Plasma APRIL as a Potential Biomarker in the Diagnosis of Obstructive Sleep Apnea

Background: Obstructive sleep apnea (OSA) is the most common type of sleep breathing disorder and is characterized by chronic intermittent hypoxia, which could cause inammation and NF-KB-dependent inammatory pathways activation. Circulating APRIL might play an important role in promoting inammation and NF-KB-dependent inammatory pathways activation. We explored the role of APRIL as a potential mechanism of inammation in OSA patients. Methods: After detailed sleep evaluated, venous blood and demographic data were collected from 155 subjects with varying severity of OSA and 52 control subjects. Plasma levels of APRIL were measured by human Magnetic Luminex assay. Results: Plasma APRIL levels were signicantly higher in OSA subjects compared with control subjects. Categorization of the OSA subjects into mild, moderate, and severe OSA subgroups showed that plasma levels of APRIL increased with severity. After adjusting confounding factors, found that increased plasma APRIL levels were conferred a higher odds ratio (OR) of OSA. Moreover, plasma APRIL levels were positively associated with the apnea-hypopnea index. Furthermore, plasma APRIL showed higher discriminatory accuracy in predicting the presence of OSA. Conclusions:Plasma APRIL levels were signicantly associated with the occurrence of OSA and its severity. APRIL could be a plasma biomarker with a positive diagnostic value for inammation and NF-KB–dependent inammatory pathways activation in subjects with OSA. Monocyte

activation of NF-KB-dependent in ammatory pathways by intermittent hypoxia and reoxygenation in OSA may be an important molecular mechanism of cardiovascular complication [4,19,20]. In addition, several studies showed that monocyte chemotactic protein-1 (MCP-1) and L-selectin as in ammation mediators can be regulated via the activation of NF-KB signaling pathways, and they were signi cantly elevated in patients with OSA [21][22][23][24].
The purpose of this study in patients with OSA was to evaluate whether levels of APRIL, MCP-1, and Lselectin by plasma are elevated; to identify whether the plasma APRIL levels were associated with the presence and severity of OSA and whether plasma APRIL could predict in ammation and NF-KBdependent in ammatory pathways activation in OSA patients.

Study design and participants
We conducted an observational cross-sectional study in Beijing An-Zhen hospital. The study protocol was approved by the Medicine Ethics Committee of Beijing An-Zhen Hospital and adhered to the principles laid out in the Declaration of Helsinki. The project was approved by the Chinese Clinical Trial Registry (No. ChiCTRROC-17011027). In total, 240 participants were recruited and underwent overnight full PSG in Sleep Laboratory between March 2017 and November 2017. The study ow chart is shown in Additional le 1: Figure S1. Participants with central sleep apnea syndrome, chronic obstructive pulmonary disease, cancer, hepatic dysfunction, interstitial lung disease, acute infectious diseases, or abnormal renal function were excluded [25]. Participants refused to sign the consent form, or with incomplete information were also excluded. Finally, we selected a total of 207 eligible participants. Data on participants' baseline demographics were extracted from self-administered questionnaires. Weight (kilograms), height (meters), blood pressure and left ventricular ejection fraction were measured or calculated using standardized and reproducible study protocols. Body mass index (BMI) (kg/m 2 ) was calculated by the formula of weight divided by height squared. Current smokers were de ned as participants who were smoking or stopping smoking for less than one year before enrollment in this study. Drinkers were subjects who consumed alcohol more than three days a week. The de nitions of coronary heart disease (CHD), diabetes mellitus, hypercholesterolemia, and hypertension were based on subject medical histories. Serum lipid, high-sensitivity C-reactive protein, and other routine biochemical parameters were measured in a biochemical analyzer (Hitachi-7600, Tokyo, Japan) using blinded quality control specimens in the Department of the Biochemical Laboratory at Beijing An-Zhen Hospital.

Sleep Data
All participants(n = 207) included in this study underwent detailed sleep evaluated. Neck circumference (cm) was measured between the mid-cervical spine and mid-anterior neck, by using a exible tape, with the participants in the standing position, head held erect [26]. The Epworth Sleepiness Scale (ESS), which was signed to assess the risk of daytime sleepiness, were recorded for each study participant [27]. All participants underwent overnight PSG (SOMNO screen; SOMNO medics GmbH, Randersacker, Germany) approved by the US Food and Drug Administration. Alcohol or sleeping medicines were strictly prohibited before the PSG examination. The SOMNO screen device continuously monitored and recorded abdominal and ribcage motion, respiratory effort air ow, arterial oxygen saturation, electroencephalogram, and electrocardiogram by a computerized polysomnogram. Sleep stage and cardiopulmonary events were manually classi ed and evaluated by certi ed technicians according to the 2012 American Academy of Sleep Medicine criteria [9]. Hypopnea was de ned as ≥ 3% oxygen desaturation sustaining for ≥ 10 s.
Apnea was de ned as a complete cessation of air ow or air ow decrease ≥ 90% relative to the baseline amplitude and persisting for ≥ 10 s. The apnea-hypopnea index (AHI) was calculated as the total number of hypopneas and apneas per hour of sleep. The AHI was used to categorize OSA as none (0-4.9 events/h), mild (5-14.9 events/h), moderate (15-29.9 events/h), and severe (AHI ≥ 30 events/h).

Collection And Storage Of Plasma Sample
Fasting whole blood sample was collected after the PSG examination. Procedures for plasma collection, processing, and storage have been previously described [28].

Luminex Assays
Magnetic Luminex® Assays was a magnetic bead-based antibody microarray founded upon the sandwich immunoassay principle, which can be used to facilitate the simultaneous quantitation of up to 100 soluble analytes in a single sample [29]. APRIL, MCP-1 and L-selectin concentrations were determined by the Luminex magnetic bead cytokines panel based on the competition principle (R&D Systems, Inc. Minneapolis, MN, USA). To ensure the accuracy and validity of the results, as previously described we evaluated the Luminex multiplex assay system by the standard curve and intra-assay variability. Firstly, the standard curve was critical for quantitation measurements. The assay sensitivity and detection range of cytokines were evaluated and the results were shown in Additional le 1: Table S1.
Reading out of range of the standard curve were excluded from all subsequent analyses. Secondly, to determine the precision of the standards and cytokines levels values obtained by the Luminex platforms, we calculated intra-assay performance using the coe cient of variation (CV%) to determine the precision of results. Intra-assay CV < 10% being acceptable. In addition, all plasma samples were optimally diluted to ensure cytokine levels fell within the detection dynamic range of the assay. Statistical analysis Continuous variables were tested for normality of distribution by the Kolmogorov-Smirnov test and were presented as means ± standard deviations (SD) (for normally distributed data) or presented as medians (interquartile ranges) (for asymmetrically distributed data). Categorical variables were expressed as numerals (percentages). Analyses for comparisons demographic, sleep, biological, and cytokine values between OSA and control groups were performed using Student's t-tests for normally distributed data and using the Mann-Whitney U test for asymmetrically distributed data. Differences among mild OSA, moderate OSA and severe OSA were compared by analysis of variance (ANOVA) in normally distributed variables and the Kruskal-Wallis test in asymmetrically distributed variables. Categorical variables were compared by using the chi-square test followed by Fisher exact tests. Spearman correlation coe cient was used to analyze the correlations between cytokine levels and clinical parameters. Multiple logistic regression analyses (forced entry method) were performed to assess the association between OSA and plasma cytokines levels. Multiple linear regression analyses (forced entry method) were performed to assess the association between AHI and plasma cytokines levels. In addition, receiving operator curves (ROC) were calculated for the prediction of OSA based on cytokines levels and the area under curve (AUC), sensitivity, and speci city of the optimal cutoff cytokines were recorded. An AUC of 0.5 indicated no predictive power, whereas an AUC of 1 indicated perfect prediction. A p-value < 0.05 was considered as statistically signi cant. Analyses were performed using SPSS version 25 software (SPSS Inc., Chicago, IL, USA).

Results
Baseline clinical characteristics of the study population The demographic and sleep data of the 207 individuals who participated in this study were shown in Table 1. Subjects with OSA had signi cantly higher BMIs, systolic blood pressure, diastolic blood pressure, neck circumference, ESS, AHI, the percentage of cumulative time with oxygen saturation below 90%, and arousal index than control subjects. The "Lowest SaO2" and "Mean SaO2" were lower in subjects with OSA. A greater proportion of subjects with OSA were male, had coronary heart disease or hypertension compared with the control group. There were no signi cant differences in age and left ventricular ejection fraction between the two groups. The prevalence of currents smokers, drinkers, diabetes mellitus, and hypercholesterolemia were also no differences between the two groups. Subjects with OSA had signi cantly higher triglyceride, leukocyte, uric acid, ALT, GGT, creatinine, and glucose levels compared to patients in the control group (Table 2). Other biological parameters were not a signi cant difference between the two groups. To determine whether OSA at different severity may differentially affect demographic, sleep and biological parameters, we compared the difference of several parameters among three different OSA status (Table 3). Compared to mild OSA, BMIs, diastolic blood pressure, neck circumference, ESS, AHI, the percentage of cumulative time with oxygen saturation below 90%, arousal index, triglyceride, glucose were signi cantly increased in severe OSA groups. The "Lowest SaO2" and "Mean SaO2" decreased in a stepwise fashion from mild to severe OSA patients. Increased plasma APRIL, MCP-1 and L-selectin levels in subjects with OSA To determine the expression levels of APRIL, MCP-1, and L-selectin in subjects with OSA, we used human Magnetic Luminex assay to analyze plasma cytokines in subjects with con rmed OSA and controls. Plasma APRIL, MCP-1, and L-selectin concentrations were signi cantly higher in the OSA group than those in the control group (Fig. 1). Meanwhile, we explore the levels of APRIL, MCP-1, and L-selectin among three different OSA status. We found that the APRIL concentrations were the lowest in mild OSA subjects and increased as the AHI increased, with the highest concentrations in the severe OSA group. The plasma concentrations of MCP-1and L-selectin were no signi cant differences among mild OSA, moderate OSA and severe OSA groups (Fig. 2).
ROC curve analysis for OSA and APRIL, MCP-1, and Lselectin We determined based on ROC curves analysis the optimal threshold value for the optimal meeting point between having the greatest sensitivity and speci city for predicting the occurrence of the OSA. The optimal cutoff value of plasma APRIL for the identi cation of OSA was 4.56 ng/mL with a corresponding sensitivity of 79.43% and speci city of 82.61%. The area under the curve of 0.856 with a P value < 0.001, indicating that APRIL concentration could be a predictor of OSA. The optimal cutoff value of plasma for MCP-1 the identi cation of OSA was 0.20 ng/ml (sensitivity:94.33%, speci city:47.83%) and the AUC was 0.673. The optimal cut-point for plasma L-selectin at OSA was 878.54 ng/ml (sensitivity:86.53%, speci city:43.48%) and the AUC was 0.692. Plasma APRIL showed higher discriminatory accuracy than plasma MCP-1 and L-selectin in predicting the presence of OSA (AUC, 0.856 vs 0.673, 0.692, respectively) ( Fig. 5).

Discussion
In this study, we found that subjects with OSA have signi cantly elevated plasma APRIL, MCP-1 and Lselectin levels compared with controls. Plasma APRIL concentrations were the lowest in mild OSA subjects and increased as the AHI increased, with the highest concentrations in the severe OSA group.
Plasma APRIL levels were signi cantly correlated with the AHI and the "Lowest SaO2", and by multiple regression analyses, we found APRIL were positively and independently associated with the OSA and the AHI, respectively. Meanwhile, plasma APRIL showed higher discriminatory accuracy than plasma MCP-1 and L-selectin in predicting OSA.
Sleep deprivation, sympathetic activation, and CIH was the major pathophysiologic character of OSA, which lead to the release of proin ammatory mediators and the directional migration of leucocytes.
These pathophysiological changes can stimulate the expression of acute-phase proteins and in ammatory mediators [2,5]. Meanwhile, immune cells dominate early atherosclerotic lesions, their effector molecules accelerate in ammation and NF-KB activity, which can elicit atherosclerotic lesions rupture and trigger the acute onset of arterial thrombosis [30,31]. Several studies have shown that treating OSA with continuous positive airway pressure improves in ammation and reduces of NF-KB signaling [20,32]. Even in patients with OSA who were free of overt cardiovascular disease, the physiological perturbations of in ammation and NF-KB-dependent in ammatory pathways activation often caused blood pressure and heart rate elevations, ventricular failure, myocardial infarction, and stroke [4,33]. Given its association with cardiovascular disease, it was extremely essential to advance the research agenda to explore validated tools for diagnosing OSA in the early stage.
The AASM clinical practice guidelines suggest that OSA is diagnosed based on clinical measures, including the subjective assessment of somnolence and the overnight multi-channel PSG [9]. The former is limited by low speci city and accuracy for diagnosis of OSA, the latter is expensive, labor-intensive, time-consuming tests and impractical for the clinical evaluation of large at-risk populations [10].
Therefore, new technologies such as circulating biomarkers could provide important information on clinical signi cance for diagnosing OSA.
By human Magnetic Luminex assay, we evaluated plasma APRIL, MCP-1, and L-selectin levels in subjects with con rmed OSA and controls. Intriguingly, we found that plasma APRIL, MCP-1, and L-selectin levels were signi cantly higher in the OSA group compared with the control group. In addition, plasma APRIL levels were statistically higher in subjects with severe OSA than subjects with mild OSA. APRIL(also named TNFSF13), a member of the tumor necrosis factor family, which is secreted as a soluble factor at low levels in immunological tissues especially by antigen-presenting cells, inactive B cells, T cells, monocytes, neutrophils, macrophages, and dendritic cells, as well as by epithelial cells, osteoclasts, and megakaryocytes [34][35][36]. Increased levels of APRIL have been found in several autoimmune diseases such as systemic lupus erythematosus, rheumatoid arthritis, multiple sclerosis, and immunode ciency [11][12][13][14]. Several studies have indicated that the relationship of APRIL and NF-KB signaling pathway activation is one of a complex interactive, APRIL could induce in ammatory activation by the activation of NF-KB signaling pathway [15,16], and conversely, NF-KB-dependent pathways can in uence APRIL secretion [17,18]. A large body of evidence, including experimental and clinical studies, have demonstrated that the activation of NF-KB-dependent in ammatory pathways by intermittent hypoxia and reoxygenation in OSA may be an important molecular mechanism of cardiovascular complication [4,19,20]. Moreover, both RNA and protein expression of APRIL have been shown in human plaque lymphocytes and macrophages and plasma [37,38]. Recently, Bernelot Moens et al found that APRIL transgenic mice do show potential plaque stabilizing features in advanced atherosclerotic lesions [39]. However, to the best of our knowledge, no previous studies have investigated the relationship between the circulating APRIL and the presence of OSA. The present study showed that a progressive increase in the concentrations of APRIL with the severity of OSA, increase plasma APRIL levels were signi cantly associated with the occurrence of OSA and its severity.
MCP-1 is one of the best-studied CC chemokines that is expressed by in ammatory cells and stromal cells, and its chemotactic activity could be upregulated after proin ammatory stimuli [40,41]. Evidence from animal models and in vitro experiments suggested that IH induced the synthesis and expression of MCP-1 via the activation of NF-KB signaling pathway [21,42]. Furthermore, early studies suggested that OSA-induced hypoxia would increase circulating MCP-1 levels and effective therapy for OSA could reduce the expression of MCP-1 [43,44]. Our results suggested that the plasma MCP-1 level was signi cantly higher in subjects with OSA compared with subjects without OSA, but there were no statistical differences in MCP-1 levels among the mild OSA, moderate OSA, and severe OSA.
L-selectin is a cell membrane-surface receptor on leukocytes, which mediated the rolling and tethering of leukocytes into the injured tissues independently of other adhesion molecules [24]. Early studies found that the monocyte/leukocyte induced effects on expression of the L-selectin through the NF-KB signaling pathway [23]. Meanwhile, some evidence demonstrated that circulating L-selectin levels were signi cant and independent indicators of accentuated atherogenesis in patients with OSA [24,45]. However, some studies showing the correlation between circulating L-selectin and OSA were not detected, probably due to the de ning criteria of the OSA were inconsistent [46,47]. In the present study, we found that the circulating L-selectin levels were signi cantly increased in the OSA group, and it was independently associated with the occurrence of the OSA and its severity. Of note, no signi cant differences in L-selectin levels were observed among the mild OSA, moderate OSA, and severe OSA.
And intriguingly, we also found that plasma APRIL levels were signi cantly correlated with plasma MCP-1 and L-selectin levels. Of note, in terms of AUC, the discriminatory accuracy of plasma APRIL signi cantly exceeded those of plasma MCP-1and L-selectin. These interesting ndings implied that plasma APRIL levels might be related to in ammation and NF-KB-dependent in ammatory pathways activation induced by OSA.
Care was taken to avoid bias in this study, the Luminex experiment was performed according to the manufacturer's instructions by a trained experimenter who was unaware of patients' clinical data.
Moreover, in the statistical analysis, adjustments were made for the confounding effects of risk factors for plasma APRIL levels and OSA/AHI. However, our study had several certain limitations. First, we included only newly diagnosed, and untreated OSA. Second, our results didn't elucidate the speci c mechanisms of APRIL in OSA.

Conclusions
In conclusion, we found that plasma APRIL levels were signi cantly elevated in patients with OSA and were associated with the occurrence of OSA and its severity. Plasma APRIL levels provided higher discriminatory accuracy levels for patients with OSA. APRIL could be a biomarker with a positive diagnostic value for in ammation and NF-KB-dependent in ammatory pathways activation in patients with OSA. Future studies are needed to highlight underlying mechanisms and the predictive value of APRIL for outcome in OSA patients.

Declarations
Additional File Additional le 1: Table S1: Evaluation of Luminex assay standard curves and intra-assay variability for APRIL, MCP-1, and L-selectin levels determined using the Luminex. Table S2: Association between OSA and MCP-1 levels in univariate and multiple logistic regression models. Table S3: Association between OSA and L-selectin levels in univariate and multiple logistic regression models. Table S4: Association between levels of AHI and MCP-1 levels in univariate and multiple linear regression models. Table S5: Association between levels of AHI and L-selectin levels in univariate and multiple linear regression models. Figure S1: Flow diagram of participants recruitment in this study.    Scatterplots of MCP-1 and L-selectin plasma levels vs APRIL plasma levels. r: Spearman correlation coe cients; APRIL, A proliferation-inducing ligand; MCP-1, Monocyte chemotactic protein-1.

Figure 5
Receiving operator curves using plasma APRIL, MCP-1, and L-selectin levels for prediction of OSA. The AUC for the plasma APRIL levels (0.856) was higher than plasma MCP-1 (0.673), and L-selectin (0.692).

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. additional le20200129.docx