Narcolepsy is a complex neurological disorder characterized by excessive daytime sleepiness (Overeem et al. 2001). In the majority of cases, cataplexy is also present - a dramatic symptom in which there is sudden bilateral loss of postural muscle tone in association with intense emotions such as laughter.
Other rarer symptoms of narcolepsy include vivid dream-like experiences during the transition between wakefulness and sleep (hypnagogic hallucinations) and sleep paralysis, in which patients are subjectively awake and conscious but unable to move during the onset of sleep or on waking.
Subsequent detailed genetic analysis including linkage analysis and positional cloning led to the identification of specific deletions of the hypocretin receptor 2 gene (Lin et al. 1999). Genetic knock-out studies of the same gene in mice gave a phenotype with some similarity to narcolepsy (Chemelli et al. 1999). There is now substantial evidence to link the 'wake promoting' neuropeptide hypocretin with narcolepsy. Post mortem studies of affected individuals show the destruction of hypothalamic neurones secreting hypocretin (Peyron et al. 2000) while cerebrospinal fluid (CSF) levels of hypocretin are characteristically low or undetectable at the time of making a clinical diagnosis of narcolepsy (Nishino et al. 2000).
How does the hypocretin peptide relate to the very strong genetic association between MHC class II alleles and narcolepsy? The initial association was reported for Japanese patients with HLA-DR2 found in all 40 patients with narcolepsy (Juji et al. 1984). The association was also found among Caucasians, with subsequent studies showing narcolepsy to be most strongly associated with HLA-DQB1*0602 across a broad range of ethnic groups - a feature unusual among HLA disease associations - such that this allele is carried by 88-98% of cases of narcolepsy, but only if cataplexy is present (Mignot et al. 1997). The haplotype is also relatively common among the general population, typically at 10% frequency, indicating incomplete penetrance and the multifactorial nature of the condition. Indeed among monozygotic twins, only about one-third are concordant, indicating a major environmental contribution. Other MHC genes have been implicated such as TNF, while a recent genome-wide association study among Japanese adults highlighted chromosome 21q22.3 where three narcolepsy candidate genes were identified (Kawashima et al. 2006). However it is HLA-DQB1*0602 that shows the strongest association - one of the strongest identified for any disorder - and the finding that a very closely related allele, HLA-DQB1*06011, protected against narcolepsy (Mignot et al. 2001) prompted structural studies comparing these HLA molecules.
The solving of the crystal structure of HLA-DQB1*0602 complexed with hypocretin has highlighted some potentially functionally important differences with DQB1*06011 (Siebold et al. 2004). The two molecules differ only at nine amino acid residues in the p chain, of which three may modulate T cell receptor recognition while the remaining five are found in the peptide binding groove. Of particular note are polymorphisms altering the P4 pocket of the groove which define the particular peptide side chains that can fit - for DQB1*06011 the pocket becomes closed up such that hypocretin is prevented from binding.
The molecular mechanism relating DQB1*0602 and narcolepsy remains incompletely understood. An autoimmune process with selective destruction of hypocretin-containing neurones in the hypothalamus is an attractive hypothesis but unproven. The environmental trigger that may be required for such an autoimmune reaction is unknown. The finding of functional autoantibodies in serum and CSF from narcolepsy patients with the *0602 haplotype is intriguing. Immunoglobulin from patient serum led to narcolepsy-like behaviour in mice (Smith et al. 2004) while that from CSF was shown to bind to rat hypothalamic proteins. Further work is required, but the HLA association has provided a clinically useful diagnostic tool and important insights into disease pathophysiology.
The study of genetic factors underlying coeliac disease has provided important insights into the pathophysiology of this disorder, in particular the role of MHC class II alleles in determining disease susceptibility (reviewed in Kagnoff 2007). Coeliac disease (OMIM 212750) is a common, chronic, inflammatory disease affecting the small bowel, which often presents with symptoms of weight loss, malnutrition, and diarrhoea. The disease is activated in susceptible individuals by eating wheat gluten, together with other similar proteins found in rye and barley; examination of the small intestine in affected people shows damage to the lining intestinal mucosa resulting in malabsorption of nutrients.
Genetic factors have been recognized for many years to be very important in defining susceptible individuals. The disease affects families and shows a very high (70-75%) concordance among monozygotic twins (Greco et al. 2002). The strongest genetic association is with HLA-DQ, which is necessary but not sufficient for the disease to occur (Kagnoff 2007). Heterodimers of HLA-DQ2 are present in 90-95% of patients with coeliac disease. Here the p chain of the class II molecule is encoded by DQB1*0201 or *0202 and the a chain by DQA1*05. The remaining 5-15% of patients have DQ8 heterodimers encoded by DQB1*0302 and DQA1*03 (encoding the p and a chain, respectively). Coeliac disease is extremely rare in Japan where the DQ2 susceptibility alleles are very uncommon.
How does possession of particular HLA-DQ2 and -DQ8 alleles relate to susceptibility to coeliac disease? The answer appears to lie once more in the peptide binding groove of the class II molecule (Kim et al. 2004; Bergseng et al. 2005). At first sight the situation appears contradictory. The binding grooves of DQ2 and DQ8 favour binding of negatively charged protein residues at key anchor positions, making them highly unlikely to bind the praline- and glutamine-rich gluten peptides. However, a specific deamidating enzyme, transglutaminase 2 (TGase), was found to be upregulated in the inflamed intestine, which converts neutral glutamine to negatively charged glutamic acid (Dieterich et al. 1997). Moreover structural studies showed that the praline-rich gluten peptide with deaminated glutamines would bind DQ2 without the expected disruption of hydrogen bonds. The DQ2 and DQ8 heterodimers on antigen presenting cells thus bind and present the gluten peptides to CD4-positive T cells in the lining of the intestine, activating them (Fig. 12.7).
The gluten peptides, by the nature of their biochemical make up, are resistant to proteolytic digestion and occur as relatively large peptides in the small intestine. The presence of multiple DQ binding epitopes is thought to explain the greater T cell stimulating ability of larger peptides. Other genetic and immunological factors are likely to be involved in disease pathogenesis as well as concurrent infection, for example with enteric viruses, which is thought to be a key event in triggering disease (Kagnoff 2007). This may involve exposure to gluten peptides at a time of intestinal inflammation with increased permeability due to viral infection associated with interferon production and immunological activation. What has become clear is that a genetic background of HLA-DQ2 or -DQ8 alleles is required for the disease to occur, and that understanding how these specific molecules interact with the peptide antigens of gluten has proved very valuable to our knowledge of disease pathogenesis.
The MHC has the strongest evidence of linkage and genetic association of any genomic region with susceptibility to, and protection from, the juvenile onset form of diabetes - which is characterized by autoimmune destruction of pancreatic p cells and insulin deficiency. Type 1 diabetes (OMIM 222100; also known as insulin-dependent diabetes) shows strong familial clustering: siblings of affected individuals are at a 15-fold increased risk of the disease (Spielman et al. 1980) with the largest proportion of this increased risk due to inherited factors within the MHC. All reported genome-wide linkage scans have shown strongest linkage to markers in the MHC although non-MHC associations are also found (Box 12.8). It is thought that genetic variation within the MHC accounts for 40% of the observed familial clustering, with relative risks of disease between 3 and 50 depending on the specific genotypes and haplotypes involved (Lambert et al. 2004). Type 1 diabetes is, however, a multifactorial disease with strong environmental influences as demonstrated by the marked rise in incidence and the earlier age of onset over the last 50 years, as well as the incomplete concordance observed in monozygotic twins (Hirschhorn 2003). The role of inherited factors in type 1 diabetes has been of great research interest in the quest for an improved understanding of disease pathogenesis and the possibility of primary prevention and early intervention through ascertainment of genetic risk markers.
Uptake of gluten peptides across gut epithelial cells? due to altered permeability
Figure 12.7 Pathogenesis of coeliac disease. Illustration of events following ingestion of gluten in a genetically susceptible person with presentation by HLA-DQ2 and -DQ8 class II molecules to CD4+ T cells of gluten peptides which have undergone glutamine deamidation. Redrawn with permission from Kagnoff (2007) .
Other genomic regions have also been associated with disease susceptibility outside the MHC, notably genetic variation at the INS (encoding insulin) gene locus on chromosome 11p15 (Section 7.3.3) (Bell et al. 1984), CTLA4 (encoding cytotoxic T lymphocyte antigen 4 gene) at 2q33 (Ueda et al. 2003), PTPN22 (encoding lymphoid protein tyrosine phosphatase, a suppressor of T cell activation) at 1p13
(Bottini et al. 2004), IL2RA (encoding interleukin-2 receptor a chain) at 10p15 (Vella et al. 2005), and IFIH1 (innate immunity viral RNA receptor gene region) at 2q24 (Smyth et al. 2006). Recent genome-wide association studies have demonstrated a number of other loci notably at 12q13, 12q24, 16p13, 18p11, and 18q22 (Section 9.3) (Hakonarson et al. 2007; Todd et al. 2007; WTCCC 2007).
The role of the MHC in determining susceptibility to diabetes has been the subject of intense research for over 20 years (Nerup et al. 1974; Cudworth and
Woodrow 1975). Almost all patients with type 1 diabetes were found to have either HLA-DR3 or -DR4 compared to about half of the unaffected population. Early studies using serological tests for HLA antigens and restriction fragment length polymorphisms (RFLPs) as markers demonstrated that the strongest association was with the linked HLA-DQ genes. In 1987 Todd and colleagues sequenced the major expressed polymorphic MHC class II gene products from affected patients and controls (Todd et al. 1987). Their work showed the amino acid sequence of the DQ p chain directly correlated with susceptibility to disease, in particular the amino acid found at position 57. The presence of aspartic acid (Asp) at position 57 in one copy of the gene was associated with protection from disease: having two Asp 57 positive DQ p alleles conferred almost complete protection from type 1 diabetes. The non-obese diabetic (NOD) mouse strain, a classic spontaneous autoimmune animal model for human type 1 diabetes, was found to have Ser at this position while all other mouse p chains have Asp. The relationship was, however, found not to be absolute, with particular combinations of DQ and DR conferring susceptibility (Erlich et al. 1990).
The situation is thus complex and remains incompletely understood. Specific polymorphisms and haplo-typic combinations across the HLA-DR and -DQ region, mainly involving amino acid substitutions in the a chain of DR, and both a and p chains of DQ, are associated with disease and show a hierarchy of susceptibility (Lambert et al. 2004). Thus the HLA-DR2 haplotype, DRB1 *1501 -DQA1 *0102-DQB1 *0602, is extremely rare among children with type 1 diabetes (<1%) while present in 20% of the unaffected population (Noble et al. 1996). Most Caucasian patients with type 1 diabetes will possess one of the two commonest susceptibility haplotypes (lacking Asp at position 57 of the p chain), of DR3 ( DRB1*0301-DQA1*0501-DQB1*0201) and DR4 (DRB1*0401-DQA1*0301 -DQB1*0302), with heterozygosity for these haplotypes associated with a 15-fold relative risk and earlier onset of disease.
How do genotypic and haplotypic differences relate to mechanism of action at a molecular level? The answer, or at least part of it, may lie in how these differences change the composition of the peptide binding groove of DR and DQ molecules. Specific features of peptide binding pockets P1, P4, and P9 appear critical. As discussed earlier, residue 57 of the DQ p chain correlates well with disease susceptibility. The substitution of Ala57p for Asp57p at this site within the P9 pocket results in a change in the basic charge, the accessible volume of the pocket, and disrupts hydrogen bond formation with the peptide backbone (Jones et al. 2006). Such changes in structure of the binding groove could have profound consequences for peptide recognition and function. They have been postulated to alter disease susceptibility through mechanisms such as tolerance (presentation of autoantigens in the thymus) as well as peripheral antigen presentation.
Finally, there is also evidence of association of type 1 diabetes with variation in MHC class I genes. The major association has been robustly demonstrated to be with HLA-DQB1 and HLA-DRBT, however this does not fully explain the observed MHC association with disease. A recent study combined a large family study of affected sib pairs together with case-control cohorts, which included individuals genotyped in the Wellcome Trust Case Control Consortium (Section 9.3.2) (Nejentsev et al. 2007). Dense genotyping of 254 polymorphic loci and 1475 SNPs within the MHC allowed Nejentsev and colleagues to define associations with HLA-B (combined P value = 2.01 X 10-19) and HLA-A (P = 2.35 X 10-13) independently of the previously resolved MHC class II associations. The major effect was with HLA-B*39, which showed strong association with disease risk and lower age at diagnosis. Possession of this allele was associated with a relative risk among families of 3.5 (95% CI 2.2-5.7) and among the case-control cohort of 2.4 (95% CI 1.43.9) (Nejentsev et al. 2007).
Was this article helpful?