Intestinal samples from 17 NEC children (average age: 7.04 days, median age: 3 days) and 22 children with intestinal atresia (IA) (average age: 13.47 days, median age: 7 days) were collected in the General Surgery Department of Shanghai Children's Hospital from June of 2014 to December of 2018. Among these NEC children's intestinal samples, both necrotic and uninflamed intestinal tissues were collected. For IA samples, only uninflamed intestinal tissues were collected. The acquisition of the tissue samples was approved by the Institutional Review Board at Shanghai Children's Hospital, and written informed consent was obtained from each patient's parents before inclusion in the study.

C57BL/6J mice (7 days old) weighing about 4 g were purchased from Shanghai Jiesijie Experimental Animal Co., Ltd. (Shanghai, China). NEC was induced using formula gavage 6 times/day (15 g Similac Advance infant formula [Abbott Laboratories, Chicago, USA] in 75 ml canine milk replacement, using a 1.9-French angio-catheter placed into the mouse esophagus under direct vision), as well as hypoxia (99% N2) for 60 s and cold (4°C) for 10 min in a hypoxic chamber (Billups-Rothenberg, CA, USA) twice daily for 4 days. The protocol was partly modified as described previously (Leaphart et al., 2007a; Sodhi et al., 2010; Afrazi et al., 2014). Several studies have shown that a NEC animal model can be induced by this protocol that resembles human NEC (Nadler et al., 2000; Cetin et al., 2004; Qureshi et al., 2005; Leaphart et al., 2007b). Control mice were fed by a dam. Terminal ileum samples were harvested as soon as animals died from NEC or were euthanized by decapitation at the end of the experimental period for subsequent analysis. All experiments were approved by the Institutional Review Board at our hospital.

Adenovirus was constructed by Shanghai Obio Technology company (Shanghai, China). Positive adenovirus was pDKD-CMV-mcherry-U6-miR30(mmu-miR-146a-5p), and scrambled adenovirus was pDKD-CMV-mcherry-U6-shRNA. NEC and control mice were divided into four groups for adenovirus intraperitoneal injection: control mice injected with scrambled adenovirus (Scrambled +ctrl), NEC mice injected with scrambled adenovirus (Scrambled +NEC), control mice injected with positive adenovirus (miR-146a-5p-OE+ctrl), and NEC mice injected with positive adenovirus (miR-146a-5p-OE+NEC). A total volume of 50 μl containing 1.58 × 10⁹ TU adenoviruses was intraperitoneally injected 24 h before NEC was conducted, which is in accordance with previous research (Nomura et al., 2006). Animals' daily body weight and survival times were recorded. Samples were harvested when animals died from NEC or were euthanized by decapitation at the end of the experimental period for subsequent analysis.

THP-1 cells were kindly provided by Stem Cell Bank, Chinese Academy of Sciences. THP-1 cells were cultured in RPMI supplemented with 10% FBS (fetal bovine serum) and 50 μM 2-mercaptoethanol (Sigma-Aldrich, MO, USA), in a humidified atmosphere of 37°C and 5% CO₂. For experiments, a total of 1 × 10⁶ cells were plated on a 12-well plate overnight. THP-1 cells were differentiated at 24 h using 0.5 mM phorbol 12-myristate 13-acetate (PMA) (Sigma-Aldrich, MO, USA). The following day, medium was replaced by Opti-DMEM, and cells were pre-incubated with 80 nM miR-146a-5p overexpression sequence (mimic: sense: 5′-UGAGAACUGAAUUCCAUGGGUU-3′, antisense: 5′-CCCAUGGAAUUCAGUUCUCAUU-3′) or miR-146a-5p knockdown sequence (inhibitor: 5′-AACCCAUGGAAUUCAGUUCUCA-3′) (Gene-Pharma, Shanghai, China), respectively, and their scrambled control sequences were preincubated under the same conditions (mimic scrambled: sense: 5′-UUCUCCGAACGUGUCACGUTT-3′, antisense: 5′-ACGUGACACGUUCGGAGAATT-3′), inhibitor scrambled sequence: (nc-in: 5′-CAGUACUUUUGUGUAGUACAA-3′) (Gene-Pharma, Shanghai, China). After 48 h, cells were treated for 3 h with 0.5 μg/ml LPS (Sigma-Aldrich, MO, USA) and 5 mM ATP (Sigma-Aldrich, MO, USA), 1.0 μg/ml LPS and 5 mM ATP, and 10 μg/ml LPS and 5 mM ATP, respectively, and ATP was added during the last 30 min of LPS stimulation. The supernatant medium and treated cells were collected for subsequent analysis.

The collected terminal ileum samples were fixed in 4% paraformaldehyde solution, embedded in paraffin, cut into sections, stained with hematoxylin and eosin (HE), and pathologically graded according to the published scoring criteria by two expert pathologists who were blind to the experimental groups (Sheng et al., 2013, 2014; Ginzel et al., 2016). The grading system was as follows: Grade 0: normal ileum, Grade 1: mild injury to the tip of the villus, Grade 2: partial loss of the villus, Grade 3: severe submucosal injury, and Grade 4: complete necrosis. A grade equal to or >2 suggests the occurrence of NEC.

Human samples and mouse terminal ileums were fixed in hybridization in situ fixing solution (Servicebio, Wuhan, China), embedded in paraffin, and cut into 5-μm-thick sections. Next, the sections underwent hybridization in situ with an RNA FISH kit (Gene-Pharma, Shanghai, China). Sections were deparaffinized twice and hydrated through decreasing concentrations of ethanol (100, 95, 90, 80, and 70%). After that, sections were incubated with proteinase K at 37°C for 20 min and hydrated again through increasing concentrations of ethanol (70, 80, 90, and 100%). Next, sections were denaturation at 78°C for 8 min and then were incubated with miR-146a-5p probe (sequences: 5′-AACCCATGGAATTCAGTTCTCA-3′-FAM) (Gene-Pharma, Shanghai, China) at 4°C overnight. Subsequently, sections were incubated with immunofluorescence primary antibody (F4/80: Abcam, MA, USA) at 4°C overnight and then with secondary antibody at room temperature for 2 h. Next, sections were stained with 4′,6-diamidino-2-phenylindole (DAPI). Finally, sections were observed under a fluorescence microscope by two expert pathologists who were blind to the experimental groups, and miR-146a-5p positive cells, miR-146a-5p, and F4/80 double-positive cells were counted in five random scope fields, and an average of five fields was calculated.

For immunohistochemical staining, 5-μm-thick paraffin-embedded sections were deparaffinized and hydrated through decreasing concentrations of ethanol (100, 85, and 75%) after which peroxidase activity was blocked using 3% H₂O₂ in methanol for 15 min at room temperature. Heat-induced antigen retrieval was conducted in 10 mM sodium citrate solution (pH = 6.0) for about 2 min until boiling in a microwave. Sections were washed three times with PBS and blocked with 5% bovine serum albumin for 60 min at room temperature. Sections were incubated overnight at 4°C with primary antibodies specific for NLRP3 and Caspase-1 (Abcam, Cambridge, United Kingdom), followed by incubation with horseradish peroxidase (HRP)-coupled secondary antibodies with 50 mM Tris-HCl buffer (pH 7.4) at room temperature for 1 h. The color was developed by a 15 min incubation at room temperature with DAB solution (Sangon Biotech Co., Ltd., Shanghai, China), and then the sections were weakly counterstained with hematoxylin for 10 min at room temperature. Negative controls were included using the replacement of the primary antibody with PBS. The sections were lightly counterstained with hematoxylin.

The supernatant medium from the treated THP-1 cells was collected, and IL-1β, IL-6, IL-10, IL-18, and TNF-α were measured using the appropriate ELISA kits (Lianke Bio Co., Ltd., Hangzhou, China).

Total protein was isolated from cell and tissues using a radioimmunoprecipitation assay (RIPA) lysis buffer (Thermo Scientific, Waltham, MA, USA). Subcellular fraction protein was extracted using a membrane, nuclear, and cytoplasmic protein extraction kit (Sangon Biotech Co., Ltd., Shanghai, China). Protein concentrations of samples were determined using the BCA Protein Assay Kit (Thermo Scientific). Then, protein samples were subjected to SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) membranes, which were incubated in 5% BSA for 1 h at room temperature. Next, PVDF membranes were incubated overnight at 4°C with the following primary antibodies: NLRP3 (Abcam, MA, USA), pro-Caspase-1 (Santa Cruz Biotechnology, TX, USA), Caspase-1 p10 (Santa Cruz Biotechnology, TX, USA), β-actin (Abcam), CLIC4 (Abcam), histone H3 (Proteintech, Rosemont, USA), and N-cadherin (Abcam). Next, membranes were incubated with IgG HRP-conjugated secondary antibodies (Cell Signaling Technology, MA, USA) for 2 h at room temperature. Protein bands were detected using a ChemiDoc-It system (Tanon, Shanghai, China) and an ECL kit (Bio-Rad, CA, USA). Protein levels were determined using ImageJ (National Institutes of Health, MD, USA). The density of each band was normalized to its respective loading control (β-actin, histone H3, and N-cadherin).

Total RNA was extracted from the treated THP-1 cells using Trizol (Life Technologies, CA, USA) following the manufacturer's instructions. For miR-146a-5p, first-strand cDNA was synthesized using a stem-loop method reverse transcription kit (Sangon Biotech Co., Ltd.), and other genes' cDNA were reversed using a Takara reverse transcription kit (Takara Bio, CA, USA). A quantitative mRNA kit (Takara Bio) and microRNA qPCR kit (Sangon Biotech Co., Ltd.) were used for mRNA and microRNA quantitative real-time PCR analysis, respectively, using the SYBR Green fluorescence system (Roche, MA, USA). The following primers were used:

β-actin: forward: 5′-TCGTGCGTGACATTAAGGAGAAGC-3′,

reverse: 5′-GGCGTACAGGTCTTTGCGGATG-3′

U6: forward: 5′-AGAGAAGATTAGCATGGCCCCTG-3′,

reverse: 5′-ATCCAGTGCAGGGTCCGAGG-3′

NLRP3: forward: 5′-ATGCTGCCTGTTCTCATGGATTGG-3′,

reverse: 5′-GCTTCTGGTTGCTGCTGAGGAC-3′

Caspase-1: forward: 5′-ATGGACAAGTCAAGCCGCACAC-3′

reverse: 5′-TCCCACAAATGCCTTCCCGAATAC-3′

CLIC4: forward: 5′-GGCCAGAGGCTAATGAAGCACTG-3′,

reverse: 5′-GGCCACCACCTTGACAATATGCAG-3′

IL-1β: forward: 5′-GCGGCATCCAGCTACGAATCTC-3′,

reverse: 5′-CGGAGCGTGCAGTTCAGTGATC-3′

IL-18: forward: 5′-TGGCTGCTGAACCAGTAGAAGA-3′,

reverse: 5′-TGGTCCGGGGTGCATTATCT-3′

IL-6: forward: 5′-GGTGTTGCCTGCTGCCTTCC-3′,

reverse: 5′-GCTCTGGCTTGTTCCTCACTACTC-3′

IL-10: forward: 5′-ACTGCTCTGTTGCCTGGTCCTC-3′,

reverse: 5′-GCCTTGATGTCTGGGTCTTGGTTC-3′

TNF-α: forward: 5′-TGCTCCTCACCCACACCATCAG-3′,

reverse: 5′-TCCCAAAGTAGACCTGCCCAGAC-3′

miR-146a-5p: forward: 5′-CGCGTGAGAACTGAATTCCA-3′,

reverse: 5′-AGTGCAGGGTCCGAGGTATT-3′

Specific mRNA expression or miR-146a-5p levels were normalized relative to β-actin mRNA or U6 levels, respectively, using the comparative 2^−ΔΔCt method.

Statistical analysis was performed using the SPSS21.0 statistical software. The descriptive data are expressed as mean ± SD. The groups were compared using a t-test when two groups were compared, or a one/two-way analysis of variance (ANOVA) when more than two groups were compared. The log-rank test was used for survival analysis. Differences at a value of P < 0.05 were considered statistically significant.

The pathological grade scores of NEC mouse terminal ileum intestine were assessed and ranged according to HE as follows: Grade 0: normal ileum, Grade 1: mild injury to the tip of the villus, Grade 2: partial loss of the villus, Grade 3: severe submucosal injury, and Grade 4: complete necrosis. A grade equal to or higher than 2 suggested the occurrence of NEC (Figure 1A). In previous studies, miR-146a-5p played a key role in regulating macrophage function. The miR-146a-5p was detected by FISH in the intestinal tissues of NEC mice. The miR-146a-5p was merged with F4/80 detecting by immunofluorescence, and the miR-146a-5p and F4/80 double-positive cells were counted. We found that miR-146a-5p was significantly increased in the NEC group (Figures 1B,C). The degree of intestinal destruction varied greatly between inflammatory and unaffected intestine. We assessed the pathological grade score of NEC inflammatory and unaffected intestine by HE (Figure 2A). The level of miR-146a-5p and F4/80 double-positive cells was significantly higher in the NEC inflamed intestine than in the unaffected intestine and samples showing intestinal atresia (Figures 2B,C). The number of miR-146a-5p and F4/80 double-positive cells also demonstrated positive correlation to the HE pathological grade score in the NEC human samples (Figure 2D).

NLRP3 inflammasome activation was reported to be an important regulator in NEC (Yin et al., 2020). Thus, we detected the expression level of NLRP3 inflammasome relative proteins, and their downstream inflammatory factors IL-1β and IL-18. Our results showed that mRNA of NLRP3, Caspase-1, IL-1β, and IL-18 was higher in the NEC group than that in the control group (Figures 3A–D). ELISA showed that IL-1β and IL-18 both increased in NEC (Figures 3E,F). The protein expression levels of NLRP3, pro-Caspase-1, and Caspase-1 p10 were all higher in the NEC mice (Figure 3G). In children's samples, both NLRP3 and Caspase-1 were increased in NEC, and they were expressed both in the epithelium and lamina propria (Figure 3H).

The above data demonstrated that NLRP3 inflammasome enzymatic protein caspase-1, as well as their downstream IL-1β and IL-18, were increased in NEC.

MiR-146a-5p was considered to be a negative regulator in innate immune responses in previous studies. In this study, Figure 2 show that miR-146a-5p-positive macrophages were associated with a worse pathological grade. Therefore, we investigated the role of miR-146a-5p mimics in the LPS/ATP-induced NLRP3 inflammasome activation. A THP-1 cell line was trans-ducted with miR-146a-5p overexpression sequence for 48 h before LPS/ATP stimulation. After miR-146a-5p overexpression, miR-146a-5p expression level was confirmed (Supplementary Figure 1). The results showed that NLRP3 inflammasome relative proteins, IL-1β, IL-18, IL-6, IL-10, and TNF-α were increased after LPS/ATP incubation (Figure 4). MiR-146a-5p overexpression had no significant effect on NLRP3 mRNA and protein expression (Figures 4A,J), but inhibited the mRNA expression of Caspase-1, IL-1β, and IL-18 at 0.5 and 1.0 μg of LPS, but not at 10 μg of LPS (Figures 4B–D). ELISA showed that IL-1β and IL-18 were downregulated by miR-146a-5p overexpression at 0.5 and 1.0 μg of LPS, but not at 10 μg of LPS (Figures 4E,F), while IL-6, IL-10, and TNF-α were downregulated by miR-146a-5p overexpression at a different degree at a gradient of LPS concentration (Figures 4G–I). In addition, protein levels of pro-Caspase-1 and Caspase-1 p10 were downregulated by miR-146a-5p overexpression, and NLRP3 showed no significant changes (Figure 4J).

Next, we determined the effect of miR-146a-5p knockdown in the LPS/ATP-induced NLRP3 inflammasome activation. However, miR-146a-5p knockdown showed no effects on the NLRP3 inflammasome or IL-1β and IL-18 mRNA (Supplementary Figures 2A–D) and protein (Supplementary Figure 2J) expression. Additionally, ELISA demonstrated that miR-146a-5p knockdown had no significant effect on NLRP3 downstream inflammatory factors IL-1β and IL-18, however, it promoted IL-10 and TNF-α expression (Supplementary Figures 2E–I).

All these data indicated that only miR-146a-5p overexpression decreased LPS/ATP-induced NLRP3 inflammasome enzymatic protein caspase-1 and downstream inflammatory factors expression.

Numerous studies suggest that CLIC4 membrane expression is closely relative to LPS-induced NLRP3 activation (He et al., 2011; Malik et al., 2012; Domingo-Fernandez et al., 2017; Tang et al., 2017). The expression of CLIC4 was significantly increased in NEC human inflamed intestinal tissues, but not in NEC unaffected tissues and intestinal atresia tissues (Figure 5). Next, we investigated whether miR-146a-5p overexpression inhibited LPS/ATP-induced CLIC4 expression level in membrane. The expression level of miR-146a-5p was confirmed after miR-146a-5p overexpression sequence incubation (Supplementary Figure 1). The expression of CLIC4 was increased after LPS/ATP incubation (Figure 6), and miR-146a-5p overexpression inhibited CLIC4 mRNA and protein expression (Figures 6A–C). However, the inhibitory effect of miR-146a-5p overexpression on CLIC4 was not as significant across the whole cell (Figure 6C, first band of the whole cell). Malik et al. showed that CLIC4 nuclear translocation regulated macrophage deactivation, while nuclear-targeted CLIC4 overexpression downregulated IL-β (Malik et al., 2012). To investigate these possibilities, we detected CLIC4 protein in the membrane, cytoplasm, and nucleus of the treated cells. CLIC4 protein was increased and mainly expressed on the membrane of the treated cells (Figure 6C). In addition, CLIC4 was significantly downregulated by miR-146a-5p mimics in the membrane and showed a more obvious difference than that in the whole cell, cytoplasm, and nucleus (Figures 6D–G).

Next, we sought to confirm miR-146a-5p knockdown's effect on CLIC4 after LPS/ATP incubation. The expression level of miR-146a-5p was confirmed after miR-146a-5p knockdown sequence incubation (Supplementary Figure 3A). PCR showed that miR-146a-5p knockdown had no effect on CLIC4 mRNA and protein expression (Supplementary Figures 3B,C). Western blot showed that miR-146a-5p knockdown had no effect on CLIC4 protein expression across the whole cell, membrane, cytoplasm, and nucleus (Supplementary Figures 3D–H).

Collectively, miR-146a-5p overexpression decreased NLRP3 inflammasome upstream protein CLIC4 membrane expression.

To further validate the effects of miR-146a-5p in NEC development, we constructed a miR-146a-5p overexpression adenovirus. Neonatal mice were trans-ducted with the control and positive adenovirus by intraperitoneal injection. Then, these neonatal mice were randomly allocated to the control or NEC group. The miR-146a-5p expression level in the terminal ileum was confirmed (Supplementary Figure 4). The overall survival rate of the NEC group was significantly lower than that of the control group. MiR-146a-5p overexpression adenovirus transduction markedly improved the NEC survival rate (Figure 7A). In addition, miR-146a-5p attenuated the NEC-induced weight loss and alleviated intestinal damage (Figures 7B,C; Table 1). Representative images of gross morphology demonstrated that edema, congestion, necrosis, and reddish-black coloring were more severe in the Scrambled + NEC group than those in the miR-146a-5p-OE + NEC group (Figure 7D). These data indicated that miR-146a-5p protected neonatal mice from NEC intestinal injury and promoted digestive tract growth. Furthermore, miR-146a-5p significantly reduced NLRP3 inflammasome mRNA and protein (Figures 7E–I) expression as well as IL-1β, IL-18 (Figures 7G–K) secretion in vivo.

In summary, all these results suggested a protective role of miR-146a-5p in NEC development by inhibiting NLRP3 inflammasome downstream inflammatory factors and CLIC4.