© 2004 American Thoracic Society DOI: 10.1165/rcmb.f280
ADAM33Where Are We Now?Channing Laboratory, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts Address correspondence to: Scott T. Weiss, M.D., M.S., Channing Laboratory, Brigham and Women's Hospital, 181 Longwood Ave., Boston, MA 02115. E-mail: scott.weiss{at}channing.harvard.edu
Abbreviations: metalloproteinase, MP • single nucleotide polymorphism, SNP ADAM33 was the first asthma gene identified by positional cloning (1). Identification of this gene marked the entry of asthma research into the genomic era, and provided new insights to the underlying molecular defects that contribute to asthma pathogenesis. However, several features of the original report raised questions regarding the generalizability of the findings (2). Subsequently, other groups studying populations with asthma have attempted to replicate the genetic associations with ADAM33. Suffice it to say that these studies have found limited evidence of replication for the gene as an asthma gene. Howard and colleagues reported evidence of association of ADAM33 and asthma in four cohorts: a U.S. white cohort, a Dutch white cohort, an African-American cohort, and a Hispanic cohort (3). Single nucleotide polymorphism (SNP) associations were observed in each population, suggesting that the initial observations by Van Eerdewegh and coworkers can be generalized, and that ADAM33 variation is important across ethnicities (1). However, no single SNP or haplotype was identified that consistently demonstrated significant association across the four populations tested. In a second study, Lind and colleagues were unable to demonstrate association either with asthma or several intermediate quantitative phenotypes with ADAM33 variants in a family-based study of 583 Hispanic trios (4). We too attempted, and failed, to detect association in the CAMP population, a family-based study of white, African-American, and Hispanic trios representative of North American children with mild-to-moderate asthma (5). Failure to replicate genetic associations is a common occurrence in complex trait genetics. Although insufficient statistical power is one reason for lack of replication (although likely not the reason for nonreplication of ADAM33 in the Lind and Raby studies), genotypic and phenotypic heterogeneity across study populations, as well as unrecognized differences in environmental interactions, are more typical causes. The genetic findings related to ADAM33 published so far suggest that this gene is a weak-to-moderate asthma and airways responsiveness susceptibility gene. Because asthma is polygenic, no individual gene will have large effects in isolation. Therefore, the full measure of this gene's importance cannot be judged by its effect alone, as it may have greater effects when considered together with other genes (epistatic interaction) or with the relevant environmental exposure. As a result, once a putative susceptibility gene is identified, focus should turn to determining its molecular function, and the precise impact that disease-associated genetic variation has on that function. Unfortunately, determining gene function, particularly as it relates to a complex disease, is equally as challenging as determining genetic association. Shapiro and Owen have eloquently described potential putative mechanisms by which ADAM33 would be involved in airway hyperresponsiveness (6). ADAM33 is primarily expressed in smooth muscle, and within lung tissue localizes to both vascular and bronchial smooth muscle (1, 7). This observation, along with the initial findings of association and linkage among individuals with asthma and with airways hyperresponsiveness, suggest that ADAM33 primarily influences airways hyperresponsiveness (8). This is further supported by the fact that the Adam33 murine homolog maps to a region of mouse chromosome 2 known as an innate airways responsiveness quantitative trait locus (9). Others have speculated that ADAM33 may also have important cytokine-stimulating effects, given that other ADAM proteins (ADAM10 and ADAM17) appear to interact with inflammatory cytokines (10). Despite this circumstantial evidence, little empiric data has been reported to strengthen one hypothesis over another. In the current issue of the AJRCMB, Powell and coworkers examined human airway fibroblasts and show that these cells contain a large number of alternatively spliced forms of ADAM33 (11). Their observations, complementing those reported by Umland and colleagues in the AJRCMB last year (7), provide important information regarding ADAM33 function. Using both quantitative RT-PCR and Western blot to characterize ADAM33 mRNA and protein expression, Powell and colleagues made several important observations: (i) ADAM33 shows a high degree of variability due to complex differential splicing; (ii) the majority of this variation occurs in the 5' end of the gene—the 3' domains (including the EGF, transmembrane, and cytoplasmic domains) are present in all transcripts; (iii) most transcripts observed in the nucleus are also present in the cytoplasm and are detectable as mature protein by Western blot, suggesting that each isoform has functional relevance; and (iv) more than 95% of ADAM33 transcripts do not encode a functional metalloproteinase (MP) domain. This last observation is perhaps the most revealing. Because of the gene's genomic similarity to other ADAM genes (with strongest homology to ADAMs 12, 15, and 19), it was assumed that ADAM33 was primarily a protease localized to the cell surface. The absence of functional MP domains in mature transcript suggests that this may not be the case. Moreover, several of the MP-lacking variants are structurally similar to a synthetic ADAM12-S minigene that induces myogenesis (12). As Powell and colleagues suggest, it is possible that these naturally occurring variants similarly induce airway smooth muscle proliferation and hypertrophy, thus providing a molecular link to the genetic association of ADAM33 with asthma and airways responsiveness. Given this new data, we are certainly getting closer to understanding how genetic variation in ADAM33 results in increased susceptibility to asthma. But we are not there yet. What remains to be described is which of the possible functions of ADAM33 is most pertinent in asthma patients, and how the SNPs associated with asthma in the studies by Van Eerdewegh and coworkers and by Howard and colleagues interfere with this functioning. Can the positive genetic association be replicated in large studies, including longitudinal studies of the development of asthma in early life? Do the MP-lacking forms of ADAM33 indeed induce myoproliferation of bronchial smooth muscle? Do the associated SNPs alter which isoforms are expressed? Do the SNPs disrupt transcript stability or processing? What are the important genetic and environmental modifiers of ADAM33's molecular function, which could explain why the genetic associations are not observed across all populations studied? Answering these questions will not be easy, and will require a multidisciplinary approach. Murine and in vitro model systems should be developed to evaluate the function of each of these isoforms, including transcripts differing by SNP. Tissue samples from numerous individuals should be obtained to evaluate the effects of each SNP and functional haplotypes for their effects on differential splicing and transcript stability. With respect to the latter, it is interesting to note that although most of the splicing diversity in ADAM33 is in the 5' coding regions of the gene, most of the SNPs associated with asthma are localized to the 3' end, including coding exons S and T, and several SNPs in the 3' UTR. It is likely that these SNPs, along with several intronic SNPs in intron ST, regulate ADAM33 gene expression in some manner. Testing these hypotheses will only be possible by assessing transcripts from multiple individuals.
The ADAM33 saga demonstrates how difficult it is to actually go from gene sequence to functional effect, particularly for complex traits, and bears witness to the fact that neither human genetics nor molecular biology alone can provide all the answers for complex traits such as asthma and chronic obstructive pulmonary disease. Only through the integration across both fields of study will we be able to understand how genes influence health. Despite these difficulties, it is worth noting that the number of positionally cloned asthma genes now stands at four: ADAM33, PHF11, DPP10, and GPRA (13–15). There are also now
Conflict of Interest Statement: B.A.R. has no declared conflicts of interest; S.T.W. has no declared conflicts of interest. Received in final form April 29, 2004
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
20 genes identified by genetic association that have been replicated in two or more populations (
Top
References
