It is thought that the alleles that determine the two major isotypes of MHC class II molecule, HLA-DQ or -DR, control the occurence of diabetes (This locus is known as IDDM1 in diabetes parlance, and is on chromosome 6) (Cruse 1995; Davies and others 1994).  Though having one allele or the other does not predispose one to being a diabetic, the interplay between the two alleles does fall into a nice continuum.  One is most susceptible to diabetes if he / she has the DQ allele with either amino acids serine, alanine, or valine at positon 57 of the Beta chain.  If the patient has the DQ allele but has the gene for aspartic acid at postion 57 of both chromosomes, then he / she has a varying degree of immunity to IDDM.  The varying degree of immunity given to DQ+ ASP+ indivuals is always less than immunity displayed by DR+ individuals.  These individuals have the greatest immunity of all, especially those individuals positive for the aspartic acid allele.  However, it is important to remember that IDDM is polygenic and there are many factors that can lead to the onset of the disease, so that regardless of an individuals MHC II isotype, he / she is always susceptible to the disease.  Furthermore, a polymorphism at postion 74 of the DR allele seems to strongly affect DR+ individual's predisposition to diabetes, though presently little else is known about this particular polymorphism (Tisch et al. 1996; HLA-disease associations and transplantation).  The following table shows the possible combinations of HLA alleles and amino acids at position 57, and the resulting IDDM susceptibility.
 
 

Locus	Allele	Susceptible	Resistant
DQ	201	alanine
DQ	302	alanine
DQ	303		aspartic acid
DQ	301		aspartic acid
DQ	502	serine
DQ	602		aspartic acid
DR	405	serine
DR	403		aspartic acid, glutamic acid
DR	401	aspartic acid

Figure 1.  The table should be interpreted as follows:  For row 1, the locus is DQ, and the allele is 201.  This particular allele codes for alanine at postion 57 of the Beta chain of the MHC II molecule, making the patient susceptible to IDDM.  For row 2, allele 302 also codes for alanine and elicits the same result.  For row 3, allele 303 in the same locus codes for aspartic acid, conferring immunity on the host, etc (Tisch 1996).

It is believed that MHC alleles susceptible to autoantigen specific T cells, particularly Th1 cells specific for B cell islet antigens, mediate IDDM susceptibility.  These susceptible cells bind antigens that elicit a primarily Th1 cell response.  The resistant alleles, like those of the DR isotype expressing aspartic acid, elicit a primarily Th2 cell response.  Studies support this hypothesis, as nonobese diabetic mice (because human testing would be unethical, animal models are used to better understand IDDM.  The most common model uses mice infected with IDDM, also known as nonobese diabetic mice (NOD mice)) expressing aspartic acid in position 57 in the I-A chain of their MHC II, showed a low to nonexistent rate of diabetes occurence.  However, since many of the epitopes of autoantigens bound by autoantibodies in IDDM have yet to be discovered, it is impossible to present convincing proof that susceptible MHC II's present different antigen subsets than unsusceptible MHC II's (Eisenbarth and others 1996; Tisch and others 1996).

Unlike the genetic factors the contribute to the disease, namely being a carrier for the HLA-DQ allele in the IDDM1 region, the environmental factors that cause type 1 diabetes are relatively unknown.  One of the leading and more realised hypotheses is that viral infection mediates the introduction of a superantigen into the patients body, and the activation of a wide assortment of T cells, some autoreactive, triggers the onset of IDDM.  Evidence supporting this hypothesis comes from the finding that breast-fed mice are at a lower risk for type 1 diabetes than mice nursed with formula.  In this instance it is thought that some milk-borne virus, such as mouse mammary tumour virus (in humans, cocksackie B virus is a candidate), is transferred to the child and causes superantigen-induced clonal deletion of potentially autoreactive variable regions in T cells.  Thus, if one is exposed to the superantigen at an early age, then all, or at least enough, of the autoreactive T cells are inactivated, providing protection against IDDM.  However, if exposure is later in life, autoreactive clones are not deleted, but instead activated, eventually resulting in beta cell destruction (Conrad and others 1994; Tisch 1996).

The pathology of IDDM progresses in stages.  The first stage is called peri-insulitis and its onset varies from 4 to 6 weeks of age.  Currently, the factors that contribute to the onset of peri-insulitis are unknown.  However, they are hypothesized to be lack of sufficient amount of antigen or APC in beta cell islets until 3 weeks of age, or inhibition of the progress of the disease by an unknown lymphoid cell until 3 weeks of age.  Peri-insulitis is characterised by an accumulation of macrophages, dendritic cells, and B and T lymphocytes in the periductal areas of the pancreas but outside the islet cell areas.  After the inlets are invaded, the next stage of diabetes, intra-insulitis, is reached.  This stage is dependent upon recognition of the B cell autoantigens.  As this stage progresses additional B cell destruction occurs and other B cell specific T cells that bind different autoantigens are recruited.  Finally, at the 18-20 week point, after roughly 95 % of all insulin producing pancreatic beta cells have been destroyed, full-on overt diabetes develops.  The signals that trigger overt diabetes are currently unknown, but are thought to include effector and regulatory T cell interactions and specific autoantigen recognition (Tisch and others 1996; Katz and others 1993).  To see a normal pancreatic islet versus an infected islet, click here.

Many of the specific autoantigens that bind to autoantibodies and trigger intra-insulitis are currently unknown or vaguely defined.  However, researchers have identified the big three.  They are:  an isoform of glutamic acid decarboxylase (GAD65), insulin, and heat shock protein 60 (HSP 60) (Tisch et al. 1996).  GAD65 is an enzyme that synthesizes the neurotransmitter GABA from glutamic acid (Baekkeskov 1990).  Recognition of GAD65 by autoantibodies is thought to occur very early in the disease process (anti-GAD65 antibodies have been found in individuals with very little islet inflammation), and may even mediate the transition from peri-insulitis to intra-insulitis.  Some evidence that supports this claim is animals with intra-insulitis that are administered GAD65 do not progress to overt diabetes and in many cases even show signs of protection from the diabetogenic response.  It is thought that this immunity is mediated by regulatory T cells that bind GAD65 and secrete lymphokines that suppress the diabetogenic response.  The presence of antibodies to GAD65 is the most reliable method of diagnosis of type 1 diabetes (Tisch and others 1996; Diabetes Mellitus and Autoimmunity).

Insulin is the second of the big three beta cell autoantigens.  Insulin has been shown to accelerate diabetes in animal models (NOD mice) by activating insulin specific CD4 T cells, and insulin is also thought to play an unknown role in events in the later stages of diabetes (late intra-insulitis).  Like GAD65, insulin also is effective at curbing the progress of the disease, though only in individuals not exhibiting symptoms of intra-insulitis.  Furthermore, antibodies to insulin do not appear until late in the pathogenisis of the disease, and, for this reason, are poor indicators of the presence of the disease (Tisch et al. 1996; Diabetes Mellitus and Autoimmunity).

Finally, the third major contributor to the diabetogenic response is heat shock protein 60 (HSP60) (Cruse 1995).  HSP60's role in the progression of diabetes remains unclear.  Making matters even more confusing, HSP60 specific Thelper cells have been shown to both block and accelerate the disease in NOD recipients.  However, treatment with HSP60 does protect animals from the disease (tisch et al. 1996).

Other pancreatic beta cell markers that are less well-defined than the three previous markers but, none the less, do exhibit autoreactivity, are:  a 37 and a 40 kDa fragment derived from tyrosine phosphotase 1A-2 and discovered on beta cells, a protein, p69, and unidentified components of the B cell secretory granule (Tisch et al 1996; Diabetes Mellitus and Autoimmunity).  The following table lists all the possible autoantigens that contribute to the diabetogenic response, and whether they induce a primarily Th2 antibody mediated response or killer T cell response.
 
 

autoantigen	antibody	T Cell response
insulin	yes	yes
GAD65	yes	yes
ICA 105 (IA-2)	yes	unknown
Carboxypeptides H	yes	yes
Peripherin	yes	yes
HSP60	yes	yes
p69	yes	unknown
ICA 512	yes	unknown
52 kDa Ag	yes	unknown
gangliosides	yes	unknown
38 kDa secretory granule antigen	unknown	yes

Figure 2.  The 11 autoantigens in IDDM and the T cell responses they elicit.

The T cell response to autoantigens that trigger the diabetogenic response is only slightly better understood than the antigens themselves.  It is known that both CD8 and CD4 T cells are required for islet infiltration and destruction of beta cells.  However, researchers are not clear on the roles the two different cells play.  One hypothesis is that CD8 T cells are required to infiltrate the pancreatic beta cell islets and cause cell injury, which then leads to the priming of CD4 T cells that release lymphokines.  These lymphokines, including TNF-alpha and IFN-gamma, are directly toxic to B cells and can also recruit nonspecific effector cells to amplify the T cell response, particularly macrophages.  The activated macrophages then relase NO and oxygen radicals that further damage the beta cell.  This hypothesis is supported by the lack of the diabetogenic response in mice lacking CD8 T cells and the appearance of CD8 T cells in the pancreatic islets in these same mice just prior to CD4 T cells (Tisch et al. 1996; Eisenbarth and others 1996)).

Though the specific interaction between CD8 and CD4 T cells is unknown, the interaction between Th1 and Th2 cells is more clear.  Specifically, the two subsets oppose one another via the cytokines they produce.  Th1 cells produce IL-2, IFN-gamma, and TNF-alpha, and through these cytokines they promote the development of diabetes and down-regulate Th2 production.  Conversely, Th2 cells play a down-regulatory role in the progression of the disease, and also inhibit Th1 production.  The balance between these two types of cells as determined by the ratio of their cytokines (IFN-gamma to IL-4) is one of the most important factors in determing the progress of IDDM.  A high ratio of IFN-gamma to IL-4 causes the destruction of pancreatic islets, while a low ratio of the same cytokines has the reverse effect, preventing the destruction of the pancreatic islets.  Unfortunately the factors that cause this ratio to fluctuate are unknown.  Furthermore, the Th1 / Th2 separation is not set in stone, as IL-10, a cytokine secreted by Th2, has been shown to promote the progress of the disease in NOD mice, and dual antogonist / agonist CD4 T cells that lack the classical pattern of Th1 or Th2 cytokine release have also been discovered. (Tirsch et al 1996; Katz et al. 1995).

In the past immuosuppressive drugs, such as cyclosporine, which inhibits T cell activation (Janeway and others 1999), and drugs that inhibit cell division, such as imuran, were used to slow the progress of diabetes.  However, these drugs had adverse side effects and, inhibited the function of the entire immune system, rather than selectively targetting only autoimmune reactive T cells.  In order to selectively target only those T cells that are autoreactive, monoclonal antibodies specific for autoantigens recognzied by effector T cells must be used.  Monoclonal antibodies have been made to CD4, CD3, MHC II, B7, and L-selectin and VLA-4, which are both responsible for homing to the pancreas.  Though many of these monoclonal antibodies were able to provide immunity to the disease in mice, they still lacked the desired selectivity and also exhibited adverse side effects (immunogenicity) in humans.  The final approach that the researcher's took, and the one currently in use today, was to administer specific antigen, such as GAD65, insulin, or HSP60, to the patient.  The antigen then reacts with antigen-specific Th2 cells that release IL-4, IL-10, and TGFbeta that suppress Th1 cell production.  However, even this treatment has its drawbacks as sustained, long-term insulin treatment can result in venal, vascular, and retinal complications (Cruse 1995; Tisch et al. 1996).

Baekkeskov S, Henk-Jan A, et al.  13 September 1990.  Identification of the 64K autoantigen in insulin-dependent diabetes as
    the GABA-synthesizing enzyme glutamic acid decarboxylase. Nature 347:  151-156.
Bennett ST, Todd JA.  1996.  Human Type 1 diabetes and the insulin gene.  Annual Review of Genetics 30:  343-370.
Chase P.  1999.  Understanding Insulin Dependent Diabetes.
    <http://diabetes.about.com/health/diabetes/gi/dynamic/offsite.htm?site=http://www.uchsc.edu/misc/diabetes/UIDDM.html>
    Accessed 2000 Apr 20.
Conrad B, Weldman E, et al.  22 Sep 1994.  Evidence for superantigen involvement in isnulin-dependent diabetes mellitus
    aetiology.  Nature 371:  351-355.
Cruse, Julius M.  Illustrated Dictionary of Immunology.  Boca Raton:  CRC Press; 1995.
Davies JL, et al.  A genome-wide search for human type 1 diabetes susceptibility genes.  Nature  371(8):  130-136.
Diabetes Mellitus and Autoimmunity.  <http://www.kronus.com/techserv/tech-diabetes.html>  Accessed 2000 Apr 20.
Eisenbarth GS, Lafferty KJ.  Type 1 Diabetes:  Molecular, Cellular, and Clinical immunology.  Oxford:  Oxford University
    Press; 1996.
HLA-disease associations and transplantation.  <http://rheumb.bham.ac.uk/teaching/clinimmunol/ci_chap6.htm>  Accessed
    2000 Apr 20.
Janeway C, Travers P, Walport M, Capra D.  Immunobiology.  New York:  Garland Publishing; 1999.
Katz JD, Wang B, Haskins K, Benoist C, Mathis D.  24 Sep 1993.  Following a Diabetogenic T cell from Genesis through
    Pathogenesis.  Cell 74:  1089-1100.
Katz JD, Benoist C, Mathis D.  26 May 1995.  T Helper Cell Subsets in Insulin-Dependent Diabetes.  Science 268:
    1185-1188.
Tisch R, McDevitt H.  3 May 1996.  Insulin-Dependent Diabetes Mellitus.  Cell  85:  291-296.
Wilson, Lynn.  Focus on...Diabetes Mellitus.  <http://www.pharminfo.com/pubs/msb/diabetes.html>  Accessed 2000 Apr 20.
Young S.  1999.  HLA disease associations and transplantation.
    <http://rheumb.bham.ac.uk/teaching/clinimmunol/ci_chap6.htm>  Accessed 2000 Apr 20.