology measured again.
RESULTS Fifty-three (95%) of the 56 volunteers completed the study. One 6.3 ng vaccinee, one 12.5 ng vaccinee, and one placebo recipient dropped out of the study after 6 months of follow-up for reasons unrelated to vaccination. IgM antibody was detected by the modified HAVABM in all but one of the vaccinees within 4 weeks of the first vaccination. All vaccinees had detectable IgM antibody to HAV within 1 week after the second dose. By 4 weeks after a single 6.3, 12.5 or 25 ng injection given to ten vaccinees each, seven, nine and ten vaccinees, respectively, had antibody detectable by modified HAVAB. All vaccinees had seroconverted by week 6 (2 weeks after the second inoculation). Antibody levels increased approximately 20-fold in all dose groups after the 6-month boost (Figure 1). The magnitude of the antibody response was dose dependent, with higher antibody levels in recipients of higher doses of vaccine (p = 0.05, linear regression of log titres, week 32). Of the 6.3 ng vaccinees, ten received their second inoculation at week 2 and ten at week 4. All vaccinees boosted at week 2 seroconverted by week 4 as compared with seven of the 6.3 ng vaccinees who had not yet received their second inoculation. Geometric mean antibody levels from week 8 to 24 were similar in those receiving their boost at 4 weeks (GMT range: 50-l 13 mIU) or at 2 weeks (GMT range: 70-90 mIU) (Figure 2). After the third immunization (week 24), geometric mean antibody levels at week 32 were not significantly different in volunteers receiving 2-week -
‘I
7i
10000
1
z
1000
L
1
0
8
I
16 Weeks
after
24
L
32
40
vaccination
Figure 1
Geometric mean antibody responses by modified HAVAB in volunteers vaccinated with 6.3 (IJ). 12.5 (0) or 25 (0) ng of inactivated hepatitis A vaccine at 0, 4 and 24 weeks
Vaccine, Vol. 10, Issue
10, 1992
669
Inactivated 7_
hepatitis
A vaccine: E.F. Ellerbeck
et al. Table 2 Percentage of hepatitis A vaccine and placebo recipients reporting pain or tenderness at the injection site during any of 4 days after the day of vaccination
10000 c
Injection
x
I”
11
’
0
16
24
32
40
vaccination
Figure 2 Geometric mean antibody responses by modified HAVAB in volunteers vaccinated with 6.3 ng of inactivated hepatitis A vaccine at either (0) 0, 4 and 24 weeks or (a) 0, 2 and 24 weeks
Table 1 Antibody titres to hepatitis A virus by modified hepatitis A virus antibody (HAVAB) and neutralization assays at week 32 after three 12.5 ng injections of hepatitis A vaccine at 0, 4 and 24 weeks Anti-HAV titre by
134 136 137 136 139 141 143 146 147
Modified HAVAB (mlU ml-‘) 2942
1653 5296 10242 12506 6151 9365 3873 20583
Neutralization assay” 2046 4096 4096 6192 16364 6192 8192 8192 > 3276%
“Titre expressed as reciprocal of highest serum dilution with neutralizing antibody
(5464 mIU) or 4-week (3425 mIU) boosts (p = 0.18, Student’s t test). Among recipients of 6.3, 12.5 or 25 ng doses (n = 10 each), hepatitis A antibody was detected by the standard HAVAB in one, two and eight vaccinees, respectively, within 4 weeks of a single dose of vaccine. After three doses of vaccine, all HAV vaccinees had seroconverted by standard HAVAB. Individuals vaccinated at weeks 0, 4 and 24 were screened for neutralizing antibody. Neutralizing antibody, in recipients of 6.3, 12.5 or 25 ng doses (n = 10 each), was detected 2 weeks after the first inoculation in seven, nine and seven vaccinees, respectively. By 4 weeks after a single inoculation, all 30 vaccinees tested had detectable neutralizing antibody. At week 32, neutralizing antibody titres ranged from 1: 2048 to > 1: 32768 in the 12.5 and 25 ng vaccine.es and from 1:64 to > 1:4096 in the 6.3 ng vaccinees. End-point titrations of neutralizing antibody at week 32 in 12.5 ng vaccinees were highly correlated with modified HAVAB antibody levels ( Table 1; Pearson correlation coefficient (I) = 0.93, p < 0.001). Nineteen vaccinees were available for follow-up 24 months after their initial vaccination. All of these vaccinees had HAV antibody detectable by the modified HAVAB assay. HAV antibody levels ranged from 42-2756 mIU (GMT = 522 mIU) for 6.3 ng vaccinees (n = 11 ), 309-2498 mIU (GMT = 1062 mIU) for 12.5 ng
670
Vaccine,
1st
2nd
3rd
Placebo 6.3 12.5 25
19 15 0 40
13 20 10 50
33 20 11 50
(16) (20) (10) (10)
I
8
Weeks after
Subject no.
Dose” (ng) (n)
Vol. 10, Issue 10, 1992
“Volume of inoculum varied with dose: 6.3 ng (0.5 ml), 12.5 ng (1.0 ml) and 25 ng (2.0 ml). Placebo recipients received 0.5-2.0 ml depending upon the cohort in which they were enrolled
vaccinees (n = 4), and 646-7627 mIU (GMT = 1701 mIU) for 25 ng vaccinees (n = 4). No significant systemic reactions were noted after vaccination. Transient ALT elevations ( >40 IU) were noted in two of 40 HAV vaccinees (range 43-52 IU) and one of 16 placebo recipients (70 IU). Subjects reported symptoms of headache, dizziness, dyspepsia, myalgias or other mild non-localized complaints during the 5 days following three (6% ) of 47 placebo inoculations and eight (14%)of6.3ng(n=59),four(14%)of12.5ng(n=29) and two (7% ) of 25 ng (n = 30) HAV inoculations. None of these symptoms interfered with the usual daily activities of the volunteers. Injection site pain and tenderness persisted beyond the day of vaccination in O-50% of HAV vaccinees and 13-33% of placebo recipients in the different groups (Table 2). These reactions were reported more frequently in the 25 ng HAV vaccine group (40-50%) than in the 6.3 and 12.5 ng HAV vaccine groups (O-20%). The volume of inoculum, however, also varied with the dose. The three placebo recipients who received the same volume of inoculum as the 25 ng vaccine recipients reported pain or tenderness beyond the day of injection after five (62% ) of eight inoculations.
DISCUSSION This inactivated HAV vaccine was highly immunogenic at doses of 6.3 to 25 ng. The magnitude of the antibody response was dose-dependent, with higher doses of vaccine inducing higher antibody levels and earlier seroconversion. Low levels of neutralizing antibody were detected 2-4 weeks after a single 6.3, 12.5 or 25 ng dose of vaccine. Four weeks after vaccination, neutralizing antibody titres were equal to or greater than those found in immune globulin recipients 3 months after a 0.02 ml kg- ’ inoculation (unpublished data), a time when immune globulin recipients are still protected from Early seropositivity in HAV infection by HAV2-‘. vaccinees may be crucial for protecting travellers leaving on short notice to HAV endemic countries. The 6-month boost induced a 15-30-fold increase in HAV antibody levels. Neutralizing antibody titres after this boost were comparable to or higher than those reported in previous studies of inactivated HAV vaccines18,1g,21. Although antibody titres were substantially higher than those seen in a trial of the live, attenuated HAV vaccine derived from the F’ strain of CR326FZo, the persistence of the immune response to both inactivated and live attenuated HAV vaccines remains to be determined. The antibody response seen
Inactivated
with a 0, 4 and 24 week dosing schedule suggests that vaccination schedules for HAV could be coordinated with hepatitis B immunization schedules or that the HAV vaccine might be formulated in combination with hepatitis B vaccine”. Using 25 ng of vaccine, a two-dose vaccination schedule may also be possible. After a single 25 ng vaccination, antibodies can persist up to 6 months and a boost given at that time can induce antibody levels comparable to those seen after the third vaccine dose in our study (Shouval et al., unpublished results). The optimal dose and schedule of vaccination may depend upon the rapidity with which seroconversion is desired, compatibility with other vaccine schedules, and vaccine cost. This HAV vaccine provides a margin of safety not found in most other inactivated vaccines in that it is derived from an attenuated strain of HAV previously shown to be well tolerated without evidence of hepatotoxicity in human volunteers’2*20. All dose levels of the inactivated vaccine were well tolerated. The mild local reactions reported more frequently by both vaccine and placebo recipients in the 25 ng cohort were due probably to the two to four times larger volume of injection given to this cohort. The total of the different types of non-localized complaints reported was higher in HAV vaccinees than in placebo recipients, but these could not be linked to vaccination; the complaints were mild, unrelated to dose, and no particular type of symptom predominated. This inactivated HAV vaccine can induce high levels of antibodies detectable by both HAVAB and virus neutralization assays. Long-term follow-up of volunteers from this and other studies is needed to assess the duration of the immune response, and to determine whether, as in one study of natural disease, seroconversion with intact immune memory indicates long-lasting protection from disease regardless of the antibody titre 27. Additional studies are warranted to evaluate the safety of this vaccine in a larger series and to establish protective vaccine efficacy.
7
6
9
IO
11
12
13
14
15
16
17
16
19 20
21 22
23
ACKNOWLEDGEMENTS This work was supported by Merck Sharp & Dohme Research Laboratories. The authors are grateful to Janine Sherman at the Johns Hopkins University and Mary McCaughtry at Merck Sharp & Dohme Research Laboratories for assistance with this study.
24
25
26
REFERENCES 27 Lemon, SM. Type A viral hepatitis: new developments in an old disease. New Engl. J. Med. 1965, 313, 105%1067 Stokes, J. Jr and Neefe, JR. The prevention and attenuation of infectious hepatitis by gamma globulin. J. Am. Med. Assoc. 1945, 127,144-145 Hall, W.T., Madden, D.L., Mundon, F.K., Brandt, D.E.L. and Clarke, N.A. Protective effect of immune serum globulin (ISG) against hepatitis A infection in a natural epidemic. Am. J. Epidemiol. 1977. 106, 72-75 Woodson, R.D. and Clinton, J.J. Hepatitis prophylaxis abroad: effectiveness of immune serum globulin in protecting Peace Corps volunteers. J. Am. Med. Assoc. 1969, 288, 1053-1056 Havens, W.P. and Paul, J.R. Prevention of infectious hepatitis with gamma globulin. J. Am. Med. Assoc. 1945, 128, 270-272 Stokes, J. Jr, Farquhar, J.A., Drake, M.E., Capps, R.B., Ward, C.S. Jr and Kitts, A.W. Length of protection by immune serum globulin
26
hepatitis A vaccine: E.F. Ellerbeck et al.
(gamma globulin) during epidemics. J. Am. Med. Assoc. 1951,147, 714-719 Conrad, ME. and Lemon, SM. Prevention of endemic icteric viral hepatitis by administration of immune serum gamma globulin. J. Infect. Dis. 1967. 156,56-63 Provost, P.J. and Hilleman, MR. Propagation of human hepatitis A virus in cell culture in vitro. Proc. Sot. Exp. Viol. Med. 1979, 160, 213-221 Provost, P.J. and Hilleman. MR. An inactivated hepatitis A virus vaccine prepared from infected marmoset liver. Proc. Sot. Exp. No/. Med. 1976,188, 20-203 Provost, P.J., Banker, F.S., Giesa, P.A.. McAleer, W.J., Buynak, E.B. and Hilleman, M.R. Progress toward a live, attenuated human hepatitis A vaccine. Proc. Sot. Exp. No/. Med. 1962, 170, 6-14 Provost, P.J., Conti, P.A., Giesa, P.A. et a/. Studies in chimpanzees of live, attenuated hepatitis A vaccine candidates. froc. Sot. Exp. Biol. Med. 1963,172, 357-363 Provost, P.J., Bishop, R.P., Gerety, R.J. et a/. New findings in live, attenuated hepatitis A vaccine development. J. Med. Viral. 1966,28, 165-175 Provost, P.J., Hughes, J.V.. Miller, W.J., Giesa. P.A., Banker, F.S. and Emini, E.A. An inactivated hepatitis A viral vaccine of cell culture origin. J. Med. Viral. 1966, 19, 23-31 Binn, L.N., Bancroft, W.H., Lemon, S.M. et a/. Preparation of a prototype inactivated hepatitis A virus vaccine from infected cell cultures. J. Infect. Dis. 1966, 153, 749-756 Karron, R.A., Daemer, R., Ticehurst. J. et a/. Studies of prototype live hepatitis A virus vaccines in primate models. J. Infect. Dis. 1966, 187,33-345 Sjogren, M.H., Eckels, K.H., Binn, L.N. et a/. Safety and immunogenicity of an inactivated hepatitis A vaccine. In: Viral Hepatitis and Liver Disease (Ed. Zuckerman, A.J.) Alan R. Liss, New York, 1966, pp. 94-S Mao, J.S., Dong, D.X., Zhang, H.Y. et a/. Primary study of attenuated live hepatitis A vaccine (HZ strain) in humans. J. Infect. Dis. 1969, 158,621-624 Flehmig, B., Heinricy, U. and Pfisterer, M. lmmunogenicity of a killed hepatitis A vaccine in seronegative volunteers. Lancet 1969, I, 1039-1041 Flehmig, B., Heinricy, U. and Pfisterer. M. Simultaneous vaccination for hepatitis A and B. J. Infect. Dis. 1990, 181, 665-666 Midthun, K., Ellerbeck, E., Gershman, K. et a/. Safety and immunogenicity of a live, attenuated hepatitis A virus vaccine in seronegative volunteers. J. Infect Dis. 1991, 183, 735-739 Sjogren, M.H., Hoke, C.H.. Binn, L.N. et a/. lmmunogenicity of an inactivated hepatitis A vaccine. Ann. Intern. Med. 1991,114,470-471 Kusov. Y.Y.. Kazachkov, Y.A., Elbert, L.B. et a/. Characteristics of inactivated hepatitis A vaccine prepared from virus propagated in heteroploid continuous monkey cell line (letter). Vaccine 1990, 8, 513-514 Andre, F.E., Hepburn, A. and D’Hondt, E. Inactivated candidate vaccines for hepatitis A. Prog. Med. Viral. 1990, 37, 72-95 Ellerbeck, E., Lewis, J., Midthun, K. eta/. Safety and immunogenicity of an inactivated hepatitis A virus vaccine. In: Viral Hepatitis and Liver Disease (Eds Hollinger, F.B., Lemon, S.M. and Margolis, H.S.) Williams & Wilkins, Baltimore, 1991, pp. 91-93 Lewis, J.A., Armstrong, M.E., Larson, V.M. et a/. Use of a live, attenuated hepatitis A vaccine to prepare a highly purified formalin-inactivated hepatitis A vaccine. In: Viral Hepatitis and Liver Disease (Eds Hollinger, F.B., Lemon, S.M. and Margolis, H.S.) Williams & Wilkins, Baltimore, 1991, pp. 94-97 Gerety, R.J., Smallwood, L.A., Finlayson, J.S. and Tabor, E. Standardization of the antibody to hepatitis A virus (anti-HAV) content of immunoglobulin. Dev. Biol. Stand. 1963, 84, 411-416 Krah, D.L., Amin, R.A., Nalin, D.R. and Provost, P.J. A simple antigen-reduction assay for the measurement of neutralizing antibodies to hepatitis A virus. J. Infect. Dis. 1991, 163, 634-637 Villarejos, V.M., Serra, C.J., Anderson-Visona, K. and Mosley, J.W. Hepatitis A virus infection in households. Am J. Epidemiol. 1992, 115, 577-586
APPENDIX Previous trials of the Merck hepatitis A vaccine have used doses expressed as 100 to 800 ng of virus antigen, based on a hepatitis A virus antigen radioimmunoassay of the vaccine relative to a reference standard (lot 440), which was purified and characterized in 1982. Relative
Vaccine,
Vol. 10, Issue 10, 1992
671
Inactivated
hepatitis
A vaccine: E.F. Ellerbeck
et
al.
to this standard, the 6.3, 12.5 and 25 ng vaccine doses used in this study contained 100, 200 or 400 ng of virus antigen, respectively. The dosages reported in the present study are based on a new reference standard which was prepared using the current virus purification process. This new reference material has been subjected to extensive analytical testing, including quantitative amino acid analysis to determine protein concentration. Based
672
Vaccine, Vol. 10, Issue 10, 1992
on comparative analyses, including radioimmunoassay, of the purified virus bulk used to prepare the vaccine for this study and our reference standard, it was determined that the 400 ng vaccine dose contained approximately 25 ng of hepatitis A viral protein. Other doses contained proportionally analogous amounts of viral protein, e.g. 200 ng (old dosage estimate) is equivalent to 12.5 ng of hepatitis A viral protein.
