Changes in Expression of the CLOCK Gene in Obstructive Sleep Apnea Syndrome Patients Are Not Reverted by Continuous Positive Airway Pressure Treatment

Purpose Metabolic syndrome and cardiovascular disease are strongly associated with obstructive sleep apnea syndrome (OSAS), which causes substantial changes to normal circadian physiological functions, including metabolic pathways. Because core clock genes are known to be modulated by sleep/vigilance cycles, we asked whether the expression level of mRNA coding for clock genes is altered in non-treated OSAS patients and if it can be corrected by standard continuous positive airway pressure (CPAP) treatment. Methods Peripheral blood was collected from male patients diagnosed with severe OSAS (apnea-hypopnea index ≥ 30/h) before and after treatment initiation. qPCR was used to measure mRNA levels of genes associated with the central circadian pacemaker including CLOCK, BMAL1, Cry1, Cry2, and three Period genes (Per 1, 2, 3) in peripheral blood mononuclear cells (PBMCs). Results We found statistically significant differences for CLOCK (p-value = 0.022) expression in PBMCs of OSAS patients which were not reverted by treatment with CPAP. We have also found a substantial decrease in the slow wave sleep (SWS) content in OSAS patients (p-value < 0.001) that, contrary to REM sleep, was not corrected by CPAP (p-value = 0.875). Conclusion CPAP treatment does not correct substantial changes in expression of core clock genes in OSAS patients. Because CPAP treatment is also unable to normalize the SWS in these patients, it is likely that additional therapeutic interventions that increase SWS content and complement the benefits of CPAP are required to more effectively reduce the known increased cardiovascular risk associated with OSAS patients.

by sleep/vigilance cycles (3). Importantly, single genetic disruption of the core clock genes CLOCK (4) and BMAL1 (5, 6) cause fullblown metabolic syndrome phenotypes in mice. Milder metabolic syndrome phenotypes have also been observed in double or triple, but not single, inactivation of additional core clock genes, including Per1, 2, 3, and Cry1 and 2 (7). In this work, we investigated whether the expression level of mRNA coding for clock genes is altered in non-treated OSAS patients and if it can be corrected by continuous positive airway pressure (CPAP) treatment.

Measurement of gene expression
We have used quantitative RT-PCR to systematically quantify the expression of the core clock genes in peripheral blood cells of severe OSAS patients in comparison with that of matched nonaffected control volunteers, before and after 1 month of treatment with CPAP. Several studies have demonstrated that clock gene expression profiles in human peripheral blood mononuclear cells (PBMCs) can be used as surrogate measures of central circadian rhythm expression changes in patients and healthy subjects (8)(9)(10)(11). Therefore, peripheral blood was collected from the study subjects before and after 1 month of treatment initiation. Collections were always performed between 8 and 10 a.m. Blood was then used to perform routine biochemical analyses (including cortisol levels) and to isolate PBMCs. The following hematologic and biochemical parameters were investigated in all participants: full peripheral blood cell count; hepatic enzymes (aspartate transaminase, alanine transaminase, and gamma-glutamyltransferase); renal parameters (creatinine, blood urea nitrogen); standard metabolic parameters (uric acid, glucose, total cholesterol, low density lipoprotein, high density lipoprotein, triglycerides, and C-reactive protein). RNA was isolated and quantitative RT-PCR used to measure mRNA levels of genes associated with the central circadian pacemaker including CLOCK, BMAL1, two Cry (Cry 1, 2), and three Period genes (Per 1, 2, 3). The selected patients were then evaluated 1 month after therapy initiation with CPAP and the mRNA level of the same genes tested. The sequence of the oligos used for RT-PCR can be found Table S2 in the Supplementary Material.

statistical analysis
A non-parametric approach was chosen for the data analysis. We used median and quartiles for the descriptive statistical analysis.
To test hypotheses related to differences between the control group and the patient group, we used the Wilcoxon signed-rank test. To investigate relationships between gene expression and other parameters, we implemented a correlation analysis based on Spearman's rank correlation coefficient.

resUlTs
Using Wilcoxon signed-rank test, we found statistically significant differences between patients and control volunteers in the ESS (W = 51.5, p-value = 0.007) and in polysomnography parameters  These comparisons did not reveal any significant differences.
On the comparison between control volunteers and patients before treatment initiation, we found statistically significant differences for CLOCK (W = 15, p-value = 0.022) expression in PBMCs of patients, but not for Cry 1 and 2 or the three Period genes tested ( Table 2). The differences of BMAL1 expression between control volunteers and patients nearly reached statistical significance (W = 20, p-value = 0.068), an outcome that is likely, should the population size be larger.
Patient treatment with CPAP did not change abnormal expression of clock genes ( Table 3). It also did not improve the reduced levels of SWS observed in non-treated OSAS patients (p-value = 0.875, Table 4), contrary to REM sleep. The persistence of the abnormal expression pattern of the CLOCK gene cannot be attributed to the lack or sub-optimal adherence levels to CPAP treatment as demonstrated by the metrics of CPAP adherence of our patients ( Table 5; Table S1 in Supplementary Material). Our data show that adherence to CPAP treatment is well above the minimum required level recommended by the most recent proposal of the American Thoracic Society (12) that points to the need of CPAP use in more than 70% of nights and for more than 4 h per night.

DiscUssiOn
In our study, as expected, control volunteers and OSAS patients could be easily distinguished by the statistical analysis of the ESS results and polysomnography parameters. They could also be distinguished on the basis of expression levels of CLOCK, but not any of the other core clock genes, with the possible exception of  BMAL1. Interestingly, these results are strikingly reminiscent of metabolic syndrome phenotypes found in the knockout of the corresponding genes in mice as described earlier. The levels of expression of the CLOCK gene were altered in OSAS patients before and after CPAP treatment, suggesting that either the standard therapeutic management of these patients is insufficient to correct the disruption of the circadian patterns of genes with expression conditioned by cycles of sleep/vigilance or changes in severe sleep apnea patients are structural, for example, as a consequence of irreversible cerebral vascular lesions. Before patients were treated, we observed consistently and substantially reduced levels of REM and SWS content as compared to healthy controls. Perhaps significantly, while all the patients who initiated CPAP recovered to normal levels of REM sleep, SWS content remained invariably abnormal, at much lower levels than controls ( Table 4). It was previously shown that even in young healthy adults, selective suppression of SWS causes pronounced decrease of insulin sensitivity, reduced glucose tolerance, and higher diabetes risk (13).
While it is clear that CPAP substantially improves cognitive and behavioral aspects of the disease (14), metabolic parameters and other OSAS-related risk factors are only partially modified (15), which might explain the incomplete reduction of cardiovascular risk in OSAS. These facts point to the need to complement CPAP therapy with other therapy interventions that are aimed at the sleep structure and circadian rhythm normalization and are likely to have a substantial impact on the natural history of the disease. Drugs that effectively improve the SWS content of patients might constitute prime candidates for this goal.

study limitations
While our major conclusions are supported by robust statistically significant results, our study would benefit from increasing the size of the population of both patients and controls. It is likely that with more samples, statistically significant associations will be found for additional clock genes, especially in the case of BMAL1, where a trend is already clear with the current analysis.

inFOrMeD cOnsenT
Informed consent was obtained from all individual participants included in the study.

eThics sTaTeMenT
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional (Centro Hospitalar Lisboa Norte, EPE) and national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.