A total of 60 neonates were recruited in Shengli Oilfield Central Hospital from August 2023 to August 2024, including 30 neonates diagnosed with sepsis and 30 healthy neonates as controls. This study was approved by the Research Ethics Committee of Shengli Oilfield Central Hospital (0023427), and written informed consent was obtained from all neonatal guardians. The diagnosis of NS is mainly based on the clinical manifestations, medical history and laboratory examination results of neonates (21). Clinical manifestations meeting three or more of the following symptoms can be determined: (1) hyperthermia or hypothermia; (2) cardiac dysfunction, including tachycardia, brady cardia, and blood pressure reduction; (3) neurological disorders, occurrence of drowsiness, convulsions, epilepsy; (4) apnea, tachypnea, cyanosis; (5) gastrointestinal disorders, such as diarrhea and abdominal distension. Thirty healthy newborns who were born in this hospital during the same period were selected as the control group. Exclusion criteria: (1) premature neonates; (2) congenital malformation or chromosomal abnormality. Blood samples from NS patients and the control group were collected before systematic treatment. Blood samples were centrifuged at 1,200 g for 10 min, and the supernatant was removed and stored at −80 ℃.

Human lung epithelial cell line A549 (from the cell Bank of Shanghai Institute of Biochemistry and Cell Biology) was cultured in DMEM/F12 medium containing 10% fetal bovine serum (FBS, AusGeneX, Australia) and 1% double antibody (PWL062, Meilun, China), and cultured in 37 ℃, 5% CO₂ incubator.

LPS (ST1470, Biyuntian, China) 10 mg was prepared into 1 mg/ml mother liquor with sterile water and stored at 4 ℃. When used, the medium was diluted to 10 μg/ml working solution. To investigate the functional role of EMX2OS in ALI, A549 cells were exposed to 10 μg/ml LPS for 24 h.

In this study, the small interfering RNA against EMX2OS (si-EMX2OS) and negative controls (scramble) were provided by Santa Cruz Biotechnology, and cell transfection was performed using a Lipofectamine 3000 transfection reagent (Thermo Fisher Scientific). Cells were inoculated into 6-well plates, and cultured to 60% density, and the above vectors were mixed with transfection reagent Lipofectamine 3000 and dropped into the cells. After 6 h, cells were collected for subsequent experiments.

RNA samples were extracted from A549 cells and blood using the AG RNAex Pro kit (AG21102, Accurate, China) in accordance with the manufacturer's protocol. The concentration and purity of the RNA were determined using the NanoDrop™ 2000 spectrophotometer (Thermo Scientific™, USA). The OD260/280 ratios for all samples fell within the range of 1.9–2.1. One microgram of total RNA was reverse transcribed into cDNA using the Evo-M-MLV RT Kit with gDNA Clean for qPCR (AG11705, Accurate, China) following the manufacturer's protocol. The RT-PCR was performed using the SYBR Green qPCR Master Mix (Thermo Fisher Scientific, Shanghai, China) by the ABI PRISM 7300 real-time PCR system (Applied Biosystems). The endogenous controls of EMX2OS and AKT3 were GAPDH, and the endogenous controls of miR-654-3p was U6. The thermal cycle was set as follows: 95 ℃ for 1 min and 40 cycles at 95 ℃ for 15 s, 58 ℃ for 20 s and 72 ℃ for 20 s. The final expression values were calculated by the 2^−△△Ct method for subsequent analysis.

CCK-8 assay was used to detect the viability of A549 cells in different treatment groups. At different time points after cell transfection, 10 ul CCK-8 reagent (TransGen Biotech, Beijing, China) was added to each well of the 96-well plate, and after continued culture for 2 h, the absorbance value at 450 nm was measured using a microplate reader to calculate cell viability.

Bioinformatics software (ENCORI/starBase available at https://rnasysu.com/encori/) was used to predict the targeted binding relationship between EMX2OS and miR-654-3p, and between miR-654-3p and AKT3. ENCORI/starBase: it is an extended version developed on the basis of starBase, integrating more comprehensive RNA interaction data and analysis functions. It is a comprehensive RNA interaction reference map. The wild-type (Wt) and mutant (Mut) miR-654-3p binding site sequences in the EMX2OS and AKT3 3′-UTR were integrated into the pmirGLO vector for the generation of the Wt vector and Mut vector. miR-654-3p mimic or inhibitor was co-transfected with EMX2OS-Wt or EMX2OS-Mut and AKT3-Wt or AKT3-Mut into cells with Lipofectamine 3000. After 48 h, cells were harvested using the lysate from a commercial kit and relative luciferase activity was measured.

According to the Ferrozine colorimetric assay kit (TC0983, Leagene), the cell homogenate supernatants were used to measure iron level. Next, the cells were centrifuged at 13,000 g for 10 min at 4 °C to obtain the supernatant. In a centrifuge tube, the supernatant was incubated with buffer for 10 min at 37 °C. The OD was determined at 562 nm. The sample size for each experimental group was n = 6. Additionally, each group of experiments was independently repeated three times.

The content of lipid oxidation in cells was detected by MDA detection kit (S0131S, Beyotime) following the manufacturer's instructions. After the cells were broken by ultrasound, the supernatant was collected, and MDA detection reagents were added followed by 30 min incubation at 90 ℃. The absorbance of the supernatant at 532 nm and 600 nm was detected, and MDA levels were calculated. The sample size for each experimental group was n = 6. Additionally, each group of experiments was independently repeated three times.

GSH (glutathione) assay kit (S0053, Beyotime) was utilized to measure the level of GSH. After the cells were broken by ultrasound, the supernatant was collected, and GSH detection reagents were added. The GSH concentration was obtained by measuring absorbance at 412 nm.

The diagnostic value of EMX2OS for NS was evaluated using the ROC curve analysis: ROC curve was potted based on the detection value of EMX2OS, and the diagnosis results of NS were assigned values for statistical analysis (GraphPad Prism, version 9.5.1) (1 for having the disease, 0 for not having the disease). The area under the ROC curve (AUC) was calculated. The closer the AUC value is to 1, the better the diagnostic efficacy of EMX2OS is (AUC > 0.7 indicates certain diagnostic value, AUC > 0.8 indicates good diagnostic value, and AUC > 0.9 indicates excellent diagnostic value). Determine the optimal cut-off value: By calculating (sensitivity + specificity), select the EMX2OS detection value corresponding to the maximum Youden index as the optimal cut-off value. Based on the optimal critical value, the sensitivity and specificity of EMX2OS in diagnosing NS were calculated respectively to comprehensively evaluate its diagnostic efficacy.

All assays were conducted in triplicates, using GraphPad Prism 9.0 and SPSS 27.0 for statistical analysis. All the measurement data in this study were normally distributed. Data were presented as the mean ± SD. The statistical tests of two-tailed student's t-test or one-way analysis of variance (ANOVA) followed by Tukey test were employed to ascertain the significant differences between the groups. ROC curve was drawn to evaluate the diagnostic value of EMX2OS for NS. Each group of experiments was independently repeated three times. P < 0.05 indicates that the difference is statistically significant.

The expression level of EMX2OS in the serum of neonates with sepsis was significantly higher than that of healthy neonates (Figure 1A, P < 0.001). The ROC curve was drawn to evaluate the diagnostic value of EMX2OS for NS (Figure 1B), and the results showed that the area under the ROC curve (AUC) was 0.871, the sensitivity was 73.3%, and the specificity was 86.7%.

The ALI model was established by LPS treatment, and the expression of EMX2OS was significantly higher than that in the control group (Figure 2A, P < 0.05). The viability of A549 cells in the LPS model group was significantly decreased (Figure 2B, P < 0.01), while the iron ion content was increased (Figure 2C, P < 0.001), the MDA level was significantly increased (Figure 2D, P < 0.001), but the GSH level was decreased (Figure 2E, P < 0.05).

We identified the binding site of EMX2OS to miR-654-3p through a bioinformatics database (starBase) (Figure 3A). The results of the dual-luciferase reporter assay showed (Figure 3B) that after transfection with miR-654-3p mimic, the luciferase activity of the EMX2OS-Wt reporter gene significantly decreased (P < 0.001); while after transfection with miR-654-3p inhibitor, the luciferase activity of EMX2OS-Wt significantly increased (P < 0.001). Regardless of whether miR-654-3p mimic or inhibitor was used for transfection, the luciferase activity of the EMX2OS-Mut reporter gene did not show any statistical difference compared with the respective negative controls (mimic NC and inhibitor NC) (P > 0.05).

As shown in Figure 3C, the expression of EMX2OS was significantly downregulated in cells transfected with si-EMX2OS (P < 0.01). si-EMX2OS group: the expression of miR-654-3p was significantly upregulated (Figure 3D, P < 0.001), suggesting that the knockdown of EMX2OS might relieve the inhibitory effect on miR-654-3p. miR-654-3p inhibitor group: the expression of miR-654-3p was expectedly downregulated (Figure 3D, P < 0.01), verifying the effectiveness of the inhibitor. si-EMX2OS+miR-654-3p inhibitor group: compared with the single knockdown of EMX2OS, the combined inhibitor treatment could reverse the high expression of miR-654-3p induced by si-EMX2OS (Figure 3D, P < 0.001). The si-EMX2OS treatment significantly enhanced the cell proliferation ability (Figure 3E, P < 0.01), while the miR-654-3p inhibitor alone inhibited proliferation (Figure 3E, P < 0.01). After combining si-EMX2OS with the miR-654-3p inhibitor, the proliferation ability partially recovered (Figure 3E, P < 0.01).

Compared with the si-NC group, the intracellular iron content and MDA level in the si-EMX2OS group were significantly decreased (Figures 4A,B, P < 0.01), and the GSH content was increased (Figure 4C, P < 0.05). However, miR-654-3p inhibitor group had significantly increased iron content and MDA level (Figures 4A,B, P < 0.05), and decreased GSH content (Figure 4C, P < 0.05). Notably, the levels of iron and MDA in the si-EMX2OS+miR-654-3p inhibitor group were significantly higher than those in the si-EMX2OS group, and the GSH content was significantly lower (Figures 4A–C, P < 0.05).

Next, we identified a potential interaction between miR-654-3p and AKT3 (Figure 5A). miR-654-3p mimics inhibited and miR-654-3p inhibitors increased the luciferase activity of AKT3 3’UTR (Figure 5B, both P < 0.001). In addition, AKT3 expression was significantly promoted by miR-654-3p inhibitor, but inhibited by reduced EMX2OS, and its expression was successfully decreased by si-AKT3 (Figure 5C, all P < 0.01).

Compared with the Mock group, the viability of ALI model cells in the si-EMX2OS group and the si-AKT3 group were significantly increased (Figure 5D, P < 0.01). However, the cell viability in the miR-654-3p inhibitor group was significantly decreased (Figure 5D, P < 0.01). Notably, the cell viability of the si-AKT3+miR-654-3p inhibitor group was significantly higher than that of the miR-654-3p inhibitor group (Figure 5D, P < 0.01).

Compared with the Mock group, the intracellular iron content and MDA level were significantly decreased, and the GSH content was significantly increased (Figures 6A–C, P < 0.05) after EMX2OS and AKT3 knockdown. However, the miR-654-3p inhibitor treatment group showed the opposite trend, with increased iron and MDA levels and decreased GSH content (Figure 6A, P < 0.05). In addition, co-transfection of si-AKT3 significantly reversed the promoting effect of miR-654-3p inhibitor on ferroptosis, as shown by the decrease of iron and MDA levels and the increase of GSH content (Figure 6A, P < 0.05).