Circulating miRNA Proles Associated with Lung Function Decline

Background: Lung function decline and miRNAs are known to involve in the pathogenesis process of respiratory diseases. However, the association between miRNA proles and lung function decline was rarely elucidated. This study aimed to compare the expression pattern of plasma miRNAs among people with rapid lung function decline and healthy controls. Methods: Ten controls and 10 cases with rapid lung function decline in last 3 years were enrolled from the Wuhan-Zhuhai cohort. The miRNA sequencing and bioinformatics analysis were carried out to explore miRNAs expression and possible signal pathways. Pearson correlation analysis was carried out to assess the associations between lung function and miRNAs expression levels. Results: A total of 1209 detected miRNAs were differently expression in the reaseach participants. Among them, 17 miRNAs (miR-6749-5p, miR-6797-3p, miR-4468, miR-4301, miR-629-3p, miR-4713-3p, miR-486-3p, miR-450a-1-3p, miR-4732-5p, miR-514a-5p, miR-193b-5p, miR-6749-5p, miR-3168, miR-4691-5p, miR-6730-5p, miR-184, miR-486-5p) were signicantly down-regulated in cases with reduced lung function. The expression levels of these miRNAs, except miR-514a-5p and miR-3168, were signicantly correlated with lung function parameters. Through bioinformatics analysis, the Wnt signaling pathway, FoxO signaling pathway, MAPK signaling pathway, Pathway in cancer, inammatory mediator regulation of TRP channels, and TGF-β signaling pathway maybe involved in mediating the association between 17 identied miRNAs and lung function decline. Conclusions: The plasma miRNA prole of people with rapid lung function decline are different when compared with healthy controls. These miRNAs might become research candidates or biomarkers in the early progression of respiratory diseases.


Introduction
Respiratory diseases with poor outcomes of lower respiratory tract infections, chronic obstructive pulmonary disease (COPD), asthma, and lung cancer resulted in a large health burden and consistently rank the second leading cause of death worldwide 1,2 . Pulmonary function test is an objective and effective method to assess the healthy status of respiratory system 3 . Little lung function decline is a normal feature of aging, but excessive decline during adulthood may lead to persistent air ow obstruction and is a common predictor for the development of respiratory diseases, including COPD, asthma, and cystic brosis [4][5][6] . Previous published study already found that rapid forced expiratory volume in 1s (FEV 1 ) decline was independently associated with both COPD-related hospitalizations and all-cause deaths 7 . The risk of rapid lung function decline for COPD-related hospitalizations was highest among people with normal FEV 1 / forced vital capacity (FVC) (≥ 70%), which indicated that the pathogenesis of COPD may begin much earlier as people developed some pulmonary abnormalities, like, FEV 1 declines more than 60 mL per year or FEV 1 /FVC less than normal level 7,8 . Unfortunately, the underlying mechanisms for lung function decline in the pathological process of respiratory diseases remain to be fully elucidated.
MicroRNA (miRNA) is a kind of small non-coding RNAs, which plays an important role in regulating gene expression by forming complementary base pairs within the 3' untranslated regions of targeted mRNAs, participating in cell proliferation, differentiation, survival, and apoptosis 9 . Considering the fact that one miRNA could regulate up to hundreds of target genes, growing evidences demonstrated that miRNAs were involved in diverse biologic pathways, gene-disease interactions, and some diseases including respiratory diseases [10][11][12] . In 2018, Conickx and colleagues revealed that miR-218-5p was signi cantly down-regulated in COPD patients and strongly correlated with lung function decline 13 . Besides, Zhang and colleagues found that overexpressed miR-486-5p might enhance the Toll-like receptor 4 triggered in ammatory responses in COPD patients by targeting the HAT1 gene 14 . After Pro ling the expression of miRNAs in 10 idiopathic pulmonary brosis (IPF) patients, Kusum and colleagues reported that let-7d was signi cantly reduced in IPF patients and positively correlated with pulmonary functions 15 . However, most of these studies investigated the association between miRNA pro les and lung function in serious patients, but not in early stage of respiratory diseases or normal subjects with slight injury.
Since plasma microRNAs were proved to be a kind of noninvasive biomarker re ecting pulmonary health 16 , we conducted the present study to compare the expression pattern of plasma miRNAs among people with rapid lung function decline and healthy controls. Pulmonary parameter FEV 1 /FVC% under 80% and FEV 1 declined more than 180 ml in last three years were regarded as rapid lung function decline. The plasma miRNA sequencing and bioinformatics analysis were carried out to explore miRNAs expression and possible signal pathways.

Study population
The study participants were derived from the Wuhan-Zhuhai cohort, which has been described in detail elsewhere 17 . Brie y, a total of 4812 Wuhan or Zhuhai residents were recruited from 2011 to 2012, with follow-up every three years. The participants who completed the investigation in both 2014 and 2017 and never smoking were considered for this study. Cases with rapid lung function decline in this study were de ned as: FEV 1 /FVC% in 2014 was greater than 80% and FEV 1 /FVC% in 2017 was less than 80%, and FEV 1 decreased by more than 180 mL. The control group is de ned as those whose FEV 1 /FVC% exceeded 80% in both 2014 and 2017, and the FEV1 decreased less than 180 mL after matching age, gender, and BMI (body mass index). As a result, a total of 10 pairs of case-control were randomly enrolled in this study. A standardized questionnaire was used to collect demographic characteristics, including age, gender, passive smoking status, alcohol consumption, physical activity, and family income by trained investigators. Anthropometric data including weight, height, and physical examination were measured by specialists. All participants in this study gave written, informed consent and the study was approved by the Ethics and Human Subject Committee of Tongji Medical College, Huazhong University of Science and Technology.

Pulmonary function tests
Lung function test was conducted by a specialist using the electronic spirometer (Chestgraph HI-101, CHEST Ltd., Japan) according to the American Thoracic Society as described previously 18 . Three acceptable volume-time curves of lung function parameters, including FVC, FEV 1 , maximal mid-expiratory ow, and peak expiratory ow were obtained.

Sample collection and RNA isolation
Approximately 5 mL fasting venous blood was collected into an EDTA tube for each person according to the standard method. Blood samples were centrifuged at 1500x g for 10 minutes at room temperature to spin down the blood cells. The supernatant plasma was isolated and stored at -80°C until analysis.
Total RNAs were extracted and puri ed from 500 µL plasma samples using the miRNeasy Plasma Kit (Qiagen, Hilden, Germany), as described in the manufacturer's protocol. QIAseq miRNA Library QC spikeins (Qiagen, Hilden, Germany) were added during the process of RNA extraction to provide external controls for the quality of the RNA isolation, library preparation, and sequencing analysis.

miRNA pro ling and qPCR quality
The integrity and concentration of puri ed RNA were analyzed with an Agilent Bioanalyzer 2100 before library preparation. Then acceptable RNAs were used as a template for the QIAseq miRNA library kit (Qiagen, Hilden, Germany), which ligated speci c adapters to the 3' and 5' sides of miRNA, reversely transcripted miRNAs to cDNA, and processed library ampli cation. High-sensitivity DNA chips and Illumina HiSeq 2500 instrument (Illumina, CA, USA) were conducted to assess the quality of the libraries and generate sequencing reads, respectively. The results of sequencing reads were validated by RT-PCR according to the QIAseq miRNA Library QC PCR panel kit (Qiagen, Hilden, Germany). U6 RNA was regarded as the reference gene to normalize the expression of ve plasma miRNAs in RT-PCR and RNA sequencing. All procedures for next-generation sequencing (NGS) were performed by CapitalBio Technology (Beijing, China).

Bioinformatics analysis
Three miRNA target databases, including Miranda (http://mirdb.org/miRDB/), TargetScan (http://www.targetscan.org/), and miRPathDB (http://mpd.bioinf.uni-sb.de/) were used to predict the target genes of miRNAs associated with lung function decline. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway was carried out to analyze those target genes which were associated with at least two miRNAs in three databases.

Statistical analysis
Continuous variables were shown as the mean ± standard deviation and categorical variables were expressed as a number and percentage. Chi-square test and Student's t-test were respectively carried out to compare categorical and continuous characteristics between cases and controls. The row data of each miRNA sequencing were rst processed and mapped with human genome version hg19 using bowtie software. Only miRNAs with expression levels of > 1 counts per million (CPM) in more than half of the samples were considered detectable miRNAs and then normalized by "EdgeR" package to identify differentially expressed miRNAs between two groups. Pearson correlation analysis was carried out to assess the correlation between log-transformed CPM by sequencing and CT value by qPCR, as well as associations between lung function and miRNAs expression levels. A false discovery rate (FDR) adjusted P-value of < 0.10 was considered statistically signi cant. All statistical analyses were conducted using SAS 9.4 (SAS Institute, NC, USA) and R version 3.4.3 (R core team, 2016).

Basic characteristics
The basic characteristics and lung function of 20 individuals was presented in Table 1. In this study, the most of the participants were female, accounting for 90% of the objects, and the mean age of the cases and controls were 58.0 and 57.8 years old respectively. There were no differences noted with among demographic characteristics, including age, height, weight, BMI, physical activity, passive smoking, drinking, and income between case and control groups. In 2014, the lung function levels of the cases (FEV 1 was 2.65 ± 0.43 L, FVC was 2.28 ± 0.35 L, FEV 1 /FVC% was 85.3 ± 4.5%) and controls (2.69 ± 0.51 L, 2.25 ± 0.43 L, 83.7 ± 2.5 %) were at the same level (all P > 0.05). While in 2017, FEV 1 and FEV 1 /FVC% for cases (1.91 ± 0.29 L, 74.2 ± 2.4 %) were signi cantly lower than those in controls (2.38 ± 0.43 L, 86.1 ± 3.7 %) (P < 0.05). The reduced FEV 1 and FEV 1 /FVC% of the case group (0.37 ± 0.21 L, 11.9 ± 6.7 %) were signi cantly higher than that of the control group (-0.13 ± 0.08 L, -2.4 ± 4.4 %) (P < 0.05).

miRNAs expression pro les
After processing the data of RNA sequencing, the composition of detected small RNAs (sRNA) was classi ed into different categories, including miRNA, piRNA, rRNA, tRNA, mRNA, novel RNA, and others ( Fig. 1). The proportions of each kind of sRNA were similar in case and control groups, in which miRNAs made up more than half of the sRNA (68.9% in the case group and 57.4% in the control group). mRNA accounted for the least in both case and control groups, while tRNA was lower in cases than in controls but not statistically different. In general, all kinds of sRNAs were not signi cantly different between the two groups.
A total of 1209 miRNAs were e ciently detected in plasma from the research participants. Compared with the expression levels in healthy controls, 8 up-regulated and 81 down-regulated miRNAs were identi ed to differently expressing in the case group with an absolute log 2 (Fold change) > 1.069 (1 unit of standard deviation) and P-value < 0.05 (Fig. 2). Notably, there were 17 miRNAs with FDR lower than 0.1 that were considered as signi cantly differential expressed miRNAs, as shown in Table 2, all of which had lower expression in the cases than in the controls. Additional 24 miRNAs showed an 0.10 ≤ FDR < 0.2 in Table S1, including 23 down-regulated and 1 up-regulated miRNA. Besides, 5 plasma miRNAs, including miR-23a-3p, miR-30c-5p, miR-103a-3p, miR-191-5p, and miR-451a, were examined to validate the sequencing results by quantitative RT-PCR ( Figure S1). Results demonstrated that the expression levels of miRNAs using miRNA pro les were highly consistent with those derived from RT-PCR, with correlation coe cients ranged from 0.88 to 0.97.

Enrichment biological function analysis
To explore the potential biological mechanisms underlying the role of these 17 differential expressed miRNAs in lung function decline, we performed targeted gene prediction and pathway analysis. A total of 5351 unique genes were predicted to have interaction with these miRNAs in three databases. After the targeted genes were ltered for high-con dence predictions and predicted frequency, 1379 target genes were involved in the following KEGG signaling pathway analysis. As presented in Fig. 3, those target genes were mainly enriched in 6 potential pathways, including the Wnt signaling pathway, FoxO signaling pathway, MAPK signaling pathway, Pathway in cancer, in ammatory mediator regulation of TRP channels, and TGF-β signaling pathway.

Discussion
In the present study, we conducted RNA sequencing to explore the pro le of plasma miRNAs in people with rapid lung function decline. Using the NGS miRNA platform, a total of 1209 miRNAs was identi ed and 17 of them were signi cantly down-regulated in cases with lung function decline when compared with controls. In addition, the expression levels of these 17 miRNAs, except miR-514a-5p and miR-3168, were found to be signi cantly correlated with lung function parameters. Bioinformatics analyses predicted that these 17 miRNAs were mainly enriched in 6 associated pathways, including the Wnt signaling pathway, FoxO signaling pathway, MAPK signaling pathway, Pathway in cancer, in ammatory mediator regulation of TRP channels, and TGF-β signaling pathway.
To our knowledge, this is the rst study to pro le the expression pattern of plasma miRNAs associated with reduced lung function. Lung function, a vital indicator for the evaluation of respiratory health, is widely used in the clinical diagnosis because of its simplicity and reproducibility 19 . Recent studies revealed that dysregulated miRNAs were involved in the pathological process of respiratory disesases due to the role of multiple gene regulation in apoptosis, autophagy, cellular differentiation, endothelial cell proliferation, and airway in ammation 20,21 . Most of these studies focused on patients who already suffered from respiratory diseases, but not people with normal lung function at baseline [22][23][24][25] . Therefore, it is necessary to evaluate the association between different miRNAs expression pro les and lung function decline in early stage. Accordingly, we found 17 miRNAs signi cantly down-regulated in people with rapid lung function decline. In this study, we selected cases with rapid lung function decline through a cohort that was followed for 3 years. Both cases and controls had lung function paramaters before and after 3 years. At the same time, in order to avoid direct interference of smoking on miRNA, the study participants excluded cigarette smokers. Therefore, our results could better observe the speci c miRNAs related to lung function decline.
Of 17 differential expressed miRNAs identi ed in this study, miR-629-3p was previously reported to be signi cantly down-regulated in plasma from COPD patients with a fold change of 0.38 when compared with healthy controls 26  miR-486 is a highly conserved miRNA in mammals, which is located at Chr:8p11 and processed into two mature miRNAs: miR-486-5p and miR-486-3p. A previous study by Ji and colleagues reported that decreased miR-486-5p expression was linked to the progression of pulmonary brosis in both humans and mice 33 . Over-expression of miR-486-5p could attenuate pulmonary brosis and repress TGF-β1induced brogenesis, which was proved in non-small cell lung cancer and renal cell carcinoma 34,35 .
These ndings demonstrated that miR-486-5p has a strong anti-brotic activity in lung tissues. Interestingly, the protective role of miR-486-5p was also found in human lung alveolar epithelial cells exposed to PM2.5 36 . miR-486-5p may play role through PTEN and FOXO1 pathway which mediated the protective effects by reducing cell apoptosis and reactive oxygen species generation. Similarly, the study by Chai and colleagues concluded that miRNA-486-5p improved nucleus pulposus cell viability, and inhibited in ammation cytokines and ECM degradation partly via inhibition of FOXO1 expression 37 . Thus, down-regulation of miR-486 may be one of the regulators which mediated the decline of lung function and the detailed mechanism needs more research in the future.
miR-184 was also reported to play a protective role in lung tissue 38 . Using microRNA expression array among 25 healthy and IPF patients, miR-184 expression levels was down-regulated in IPF patient tissues. Further in vitro study investigated that miR-184 could bind to the smad2-and akt-3'UTRs and repress the downstream signaling pathways triggered by TGF-β 38 . However, Cristina and colleagues reported contradictory results that miR-184 acted a downstream effector of albuminuria through LPP3 to promote tubulointerstitial brosis 39 . This difference may be attributed to the type 2 diabetes rat model used by Christina, who studied the kidneys in the late stage of the disease. Further research is needed to investigate whether brosis is involved in the association between miR-184 and lung function decline.
The other miRNAs we found in this study have not previously been reported and were novel objects to study the association with respiratory diseases in the future.
To gain insight into the potential functional importance of these 17 miRNAs, we analyzed the signaling pathways with three databases. Most of the associated pathways have been revealed to be involved in the pathological process of respiratory diseases. The developmental WNT pathway is fundamental for lung development, and altered WNT activity has been reported to contribute to lung epithelium injury repair 40  reported to participate in in ammatory response by aggravating oxidative stress in lung tissue 43 . The in ammatory mediator regulation refers to PGE2, bradykinin, ATP, and pro-in ammatory cytokines that are generated during acute lung tissue injury 44 . As previously mentioned, TGF-β is crucial in pulmonary brosis and emphysema via mediating broblast activation, ECM synthesis, and immune response 45 . Further researches are needed to reveal whether these associated signaling pathways were underlying the pathogenesis of lung function decline and the early stage of respiratory diseases including COPD, asthma and pulmonary brosis.
Our study have some strengths. Firstly, this study was based on a cohort study and participants enrolled were follow-up in three years. Then, we used NGS technology in miRNAs pro ling to provide an open and objective result. However, some limitations of our study need to be considered. First, the number of participants enrolled in this study is relatively small, and further studies with a large population are needed to validate our ndings. Second, the expressing levels of miRNAs we identi ed were not all proved by RT-PCR. To validate the sequencing results, we measured the expression of 5 endogenous miRNAs by RT-PCR and found a high correlation (R 2 ≥ 0.88) with sequencing results. Third, it's hard to determine the true origin of miRNAs we identi ed in plasma. But according to The Human miRNA Tissue Atlas (website https://ccb-web.cs.uni-saarland.de/tissueatlas/) 46 , most miRNAs we identi ed were expressed in lung tissue.

Conclusion
Our results suggest that several plasma miRNAs associated with lung function decline, which might become research candidates or biomarkers in the early progression of various respiratory diseases. Further studies are encouraged to validate our ndings in more robust studies. 6. Abbreviations

Declarations
Ethics approval and consent to participate The study was approved by the Ethics and Human Subject committee of Tongji Medical College, and the approval ID was [2016] IEC S128. Written informed consent was obtained from all participants.

Consent for publication
No applicable.

Availability of data and material
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.  The volcano plot for the 1209 miRNAs up-and down-regulated in cases with lung function decline compared with controls. The X-axis is log2 fold-change and the Y-axis is log10 P-value.