INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Sarcoidosis is a systemic inflammatory disorder with a marked imbalance of the cellular immune response. High activity of macrophages and CD4+ helper T cells results in the formation of specific noncaseating granulomas that can be found in a variety of affected organs. From the characteristics of the immune cells involved, an exogenous agent is thought to cause the disease. However, despite intensive efforts no conclusive evidence for a pathogen has been detected in the center of the granulomas (1, 2). There are numerous reports on familial clustering of sarcoidosis (3) and on different prevalence rates in different ethnic groups (9, 10) which supports the idea of an inherited susceptibility. Many studies have focused on HLA genes as candidates that confer this predisposition (7, 11).

Chronic beryllium disease is another granulomatous disorder that shows a striking similarity to sarcoidosis (21). However, a dysregulated immune reaction in response to beryllium has been identified as the cause of beryllium disease. Variants of the HLA-DPB1 gene, which codes for the HLA-DP beta chain, have been identified as a major risk factor for chronic beryllium disease (22). Studies of HLA-DPB1 in sarcoidosis, however, have shown conflicting results (13, 16, 18, 19, 23).

There are a vast number of polymorphic DNA sites throughout the human genome. One type of these polymorphisms, with a variable number of dinucleotide (CA) repeats (so- called microsatellites), is widely used to label chromosomes and to follow their segregation in families (24). We have employed this method to test for linkage of the MHC region and sarcoidosis in families with siblings suffering from this disease.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A sample of 55 German families (white individuals) with siblings suffering from sarcoidosis was drawn from a DNA bank of sarcoidosis families. Details of these families are listed in Table 1. Most families had been recruited by inquiries among members of the German support group Deutsche Sarkoidose-Vereinigung (8). Others were identified from the author's patient archives. All patients and their physicians were interviewed by telephone or questionnaire. Eighty-six of 122 patients reported biopsies from affected organs and in 59 cases the reports were available. The remaining patients showed characteristic radiologic signs in combination with clinical symptoms, laboratory parameters, and course of disorder consistent with the diagnosis of sarcoidosis. The parents of affected siblings were included in the study whenever possible (see Table 1). In families with one or both parents missing, one or two unaffected siblings were analyzed, if available, in order to reconstruct the parental genotypes.

Seven polymorphisms that flank and cover the MHC region were selected from a microsatellite database (http://www.cephb.fr). The positions of these markers on chromosome 6 are included in Figure 1. Genotypes for the microsatellite markers were determined by a polymerase chain reaction (PCR) amplification technique with M13 tailed primers and reaction details from the same database. Fluorescent PCR products were separated according to size and detected on an automated sequence analysis apparatus (25). HLA-DPB1 genotypes were determined with a commercially available kit (INNO LiPA HLA-DPB; Innogenetics, Ghent, Belgium) or with a sequence specific PCR technique (26).

View larger version (11K): [in this window] [in a new window]   Figure 1.   Result of multipoint nonparametric linkage (NPL) analysis of seven microsatellite markers and HLA-DPB1 variants in 55 sarcoidosis families. Distances between polymorphic loci are given in cM (centiMorgan).

Parametric and nonparametric linkage analysis and haplotype construction was performed by use of the GENEHUNTER 2.0 linkage calculation programme (27). Transmission rates for HLA-DPB1 alleles were determined with the Extended Transmission Disequilibrium Test (ETDT) software (28). Eighty-four independent founders of the family panel served as a control group for allele frequencies. Forty-three of these founders were analyzed and 41 could be reconstructed unambiguously from the family data. Reference population data for west Europe were adopted from the literature (29). Allele frequencies were compared with a Monte Carlo test for highly polymorphic loci, CLUMP (30).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Fifty-five sarcoidosis families with 78 pairs of affected siblings (Table 1) were genotyped for seven microsatellite markers and for the HLA-DPB1 polymorphism. Each sib pair could have one or two or no parental allele of each marker in common, or they could be uninformative if the parents were homozygous for the marker. In the absence of genetic linkage between the marker and the disease under scrutiny one would expect an average of 1.0 shared alleles per pair, and as many shared alleles as not shared alleles in the study sample. Predominance of shared alleles is an indicator of linkage between the disease and the tested chromosomal region.

Nonparametric linkage (NPL) analysis of the GENEHUNTER linkage calculation programme is based on this principle. Additional information from haplotypes of adjacent loci is included if multipoint linkage is performed. The result of the multipoint NPL analysis of the genotype data set is shown in Figure 1. There was evidence of linkage (NPL score > 2.5; p < 0.006) for the entire MHC region, with a maximum NPL score of 3.2 (p = 0.0008) for the marker D6D1666 that resides in the MHC Class III gene cluster.

By contrast, parametric linkage (PL) calculation compares the segregation of marker alleles in families with assumed models of inheritance for the disease under scrutiny. Multipoint PL analysis of the genotype data excluded linkage under models of a dominant or recessive gene mutation (assumed mutation frequency in the general population varying from 1 in 33 to 1 in 10,000) within the MHC region.

Haplotypes of the eight polymorphisms were constructed and compared for the 78 affected sib pairs. Ten sib pairs had no parental HLA-DPB1 allele in common, 32 sib pairs shared one HLA-DPB1 allele, and 36 sib pairs shared two HLA-DPB1 alleles (expected under the null hypothesis of no genetic linkage: 19.5/39/19.5 sib pairs with no/one/two shared HLA-DPB1 alleles; chi-squared 19.84, 2 df, p < 0.00005). There was a total of 104 shared HLA-DPB1 alleles. The distribution of these HLA-DPB1 alleles is given in Table a2a. In families with more than two affected siblings the number of shared HLA-DPB1 alleles was divided by the number of sib pairs constructed from these families (see Table 1) to correct for multiple counting. There was no significant deviation from allele frequencies of a control group of 84 founders of the family panel. The founder group did not differ from allele frequencies of a reference population from Western Europe (29). Monte Carlo analysis using CLUMP (30) gave a p value of 0.985 (using 2,000 replicates)

In addition to single allele analysis, HLA-DPB1 alleles could be grouped with regard to concordance of the gene product for single amino acids. For instance, HLA-DPB1*0201, *0601, and others code for glutamic acid at position 69 of the HLA-DP beta chain, i.e., they are HLA-DPB1 Glu69+. The variability of the HLA-DP beta chain is mainly defined by six hypervariable regions of the HLA-DPB1 gene, forming alleles that are positive or negative for Phe9, Lys11, Phe35, Val36, Asp55, Glu56, Lys65, Glu69, Met76, and Gly85 (31). The results for these allele groups are shown in Table b2b. Again, there was no difference from the founder group.

Problems in the definition of a correct control group and the handling of complex families are overcome by use of the transmission disequilibrium test (TDT). Informative trios, including one affected offspring and both parents, are extracted from the families data for this test. The transmission rate of both alleles from a heterozygote parent is expected to be equal (at equilibrium) if the tested marker is not genetically linked to the disorder. Conversely, transmission disequilibrium indicates involvement of the tested marker, regardless of the mode of inheritance. A multiallelic TDT (genotype-wise) using ETDT (28) gave a borderline significant result with a p value of 0.046 (Monte Carlo analysis, after correcting for two tests having been performed, i.e., allelewise and genotypewise TDT; the nominal p value was estimated at 0.023). The results of the TDT for the alleles individually are listed in Table 3. The transmission disequilibrium was most marked for the Glu69+ allele HLA-DPB1*0201 (passed/not passed alleles: 14/4; chi-squared 5.56; 1 df; p = 0.019, not corrected for multiple comparisons). If all HLA-DPB1 Glu69+ alleles were considered together, no significant transmission disequilibrium was detected (passed/not passed alleles: 19/12).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Familial clustering of sarcoidosis is well documented (3, 10). The extent of familial occurrence is influenced by ethnic factors. According to recent reports from Europe it is low in the Spanish population (32) and high in The Netherlands (33). In Germany, a nationwide chest radiography screening program revealed familial occurrence in first degree relatives in 92 of 2,471 patients (3.7%) with intrathoracic sarcoidosis (4). Among the members of the German support group Deutsche Sarcoidose-Vereinigung, 49 of 651 patients (7.5%) had an affected first- or second-degree relative (8).

HLA genes have long been suspected to be involved in the pathogenesis of sarcoidosis. Linkage to HLA Class I genes was ruled out for an exceptionally large pedigree with eight patients with sarcoidosis from northern Sweden (11). A study of 18 African-American sib pairs showed a low probability of linkage to two microsatellite polymorphisms (D6S89, D6S109) close to the MHC (14). Several investigators have reported associations in studies of unrelated patients with sarcoidosis (12, 15, 17, 20). The detection of a strong association between chronic beryllium disease and HLA-DPB1 Glu69+ alleles (22) stimulated the analysis of this polymorphism in sarcoidosis (13, 16, 18, 19). The results were conflicting, but a comparatively strong influence as in beryllium disease was ruled out for sarcoidosis. Maliarik and colleagues (18) found a more pronounced over-representaion of HLA-DPB1 Phe35, Val36, Asp55, and Glu56 in African-American patients with sarcoidosis. Interestingly, these associations were weaker in a subgroup of patients with a family history of sarcoidosis.

Our study has shown that the MHC region is linked to the pathogenesis of sarcoidosis with a significant multipoint NPL score over all of the region. A peak at D6S1666 from the Class III region might point to a stronger influence of this gene cluster. Among the genes of the Class III region, the TNF genes and their key function in cellular immunity are of special interest. Overrepresentation of the TNFA2 allele of a polymorphic site in the promotor region of the TNF alpha gene was found in patients suffering from acute sarcoidosis (Löfgren syndrome), but not in the more frequent group of patients without Löfgren syndrome (34). In Japanese patients with sarcoidosis, analysis of TNF polymorphisms gave no significant results (35).

Our findings for the candidate gene HLA-DPB1 and for DPB1 Glu69+ Berylliosis risk alleles are conflicting. The prevalence of the most frequent Glu69+ allele DPB1*0201, and of all Glu69+ alleles together showed no significant difference between founders of our study and a reference population from West Europe (29). DPB1*0201 was not overrepresented among DPB1 alleles on chromosomes shared by affected siblings.

However, in eight informative families with DPB1*0201 and in one branch of a complex family, the TDT revealed a significant transmission of this allele to affected offspring. This result might have occurred by chance in the small number of DPB1*0201 positive families. Alternatively, it could document heterogeneity of sarcoidosis with a meaning of DPB1*0201 in only a subgroup of families. Finally it could be the consequence of linkage disequilibrium between the susceptibility gene and the HLA-DPB1 locus. In this case DPB1*0201 could often but not always be linked with the predisposing variant of the susceptibility gene.

Linkage disequilibrium is effective only over short distances on the chromosome and would not extend beyond the Class II region. Published studies of HLA-DR and HLA-DQ have so far shown either no or only a weak association or an association confined to specific subgroups of patients with sarcoidosis (12, 15, 17, 20). HLA-DM and HLA-DO lie closer to the HLA-DPB1 locus but have only once been tested (in Japanese patients with sarcoidosis) (36), even though their function on the effectiveness of cellular immune response makes them good candidates for sarcoidosis susceptibility. The antigen processing genes TAP1 and TAP2 are located between HLA-DM and HLA-DO. Associations were found between these genes and patients with sarcoidosis from the United Kingdom and from Poland, but different alleles were associated in either population (23). This result was considered to be a consequence of linkage disequilibrium, but significant linkage disequilibrium between TAP genes and HLA-DPB1 was excluded for the UK patients and control subjects.

Our study has disclosed the most significant evidence of involvement of the MHC region in the pathogenesis of sarcoidosis published so far. From our parametric linkage data is it clear that no single gene with a simple mode of inheritance can account for this effect. Published associations between genes from the MHC and sarcoidosis as well as our TDT data for HLA-DPB1 alone cannot explain the observed excess of MHC haplotype sharing in sarcoidosis sib pairs. Multiple additive MHC gene effects or an as-yet unknown susceptibility gene in linkage disequilibrium with HLA Class II genes, perhaps together with a contribution of genes from the MHC Class III region, might be considered. In any case, the MHC is a most promising target to search for genes that predispose to sarcoidosis.