© 2003 American Thoracic Society
Seeking Modifier Genes in Cystic FibrosisThe Mike McMorris Cystic Fibrosis Research and Treatment Center Department of Pediatrics The Children's Hospital University of Colorado School of Medicine Denver, Colorado We need new treatments for cystic fibrosis (CF). Although life expectancy is increasing, most patients succumb in young adulthood from progressive lung disease (1). We know that lung disease in CF is caused by mutations in the gene coding for the cystic fibrosis transmembrane conductance regulator. But mutations in this gene do not explain the striking variability in clinical course. Patients with identical CF genotypes can have marked differences in lung function (2). Environmental influences likely modify clinical course. It is likely, however, that variations in genes other than the gene coding for cystic fibrosis transmembrane conductance regulator also modify clinical course. Identification of these "modifier genes" could lead to development of new treatments.
The case for the existence of modifier genes in CF is growing stronger. Genetic variation in a region on human chromosome 19 is associated with meconium ileus, neonatal intestinal obstruction that occurs in some infants with CF (3). Putative modifier genes of pulmonary outcome in CF have also been reported, including genes that code for tumor necrosis factor, transforming growth factor, In one report, Grasemann and coworkers (pp. 390–394) (6) continue their search for modifiers of the nitric oxide pathway in CF. The nitric oxide pathway is of interest because of links to host defense and to inflammation. In addition, abnormalities in cystic fibrosis transmembrane conductance regulator have been associated with abnormalities in nitric oxide metabolism (7). Grasemann and coworkers find that a polymorphism in the endothelial nitric oxide synthase gene leads to increased nitric oxide in the airway of female patients with CF. This polymorphism is also associated with less frequent colonization with Pseudomonas aeruginosa, the major pathogen in CF. The implications are that the endothelial nitric oxide synthase gene is a modifier of lung disease in CF and that increased airway nitric oxide may be beneficial in CF. This report, however, raises a number of questions about studies examining association of candidate modifier genes with clinical outcomes. Definition of phenotype is an important issue. Grasemann and coworkers use culture status, defined as presence or absence of P. aeruginosa, as a key characteristic of their phenotype. But we are not told how many culture results were examined to determine this characteristic. Cultures are sometimes negative even in patients who have multiple positive cultures on other occasions, indicating the need for careful attention to definition of culture status. In addition, the strongest statistical finding uncovered in this study is the baseline sex difference in the frequency of polymorphisms in the endothelial nitric oxide synthase gene. The polymorphism that is expected to result in higher airway nitric oxide is less frequent in females than in males. The authors do not provide an explanation for this. Could it be a survivor effect (i.e., females with the high airway nitric oxide synthase polymorphism succumbed early in their course leading to a decrease in prevalence)? Or could the baseline sex difference be a statistical vagary of this small group of 70 patients? Furthermore, we are not given odds ratios or confidence limits. It is difficult, therefore, to determine the significance of associations between polymorphisms and pulmonary outcome. Finally, the authors have previously reported that polymorphisms in the neuronal nitric oxide synthase gene are associated with nitric oxide levels in CF (8). We are not given information regarding the genotype status of this isoform in patients in the current report. Will the association of pulmonary outcome with polymorphisms of the endothelial nitric oxide synthase gene hold up once polymorphisms in the neuronal isoform are taken into account? In the other report, Arkwright and coworkers (pp. 384–389) (9) find that a polymorphism in the gene coding for angiotensin converting enzyme but not polymorphisms in four candidate cytokine genes is associated with development of liver disease in CF. This is an important finding because it points to control of the hepatic circulation as a potential target for treatment—possibly with angiotensin converting enzyme inhibitors. In addition, the angiotensin converting enzyme gene is shown to modify development of lung disease, suggesting new approaches to pulmonary treatment. We are provided with more details of phenotype definition in this study than in the study of Grasemann and coworkers (6). In addition, Arkwright and coworkers (9) consider several candidate modifier genes at once, allowing comparison between candidates. Odds ratios and confidence limits are also provided. Association studies such as those reported in this issue examine the relationship between phenotype and polymorphisms in candidate modifier genes in affected individuals. There are important limitations to this approach. Associations do not prove that the candidate genes studied are the cause of differences in phenotype because differences may be due to other genes that travel with the candidates. In homogeneous population groups, as may be the case in CF, there can be association of genes at even distant loci. False positive association studies have been noted (10). Several other approaches to genetic epidemiology allow more robust inferences than association studies. These approaches include study of the affected individual along with the parents ("trios"), as well as study of siblings and twins. As increasing numbers of genes are associated with outcome, we will also need to apply statistical techniques that account for polymorphisms in many genes at once. Two other points regarding the search for modifier genes in CF are worth noting. It is clear that we need a standardized definition of phenotype to allow more meaningful comparisons of modifier genes. In addition, a recent study in twins and siblings indicates that functions directly related to cystic fibrosis transmembrane conductance regulator, membrane ion transport, and/or intracellular trafficking of mutant protein are subject to modifier effects (11). Identification of modifier genes for these functions directly related to cystic fibrosis transmembrane conductance regulator may yield intriguing insights into the pathophysiology of CF as well as providing therapeutic targets. REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1 antitrypsin, and mannose-binding lectin (
