Genomewide Association Between GLCCI1 and Response to Glucocorticoid Therapy in Asthma Kelan G
Total Page:16
File Type:pdf, Size:1020Kb
The new england journal of medicine original article Genomewide Association between GLCCI1 and Response to Glucocorticoid Therapy in Asthma Kelan G. Tantisira, M.D., Jessica Lasky-Su, Sc.D., Michishige Harada, Ph.D., Amy Murphy, Ph.D., Augusto A. Litonjua, M.D., Blanca E. Himes, Ph.D., Christoph Lange, Ph.D., Ross Lazarus, M.B., B.S., Jody Sylvia, B.S., Barbara Klanderman, Ph.D., Qing Ling Duan, Ph.D., Weiliang Qiu, Ph.D., Tomomitsu Hirota, Ph.D., Fernando D. Martinez, M.D., David Mauger, Ph.D., Christine Sorkness, Pharm.D., Stanley Szefler, M.D., Stephen C. Lazarus, M.D., Robert F. Lemanske, Jr., M.D., Stephen P. Peters, M.D., Ph.D., John J. Lima, Pharm.D., Yusuke Nakamura, M.D., Ph.D., Mayumi Tamari, M.D., Ph.D., and Scott T. Weiss, M.D. Abstract Background The response to treatment for asthma is characterized by wide interindividual vari- From Channing Laboratory (K.G.T., J.L.-S., ability, with a significant number of patients who have no response. We hypothe- A.M., A.A.L., B.E.H., R.L., J.S., B.K., Q.L.D., W.Q., S.T.W.) and the Pulmonary Division sized that a genomewide association study would reveal novel pharmacogenetic (K.G.T., A.A.L.), Brigham and Women’s determinants of the response to inhaled glucocorticoids. Hospital and Harvard Medical School; and Harvard School of Public Health (C.L.) Methods — all in Boston; the Center for Genomic We analyzed a small number of statistically powerful variants selected on the basis Medicine, Institute of Physical and Chemi- cal Research (RIKEN), Kanagawa, Japan of a family-based screening algorithm from among 534,290 single-nucleotide poly- (M.H., T.H., Y.N., M.T.); the Department morphisms (SNPs) to determine changes in lung function in response to inhaled of Medicine, University of California, San glucocorticoids. A significant, replicated association was found, and we character- Francisco, San Francisco (S.C.L.); the Department of Pediatrics, University of ized its functional effects. Wisconsin School of Medicine and Public Results Health (R.F.L.), and the University of Wisconsin (C.S.) — both in Madison; the We identified a significant pharmacogenetic association at SNP rs37972, replicated in Center for Human Genomics, Section of four independent populations totaling 935 persons (P = 0.0007), which maps to the Pulmonary, Critical Care, Allergy and Im- GLCCI1 munologic Diseases, Wake Forest Univer- glucocorticoid-induced transcript 1 gene ( ) and is in complete linkage dis- sity School of Medicine, Winston-Salem, equilibrium (i.e., perfectly correlated) with rs37973. Both rs37972 and rs37973 are NC (S.P.P.); Nemours Clinic, Jacksonville, associated with decrements in GLCCI1 expression. In isolated cell systems, the FL (J.J.L.); Arizona Respiratory Center and Department of Pediatrics, University rs37973 variant is associated with significantly decreased luciferase reporter activity. of Arizona, Tucson (F.D.M.); Penn State Pooled data from treatment trials indicate reduced lung function in response to in- Hershey College of Medicine, Hershey, PA haled glucocorticoids in subjects with the variant allele (P = 0.0007 for pooled data). (D.M.); and the Department of Pediatrics, University of Colorado Health Sciences Overall, the mean (±SE) increase in forced expiratory volume in 1 second in the Center, Denver (S.S.). Address reprint re- treated subjects who were homozygous for the mutant rs37973 allele was only about quests to Dr. Tantisira at Channing Labo- one third of that seen in similarly treated subjects who were homozygous for the ratory, 181 Longwood Ave., Boston, MA 02115, or at [email protected]. wild-type allele (3.2±1.6% vs. 9.4±1.1%), and their risk of a poor response was sig- nificantly higher (odds ratio, 2.36; 95% confidence interval, 1.27 to 4.41), with geno- Drs. Tantisira and Lasky-Su contributed type accounting for about 6.6% of overall inhaled glucocorticoid response variability. equally to this article. Conclusions This article (10.1056/NEJMoa0911353) was published on September 26, 2011, at A functional GLCCI1 variant is associated with substantial decrements in the response NEJM.org. to inhaled glucocorticoids in patients with asthma. (Funded by the National Insti- N Engl J Med 2011. tutes of Health and others; ClinicalTrials.gov number, NCT00000575.) Copyright © 2011 Massachusetts Medical Society. 10.1056/nejmoa0911353 nejm.org 1 The New England Journal of Medicine Downloaded from nejm.org at WAKE FOREST UNIV HLTH SCIENCES on September 27, 2011. For personal use only. No other uses without permission. Copyright © 2010 Massachusetts Medical Society. All rights reserved. The new england journal of medicine sthma is a complex genetic syn- Methods drome that affects 300 million persons worldwide.1 The response to treatment is Study Design and Screening and Replication A Cohorts also genetically complex and is characterized by high intraindividual repeatability2 and high in- Figure 1 provides an overview of the study design. terindividual variability,3 with up to 40% of pa- Detailed descriptions of the methods (including tients with asthma having no response to therapy. the population descriptors, functional character- Inhaled glucocorticoids are the most widely pre- ization, and statistical approaches used) can be scribed medications for controlling asthma. Levels found in the Supplementary Appendix, available of endogenous glucocorticoids are heritable and with the full text of this article at NEJM.org. vary, both at baseline and in response to environ- In the CAMP,19,20 a randomized, controlled trial, mental perturbation.4-6 Moreover, studies in fami- we followed 1041 children with asthma who were lies with conditions other than asthma have shown 5 to 12 years of age at the onset of the study both familial segregation and heritability in re- and who received treatment for a mean period of sponses to glucocorticoid medications.7,8 Given 4.6 years. The children were randomly assigned to the heritability within the therapeutic class of treatment with inhaled budesonide, nedocromil glucocorticoids, as well as the high degrees of sodium, or placebo. From this study, we selected between-patient variability and within-patient re- 422 white, non-Hispanic participants and their peatability in the response to inhaled glucocorti- parents for genotyping on the HumanHap550 v3 coids for the treatment of asthma, it is likely that BeadChip (Illumina); of this group, 118 trios (con- this response has a genetic basis. sisting of a child and his or her two parents) were To date, pharmacogenetic investigations in randomly assigned to budesonide. These trios con- asthma have focused on candidate genes.9-13 A stituted the family-based screening cohort in which powerful family-based screening algorithm14 for we assessed the longitudinal response to gluco- genomewide association studies has identified corticoid therapy. All research involving data col- novel genetic loci that contribute to obesity15,16 lected from the CAMP Genetics Ancillary Study and Alzheimer’s disease.17 The algorithm uses was conducted at the Channing Laboratory of parental genotype information to rank single- the Brigham and Women’s Hospital according to nucleotide polymorphisms (SNPs) that have the the appropriate CAMP policies and regulations greatest potential power for association. A small for human-subjects protection. subset of statistically powerful SNPs can then be To replicate our initial findings, we genotyped tested in the probands with the use of the family- DNA obtained from subjects with asthma who based association test (FBAT).18 Identifying mark- were enrolled in three clinical trials: the 6-week ers in this fashion limits the potential for false- common run-in period during which the subjects positive associations and increases the likelihood were using inhaled glucocorticoids in the Sal- of replication in subsequent studies. meterol or Corticosteroids (SOCS) trial21 and We hypothesized that a genomewide asso- the Salmeterol ± Inhaled Corticosteroids (SLIC) ciation study would identify novel variants as- trial,22 the Adult Study,23 and the Leukotriene sociated with the response to inhaled glucocor- Modifier or Corticosteroid or Corticosteroid– ticoids for asthma. We tested this hypothesis Salmeterol (LOCCS) trial (ClinicalTrials.gov num- with the use of the family-based screening algo- ber, NCT00156819).24 rithm in subjects randomly assigned to inhaled After the initial replication phase, an additional glucocorticoids in the Childhood Asthma Man- replication was performed that was limited to the agement Program (CAMP).19,20 Through screen- variant associated in each of the populations with ing, we identified SNPs that offered the great- the use of data from the Childhood Asthma Re- est power for a replicable association with the search and Education (CARE) Network trials,25,26 longitudinal response to inhaled glucocorti- archived on the database of Genotypes and Pheno- coids, measured as a change in forced expira- types (dbGaP) (www.ncbi.nlm.nih.gov/gap) with- tory volume in 1 second (FEV1). After screening, in the SNP Health Association Resource (SHARe) we tested the association of the highest-powered Asthma Resource Project (SHARP). SNPs in four additional, independent popula- Table 1 summarizes these populations, most tions drawn from clinical trials involving sub- of which have been described in detail in re- jects with asthma. ports on previous pharmacogenetic studies.27,28 2 10.1056/nejmoa0911353 nejm.org The New England Journal of Medicine Downloaded from nejm.org at WAKE FOREST UNIV HLTH SCIENCES on September 27, 2011. For personal use only. No other uses without permission. Copyright © 2010 Massachusetts Medical Society. All rights reserved. GLCCI1 in Asthma Response to Glucocorticoid Therapy All subjects or their legal guardians provided writ- ten informed consent to participate in the study Genotyping protocols and ancillary genetic testing. of CAMP trios Drug responses and outcomes FBAT screen for power In the CAMP, the greatest differences in FEV1 that Genotype: expected genotypes from parental DNA were attributable to budesonide were seen during Phenotype: post-ICS FEV 20 1 the first 16 months of treatment.