Mechanism of vaccine induced diabetes and autoimmunity
The mechanisms by which vaccines alter the risk of an autoimmune disease such as insulin dependent diabetes (IDDM) is not known but there are probably several mechanisms. It is known that killed vaccines can induce the formation of autoantibodies and autoimmunity in humans. Research has focused primarily on vaccines containing molecules that immunologically mimic autoantigens. The neural tissue derived rabies vaccine has historically been associated with most frequently inducing autoimmunity, a condition which has been attributed to the contamination of the vaccine with neural antigens that immunologically cross react to human antigens. Other vaccines including the whole cell pertussis vaccine are known to contain autoantigens (Giudice, Gervaix, Costantinoet al.1993). The presence of these autoantigens however does not readily explain the wide divergence of autoantibodies that arise after vaccination (Hackett & Beech, 1960; Meiselas, Zingale, Leeet al.1961; Caspary, Field & Ball, 1964; Huang,YP, Gauthey, Michelet al.1992) although cross reactivity could have some role (Sutjita, Hohmann, Comacchioet al.1988). Live viral vaccines have been suggested to cause autoimmunity by infecting the recipient's tissue and inducing an autoimmune response as is proposed with insulin dependent diabetes associated with mumps and rubella (Sinaniotis, Daskalopoulou, Lapatsaniset al.1975; Bodansky, Dean, Grantet al.1990; Fescharek, Quast, Maasset al.1990) vaccines. Recently several investigators working on vaccines to control fertility used the diphtheria and tetanus toxoids to break immunological tolerance to HCG, human chorionic gonadotropin, in humans when the hormone is associated with the vaccine (Griffin, 1994). The ability of vaccines to induce autoimmunity to autoantigens that associate with the vaccine can explain the development of a number of autoimmune diseases following immunization.
Vaccines to prevent pregnancy which act by inducing an autoimmune response to HCG have been used in at least four human clinical trials (Griffin, 1994). These vaccines consist of human HCG holoprotein or peptides covalently bound to either a diphtheria or tetanus toxoid, the chief component of the diphtheria and tetanus vaccine. The vaccine toxoids were successful in breaking immunological tolerance and inducing autoimmunity to the human hormone as demonstrated by the development of anti-HCG autoantibodies in the recipients. Detailed studies in animals show that the association of beta HCG with vaccine toxoids greatly increase the immune response to HCG as does the use of alum based adjuvants which are commonly used in vaccines (Schutze, LeClerc, Jolivetet al.1987). HCG peptides were used instead of the holoprotein in order to restrict the development of autoantibodies to epitopes unique to HCG since HCG shares epitopes with other protein hormones. The ability of the vaccines antigens to induce autoimmunity is not limited to HCG since a vaccine comprising a diphtheria toxoid covalently linked to a peptide from human gastrin was able to induce antigastrin antibodies in humans (Anonymous, 1994) .
Animal and human experiments show that vaccine antigens, in killed vaccines, do not have to be covalently attached to autoantigens to induce an autoimmune response. Autoimmunity to the testis and thyroid have been induced in both humans and animals when autoantigens have been administered with Freund's complete adjuvant, (Mancini, 1962; Gerfo, Feind, Weberet al.1983). Autoimmunity has been induced in animals when Freund's complete adjuvant and the autoantigen are administered in different locations but share the same draining lymph nodes (Katsh, 1964; Levine & Wenk, 1967). Animal studies of this phenomenon show the induction of autoimmunity is not limited to the use of Freund's complete adjuvant since the administration of the swine flu vaccine in combination with an neural extract has lead to the development of autoimmune neuritis and the administration of the pertussis vaccine with thyroid extract has lead to the development of autoimmune thyroiditis in rodents (Hjorth, Bonde, Pineret al.1984; Greiner, Mordes, Handleret al.1987) . In the later case a depot type adjuvant, Freund's incomplete adjuvant, was necessary for the induction of autoimmunity. Administration of the pertussis vaccine in the absence of autoantigens has been shown to exacerbate smoldering autoimmunity in rodents (Levine & Wenk, 1966).
The ability of vaccines to induce an autoimmune response to antigens that are in proximity to them explains the induction of autoimmunity in humans following immunization. The pertussis vaccine is known to be an effective adjuvant, stimulating an immune response to antigens that are in close proximity to it (Greenberg & Fleming, 1947). Vaccine antigens become closely association with immunoglobulins after entering the body and this explains why rheumatoid factor, an autoantibody against IgG antibodies, frequently develops after immunization (Aho, Konttinen, Rajasalmiet al.1962; Aho, Somer & Salo, 1967; Palit, Chattopadhyay, Malaviyaet al.1977; Welch, Fong, Vaughanet al.1982). Antigens from killed vaccines associate with a number of other autoantigens besides IgG after being administered. For example antigens from the DTP and other killed vaccines are known to circulate in the blood stream (Lewis, Jordan, Cherryet al.1986) and associate with the membranes of blood cells causing acute lysis of these cells (Brown, Blecher, Frenchet al.1973; Haneberg, Matre, Winsneset al.1978) . This explains why some who receive vaccines develop an autoimmune response to these cells (Klemparskaya, 1973; Zupanska, Lawkowicz, Gorskaet al.1976; Kelton, 1981). Solid organs may also be effected as well. Vaccination causes myocarditis in 3% of health patients (Helle, Koskenvuo, Heikkilaet al.1978; Amsel, Hanukoglu, Friedet al.1986) . Part of this can be explained by circulating antigen precipitating in the heart tissue however people often develop autoantibodies to myocardial tissue after damage to the heart (Rose, Herskowitz, Neumannet al.1988) and this response may be exacerbated by a vaccine draining into a lymphnode where the autoimmune process is developing. In either case the myocarditis induced by vaccination can lead to chronic autoimmune destruction of the myocardial tissue.
The mechanism for inducing autoimmunity appears to involve a lymphokine drive phenomenon where the vaccine activates the immune systems and an immune response develops to autoantigens that are attached to MHC molecules on the same antigen presenting cells as the vaccine toxoids or MHCs on adjacent antigen presenting cells. This phenomenon appears to occur in the draining lymph nodes (Katsh, 1964; Levine & Wenk, 1967) and is likely to involve both the direct activation of macrophages (Mannhalter, Neychev, Zlabingeret al.1985) , the release of lymphokines capable of inducing autoimmunity (Gearing, Bird, Wadhaet al.1987) and the up regulation of lymphokine receptors on cells (Tvede, Heilmann & Christensen, 1989). Animal studies indicate that if the recipient has a smoldering autoimmune disease that the vaccines do not have to be closely associated with autoantigens to exacerbate disease (Levine & Wenk, 1966). This is especially troublesome since subclinical autoimmunity occurs in at least 2-3 % of children based on the presence of autoantibodies (Landin-Olsson, Karlsson, Dahlquistet al.1989b) and is even more common in adults. While naturally acquired foreign antigens such as infections (Rocken, Urban & Shevach, 1992) may also lead to induction of autoimmunity, animal and human experiments have shown the use of depot type adjuvants as well as repetitive administrations intensifies the induction of autoantibodies (Schutze, LeClerc, Jolivetet al.1987; Griffin, 1994). This indicates vaccination is a greater risk factor for autoimmunity than many natural infections of short duration.
Vaccines may alter the incidence of IDDM through mechanisms discussed above as well as several other methods. Cytokines released following immunization may exacerbate subclinical anti-islet cell autoimmunity however they may be directly toxic to the islet cells (Huang,X, Yuan, Goddardet al.1995). It is also possible that immunization later in life causes the release of diabetes inducing viruses which have chronically infected the host (Stanley, Ostrowski, Justementet al.1996). Furthermore, certain vaccines such as the pertussis vaccine (Giudice, Gervaix, Costantinoet al.1993) and the BCG have at least one antigen that cross react to islet cells (Elias, Markovitis, Reshefet al.1990) and these antigens may stimulate an autoimmune disease. Immunization starting at birth may have the opposite effect than immunization starting after 2 months for several reasons. Interferon released following immunization at birth may inhibit vertically transmitted Coxsakie B virus infections which have been attributed to causing 27% or more cases of insulin dependent diabetes (Dahlquist, Frisk, Svanberget al.1995); (Hyoty, Hiltunen, Knipet al.1995). It is possible that immunization at birth also alters the ratios of Th1/Th2 CD4 lymphocytes (Kolb, 1997).