| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Division of Pulmonary, Critical Care, and Sleep Medicine, Mount Sinai Medical Center, New York, New York
Correspondence and requests for reprints should be addressed to Michael C. Iannuzzi, M.D., Mount Sinai Medical Center, 1 Gustave L. Levy Place, Box 1232, New York, NY 10029. E-mail: michael.iannuzzi{at}mountsinai.org
ABSTRACT
Familial aggregation and racial differences in incidence support the notion that sarcoidosis occurs in genetically susceptible hosts. Siblings of those affected with sarcoidosis have a modestly increased disease risk, with an odds ratio of about 5. HLA genes have been the most extensively studied susceptibility genes in sarcoidosis. Many other attractive candidate genes have been evaluated using the case–control study design, but few have been confirmed. Confounding by population stratification likely explains much of the failure to replicate initial findings. A genomewide scan performed in German families with follow-up fine mapping studies has yielded a highly attractive candidate gene, BTNL2 in the MHC II region on chromosome 6. BTNL2, a member of the B7 family of costimulatory molecules, likely functions to down-regulate T-cell activation. A BTNL2 single-nucleotide polymorphism associated with sarcoidosis is predicted to result in a truncated nonfunctioning protein. Association of BTNL2 with sarcoidosis has been confirmed in both white and African Americans. A genomewide scan with follow-up fine mapping studies in African American families has identified chromosome 5 as potentially harboring candidate genes. Additional linkage analysis in the African American families stratified according to genetic ancestry demonstrated that linkage signals varied according to degree of admixture. Certain chromosomal regions were also found linked to specific phenotypes. Follow-up fine mapping studies of the linked regions are underway.
Key Words: sarcoidosis • genetics • BTNL2
First described in 1877 by Jonathan Hutchinson as a "case of livid papillary psoriasis," the disease was later coined "sarcoidosis" in 1899 by Caesar Boeck. Boeck could study the disease histologically as the newer techniques of hematoxylin staining (1863) and paraffin processing (1871) became available (1). A key ingredient to advancing knowledge about sarcoidosis came from Boeck's patient's agreeing to participate and allowing for a skin biopsy. In much the same way, more recent advances in the genetics of sarcoidosis rely on patient participation and new technology including high-throughput genotyping and statistical genetic analyses.
Differences in disease incidence among racial groups and the worldwide observation of familial aggregation support that genetic susceptibility to sarcoidosis exists (2, 3). In the United States, African Americans are three and a half times more commonly affected than white Americans (3). Sarcoidosis occurs in about 35.5 per 100,000 African Americans and 11 per 100,000 white Americans. African-American women, 30 to 39 years of age, are the most commonly affected (107 per 100,000) (3). Outside the United States, from 3.6 to 9.6% of patients report that their first- or second-degree relatives also have sarcoidosis. About 3.8% of white Americans and 17% of African Americans report that a first- or second-degree relative is also affected (4).
The sibling recurrence-risk ratio (
s), which compares disease risk among siblings with the disease prevalence in the general population, provides a more quantitative measure of familial aggregation. We found a modest s of 2.2 (95% confidence interval [CI], 1.03–3.68) for African Americans (5). For comparison, s is about 4 for asthma and 20 for multiple sclerosis. In ACCESS (A Case–Control Etiologic Sarcoidosis Study), which evaluated 706 cases and matched control subjects, the odds ratio for siblings was 5.2 (95% CI, 2.1–15.9) (6).
The search for sarcoidosis susceptibility genes has generally relied on the candidate gene approach (7). Investigators have selected genes for study that fit into the prevailing disease model. Sarcoidosis is thought to be a dysregulated response to an inhaled antigen that involves antigen-presenting cells, T cells (primarily a helper T-cell type 1 polar response), and cytokine and chemokine release resulting in cell recruitment and the formation of granulomas in involved organs.
HLA genes have been the best studied candidate genes in sarcoidosis because major histocompatibility (MHC) class II proteins are responsible for antigen presentation to T cells. Four attractive non-HLA candidate genes that have been studied are caspase recruitment domain family, member 15 (CARD15); natural resistance–associated macrophage protein-1 (NRAMP1); angiotensin-converting enzyme (ACE); and chemokine (C–C motif) receptor-2 (CCR2).
CARD15, alias nucleotide oligomerization domain protein-2 (NOD2), located on chromosome 16, is the gene responsible for Blau syndrome (8, 9) and the gene identified in the inflammatory bowel disease locus IBD1 (10). NOD2, encoded by CARD15, recognizes peptidoglycan, a component of bacterial cell walls, and is expressed mainly by antigen-presenting cells and epithelial cells (11). Activation of NOD2 leads to nuclear factor (NF)-
B activation (11). The key features of Blau syndrome are listed in Table 1. Other reported manifestations include neuropathy, hepatitis, and bowel inflammation. Using exclusion mapping among sarcoid-affected sib pairs, the Blau syndrome/IBD1 locus was found not to confer risk for sarcoidosis (12). Further eliminating CARD15 as a sarcoidosis susceptibility gene, Schurmann and coworkers (13) evaluated four main coding CARD15 polymorphisms associated with increased risk of Crohn's disease in both case–control and family-based samples and concluded that CARD15 mutations play no role in sarcoidosis.
|
Epithelioid cells in granulomas produce ACE and serum ACE (SACE) levels are thought to reflect granuloma burden. The ACE gene insertion (I)/deletion (D) polymorphism partially accounts for the SACE level variation and investigators have proposed that reference values to interpret SACE levels should be adjusted according to genotype (19). Studies supporting a role for ACE gene polymorphisms in either susceptibility or disease severity have been inconsistent (20, 21).
Several chemokines play pivotal roles in granuloma formation. The V64I polymorphism in CCR2, a receptor for monocyte chemotactic protein-1, has been reported to be associated with sarcoidosis in Japanese and Czech populations (22, 23). Valentonyte and colleagues studied CCR2 polymorphisms more extensively, using both case–control and family-based study designs, and found no support for an association of CCR2 with sarcoidosis risk (24).
Although choosing candidate genes based on the prevailing model of sarcoidosis pathogenesis and employing a case–control study design is convenient, failure to replicate findings is commonplace. The likely reason for failure to replicate is that the case–control design is susceptible to a form of confounding known as population stratification. Population stratification arises when the gene under study shows marked variation in allele frequency across subgroups that also differ in disease risk. Because infection and immune surveillance appear to be under strong negative population pressure, immune response genes seem particularly subject to confounding by population stratification. The problem of population stratification can be overcome by family-based case–control designs using sibling and parents as controls. Parental alleles not transmitted to affected family members are used as the control alleles, thus controlling for genetic background (Figure 1). Using a family-based study design, we found that previously reported associations for several candidate genes with sarcoidosis could not be confirmed (20). Another potential cause of replication failure is that certain candidate genes might influence phenotype more than susceptibility. For example, although the CCR2 association has not been consistently replicated, Spagnolo and coworkers (25) did report an association with Löfgren's syndrome. Other studies likely did not include samples with sufficient patients with Löfgren's syndrome to also detect an association.
|
BTNL2, also known as "butyrophilin-like 2" and "BTL-2," is a butyrophilin gene that belongs to the immunoglobulin gene superfamily and is related to the B7 family (28, 29). Jack and Mather were the first to identify butyrophilin, initially cloned from cattle mammary epithelial cells (30). This gene was localized to the MHC class II region in humans.
To determine the consistency of the BTNL2 gene as a sarcoidosis risk factor across different populations, we characterized variation in the BTNL2 exon/intron 5 region in an African-American family sample that consisted of 219 nuclear families (686 individuals) and in 2 case–control samples (295 African-American matched pairs and 366 white American matched pairs) (31). We confirmed the importance of BTNL2 in white Americans and found additional SNPs that were associated with sarcoidosis in African Americans. BTNL2 appears to have moderate influence on individual disease risk (odds ratio of 1.6 in heterozygotes and 2.8 in homozygotes). The population attributable risk of 23% for heterozygotes and homozygotes indicates a significant contribution at the population level.
One question regarding BTNL2 as a sarcoidosis risk factor is whether it is independent of HLA-DRB risk alleles. HLA-DRB and BTNL2 are in linkage disequilibrium. Linkage disequilibrium is the nonrandom association of alleles physically close on a chromosome. HLA-DRB lies about 180 kb centromeric to BTNL2. On the basis of regression models, BTNL2 appears to be an independent risk factor (27, 31). In the case of African Americans, in whom the BTNL2-conferred sarcoidosis risk is less significant than for white Americans, a negative interaction with HLA-DR appears to exist (31). BTNL2 has been found not to be associated with Wegener's granulomatosis (32).
Eleven centers joined together in an NHLBI-sponsored effort entitled the Sarcoidosis Genetic Analysis Consortium (SAGA). This group performed a 380-microsatellite scan across 22 autosomes in 519 African American sib pairs and found 15 markers with p values less than 0.05 (33). Fine mapping studies indicated that the linkage peak with the highest statistical significance, located on chromosome 5, remained significant (34). We chose to perform our linkage scan in African-American families because African Americans are more commonly and severely affected with sarcoidosis and report having an affected family member more often than do white Americans. Further, 11 centers participated in recruitment, allowing us to quickly obtain a relatively large sample size. One disadvantage to this strategy is that African Americans are admixed with white Americans and other populations to varying degrees, with possible admixture among our participating centers ranging from 12% in South Carolina to 20% in New York (35).
We performed linkage analysis stratified by genetically determined ancestry, using the data from the 380 microsatellite markers genotyped in the genome scan. We clustered the African-American families into subpopulations based on ancestry similarity. Evidence of two genetically distinct groups was found: stratified linkage results suggest that one subpopulation of families contributed to previously identified linkage signals at 1p22, 3p21–14, 11p15, and 17q21, and a second subpopulation of families contributed to those found at 5p15–13 and 20q13. Our findings support the presence of sarcoidosis susceptibility genes in regions previously identified, but indicate that these genes are likely to be specific to ancestral groups that have combined to form modern-day African Americans.
In addition to finding that the chromosome 5 signal remained significant after fine mapping and stratifying by ancestry, we further analyzed the genome scan linkage signals and found linkage of certain phenotypes to different chromosomes rather than to disease presence. For example, we found that cardiac/renal involvement was linked to chromosome 18q22 (LOD, 6.31; p = 0.03) (Rybicki and coworkers, manuscript in review). A genome scan of diabetic nephropathy in African Americans produced a strong linkage signal in this same region, with a maximum logarithm of odds score of 3.72 in families distributed to subsets on the basis of age at diabetes diagnosis (36).
Even though it has been more than 100 years since Hutchinson and Boeck first described sarcoidosis, the cause and why certain populations are more susceptible remain unknown. Many candidate gene studies with case–control designs have been reported, but with the exception of HLA genes few have been adequately replicated. Two genome scans have been reported and one has yielded a likely candidate gene, BTNL2, that has been replicated in large studies. We await follow-up of the scan performed in African Americans, which may explain why African Americans are more commonly and severely affected.
FOOTNOTES
Conflict of Interest Statement: M.C.I. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript.
(Received in original form June 28, 2006; accepted in final form March 20, 2007)
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
