Long non-coding RNA MALAT1 exacerbates acute respiratory distress syndrome by upregulating ICAM-1 expression via microRNA-150-5p downregulation

Acute respiratory distress syndrome (ARDS) is a severe form of acute lung injury in which severe inflammatory responses induce cell apoptosis, necrosis, and fibrosis. This study investigated the role of lung adenocarcinoma transcript 1 (MALAT1) in ARDS and the underlying mechanism involved. The expression of MALAT1, microRNA-150-5p (miR-150-5p), and intercellular adhesion molecule-1 (ICAM-1) was determined in ARDS patients and lipopolysaccharide (LPS)-treated human pulmonary microvascular endothelial cells (HPMECs). Next, the interactions among MALAT1, miR-150-5p, and ICAM-1 were explored. Gain- or loss-of-function experiments in HPMECs were employed to determine cell apoptosis and inflammation. Furthermore, a mouse xenograft model of ARDS was established in order to verify the function of MALAT1 in vivo. MALAT1 and ICAM-1 were upregulated, while miR-150-5p was downregulated in both ARDS patients and LPS-treated HPMECs. MALAT1 upregulated ICAM-1 expression by competitively binding to miR-150-5p. MALAT1 silencing or miR-150-5p overexpression was shown to suppress HPMEC apoptosis, decrease the expressions of pro-inflammatory cytokines (IL-6, IL-1β and TNF-α) and E-selectin in HPMECs, as well as alleviated lung injury in nude mice. These findings demonstrated that MALAT1 silencing can potentially suppress HPMEC apoptosis and alleviate lung injury in ARDS via miR-150-5p-targeted ICAM-1, suggestive of a novel therapeutic target for ARDS.

AGING Emerging evidence has confirmed the relevance of lncRNAs to respiratory diseases and thus targeting lncRNAs has been highlighted as a novel therapeutic strategy for respiratory disease treatment [7,8]. Lung adenocarcinoma transcript 1 (MALAT1) is an abundant lncRNA localized to nuclear speckles, which consists of a series of pre-mRNA processing factors [9]. The expression of MALAT1 is strongly regulated in lung adenocarcinoma and other physiological processes [10]. A high expression level of MALAT1 is associated with increased ARDS risk, disease severity, and increased mortality in patients with sepsis [11]. LncRNAs are likely to mediate the function of microRNAs (miRNAs or miRs) by regulating their gene expression, acting as endogenous sponges [12]. In particular, the overexpression of MALAT1 has been reported to sponge miR-150-5p and downregulate its expression in chondrocytes [13]. Multiple miRs have also been implicated in the progression of respiratory diseases [14]. miR-150 can exert protection against cigarette smoke-induced lung inflammation and resist airway epithelial cell apoptosis via downregulation of p53 [15]. Moreover, miR-150-5p is capable of suppressing the expression of proinflammatory cytokines in patients with ischemic stroke [16]. In addition, another ARDS-related factor, intercellular adhesion molecule-1 (ICAM-1) has been found to help ameliorate lung inflammation in a mouse model of ARDS after its expression is reduced [17]. ICAM-1 is a type of cell surface adhesion receptor that can promote multiple effectors/target cell interactions in tissues impacted by inflammatory or immune processes [18] and is involved in elevating the permeability of PMECs [19]. Herein, the current study was conducted to explore whether the MALAT1/miR-150-5p/ICAM-1 signaling axis is involved in mediating the biological functions of PMECs and lung injury following ARDS as well as the underlying mechanisms.

Downregulation of MALAT1 suppresses apoptosis of human pulmonary microvascular endothelial cells (HPMECs) and decreases expression of proinflammatory cytokines and adhesion factors in ARDS
The expression of MALAT1 was initially determined in peripheral blood samples of 46 healthy controls and 46 patients with ARDS by reverse transcription quantitative polymerase chain reaction (RT-qPCR). Compared with healthy controls, the expression of MALAT1 was increased in patients with ARDS caused by different etiologies (p < 0.05) ( Figure 1A). Thereafter, HPMECs were treated with lipopolysaccharide (LPS) to induce inflammation. RT-qPCR revealed that the expression of MALAT1 was increased in HPMECs upon treatment with LPS (p < 0.05) ( Figure 1B). Fluorescence in situ hybridization (FISH) revealed that MALAT1 was mainly localized in the cytoplasm with minimal expression in the nucleus ( Figure 1C). In addition, the expression of MALAT1 was found to be increased in LPS-treated HPMECs transfected with overexpression (oe)-MALAT1, and conversely, it was decreased upon MALAT1 silencing (short hairpin RNA [sh]-MALAT1-1 or sh-MALAT1-2) (p < 0.05) ( Figure 1D).
Subsequently, the expression of proinflammatory factors (interleukin-6 [IL-6], IL-1β and tumor necrosis factor-α [TNF-α]) and endothelial cell adhesion molecules (E-selectin and ICAM-1) was detected by enzyme-linked immunosorbent assay (ELISA) ( Figure  1E) and immunofluorescence ( Figure 1F). Meanwhile, the apoptosis of HPMECs was analyzed by terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL) assay ( Figure 1G), and the expression of B-cell lymphoma 2 (Bcl-2), Bcl-2associated X protein (Bax), and cleaved caspase 3 was determined by Western blot analysis ( Figure 1H). The results revealed that overexpression of MALAT1 resulted in a notable increase in the expression of IL-6, IL-1β, TNF-α, E-selectin, ICAM-1, Bax and cleaved caspase 3, as well as in apoptosis rate, whereas the expression of Bcl-2 was decreased in HPMECs (p < 0.05). On the other hand, silencing MALAT1 led to a marked decrease in the expression of IL-6, IL-1β, TNF-α, E-selectin, ICAM-1, Bax and cleaved caspase 3 along with lowered apoptosis rate, while the expression of Bcl-2 was increased (p < 0.05). Taken together, the data indicated that silencing MALAT1 could inhibit apoptosis of HPMECs and reduce the expression of pro-inflammatory cytokines and adhesion factors.

Overexpression of miR-150-5p reverses the promoting effects of MALAT1 on the progression of ARDS
The biological prediction website starBase (http://starbase.sysu.edu.cn/) predicted a binding site between miR-150-5p and MALAT1 (Figure 2A). The expression of miR-150-5p was lower in 46 patients with ARDS than that in the healthy controls (p < 0.05) ( Figure 2B). The expression of miR-150-5p was also decreased in HPMECs following LPS treatment (p < 0.05) ( Figure 2C). Therefore, it was speculated that MALAT1 might affect ARDS by binding to miR-150-5p. To verify this speculated relationship, dual luciferase reporter assay was conducted, and the results ( Figure 2D) showed miR-150-5p mimic transfection led to a decrease in the luciferase activity of MALAT1-wild type (WT) as compared with negative control (NC) transfection, (p < 0.05), while the luciferase activity of MALAT1-mutant (MUT) showed no changes (p > 0.05). RNA binding protein immunoprecipitation (RIP) and AGING RNA pull-down assays showed that miR-150-5p-WT could combine more MALAT1 in comparison to miR-150-5p-MUT and Bio-NC (p < 0.05), which further confirmed that MALAT1 could bind to miR-150-5p. Moreover, the expression of miR-150-5p was upregulated in LPS-treated HPMECs upon miR-150-5pmimic transfection, which was negated by dual transfection with miR-150-5p mimic and oe-MALAT1 (p < 0.05) ( Figure 2F). These results demonstrated that MALAT1 could bind with miR-150-5p and consequently reduce its expression.

AGING
(p < 0.05). As illustrated in Figure 2I, 2J, TUNEL assay, RT-qPCR and Western blot analysis demonstrated that HPMECs treated with miR-150-5p-mimic showed a decrease in apoptosis rate and the expression of Bax and cleaved caspase 3, yet an increase in Bcl-2 expression, as compared with cells treated with miR-150-5p-mimic NC (p < 0.05). However, co-transfection with miR-150-5p mimic and oe-MALAT1 led to a converse trend in expression of the aforementioned factors (p < 0.05). Overall, the above diagrams served to illustrate that miR-150-5p overexpression was capable of rescuing the stimulated apoptosis of HPMECs and increased expression of pro-inflammatory cytokines and adhesion factors induced by MALAT1 overexpression.

Downregulation of MALAT1 or overexpression of miR-150-5p alleviates lung injury
Finally, we aimed to explore the role of MALAT1 and miR-150-5p in ARDS in vivo. Hematoxylin-eosin (HE) staining analysis showed that mouse lung tissues underwent a series of typical pathological changes after AGING treatment with LPS ( Figure 5A, 5B). Next, the partial pressure of oxygen (PaO2) ( Figure 5C-5E) in arterial blood, bronchoalveolar lavage (BAL) cell count, and neutrophil cell count, as well as the concentration of total protein, albumin and immunoglobulin M (IgM) in mice ( Figure 5D-5G) were measured. In addition, LPStreated mice presented with increases in pathological score, total inflammatory cells and neutrophils in BAL, and concentration of total protein, albumin, and IgM in BAL while PaO2 was decreased (p < 0.05). These findings indicated that LPS induced pneumonia and lung injury in mice.
Subsequently, the mice were injected with different lentiviruses to alter the expression of miR-150-5p and MALAT1. As shown in Figure 5F-5J, as compared with sh-NC or agomiR-150-5p NC treatment, mice treated with sh-MALAT1 or agomiR-150-5p showed reductions in lung injury, pathological score, total inflammatory cells and neutrophils in BAL, total protein, albumin and IgM in BAL, whereas PaO2 was elevated (p < 0.05). These results demonstrated that the downregulation of MALAT1 or overexpression of miR-150-5p could alleviate LPS-induced pulmonary inflammation and enhance lung permeability. The data were measurement data and expressed as mean ± standard deviation. The data between two groups were compared using unpaired t-test and those among multiple groups were analyzed by one-way ANOVA, with Tukey's post hoc test. The cell experiment was repeated three times independently. were measurement data and expressed as mean ± standard deviation. The data between two groups were analyzed by unpaired t-test and those among multiple groups were analyzed by one-way ANOVA, with Tukey's post hoc test. The cell experiment was repeated three times independently.

DISCUSSION
ARDS is a multifactorial disorder and a severe manifestation of acute lung injury characterized by high morbidity and mortality rates [20]. Studies have documented the prognostic potential of lncRNAs in ARDS [21,22]. The present study aimed to explore the effects of MALAT1 on the progression of ARDS. Our results demonstrated that downregulation of MALAT1 could potentially decrease the expression of ICAM-1 by AGING up-regulating miR-150-5p, thereby suppressing the apoptosis of HPMECs and alleviating lung injury in ARDS mice.
Our initial findings demonstrated upregulated MALAT1 and ICAM-1 alongside downregulated miR-150-5p in peripheral blood samples of patients with ARDS and LPS-treated HPMECs. LncRNAs have been found to be upregulated in diverse intricate human diseases such as pulmonary fibrosis and lethal lung developmental disorders [23,24]. Similar to our findings, MALAT1 has been shown to be significantly increased in the peripheral blood mononuclear cells with infection exposure and treated with LPS in neonatal respiratory distress syndrome [25]. In addition, a recent study has revealed that MALAT1 is highly expressed in the plasma of ARDS patients and peripheral blood mononuclear cells [26]. Patients with ARDS show significantly reduced serum miR-150 expression, which has been found negatively associated with disease severity and 28-day survival [27]. Furthermore, the inhibition of ICAM-1 has been found to help ameliorate LPS-induced ARDS in rats [28], which suggests an increased ICAM-1 level in ARDS.
We also found that MALAT1 could downregulate miR-150-5p by competitively binding to it, thereby upregulating the expression of ICAM-1, a target of miR-150-5p. LncRNAs have been widely reported to serve as miR sponges, thereby regulating the expression of their target genes of miRs by serving as competing endogenous RNAs (ceRNAs) [29]. For instance, MALAT1 is able to sponge miR-124 and thus facilitates cell apoptosis in Parkinson's disease [30]. In addition, MALAT1 has been found to promote cell apoptosis in testicular ischemia-reperfusion injury through upregulation of TRPV4 expression by sponging miR-214 [31]. By functioning as a sponge or as a competing endogenous noncoding RNA for miR-150-5p, lncRNA FOXD3-AS1 promotes hyperoxia-induced lung epithelial cell death [32]. In accordance with the findings from the current study, miR-150 has been demonstrated to be a target of MALAT1 in airway smooth muscle cells and MALAT1 can augment the expression of eIF4E, a target of miR-150, as a ceRNA for miR-150 [33].
Another key observation of the current study indicated that downregulation of MALAT1 or overexpression of AGING miR-150-5p inhibited the apoptosis of HPMECs, suppressed the inflammatory response (reflected by decreased expression of Bax, cleaved caspase 3, IL-6, IL-1β, TNF-α and increased expression of Bcl-2) and decreased the expression of adhesion factors ICAM-1 and E-selectin in vitro, and also alleviated lung injury in ARDS in vivo. These observations are supported by existing evidence. The downregulation of MALAT1 has earlier been shown to inhibit the apoptosis of hypoxic/reoxygenated human umbilical vein endothelial cells in vitro [34], which is consistent with our findings. TNF-induced miRs can mediate TNF-induced expression of both E-selectin and ICAM-1 in human endothelial cells and thus provide a feedback control of inflammation [35]. The upregulation of exogenous miR-150 is capable of contributing to an enhancement in HMEC-1 cell migration in atherosclerosis [36], which is aligned with our findings. Overexpression of MALAT1 promotes cell apoptosis and then deteriorates lung injury through sponging of miR-425 during ARDS [26]. In LPS-induced acute kidney injury, silencing MALAT1 elevates the expression of miR-146a, leading to repression of the pro-inflammatory nuclear factor-κB (NF-κB) pathway and its downstream transcription factors [37]. Increased miR-150 decreases neutrophil counts and production of inflammatory cytokines IL-1β, IL-6, and TNF-α along with levels of total protein, albumin and IgM in the BAL fluid in LPS-induced acute lung injury mice in vivo, as well as alleviating LPS-induced A549 cell apoptosis in vitro [27]. Furthermore, blocking pulmonary ICAM-1 expression potentially alleviates lung injury in diet-induced pancreatitis [38]. ICAM-1 has also been reported to induce IL-6 and TNF-α production in a Kupffer celldependent manner during liver regeneration following hepatectomy [39]. In addition, retreatment with anti-ICAM-1 antibody can reduce the extent of acinar cell damage and inhibit their apoptosis [40]. On the basis of the aforementioned information, we reasoned that MALAT1 could upregulate the expression of ICAM-1 by binding to miR-150-5p, thereby suppressing the apoptosis of HPMECs and alleviating lung injury in ARDS.
To conclude, the present study demonstrated that MALAT1 silencing could suppress HPMEC apoptosis and alleviate lung injury in ARDS via the miR-150-5pmediated ICAM-1 axis (Figure 6), indicating that targeting MALAT1 may serve as a promising therapeutic strategy for ARDS. Our work thus identifies the evolutionarily MALAT1/miR-150-5p/ICAM-1 signaling as a major molecular pathway in the control of ARDS. Further investigations into the interaction between MALAT1, miR-150-5p, and ICAM-1 are still required to fully elucidate the specific mechanisms of MALAT1 in ARDS.

Clinical blood collection
We enrolled forty-six patients with ARDS who were hospitalized at the respiratory department, emergency department, department of critical care medicine or general surgery department of The First Affiliated Hospital of Zhengzhou University from January 2017 to June 2018 in this study. Based on their PaO2/FiO2 ratios on the day when they were diagnosed with ARDS, the patients were classified into 3 groups: mild (200 < PaO2/FiO2 ≤ 300 mm Hg; n = 15), moderate (100 < PaO2/FiO2 ≤ 200 mm Hg; n = 22) and severe (PaO2/FiO2 < 100 mm Hg; n = 9) groups. Meanwhile, 46 age-matched healthy individuals were recruited as healthy controls. Based on the data obtained from medical history and physical examination, none of the healthy controls had any symptoms or signs of pulmonary discomfort, history of underlying diseases, or pulmonary abnormalities in routine chest X-ray examination. Peripheral venous blood (2 mL) was collected from all participants, which was then subjected to ethylenediaminetetraacetic acid (EDTA) anticoagulation and centrifuged at 3000 rpm for 5 minutes. The upper-layer serum was stored at -80°C.

Fish
The subcellular localization of MALAT1 and miR-150-5p in HPMECs was detected using a FISH kit [RiboTM lncRNA FISH Probe Mix (Red)] in accordance with the kit instructions. In short, cells were inoculated into a 24well culture plate at a density of 6 × 10 4 cells/well. When cell confluence reached about 80%, the cells were rinsed with phosphate buffer saline (PBS) and fixed with 4% paraformaldehyde at room temperature. After being treated with protease K (2 μg/mL), glycine, and ethylphthalide reagent, the cells were incubated with 250 μL pre-hybridization solution at 42°C for 1 hour.

RT-qPCR
Total RNA was extracted from the cultured cells using a TRIzol reagent (Invitrogen, Carlsbad, CA, USA), and the purity and concentration of the extracted RNA were measured using a Nano-Drop ND-1000 spectrophotometer. The obtained RNA was then reverse transcribed into complementary DNA (cDNA) according to the instructions of the PrimeScript RT reagent kit (Takara, Otsu, Shiga, Japan). Next, RT-qPCR was performed with a SYBR®Premix Ex TaqTM II kit (Takara, Otsu, Shiga, Japan) on an ABI 7500 instrument (Applied Biosystems, Foster City, CA, USA). The primer sequences (Table 1) for MALAT1, miR-150-5p, ICAM-1 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were designed and synthesized by Shanghai Sangon Biotechnology Co., Ltd. (Shanghai, China). The relative expression of the target gene was calculated using the 2 -ΔΔCt method with GAPDH serving as the internal reference [42][43][44][45].

RIP assay
The binding of MALAT1 and Argonaute2 (AGO2) protein was detected using a RIP kit (Millipore Corp of Billerica, Massachusetts, USA). Briefly, HPMECs were lysed with radioimmunoprecipitation assay (RIPA) lysis buffer (P0013B, Beyotime Biotechnology Co., Shanghai, China) on ice bath for 5 minutes, followed by centrifugation at 14000 rpm for 10 minutes to collect the supernatant. One portion of the cell extract was used as input, and the other portion was co-precipitated with antibody. Next, 50 μL magnetic beads were collected from each co-precipitation reaction system, washed, and re-suspended in 100 μL RIP wash buffer. Next, 5 μL antibody was added to the magnetic beads and incubated for 30 minutes at room temperature for binding. The magnetic beads-antibody complex was washed, resuspended in 900 μL RIP wash buffer, and then incubated overnight at 4°C with 100 μL cell extract. The samples were subsequently placed on magnetic pedestals to collect bead-protein complexes. The precipitated complex and Input were treated with protease K, followed by extraction of RNA for subsequent RT-qPCR detection. The antibodies used in RIP included rabbit anti-human AGO2 (ab186733; dilution ratio of 1 : 50, Abcam Inc., Cambridge, MA, USA) and rabbit antihuman immunoglobulin G (IgG; ab109489; dilution ratio of 1 : 100, Abcam Inc., Cambridge, MA, USA) which served as NC.

RNA-pull down assay
Biotinylated WT-bio-miR-132 and MUT-bio-miR-132 (50 nM each) were each transfected into HPMECs. After 48 hours, the cells were collected and incubated with specific lysis buffer (Ambion, Austin, Texas, USA) for 10 minutes. Next, the lysate was co-incubated with M-280 streptavidin magnetic beads (S3762, Sigma-Aldrich Chemical Company, St Louis MO, USA) that had been pre-coated with RNase-free bovine serum albumin (BSA) and yeast transfer RNA (tRNA; RNABAK-RO, Sigma-Aldrich Chemical Company, St Louis MO, USA) at 4°C for 3 hours. Finally, the enrichment of MALAT1 was detected using RT-qPCR.

ELISA
The levels of IL-6, IL-1β, and TNF-α were determined in cell lysate in each group using ELISA kits (Rapidbio, Inc., West Hills, CA, USA) according to the kit instructions. In brief, the antigen was diluted with coating diluent at a dilution ratio of 1 : 20. Next, 100 μL standard dilution solution was added to each well, followed by overnight incubation at 4°C. The diluted samples were added into the reaction wells of the ELISA plate (100 μL in each well). Negative and positive controls were set. Each well was added with 100 μL enzyme conjugates that had been diluted with sample dilution solution, followed by 30-minute reaction at 37°C. Next, 100 μL horseradish peroxidase substrate solution was added, followed by coloration at 37°C for 10 to 20 minutes. In the event of significant color changes in the positive control or slight color changes in the NC, 50 μL termination solution was added to each well to terminate the reaction. The optical density (OD) of each well was then measured at a wavelength of 450 nm using the Spectramax M5 microplate reader (Molecular Devices, Sunnyvale, CA, USA).

TUNEL assay
TUNEL staining was performed using an in situ apoptosis detection kit (Chemicon International Inc., Temecula, CA, USA). In brief, the transfected HPMECs were pretreated with 10 nmol/L docetaxel (DTX) for 24 hours. The cells were rinsed and stained according to the kit instructions. A fluorescence microscope (Axiovert 200, Carl Zeiss, Oberkochen, Germany) was used to observe and photograph the cells. The number of positive-stained cells was counted under an EVOS FL microscope (Thermo Fisher Science, Waltham, MA, USA).

Immunofluorescence
Cells were fixed with 4% paraformaldehyde and then incubated with monoclonal antibodies against ICAM-1 and E-selectin (Abcam Inc., Cambridge, UK), respectively. The cells were then incubated with fluorescent dye-labeled secondary antibody (AF555; Abcam Inc., Cambridge, MA, USA), followed by DAPI staining. The fluorescence intensities of AF555 and AF647 were recorded at an excitation wavelength of 552 nm and 638 nm and an emission wavelength of 570 nm and 665 nm, respectively. At last, the protein levels of E-selectin and ICAM-I in cells were determined.

Establishment of acute lung injury mouse models
A total of 40 C57BL/6 WT mice purchased from Vital River (Beijing, China) were anesthetized and underwent oral intubation. Next, mouse models of acute lung injury were established by intratracheal instillation of LPS (1.5 mg/kg). The healthy control mice were injected with NaCl solution via the tail vein. The lentiviruses of AGING agomiR-150-5p-NC, agomiR-150-5p (20 mg/kg; Sigma-Aldrich Chemical Company, St Louis MO, USA) and sh-MALAT1 NC, sh-MALAT1 (10 mg/kg; GenePharma, Shanghai, China) were intratracheally injected into the mice (n = 8 for each treatment). One day after injection, the mice were stimulated by LPS. On the 21 st day after LPS stimulation, the mice were euthanized and their lung tissues were collected.

HE staining and arterial blood gas analysis
Paraffin-embedded lung tissues were sliced into 5-μmthick sections, dewaxed with xylene, and hydrated with gradient ethanol. The sections were stained with hematoxylin for 7 minutes, treated with 95% ethanol for 5 seconds, hydrated with gradient ethanol, and stained with eosin for 1 minute. The sections were then cleared with xylene, dried, and finally mounted with neutral balsam. The histological changes were observed under an optical microscope. Lung injury scores were assessed based on the following criteria [48]: (a) alveolar congestion; (b) erythrocyte exudation; (c) neutrophil exudation or aggregation in alveoli; (d) alveolar wall thickening and hyaline membrane formation. Based on severity, the disease was scored as the following four grades: grade 0, non-invasive or very mild; grade 1, mild; grade 2, moderate; grade 3, severe; grade 4, very serious. The sum of the scores for each item was considered as the total of lung injury score. Arterial blood samples were collected and the PaO2 was measured using an automated blood gas analyzer (GEMPREMIRE3000, MA, USA).

Cell counting and measurement of albumin and IgM
BAL was conducted by injecting 0.9% sodium chloride containing 0.5 mL of EDTA (0.6 mg/L) into the lung of the mice. Bronchoalveolar lavage fluid (BALF) was retrieved via aspiration and then placed on ice. The total number of BALF cells was counted using a blood cell analyzer. The neutrophil count was calculated by multiplying the percentage of neutrophils in BALF and the number of total cells. BALF was retrieved and centrifuged at 800 g for 6 minutes. The levels of albumin and IgM were determined using a mouse specific albumin ELISA kit (ALPCO Diagnostic, Salem, NH, USA) and a mouse specific IgM ELISA kit purchased from Bethyl Laboratory (Montgomery, TX, USA), respectively.

Statistical analysis
Statistical analyses were conducted using the SPSS 21.0 statistical software (IBM Corp. Armonk, NY, USA). Measurement data obeying normal distribution were summarized as mean ± standard deviation. Comparisons between two groups were conducted using unpaired t-test and those among multiple groups were conducted by one-way analysis of variance (ANOVA), followed by Tukey's post hoc tests with corrections for multiple comparisons. A value of p < 0.05 indicated the difference was statistically significant.

Ethics statement
The current study was conducted after approval by the ethics committee of The First Affiliated Hospital of Zhengzhou University (NO. 201611042). Signed written informed consent was obtained from all participants or their guardians prior to sample collection. Animal experiments were conducted according to the international convention on laboratory animal ethics and relevant national regulations (NO. 201812009), and all efforts were made to minimize the suffering of the included animals.

AUTHOR CONTRIBUTIONS
Meng-Ying Yao, Wei-Hong Zhang, Wen-Tao Ma, Qiu-Hong Liu, Li-Hua Xing and Gao-Feng Zhao designed the study. Meng-Ying Yao and Wei-Hong Zhang collated the data, carried out data analyses and produced the initial draft of the manuscript. Wen-Tao Ma, Qiu-Hong Liu, Li-Hua Xing and Gao-Feng Zhao contributed to drafting the manuscript. All authors have read and approved the final submitted manuscript.