Progression index; time to EDSS 6.0, time to secondary-progressive phase

Normalized brain volume and T2-weighted lesion load on MRI Changes in brain volume and in Tl- and T2-weighted lesion loads on MRI Raw EDSS; progression index EDSS; extent of demyelination and brain atrophy on MRI Progression index; time to EDSS 6.0; lesion loads and brain volume on MRI Progression index s2 associated with less severe disease No in women

More pronounced brain damage in Yes s4 carriers (even in earliest disease stages)

More pronounced tissue destruction Yes in s4 carriers

Role of s4 in MS progression may be Yes limited to initial disease stages

No effect of alleles on clinical or No

MRI severity parameters

No relation between alleles and No clinical or MRI measures of disease severity

Faster disease progression in e4 Yes carriers aCases of "Asian-type" MS were excluded.

Abbreviations'. MS, multiple sclerosis; EDSS, Expanded Disability Status Scale; RR, relapsing-remitting; BO, bout-onset (RR or secondary progressive); MRI, magnetic resonance imaging; PP, primary progressive; 'H-MRS, proton magnetic resonance spectroscopy; SNPs, single-nucleotide polymorphisms. Source: From Ref. 125.

complex diseases. As Carlson et al. (86) have explained, concentrating on proximal phenotypes improves the "signal-to-noise ratio'' of any genetic factor contributing to the overall phenotype (provided the contribution is mediated by the proximal phe-notype in question).

In this spirit, Kikuchi et al. (135) have recently described two separate "subpopulations'' of patients with "Western-type" MS in Japan: in the first, in which women outnumber men by a factor of five to one, MS is associated with the HLA-DR15 and the presence of oligoclonal bands (OCBs) in the cerebrospinal fluid (CSF); in the second, in which women are still more common, but only by a factor of two to one, MS is associated with HLA-DR4 and the absence of OCB. We have now confirmed, in a Swedish dataset, the associations of DR15 and DR4 with, respectively, OCB-positive and OCB-negative MS (136). Moreover, we have also demonstrated that HLA-DR15 is associated with an earlier age at onset in MS (137,138), a finding corroborated in two subsequent studies (139,140). In the latter two studies, the association between DR15 and MS susceptibility was stronger in females than in males. We have also observed that over 80% of our OCB-positive MS patients with an age at onset under 21 years are carriers of DR15 (99). Thus, another aspect of the MS phenotype, age at onset, could perhaps be incorporated into the scheme proposed by Kikuchi et al. (141): early-onset MS cases are typically OCB-positive, DR15-positive females, while late-onset cases are often OCB-negative, DR4-positive males.

This strategy of stratification—based on clinically, paraclinically, demographic-ally, and immunogenetically defined intermediate phenotypes—may in the future facilitate the identification of non-HLA genetic risk factors (or even gender-specific non-genetic risk factors) in MS. Indeed, in recent studies from Japan (141) and Finland (142), after stratification for gender and HLA class II phenotype, genotypes at an intronic SNP in the gene encoding estrogen receptor 1 were shown to confer, respectively, 16- and 19-fold risks for the development of MS in HLA-DR15-positive females. If this association turns out to be reproducible, it would strengthen the suspicion of a hormonal basis for the gender bias in MS; suggest the importance of immunoendocrine crosstalk in MS; imply the existence of separate genetic risk factors in men and women and in carriers and noncarriers of DR15; and, perhaps most importantly, designate a potential target for pharmacological therapy (or prophylaxis).

Another strategy used in genetic studies to decrease the heterogeneity of the MS phenotype is to investigate geographically isolated populations with a high prevalence of MS, or multigenerational families in which MS appears to be inherited in a Mendelian fashion. The rationale behind this strategy assumes that, within each isolated population or within each family displaying Mendelian inheritance, the same combination of genetic and environmental factors—which represents a subset of all the risk factors in the total MS population—is contributing to disease risk.

In 1994, Binzer et al. (143) reported that, in Overkalix, Sweden, 12 of the village's 4744 inhabitants suffered from MS (corresponding to a prevalence of 253 cases per 100,000 persons); through church archives, it could be shown, in the majority of cases, that the MS patients descended from the same 18th-century ancestral couple. In a subsequent genetic study of this population (144)—a genome-wide screen, followed by analysis of the transmission of alleles within familial trios [by the transmission-disequilibrium test (TDT)]—we found that 12 of 15 affected subjects carried some portion of a conserved haplotype on chromosome 17p11. We later performed genome-wide TDT analysis on MS patients and healthy family members from another geographically isolated population in Varmland County, Sweden (145), and identified five regions that appeared to be associated with MS (on chromosomes 2q23-31, 6p24-21, 6q25-27, 14q24-32, 16p13-12, and 17q12-24).

In both AD and Parkinson's disease—which, like MS, are common neurological disorders believed to be caused in the majority of cases by the interaction of several genes with unknown environmental factors—the identification of rare disease forms inherited in a classic Mendelian fashion has helped investigators elucidate generalizable pathogenetic mechanisms. We have recently characterized a consanguineous family of Middle Eastern origin exhibiting multiple cases of MS and performed a genome-wide screen on nine family members now residing in Sweden (146). Based on the presence of consanguinity, our a priori hypothesis was that the disease was being transmitted in an autosomal recessive manner in the pedigree; however, we found no chromosomal region for which all affected family members were homozy-gous by descent. Yet, there are indications that MS may not be a straightforwardly autosomal recessive trait in our pedigree; e.g., an LOD score of 1.7 was found on the X chromosome, suggestive of an X-linked trait partially penetrant in females.

At the same time, consanguinity is known to increase the likelihood that non-Mendelian, oligogenic traits will occur multiple times within the same family; in a recent genome-wide screen of 16 members of a large inbred Amish kinship, 7 of whom had MS, Vitale et al. (147) reported a maximum LOD score of 2.7, conditional on the presence of HLA-DR15, for a locus on chromosome 12p12, suggesting a two-locus inheritance model in the pedigree. Meanwhile, in another recent genome-wide screen of a seemingly Mendelian multigenerational MS kinship, Dyment et al. (148) found, by parametric analysis, no evidence for linkage to the HLA-DRB1 locus, but, by TDT, significant association with the DRB1*15 allele. The authors conclude, in cliffhanger fashion, that DRB1 "is ... not the hypothetical single gene acting to determine MS within this family,'' but rather a "modifier" of either pene-trance or some unnamed phenotypic trait.

It is important to note that it is uncertain to what extent loci identified in population isolates or Mendelian families will be of relevance to sporadic MS cases from the general population. Indeed, according to Terwilliger et al. (149), this is the central paradox of complex-disorder gene mapping: the simpler one makes the localization of susceptibility genes, the more difficult it becomes to estimate the contribution of any localized genes to the total risk of the disease in question. Indeed, there is growing consensus that the future of gene mapping in complex disorders lies, not in the investigation of upstream phenotypes or exceptional pedigrees, but rather in SNP-based genome-wide association studies of large datasets of unexceptional cases and controls.

It has been estimated that there are 15 million SNPs in the human genome (150). There is agreement among researchers that only a subset of these SNPs needs to be genotyped in a genome-wide LD screen, but there is disagreement regarding the best way to select this subset (151). Proponents of "map-based" screening favor a subset of anonymous SNPs, each in LD with a highly conserved ancestral haplotype (152), while proponents of "sequence-based" screening favor SNPs with functional consequences, such as those that encode amino-acid substitutions or disrupt splice sites (153). The relative merits of the two approaches have been outlined in a recent review (150).

At the same time, Merikangas and Risch (154) have argued that, from a public-health standpoint, not all complex disorders are worthy of the "expensive and laborious tools of molecular genetics.'' They propose that, with regard to the mapping of germline variants that increase disease risk, complex disorders that are ''highly amenable to environmental modification"—such as nicotine dependence and AIDS— should be given much lower priority than disorders for which the hypothesized environmental risk factors remain unknown, such as schizophrenia and MS. However, given the recent breakthroughs in MS epidemiology—in particular, findings regarding the influence of smoking (155,156), sunlight exposure (157,158), vitamin D intake (159), and the immune response to common herpes viruses (160,161) on disease susceptibility—it is uncertain how long MS will remain in the latter, prioritized category.

Keeping abreast of all the genetic studies performed in MS—many of them underpowered or otherwise flawed in their design, and nearly all of them inconclu-sive—has been rightly described as ''a Sisyphean task" (51). Still, we agree with Saw-cer and Compston (162), the principal instigators of the GAMES initiative, that the current state of MS genetics is a ''cup half full,'' rather than a ''cup half empty." To paraphrase the Belorussian Talmudist Saul Lieberman, although futility is futility, the history of futility is scholarship.

0 0

Post a comment