| | Genome-wide gene expression profiling in children with non-obese obstructive sleep apnea☆Received 13 September 2007; received in revised form 31 October 2007; accepted 3 November 2007. Abstract BackgroundObstructive sleep apnea (OSA) is a multi-factorial and highly prevalent disorder in which both genetic and environmental factors may be involved. If left untreated, OSA may lead to significant cardiovascular and neurocognitive and behavioral morbidities. We hypothesized that pediatric OSA would lead to altered gene expression in circulating leukocytes. Methods and resultsOligonucleotide-based microarray technology was used to identify mRNAs that may be differentially regulated in non-obese children with polysomnographically-established OSA compared to matched control children. Total morning blood RNA from 40 children (20 OSA and 20 controls) was extracted, labeled, and hybridized onto independent oligonucleotide-based microarrays. Of the 44,000 transcripts, 1217 transcripts were differentially expressed in OSA (p-value <0.05), with 68 transcripts (38 RefSeq accession numbers, 30 ESTs) fulfilling high stringency criteria. False Discovery rate (FDR) was used to determine the significance-difference of OSA vs. normal samples. Microarray data were further validated using quantitative RT-PCR techniques. Biological pathways pertinent to the differentially expressed genes were explored and revealed prominent involvement of inflammatory pathways. ConclusionsRNA derived from peripheral leukocytes confirms the presence of altered expression of functionally relevant gene clusters in pediatric OSA. Large-scale genomic approaches may provide further insights into adaptive and end-organ injury related mechanisms in the context of OSA in children. 1. Introduction  Obstructive sleep apnea (OSA) is a relatively prevalent disorder across the lifespan, in which both genetic and environmental factors may be involved [1]. OSA is characterized by episodic partial or complete upper airway obstruction during sleep in association with loud snoring, altered gas exchange, and sleep fragmentation. This condition may affect up to 3% of otherwise healthy school-aged children [2], and has been associated with substantial cardiovascular, metabolic, and neurocognitive morbidities [3], [4], [5], [6]. Indeed, schooling and behavioral problems such as restlessness, inattention and impulsiveness, aggressive behavior, excessive daytime sleepiness, and poor test performances have been repeatedly reported in children with OSA [6], [7], [8], [9], [10]. In addition, systemic and pulmonary hypertension, reduced somatic growth, mood disturbances, and decreased quality of life may develop in pediatric OSA patients [11], [12]. In children, hypertrophy of the tonsils and adenoids in the upper airway is the most frequent and prominent abnormality associated with OSA, such that surgical extirpation of the enlarged upper airway lymphoid tissue is usually the initial management approach [13], [14]. The definitive diagnosis of OSA currently requires an overnight polysomnographic evaluation in a sleep laboratory, and is therefore an onerous and labor intensive procedure, such that delays in the timely diagnosis and treatment are frequent occurrences. Although much has been learned on the pathophysiology and consequences of pediatric OSA in the last 30 years, the mechanisms and specific genes associated with such processes remain poorly defined. In the last decade, development of high-throughput technologies, such as gene expression profiling using microarrays, has become a fundamental approach for identifying potential diagnostic and therapeutic targets for many diseases. The information gained by such a non a priori approach offers an unprecedented opportunity to fully characterize biological processes, since DNA microarrays yield a simultaneous measurement of gene expression levels for thousands of genes or for the whole genome, thereby allowing for analysis of differential gene expression patterns [15], [16]. Several microarray platforms are currently available [15], [16], [17], [18] and use different approaches to the construction, layout, optimization, hybridization, image acquisition and data extraction methods. Several reviews using different microarray platforms have been reported and summarize the major and most salient points related to their inherent advantages and limitations [19], [20]. Although microarrays have been used to identify molecular signatures for many diseases, such as asthma [21], and heart disease [22], there have been only a few reports on the use of this technology in sleep-related issues [23], [24]. Furthermore, we are aware of only one published study in four adult patients with OSA in whom microarrays were employed. In this report, differential expression of genes mediating oxidative stress was found and postulated as playing an important role in end-organ injury associated with OSA [25]. Clearly, the characterization of specific genes and particular biological pathways in children with OSA using high-throughput gene expression may further enhance our knowledge of this condition in the pediatric age range and allow for reliable and convenient clinical approaches for diagnosis and treatment of these children in the future. Royce et al. [26] reported that the applications of microarrays span from the bench to the bedside, thereby providing tools that require less effort, expense, and sample amount than any other technology. Based on aforementioned considerations, we hypothesized that specific changes in gene expression would occur in children with OSA. Therefore, the aims of the present study were to examine global changes in gene expression profiles in children with OSA, identify differentially expressed genes, and validate some of the latter using other molecular techniques such as QRT-PCR. 2. Materials and methods 2.1. Subjects Approval for the study was obtained from the institutional review board of the University of Louisville School of Medicine. Parental informed consent and child assent, in the presence of a parent, were obtained. Consecutive pre-pubertal non-obese children between the ages of four and nine years of age with a polysomnographic diagnosis of OSA (see below) were identified and recruited to the study. Control subjects were recruited from an ongoing large-scale population study and were initially screened for symptoms of OSA using a validated questionnaire [27]; they were invited to participate if they had no symptoms of sleep disordered breathing and their overnight polysomnogram was within normal limits (see below). Children were excluded if they had any chronic medical condition, were receiving medications, and if they had any genetic or craniofacial syndromes. The control children were non-obese and were matched for age, gender and ethnicity. In addition, chronic medical conditions such as Down syndrome, craniofacial or known genetic syndromes, a known episode of infection in the eight weeks preceding the sleep study, asthma or allergies receiving specific therapy (desensitization, leukotriene inhibitors, steroids (topical or systemic)) were excluded from the study. 2.2. Overnight polysomnography A standard overnight multichannel polysomnographic evaluation was performed in the sleep laboratory as described previously [28]. Sleep architecture was assessed by standard techniques [29]. The proportion of time spent in each sleep stage was expressed as percentage of total sleep time (%TST). Central, obstructive, and mixed apneic events were counted. Obstructive apnea was defined as the absence of airflow with continued chest wall and abdominal movement for duration of at least two breaths [28], [30]. Hypopneas were defined as a decrease in oronasal flow of ⩾50% with a corresponding decrease in SpO2 of ⩾4% and/or arousal [28]. The obstructive apnea/hypopnea index [9] was defined as the number of apneas and hypopneas per hour of TST. Arousals were defined as recommended by the American Sleep Disorders Association Task Force report [31], [32] and included respiratory-related (occurring immediately following an apnea, hypopnea, or snore), technician-induced, and spontaneous arousals. Arousals were expressed as the total number of arousals per hour of sleep time (arousal index). Control children required the presence of an AHI ⩽1 in the absence of a history of snoring and when no snoring was detected during the sleep study. Habitually snoring children with AHI>2/h TST were considered to have OSA [28]. 2.3. Body mass index Height and weight were obtained using standard techniques from each child. BMI was then calculated (body mass/height2) and was expressed as BMI z score using an online BMI z score calculator (http://www.cdc.gov/epiinfo/). Children with BMI z score values exceeding 1.20 were classified as fulfilling the criteria for overweight/obesity [33] and were excluded from this study. For demographic and sleep measures, data are presented as means SEM unless otherwise indicated. All analyses were conducted using SPSS software (version 11.5; SPPS, Inc., Chicago, IL). Comparisons of demographics were made with independent t-tests or by analysis of the variance (ANOVA) followed by post hoc comparisons, with p-values adjusted for unequal variances when appropriate (Levene’s test for equality of variances). All p-values reported are two-tailed with statistical significance set at <0.05. 2.4. RNA isolation and quality control Following the sleep study, fasting peripheral blood samples were drawn from all the children within the first hour after awakening in vacutainer tubes containing EDTA (Becton Dickinson, Franklin Lakes, NJ, USA). Total RNA was extracted from the peripheral leukocytes in the blood samples using PAXgene Mini columns and were DNase treated (Qiagen, Valencia, CA) according to the manufacturer protocol. The RNA quality and integrity were determined using the Eukaryote Total RNA Nano 6000 LabChip assay (Agilent Technologies) on the Agilent 2100 Bioanalyzer using Agilent’s RNA Integrity Number [34] software. All RNA samples showed A260/280 ratios between 1.9 and 2.1. RNA samples were quantified by measuring A260nm on a UV/vis spectrophotometer (ND-1000, NanoDrop Technologies, Wilmington, DE). The yield of RNA from peripheral blood leukocytes varies, but typically, the amount of RNA isolated revolves around 2–3μg from 2.5-mL whole blood. The RIN is a software tool that scans the peaks of RNA electropherograms for RNA intactness. All the purified samples were stored at −80°C until further analyses. 2.7. Microarray data analyses The microarray slides were scanned immediately after hybridization using an Agilent dual-laser Microarray Scanner, which has powerful auto-focusing for high resolution feature scanning and high-throughput analysis of multiple slides. The digitized images were acquired using Agilent Feature Extraction (FE) software v.9.5 with settings that select the locally weighted linear regression curve fit (lowess) normalization method after background subtraction. Data were filtered by first excluding the features automatically flagged by Agilent Feature Extraction Software. The SpotAnalyzer and the PolyOutlierFlagger algorithms define the features of flag outliers. Spots are flagged if pixel variation is too high or if the spot intensity is found to be an outlier. Images were captured and saved by Feature Extraction software v. 9.5 (Agilent) and then analyzed by GeneSpring v. 7.3 software (Agilent). All saturated and irregular spots were filtered out during data analysis using GeneSpring software. A list of differentially expressed genes were identified using a false discovery rate (FDR) correction [35]. Briefly, microarray data were up-loaded into the GeneSpring software, and differentially expressed genes were identified from the normalized data. The customary p-value cut-off was selected as <0.05 using the Benjamini and Hochberg method. However, to increase the stringency of our gene selection criteria, the data were further filtered based on a p-value cut-off of <0.01. The fold change values for the differentially expressed (up- and down-regulated) genes were calculated by ratios of intensities between the two groups (OSA vs. controls). The t-test p-values (in which two-sample unequal variances were used) were utilized to detect the significance of differences between the two groups. Log ratios of intensities for individual genes were determined for each OSA/control from which the mean value of log ratios for each sample group was obtained. Data were subjected to log transformation to improve the characteristics of the distribution of the expression values (to make the resulting variable normally distributed). The quality for each microarray experiment was evaluated individually based on the spike-in controls. Agilent FE software electronically generates a QC report to evaluate the results. The QC Tools report was used to generate one report for all the arrays used in the experiments (40 arrays). The system sensitivity and dynamic range of the each microarray was measured by the performance of the spike-in controls. The number of biological replicates planned a priori for these experiments would predict a high level of reliability for the selection criteria stipulated above. 2.8. Real-time RT-PCR Quantitative real time RT-PCR analyses were performed for a set of six selected genes using ABI PRISM 7500 System (Applied Biosystems, Foster City, CA). The same total RNA was used for both microarray and RT-PCR experiments. cDNA synthesis was performed using a High-Capacity cDNA Archive Kit (Applied Biosystems, Foster City, CA). A housekeeping gene, ribosomal 18S rRNA, was used as a reference gene to normalize the expression ratios for the gene of interest. The primer sequences of the gene of interest were designed to be within the same region of the microarray sequence probes. One microgram of total RNA from OSA and control group samples was used to generate cDNA templates for RT-PCR, and a TaqMan® Master Mix Reagent Kit (Applied Biosystems, Foster City, CA) was used to amplify and quantify each transcript of interest in 25μl reactions. Triplicate PCR reactions were performed in 96-well plates for each gene in parallel with the 18S rRNA. The steps involved in the reaction program included: the initial step of 2min at 50°C; denaturation at 95°C for 10min, followed by 45 thermal cycles of denaturation (15s at 95°C) and elongation (1min at 60°C). The expression values were obtained from the cycle number (Ct value) using the Biosystems analysis software. All the genes of interest and 18S rRNA were performed in triplicates to determine the Ct-diff. These Ct values were averaged and the difference between the 18S Ct (Avg) and the gene of interest Ct (Avg) was calculated (Ct-diff). The relative expression of the gene of interest was analyzed using the  method [36]. Quantitative results are expressed as mean±standard deviation (SD). Statistical significance was evaluated by the Student’s t-test. 2.9. Biological pathways Microarrays generate large quantities of gene expression data in an attempt to understand complicated biological systems. To identify groups of genes that are similarly regulated across the biological samples, a variety of mathematical methods have been developed that include parametric calculations and non-parametric calculations. Efficient analysis of the biology behind microarray data requires standardized vocabularies and ontologies for describing genes, gene products, and their functions. The Gene Ontology (GO) is a common controlled vocabulary of terms and phrases describing the function of genes and gene products (http://www.geneontology.org/). This approach takes a list of differentially expressed genes and uses statistical analysis to identify the GO categories (biological processes, cellular components, and molecular functions). The potential of GO in biological reasoning is now increasingly recognized in biomedicine. To examine the potential GO distribution patterns of those genes showing differential expression between OSA children and controls, the differentially expressed genes generated by microarray were imported into PathwayStudio software (Ariadne Genomics Inc., Rockville, MD) to build biological pathways. This software extracts functional information on specific genes from the ResNet database using a natural language processing algorithm called MedScan. Data analyzed through this program could then be resolved into cogent models of the specific biological pathways activated under the experimental conditions used in the microarray analyses. This PathwayStudio is a software tool for biological pathway analysis based on Gene Ontology (GO) designed to describe key aspects of the molecular function, biological process and cellular component of gene products. Selected differentially expressed genes were imported into the PathwayStudio to identify all known relationships between the differentially expressed genes. This software utilizes a proprietary database containing a comprehensive set of pathway and molecular interactions uniquely representative of PubMed in its entirety (14,000,000 abstracts) with protein interactions to generate a biological association network of known protein interactions. 3. Results 3.2. RNA quality control The quality of microarrays data is largely dependent on the quality of RNA. Thus, we analyzed total RNA from OSA and control samples that were used in microarray experiments. A representative example of an electropherogram is shown in Fig. 1. Overall quality for all RNA samples was very high as determined by gel capillary electrophoresis and RNA Integrity Numbers (RIN) [34] (Fig. 1A and B). All samples passed quality control criteria and were subsequently used in analyses. The ratio for 28S/18S rRNA for all samples was between 1.8 and 2.1, and their RIN was in the 9.6–9.8 range. Of note, RIN was rated from 1 to 10 with 1 being the most degraded and 10 the most intact. The total RNA extracted from 20 children with OSA and 20 matched controls were then hybridized onto whole human genome arrays. 3.3. Gene expression profiling analysis To identify and characterize the changes of gene expression levels in children with OSA and control groups, a whole human genome long oligouncleotide-based microarray (60-mer) containing 44,000 transcripts was utilized. Of the 44K transcripts spotted on the microarrays, there were about 41K transcripts representing the whole human genome. Data analysis started by background subtraction following filtration and normalization. Of the 41K transcripts, following data filtration for all 40 microarrays used in this study, 38K transcripts were detected in all microarrays. The data were normalized to the median net signal intensities (raw data-background) for each spot on the arrays. The results of repeated hybridizations were analyzed by GeneSpring 7.3 software, and signal intensities (excluding outliers and flagged spots) were initially filtered for genes/ESTs commonly expressed in all slides. Of the 44,000 transcripts, there were 1217 transcripts (629 RefSeq accession numbers, 588 ESTs) in OSA that were differentially expressed (p<0.05). Among those differentially expressed genes, there were 68 transcripts (38 RefSeq accession numbers, 30 ESTs) that displayed differential expression using markedly stringent criteria (p<0.01). Table 3 shows the top 10 differentially expressed up-and down-regulated genes in OSA including their RefSeq accession numbers and chromosome location. | | |  | Description | GeneName | RefSeq # | Fold change | p-value | Chromosome | |
|---|
| Up-regulated | | | RAP1 interacting factor homolog | RIF1 | NM_018151 | 1.22 | 0.01 | 2q23.3 | | | Serologically defined colon cancer antigen 8 | SDCCAG8 | NM_006642 | 1.22 | 0.002 | 1q43-q44 | | | Spectrin repeat containing, nuclear envelope 2 | SYNE2 | NM_015180 | 1.23 | 7E-05 | 14q23.2 | | | UBX domain containing 2 | UBXD2 | NM_014607 | 1.23 | 0.009 | 2q21.3 | | | Ninein (GSK3B interacting protein) | NIN | NM_182944 | 1.26 | 0.008 | 14q22.1 | | | Kinesin family member 3C | KIF3C | NM_002254 | 1.27 | 0.007 | 2p23 | | | Oxysterol binding protein-like 8 | OSBPL8 | NM_020841 | 1.33 | 0.01 | 12q14 | | | Interleukin 27 | IL27 | NM_145659 | 1.96 | 0.007 | NA | | | Protease, serine, 36 | PRSS36 | NM_173502 | 2.20 | 0.009 | 16p11.2 | | | Malonyl-CoA:acyl carrier protein transacylase, mitochondrial | MT | NM_014507 | 2.80 | 0.005 | 22q13.31 | | | | | | Down-regulated | | | MEGF11 protein | MEGF11 | NM_032445 | −1.25 | 0.007 | 15q22.31 | | | Transgelin | TAGLN | NM_001001522 | −1.25 | 0.006 | 11q23.2 | | | Serpin peptidase inhibitor, clade A (alpha-1 antiproteinase, antitrypsin) | SERPINA10 | NM_016186 | −1.20 | 0.01 | 14q32.13 | | | Hypothetical protein LOC92270 | LOC92270 | NM_001017971 | −1.20 | 0.0002 | 5q14.2 | | | Chromosome 20 open reading frame 54 | C20orf54 | NM_033409 | −1.20 | 0.0039 | 20p13 | | | Pyrroline-5-carboxylate reductase 1 | PYCR1 | NM_153824 | −1.18 | 0.006 | 17q25.3 | | | Potassium intermediate/small conductance calcium-activated channel | KCNN1 | NM_002248 | −1.18 | 0.009 | 19p13.1 | | | Centrosomal protein 68kDa | CEP68 | NM_015147 | −1.20 | 0.009 | 2p14 | | | FLJ32363 protein | FLJ32363 | NM_198566 | −1.15 | 0.01 | 5p12 | | | Growth arrest-specific 2 like 1 | GAS2L1 | NM_152236 | −1.15 | 0.004 | 22q12.2 | | | | |
3.4. Biological pathways To better understand potential pathways that might be implicated in biological processes affected among children with OSA, 47 highly significant differentially expressed genes with p-values <0.01 were imported onto the PathwayStudio software (Ariadne Genomics Inc.). The data presented in Fig. 2A shows a very complex biological pathway which indicates that many genes and gene product interactions are involved. In addition, to the 47 genes identified in microarray data, there were approximately 200 candidate genes that were associated with these 47 genes in the same pathways. These 247 genes led to further identification of 23 putative cellular processes (yellow color) and 32 functional processes (gold color). This complex hierarchical pathway represents an example of what could be potentially occurring at the biological level among children with OSA. Due to the high level of complexity of this pathway, we next focused more specifically on inflammatory responses, cell survival, and cell proliferation and differentiation pathways (Fig. 2B). As indicated in Fig. 2B, 151 genes were identified as being associated with the inflammatory response pathway (yellow color), and two functional classes recruiting transcription factors and cytokine activity emerged (gold color). 3.5. Functional categories of regulated genes The highly significant differentially expressed genes identified in all OSA patients were also classified based on their known involvement in specific biological processes using the Gene Ontology (GO) database (http://www.geneontology.org). Of the differentially expressed genes (1217 transcripts), there were 895 transcripts that were involved in GO and classified into three functional categories, namely 524 transcripts in biological process (58%; i.e., the biological objective to which the gene product contributes), 506 transcripts in cellular components (56%; the location in the cell where a gene product is active), and 593 transcripts in molecular function (66%; i.e., the biochemical activity of the gene product). Of the highly statistical expressed genes (68 genes) with p-value <0.01, there were 65 genes identified to be involved in GO and they were classified as follows: 37 genes (56%) in cellular processes, 43 genes (66%) in biological processes and 46 genes (70%) in molecular functions. Since multifunctional proteins may be assigned several annotations corresponding to their different functions, we further analyzed potential associations with a particular gene product, and found that among the 46 genes putatively ascribed to molecular functions, their major targets involved S-S-acyltransferase activity and protein binding and nucleotide binding. 3.6. Comparison of real-time PCR with microarray data We selected six genes identified though the oligonucleotide-based microarray analysis for additional confirmation using quantitative RT-PCR. RT-PCR analysis of mRNA expression changes was performed on two genes that were up-regulated (Malonyl-CoA:ACP acyltransferase, Interleukin 27), down-regulated (Multiple EGF-like-domains 11, Amino acid transporter), or remained unchanged in the presence of OSA (Diphthamide 4, and Zinc finger protein 430). All of the selected genes confirmed the changes in expression patterns as previously identified in the microarrays. 4. Discussion To the best of our knowledge, this study represents the first attempt to characterize genome-wide gene expression profiling analysis of circulating leukocytes among non-obese children with OSA. The major findings support the activation and regulation of several functionally-related pathways in children with OSA, a significance that, while completely unexplored, is clearly of great potential interest for future discovery. Among such pathways, up-regulation and modulation of inflammation was prominently modified by the presence of OSA, lending previous support to our contention that OSA is indeed a systemic inflammatory disorder [11]. Before we discuss the potential significance of our findings, some methodological issues deserve comment, in particular subject selection considerations. As our major aim, we used the whole genome to screen for changes in gene expression profiling. To make optimal use of the microarray technology for investigating potentially relevant biological processes in pediatric OSA, we gathered a relatively large number of very carefully matched subjects and delimited the age of our subjects to the age in which there is a peak in the prevalence of OSA. We also excluded obese children in order to better demarcate genes that are strictly involved in the OSA disease process, as opposed to genes that may be modulated by the underlying presence of obesity. Of course, future studies will be needed to better delineate the relative contributions of obesity to the systemic biological responses recruited by OSA. To this effect, we identified patients within the typical range of OSA severity as usually seen in pediatric sleep centers. However, while this approach was dictated by our interest in characterizing the more frequently encountered OSA patient prototypes, there will undoubtedly be great value in further investigation of other severity based categories, such as habitual snoring, upper airway resistance syndrome, isolated obstructive alveolar hypoventilation, and several strata of OSA severity ranging from mild to severe. An additional potential limitation of our study is that we are restricted to gene expression changes within circulating white blood cells, which may not accurately reflect the response patterns within specific end-organs, and therefore may provide only a relatively unspecific cross-sectional picture of altered gene expression. Notwithstanding, if indeed our current findings are further confirmed in larger cohorts, then the highlighted groups of genes with differential expression levels in OSA samples may serve as a specific disease signature. Finally, we applied stringent criteria for significant changes in gene expression, and thus minimized the number of false positives, with the caveat that some genes of significance that are only modified in a subset of OSA patients would have been missed using such an approach. Some of the highly up-regulated genes consistently identified in OSA children included a malonyl group from malonyl-CoA and a mitochondrial acyl carrier protein (MCAT). The MCAT protein is found exclusively in the mitochondrion, where it catalyzes the transfer of a malonyl group from malonyl-CoA to the mitochondrial acyl carrier protein. The encoded protein may be part of a fatty acid synthase complex. MCAT has not only been considered as an attractive drug target in the discovery of antibacterial agents [38], but more importantly, in the context of OSA, increased malonyl-CoA levels in muscles from obese and type 2 diabetic subjects lead to decreased fatty acid oxidation and increased lipogenesis [39]. Furthermore, MCAT levels are elevated in insulin resistant tissues [40]. Considering the potential metabolic adverse consequences of OSA in children [5], [41], [42], [43], [44], and more particularly, the lipidogenic and diabetogenic effects, both sets of consequences are ascribed to the intermittent hypoxia and sleep fragmentation that characterize OSA [45], [46], [47]. Another gene that showed highly significant changes in expression was serine 36 protease. Analysis of post-translational processing mechanisms revealed that this novel polyprotein is a secreted enzyme whose three protease domains remain an integral part of a single polypeptide chain. This gene is located on chromosome 6p11.2. The molecular function of serine 36 protease involves serine-type endopeptidase activity during cellular stress conditions [48] and is highly expressed in skeletal muscle, the liver, and heart. The third highly significant regulated gene in OSA was interleukin 27 (IL27). This gene is one of the subunits of a heterodimeric cytokine complex related to interleukin 12A (IL12A). IL27 interacts with Epstein-Barr virus induced gene 3 (EBI3), a protein similar to interleukin 12B (IL12B), and forms a complex that has been shown to drive the rapid expansion of naïve T cells, but not memory CD4(+) T cells. The complex is also found to synergize strongly with interleukin 12 to trigger interferon gamma (IFNγ) production by CD4(+) T cells. The biological effects of this cytokine are mediated by the class I cytokine receptor (WSX1/TCRR), which has major implications for a variety of systemic inflammatory responses. For example, Stumhofer et al. [49] reported that IL27 negatively regulates the development of interleukin 17-producing T helper cells during chronic inflammation of the central nervous system. The molecular function of IL27 appears particularly important in the facilitation of receptor binding, and the biological function of its gene involves positive regulation of IFNγ biosynthetic processes, regulation of T cell proliferation, and regulation of T helper 1 cell differentiation. Based on these considerations, it was not surprising that the viable theoretical interactions among the differentially expressed genes reported herein would involve multiple biological pathways in a rather highly complex and orchestrated scenario (Fig. 2). To facilitate our understanding and focus our interest on a pathway of demonstrated interest and relevance to OSA, we restricted our detailed analysis on the proposed inflammatory pathway, activated in the context of pediatric OSA (Fig. 2). As mentioned above, one of the major genes associated with the inflammatory response was IL27, and our current findings clearly justify a more extensive exploration of the mechanistic role(s) played by this recently discovered interleukin in OSA. In addition, considering the large body of work previously reported by our laboratory on the regulation of specific inflammatory proteins in pediatric OSA [50], [51], [52], [53], we explored, within the microarray, data changes in specific inflammatory genes corresponding to the proteins that have previously been investigated. As such, C-reactive protein gene expression (NM_000567) was up-regulated in OSA (p-value=0.012); interleukin 6 (NM_000600) showed no differences in mRNA expression between OSA and control children (p-value=0.96), and similarly, the tumor necrosis factor superfamily, member 2 (NM_000594), showed no changes either (p-value=0.94). However, we should point out that inferences from transcriptomic data to protein expression data in plasma should not be pursued since very little overlap may occur between these two biological assessments. Due to the complexity of the biological pathways presented in Fig. 2A and B, we highlighted three genes that show the highest number of interactions within the inflammatory network identified for pediatric OSA in this study (Fig. 2A). As such, the growth factor receptor-bound protein 2 (GRB2) is known to bind to the epidermal growth factor receptor, and contains one SH2 domain and two SH3 domains. This gene is similar to the Sem5 gene of Caenorhabditis elegans, which is involved in inflammatory signal transduction pathways. GRB2 has two alternatively spliced transcript variants encoded for different isoforms and located on chromosome 17q24-q25. The GRB3 gene is involved in several molecular functions such as SH3/SH2 adaptor activity, epidermal growth factor receptor binding, and insulin receptor substrate binding. Furthermore, Grb2 functions as an adaptor protein in signal transduction pathways in which protein–tyrosine kinases are involved [54]. Another gene highlighted in Fig. 2B, the RAS-related C3 botulinum substrate 1 (RAC1) gene, is a GTPase which belongs to the RAS superfamily of small GTP-binding proteins. Members of this superfamily appear to regulate a diverse array of cellular events, including control of cell growth, cytoskeletal reorganization, and the activation of protein kinases. Several alternatively spliced transcript variants of this gene have been described, but the full-length nature of some of these variants has not been determined. The RAC1 gene is located on chromosome 7p22. The molecular function of this gene is involved in GRP binding, GTP-dependent protein binding, GTPase activity, and nucleotide binding. The third and final gene highlighted in Fig. 2B, is cytochrome c (CYCS). This is the somatic variant that encodes for cytochrome c, a component of the electron transport chain in mitochondria. Cytochrome c is involved in initiation of apoptosis. Upon release of cytochrome c to the cytoplasm, the protein binds apoptotic protease activating factor which activates the apoptotic initiator pro-caspase 9. Many cytochrome c pseudogenes exist, and are scattered throughout the human genome. This gene is located on chromosome 7p15.2 and has several molecular functions such as an electron transporter, transferring electrons from CoQH2-cytochrome c, reductase complex and cytochrome c oxidase complex activity, heme binding, iron ion binding, metal ion binding, and protein binding. Microarray applications have been previously used in the analysis of blood-derived RNA to conduct genomic analyses of human diseases. This approach represents a rigorous and convenient alternative to traditional tissue biopsy-derived RNA, as it allows for larger sample sizes, better standardization of technical procedures, and the ability to non-invasively profile human subjects [55]. In humans, gene microarrays have successfully allowed for discovery of previously unsuspected genes or gene pathways in a veriety of disorders ranging from hypertension, cancer, neurodegenerative diseases, etc. [56], [57], [58], [59]. We should also point out experimental evidence suggesting that specific brain injury paradigms are associated with selective and distinct alterations in blood-derived gene expression patterns [60], and furthermore, gene expression profiles from both peripheral blood and nine different human tissues shared an excess of 80% of the differential gene expression profile [61], [62]. Thus, the use of peripheral blood gene expression patterns is likely to serve as a useful surrogate for gene expression in the other much less accessible tissues, and provides insights into pathophysiolgcal mechanisms underlying end-organ morbidities associated with diseases such as pediatric OSA. It has become apparent that not only the etiology but also the phenotypic expression of OSA are multi-factorial. A precise genetic foundation of OSA has been difficult to identify thus far, because it is still unknown whether some of the putative candidate genes for OSA are directly causal to the expression of the disorder or whether their role in OSA is mediated through other intermediate genes. Similarly, the phenotypic expression of OSA and its consequences are most likely determined by multiple genetic and environmental factors and their interactions. In summary, we have shown that pediatric OSA is predictably associated with altered gene expression patterns in circulating leukocytes. Furthermore, one of the important functional categories affected by the presence of OSA involves complex recruitment and interplay of genes involved in the inflammatory response in children. Considering the prominent evidence accumulated linking pediatric OSA to systemic inflammation [51], [52], [53], [63], [64], and the robust link between the magnitude of the systemic inflammatory response and the occurrence of end-organ morbidity [50], the present study provides a working road map to the discovery of novel and important pathways underlying the clinical manifestations of OSA in children. Appendix A. Supplementary data References [1]. [1]Kaparianos A, Sampsonas F, Karkoulias K, Spiropoulos K. Obstructive sleep apnoea syndrome and genes. Neth J Med. 2006;64:280–289. MEDLINE [2]. [2]Ali NJ, Pitson DJ, Stradling JR. Snoring, sleep disturbance, and behaviour in 4–5year olds. Arch Dis Child. 1993;68:360–366.
CrossRef
[3]. [3]Amin RS, Kimball TR, Bean JA, Jeffries JL, Willging JP, Cotton RT, et al. Left ventricular hypertrophy and abnormal ventricular geometry in children and adolescents with obstructive sleep apnea. Am J Respir Crit Care Med. 2002;165:1395–1399.
CrossRef
[4]. [4]Kheirandish L, Gozal D. Neurocognitive dysfunction in children with sleep disorders. Dev Sci. 2006;4:388–399. [5]. [5]de la Eva RC, Baur LA, Donaghue KC, Waters KA. Metabolic correlates with obstructive sleep apnea in obese subjects. J Pediatr. 2002;140(6):654–659. Abstract | Full Text |
Full-Text PDF (84 KB)
|
CrossRef
[6]. [6]Gozal D. Congenital central hypoventilation syndrome: an update. Pediatr Pulmonol. 1998;4:273–282. [7]. [7]Gozal D, Pope DW. Snoring during early childhood and academic performance at ages thirteen to fourteen years. Pediatrics. 2001;107:1394–1399. [8]. [8]Kaplan BJ, McNicol J, Conte RA, Moghadam HK. Sleep disturbance in preschool-aged hyperactive and nonhyperactive children. Pediatrics. 1987;80:839–844. [9]. [9]Chervin RD, Archbold KH, Dillon JE, Panahi P, Pituch KJ, Dahl RE, et al. Inattention, hyperactivity, and symptoms of sleep-disordered breathing. Pediatrics. 2002;109:449–456. [10]. [10]Corkum P, Tannock R, Moldofsky H, Hogg-Johnson S, Humphries T. Actigraphy and parental ratings of sleep in children with attention-deficit/hyperactivity disorder (ADHD). Sleep. 2001;24:303–312. MEDLINE [11]. [11]Gozal D, Kheirandish L. Oxidant stress and inflammation in the snoring child: confluent pathways to upper airway pathogenesis and end-organ morbidity. Sleep Med Rev. 2006;2:83–96. [12]. [12]Gozal D, Kheirandish L. Disorders of breathing during sleep. In: Chernick V, Kendig EL, Boat T, Wilmott R, Bush A editor. Kendig’s disorders of the respiratory tract in children. 7th ed.. Philadelphia: Saunders Elsevier; 2006;p. 1046–1070. [13]. [13]Darrow DH, Siemens C. Indications for tonsillectomy and adenoidectomy. Laryngoscope. 2002;112(8 Pt 2 Suppl. 100):6–10. [14]. [14]Lipton AJ, Gozal D. Treatment of obstructive sleep apnea in children: do we really know how?. Sleep Med Rev. 2003;1:61–80. [15]. [15]Lockhart DJ, Winzeler EA. Genomics, gene expression and DNA arrays. Nature. 2000;405:827–836. MEDLINE |
CrossRef
[16]. [16]Schena M, Shalon D, Davis RW, Brown PO. Quantitative monitoring of gene expression patterns with a complementary DNA microarray. Science. 1995;270(5235):467–470. MEDLINE [17]. [17]Lipshutz RJ, Fodor SP, Gingeras TR, Lockhart DJ. High density synthetic oligonucleotide arrays. Nat Genet. 1999;21(1 Suppl.):20–24. MEDLINE |
CrossRef
[18]. [18]Xiao J, Lucas A, D’Andrade P, Visitacion M, Tangvoranuntakul P, Fulmer-Smentek S. Performance of the Agilent microarray platform for one-color analysis of gene expression. Agilent Technical Note; 2006. [19]. [19]Hardiman G. Microarray platforms – comparisons and contrasts. Pharmacogenomics. 2004;5:487–502. MEDLINE |
CrossRef
[20]. [20]Stoughton RB. Applications of DNA microarrays in biology. Annu Rev Biochem. 2005;74:53–82. MEDLINE |
CrossRef
[21]. [21]Izuhara K, Saito H. Microarray-based identification of novel biomarkers in asthma. Allergol Int. 2006;55:361–367. MEDLINE |
CrossRef
[22]. [22]Nanni L, Romualdi C, Maseri A, Lanfranchi G. Differential gene expression profiling in genetic and multifactorial cardiovascular diseases. J Mol Cell Cardiol. 2006;41:934–948. Abstract | Full Text |
Full-Text PDF (235 KB)
|
CrossRef
[23]. [23]Cirelli C, Faraguna U, Tononi G. Changes in brain gene expression after long-term sleep deprivation. J Neurochem. 2006;98:1632–1645. MEDLINE |
CrossRef
[24]. [24]Irwin MR, Wang M, Campomayor CO, Collado-Hidalgo A, Cole S. Sleep deprivation and activation of morning levels of cellular and genomic markers of inflammation. Arch Intern Med. 2006;166:1756–1762. MEDLINE |
CrossRef
[25]. [25]Hoffmann MS, Singh P, Wolk R, Romero-Corral A, Raghavakaimal S, Somers VK. Microarray studies of genomic oxidative stress and cell cycle responses in obstructive sleep apnea. Antioxid Redox Signal. 2007;6:661–669. [26]. [26]Royce TE, Rozowsky JS, Luscombe NM, Emanuelsson O, Yu H, Zhu X, et al. Extrapolating traditional DNA microarray statistics to tiling and protein microarray technologies. Methods Enzymol. 2006;411:282–311. MEDLINE |
CrossRef
[27]. [27]Montgomery-Downs HE, O’Brien LM, Holbrook CR, Gozal D. Snoring and sleep-disordered breathing in young children: subjective and objective correlates. Sleep. 2004;27:87–94. MEDLINE [28]. [28]Montgomery-Downs HE, O’Brien LM, Gulliver TE, Gozal D. Polysomnographic characteristics in normal preschool and early school-aged children. Pediatrics. 2006;117:741–753. [29]. [29]Rechtschaffen A, Kales A. A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects. Allan Rechtschaffen, Anthony Kales, editors. 1st ed. Washington DC: Government Printing Office, National Institute of Health; 1968. [30]. [30]American Thoracic Society . Standards and indications for cardiopulmonary sleep studies in children. Am J Respir Crit Care Med. 1996;153:866–878. [31]. [31]American Sleep Disorders Association . EEG arousals: scoring rules and examples: a preliminary report from the Sleep Disorders Atlas Task Force of the American Sleep Disorders Association. Sleep. 1992;2:173–184. [32]. [32]Mograss MA, Ducharme FM, Brouillette RT. Movement/arousals. Description, classification, and relationship to sleep apnea in children. Am J Respir Crit Care Med. 1994;150(6 Pt 1):1690–1696. [33]. [33]Kuczmarski RJ, Ogden CL, Grummer-Strawn LM, Flegal KM, Guo SS, Wei R, et al. CDC growth charts: United States. Adv Data. 2000;314:1–27. [34]. [34]Khan-Farooqi HR, Prins RM, Liau LM. Tumor immunology, immunomics and targeted immunotherapy for central nervous system malignancies. Neurol Res. 2005;27:692–702. MEDLINE |
CrossRef
[35]. [35]Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc B Methodol. 1995;57:289–300. [36]. [36]Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(−Delta Delta C(T)) method. Methods. 2001;25:402–408. MEDLINE |
CrossRef
[37]. [37]Marcus CL, Gozal D, Arens R, Basinski DJ, Omlin KJ, Keens TG, et al. Ventilatory responses during wakefulness in children with obstructive sleep apnea. Am J Respir Crit Care Med. 1994;149(3 Pt 1):715–721. [38]. [38]Zhang L, Liu W, Xiao J, Hu T, Chen J, Chen K, et al. Malonyl-CoA: acyl carrier protein transacylase from Helicobacter pylori: crystal structure and its interaction with acyl carrier protein. Protein Sci. 2007;16:1184–1192. MEDLINE |
CrossRef
[39]. [39]Bandyopadhyay GK, Yu JG, Ofrecio J, Olefsky JM. Increased malonyl-CoA levels in muscle from obese and type 2 diabetic subjects lead to decreased fatty acid oxidation and increased lipogenesis; thiazolidinedione treatment reverses these defects. Diabetes. 2006;55:2277–2285. MEDLINE |
CrossRef
[40]. [40]Noland RC, Woodlief TL, Whitfield BR, Manning SM, Evans JR, Dudek RW, et al. Peroxisomal-mitochondrial oxidation in a rodent model of obesity-associated insulin resistance. Am J Physiol Endocrinol Metab. 2007;. [41]. [41]Tauman R, O’Brien LM, Ivanenko A, Gozal D. Obesity rather than severity of sleep-disordered breathing as the major determinant of insulin resistance and altered lipidemia in snoring children. Pediatrics. 2005;116:e66–e73. [42]. [42]Kaditis AG, Alexopoulos EI, Damani E, Karadonta I, Kostadima E, Tsolakidou A, et al. Obstructive sleep-disordered breathing and fasting insulin levels in nonobese children. Pediatr Pulmonol. 2005;40:515–523. MEDLINE |
CrossRef
[43]. [43]Waters KA, Sitha S, O’Brien LM, Bibby S, de Torres C, Vella S, et al. Follow-up on metabolic markers in children treated for obstructive sleep apnea. Am J Respir Crit Care Med. 2006;174:455–460.
CrossRef
[44]. [44]Redline S, Storfer-Isser A, Rosen CL, Johnson NL, Kirchner HL, Emancipator J, et al. Association between metabolic syndrome and sleep-disordered breathing in adolescents. Am J Respir Crit Care Med. 2007;176:401–408.
CrossRef
[45]. [45]Li J, Grigoryev DN, Ye SQ, Thorne L, Schwartz AR, Smith PL, et al. Chronic intermittent hypoxia upregulates genes of lipid biosynthesis in obese mice. J Appl Physiol. 2005;99:1643–1648. [46]. [46]Li J, Thorne LN, Punjabi NM, Sun CK, Schwartz AR, Smith PL, et al. Intermittent hypoxia induces hyperlipidemia in lean mice. Circ Res. 2005;97:698–706.
CrossRef
[47]. [47]Spiegel K, Knutson K, Leproult R, Tasali E, Van Cauter E. Sleep loss: a novel risk factor for insulin resistance and type 2 diabetes. J Appl Physiol. 2005;99:2008–2019. [48]. [48]Cal S, Quesada V, Llamazares M, Diaz-Perales A, Garabaya C, Lopez-Otin C. Human polyserase-2, a novel enzyme with three tandem serine protease domains in a single polypeptide chain. J Biol Chem. 2005;280:1953–1961. MEDLINE |
CrossRef
[49]. [49]Stumhofer JS, Laurence A, Wilson EH, Huang E, Tato CM, Johnson LM, et al. Interleukin 27 negatively regulates the development of interleukin 17-producing T helper cells during chronic inflammation of the central nervous system. Nat Immunol. 2006;9:937–945.
CrossRef
[50]. [50]Gozal D, Crabtree VM, Sans Capdevila O, Witcher LA, Kheirandish-Gozal L. C-reactive protein, obstructive sleep apnea, and cognitive dysfunction in school-aged children. Am J Respir Crit Care Med. 2007;176:188–193.
CrossRef
[51]. [51]Gozal D, Serpero LD, Sans Capdevila O, Kheirandish-Gozal L. Systemic inflammation in non-obese children with obstructive sleep apnea. Sleep Med. 2008;9:254–259. Abstract | Full Text |
Full-Text PDF (121 KB)
|
CrossRef
[52]. [52]Kheirandish-Gozal L, Capdevila OS, Tauman R, Gozal D. Plasma C-reactive protein in nonobese children with obstructive sleep apnea before and after adenotonsillectomy. J Clin Sleep Med. 2006;3:301–304. [53]. [53]Tauman R, Ivanenko A, O’Brien LM, Gozal D. Plasma C-reactive protein levels among children with sleep-disordered breathing. Pediatrics. 2004;113:e564–e569. [54]. [54]Pawson T, Scott JD. Signaling through scaffold, anchoring, and adaptor proteins. Science. 1997;278:2075–2080. MEDLINE |
CrossRef
[55]. [55]Tsuang MT, Nossova N, Yager T, Tsuang MM, Guo SC, Shyu KG, et al. Assessing the validity of blood-based gene expression profiles for the classification of schizophrenia and bipolar disorder: a preliminary report. Am J Med Genet B Neuropsychiatr Genet. 2005;133:1–5. MEDLINE [56]. [56]Borovecki F, Lovrecic L, Zhou J, Jeong H, Then F, Rosas HD, et al. Genome-wide expression profiling of human blood reveals biomarkers for Huntington’s disease. Proc Natl Acad Sci USA. 2005;102:11023–11028. MEDLINE |
CrossRef
[57]. [57]Lacayo NJ, Meshinchi S, Kinnunen P, Yu R, Wang Y, Stuber CM, et al. Gene expression profiles at diagnosis in de novo childhood AML patients identify FLT3 mutations with good clinical outcomes. Blood. 2004;1049:2646–2654. [58]. [58]Li Y, Elashoff D, Oh M, Sinha U, St John MA, Zhou X, et al. Serum circulating human mRNA profiling and its utility for oral cancer detection. J Clin Oncol. 2006;24:1754–1760.
CrossRef
[59]. [59]Timofeeva AV, Goryunova LE, Khaspekov GL, Kovalevskii DA, Scamrov AV, Bulkina OS, et al. Altered gene expression pattern in peripheral blood leukocytes from patients with arterial hypertension. Ann N Y Acad Sci. 2006;1091:319–335. MEDLINE |
CrossRef
[60]. [60]Sharp FR, Xu H, Lit L, Walker W, Apperson M, Gilbert DL, et al. The future of genomic profiling of neurological diseases using blood. Arch Neurol. 2006;63:1529–1536. MEDLINE |
CrossRef
[61]. [61]Liew CC, Ma J, Tang HC, Zheng R, Dempsey AA. The peripheral blood transcriptome dynamically reflects system wide biology: a potential diagnostic tool. J Lab Clin Med. 2006;147:126–132. Abstract | Full Text |
Full-Text PDF (148 KB)
|
CrossRef
[62]. [62]Sullivan PF, Fan C, Perou CM. Evaluating the comparability of gene expression in blood and brain. Am J Med Genet B Neuropsychiatr Genet. 2006;141:261–268. MEDLINE [63]. [63]Larkin EK, Rosen CL, Kirchner HL, Storfer-Isser A, Emancipator JL, Johnson NL, et al. Variation of C-reactive protein levels in adolescents: association with sleep-disordered breathing and sleep duration. Circulation. 2005;111:1978–1984.
CrossRef
[64]. [64]O’Brien LM, Serpero LD, Tauman R, Gozal D. Plasma adhesion molecules in children with sleep-disordered breathing. Chest. 2006;129:947–953. MEDLINE |
CrossRef
a Kosair Children’s Hospital Research Institute, Department of Pediatrics, University of Louisville, 570 South Preston Street, Suite 204, Louisville, KY 40202, USA b Division of Pediatric Sleep Medicine, Department of Pediatrics, University of Louisville, 571 S. Floyd Street, Suite 439, Louisville, KY 40202, USA  Corresponding author. Address: Kosair Children’s Hospital Research Institute, Department of Pediatrics, University of Louisville, 570 South Preston Street, Suite 204, Louisville, KY 40202, USA. Tel.: +1 502 852 2323; fax: +1 502 852 2215.
☆ Funding sources: A.K. is supported by University of Louisville Institutional Research Grant E0606. D.G. is supported by NIH Grants HL-65270 and HL-83075, The Children’s Foundation Endowment for Sleep Research, and by the Commonwealth of Kentucky Challenge for Excellence Trust Fund. L.K.G. is supported by a grant from the National Space Agency (NNJ05HF 06G). PII: S1389-9457(07)00375-9 doi:10.1016/j.sleep.2007.11.006 © 2007 Elsevier B.V. All rights reserved. | |
|
FULL TEXT00375-9/journalimage?img=PIIS1389945707003759.gr1.lrg.gif&fig=fig1&kwhquery=&issn=1389-9457&ishighres=false&allhighres=false&free=yes','journalimage');)
00375-9/journalimage?img=PIIS1389945707003759.gr2.lrg.gif&fig=fig2&kwhquery=&issn=1389-9457&ishighres=false&allhighres=false&free=yes','journalimage');)
