A total of 40 patients (21 male patients and 19 female patients; age range, 20–80 years; mean age, 66.5±5.8 years) admitted to the Department of Respiratory and Critical Care Medicine of the Third Affiliated Hospital of Inner Mongolia Medical University (Baotou, China) between January 1st 2015 and October 31st 2015 were enrolled in the current study. The inclusion criteria were as follows: Patients with sepsis were diagnosed based on the Sepsis 3.0 guidelines, which can be simplified as follows: Sepsis = infection + sequential organ failure assessment (SOFA) ≥2 (17). The exclusion criteria were as follows: i) Patients <18 or >80 years of age; ii) patients with liver and kidney failure; iii) patients that were pregnant or lactating; iv) patients with tumors and hematological diseases; v) an agranulocytosis score of <0.5×109/l; vi) patients with human immunodeficiency virus-related complications; and vii) patients administered glucocorticoids up to 4 weeks before enrollment or that were receiving immunotherapy after organ transplantation. A total of 40 healthy individuals that donated samples following routine physical examinations were used as a control group (20 males, 20 females; aged 20–80 years). A further 80 samples (including 40 samples from patients with sepsis and 40 samples from healthy controls, whose details are listed in Table I) were obtained from individuals admitted to the same hospital that were enrolled to the present study between April 10th 2020 and January 31st 2021. The inclusion criteria were the same as aforementioned. There was no significant difference in the mean values of sex and age between the test group and the control group. All protocols were approved by the Local Ethics Committee of the Third Affiliated Hospital of Inner Mongolia Medical University (Baotou, China), and written informed consent was obtained from all subjects who were permitted to withdraw from clinical observation at any time for any reason.

To compare the DEMs in patients with sepsis and healthy controls, 2 ml whole blood samples were collected into sterilized Eppendorf tubes. Total RNA was extracted using the mirVana miRNA Isolation kit (Ambion; Thermo Fisher Scientific, Inc.) in accordance with the manufacturer's protocol. Subsequently, the concentration and integrity of total RNA was determined using a spectrophotometer (NanoDrop™; Thermo Fisher Scientific, Inc.) and agarose gel electrophoresis. The conditions required for microarray analysis included: An average A²⁶⁰/²⁸⁰ ratio of total RNA ≥1.8, RNA content ≥1 µg/ml and clear 28S and 18S electrophoresis bands of total RNA (18).

Microarray analysis was performed by Sangon Biotech Co., Ltd. Briefly, total RNA (~200 ng) extracted from the aforementioned blood samples was subjected to dephosphorylation and labeling using the miRNA Complete Labeling and Hyb kit (Agilent Technologies, Inc.), where Cyanine3-pCp was connected to the 3′ end of RNA using T4 RNA ligase. The labeled reaction products were concentrated using a vacuum concentrator for 3 h at 45–55°C and subsequently underwent overnight hybridization using Agilent SureSelect Capture Technology (Agilent Technologies, Inc.). After washing with saline-sodium citrate at room temperature, slides were dried by air and scanned using the Agilent Scanner G2505C (Agilent Technologies, Inc.), after which the Agilent Feature Extraction (v10.7; Agilent technologies, Inc.) was utilized to extract data and analyze the hybridization results. Agilent GeneSpring (version 12.5; Agilent Technologies, Inc.) was used for quartile data normalization and to determine differences between groups. A Q-value of ≥5% and a fold change (FC) value of >2.0 or <0.5 were selected as cut-off points for DEMs. Statistical significance was determined using an unpaired t-test and was represented using a P-value. To reduce the risk of false positives, values were adjusted for multiple testing using the Benjamini-Hochberg False Discovery Rate (FDR) method. The corrected value was represented by FDR (19). FDR <0.05 was selected as the cut-off value for DEM screening. The possible miRNAs that target sepsis were predicted using TargetScan software (http://www.targetscan.org/vert_80/).

According to microarray assay, 305 miRNAs with >2-fold expression changes (patients with sepsis vs. healthy people) were detected. Subsequently, the top 18 up- and downregulated miRNAs were selected for microarray analysis verification via RT-qPCR. The primers used for RT-qPCR were synthesized by Sangon Biotech Co., Ltd. and are listed in Table II. Total RNA was extracted using RNAiso Plus (cat. no. 9108; Takara Bio, Inc.), and reverse transcribed into cDNA using miRNA First Strand cDNA Synthesis (Stem-loop Method) (cat. no. B532453; Sangon Biotech Co., Ltd.) according to the manufacturers' instructions. RT-qPCR was performed using SYBR Premix Ex Taq™ II (Takara Bio, Inc.) with U6 as an internal control for miRNA detection. The thermocycling conditions were as follows: Pre-denaturation at 95°C for 10 min, followed by 40 cycles of denaturing at 95°C for 5 sec, annealing at 60°C for 20 sec and elongation at 70°C for 10 sec. The melt curve conditions were as follows: 60°C to 95°C in increments of 0.3°C; this was used to determine the melting temperature of the detected miRNAs and primer dimers. The relative expression of DEMs was calculated using the comparative 2^−ΔΔcq method (20) and the data were analyzed using SPSS v19.0 software (IBM Corp.).

It has been reported that TDAG8 is involved in the maintenance of lysosomal function, particularly during pathogen defense (21). In addition, TLR4 can be activated by lipopolysaccharide (LPS) to induce the production of proinflammatory mediators to eradicate bacteria or other pathogens (22). Dysregulation of the host response to LPS can lead to sepsis, which is a systemic inflammatory condition (23). The current study therefore detected the expression levels of TDAG8 and TLR4 mRNA in the peripheral blood of patients with sepsis. Total RNA was extracted using RNAiso Plus, and reverse transcribed into cDNA using PrimeScript™ RT reagent Kit with gDNA Eraser (Perfect Real Time) (cat. no. R047A; Takara Bio, Inc.) according to the manufacturer instructions. The primers used for this reaction are listed in Table II. Amplification was performed using the SYBR Premix Ex Taq™ II under the following thermocycling conditions: Initial denaturation at 95°C for 30 sec; followed by 50 cycles of denaturation at 95°C for 5 sec, annealing at 60°C for 30 sec and extension at 72°C for 30 sec. The expression of TDAG8 and TLR4 mRNA was calculated using the 2^−ΔΔCq method and GAPDH was used as an internal control.

The current results revealed that the expression levels of miR-3663-3p and miR-6881-3p in patients with sepsis were significantly increased compared with healthy controls. In addition, our pre-experiment determined that TDAG8 was closely related to the regulation of the inflammatory response in patients with sepsis. That is, in the sepsis group, the correlation coefficients (r) of TDAG8 mRNA expression and IL-6 or CXCL8 concentration were 0.8455 or 0.7117, respectively, indicating that TDAG8 mRNA expression was positively correlated with IL-6 and CXCL8 levels. In the present study, TargetScan software (www.targetscan.org/vert_72/; TargetScanHuman 7.2) was applied to predict whether TDAG8 was the biological target of miR-3663 and miR-6881. After selecting ‘human’ as the species, the human gene TDAG8 (GPR65) was entered. The database was subsequently searched for miR-3663-3p and miR-6881-3p.

Sepsis often leads to multiple organ failure as a result of an uncontrolled inflammatory response (24). In the present study, ELISA was performed to detect the concentrations of various proinflammatory cytokines, including IL-6 (cat. no. CSB-E04638h; Cusabio Technology LLC), IL-21 (cat. no. CSB-E11707h; Cusabio Technology LLC), C-X-C motif chemokine ligand-8 (CXCL8) (cat. no. CSB-E04641h; Cusabio Technology LLC) and monocyte chemoattractant protein-1 (MCP-1) (cat. no. CSB-E04655h; Cusabio Technology LLC) in the peripheral blood of patients with sepsis. All procedures were performed in accordance with the respective protocols of commercially available ELISA kits (R&D Systems, Inc.). Each sample was measured three times and an average value was calculated.

Data were obtained from three independent experiments and were expressed as the mean ± SD. All statistical analyses were performed using SPSS v19.0 software. The statistical differences between two groups were detected using unpaired Student's t-test. The correlation between inflammatory response and miRNA-3663-3p or TDAG8/TLR4 mRNA expression was determined using Pearson's correlation analysis. P<0.05 was considered to indicate a statistically significant difference.

There were 21 males and 19 females in the sepsis group, aged between 20–80 years, with a mean age of 66.5±5.8 years. A total of 40 healthy individuals that donated samples following routine physical examinations were selected as the control group (20 males; 20 females; age range, 20–80 years; mean age, 67.2±4.2 years). There were no significant differences between the mean values of sex and age between the sepsis and healthy control groups (Table I).

The integrity of total RNA was assessed using agarose gel electrophoresis and subsequent 18S and 28S RNA band staining for visualization. As presented in Fig. 1, the electrophoretic bands of 28S and 18S RNA were clear. In addition, the ratio of A260 to A280 in the extracted RNA was 1.8-2.0. Therefore, total RNA samples with high purity and good integrity were used for microarray analysis.

To clarify the potential function of DEMs, microarray analysis was performed by Sangon Biotech Co., Ltd. Agilent GeneSpring was used for the analysis of DEM expression between patients with sepsis and healthy controls in order to identify characteristic DEMs. The significance of microarray analysis was denoted by FDR (FDR <0.05). As presented in Fig. 2A, the results revealed 305 miRNAs that demonstrated a >2-fold change in expression in patients with sepsis compared with in healthy individuals, including 212 upregulated and 93 downregulated DEMs. Among these, the top 18 up- or downregulated miRNAs, which exhibited significant differential expression (P<0.05; FC >2.0), were selected for further analysis. A total of 9 miRNAs demonstrated an increased expression, while 9 miRNAs exhibited a decreased expression (Table III). Hierarchical clustering of the top 18 DEMs was subsequently performed, the results of which are presented in Fig. 2B.

To validate microarray analysis data, DEM expression levels were detected by RT-qPCR. As presented in Fig. 2C, the results of RT-qPCR were almost consistent with those obtained through microarray analysis. Similar trends in the following miRNAs were detected: miR-625-5p, miR-6786-5p, miR-17-3p, miR-501-3p, miR-6756-5p, miR-501-5p, miR-6786-5p, miR-3663-3p, miR-4694-5p, miR-6881-3p, miR-592, miR-491-5p, miR-1260b, miR-3591-3p, miR-6514-3p and miR-491-3p. However, the expression levels of miR-892b and miR-362-5p differed to that of microarray analysis; therefore, a larger sample size for RT-qPCR analysis is required to validate these results. Integrated analysis of miRNA and mRNA expression profiles indicated that among the highly expressed miRNAs, the expression levels of miR-3663-3p, miR-6881-3p, miR-625-5p, miR-3591-3p and miR-6514-3p were significantly different compared with the control group [FC (abs) >10]. Notably, the expression levels of miR-3663-3p were significantly increased in the peripheral blood of patients with sepsis according to RT-qPCR. TargetScan software was used to evaluate the putative target genes of miR-3663-3p. The results revealed that TDAG8 was predicated to be a potential target gene of miR-3663-3p (data not shown), as evolutionary conservation was demonstrated. As the TLR4 signaling pathway has been reported to serve a role in sepsis (25,26) and TDAG8 has been reported to be involved in regulation of cell functions associated with airway inflammation (27), TDAG8 and TLR4 were selected for further experimentation.

To further clarify whether TDAG8 and TLR4 are involved in the occurrence and development of sepsis, the mRNA expression levels of TDAG8 and TLR4 were determined in human blood samples via RT-qPCR. As presented in Fig. 3, significant differences were demonstrated in both TDAG8 and TLR4 mRNA levels between healthy controls and patients with sepsis. More specifically, the expression levels of serum TDAG8 (1.0±0.02 vs. 2.7±0.13) and TLR4 (1.0±0.08 vs. 1.5±0.10) mRNA in patients with sepsis were significantly higher compared with healthy controls. The results indicated that TDAG8 and TLR4 mRNA expression is upregulated in patients with sepsis.

Given the consideration that high expression levels of TLR4 can result in the production of chemokines and proinflammatory cytokines (28), serum IL-6, IL-21, CXCL8 and MCP-1 levels were determined in the current study by performing ELISA. As presented in Fig. 4, serum IL-6, IL-21, CXCL8 and MCP-1 levels in the healthy control group were 2.12±0.18, 48.50±7.9, 0.15±0.04 and 8.25±1.45 ng/l, while levels in patients with sepsis were 6.24±0.92, 137.0±29.20, 0.28±0.05 and 30.40±5.30 ng/l, respectively. The results demonstrated that proinflammatory cytokines IL-6 and IL-21 and CXCL8 and MCP-1 chemokines were significantly increased in patients with sepsis (t=6.89, 6.07, 11.8 and 9.03, respectively), suggesting an increased inflammatory response.

To clarify the relationship between miR-3663-3p and the secretion of proinflammatory cytokines, linear correlation analysis was performed. The results revealed that, in patients with sepsis, the correlation coefficients (r) of miR-3663-3p expression and IL-6, IL-21, CXCL8 and MCP-1 levels were 0.8352, 0.8976, 0.6633 and 0.7661, respectively. Furthermore, the correlation coefficients (r) of the same miRNA with TDAG8 and TLR4 mRNA expression levels were 0.7895 and 0.8622, respectively (Fig. 5A). The results indicated a positive correlation between miRNA-3663-3p and the inflammatory response. As the secretion of inflammatory factors is closely associated with the activation of TLR4 (29), correlation analysis was performed between the expression levels of TDAG8/TLR4 mRNA and the production of IL-6, IL-21, CXCL8 and MCP-1. As presented in Fig. 5B and C, the correlation coefficients (r) of TDAG8 mRNA with the contents of IL-6, IL-21, CXCL8 and MCP-1 were 0.856, 0.914, 0.515 and 0.902, respectively. Additionally, the correlation coefficients (r) of TLR4 mRNA with the contents of IL-6, CXCL8, IL-21 and MCP-1 were 0.940, 0.946, 0.715 and 0.890, respectively. The data indicated a positive correlation between TDAG8/TLR4 mRNA expression and proinflammatory cytokine/chemokine secretion. TDAG8 mRNA expression also revealed a positive correlation with TLR4 mRNA expression, with a correlation coefficient value (r) of 0.878).

In the current study, miR-3663-3p and the expression of various cytokines/chemokines (IL-6, IL-21, CXCL8 and MCP-1) were selected for sensitivity and specificity analysis using a receiver operating characteristic (ROC) curve. ROC curves revealed that miR-3663-3p, IL-6, IL-21, CXCL8 and MCP-1 levels, along with TDAG8 and TLR4 mRNA had area under the curve values of 0.908, 0.912, 0.959, 0.944, 0.996, 0.952 and 0.815, respectively, when distinguishing patients with sepsis from healthy controls (Fig. 5D). When selecting a cut-off value of 0.02 for miR-3663-3p, as determined via ROC curve analysis, the diagnostic sensitivity and specificity were determined to be 100%. The cut-off points, diagnostic sensitivities and diagnostic specificities of IL-6, IL-21, CXCL8 and MCP-1 are listed in Table IV. The area under the ROC curve of IL-6, IL-21, CXCL8 and MCP-1 were >0.9, indicating that they may be used to provide an earlier warning of sepsis when combined with miR-3663-3p. To validate these findings, an additional 80 samples (40 samples for healthy controls and 40 samples for patients with sepsis) were obtained for RT-qRCR analysis. The results revealed that the expression levels of miR-3663-3p were significantly increased in patients with sepsis compared with healthy controls (P<0.01; Fig. 5E). The data indicated that miR-3663-3p could serve as a potentially powerful diagnostic and predictive biomarker for sepsis, and that miRNA analysis combined with the measurement of inflammatory cytokine secretion may be a reliable approach for the fast diagnosis and early identification of sepsis.