Glucocorticoid-induced Changes in Gene Expression ... - ATS Journals

5 downloads 57 Views 667KB Size Report
May 15, 2013 - Conclusions: Oral prednisolone changes the transcriptomic profile of .... After rinsing in phosphate-buffered saline, the biopsy specimens were.
Glucocorticoid-induced Changes in Gene Expression of Airway Smooth Muscle in Patients with Asthma ¨ nberg4, Thais Mauad5, Kees Fluiter6, Ching Yong Yick1, Aeilko H. Zwinderman2, Peter W. Kunst3, Katrien Gru Elisabeth H. Bel1, Rene´ Lutter1,7, Frank Baas6, and Peter J. Sterk1 1 Department of Respiratory Medicine, 2Department of Biostatistics, 6Department of Genome Analysis, and 7Department of Experimental Immunology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands; 3Department of Respiratory Medicine, Admiraal De Ruyter Hospital, Goes, The Netherlands; 4Department of Pathology, VU Medical Center, Amsterdam, The Netherlands; and 5Department of Pathology, Sa˜o Paulo University Medical School, Sa˜o Paulo, Brazil

Rationale: Glucocorticoids are the mainstay of asthma therapy. However, it is unclear whether the benefits of glucocorticoids in asthma are merely based on antiinflammatory properties. Glucocorticoids may also alter gene expression of airway smooth muscle (ASM). We hypothesized that the gene expression profile of the ASM layer in endobronchial biopsies of patients with asthma is altered by oral glucocorticoid therapy as compared with placebo. Objectives: First, we investigated the change in ASM transcriptomic profile in endobronchial biopsies after 14 days of oral glucocorticoid therapy. Second, we investigated the association between changes in ASM transcriptomic profile and lung function. Methods: Twelve steroid-free patients with atopic asthma were included in this double-blind intervention study. Endobronchial biopsies were taken before and after 14 days of oral prednisolone (n ¼ 6) or placebo (n ¼ 6). RNA of laser-dissected ASM was sequenced (RNASeq) using GS FLX1 (454/Roche). Gene networks were identified by Ingenuity Pathway Analysis. RNA-Seq reads were assumed to follow a negative binomial distribution. At the current sample size the estimated false discovery rate was approximately 3%. Measurements and Main Results: Fifteen genes were significantly changed by 14 days of oral prednisolone. Two of these genes (FAM129A, SYNPO2) were associated with airway hyperresponsiveness (provocative concentration of methacholine causing a 20% drop in FEV1: r ¼ 20.740, P , 0.01; r ¼ 20.746, P , 0.01). Pathway analysis revealed three gene networks that were associated with cellular functions including cellular growth, proliferation, and development. Conclusions: Oral prednisolone changes the transcriptomic profile of the ASM layer in asthma. This indicates that in parallel to antiinflammatory properties, glucocorticoids also exert effects on gene expression of ASM, which is correlated with improved airway function. Keywords: asthma; airway smooth muscle; RNA-Seq; glucocorticoids; prednisolone

(Received in original form October 19, 2012; accepted in final form February 27, 2013) Author Contributions: C.Y.Y. performed the study procedures, participated in the design of the study, wrote the manuscript, and performed the statistical analyses with A.H.Z. A.H.Z. as biostatistician assisted in designing the study, the analysis plan, power calculations, and statistical analysis and participated in drafting the manuscript. P.W.K. performed all bronchoscopic procedures. K.G. and T.M. participated in the design of the study. K.F. designed the primers for the qPCR validation of the gene sequencing data. E.H.B., R.L., and F.B. participated in the design of the study and its coordination. P.J.S. conceived the study, participated in the design of the study and its coordination. In addition, all coauthors helped to draft the manuscript. All authors read and approved the final manuscript. Correspondence and requests for reprints should be addressed to Ching Yong Yick, M.D., Department of Respiratory Medicine (F5-260), Academic Medical Center, University of Amsterdam, Meibergdreef 9, 1105 AZ, Amsterdam, The Netherlands. E-mail: [email protected] Am J Respir Crit Care Med Vol 187, Iss. 10, pp 1076–1084, May 15, 2013 Copyright ª 2013 by the American Thoracic Society Originally Published in Press as DOI: 10.1164/rccm.201210-1886OC on March 14, 2013 Internet address: www.atsjournals.org

AT A GLANCE COMMENTARY Scientific Knowledge on the Subject

Airway smooth muscle (ASM) plays an important role in the pathophysiology of asthma and is therefore a potential target in the treatment of this disease. Glucocorticoids are used in asthma therapy, leading to improvements in clinical status, exacerbation rate, and lung function. However, it is unclear whether the benefits of glucocorticoids in asthma are due to antiinflammatory properties or whether glucocorticoids also exert other actions within the airways. Glucocorticoids may also affect the ASM phenotype by altering its gene expression. What This Study Adds to the Field

Glucocorticoids not only have antiinflammatory properties, they also change the gene expression profile of the ASM layer in asthma, which appears to be correlated with improved airway hyperresponsiveness. Better understanding of the pharmacodynamics of glucocorticoids in ASM may help advance asthma therapy.

Asthma is a respiratory disease affecting millions of people, posing a considerable burden on health care systems globally (1). The functional characteristics observed in asthma are well described including airway hyperresponsiveness (2), variable airway obstruction (3), and altered airway dynamics (4). Structural changes of the airway wall, collectively called airway remodeling, are likely to contribute to these functional changes (5). More specifically, airway smooth muscle (ASM) mass, elevated mast cell counts, and extracellular matrix (ECM) deposition within the ASM layer are associated with the clinical expression and severity of asthma (6–9). This suggests that ASM plays an important role in the pathophysiology of asthma and therefore in being a potential target in the treatment of this disease. Glucocorticoids are presently the most effective maintenance therapy in asthma leading to improvements in the clinical status, exacerbation rate, and lung function (1). However, it is unclear whether these benefits are based solely on antiinflammatory properties or whether glucocorticoids also exert other local actions within the airways (10). In vitro studies suggest that glucocorticoids may change gene transcription in ASM cells (11), leading to altered transcription of contractile elements, cytoskeleton, cell surface molecules, cytokines, and mediators (12). Indeed, glucocorticoids influence the contractile properties and glucocorticoid receptor–related genes of ASM in vitro (13, 14), and airway dynamics in vivo (15). This suggests that the ASM phenotype is influenced by glucocorticoids (16).

Yick, Zwinderman, Kunst, et al.: Prednisolone and ASM Gene Expression in Asthma

To our knowledge there is no study on the effect of glucocorticoids on the ASM gene expression profile in patients with asthma in vivo. Gene expression profiles of ASM can be investigated with microarrays or by transcriptomics using next-generation, high-throughput sequencing (RNA-Seq) (17). The latter allows a detailed examination of the ASM gene expression profile and provides elaborate insight into the gene pathways that regulate various cellular and molecular processes (18). We hypothesized that the transcriptomic profile of the ASM layer in endobronchial biopsies of patients with atopic asthma is altered by oral glucocorticoid therapy as compared with placebo. First, we examined the change in ASM transcriptomic profile in endobronchial biopsies taken before and after 14 days of oral glucocorticoid therapy. Second, we investigated the association between the changes in ASM transcriptomic profile and those in airway function.

METHODS Design The current randomized, double-blind, parallel, placebo-controlled intervention study comprised four visits. At visit 1, patients with asthma were screened according to the inclusion and exclusion criteria before enrollment. In addition, spirometry and a methacholine bronchoprovocation test were performed. At visit 2, FEV1 reversibility was measured and endobronchial biopsies were collected during bronchoscopy. On the same day as the bronchoscopy, patients with asthma started their 14-day course of placebo or prednisolone. Patients with asthma were prescribed oral prednisolone at a dose of 0.5 mg/kg/day. The dosage and dosing scheme were based on international recommendations for the treatment of acute exacerbations (1). On Day 11 after visit 2, spirometry and a methacholine bronchoprovocation test were performed. Finally, at visit 4 (Day 15 after visit 2) FEV1 reversibility was measured and endobronchial biopsies were collected by bronchoscopy. Both the bronchoscopy procedures on visit 2 and on Day 15 after visit 2 were performed in the morning and consistently during one of the following three time slots: 09:00 A.M., 10:00 A.M., and 11:15 A.M.

Subjects The study population consisted of 12 steroid-free patients with atopic asthma distributed into a prednisolone study group (n ¼ 6) or placebo study group (n ¼ 6). The sample size was based on an a priori power and false discovery calculation (see below). Participants were recruited by the Department of Respiratory Medicine of the Academic Medical Center Amsterdam (Amsterdam, The Netherlands). The inclusion criteria were as follows: age, 18–50 years; nonsmoking or stopped for longer than 12 months with a smoking history of no more than 5 packyears; controlled disease according to Global Initiative for Asthma (GINA) guidelines (1); no exacerbations within 6 weeks before participation; steroid-naive or stopped using glucocorticoids by any dosing route for at least 8 weeks before participation; airway hyperresponsiveness defined by a methacholine PC20 (provocative concentration of methacholine causing a 20% drop in FEV1) not greater than 8 mg/ml; postbronchodilator FEV1 greater than 70% of predicted; no pulmonary diseases other than asthma and no pulmonary medication except inhaled short-acting b2-agonists as rescue therapy; and atopy defined by a positive skin prick test. This study was approved by the Medical Ethics Committee of the Academic Medical Center Amsterdam and was part of a study registered in the Netherlands Trial Register (The Role of the Airway Smooth Muscle Layer in Asthma; NTR1306).

Study Procedures Standardized spirometry was performed according to the recommendations of the European Respiratory Society (19). The methacholine bronchoprovocation test by the standardized tidal volume method with a maximal dose of 16 mg/ml was performed to measure the PC20 (2). FEV1 reversibility was determined by the difference in FEV1 preand postinhalation of 400 mg of salbutamol. The skin prick test for

1077

determining the atopic status of patients with asthma included 12 common aeroallergen extracts (ALK, Hørsholm, Denmark). Endobronchial biopsies were collected as previously described (8) and according to international recommendations (20). In short, lignocaine was sprayed in the nose and throat. During bronchoscopy, additional lignocaine solution was instilled into lung segments to dampen the cough reflex. A cup forceps (KW2411S; PENTAX, Breda, The Netherlands) was placed laterally to the bronchial carina to maximize the proportion of ASM in the biopsy specimens. Four endobronchial biopsies per subject were collected at B 7–9 of the right lung.

Isolation of RNA from ASM After rinsing in phosphate-buffered saline, the biopsy specimens were incubated overnight in RNAlater (Qiagen, Venlo, The Netherlands). Afterward, the biopsies were embedded in Tissue-Tek (Sakura Finetek, Alphen aan den Rijn, The Netherlands) and frozen in liquid nitrogen– cooled isopentane. The frozen biopsies with an average size of 1–2 mm3 were entirely cut up into 10-mm sections in an RNase-free cryostat and mounted on polyethylene naphthalate (PEN) membrane–covered microscope slides (Carl Zeiss, Sliedrecht, The Netherlands). The slides were stained with hematoxylin and eosin followed by laser capture microdissection (LCM) of ASM at 310 magnification (AS LMD; Leica, Wetzlar, Germany). The captured ASM sections from the four frozen biopsies per subject were pooled into the same microcentrifuge tube. Total RNA was isolated with TRIzol (Invitrogen, Carlsbad, CA) with the addition of RNase-free glycogen (UltraPure glycogen; Invitrogen) as carrier.

Complementary DNA Preparation and RNA-Seq The Ovation RNA-Seq system (NuGEN, San Carlos, CA), which includes an isothermal, linear amplification step (21), and an Agencourt RNAClean XP kit (Beckman Coulter, Brea, CA) were used to prepare amplified and purified complementary DNA (cDNA) from the isolated RNA. cDNA libraries greater than 350 bp in size were constructed with the SPRIworks fragment library system II (Beckman Coulter) with the addition of GS FLX Titanium Rapid Library multiplex identifier (MID) adaptors (Roche, Penzberg, Germany) to allow multiplexing of samples. The GS FLX Titanium emPCR kit (Lib-L; Roche) was used to prepare enriched DNA library beads by emulsionbased clonal amplification. Afterward, the enriched DNA beads were loaded on a two-region PicoTiterPlate device, with each region containing 2 million DNA beads. The GS FLX1 system (454/Roche) was used for the sequencing run (22).

Sequence Data Analysis The GS Reference Mapper version 2.6 (Roche) was used to map the sequencing reads against the human genome (hg19) (23). Gene network analysis was performed with the Ingenuity Pathway Analysis (IPA) application (24) by uploading the significantly up- or downregulated genes. A network score was generated by IPA for each gene network found. This IPA score is displayed as the negative log of the P value of that specific network and gives the likelihood that the set of genes in this network could be explained by chance alone. In other words, networks with an IPA score equal to or greater than 2 have at least 99% confidence that it is not generated by chance. Read numbers of reference genes found before and after 14 days of study medication were compared including glyceraldehyde-3-phosphate dehydrogenase (GAPDH), b-actin (ACTB), b2-microglobulin (B2M), and lactate dehydrogenase A (LDHA). RNA-Seq data were uploaded to the Gene Expression Omnibus (GEO; GSE40996). Quantitative RT-PCR (qPCR) was used to validate gene expression data for FAM129A and SYNPO2 in the cDNA. The qPCRs were performed on a Roche LightCycler 480 using LightCycler 480 probes master and the universal ProbeLibrary system (Roche, Woerden, The Netherlands) on the basis of the reported cDNA sequences. For FAM129A we used forward primer ATGCTGTGGAGAGCTATGAGAA and reverse primer GGCTGGAAGAATTCGACATTTA (Roche universal probe nr. 17), and for SYNPO2 we used forward primer ACAGCAGACCTCA CAAGCAC and reverse primer CACTTGTTTTTCTGACAGGCTTT (Roche universal probe no. 63). Values were corrected using control

1078

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

VOL 187

2013

TABLE 1. CHARACTERISTICS OF PATIENTS WITH ASTHMA RECEIVING PREDNISOLONE OR PLACEBO Prednisolone

Placebo

Pre Subjects, n Age, yr* FEV1 postbronchodilator, L* FEV1 postbronchodilator, %pred† FEV1/FVC postbronchodilator, %pred† PC20, mg/ml‡

Post

Pre

Post

6 25 (21–27) 4.10 110 102 2.33

(3.49–4.60) (9) (10) (0.91–5.96)

6 24 (20–26) 4.15 111 102 3.20

(3.29–4.82) (10) (9) (1.03–9.91)

4.21 104 98 2.05

(3.45–4.99) (11) (8) (0.88–4.78)

4.27 105 99 2.02

(3.54–5.26) (10) (7) (0.97–4.21)

Definition of abbreviations: %pred ¼ percentage of the predicted value; PC20 ¼ provocative concentration of methacholine causing a 20% drop in FEV1; Post ¼ poststudy medication; Pre ¼ prestudy medication. * Mean (minimum–maximum). y Mean (SD). z Geometric mean (95% confidence interval).

hATPase 6 to calibrate hATPase 6 forward primer CATAATGACC CACCAATCACA and reverse primer GAGAGGGCCCCTGTTAGG (Roche universal probe no. 23). Samples were run in triplicate.

Statistical Analysis Subject characteristics were analyzed using unpaired t tests or MannWhitney U tests. Because the data concerned countable numbers, the comparison per gene was based on the Poisson–g mixture, which is also known as the negative binomial distribution. This distribution has two parameters only and the entire procedure is similar to a Student t test. It is implemented in the R-package DESeq, which is dedicated to the analysis of RNA sequencing data and includes normalization of RNASeq libraries based on size factors (25). Therefore, this R-package was used for comparison of the numbers of reads per gene between the prednisolone and placebo study groups. Spearman correlation coefficients were used to examine whether correlations existed between individual genes and airway function. For evaluating the association between airway function and gene expression, a P value less than 0.05 was considered statistically significant. Statistical analyses were performed with SPSS 20 (IBM Corporation, New York, NY) and R statistical program version 2.15.0.

Sample Size Evidence indicates that the number of genes in inflammatory pathways is already greater than 10,000 (26). Therefore, we conservatively assumed that the pathways associated with asthma altogether consist of at least 10,000 genes, which is approximately one-third of the 31,227 genes included in the hg19 human genome database tested. We presumed that the average fold change in gene expression in the prednisolone study group relative to the placebo study group was approximately 1.2, which coincides with a Cohen effect size of about 0.5. To correct for multiple testing, a significance level of 0.0001 was used in the statistical analysis. With these assumptions, we expected to find 68 true and 2 false significant genes (27, 28). This corresponds to a false discovery rate of approximately 3%.

RESULTS Subject Characteristics

The characteristics of the patients with asthma receiving either oral prednisolone or placebo were not significantly different at baseline (Table 1). All subjects completed the study and none withdrew because of any adverse effects of the study medication or procedures. Gene Sequencing by RNA-Seq

The average size of the endobronchial biopsy specimens ranged from 1 to 2 mm3. ASM was present in all biopsies and was successfully collected by LCM. Characteristics of the biopsies and of the RNA-Seq run were comparable between the two study groups (Table 2). For the prednisolone study group, the transcriptomes found by RNA-Seq were mapped to 13,076 and 11,804 unique genes before and after 14 days of oral prednisolone, respectively. These were 12,266 and 12,462 genes for the placebo study group before and after 14 days of placebo, respectively. Figure 1 presents the overview of the read numbers per gene before and after oral prednisolone or placebo. For the prednisolone study group, 85 genes had significantly different read numbers after 14 days of oral prednisolone than before taking the study medication. To clarify whether the change in read numbers of these 85 genes could be explained by the use of oral prednisolone, we compared it with the change in read numbers of the same 85 genes found in the placebo study group. Fifteen of these 85 genes maintained a significant change in read numbers after comparison with the placebo study group (Table 3; Figure 2). These 15 genes did not comprise inflammatory cytokines. No significant differences in read numbers of reference genes were found before and after 14 days of study medication. qPCR was used to confirm the data found by RNA-Seq. For the

TABLE 2. SAMPLE CHARACTERISTICS

RNA yield, ng; four biopsies per subject* cDNA yield, ng; 50-ng RNA input* cDNA library concentration, 3 109 mol/µl* RNA-Seq read number, n Before study medication After 14 d of study medication Read number mapped to hg19, % Before study medication After 14 d of study medication

Prednisolone

Placebo

126 (80–213) 1,863 (791–4,215) 1.02 (0.84–1.21)

171 (105–236) 1,717 (433–5,040) 1.24 (0.57–1.83)

427,628 327,857

319,192 357,275

92 93

92 92

Definition of abbreviations: cDNA ¼ complementary DNA; hg19 ¼ human genome version 19; RNA-Seq ¼ next-generation, high-throughput RNA sequencing. * Data are shown as means (range).

Yick, Zwinderman, Kunst, et al.: Prednisolone and ASM Gene Expression in Asthma

1079

and third ranking networks both have an IPA score of 3 and were associated with the network functions cellular movement, function, and maintenance, and endocrine system disorders. Transcriptomic Profile and Lung Function

To explore the effects of changes in transcriptomic profile after 14 days of oral prednisolone on airway function, we examined the relation between the 15 genes significantly changed by oral prednisolone and the change in lung function measurements. Among these, FAM129A (family with sequence similarity 129, member A) and SYNPO2 (synaptopodin 2) were correlated with the methacholine PC20. After a 14-day course, oral prednisolone induced a positive change, in other words an increase in read numbers of these two genes in all patients of the prednisolone study group, whereas this was not the case with placebo (Figure 5). Irrespective of this, an increase in the change in read numbers of both FAM129A (r ¼ 20.740, P , 0.01; Figure 5A) and SYNPO2 (r ¼ 20.746, P , 0.01; Figure 5B) per se was associated with a decrease in the fold change in methacholine PC20, in other words, reduced airway responsiveness. Figure 1. Gene read numbers before and after 14 days of prednisolone or placebo. Read numbers of individual unique genes found by RNASeq (next-generation, high-throughput RNA sequencing) before (x axis) and after (y axis) 14 days of study medication are shown for the prednisolone and placebo study groups.

gene FAM129A the change in mean crossing point (CP) value, a measure for the amount of target RNA in the specimen, from pre– to post–study medication was 20.98 and 0.62 for the prednisolone study group and placebo study group, respectively. The change in mean CP value of the gene SYNPO2 was 21.06 and 0.55 for the prednisolone study group and placebo study group, respectively (Figure 3). Pathway Analysis

The 15 genes that were significantly changed after 14 days of oral prednisolone were used to explore gene networks by the IPA application. In total three gene networks were found with an IPA score equal to or greater than 2. The highest scoring network had an IPA score of 32 and was associated with the network functions cellular growth, proliferation, and development (Figure 4). Key components in this network are ERK1/2 (extracellular signal–regulated kinase 1/2), UBC (ubiquitin C), and PPP2R1B (protein phosphatase 2, regulatory subunit A, b). The second

DISCUSSION The results of this study show that the transcriptomic profile of the ASM layer in the airway wall of patients with asthma changes after the use of oral prednisolone. Eighty-five unique genes in total had significantly different read numbers after 14 days of oral prednisolone use. After comparison with the placebo study group, 15 of these 85 genes maintained a significant change in read numbers. Pathway analysis revealed three gene networks that were associated with cellular functions including cellular growth, proliferation, and development. Remarkably, 2 of the 15 genes that were significantly changed by the use of oral prednisolone were correlated with airway hyperresponsiveness. These results indicate that glucocorticoids exert their effect at the transcriptomic level in ASM that is associated with improved airway hyperresponsiveness. To our knowledge this is the first study to show that oral prednisolone changes gene expression of ASM isolated from in vivo endobronchial biopsy specimens of patients with asthma. Gene networks that were found to be changed by glucocorticoids, using pathway analysis, included ERK1/2, UBC, and PPP2R1B as key nodes. In vitro experiments have shown that inhibition of ERK1/2 leads to a reduction in the level of ASM proliferation induced by chemokines (29). UBC, also known as ubiquitin C, is

TABLE 3. AIRWAY SMOOTH MUSCLE GENES AFFECTED BY ORAL PREDNISOLONE Gene Symbol

Gene Name

NCBI Gene ID

Log2 Fold Change*

P Value

PPP2R1B SCUBE3 ADAM22 GCC2 ACTA2 KIAA0319 FAM129A MYOCD NTN1 SYNPO2 CCDC30 KIF5C RBM14-RBM4 CATSPERB GRB14

Protein phosphatase 2, regulatory subunit A, b Signal peptide, CUB domain, EGF-like 3 ADAM metallopeptidase domain 22 GRIP and coiled-coil domain containing 2 Actin, a2, smooth muscle, aorta KIAA0319 Family with sequence similarity 129, member A Myocardin Netrin 1 Synaptopodin 2 Coiled-coil domain containing 30 Kinesin family member 5C RBM14-RBM4 readthrough Catsper channel auxiliary subunit b Growth factor receptor–bound protein 14

5519 222663 53616 9648 102620 9856 116496 93649 9423 171024 728621 3800 100526737 611169 2888

3.18 2.22 2.10 2.07 1.56 1.39 1.35 1.31 1.29 0.92 0.90 21.60 22.77 22.89 23.16

0.03 0.05 0.02 0.02 0.02 ,0.01 0.03 0.02 0.03 0.02 0.02 0.03 0.02 0.03 0.03

* Log2 fold change after versus before 14 days of oral prednisolone: a positive number signifies up-regulated after 14 days of oral prednisolone, whereas a negative number indicates down-regulation.

1080

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

Figure 2. Changes in read number per gene. The change in read number from before to after 14 days of study medication [D read number (post 2 pre)] per gene per subject is depicted together with the mean D value for the prednisolone study group (dotted green lines) and placebo study group (solid blue lines). The mean D value of each of these 15 genes was significantly different between the prednisolone and placebo study groups.

a signaling molecule involved in various cellular processes. It has been proposed to be involved in the regulation of tumor necrosis factor–induced nuclear factor-kB signaling, which plays an important role in inflammatory processes (30). PPP2R1B is a subunit of protein phosphatase 2 and has been implicated in the regulation of Akt (v-Akt murine thymoma viral oncogene homolog 1) activity in cardiomyocytes (31). Akt itself is a protein kinase that is involved in the PI3K (phosphatidylinositol 3kinase)/Akt/mTOR (mammalian target of rapamycin) signaling pathway and has been reported to be one of the complexes

VOL 187

2013

through which YKL-40 increases bronchial smooth muscle cell proliferation and migration (32). Therefore, the genes whose expression was changed by the use of oral prednisolone are involved in the regulation of ASM structure and cellular processes in the ASM layer. Fifteen genes had a significant change in expression after the use of prednisolone compared with placebo. Interestingly, these 15 genes also changed in the placebo group (Figure 2). As mentioned previously, pathway analyses revealed that these 15 genes were associated with various inflammatory and cell regulatory processes. It may be inferred that the changes in gene expression observed in the placebo study group reflect a natural course in biological activity in patients with controlled, mild asthma not receiving asthma therapy. Indeed, asthma is a dynamic disease with its inflammatory and airway remodeling processes being dependent on disease severity and activity. When prednisolone was prescribed, the expression of these genes was additionally changed, indicating that the course of biological activity was adjusted by glucocorticoids. The exact biological effects of each of these 15 genes have yet to be established. The genes GILZ (TSC22D3) and FKBP5 were found to be up- and down-regulated, respectively, in the ASM samples of the prednisolone group compared with placebo. This confirms part of the results presented by previous ASM gene expression studies using real-time RT-PCR (33). However, both genes did not reach statistical significance in the current study. This may be explained by the difference between the method of gene expression analysis (RNA-Seq vs. real-time RT-PCR), and tissue used (in vivo biopsies vs. cultured ASM cells). The changes in read number of FAM129A and SYNPO2 were associated with fold changes of PC20. In the placebo group, the changes in these genes ranged from positive to negative, whereas those in the prednisolone group were all positive (Figure 5). In other words, prednisolone induced an increase in gene expression level resulting in a positive change in read numbers (D). If FAM129A and SYNPO2 are truly affecting the PC20, the association between changes in these genes and those in PC20 should be consistent regardless of the intervention. Apparently,

Figure 3. Crossing point (CP) values by quantitative PCR (qPCR) before (Pre) and after (Post) study medication. qPCR was used to confirm the data found by RNA-Seq (next-generation, high-throughput RNA sequencing). The CP values (y axis) of (A) FAM129A and (B) SYNPO2 before and after study medication are depicted for the prednisolone and placebo study groups. The change in mean CP value of the gene FAM129A was 20.98 and 0.62 for the prednisolone and placebo study groups, respectively. The change in mean CP value of SYNPO2 was 21.06 and 0.55 for the prednisolone and placebo study groups, respectively.

Yick, Zwinderman, Kunst, et al.: Prednisolone and ASM Gene Expression in Asthma

1081

Figure 4. Gene network identification by Ingenuity Pathway Analysis (IPA). The highest scoring gene network with an IPA score of 32 is shown. Associated network functions included cellular growth, proliferation, and development. Each node in the network is a gene involved in this particular network. Red denotes up-regulation, green denotes down-regulation. Interactions between the various genes in the network are displayed as solid lines (direct interaction) or dotted lines (indirect interaction) connecting the various nodes. The more connecting lines converge on a node, the more interactions that node has, rendering it a key component in the network.

for both genes an increase in read numbers was correlated with a decrease in fold change PC20 and thereby with reduced airway responsiveness. These data indicate that airway responsiveness varies with FAM129A and SYNPO2 expression. Neither gene has been linked directly to asthma in the literature. FAM129A has been primarily associated with thyroid tumors (34). Although the exact function of FAM129A is not known yet, it tends to regulate protein phosphorylation and thereby potentially affects many cellular processes. In the case of SYNPO2, also known as myopodin, previous evidence has shown that it is associated with the early development of skeletal muscle cells, more specifically the sarcomeric Z-disc (35). In addition, SYNPO2 has been proposed as a serum response factor regulating the phenotype of cardiac smooth muscle cells (36, 37). This may imply that SYNPO2 is also important in the development and functioning of the ASM layer in asthma. Taken together, oral glucocorticoids not only have antiinflammatory properties,

they also exert their effect at the transcriptomic level in the ASM layer that was associated with a reduction in airway hyperresponsiveness. Whether this beneficial effect on airway function occurs secondary to the antiinflammatory properties or is primarily due to changes at the transcriptomic level of ASM, or a combination of both, remains to be established. We did not find evidence of changed transcription of inflammatory cytokines between the prednisolone and placebo study groups. This suggests that ASM is directly responsive to glucocorticoids, which has been demonstrated in vitro (11, 12). The strength of our study seems to be the inclusion of wellcharacterized patients with asthma with stable disease without glucocorticoids by any route of administration. Furthermore, none of the subjects used inhaled short-acting b2-agonists as rescue therapy during the study period. Therefore, the transcriptomic profile of the ASM layer was not affected by differences in medication use other than the study medication. In

1082

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

VOL 187

2013

Figure 5. Correlation of FAM129A and SYNPO2 with airway hyperresponsiveness. The changes (D) in read number of (A) FAM129A and (B) SYNPO2, and fold changes in PC20 (provocative concentration of methacholine causing a 20% drop in FEV1) after 14 days of study medication are depicted on the x axis and y axis, respectively. An increase in D read numbers of both genes was correlated with a decrease in fold change in PC20, in other words, reduced airway responsiveness.

addition, the study procedures for measuring lung function were identical and took place at identical time intervals preceding the bronchoscopy. Therefore, the change in ASM transcriptomic profile is not due to differences in lung function testing procedures. Importantly, a detailed a priori sample size calculation showed that the current false discovery rate was limited to 3%. This indicates that the 15 genes that changed their expression by glucocorticoids as compared with placebo were unlikely to be detected by chance. This is strengthened by the observed associations between changes in gene expression and changes in PC20. We used RNA-Seq as a powerful tool for the quantification of gene expression. The implementation of RNA-Seq for transcriptomic analysis of the ASM layer in the current study is novel and represents an unbiased way of gene expression profiling compared with microarrays (17). The detection of nucleotide sequences by RNA-Seq is not limited to predefined probe sets, which is an advantage for gene discovery. Degradation of RNA by RNases, which are abundantly present either in the cell itself or in the environment, may hamper gene expression analysis. However, by implementing strict operating procedures in our workflow and combining the Ovation RNA-Seq system with the GS FLX system we prevented this as much as possible. First, the Ovation kit requires an input of at least 500 pg for successful processing of RNA into amplified cDNA, which was appropriate for the amount of RNA isolated in the current study (Table 2). Second, isothermal, linear amplification is initiated at both the 39 end and randomly throughout the target transcriptome. Besides maximizing the coverage as reads are distributed across the whole transcriptome, this will minimize the potentially detrimental effects of degraded RNA on gene expression analysis as well. In addition, with 92% of the sequence reads mapped to the reference genome resulting in the identification of approximately 12,000 unique genes based on contigs with a median length of 260 bp, the coverage in the current study is approximately 213 (38). A few clinical aspects of our findings should be taken into account. First, prednisolone administered by the oral route was used as study medication. It is possible that inhaled glucocorticoids will yield different effects on ASM gene expression in

asthma, due to the local deposition in the lungs as compared with the systemic effect of prednisolone taken orally. However, the amount of inhaled glucocorticoids reaching the airways is highly dependent on the individual inhalation technique of a patient (39). Moreover, the deposition of inhaled glucocorticoids is not evenly distributed throughout the tracheobronchial tree and therefore may not represent transcriptomic profiles from biopsies taken at a specific location. We aimed to circumvent these aspects by using oral prednisolone as study medication. Second, the inflammatory and airway remodeling processes in asthma are dynamic and the extent of these processes may also vary at different locations in the bronchial tree (7). In addition, the composition of the ASM layer also varies between patients with asthma with varying disease severity (9). Evidence shows that the expression of ECM proteins and inflammatory cells inside the ASM layer itself tends to be associated with the activity of airway inflammation (40). This may also be reflected in the sample characteristics shown in Table 2 that vary between and within the study groups. The yield of RNA was dependent on the amount of ASM present within the biopsy specimens. To ensure a representative ASM transcriptomic profile the collection and processing of all four biopsies collected for each patient were performed in a standardized manner, including cutting up all four biopsies from each patient entirely, and isolating the ASM present within those sections by LCM. Furthermore, we included well-characterized patients with asthma based on strict inclusion criteria regarding airway function and medication usage to form a study population that is as homogeneous as possible in disease severity. The absence of differences in the expression of gene transcripts of, for example, inflammatory cytokines at baseline between and within the groups in the current study suggests that the disease severity and the ASM transcriptomic profiles of the study participants were comparable. Therefore, we believe that the carefully chosen inclusion and exclusion criteria were successful in minimizing the biological variability between the study participants as much as possible. Still, the transcriptomic profiles of the ASM layer from biopsies collected at B 7–9 of the right lung in the current study may

Yick, Zwinderman, Kunst, et al.: Prednisolone and ASM Gene Expression in Asthma

differ from other locations of the bronchial tree. In addition, such profiles may also deviate in patients with more severe asthma, due to a different composition of the ASM layer (9). Therefore, similar studies in severe asthma focusing on other locations of the airways are needed to complement our current study to provide a more comprehensive overview of ASM gene expression. In addition, studies implementing long-term glucocorticoid treatment should further clarify the specific pharmacodynamic properties of glucocorticoids on the gene expression of the ASM layer in asthma. The current transcriptomics study revealed that several genes, previously not considered to be associated with asthma, were changed by glucocorticoid treatment. In addition, we found gene networks that may clarify the effect of glucocorticoids on ASM gene expression in addition to their well-known antiinflammatory effect. These results suggest that the glucocorticoid-induced attenuation of airway hyperresponsiveness in the current study and the previously reported steroid-induced increase in bronchodilation by deep inspiration (15) are associated with the response of ASM itself to glucocorticoids. Nevertheless, the present data are limited to transcriptomics. Additional studies are needed to link these data with the protein level, which requires a multiscale, systems biology approach. This will allow further clarification of the biochemical processes in asthma from the translation of DNA to proteins via RNA, to the posttranslational modification and regulation of these proteins through the implementation of various “omics” technologies (18). Furthermore, it remains important to validate the gene sequencing results both internally and externally, ideally with the same gene sequencing platform that was used, which may be more appropriate than validation by qPCR alone. Our results may have important implications for future asthma treatment (41). By mapping out the key glucocorticoid-responsive elements in ASM, targeted interventions can be developed that enhance the beneficial effects of glucocorticoids while diminishing their unwanted and potentially harmful side effects. As glucocorticoids are the mainstay therapy for asthma, this development may have significant clinical impact. Conclusions

The transcriptomic profile of the airway smooth muscle layer in asthma was changed by a 14-day course of oral prednisolone. The expression of 15 novel genes that had not been linked to asthma previously was found to be altered by prednisolone. Two of these genes, FAM129A and SYNPO2, were associated with an improvement in airway hyperresponsiveness (methacholine PC20). Pathway analysis revealed three gene networks that were associated with cellular functions including cellular growth, proliferation, and development. These results indicate that glucocorticoids also exert their effect on the gene expression of airway smooth muscle in conjunction with their well-known antiinflammatory properties. Increased understanding of the direct or indirect pharmacodynamics of glucocorticoids in airway smooth muscle may help in developing novel medications with enhanced beneficial effects, but diminished side effects. Author disclosures are available with the text of this article at www.atsjournals.org. Acknowledgment: The authors thank Jeroen Vreijling (Department of Genome Analysis, Academic Medical Center, Amsterdam, The Netherlands) for support in performing the qPCRs.

References 1. Global Initiative for Asthma (GINA). GINA Report, Global Strategy for Asthma Management and Prevention [updated 2011 Dec; accessed 2013 Feb 10]. Available from: http://www.ginasthma.com

1083

2. Sterk PJ, Fabbri LM, Quanjer PH, Cockcroft DW, O’Byrne PM, Anderson SD, Juniper EF, Malo J-L. Airway responsiveness: standardized challenge testing with pharmacological, physical and sensitizing stimuli in adults. Report Working Party Standardization of Lung Function Tests, European Community for Steel and Coal. Official Statement of the European Respiratory Society. Eur Respir J Suppl 1993;16:53–83. 3. Reddel H, Jenkins C, Woolcock A. Diurnal variability—time to change asthma guidelines? BMJ 1999;319:45–47. 4. Slats AM, Janssen K, van Schadewijk A, van der Plas DT, Schot R, van den Aardweg JG, de Jongste JC, Hiemstra PS, Mauad T, Rabe KF, et al. Bronchial inflammation and airway responses to deep inspiration in asthma and chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2007;176:121–128. 5. Mauad T, Bel EH, Sterk PJ. Asthma therapy and airway remodeling. J Allergy Clin Immunol 2007;120:997–1009. 6. Brightling CE, Bradding P, Symon FA, Holgate ST, Wardlaw AJ, Pavord ID. Mast-cell infiltration of airway smooth muscle in asthma. N Engl J Med 2002;346:1699–1705. 7. Araujo BB, Dolhnikoff M, Silva LF, Elliot J, Lindeman JH, Ferreira DS, Mulder A, Gomes HA, Fernezlian SM, James A, et al. Extracellular matrix components and regulators in the airway smooth muscle in asthma. Eur Respir J 2008;32:61–69. 8. Yick CY, Ferreira DS, Annoni R, von der Thüsen JH, Kunst PW, Bel EH, Lutter R, Mauad T, Sterk PJ. Extracellular matrix in airway smooth muscle is associated with dynamics of airway function in asthma. Allergy 2012;67:552–559. 9. James AL, Elliot JG, Jones RL, Carroll ML, Mauad T, Bai TR, Abramson MJ, McKay KO, Green FH. Airway smooth muscle hypertrophy and hyperplasia in asthma. Am J Respir Crit Care Med 2012;185:1058–1064. 10. Sterk PJ, Yick CY, Slats AM. The secret life of steroids in asthma. Eur Respir J 2008;32:1135–1137. 11. Hakonarson H, Halapi E, Whelan R, Gulcher J, Stefansson K, Grunstein MM. Association between IL-1b/TNF-a induced glucocorticoid-sensitive changes in multiple gene expression and altered responsiveness in airway smooth muscle. Am J Respir Cell Mol Biol 2001;25:761–771. 12. Ammit AJ, Burgess JK, Hirst SJ, Hughes JM, Kaur M, Lau JY, Zuyderduyn S. The effect of asthma therapeutics on signalling and transcriptional regulation of airway smooth muscle function. Pulm Pharmacol Ther 2009;22:446–454. 13. Lakser OJ, Dowell ML, Lega F, Chen B, Lavoie T, Ferreira C, Pinto LH, Dulin NO, Kogut P, Churchill J, et al. Steroids augment relengthening of contracted airway smooth muscle: potential additional mechanism of benefit in asthma. Eur Respir J 2008;32: 1224–1230. 14. Masuno K, Haldar SM, Jeyaraj D, Mailloux CM, Huang X, Panettieri RA Jr, Jain MK, Gerber AN. Expression profiling identifies Klf15 as a glucocorticoid target that regulates airway hyperresponsiveness. Am J Respir Cell Mol Biol 2011;45:642–649. 15. Slats AM, Sont JK, van Klink HCJ, Bel EH, Sterk PJ. Improvement of bronchodilation following deep inspiration after a course of high-dose oral prednisone in asthma. Chest 2006;130:58–65. 16. Halayko AJ, Tran T, Ji SY, Yamasaki A, Gosens R. Airway smooth muscle phenotype and function: interactions with current asthma therapies. Curr Drug Targets 2006;7:525–540. 17. Metzker ML. Sequencing technologies—the next generation. Nat Rev Genet 2010;11:31–46. 18. Auffray C, Adcock IM, Chung KF, Djukanovic R, Pison C, Sterk PJ. An integrative systems biology approach to understanding pulmonary diseases. Chest 2010;137:1410–1416. 19. Miller MR, Hankinson J, Brusasco V, Burgos F, Casaburi R, Coates A, Crapo R, Enright P, van der Grinten CP, Gustafsson P, et al.; ATS/ ERS Task Force. Standardisation of spirometry. Eur Respir J 2005;26: 319–338. 20. Busse WW, Wanner A, Adams K, Reynolds HY, Castro M, Chowdhury B, Kraft M, Levine RJ, Peters SP, Sullivan EJ. Investigative bronchoprovocation and bronchoscopy in airway diseases. Am J Respir Crit Care Med 2005;172:807–816. 21. Kurn N, Chen P, Heath JD, Kopf-Sill A, Stephens KM, Wang S. Novel isothermal, linear nucleic acid amplification systems for highly multiplexed applications. Clin Chem 2005;51:1973–1981.

1084

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

22. Weterman MA, Sorrentino V, Kasher PR, Jakobs ME, van Engelen BG, Fluiter K, de Wissel MB, Sizarov A, Nürnberg G, Nürnberg P, et al. A frameshift mutation in LRSAM1 is responsible for a dominant hereditary polyneuropathy. Hum Mol Genet 2012;21:358–370. 23. UCSC Genome Bioinformatics Group. UCSC Genome Bioinformatics [hg19 assembly 2009 Feb; accessed 2013 Feb 10]. Available from: http://genome.ucsc.edu 24. Ingenuity Systems [accessed 2013 Feb 10]. Available from: http://www. ingenuity.com 25. Anders S, Huber W. Differential expression analysis for sequence count data. Genome Biol 2010;11:R106. 26. Loza MJ, McCall CE, Li L, Isaacs WB, Xu J, Chang BL. Assembly of inflammation-related genes for pathway-focused genetic analysis. PLoS ONE 2007;2:e1035. 27. Ferreira JA, Zwinderman AH. On the Benjamini-Hochberg method. Ann Stat 2006;34:1827–1849. 28. Ferreira JA, Zwinderman AH. Approximate sample size calculations with microarray data: an illustration. Stat Appl Genet Mol Biol 2006;5:Article25. 29. Halwani R, Al-Abri J, Beland M, Al-Jahdali H, Halayko AJ, Lee TH, AlMuhsen S, Hamid Q. CC and CXC chemokines induce airway smooth muscle proliferation and survival. J Immunol 2011;186:4156–4163. 30. Verhelst K, Verstrepen L, Carpentier I, Beyaert R. Linear ubiquitination in NF-kB signaling and inflammation: what we do understand and what we do not. Biochem Pharmacol 2011;82:1057–1065. 31. Zuluaga S, Alvarez-Barrientos A, Gutiérrez-Uzquiza A, Benito M, Nebreda AR, Porras A. Negative regulation of Akt activity by p38a MAP kinase in cardiomyocytes involves membrane localization of PP2A through interaction with caveolin-1. Cell Signal 2007;19:62–74. 32. Bara I, Ozier A, Girodet PO, Carvalho G, Cattiaux J, Begueret H, Thumerel M, Ousova O, Kolbeck R, Coyle AJ, et al. Role of YKL-40 in bronchial smooth muscle remodeling in asthma. Am J Respir Crit Care Med 2012;185:715–722.

VOL 187

2013

33. Kelly MM, King EM, Rider CF, Gwozd C, Holden NS, Eddleston J, Zuraw B, Leigh R, O’Byrne PM, Newton R. Corticosteroid-induced gene expression in allergen-challenged asthmatic subjects taking inhaled budesonide. Br J Pharmacol 2012;165:1737–1747. 34. Matsumoto F, Fujii H, Abe M, Kajino K, Kobayashi T, Matsumoto T, Ikeda K, Hino O. A novel tumor marker, Niban, is expressed in subsets of thyroid tumors and Hashimoto’s thyroiditis. Hum Pathol 2006;37:1592–1600. 35. Linnemann A, van der Ven PF, Vakeel P, Albinus B, Simonis D, Bendas G, Schenk JA, Micheel B, Kley RA, Fürst DO. The sarcomeric Z-disc component myopodin is a multiadapter protein that interacts with filamin and a-actinin. Eur J Cell Biol 2010;89:681–692. 36. Du KL, Ip HS, Li J, Chen M, Dandre F, Yu W, Lu MM, Owens GK, Parmacek MS. Myocardin is a critical serum response factor cofactor in the transcriptional program regulating smooth muscle cell differentiation. Mol Cell Biol 2003;23:2425–2437. 37. Cao D, Wang C, Tang R, Chen H, Zhang Z, Tatsuguchi M, Wang DZ. Acetylation of myocardin is required for the activation of cardiac and smooth muscle genes. J Biol Chem 2012;287:38495–38504. 38. Lander ES, Waterman MS. Genomic mapping by fingerprinting random clones: a mathematical analysis. Genomics 1988;2:231–239. 39. Cochrane MG, Bala MV, Downs KE, Mauskopf J, Ben-Joseph RH. Inhaled corticosteroids for asthma therapy: patient compliance, devices, and inhalation technique. Chest 2000;117:542–550. 40. Slats AM, Janssen K, van Schadewijk A, van der Plas DT, Schot R, van den Aardweg JG, de Jongste JC, Hiemstra PS, Mauad T, Rabe KF, et al. Expression of smooth muscle and extracellular matrix proteins in relation to airway function in asthma. J Allergy Clin Immunol 2008; 121:1196–1202. 41. Siddiqui S, Redhu NS, Ojo OO, Liu B, Irechukwu N, Billington C, Janssen L, Moir LM. Emerging airway smooth muscle targets to treat asthma. Pulm Pharmacol Ther 2013;26:132–144.