- Email: [email protected]
Whole blood hypoxia-related gene expression reveals novel pathways to obstructive sleep apnea in humans
Respiratory Physiology & Neurobiology 189 (2013) 649–654
Contents lists available at ScienceDirect
Respiratory Physiology & Neurobiology journal homepage: www.elsevier.com/locate/resphysiol
Short communication
Whole blood hypoxia-related gene expression reveals novel pathways to obstructive sleep apnea in humans Juliana C. Perry ∗ , Camila Guindalini, Lia Bittencourt, Silverio Garbuio, Diego R. Mazzotti, Sergio Tufik Departamento de Psicobiologia, Universidade Federal de São Paulo, Brazil
a r t i c l e
i n f o
Article history: Accepted 17 August 2013 Keywords: Hypoxia Obstructive sleep apnea Gene expression CPAP Sleep
a b s t r a c t In this study, our goal was to identify the key genes that are associated with obstructive sleep apnea (OSA). Thirty-five volunteers underwent full in-lab polysomnography and, according to the sleep apnea hypopnea index (AHI), were classified into control, mild-to-moderate OSA and severe OSA groups. Severe OSA patients were assigned to participate in a continuous positive airway pressure (CPAP) protocol for 6 months. Blood was collected and the expression of 84 genes analyzed using the RT2 ProfilerTM PCR array. Mild-to-moderate OSA patients demonstrated down-regulation of 2 genes associated with induction of apoptosis, while a total of 13 genes were identified in severe OSA patients. After controlling for body mass index, PRPF40A and PLOD3 gene expressions were strongly and independently associated with AHI scores. This research protocol highlights a number of molecular targets that might help the development of novel therapeutic strategies. © 2013 Elsevier B.V. All rights reserved.
1. Introduction
2. Materials and methods
Obstructive sleep apnea (OSA) is characterized by repetitive obstruction of the upper airway, resulting in pauses in breathing and subsequent oxygen desaturation. The majority of sleep disorders, including OSA, are caused by complex interactions between genes and the environment, leading studies to focus on the extent to which genes predetermine susceptibility to intermittent apneas, as well as on the effects of hypoxia and sleep fragmentation on the expression of candidate genes (Arnardottir et al., 2009). Previous studies have identified a key area of genetic influence as the cellular reaction to hypoxia (Arnardottir et al., 2009). Knowledge of the genetic factors underlying the phenotypic traits following hypoxia insults can reveal how cells confer vulnerability or resistance to hypoxia and modulate responsiveness to therapeutic interventions aimed at restoring respiratory, cardiovascular, cognitive, and metabolic functions (Cassavaugh and Lounsbury, 2011). Here, our goal was to identify the genes that are associated with hypoxia that play a role in OSA. This approach resulted in the identification of novel candidate genes of OSA, pointing the way to how genetic factors may modulate apneic susceptibility, as well as help the potential development of novel therapeutic targets.
2.1. Study population and polysomnography A total of 35 volunteers underwent full in-lab polysomnography (PSG; EMBLA System N7000, software RemLogic Version 2.0, CO, USA) during their habitual sleep time. Standard montage and criteria were used for scoring sleep stages, arousals, leg movements, and respiratory events according to the guidelines from the American Academy of Sleep Medicine, including the recommend rule for hypopneas – a decrease of ≥30% in the nasal pressure signal of baseline, with a duration ≥10 s, with a desaturation ≥4% associated. A total of 11 individuals were selected as controls (AHI < 5), 10 patients were diagnosed with mild-moderate (AHI 5 to <30), and 14 patients with severe OSA (AHI > 30 and clinical complaints). The research protocol was approved by the institution’s ethics committee (CEP no. 985/08), and informed consent was obtained from each patient. The total number of patients recruited, inclusion and exclusion criteria, primary and secondary endpoints of the clinical trial were registered in ClinicalTrials.gov (identifier NCT01392339). Data collection, processing and analysis were completed in 10 months. 2.2. CPAP treatment
∗ Corresponding author at: Departamento de Psicobiologia, Universidade Federal de São Paulo, Brazil Rua Napoleão de Barros, 925 Vila Clementino, 04024-002 São Paulo, SP, Brazil. Tel.: +55 11 2149 0155; fax: +55 11 5572 5092. E-mail address: [email protected] (J.C. Perry). 1569-9048/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.resp.2013.08.012
Patients with severe OSA underwent a continuous positive airway pressure (CPAP) protocol for 6 months to examine the influence of treatment on gene expression. All patients underwent a pressure titration PSG to find the ideal pressure to eliminate obstructive apnea, hypopnea, respiratory effort related arousal and
650
J.C. Perry et al. / Respiratory Physiology & Neurobiology 189 (2013) 649–654
snoring. After that they received a CPAP device (REMstar® Plus; Respironics Inc., Murrysville, PA) that allowed for pressure variance between 4 and 20 cmH2 O. A 20-min latency period was set before gradually increasing the pressure to its ideal value. The devices had a time-of-use function that measured the amount of time participants had effective delivery of treatment pressure (mask-on pressure). Patients also completed a sleep diary that consisted of questions about their use of CPAP and their amount of sleep per night. From that a mean percentage of CPAP use was calculated as a ratio of total sleep time over the number of hours using the device, as reported in the diary. A trained nurse contacted patients after 1 week, and 1, 3 and 6 months to check the adherence of treatment. 2.3. PCR array and data analysis Blood samples were collected at 8:00 AM. Blood was collected in PAXgene® tubes (PreAnalytiX GmbH, Hombrechtikon,
Switzerland), and total RNA was isolated according to the manufacturer’s directions. To evaluate gene expression profile, the RT2 ProfilerTM hypoxia signaling pathway PCR array, consisting of 84 genes known to be involved in the hypoxic response (Catalog #PAHS-032, SABiosciences, Frederick, MD, USA) was used according to the standard protocol. Gene expression associated with a p-value of <0.001 or absolute fold change >2 were taken as differently expressed between groups. 2.4. Functional annotation and network analysis of profile genes Functional annotation and network analysis of the profile genes was performed using Ingenuity Pathway Analysis (IPA) software (Ingenuity Systems Inc., Redwood City, CA). Statistical significance for enrichment of functional groups was based on Fisher’s exact test and corrected for multiple testing.
Table 1 List of differently expressed genes identified in mild-moderate and severe OSA patients, as well as after 6 months of CPAP treatment. Group
Gene Name
Control × OSA 5 to <30 DAPK3 KAT5
Description
FC
Death-associated protein kinase 3 K(lysine) acetyltransferase 5
−1.2 −1.2
Angiopoietin-like 4 Death-associated protein kinase 3 E1A binding protein p300 K(lysine) acetyltransferase 5 KH-type splicing regulatory protein Molybdenum cofactor synthesis 3 Metallothionein 3 ARD1 homolog A, N-acetyltransferase Notch homolog 1, translocation-associated Procollagen-lysine, 2-oxoglutarate 5-dioxygenase 3 PRP40 pre-mRNA processing factor 40 homolog A Ribosomal protein L28 Small nuclear ribonucleoprotein 70 (U1)
−1.6 −1.4 −2.1 −1.3 −1.4 −1.2 −1.8 −1.2 −2.4 −1.4 −1.3 −1.4 −1.3
Angiopoietin-like 4 Aryl-hydrocarbon receptor nuclear translocator 2 BCL2-associated X protein Basic helix-loop-helix family, member e40 Catalase Cystatin B (stefin B) Death-associated protein kinase 3 Endothelin converting enzyme 1 Enolase 1, (alpha) Hypoxia inducible factor 1, alpha subunit Hypoxia inducible factor 1, alpha subunit inhibitor Heme oxygenase (decycling) 1 Interleukin 6 signal transducer (gp130, oncostatin M receptor) K(lysine) acetyltransferase 5 KH-type splicing regulatory protein Mannosidase, alpha, class 2B, member 1 Molybdenum cofactor synthesis 3 Nudix (nucleoside diphosphate linked moiety X)-type motif 2 Peroxisome proliferator-activated receptor alpha Protein phosphatase 2 (formerly 2A), catalytic subunit, beta isoform Protein kinase, AMP-activated, alpha 1 catalytic subunit Proteasome (prosome, macropain) subunit, beta type, 3 Pentraxin-related gene, rapidly induced by IL-1 beta Ribosomal protein L28 Ribosomal protein S2 SUMO1 activating enzyme subunit 1 Solute carrier family 2 (facilitated glucose transporter), member 1 Sjogren syndrome/scleroderma autoantigen 1 SMT3 suppressor of mif two 3 homolog 2 (S. cerevisiae) Tubulin, alpha 4a Uncoupling protein 2 (mitochondrial, proton carrier) Vascular endothelial growth factor A
−1.9 3.5 −2.1 −2.9 −2.5 −2.0 −2.2 −6.0 −2.2 −2.4 −2.3 −1.8 −1.7 −1.8 −2.2 −2.4 −1.8 −1.4 −1.9 −2.3 −1.4 −3.2 −2.0 −2.3 −1.9 3.6 −2.8 4.2 −2.1 −2.3 −2.5 −1.6
Control × OSA >30 ANGPTL4 DAPK3 EP300 KAT5 KHSRP M0CS3 MT3 NAA10 N0TCH1 PL0D3 PRPF40A RPL28 SNRNP70 OSA>30 × CPAP treatment ANGPTL4 ARNT2 BAX BHLHE40 CAT CSTB DAPK3 ECE1 EN01 HIF1A HIF1AN HM0X1 IL6ST KAT5 KHSRP MAN2B1 M0CS3 NUDT2 PPARA PPP2CB PRKAA1 PSMB3 PTX3 RPL28 RPS2 SAE1 SLC2A1 SSSCA1 SUM02 TUBA4A UCP2 VEGFA
FC
CPAP −1.9 −2.2 −1.1 −1.8 −2.2 −1.8 −1.0 −1.4 1.3 −1.6 −1.1 −2.3 1.5
* *
NS * * *
NS NS NS NS NS *
NS
Annotation of genes identified by Student’s t-test (p < 0.001) or absolute fold change in gene expression of >2 were taken as significant. FC, fold change; NS, not significant; OSA, obstructive sleep apnea; CPAP, continuous positive airway pressure. * <0.001
J.C. Perry et al. / Respiratory Physiology & Neurobiology 189 (2013) 649–654
651
Fig. 1. Molecular signature induced by severe OSA. The panel shows a network integrating multi-level measurements in mild-moderate OSA (orange lines) and severe OSA (gray lines). Data are standardized to represent above average values in red and below average values in green. AXIN2: axis inhibition protein 2; BBC3: BCL2 binding component 3; BTRC: beta-transducin repeat containing E3 ubiquitin protein ligase; CCND1: cyclin D1; CHGA: chromogranin A; CLOCK: clock circadian regulator; CTNNB1: catenin (cadherin-associated protein), beta 1; DDX5: DEAD (Asp-Glu-Ala-Asp) box helicase 5; ELAVL1: ELAV-like 1 (Hu antigen R); ENO2: enolase 2; EPHB2: Eph receptor B2; FBXW7: F-box and WD repeat domain containing 7; GATA3: GATA binding protein 3; HEY1: hairy/enhancer-of-split related with YRPW motif 1; ILF3: interleukin enhancer binding factor 3; ING3: inhibitor of growth family, member 3; ING4: inhibitor of growth family, member 4; ING5: inhibitor of growth family, member 5; JMY: junction mediating and regulatory protein, p53 cofactor; MAML1: mastermind-like 1; MDM2: MDM2 oncogene, E3 ubiquitin protein ligase; NR3C1: nuclear receptor subfamily 3, group C, member 1; PPARG: peroxisome proliferator-activated receptor gamma; PRNP: prion protein; TH: tyrosine hydroxylase; TNFRSF10B: tumor necrosis factor receptor superfamily, member 10b; TP53: tumor protein p53; UBE2C: ubiquitin-conjugating enzyme E2C; see Table 1 for expansion of the other abbreviation.
2.5. Statistical analysis Values shown are expressed in mean ± standard error of mean (SEM). Sleep pattern data were analyzed using one-way analysis of variance (ANOVA) followed by post hoc Tukey tests. The 2−Ct values for human and animal data were compared using independent t-tests. Covariance analyses were used to analyze human gene expression and to adjust for the effects of CPAP adherence. The impact of measured 2−Ct values and subject characteristics, controlling for body mass index (BMI), was evaluated using multiple linear regression analysis with AHI as the dependent outcome. p < 0.05 were considered statistically significant. Gene expression results were adjusted using the Bonferroni correction for multiple comparisons (p < 0.001). 3. Results 3.1. Sleep pattern – evaluation of apnea-hypopnea index The control, mild-to-moderate, and severe OSA groups were comparable in age (Control 46.4 ± 2.6; OSA 5 to <30 41.7 ± 2.9; OSA >30 46.2 ± 2.3 years), sleep efficiency, percentage spent in
Table 2 Results of AHI levels prediction by the expression of the genes differently regulated in 35 severe OSA patients, as calculated by multiple linear regression models. Gene name
Coefficient
Std. error
t
p-level
ANGPTL4 DAPK3 EP300 KAT5 KHSRP M0CS3 MT3 NAA10 N0TCH1 PLOD 3 PRPF40A RPL28 SNRNP70
0.0034 0.0742 0.1932 0.0436 0.0992 0.0140 0.0009 0.1414 0.0510 0.0230 0.0160 0.0240 0.0595
0.0012 0.0148 0.0572 0.0104 0.0257 0.0031 0.0003 0.0226 0.0655 0.0092 0.0047 0.0058 0.0187
2.72 5.01 3.37 4.17 3.86 4.4 2.52 6.24 0.77 2.48 3.41 4.09 3.18
<0.010 <0.001 <0.001 <0.001 <0.001 <0.001 <0.020 <0.001 0.440 <0.020 <0.001 <0.001 <0.004
The results described in the table above were controlled for body mass index. Statistically significant (p < 0.04) were observed in PRP40 pre-mRNA processing factor 40 homolog A (PRPF40A) and procollagen-lysine, 2-oxoglutarate 5-dioxygenase 3 (PLOD3). OSA, obstructive sleep apnea; AHI, apnea hypopnea index.
652
J.C. Perry et al. / Respiratory Physiology & Neurobiology 189 (2013) 649–654
Fig. 2. Molecular signature induced by CPAP treatment. Panel A (cell death, cell cycle, cellular growth and proliferation) and Panel B (embryonic, organ and organismal development) show networks integrating multi-level measurements in sleep apneic patients after 6 months CPAP treatment. Data are standardized to represent above average values in red and below average values in green. APEX1: APEX nuclease; AR: androgen receptor; BCL2L1: BCL2-like 1; CDKN1A: cyclin-dependent kinase inhibitor 1A; CDKN2A: cyclin-dependent kinase inhibitor 2A; CTNNB1: catenin (cadherin-associated protein), beta 1; CUL4A: cullin 4A; DDB2: damage-specific DNA binding protein 2; ELAVL1: ELAV (embryonic lethal, abnormal vision, Drosophila)-like 1 (Hu antigen R); EP300: E1A binding protein p300; ERBB2: v-erb-b2 erythroblastic leukemia viral oncogene homolog 2, neuro/glioblastoma derived oncogene homolog; ESR1: estrogen receptor 1; EZH2: enhancer of zeste homolog 2; HSPD1: heat shock 60 kDa
J.C. Perry et al. / Respiratory Physiology & Neurobiology 189 (2013) 649–654
NREM sleep stage N1, and REM sleep; however, the BMI was higher in the severe OSA group (Control 26.1 ± 0.7; OSA 5 to <30 26.6 ± 0.7; OSA >30 32.2 ± 1.4 Kg/m2 ; p < 0.05). The AHI was 2.4 events/hour in the control group, 13.2 in the patients with mildto-moderate OSA, and 56.6 events/hour in the patients with severe OSA (p < 0.001). OSA patients exhibited higher percentage of NREM sleep stage N2 (65.6% vs 49.1%; p < 0.001), arousal index (52.8 vs 17.2; p < 0.001), lower percentage of NREM sleep stage N3 (11.2% vs 25.2%; p < 0.001) and minimum oxygen saturation during sleep (74.8% vs 90.3%; p < 0.001) compared to the control group. Epworth Sleepiness Scale (ESS) score was only higher in the severe OSA group when compared to mild-to-moderate OSA group (12.7 vs 6.1; p < 0.02); however, the value 12.7 was considered mild somnolence in the ESS, which might have contributed to the lack of difference compared to the control group. Following CPAP treatment, patients with severe OSA showed normalized sleep pattern (N2 46.4%; N3 25.2%; p < 0.001), arousal index (16.2; p < 0.001), AHI (5.9; p < 0.001), ESS index (9.1; p < 0.001) and minimum oxygen saturation during sleep (89.5%; p < 0.001) compared to baseline condition. Two volunteers showed no adherence to treatment. 3.2. Gene regulation and molecular signatures of OSA identified by genetic profiling In the group of patients with mild-to-moderate OSA, expression of two genes was found to be down-regulated: death-associated protein kinase 3 (DAPK3) and K(lysine) acetyltransferase 5 (KAT5) genes compared to healthy individuals. We identified 13 genes that form the gene expression profile associated with severe OSA in humans and 32 genes associated with the effects of CPAP treatment (Table 1). The direct associations between gene expression probe-set showed that differentially expressed genes associated with severe OSA were related to cell death, cancer, cellular growth and proliferation, and include the down-regulation of ANGPTL4, EP300, KAT5, DAPK3, KHSRP, NAA10, NOTCH1 and SNRNP70 genes (Fig. 1 gray lines). This molecular signature also overlapped that found in mild-moderate apnea (Fig. 1 – orange lines), where nuclear receptor subfamily 3, group C, member 1 (NR3C1) related to KAT5 and DAPK3 genes were differently expressed, making a link between mild-moderate and severe apnea. Network analysis of profile genes after 6 months of CPAP treatment indicated two main functional networks (Fig. 2). The first network (27 genes) was enriched for genes associated with cell death, cell cycle and cellular growth and proliferation, and included ANGPTL4, BAX, BHLHE40, CAT, DAPK3, ENO1, HMOX1, IL6ST, KAT5, KHSRP, PPP2CB, PTX3, SLC2A1 and SUMO2. The second network (7 genes) included genes important for embryonic, organ, and organismal development (HIF1A, HIF1AN, PPARA, TUBA4A and VEGFA). 3.3. Analysis of molecular profiling data to understand targets of disease Multiple linear regression modeling identified independent and strong associations between the AHI score and the gene expression level of PRP40 pre-mRNA processing factor 40 homolog A (PRPF40A) and procollagen-lysine, 2-oxoglutarate 5-dioxygenase 3 (PLOD3), after controlling for BMI (p < 0.04; Table 2). After CPAP treatment,
653
no significant difference was observed in PRPF40A or PLOD3 gene expression using adherence to CPAP as covariate in the model. No significant differences were obtained between the sleep fragmentation and time of desaturation >90% in the genes expression including BMI (data not shown).
4. Discussion Here we show direct correlation of physiological and molecular markers for OSA. The comparison between mild-to-moderate and severe OSA patients to healthy individuals showed downregulation of DAPK3 and KAT5 genes, two molecules associated with induction of apoptosis. Taken together, we propose that the regulation of these key genes by OSA may, in part, underlie preconditioning. Future studies should examine this possibility. Moreover, our data identified independent and strong associations between AHI levels and the expression of two genes: PLOD3 and PRPF40A, related to collagen biosynthesis and splicing efficiency enhancement, respectively. Interestingly, the association between these genes and OSA severity was no longer significant after CPAP treatment for 6 months. Prospective studies noted that compared with healthy controls matched for age, sex, and weight, the increased risk of death associated with severe sleep-disordered breathing was statistically significant (Marin et al., 2005). Moreover, it is well known that hypoxia results from a multitude of causes that lead to decreased oxygen delivery or availability (Cassavaugh and Lounsbury, 2011). Thus, the additional hypothesis to be tested in this study was whether the process of hypoxia preconditioning is partially mediated by genetic remodeling. Our results show that patients with mild-moderate OSA showed a down-regulation of only two genes expression: DAPK3 and KAT5. The DAPK3 is a pro-apoptotic gene that induces morphological changes when there is increased expression in mammalian cells (Kawai et al., 1998) and the KAT5 is associated with various processes such as cell signaling, DNA repair and apoptosis (Carrozza et al., 2003). Surprisingly, these two genes also showed down-regulation in patients with severe OSA. Under these conditions, it is reasonable to question whether the change in expression of these genes is a consequence of state or is part of the regulatory mechanisms, occurring as a result of OSA. In this sense, our data suggest that the processes of preconditioning in response to hypoxia could initiate during mild-moderate apnea. These findings raise the possibility that a molecular mechanism triggered by tissue hypoxia may be the basis of preconditioning in OSA patients. Accordingly, a higher mortality risk is associated with OSA in young patients and sharply declines with increasing age (Marin et al., 2005). According to the preconditioning hypothesis, the lower risk of mortality related to increasing age in patients with OSA may be explained by cardiovascular and cerebrovascular protection conferred by ischemic preconditioning resulting from the nightly cycle of hypoxia-reoxygenation (Lavie and Lavie, 2006). However, the exact mechanisms involved in triggering such alterations remain to be clarified. In addition, such a finding would imply that preconditioning may significantly alter the natural course of OSA. Molecular techniques are increasingly applied to determine the contribution of genes to sleep and its disorders (Arnardottir
protein 1; ID1: inhibitor of DNA binding 1, dominant negative helix-loop-helix protein; IL6: interleukin 6; ILF3: interleukin enhancer binding factor 3; ING3: inhibitor of growth family, member 3; INHBB: inhibin, beta B; ITGA1: integrin, alpha 1; MMP9: matrix metallopeptidase 9; MOAP1: modulator of apoptosis 1; MUC1: mucin 1, cell surface associated; MYC: v-myc myelocytomatosis viral oncogene homolog; NAA10: N(alpha)-acetyltransferase 10, NatA catalytic subunit; NCOA2: nuclear receptor coactivator 2; NR3C1: nuclear receptor subfamily 3, group C, member 1; PMAIP1: phorbol-12-myristate-13-acetate-induced protein 1; PPARG: peroxisome proliferator-activated receptor gamma; PRKCE: protein kinase C, epsilon; SMAD3: SMAD family member 3; SMAD4: SMAD family member 4; STAT3: signal transducer and activator of transcription 3; STRN3: striatin, calmodulin binding protein 3; TCF4: transcription factor 4; TFAP2A: transcription factor AP-2 alpha; TIA1: TIA1 cytotoxic granule-associated RNA binding protein; TIAL1: TIA1 cytotoxic granule-associated RNA binding protein-like 1; TJP1: tight junction protein 1; TP53: tumor protein p53; YY1: YY1 transcription factor; see Table 1 for expansion of the other abbreviation.
654
J.C. Perry et al. / Respiratory Physiology & Neurobiology 189 (2013) 649–654
et al., 2009). The diagnosis of OSA is based on the combination of characteristic clinical features in addition to compatible findings on instrumental tests in which multiple physiologic signals are monitored simultaneously during a night of sleep. The high prevalence of OSA (Tufik et al., 2010), its clinical importance, and difficulty of access to the high-cost diagnostic PSG has motivated the investigation of alternative diagnostic methods; the use of biochemical or genetic tests for diagnosis, low cost compared to PSG, has increased (Arnardottir et al., 2009). Our results showed independent and strong associations between the apnea-hypopnea index and PRPF40A and PLOD3 genes expression after adjusting for confounders, such as BMI. The lack of correlation between the expression of these two genes and BMI in the severe apnea patients suggests a specific regulation of these genes by the hypoxia insult. Moreover, after CPAP treatment, the expression of PRPF40A and PLOD3 genes was not significantly different from the pretreatment condition. PLOD3 is a multifunctional enzyme of collagen biosynthesis, which catalyzes the post-translational modifications of lysine in collagen and collagen-like proteins (Heikkinen et al., 2011). Recently, deficiency of PLOD3 causes growth arrest, revealing the importance of this molecule for cell growth and viability, and highlighting this gene as a potential target for medical applications (Wang et al., 2009). Moreover, FBP11/HYPA, a mammalian homolog of yeast splicing factor PRPF40A acts to enhance the splicing efficiency in mammalian cells (Lin et al., 2004). Nevertheless, additional studies are required to determine the exact mechanisms for the hypoxia – lowering expression of PRPF40A and PLOD3 genes. Although the importance of a molecular signature for OSA has been recognized (Arnardottir et al., 2009), only genetic analyses such as that described here can identify the precise molecular alterations that may be responsible for CIH dysregulation in both apneic patient and animal models. We found networks associated with cellular growth and proliferation (EP300 and NOTCH1), regulation of target genes that induce cell cycle arrest, apoptosis, senescence, DNA repair, or changes in metabolism (tumor protein p53 – TP53) and cell growth and adhesion between cells (cadherin-associated protein – CTNNB1 – the protein encoded by this gene is part of a complex of proteins that constitute adherent junctions; Fig. 1). Interestingly, there was a connection between severe and mildmoderate OSA. The molecular signature of mild-severe OSA was related to KAT5 and DAPK3 genes and NR3C1 gene, which encodes glucocorticoid receptor. The analysis of the effect of CPAP treatment showed activation of the HIF1A, VEGFA, CTNNB1 and TP3 genes. Moreover, two genes, TP3 and ESR1 (estrogen receptor 1) interconnected both identified CPAP networks (Fig. 2). Future studies should expand the molecular characterization of OSA begun in the present study. 5. Conclusions Here, we propose for the first time integration between the molecular and physiological responses following a hypoxia insult
within a network inference framework: a gene toolbox. In developing such a framework, we endeavor to define checkpoints spanning clinical application, in an attempt to help the design of novel strategies in the search for effective targets to treat OSA. Authors’ contribution (1) JCP, LB, SG, and DRM performed the conception and design of the study, or acquisition of data, or analysis and interpretation of data; (2) JCP, CG, LB, SG, DRM and ST involved in drafting the article or revising it critically for important intellectual content; and (3) JCP, CG, LB, SG, DRM, and ST involved final approval of the version to be submitted. Conflict of interest The authors declare that there is no conflict of interest. Acknowledgements This work was supported by grants from Associac¸ão Fundo de Incentivo à Pesquisa (AFIP), CNPq (#558924/2008-5 to JCP, CG, LRAB, ST) and FAPESP (#98/14303-3 to ST). We are grateful to Leiko K. Zanin for contacting patients and for verifying the adherence of CPAP treatment, Dr. Altay Alves Lino de Souza for statistical assistance, and all the efforts of Renata Pellegrino and Diva Maria Lima in helping the collection of data. References Arnardottir, E.S., Mackiewicz, M., Gislason, T., Teff, K.L., Pack, A.I., 2009. Molecular signatures of obstructive sleep apnea in adults: a review and perspective. Sleep 32, 447–470. Carrozza, M.J., Utley, R.T., Workman, J.L., Cote, J., 2003. The diverse functions of histone acetyltransferase complexes. Trends Genet. 19, 321–329. Cassavaugh, J., Lounsbury, K.M., 2011. Hypoxia-mediated biological control. J. Cell. Biochem. 112, 735–744. Heikkinen, J., Risteli, M., Lampela, O., Alavesa, P., Karppinen, M., Juffer, A.H., Myllyla, R., 2011. Dimerization of human lysyl hydroxylase 3 (LH3) is mediated by the amino acids 541–547. Matrix Biol. 30, 27–33. Kawai, T., Matsumoto, M., Takeda, K., Sanjo, H., Akira, S., 1998. ZIP kinase, a novel serine/threonine kinase which mediates apoptosis. Mol. Cell. Biol. 18, 1642–1651. Lavie, L., Lavie, P., 2006. Ischemic preconditioning as a possible explanation for the age decline relative mortality in sleep apnea. Med. Hypotheses 66, 1069–1073. Lin, K.T., Lu, R.M., Tarn, W.Y., 2004. The WW domain-containing proteins interact with the early spliceosome and participate in pre-mRNA splicing in vivo. Mol. Cell. Biol. 24, 9176–9185. Marin, J.M., Carrizo, S.J., Vicente, E., Agusti, A.G., 2005. Long-term cardiovascular outcomes in men with obstructive sleep apnoea-hypopnoea with or without treatment with continuous positive airway pressure: an observational study. Lancet 365, 1046–1053. Tufik, S., Santos-Silva, R., Taddei, J.A., Bittencourt, L.R., 2010. Obstructive sleep apnea syndrome in the Sao Paulo Epidemiologic Sleep Study. Sleep Med. 11, 441–446. Wang, C., Kovanen, V., Raudasoja, P., Eskelinen, S., Pospiech, H., Myllyla, R., 2009. The glycosyltransferase activities of lysyl hydroxylase 3 (LH3) in the extracellular space are important for cell growth and viability. J. Cell. Mol. Med. 13, 508–521.
Contents lists available at ScienceDirect
Respiratory Physiology & Neurobiology journal homepage: www.elsevier.com/locate/resphysiol
Short communication
Whole blood hypoxia-related gene expression reveals novel pathways to obstructive sleep apnea in humans Juliana C. Perry ∗ , Camila Guindalini, Lia Bittencourt, Silverio Garbuio, Diego R. Mazzotti, Sergio Tufik Departamento de Psicobiologia, Universidade Federal de São Paulo, Brazil
a r t i c l e
i n f o
Article history: Accepted 17 August 2013 Keywords: Hypoxia Obstructive sleep apnea Gene expression CPAP Sleep
a b s t r a c t In this study, our goal was to identify the key genes that are associated with obstructive sleep apnea (OSA). Thirty-five volunteers underwent full in-lab polysomnography and, according to the sleep apnea hypopnea index (AHI), were classified into control, mild-to-moderate OSA and severe OSA groups. Severe OSA patients were assigned to participate in a continuous positive airway pressure (CPAP) protocol for 6 months. Blood was collected and the expression of 84 genes analyzed using the RT2 ProfilerTM PCR array. Mild-to-moderate OSA patients demonstrated down-regulation of 2 genes associated with induction of apoptosis, while a total of 13 genes were identified in severe OSA patients. After controlling for body mass index, PRPF40A and PLOD3 gene expressions were strongly and independently associated with AHI scores. This research protocol highlights a number of molecular targets that might help the development of novel therapeutic strategies. © 2013 Elsevier B.V. All rights reserved.
1. Introduction
2. Materials and methods
Obstructive sleep apnea (OSA) is characterized by repetitive obstruction of the upper airway, resulting in pauses in breathing and subsequent oxygen desaturation. The majority of sleep disorders, including OSA, are caused by complex interactions between genes and the environment, leading studies to focus on the extent to which genes predetermine susceptibility to intermittent apneas, as well as on the effects of hypoxia and sleep fragmentation on the expression of candidate genes (Arnardottir et al., 2009). Previous studies have identified a key area of genetic influence as the cellular reaction to hypoxia (Arnardottir et al., 2009). Knowledge of the genetic factors underlying the phenotypic traits following hypoxia insults can reveal how cells confer vulnerability or resistance to hypoxia and modulate responsiveness to therapeutic interventions aimed at restoring respiratory, cardiovascular, cognitive, and metabolic functions (Cassavaugh and Lounsbury, 2011). Here, our goal was to identify the genes that are associated with hypoxia that play a role in OSA. This approach resulted in the identification of novel candidate genes of OSA, pointing the way to how genetic factors may modulate apneic susceptibility, as well as help the potential development of novel therapeutic targets.
2.1. Study population and polysomnography A total of 35 volunteers underwent full in-lab polysomnography (PSG; EMBLA System N7000, software RemLogic Version 2.0, CO, USA) during their habitual sleep time. Standard montage and criteria were used for scoring sleep stages, arousals, leg movements, and respiratory events according to the guidelines from the American Academy of Sleep Medicine, including the recommend rule for hypopneas – a decrease of ≥30% in the nasal pressure signal of baseline, with a duration ≥10 s, with a desaturation ≥4% associated. A total of 11 individuals were selected as controls (AHI < 5), 10 patients were diagnosed with mild-moderate (AHI 5 to <30), and 14 patients with severe OSA (AHI > 30 and clinical complaints). The research protocol was approved by the institution’s ethics committee (CEP no. 985/08), and informed consent was obtained from each patient. The total number of patients recruited, inclusion and exclusion criteria, primary and secondary endpoints of the clinical trial were registered in ClinicalTrials.gov (identifier NCT01392339). Data collection, processing and analysis were completed in 10 months. 2.2. CPAP treatment
∗ Corresponding author at: Departamento de Psicobiologia, Universidade Federal de São Paulo, Brazil Rua Napoleão de Barros, 925 Vila Clementino, 04024-002 São Paulo, SP, Brazil. Tel.: +55 11 2149 0155; fax: +55 11 5572 5092. E-mail address: [email protected] (J.C. Perry). 1569-9048/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.resp.2013.08.012
Patients with severe OSA underwent a continuous positive airway pressure (CPAP) protocol for 6 months to examine the influence of treatment on gene expression. All patients underwent a pressure titration PSG to find the ideal pressure to eliminate obstructive apnea, hypopnea, respiratory effort related arousal and
650
J.C. Perry et al. / Respiratory Physiology & Neurobiology 189 (2013) 649–654
snoring. After that they received a CPAP device (REMstar® Plus; Respironics Inc., Murrysville, PA) that allowed for pressure variance between 4 and 20 cmH2 O. A 20-min latency period was set before gradually increasing the pressure to its ideal value. The devices had a time-of-use function that measured the amount of time participants had effective delivery of treatment pressure (mask-on pressure). Patients also completed a sleep diary that consisted of questions about their use of CPAP and their amount of sleep per night. From that a mean percentage of CPAP use was calculated as a ratio of total sleep time over the number of hours using the device, as reported in the diary. A trained nurse contacted patients after 1 week, and 1, 3 and 6 months to check the adherence of treatment. 2.3. PCR array and data analysis Blood samples were collected at 8:00 AM. Blood was collected in PAXgene® tubes (PreAnalytiX GmbH, Hombrechtikon,
Switzerland), and total RNA was isolated according to the manufacturer’s directions. To evaluate gene expression profile, the RT2 ProfilerTM hypoxia signaling pathway PCR array, consisting of 84 genes known to be involved in the hypoxic response (Catalog #PAHS-032, SABiosciences, Frederick, MD, USA) was used according to the standard protocol. Gene expression associated with a p-value of <0.001 or absolute fold change >2 were taken as differently expressed between groups. 2.4. Functional annotation and network analysis of profile genes Functional annotation and network analysis of the profile genes was performed using Ingenuity Pathway Analysis (IPA) software (Ingenuity Systems Inc., Redwood City, CA). Statistical significance for enrichment of functional groups was based on Fisher’s exact test and corrected for multiple testing.
Table 1 List of differently expressed genes identified in mild-moderate and severe OSA patients, as well as after 6 months of CPAP treatment. Group
Gene Name
Control × OSA 5 to <30 DAPK3 KAT5
Description
FC
Death-associated protein kinase 3 K(lysine) acetyltransferase 5
−1.2 −1.2
Angiopoietin-like 4 Death-associated protein kinase 3 E1A binding protein p300 K(lysine) acetyltransferase 5 KH-type splicing regulatory protein Molybdenum cofactor synthesis 3 Metallothionein 3 ARD1 homolog A, N-acetyltransferase Notch homolog 1, translocation-associated Procollagen-lysine, 2-oxoglutarate 5-dioxygenase 3 PRP40 pre-mRNA processing factor 40 homolog A Ribosomal protein L28 Small nuclear ribonucleoprotein 70 (U1)
−1.6 −1.4 −2.1 −1.3 −1.4 −1.2 −1.8 −1.2 −2.4 −1.4 −1.3 −1.4 −1.3
Angiopoietin-like 4 Aryl-hydrocarbon receptor nuclear translocator 2 BCL2-associated X protein Basic helix-loop-helix family, member e40 Catalase Cystatin B (stefin B) Death-associated protein kinase 3 Endothelin converting enzyme 1 Enolase 1, (alpha) Hypoxia inducible factor 1, alpha subunit Hypoxia inducible factor 1, alpha subunit inhibitor Heme oxygenase (decycling) 1 Interleukin 6 signal transducer (gp130, oncostatin M receptor) K(lysine) acetyltransferase 5 KH-type splicing regulatory protein Mannosidase, alpha, class 2B, member 1 Molybdenum cofactor synthesis 3 Nudix (nucleoside diphosphate linked moiety X)-type motif 2 Peroxisome proliferator-activated receptor alpha Protein phosphatase 2 (formerly 2A), catalytic subunit, beta isoform Protein kinase, AMP-activated, alpha 1 catalytic subunit Proteasome (prosome, macropain) subunit, beta type, 3 Pentraxin-related gene, rapidly induced by IL-1 beta Ribosomal protein L28 Ribosomal protein S2 SUMO1 activating enzyme subunit 1 Solute carrier family 2 (facilitated glucose transporter), member 1 Sjogren syndrome/scleroderma autoantigen 1 SMT3 suppressor of mif two 3 homolog 2 (S. cerevisiae) Tubulin, alpha 4a Uncoupling protein 2 (mitochondrial, proton carrier) Vascular endothelial growth factor A
−1.9 3.5 −2.1 −2.9 −2.5 −2.0 −2.2 −6.0 −2.2 −2.4 −2.3 −1.8 −1.7 −1.8 −2.2 −2.4 −1.8 −1.4 −1.9 −2.3 −1.4 −3.2 −2.0 −2.3 −1.9 3.6 −2.8 4.2 −2.1 −2.3 −2.5 −1.6
Control × OSA >30 ANGPTL4 DAPK3 EP300 KAT5 KHSRP M0CS3 MT3 NAA10 N0TCH1 PL0D3 PRPF40A RPL28 SNRNP70 OSA>30 × CPAP treatment ANGPTL4 ARNT2 BAX BHLHE40 CAT CSTB DAPK3 ECE1 EN01 HIF1A HIF1AN HM0X1 IL6ST KAT5 KHSRP MAN2B1 M0CS3 NUDT2 PPARA PPP2CB PRKAA1 PSMB3 PTX3 RPL28 RPS2 SAE1 SLC2A1 SSSCA1 SUM02 TUBA4A UCP2 VEGFA
FC
CPAP −1.9 −2.2 −1.1 −1.8 −2.2 −1.8 −1.0 −1.4 1.3 −1.6 −1.1 −2.3 1.5
* *
NS * * *
NS NS NS NS NS *
NS
Annotation of genes identified by Student’s t-test (p < 0.001) or absolute fold change in gene expression of >2 were taken as significant. FC, fold change; NS, not significant; OSA, obstructive sleep apnea; CPAP, continuous positive airway pressure. * <0.001
J.C. Perry et al. / Respiratory Physiology & Neurobiology 189 (2013) 649–654
651
Fig. 1. Molecular signature induced by severe OSA. The panel shows a network integrating multi-level measurements in mild-moderate OSA (orange lines) and severe OSA (gray lines). Data are standardized to represent above average values in red and below average values in green. AXIN2: axis inhibition protein 2; BBC3: BCL2 binding component 3; BTRC: beta-transducin repeat containing E3 ubiquitin protein ligase; CCND1: cyclin D1; CHGA: chromogranin A; CLOCK: clock circadian regulator; CTNNB1: catenin (cadherin-associated protein), beta 1; DDX5: DEAD (Asp-Glu-Ala-Asp) box helicase 5; ELAVL1: ELAV-like 1 (Hu antigen R); ENO2: enolase 2; EPHB2: Eph receptor B2; FBXW7: F-box and WD repeat domain containing 7; GATA3: GATA binding protein 3; HEY1: hairy/enhancer-of-split related with YRPW motif 1; ILF3: interleukin enhancer binding factor 3; ING3: inhibitor of growth family, member 3; ING4: inhibitor of growth family, member 4; ING5: inhibitor of growth family, member 5; JMY: junction mediating and regulatory protein, p53 cofactor; MAML1: mastermind-like 1; MDM2: MDM2 oncogene, E3 ubiquitin protein ligase; NR3C1: nuclear receptor subfamily 3, group C, member 1; PPARG: peroxisome proliferator-activated receptor gamma; PRNP: prion protein; TH: tyrosine hydroxylase; TNFRSF10B: tumor necrosis factor receptor superfamily, member 10b; TP53: tumor protein p53; UBE2C: ubiquitin-conjugating enzyme E2C; see Table 1 for expansion of the other abbreviation.
2.5. Statistical analysis Values shown are expressed in mean ± standard error of mean (SEM). Sleep pattern data were analyzed using one-way analysis of variance (ANOVA) followed by post hoc Tukey tests. The 2−Ct values for human and animal data were compared using independent t-tests. Covariance analyses were used to analyze human gene expression and to adjust for the effects of CPAP adherence. The impact of measured 2−Ct values and subject characteristics, controlling for body mass index (BMI), was evaluated using multiple linear regression analysis with AHI as the dependent outcome. p < 0.05 were considered statistically significant. Gene expression results were adjusted using the Bonferroni correction for multiple comparisons (p < 0.001). 3. Results 3.1. Sleep pattern – evaluation of apnea-hypopnea index The control, mild-to-moderate, and severe OSA groups were comparable in age (Control 46.4 ± 2.6; OSA 5 to <30 41.7 ± 2.9; OSA >30 46.2 ± 2.3 years), sleep efficiency, percentage spent in
Table 2 Results of AHI levels prediction by the expression of the genes differently regulated in 35 severe OSA patients, as calculated by multiple linear regression models. Gene name
Coefficient
Std. error
t
p-level
ANGPTL4 DAPK3 EP300 KAT5 KHSRP M0CS3 MT3 NAA10 N0TCH1 PLOD 3 PRPF40A RPL28 SNRNP70
0.0034 0.0742 0.1932 0.0436 0.0992 0.0140 0.0009 0.1414 0.0510 0.0230 0.0160 0.0240 0.0595
0.0012 0.0148 0.0572 0.0104 0.0257 0.0031 0.0003 0.0226 0.0655 0.0092 0.0047 0.0058 0.0187
2.72 5.01 3.37 4.17 3.86 4.4 2.52 6.24 0.77 2.48 3.41 4.09 3.18
<0.010 <0.001 <0.001 <0.001 <0.001 <0.001 <0.020 <0.001 0.440 <0.020 <0.001 <0.001 <0.004
The results described in the table above were controlled for body mass index. Statistically significant (p < 0.04) were observed in PRP40 pre-mRNA processing factor 40 homolog A (PRPF40A) and procollagen-lysine, 2-oxoglutarate 5-dioxygenase 3 (PLOD3). OSA, obstructive sleep apnea; AHI, apnea hypopnea index.
652
J.C. Perry et al. / Respiratory Physiology & Neurobiology 189 (2013) 649–654
Fig. 2. Molecular signature induced by CPAP treatment. Panel A (cell death, cell cycle, cellular growth and proliferation) and Panel B (embryonic, organ and organismal development) show networks integrating multi-level measurements in sleep apneic patients after 6 months CPAP treatment. Data are standardized to represent above average values in red and below average values in green. APEX1: APEX nuclease; AR: androgen receptor; BCL2L1: BCL2-like 1; CDKN1A: cyclin-dependent kinase inhibitor 1A; CDKN2A: cyclin-dependent kinase inhibitor 2A; CTNNB1: catenin (cadherin-associated protein), beta 1; CUL4A: cullin 4A; DDB2: damage-specific DNA binding protein 2; ELAVL1: ELAV (embryonic lethal, abnormal vision, Drosophila)-like 1 (Hu antigen R); EP300: E1A binding protein p300; ERBB2: v-erb-b2 erythroblastic leukemia viral oncogene homolog 2, neuro/glioblastoma derived oncogene homolog; ESR1: estrogen receptor 1; EZH2: enhancer of zeste homolog 2; HSPD1: heat shock 60 kDa
J.C. Perry et al. / Respiratory Physiology & Neurobiology 189 (2013) 649–654
NREM sleep stage N1, and REM sleep; however, the BMI was higher in the severe OSA group (Control 26.1 ± 0.7; OSA 5 to <30 26.6 ± 0.7; OSA >30 32.2 ± 1.4 Kg/m2 ; p < 0.05). The AHI was 2.4 events/hour in the control group, 13.2 in the patients with mildto-moderate OSA, and 56.6 events/hour in the patients with severe OSA (p < 0.001). OSA patients exhibited higher percentage of NREM sleep stage N2 (65.6% vs 49.1%; p < 0.001), arousal index (52.8 vs 17.2; p < 0.001), lower percentage of NREM sleep stage N3 (11.2% vs 25.2%; p < 0.001) and minimum oxygen saturation during sleep (74.8% vs 90.3%; p < 0.001) compared to the control group. Epworth Sleepiness Scale (ESS) score was only higher in the severe OSA group when compared to mild-to-moderate OSA group (12.7 vs 6.1; p < 0.02); however, the value 12.7 was considered mild somnolence in the ESS, which might have contributed to the lack of difference compared to the control group. Following CPAP treatment, patients with severe OSA showed normalized sleep pattern (N2 46.4%; N3 25.2%; p < 0.001), arousal index (16.2; p < 0.001), AHI (5.9; p < 0.001), ESS index (9.1; p < 0.001) and minimum oxygen saturation during sleep (89.5%; p < 0.001) compared to baseline condition. Two volunteers showed no adherence to treatment. 3.2. Gene regulation and molecular signatures of OSA identified by genetic profiling In the group of patients with mild-to-moderate OSA, expression of two genes was found to be down-regulated: death-associated protein kinase 3 (DAPK3) and K(lysine) acetyltransferase 5 (KAT5) genes compared to healthy individuals. We identified 13 genes that form the gene expression profile associated with severe OSA in humans and 32 genes associated with the effects of CPAP treatment (Table 1). The direct associations between gene expression probe-set showed that differentially expressed genes associated with severe OSA were related to cell death, cancer, cellular growth and proliferation, and include the down-regulation of ANGPTL4, EP300, KAT5, DAPK3, KHSRP, NAA10, NOTCH1 and SNRNP70 genes (Fig. 1 gray lines). This molecular signature also overlapped that found in mild-moderate apnea (Fig. 1 – orange lines), where nuclear receptor subfamily 3, group C, member 1 (NR3C1) related to KAT5 and DAPK3 genes were differently expressed, making a link between mild-moderate and severe apnea. Network analysis of profile genes after 6 months of CPAP treatment indicated two main functional networks (Fig. 2). The first network (27 genes) was enriched for genes associated with cell death, cell cycle and cellular growth and proliferation, and included ANGPTL4, BAX, BHLHE40, CAT, DAPK3, ENO1, HMOX1, IL6ST, KAT5, KHSRP, PPP2CB, PTX3, SLC2A1 and SUMO2. The second network (7 genes) included genes important for embryonic, organ, and organismal development (HIF1A, HIF1AN, PPARA, TUBA4A and VEGFA). 3.3. Analysis of molecular profiling data to understand targets of disease Multiple linear regression modeling identified independent and strong associations between the AHI score and the gene expression level of PRP40 pre-mRNA processing factor 40 homolog A (PRPF40A) and procollagen-lysine, 2-oxoglutarate 5-dioxygenase 3 (PLOD3), after controlling for BMI (p < 0.04; Table 2). After CPAP treatment,
653
no significant difference was observed in PRPF40A or PLOD3 gene expression using adherence to CPAP as covariate in the model. No significant differences were obtained between the sleep fragmentation and time of desaturation >90% in the genes expression including BMI (data not shown).
4. Discussion Here we show direct correlation of physiological and molecular markers for OSA. The comparison between mild-to-moderate and severe OSA patients to healthy individuals showed downregulation of DAPK3 and KAT5 genes, two molecules associated with induction of apoptosis. Taken together, we propose that the regulation of these key genes by OSA may, in part, underlie preconditioning. Future studies should examine this possibility. Moreover, our data identified independent and strong associations between AHI levels and the expression of two genes: PLOD3 and PRPF40A, related to collagen biosynthesis and splicing efficiency enhancement, respectively. Interestingly, the association between these genes and OSA severity was no longer significant after CPAP treatment for 6 months. Prospective studies noted that compared with healthy controls matched for age, sex, and weight, the increased risk of death associated with severe sleep-disordered breathing was statistically significant (Marin et al., 2005). Moreover, it is well known that hypoxia results from a multitude of causes that lead to decreased oxygen delivery or availability (Cassavaugh and Lounsbury, 2011). Thus, the additional hypothesis to be tested in this study was whether the process of hypoxia preconditioning is partially mediated by genetic remodeling. Our results show that patients with mild-moderate OSA showed a down-regulation of only two genes expression: DAPK3 and KAT5. The DAPK3 is a pro-apoptotic gene that induces morphological changes when there is increased expression in mammalian cells (Kawai et al., 1998) and the KAT5 is associated with various processes such as cell signaling, DNA repair and apoptosis (Carrozza et al., 2003). Surprisingly, these two genes also showed down-regulation in patients with severe OSA. Under these conditions, it is reasonable to question whether the change in expression of these genes is a consequence of state or is part of the regulatory mechanisms, occurring as a result of OSA. In this sense, our data suggest that the processes of preconditioning in response to hypoxia could initiate during mild-moderate apnea. These findings raise the possibility that a molecular mechanism triggered by tissue hypoxia may be the basis of preconditioning in OSA patients. Accordingly, a higher mortality risk is associated with OSA in young patients and sharply declines with increasing age (Marin et al., 2005). According to the preconditioning hypothesis, the lower risk of mortality related to increasing age in patients with OSA may be explained by cardiovascular and cerebrovascular protection conferred by ischemic preconditioning resulting from the nightly cycle of hypoxia-reoxygenation (Lavie and Lavie, 2006). However, the exact mechanisms involved in triggering such alterations remain to be clarified. In addition, such a finding would imply that preconditioning may significantly alter the natural course of OSA. Molecular techniques are increasingly applied to determine the contribution of genes to sleep and its disorders (Arnardottir
protein 1; ID1: inhibitor of DNA binding 1, dominant negative helix-loop-helix protein; IL6: interleukin 6; ILF3: interleukin enhancer binding factor 3; ING3: inhibitor of growth family, member 3; INHBB: inhibin, beta B; ITGA1: integrin, alpha 1; MMP9: matrix metallopeptidase 9; MOAP1: modulator of apoptosis 1; MUC1: mucin 1, cell surface associated; MYC: v-myc myelocytomatosis viral oncogene homolog; NAA10: N(alpha)-acetyltransferase 10, NatA catalytic subunit; NCOA2: nuclear receptor coactivator 2; NR3C1: nuclear receptor subfamily 3, group C, member 1; PMAIP1: phorbol-12-myristate-13-acetate-induced protein 1; PPARG: peroxisome proliferator-activated receptor gamma; PRKCE: protein kinase C, epsilon; SMAD3: SMAD family member 3; SMAD4: SMAD family member 4; STAT3: signal transducer and activator of transcription 3; STRN3: striatin, calmodulin binding protein 3; TCF4: transcription factor 4; TFAP2A: transcription factor AP-2 alpha; TIA1: TIA1 cytotoxic granule-associated RNA binding protein; TIAL1: TIA1 cytotoxic granule-associated RNA binding protein-like 1; TJP1: tight junction protein 1; TP53: tumor protein p53; YY1: YY1 transcription factor; see Table 1 for expansion of the other abbreviation.
654
J.C. Perry et al. / Respiratory Physiology & Neurobiology 189 (2013) 649–654
et al., 2009). The diagnosis of OSA is based on the combination of characteristic clinical features in addition to compatible findings on instrumental tests in which multiple physiologic signals are monitored simultaneously during a night of sleep. The high prevalence of OSA (Tufik et al., 2010), its clinical importance, and difficulty of access to the high-cost diagnostic PSG has motivated the investigation of alternative diagnostic methods; the use of biochemical or genetic tests for diagnosis, low cost compared to PSG, has increased (Arnardottir et al., 2009). Our results showed independent and strong associations between the apnea-hypopnea index and PRPF40A and PLOD3 genes expression after adjusting for confounders, such as BMI. The lack of correlation between the expression of these two genes and BMI in the severe apnea patients suggests a specific regulation of these genes by the hypoxia insult. Moreover, after CPAP treatment, the expression of PRPF40A and PLOD3 genes was not significantly different from the pretreatment condition. PLOD3 is a multifunctional enzyme of collagen biosynthesis, which catalyzes the post-translational modifications of lysine in collagen and collagen-like proteins (Heikkinen et al., 2011). Recently, deficiency of PLOD3 causes growth arrest, revealing the importance of this molecule for cell growth and viability, and highlighting this gene as a potential target for medical applications (Wang et al., 2009). Moreover, FBP11/HYPA, a mammalian homolog of yeast splicing factor PRPF40A acts to enhance the splicing efficiency in mammalian cells (Lin et al., 2004). Nevertheless, additional studies are required to determine the exact mechanisms for the hypoxia – lowering expression of PRPF40A and PLOD3 genes. Although the importance of a molecular signature for OSA has been recognized (Arnardottir et al., 2009), only genetic analyses such as that described here can identify the precise molecular alterations that may be responsible for CIH dysregulation in both apneic patient and animal models. We found networks associated with cellular growth and proliferation (EP300 and NOTCH1), regulation of target genes that induce cell cycle arrest, apoptosis, senescence, DNA repair, or changes in metabolism (tumor protein p53 – TP53) and cell growth and adhesion between cells (cadherin-associated protein – CTNNB1 – the protein encoded by this gene is part of a complex of proteins that constitute adherent junctions; Fig. 1). Interestingly, there was a connection between severe and mildmoderate OSA. The molecular signature of mild-severe OSA was related to KAT5 and DAPK3 genes and NR3C1 gene, which encodes glucocorticoid receptor. The analysis of the effect of CPAP treatment showed activation of the HIF1A, VEGFA, CTNNB1 and TP3 genes. Moreover, two genes, TP3 and ESR1 (estrogen receptor 1) interconnected both identified CPAP networks (Fig. 2). Future studies should expand the molecular characterization of OSA begun in the present study. 5. Conclusions Here, we propose for the first time integration between the molecular and physiological responses following a hypoxia insult
within a network inference framework: a gene toolbox. In developing such a framework, we endeavor to define checkpoints spanning clinical application, in an attempt to help the design of novel strategies in the search for effective targets to treat OSA. Authors’ contribution (1) JCP, LB, SG, and DRM performed the conception and design of the study, or acquisition of data, or analysis and interpretation of data; (2) JCP, CG, LB, SG, DRM and ST involved in drafting the article or revising it critically for important intellectual content; and (3) JCP, CG, LB, SG, DRM, and ST involved final approval of the version to be submitted. Conflict of interest The authors declare that there is no conflict of interest. Acknowledgements This work was supported by grants from Associac¸ão Fundo de Incentivo à Pesquisa (AFIP), CNPq (#558924/2008-5 to JCP, CG, LRAB, ST) and FAPESP (#98/14303-3 to ST). We are grateful to Leiko K. Zanin for contacting patients and for verifying the adherence of CPAP treatment, Dr. Altay Alves Lino de Souza for statistical assistance, and all the efforts of Renata Pellegrino and Diva Maria Lima in helping the collection of data. References Arnardottir, E.S., Mackiewicz, M., Gislason, T., Teff, K.L., Pack, A.I., 2009. Molecular signatures of obstructive sleep apnea in adults: a review and perspective. Sleep 32, 447–470. Carrozza, M.J., Utley, R.T., Workman, J.L., Cote, J., 2003. The diverse functions of histone acetyltransferase complexes. Trends Genet. 19, 321–329. Cassavaugh, J., Lounsbury, K.M., 2011. Hypoxia-mediated biological control. J. Cell. Biochem. 112, 735–744. Heikkinen, J., Risteli, M., Lampela, O., Alavesa, P., Karppinen, M., Juffer, A.H., Myllyla, R., 2011. Dimerization of human lysyl hydroxylase 3 (LH3) is mediated by the amino acids 541–547. Matrix Biol. 30, 27–33. Kawai, T., Matsumoto, M., Takeda, K., Sanjo, H., Akira, S., 1998. ZIP kinase, a novel serine/threonine kinase which mediates apoptosis. Mol. Cell. Biol. 18, 1642–1651. Lavie, L., Lavie, P., 2006. Ischemic preconditioning as a possible explanation for the age decline relative mortality in sleep apnea. Med. Hypotheses 66, 1069–1073. Lin, K.T., Lu, R.M., Tarn, W.Y., 2004. The WW domain-containing proteins interact with the early spliceosome and participate in pre-mRNA splicing in vivo. Mol. Cell. Biol. 24, 9176–9185. Marin, J.M., Carrizo, S.J., Vicente, E., Agusti, A.G., 2005. Long-term cardiovascular outcomes in men with obstructive sleep apnoea-hypopnoea with or without treatment with continuous positive airway pressure: an observational study. Lancet 365, 1046–1053. Tufik, S., Santos-Silva, R., Taddei, J.A., Bittencourt, L.R., 2010. Obstructive sleep apnea syndrome in the Sao Paulo Epidemiologic Sleep Study. Sleep Med. 11, 441–446. Wang, C., Kovanen, V., Raudasoja, P., Eskelinen, S., Pospiech, H., Myllyla, R., 2009. The glycosyltransferase activities of lysyl hydroxylase 3 (LH3) in the extracellular space are important for cell growth and viability. J. Cell. Mol. Med. 13, 508–521.