Table 1. Representative Reports of Extrahepatic Adverse Reactions to Hepatitis B virus infection.

Adverse Reaction/Diagnosis Reference

Systemic "Lupoid hepatitis", Systemic lupus erythematosus Borisova and Krel, 1992;
Chng et al., 1993;
Arthritis (polyarthritis, rheumatoid arthritis) McCarty and Ormiste, 1973;
Gocke, D., 1975;
Duffy et al., 1976;
Onion et al., 1971;
Wands et al., 1975;
Chistau and Helin, 1987;
Morris and Stevens, 1978;
Pease and Keat, 1985;
Tsukada et al., 1987;
Vascular Disorders (Vasculitis, polyarteritis, erythema nodosum) Gocke, D., 1975;
Sargent et al., 1976;
Duffy et al., 1976;
repo et al., 1974;
Michalak, 1977;
Maggiore, 1983;
Di Giusto and Bernhard, 1986;
Tsukada et al., 1987;
Rogerson and Nye, 1990;
Guillin Barre Syndrome Neirmeijer and Gips, 1975;
Penner et al., 1982Tsukada et al., 1987;
Tabor et al., 1987;
Demyelinating disorders (optic neuritis, demyelinating neouropathy etc.) Galli et al., 1986;
Tsukada et al., 1987;
Inoue et al., 1994;
Achiron, 1994;
Chronic Fatigue Berelowitz et al., 1995;
Glomerulonephritis Venkataseshan et al., 1990.

2. Role of MHC genes in autoimmune disorders.

There is substantial evidence that there are strong associations between autoimmune disorders and MHC molecules (See reviews by Paul, 1987; Abbas, 1994). In systemic lupus erythematosus (SLE) the presence of HLA-DR2 or HLA-DR3 haplotypes is associated with a relative risk which is doubled if both are present (Mackworth-Young and Schwartz (1988). Rheumatoid arthritis (RA) is a chronic inflammatory polyarthritis which has a strong association with MHC Class II molecules, although the concordance of disease occurs at only 15% in monozygotic twins suggesting that environmental factors may also play a role in the onset of disease in genetically susceptible individuals (Silman et al., 1993). Of patients having type I diabetes, about 95% have HLA-DR3 or DR4 haplotypes or both as compared to 40% of the general population (Rotter et al., 1983).

It has also been shown that cytotoxic T lympocytes recognize an HLA-A2-restricted epitope within the hepatitis B virus nucleocapsid antigen (Penna et al., 1991). More recently it has been shown that MHC class I-restricted responses as well as class II restricted responses may also be involved in the pathogenesis of demyelinating disorders (Pelfrey et al., 1993; Tsuchida et al., 1994). In multiple sclerosis, myelin proteins are thought to be the targets for auto-reactive T-cell responses. In studies by Tsuchida et al. (1994) it was found that self peptides derived from human myelin proteins (including sequences from human myelin basic protein, proteolipid protein, myelin associated glycoprotein and myelin oligodendrocyte glycoprotein) bind to and form stable complexes with HLA-A2. These studies are important because they demonstrate that self peptides from human myelin proteins can induce auto-reactive CD8 cytoxic lymphocytes and that these lymphocytes produce cytokines thought to be important in mediating demyelinating diseases.

3. Role of MHC genes in immune response to HBsAg of virus and vaccine.

It has long been known that there is a genetic correlation of the immune response with respect to the hepatitis B surface antigen (Milich et al., 1982, 1983). It has been clearly documented that the human antibody response to the hepatitis B surface antigen (HBsAg) vaccine is associated with the major histocompatability complex (MHC) and is inherited in a dominant fashion (Craven et al., 1985; Kramer et al., 1988; Alper et al., 1989; Varla-Leftherioti et al., 1990; Kruskall et al., 1992). It has further been reported that 5 to 10% of healthy individuals fail to respond ("non-responders") to the plasma-derived HBV vaccine (Weissman et al., 1988, Kramer et al., 1988). Because the plasma derived vaccine contains a distinct glycosylation pattern from that of the yeast-produced vaccine, these investigators carried out studies to evaluate whether the recombinant, yeast- produced vaccine would produce antibody titers in the non-responder population. These studies demonstrated that the recombinant hepatitis B vaccine (Recombivax HB) in non-responders to the plasma derived vaccine and that HLA subtyping showed a high prevalence of DR7, B8, and the combinations of DR3, DR4 and DR7.

Additional studies by Margot et al. (1992) indicate that:

1. the response to the HBsAg vaccine is MHC-linked and inherited in a dominant fashion,


2. that an abnormal or missing immune response (Ir) gene for HBsAg is a characteristic of most examples of the extended haplotype (HLA-B8, SCO1, DR3), and


3. other haplotypes also have abnormal or missing Ir genes for HBsAg.

Sktachowski, et al., (1995) have also shown that that responder groups can be divided into two subgroups: low responders and high responders. In these studies marked differences were observed between responders and non-responders in the occurrence of carriers of different MHC class I, II and III alleles. High responders were found to have different haplotypes than low responders. These findings indicate that amounts of antibody to HBsAg is genetically influenced even in patients demonstrating adequate antibody response.

Many of these studies have been summarized by Abbas (1994). His work indicates that Caucasians who are homozygous for an extended HLA haplotype containing HLA B8, DR3, Dqw2a are low responders to the HBsAg. In this review he proposed that individuals who are heterozygous for this locus are high responders presumably because the other alleles contain one or more HLA genes that confer responsiveness. Abbas et al therefore concludes that HLA typing may prove to be valuable for predicting the success of this vaccine (Abbas et al., 1994) . This proposal therefore hypothesizes that HLA typing may also prove to be valuable for predicting the failure of the vaccine, including non-response or autoimmune side effects.

D. Association of the hepatitis B surface antigen vaccine and autoimmune syndromes.

Although there have been numerous reports of the efficacy of the hepatitis B plasma and recombinant vaccines (Mast and Alter, 1983; Hollinger et al., 1986; Zahrandnik et. al., 1987; McMahon and Wainright, 1993), these reports concentrate primarily on antibody titers (e.g. responders and non-responders); and no long term follow-up has been reported. The clinical trials reported in the drug inserts cite 4 or 5 day follow-ups, which could be expected to be too short an observation period to observe long term autoimmune responses. Table 2 includes a representation of published reports of such adverse reactions although it does not include summation of the tens of thousands of reactions reported to the FDA Adverse Reaction Reporting System. We have included 12 pages representative of 900 total pages of those reaction reports (Merck vaccine only) (See Appendix 1). Note: one listing is a report by a registered nurse of 15 incidences of multiple scleroses; she inquires whether there is any known relationship with this vaccine.

To date, the most detailed study which has demonstrated a correlation of rheumatoid arthritis (RA) with the hepatitis B vaccine and a relationship to HLA subtypes is that of the two Co-P.I.s of this project . RA is not listed as a potential side effect for this vaccine in the Physician Desk Reference or package insert for the Energix (Smith Kline) vaccine. This study is summarized in the Progress Report and is provided in complete form in Appendix 2.

Table 2. Reports of Adverse Reactions of Hepatitis B Vaccine (does not include the FDA Adverse Reactions Reports).

Adverse Reaction/Diagnosis Reference

Systemic "Lupoid hepatitis", Lupus erythematosus See Appendix 2;
Tudela and Bonal, 1992;
Mamoux and Dumont, 1994;

Arthritis (polyarthritis, rheumatoid arthritis) Rogerson and Nye, 1990;
Vautier and Carty; 1993;
Gross et al., 1995;
Pope et al., 1995;
(see Appendix 2 and 3);
Biasi, 1987-1993.

Vascular Disorders (Vasculitis, polyarteritis, erythema nodosum, cryoglobulinemia) DiGuisto and Bernhard. (1986);
Goolsby, 1989;
Cockwell et al., 1990;
Rogerson and Nye, 1990;
Mathieu et al., 1996;
Caarmeli and De-Medina, 1993;
Isla, 1993;
Mathieu and Krivitsky, 1996.

Guillain Barre Syndrome Shaw et al., 1988;
Demyelinating disorders (optic neuritis, Bells Palsy, demyelinating neouropathy, multiple schlerosis etc.) Ribera and Dukta, 1983;
Shaw et al., 1988;
Herroelen et al., 1991;
Nadler (1993);
Devin et al., 1996;
Dunbar et al., (unpublished observations); Senejoux et al., 1996;
See Appendix 2; Baglivo et al., 1996;.

Diabetes mellitus
Poutasi, 1996;
Classen, 1996;

Chronic Fatigue I. Salit (1993);
Delage et al., 1993;
Other Germanaus et al., 1995.

The majority, if not all, of these reported side effects of the recombinant hepatitis B vaccine are the same as or similar to those reported as extra-hepatic manifestations of the virus infection itself. These severe adverse side effects are also associated with autoimmune responses. Some of these reported adverse reactions have been dismissed (e.g. Canadian Laboratory Centre for Disease Control report, Delage et al., 1992) because:

1. there was no clear pattern of the onset of symptoms and


2. there was no biological evidence to support the occurrences of adverse reactions (i.e. no circulating virus).

In view of substantial new information that autoimmune disease can be induced by viral molecular mimicry, anti-idiotypic antibodies, or anti-phospholipid antibodies (see discussion below), it is apparent that the dismissal of these reactions as a result of vaccination might well have been premature.

In the text "Adverse Events of Childhood Vaccines", (1993) Hauser et al. (1987), state: "The antibodies produced after infection with hepatitis B virus or after administration of plasma derived vaccine or recombinant vaccine are alike in terms of their ability to elicit protective determinants that are active against all subtypes of the virus..." and that "the results of the trials of recombinant vaccine are much the same as those of trials of the plasma-derived vaccine". They further stated that the studies were not designed to assess serious, rare adverse events, the total number of recipients were too small and the follow-up generally too short to detect rare or delayed, serious, adverse reactions. Finally it was pointed out that "overall the number of examples of adverse neurologic outcomes following receipt of hepatitis B vaccine are of concern, particularly those resulting in demyelinating neurologic disease".

In view of these observations and the more recent observations outlined in Table 2, it is medically crucial to evaluate the nature of the autoimmune reactions (i.e. risks) associated with the hepatitis B vaccine and to determine if individuals who will have these adverse reactions can be identified in advance of receiving the vaccine.

E. Possible molecular mechanisms for adverse reactions to the hepatitis B vaccine.

1. Molecular mimicry:

There has long been an awareness that viruses have developed evasion strategies by which they can successfully circumvent immune detection and/or effect. There have been recent studies of the molecular mechanisms that viruses use to circumvent the immune system (Lidbury, 1994). Molecular mimicry, which is one of the proposed mechanisms, has been characterized as the presence of one or more common epitopes, either linear or conformational, shared by host and microbial determinants (Dyrberg and Oldstone, 1986; Olesak, 1994).

The theory that molecular mimicry between viral and self antigens could, in some instances, initiate autoimmunity has gained increased acceptance in the past few years (Fujinami and Oldstone, 1985; Jahnke et al., 1985; Fujinami, 1988; Olesak, 1994; Wucherpfennig and Strominger, 1995; Gianani and Sarvetnick, 1996; Baughan et al., 1995). One study has provided evidence that while only one peptide could have been identified as a molecular mimic by sequence alignment, seven viral and one bacterial peptide efficiently activated T cell responses in cells isolated from patients with multiple sclerosis (Wucherpfennig and Strominger, 1995). These authors conclude that the diverse nature of the molecular mimicry peptides and the ubiquitous presence of some of these pathogens make it difficult to establish a direct epidemiological link between these viral infections and the occurrence of multiple sclerosis. These authors also conclude that: "Genetic modifications of viral vaccines that eliminate proven mimicry epitopes could make viral vaccines safer and reduce the frequency of post-vaccinal encephalomyelitis".

With respect to the hepatitis B virus, it has been shown by Fujinami and Oldstone (1985) that when computer aided analysis identified peptide sequences shared between the viral protein and myelin basic protein (MBP), these peptides could stimulate a T cell proliferative response directed to MBP and result in autoimmune central nervous system disease in rabbits. In fact, this study of the hepatitis B protein was used as the basis for the molecular mimicry model of autoimmune disease (Barnett et al., 1993).

It has also been shown that the conformational nature of peptides may be important for conferring molecular mimicry (Madden et al., 1993). It is clear that detailed molecular and biochemical analyses need to be carried out to identify specific peptides which might be acting as molecular mimics. The advances in computer technology and data bases now provide the tools necessary to design experiments to evaluate these hypotheses directly.

2. Anti-idiotypic antibodies.

Anti-idotypic antibodies are associated with autoimmune disorders including myasthenia gravis, diabetes, systemic lupus and Grave's disease (Nasu, et al., 1982; Rauch et al., 1985; Dwyer et al., 1983,1987; Sikorska, 1986; Schoelson et al., 1986), as has been demonstrated directly in experimental animal models (Shoenfeld, 1994; Blank et al., 1995). It has been proposed from these results that in some autoimmune diseases, especially in those in which the presumed auto-antigen is not immunogenic (e.g. DNA, cardiolipin), that antibodies against the infecting agent may carry a pathogenic idiotype of a specific auto-antibody (Schoenfeld, 1994). The latter author has further hypothesized that, if a subject is prone to autoimmunity (e.g. genetic, hormonal), the pathogenic idiotype will progress in dysregulating the immune system resulting in a clinically overt autoimmune disease.

Autoantibodies that are anti-idiotypic to anti-viral antibodies have also been observed (Plotz, 1983). Furthermore, anti-idiotypic reagents that bear an internal image capable of mimicking the hepatitis B surface antigen have been used to induce an antibody response to HBsAg in both rabbits and chimpanzees (Kennedy et al., 1986). Also in vivo injections of anti-idiotypic antibodies have been used to prime the immune response of mice to HBsAg (Kennedy et al., 1984). Given the correlation of anti-idiotypic antibodies with antigens known to be associated with molecular mimicry, the probability that such antibodies are involved in the adverse reactions associated with hepatitis B vaccination appears high and needs to be addressed in detail as is proposed in this study.

c. Anti-phospholipid antibodies.

Recently, anti-phospholipid antibodies have been shown to be associated with clinical syndromes such as SEL and thrombosis (Puurunen et al., 1996), recurrent abortion and thrombocytopenia. Other reports include a relationship to multiple sclerosis (Sugiyama and Yamamoto, 1996). It has been shown recently that there is an increased incidence of anti-cardiolipin antibodies in patients having autoimmune thyroid diseases, although these autoantibodies are thought to represent a non-specific marker of immune dysregulation. Other studies have described autoimmune phenomena in which anti-phospholipid antibodies are associated with "anti-phospholipid syndrome" (Lekarstvi, 1994), and there has been a report of a patient with mixed types of chronic active hepatitis and primary bilary cirrhosis having this syndrome (Saeki et al., 1993). In addition, it has been shown that anti-phospholipid antibodies in women with recurrent fetal loss correlate to clinical and serological characteristics of SLE (Bagger et al., 1993, Zurgil et al., 1993). One of the patients who will be participating in this study had recurrent miscarriages following the onset of adverse reactions to the hepatitis B vaccine and had noted anti-cardiolipin antibodies (See Appendix 2) A proposed mechanism for anti-phospholipid antibodies involved in autoimmune diseases has been derived from observations that heparin sulfate is a high affinity antigen of pathological significance for anti-DNA and anti-phospholipid antibodies. Interference due to binding of these antibodies with components of the cell surface and heparin sulfate, which is an extracellular matrix molecule, could play a major role in tissue injury including the vasculature system (Shibata et al., 1993). As vascular problems are a common adverse effect of the HBsAg, this will be an important parameter to evaluate.

The hepatitis B surface antigen binds to a human liver plasma membrane protein, endonexin II, a calcium dependent phospholipid binding protein (deBruin et al., 1996). This protein has further been shown to be associated with numerous tissue cell types. They conclude that the species specific distribution of the HBsAg binding protein correlates with the species tropism of hepatitis B virus infection. Because this protein is not present in species which are not infected with this virus, the binding to this surface molecule might be important in the mechanisms of cellular internalization of this HBsAg. There is also a correlation of anti-phospholipid antibodies in some patients who have been exposed to the HBsAg and the prevalence of demyelinating autoimmune disorders. Thus it is possible that the binding of HBsAg to this phospholipid binding protein may be related to cellular internalization during immune processing of the peptide and may be related to some of the reported adverse reactions.

III. Preliminary Data and Progress Report

A. Risk vs. benefit of Hepatitis B vaccine.

For any vaccine, it is critical to evaluate the risk/benefit ratio to determine the efficacy, safety and practicality of the vaccine. We have initiated investigations to evaluate the risk (i.e. adverse vaccine reactions) vs. the benefits of this vaccine. Based on well respected published values from the major text used in U.S. medical schools (Crawford, 1994), as well as values from the Centers for Disease Control (CDC) of the incidence and clinical manifestations of this disease the following estimations of vaccine efficacy have been calculated.

1. Estimation of the risk of contracting hepatitis B in U.S.

It is commonly reported by drug company brochures and Center for Disease Control (CDC) publications (Hadler and Margolis, 1993) that there are 200,000 to 300,000 cases of hepatitis B per year in the United States. There are arguments on the internet attributed to the Morbidity and Mortality Weekly Report (internet: www.i-wayco.com/niin/van/van_p5.html) argue that the number of cases may be lower. In this reference, it is reported that in 1992 there were 358 reports of Hepatitis B in New York City and 438 in Upstate New York and only 13,857 cases nationwide. The broad range of estimated cases of hepatitis B in the United States is clearly a problem in assessing the risk of contracting hepatitis B in this country. That problem is compounded by three apparent additional problems:

1. the fact that the majority of these cases appear confined to I-V drug users, sexually promiscuous persons, and medical contacts;


2. the genetic predisposition of some exposed to the hepatitis B virus to fend it off without serious illness; and


3. contraction of the disease by non-responders to the vaccine after vaccination.

Our best assessment of the risk is outlined in Table 3. As best we can we take into account the worst and best case scenarios from the above information and assume that every individual has the same risk (i.e. there is no difference in risk between the normal population and high risk categories such as drug users). This assessment is also based on the averages of clinical populations of this disease outlined by Crawford, (1994). Clearly, the risk would be greatly reduced if the distinction between the normal population and high risk categories was established. Furthermore, these numbers might vary if it was possible to assess how many non-responders to the vaccine contract the virus and have subsequent serious symptoms or die.

Table 3. Estimated relative risk of severe disease of death from Hepatitis B in United States.