Animals preparation and NS model induction
All the animal experiments were performed under the approval of the Animal Care and Use Committee of Maternal and Child Health Care Hospital of Zibo, and were carried out in accordance with the Animal Welfare Law and the guidelines for animal care and use of the National Institutes of Health. Neonatal mice (5–7 days old) bred by the C57BL/6J mice were prepared by Wanleibio (Shenyang, China) and maintained with free access to standard care, diet, and drinking. In the current study, the cecal slurry (CS) model was established to induce NS following the of method by Wynn et al.  with modifications: briefly, the adult mice were sacrificed and the cecal contents were collected, and mixed with 5% glucose solution to produce CS solution (80 mg/ml). Then 60 neonatal mice were randomly divided into four groups (15 for each group): Sham group, mice received intraperitoneal injection of 0.9% saline. CS group, mice received intraperitoneal injection of CS at a concentration of 1.3 mg/g body weight according to the study by Li et al. ; CS + Low group, mice received intraperitoneal injection of CS and low dose of berberine (50 μg); CS + High group, mice received intraperitoneal injection of CS and high dose of berberine (100 μg). 24 h after the injection, the survival rate of the mice was calculated in a 72-h period and the left mice were sacrificed using overdose (150 mg/kg BW) pentobarbital sodium. The small intestine and whole blood samples were collected from the survival mice for the subsequent assays.
Hematoxylin-eosin (H&E) staining and intestinal injury score
Intestinal injuries were detected with H&E staining: the tissues were immersed, fixed in 8% formalin for embedding, and sectioned. Then the sections were incubated with hematoxylin and eosin. The histological changes in the tissues were detected under a microscope under 200× magnification and the injury degree of the tissues were evaluated by two investigators blind to the experimental designment following previous methods . Briefly, a score of 0 represented normal mucosa; a score of 1 represented the development of subepithelial Gruenhagen’s space, vacuolization or subepithelial lifting limited to the lamina propria or tips of villi; a score of 2 represented epithelial lifting and vacuolization greater than half of the villi, villi distortion, or mucosal ulceration and disintegration of the lamina propria.
Enzyme linked sorbent immune assay (ELISA)
The blood levels of cytokines IL-6 (SEA079Hu), IL-1β (SEA563Hu), and TNF-α (SEA133Hu) were detected using corresponding kits purchased from USCN Business Co., Ltd (China).
Reverse transcription quantitative PCR (RT-qPCR)
Total RNA in small intestine tissues was collected and extracted using a RNA extraction kit (QIGEN, Duesseldorf, Germany), and was reversely transcribed into cDNA (50 ng/μl) according to the instructions of PrimeScriptTM RT reagent Kit (Beijing, China). The expression of miR-132-3p (forward: 5’- GCGCGTAACAGTCTACAGCCA-3’; reverse: 5’- AGTGCAGGGTCCGAGGTATT-3’) was detected using a Real-time PCR Detection System following routine protocol (ABI 7900, Shanghai, China) and the relative expression level of miR-132-3p was calculated using the formula of 2-△△cq in reference to Sham group (normalized as 1).
Western blotting assay
Total protein products of small intestinal tissues were extracted using the RIPA lysis buffer and were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Then the membranes were incubated with primary antibodies against different antibodies (Table S1). After washing, protein blots were developed using Beyo ECL Plus reagent. The results were scanned in the Gel Imaging System (WD-9413B, Liuyi Factory, China) and the integrated optical densities were recorded. The relative expression levels of the different proteins were calculated with Gel-Pro-Analyzer (Media Cybernetics, USA).
Inhibition of miR-132-3p in mice
To verify the interaction between miR-132-3p and berberine in the treatment of NS, the level of miR-132-3p was inhibited using specific antagomir 24 h prior to CS and berberine administrations. Then the changes in survival rate, small intestine histology, systemic inflammation, and activities of FOXA1 and NF-κB pathways were detected as described above.
All the continuous data were represented as mean ± standard deviation (SD). The difference was calculated using one-way analysis of variance (ANOVA) followed by multiple comparisons with Tukey method. Survival was compared using log-rank test. All the statistical analyses and graphic manipulation were conducted using Graphpad Prism version 6.0 (GraphPad Software, Inc., San Diego, CA) with a significant level of 0.05 (two-tailed P value).