Reverse phenotyping in sarcoidosis.

The term “phenotype,” first coined by the Danish botanist Wilhelm Johannsen in 1909, is defined as “the observable characteristics of an individual resulting from the interaction of its genotype with the environment.” The success in mapping complex disease genes and in identifying genotype and environment interactions depends on how well the phenotype is defined. Complex diseases such as sarcoidosis are likely influenced by several environmental factors and genes. Further adding to the difficulties in genetic dissection of sarcoidosis is that this is a highly heterogeneous disease, with a wide-ranging clinical course (acute, chronic, spontaneously resolved) and pattern of organ system involvement. This heterogeneity likely hampered achieving highly statistically significant linkage signals in the two genome scans thus far reported (1, 2). One way to counter the inherent difficulties of disease heterogeneity is to identify narrowly defined, accurate, stable phenotypes. Focusing on subtypes can identify more homogeneous populations for analysis. Examples of this approach include the subdivision of age of onset in Alzheimer’s disease (3) and chronic obstructive pulmonary disease (4) and aggressiveness in prostate cancer (5). High-throughput genotyping and sophisticated statistical methods now allow an additional approach to the genetic dissection of complex diseases called “reverse phenotyping” (6). In this approach, the genetic marker data are used to drive new phenotype definitions. Groups can be distinguished by higher rates of allele sharing than that seen in the traditional diagnostic categories—for example, byHLA genotype. Because a narrowly defined, stable disease phenotype can offer advantages to gene discovery in sarcoidosis, the article in this issue of the Journal (pp. 40–44) by Grunewald and Eklund (7) is of particular interest. The association of erythema nodosum (EN) with bilateral hilar lymphadenopathy (BHL), called Löfgren’s syndrome, was first recognized as an acute, benign form of sarcoidosis by Löfgren and Lundbach (8). Uveitis occurred in 13 of the original 212 patients described (6%) and the syndrome was later broadened to include fever and periarticular ankle inflammation. Grunewald and Eklund inform us that Scandinavian patients with acute onset, BHL, and marked periarticular ankle inflammation fall into one phenotype whether or not they have EN. Furthermore, disease resolution occurred in almost every patient with the HLA alleles DRB1*0301/DQB1*0201 whereas resolution occurred in only 55% of patients without the DRB1*0301/ DQB1*0201 alleles. Significantly more women were in the ENpositive group (67%) compared with the EN-negative group (27%) and more men with the same acute sarcoidosis syndrome presented without EN but with signs of bilateral ankle arthritis. In addition to being equipped with a more defined clinical phenotype—acute sarcoidosis with BHL, which may or may not include EN or bilateral ankle arthritis—this subgroup can also be more narrowly characterized by the presence of HLADRB1*0301/DQB1*0201. HLA class II genes have long been thought to play a role in sarcoidosis susceptibility and phenotype (9–11). Grunewald and Eklund previously reported that HLADR *0301 predicts a good prognosis in Scandinavian patients with pulmonary sarcoidosis (12) and is associated with a reduced expression of T-helper cell type 1 cytokines (IFNand tumor necrosis factor) and a tendency toward higher levels of transforming growth factor1 (13). Löfgren’s syndrome was noted to portend a good prognosis. Over 90% of patients with sarcoidosis with Löfgren’s syndrome will have total resolution of the chest roentgenogram and other signs of the disease within 2 years (14, 15). Some of the remaining 10% could have severe disease lasting as many as 20 years (15). To date, there has been no way to identify the subset of patients with Löfgren’s syndrome with a poor prognosis. Grunewald and Eklund have provided important long-term follow-up of their patients with Löfgren’s syndrome. They found that only one of 87HLA-DRB1*0301/DQB1*0201–positive patients had chronic disease,whereasnearlyhalf of theHLA-DRB1*0301/DQB1*0201– negative patients had chronic disease. Linkage disequilibrium, the nonrandom association between two or more alleles, exists for HLA-DRB1*0301 and DQB1*0201. A clear association for DQB1*0201 with reduced risk of disease progression has been previously reported and also found to be associated with Löfgren’s syndrome in patients in the United Kingdom (16). Using a family-based approach to study AfricanAmerican patients with sarcoidosis, Iannuzzi and colleagues (10) reported that HLA-DQB1*0201 was transmitted to affected offspring half as often as expected and that sarcoidosis risk associated with *0201 depended on exposure status. The *0201 allele showed no significant correlationwith phenotype (10). It remains to be demonstratedwhether the combinationHLA-DRB1*0301/ DQB1*0201 is associated with a low rate of disease progression in patients with sarcoidosis without Löfgren’s syndrome. Many challenges exist in the identification of genes underlying complex human diseases. As high-throughput genotyping and statistical tools reach unprecedented levels of efficiency and sophistication, the key lies in the phenotype. Grunewald and Eklund have more thoroughly clinically defined Löfgren’s syndrome, emphasizing that EN predominantly occurs in women and bilateral periarticular arthritis in men, and that this acute sarcoidosis subgroup can be further characterized according to presence of HLA-DRB1*0301/DQB1*0201, providing the first example of reverse phenotyping.