miR-19b downregulates intestinal SOCS3 to reduce intestinal inflammation in Crohn’s disease

Although aberrant microRNA (miRNA) expression has frequently been observed in inflammatory bowel disease (IBD), its biological functions and targets remain largely unknown. Present study found that miR-19b was significantly downregulated in active Crohn’s disease (CD). Using bioinformatics analysis, suppressor of cytokine signalling 3 (SOCS3), a physiological regulator of innate and adaptive immunity that controls several immuno-inflammatory diseases, was predicted to be a potential target of miR-19b. An inverse correlation between miR-19b and SOCS3 protein levels, but not mRNA, was identified in active-CD intestinal tissue samples. By overexpressing or knocking down miR-19b in Caco2 cells and HT29 cells, it was experimentally validated that miR-19b is a direct regulator of SOCS3. Using a luciferase reporter assay, it was confirmed that miR-19b directly recognizes the 3’-untranslated region (3’-UTR) of SOCS3. Furthermore, overexpression of miR-19b decreased SOCS3 expression, leading to increased production of macrophage-inflammatory protein-3α (MIP-3α) in Caco2 cells. In contrast, knockdown of miR-19b increased SOCS3 and decreased MIP-3α. Finally, intracolonically delivered miR-19b decreased the severity of colitis induced with 2,4,6-trinitrobenzene sulphonic acid (TNBS). Taken together, our findings suggest that miR-19b suppresses the inflammatory response by inhibiting SOCS3 to modulate chemokine production in intestinal epithelial cells (IECs) and thereby prevents the pathogenesis of CD.

JAK or cytokine receptors to suppress downstream signalling events 8 . Among the SOCS family members, SOCS1 and SOCS3 are key regulators of innate and adaptive immunity 6,9 . Because SOCS3 regulates multiple cytokine signalling pathways, it may be a useful therapeutic target for autoimmune disease 9,10 . SOCS3 expression is increased in inflamed tissue compared to normal tissue, and its expression is particularly high in recruited leukocytes and the epithelium 11 . Intestinal epithelial cells (IECs) have been consistently linked to IBD pathogenesis. During the course of intestinal inflammation or microbial infection, IECs exhibit a classic inflammatory response in addition to their normal absorptive and secretory functions 12,13 . High SOCS3 expression has been observed in IBD, although the role of SOCS3 in IBD remains unclear 14 . It is important to understand what molecules regulate SOCS3 expression to identify potential therapeutic targets for anti-inflammatory therapies.
MicroRNAs (miRNAs) are non-coding RNA molecules (21-23 nucleotides in length) that post-transcriptionally regulate gene expression. miRNA binding to complementary sequences in the 3-untranslated region (UTR) of target mRNA molecules results in mRNA degradation or translational inhibition 15 . Although most miRNA target genes have not been identified, miRNAs have been implicated in several cellular processes, including differentiation, proliferation, maturation, and apoptosis. Furthermore, there is accumulating evidence that miRNAs regulate inflammatory processes 16,17 . Differential miRNA expression has been described in autoimmune diseases, suggesting that miRNA regulation could be involved in autoimmune disease development or prevention, including conditions such as psoriasis, rheumatoid arthritis (RA), and systemic lupus erythematosus 18,19 . Recently, unique miRNA expression profiles have been described in active UC and CD patient epithelia 17,20 . Our previous study showed that several miRNAs are differentially expressed in the intestine of active CD patients. However, differential expression of miRNAs and their roles in epithelial disruption during IBD remain unclear. In the current study, we hypothesized that intestinal epithelia disruption is linked to abnormal SOCS3 expression and that miRNAs regulate this abnormal expression during intestinal inflammation.

SOCS3 protein, but not SOCS3 mRNA, is upregulated in active CD intestinal mucosa.
To evaluate SOCS3 expression and distribution in the intestinal mucosa, we performed quantitative RT-PCR, immunohistochemical (IHC) analysis and Western blot analysis. IHC analysis revealed that SOCS3 protein was highly expressed in the intestinal mucosa of active CD patients compared to normal controls, especially in the epithelial layer (Fig. 1a). In addition, the intensity of SOCS3 expression was significantly higher in sites with immobilized SOCS3 antibody, whereas expression was undetectable in sites labelled with isotype-matched control monoclonal antibody (mAb) (Fig. 1a). Based on Western blot analysis, SOCS3 protein was intensely expressed in the inflamed jejunum, ileum and colon mucosa of CD patients, whereas this expression was barely detectable in normal colon mucosa; there was no difference in SOCS3 expression between the jejunum, ileum and colon in CD patients (Fig. 1b). Interestingly, SOCS3 mRNA expression was not significantly different among inflamed colonic tissue and small intestinal mucosa from CD patients and normal intestinal mucosa from control subjects (Fig. 1c). This disparity between SOCS3 protein and mRNA expression in active CD suggests that SOCS3 expression is post-transcriptionally regulated.
The SOCS3 3'-UTR contains conserved miR-19b target sites. Because miRNA is an important post-transcriptional regulator of gene expression, we hypothesized that miRNAs regulate SOCS3 expression. We used the computer-aided algorithms TargetScan, PicTar and microRNA.org to predict miRNAs that could potentially target SOCS3. Using these approaches, we identified eight miRNAs (miR-218, miR-19a, miR-19b, miR-148a, miR-148b, miR-152, miR-30a, miR-30d) that could potentially bind to complementary sequences in the SOCS3 3'-UTR .We next examined the expression of these miRNAs in inflamed intestinal mucosa from active CD patients by TaqMan probe-based qRT-PCR. Among these miRNAs, miR-19b was downregulated and miR-148a was upregulated, whereas the expression levels of miR-218, miR-19a, miR-148b, miR-152, miR-30a, and miR-30d showed no changes in active CD (Fig. 2a). As shown in Figure 2b, we observed a miR-19b-SOCS3 3'-UTR hybrid; the minimum free energy value of miR-19b-SOCS3 hybridization was -27.8 kcal/mol, which is well within the range of genuinemiRNA-target pairs determined by RNAhybrid analysis 21 . Moreover, there was perfect base pairing between the seed region (the core sequence encompassing the first 2-8 bases of mature miRNA) and the cognate targets. Furthermore, miR-19b binding sequences in the SOCS3 3'-UTR were highly conserved across species (Fig. 2b). Thus, human SOCS3 is likely a direct miR-19b target. miR-19b and SOCS3 levels are inversely correlated in active CD intestinal tissues. miRNAs generally show opposite expression patterns when compared to their targets 22 . Thus, we investigated whether miR-19b was inversely correlated with SOCS3 in active CD. Using quantitative real-time PCR, we found that miR-19b expression was significantly decreased in inflamed intestinal mucosa from active CD patients compared to normal control samples (Fig. 3a). Using miRNA in situ hybridization, we detected miR-19b expression in epithelial cells and non-epithelial cells (Fig. 3b). In addition, miR-19b and SOCS3 protein expression were inversely correlated (Fig. 3c). The inverse correlation between miR-19b and SOCS3 protein level (Fig. 3c) and the disparity between miR-19b and SOCS3 mRNA levels ( Fig. 3d)  to block translation without affecting transcript levels 22 . Our results indicate that SOCS3 expression is regulated by a typical miRNA-mediated, post-transcriptional mechanism. Based on computational predictions and the inverse correlation between miR-19b and SOCS3 protein levels in active CD, these data suggest that SOCS3 is a miR-19b target. Validation of human intestinal SOCS3 as a direct miR-19b target. We further examined the correlation between miR-19b and SOCS3 by altering the miR-19b level and examining the effect on SOCS3 expression in Caco2 cells and HT29 cells. We first overexpressed miR-19b by transfecting cells with pre-miR-19b, a synthetic RNA oligonucleotide duplex mimicking the miR-19b precursor. Then, we knocked down miR-19b by transfecting cells with anti-miR-19b, a chemically modified single-stranded antisense oligonucleotide designed to specifically target mature miR-19b. We validated these changes in miR-9b expression by quantitative RT-PCR 24 h after transfection (Fig. 4a).
To determine whether miR-19b overexpression or knockdown affected the SOCS3 level, we repeated the above experiments and examined SOCS3 expression 24 h after transfection. miR-19b overexpression significantly inhibited SOCS3 protein expression, whereas miR-19b knockdown significantly upregulated SOCS3 expression compared to the ncRNA control (Fig. 4b). In contrast, miR-19b overexpression or knockdown had no effect on SOCS3 mRNA stability (Fig. 4c). SOCS3 knockdown by RNA interference significantly decreased SOCS3 protein expression ( Supplementary Fig. S1), while cells transfected with a SOCS3 overexpression plasmid expressed higher SOCS3 protein in both cell lines ( Supplementary  Fig. S1). These results demonstrate that miR-19b specifically regulates SOCS3 protein expression at the post-transcriptional level, which is a canonical mechanism of miRNA-mediated regulation.
To determine how miR-19b negatively affects SOCS3 expression, we fused full-length SOCS3 3'-UTR (1574bp) containing the presumed miR-19b binding site downstream of the firefly luciferase gene in a reporter plasmid. We co-transfected Caco2 cells and HT29 cells with the luciferase reporter and a transfection control β -galactosidase expression plasmid along with pre-miR-19b, anti-miR-19b or scrambled ncRNA. As expected, in both cell lines, miR-19b overexpression led to an approximate 50% reduction in luciferase reporter activity compared to control cells treated with scrambled ncRNA, whereas inhibition of miR-19b resulted in a significant increase in reporter activity compared to cells transfected with ncRNA (Fig. 4d). Furthermore, we introduced point mutations into the complementary sites of the SOCS3 3'-UTR to eliminate the predicted miR-19b binding site. As shown in Fig. 4d, the mutated luciferase reporter was not affected by altering the miR-19b level. These results suggest that the miRNA binding site strongly contributes to the miRNA-mRNA interaction that mediates post-transcriptional SOCS3 expression. Taken together, these results unequivocally demonstrate that miR-19 directly recognizes and binds the SOCS3 3' UTR and inhibits its expression at the post-transcriptional level.

miR-19b modulates chemokine production in Caco2 cells via SOCS3 downregulation.
To determine the function of miR-19b in IECs during intestinal inflammation, we evaluated the effects of miR-19b on chemokine production in IECs. First, chemokine levels were examined by chemokine array, and we observed alterations of chemokine production in Caco2 cells induced by IL-6 (shown in Supplementary Table S1, Table S2). We also found that the levels of IL-8 and CXCL16 were increased 72 h after IL-6 treatment. The level of macrophage-inflammatory protein-3α (MIP-3α ) was increased 24 h after IL-6 treatment and remained high until 72 h after treatment. Therefore, we focused on these three chemokines (MIP-3α , IL-8 and CXCL16) in subsequent experiments. We first transfected Caco2 cells with pre-miR-19b or anti-miR-19b and then treated the cells with IL-6, and enzyme-linked immunosorbent assays (ELISAs) were performed to measure the levels of these three chemokines. As shown in Fig. 5a, MIP-3α was significantly increased upon miR-19b up-regulation and decreased upon miR-19b down-regulation 24 h after IL-6 treatment. However, the MIP-3α level was not significantly different 72 h after transfection with pre-miR-19b or anti-miR-19b ( Supplementary Fig. S2). Similarly, the IL-8 and CXCL16 levels were not significantly different upon up-regulation or down-regulation of miR-19b after 72 h IL-6 treatment ( Supplementary Fig. S2). We then transfected cells with siRNA to deplete SOCS3 and stimulated the cells with IL-6 for 24 h, and the results revealed increased levels of MIP-3α ( Supplementary Fig. S3a). In contrast, transfection with the SOCS3 overexpression plasmid alone downregulated MIP-3α compared to control cells transfected with an empty plasmid control ( Supplementary  Fig. S3a). However, 72 h of IL-6 stimulation did not significantly affect the MIP-3α level regardless of the SOCS3 protein level ( Supplementary Fig. S3b). We next investigated whether SOCS3 overexpression Firefly luciferase reporters containing wild-type (WT) or mutant (MT) miR-19b binding sites in the SOCS3 3'-UTR were co-transfected into Caco2 cells and HT29 cells with scrambled ncRNA, pre-miR-19b or anti-miR-19b. Twenty-four hours after transfection, a luciferase assay was performed. Firefly luciferase values were normalized to β -galactosidase activity and plotted as relative luciferase activity. The results were calculated as the ratio of firefly luciferase activity in miR-19b-transfected cells to the luciferase activity in control cells. Left: Caco2 cells. Right: HT29 cells. The results are presented as the mean± SEM from three independent experiments (Caco2 group: * p = 0.013, pre-scramble vs. pre-miR-19b, *** p = 0.0000, anti-scramble vs. anti-miR-19b; HT29 group: *** p = 0.000021, pre-scramble vs. pre-miR-19b, *** p = 0.000019, anti-scramble vs. anti-miR-19b).
was sufficient to reverse the inhibitory effects of miR-19b on MIP-3α production in Caco2 cells. We co-transfected the SOCS3 overexpression plasmid and pre-miR-19b into Caco2 cells and examined the level of MIP-3α . Compared to cells transfected with pre-miR-19b, cells transfected with pre-miR-19b and the SOCS3 overexpression plasmid exhibited significantly lower production of MIP-3α , suggesting that miR-19b upregulation induced the increased expression of MIP-3α , which could be reversed by SOCS3 overexpression (Fig. 5b). Taken together, these results demonstrate that miR-19b promotes IEC chemokine production by down-regulating SOCS3 protein.
The relationship between miR-19b and SOCS3 in the pathogenesis of colitis. To investigate the relationship between miR-19b and SOCS3 in the pathogenesis of colitis, we employed the TNBS-induced colitis model and examined SOCS3 expression after intra-colonic pre-miR-19b, pre-scramble or 50% ethanol administration. As shown in Fig. 7a,b, the expression of miR-19b in TNBS-induced colitis was decreased, while the expression of SOCS3 in TNBS-induced colitis was increased compared with the control. However, SOCS3 protein in TNBS-induced colitis was decreased after intra-colonic administration with pre-miR-19b. When we analysed the alteration in SOCS3 mRNA before and after administration of pre-miR-19b, there was no significant alteration in SOCS3 mRNA after treatment with pre-miR-19b (Fig. 7c). Furthermore, the expression of MIP-3α in TNBS-induced colitis was increased Histopathological scores were calculated according to H&E staining (n = 8 mice per group; *** p = 0.0000256, TNBS group vs. PEI+ pre-miR19b group; ** p = 0.003448, PEI group vs. PEI+ pre-miR-19b group; ** p = 0.007246, PEI+ pre-scramble group vs. PEI+ pre-miR-19b group). (e), Immunofluoresence staining to determine miRNA precursor localization after intracolonic administration of the Cy3-labeled pre-miR-19b/PEI complex. Red, miR-19b-precusor; green, pan-cytokeratin (PCK); blue, DAPI nuclear staining (magnification, 400× ). All values are expressed as the mean± SEM.

Discussion
The etiology and pathogenesis of IBD remains unclear, although evidence from animal models and humans has demonstrated that the abnormal function of IECs contributes to tissue damage and the intestinal inflammatory response 23 . The intestinal epithelium is considered a constitutive component of the mucosal immune system. In our study, we provide the first report of a direct and important role for miR-19b in regulating SOCS3 expression in CD-associated inflammation and highlight its potential as a miRNA-based therapeutic candidate. miRNAs are involved in the innate and adaptive immune systems, and significant efforts have been made to determine the expression of miRNAs in autoimmune disease and their roles in the pathogenesis, diagnosis, and treatment of several diseases, including CD 24-26 . However, previous CD studies have been limited by an incomplete understanding of the mechanisms regulating gene expression.
In this study, we identified a conserved miR-19b binding site in the human SOCS3 mRNA 3'-UTR. We examined miR-19b and SOCS3 expression levels in human inflammatory intestinal tissue and normal intestinal tissue and found an inverse correlation between the miR-19b level and SOCS3 protein level. We further demonstrated this correlation by luciferase reporter assay, which showed direct binding between miR-19b and the SOCS3 3'-UTR. We further confirmed that miR-19b directly inhibited SOCS3 post-transcriptional expression by over-expressing or knocking down miR-19b in Caco2 cells and HT29 cells. Our results showed a clear inverse relationship between the cellular miR-19b level and SOCS3 protein level in these two cell lines. We also demonstrated that miR-19b was expressed in both IECs and non-epithelial cells in intestinal tissue by miRNA in situ hybridization. Together, these results indicate that miR-19b is a negative regulator of SOCS3 expression in IECs.
SOCS3-related miRNAs (miR-30a, miR-30e and miR-19b) are upregulated in atherosclerotic tissues in the mouse arteriosclerotic vascular disease (ASVD) model 27 . Previous worked showed that miR-19b negatively regulates SOCS3 expression in atherosclerotic tissue, and our study demonstrated this phenomenon in inflammatory intestinal tissue.
Some studies have reported that SOCS3 expression is specifically increased in inflamed colonic areas in UC and CD patients. Indeed, SOCS3 contributes to enhanced IEC vulnerability during UC remission, suggesting that SOCS3 may be a useful therapeutic target for IBD clinical monitoring and early mucosal healing 28 . Our results also demonstrate that SOCS3 expression is upregulated in inflamed areas of active CD, which is consistent with previous reports.
Finally, we demonstrated that SOCS3 inhibition via miR-19b overexpression promoted MIP-3α production in vitro. Specifically, overexpression of miR-19b downregulated SOCS3 and consequently increased MIP-3α , whereas miR-19b knockdown induced SOCS3 expression and inhibited MIP-3α production. MIP-3α is a CC chemokine and is predominantly expressed by small intestine and colon epithelial cells 29 . MIP-3α plays an important homeostatic function in the intestine by regulating tissue turnover and epithelial maintenance 30 , and it is possible that MIP-3α also participates in colitis development.
Because our results established miR-19b as a suppressor of inflammation in active CD, we next investigated the possible therapeutic effect of miR-19b's in vivo. We used polyetherimide to load pre-miR-19b into IECs. Compared to the control groups, pre-miR-19b treatment reduced colitis-associated weight loss and the DAI and improved histological signs of inflammation. We also found that miR-19b expression was decreased, while SOCS3 expression was increased, in TNBS-induced colitis, and this finding is similar to that seen in CD patients. Furthermore, the expression of SOCS3 was inhibited in TNBS-induced colitis model after treatment with pre-miR-19b. Together, these results demonstrate that miR-19b significantly decreased inflammation in TNBS-induced colitis, which may be due to miR-19b's targeted regulation of SOCS3.
In summary, miR-19b downregulation and SOCS3 protein upregulation may serve as a molecular signature of CD. Specifically, our results show that miR-19b suppresses its target SOCS3 to regulate MIP-3α production and mediate intestinal epithelial homeostatic function, and these alterations may influence mucosal repair following intestinal inflammation.
It should be noted that miR-19b may have additional, non-epithelial effects. For example, the study by Liu et al. suggests that miR-17 and miR-19b are the two miRNAs in the miR-17-92 cluster responsible for promoting Th17 responses 31 . Our results from miRNA in situ hybridization show that miR-19b is expressed in both IECs and non-epithelial cells in intestinal tissue; thus, both epithelial and non-epithelial miR-19b-SOCS3 signals may be important for colitis development. However, the association between miR-19b and T cells requires further investigation.

Materials and Methods
All methods were carried out in accordance with the approved guidelines.
Human tissue. The Ethics Committee of Nanjing Medical University approved all study aspects. All patients and normal control individuals provided informed consent prior to the procedure. Intestinal tissue samples were obtained by pinch biopsy during endoscopic examination of active CD patients or normal control subjects undergoing screening colonoscopies at the First Affiliated Hospital of Nanjing Medical University (Nanjing, China). Tissue samples were immediately frozen in liquid nitrogen following pinch biopsies and stored at -80°C. The clinical features of the patients are listed in Table 1. Reactions were incubated in a 96-well optical plate at 95 °C for 5 min, followed by 40 cycles at 95 °C for 15 s and 60 °C for 1 min. All reactions, including controls without template, were run in triplicate. Cycle threshold (CT) data were determined using fixed threshold settings, and the mean CT value was determined from triplicate PCRs. A comparative CT method was used to compare each condition to control reactions. U6 snRNA served as an internal control, and the relative amount of miRNA was normalized to U6 using the 2 -ΔΔCT equation, in which Δ Δ CT = (C T miR-19b -C T U6 ) target -(C T miR-19b -C T U6 ) control.
To quantify SOCS3 and GAPDH mRNA expression, 1 μ g total RNA was reverse-transcribed into cDNA using oligo(dT) 18 primers (TaKaRa) and ThermoScript reverse transcriptase (Invitrogen). The reaction conditions were as follows: 16 °C for 30 min, 42 °C for 30 min, and 85 °C for 5 min. SYBR Green dye (Invitrogen) was added to the reverse transcription (RT) product, and real-time PCR was performed.
PCR was performed in a 20 μ L reaction containing 1 μ L cDNA, 1X QuantiTect SYBR green PCR Master Mix (Applied Biosystems), and 0.5 μ M sense and antisense primers. The primer sequences were as follows: Western blot analysis. Cold RIPA lysis buffer (Beyotime, China), supplemented with a protease and phosphatase inhibitor cocktail (Thermo Scientific, USA), was used to solubilize tissues and cells on ice for 30 min. Then, cell lysates or tissue homogenates were centrifuged for 15 min (12,000 g, 4 °C). The supernatant was collected, and the protein concentration was calculated with a BCA protein assay kit (Thermo Scientific, USA). A total of 60 μ g protein from cell or tissue was separated by 10% SDS-PAGE and transferred to a polyvinylidene difluoride (PVDF) membrane (Bio-Rad Laboratories, USA), which was then blocked in 5% nonfat dried milk with PBS at room temperature for 2 h. The membrane was probed overnight at 4 °C with rabbit anti-SOCS3 antibody (1:1,000; Abcam, USA) or mouse anti-α -tubulin antibody (1:1,000, Santa Cruz Biotechnology, USA). Then, the membrane was incubated with goat-anti-rabbit or goat-anti-mouse secondary antibody, respectively (1:4,000, Bioworld Technology, China), for 60 min at room temperature. Enhanced chemiluminescence reagent was used to detected the signal on the membrane (Thermo Scientific, USA). Data were analysed by densitometry using ImageJ 1.42 software and normalized to α -tubulin expression. The blots shown in the figures were cropped for better presentation. The original whole-gel blots can be found in the supplementary data.
Immunohistochemistry. Immunohistochemistry was performed on 5-μ m sections of paraffin-embedded intestinal tissue specimens as described previously 32 . Briefly, the tissues were fixed in 4% paraformaldehyde and washed in 0.01 M PBS. Isotype-matched mAb (Iso, anti-rabbit IgG mAb, Equitech-Bio, USA) with no specific affinity to interest proteins was used as the negative control. Antibodies for SOCS3 (Abcam, USA), p-STAT3 (Cell Signaling Technology, USA) and STAT3 (Cell Signaling Technology, USA), as well as isotype control mAbs, were diluted in plain phosphate-buffered saline (PBS) at 25 μ g/mL and immobilized on the tissue slides at 4 °C overnight, as previously described 33 .
The average integrated optical density (IOD) of SOCS3 IHC staining in CD and normal control tissues was calculated with Image-Pro Plus software (version 5.0) as previously reported 34 . Five fields were analysed on each slide 35 . miRNA in situ hybridization. A  Plasmid construction and luciferase reporter assay. A luciferase reporter assay was performed as previously described 39 to test miR-19b binding to its target gene, human SOCS3. The entire human SOCS3 3'-UTR, containing a presumptive miR-19b complementary site (seed sequence, UUUGCAC), was PCR-amplified with the following primers: sense: 5'-TGGGAGCTCAATGTCAGCCCAGTAAGTATTGGCCAGT-3'; antisense: 5'-GATAAGCTT-GTGCTCTTTATTATAAATTACTGAAATGTTTC-3' . Human genomic DNA was used as a template. PCR products were cloned into the pMIR-REPORT plasmid (Ambion), and insertion was confirmed by sequencing. To test binding specificity, the miR-19b seed sequence was mutated from UUUGCAC to AAACGTG. For luciferase reporter assays, 2 μ g firefly luciferase reporter plasmid, 2 μ g β -galactosidase (β -gal) expression vector (Ambion), and 100 pmol pre-miR-19b (GenePharma, Shanghai, China), anti-miR-19b (GenePharma), and scrambled ncRNA (GenePharma) were transfected into Caco2 cells or HT29 cells in 6-well plates using Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer's instructions. The β -gal vector was used as a control. Twenty-four hours after transfection, luciferase assays were performed (Promega, Madison, WI, USA). The reported data represent three independent experiments. miRNA overexpression and knockdown. miR-19b was overexpressed using pre-miR-19b, whereas knockdown was achieved with anti-miR-19b. Synthetic pre-miR-19b and anti-miR-19b and scrambled negative control RNA were purchased from GenePharma (Shanghai, China). Caco2 cells or HT29 cells were seeded in 6-well plates and transfected with Lipofectamine 2000 (Invitrogen) when the cells reached 70% confluence. In miR-19b overexpression experiments, 100 pmol of pre-miR-19b or ncRNA was added. For miR-19b knockdown, 100 pmol of anti-miR-19b was used. After 6 h, the medium was changed to DMEM or McCoy's 5 A supplemented with 1% FBS. Cells were harvested 24 h or 48 h after transfection. The miR-19b expression level was analysed by quantitative RT-PCR, and the SOCS3 protein level was assessed by Western blot. Protein samples were normalized against α -tubulin. SOCS3 overexpression and knockdown. A mammalian expression plasmid containing the human SOCS3 open reading frame without a 3'-UTR was purchased from Invitrogen. An empty plasmid was used as a negative control. Depletion of SOCS3 activity was accomplished by transfecting Caco2 cells and HT29 cells with small interfering RNA (siRNA). siRNA sequences targeting SOCS3 were designed and purchased from GenePharma (Shanghai, China). A scrambled siRNA that did not target human SOCS3 was included as a negative control. The SOCS3 overexpression plasmid or siRNA was transfected into Caco2 cells or HT29 cells using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions.
Cytokine analysis. Cytokine chips were used to detect cytokines in Caco2 cell culture supernatants according to the manufacturer's protocol. MIP-3α in culture supernatants from human Caco2 cells using a commercial ELISA kit (Lianshuo Biological, Shanghai, China) according to the manufacturer's instructions.
For the expression of MIP-3α in the TNBS-induced colitis model, full-thickness colonic tissue specimens of mice were homogenised in PBS containing a cocktail of protease inhibitors (KeyGEN Bio TECH, Nanjing, China) supplemented with 1 mM PMSF. After centrifugation at 12,000 g for 10 min at 4 °C, the total protein level of the supernatants was determined with the BCA protein assay (Thermo Scientific, Rockford, IL, USA). To measure MIP-3α , a mouse double-antibody sandwich ELISA kit (R&D Systems, USA) was developed as described 40 . The absorbance was measured at 450 nm and compared with the respective standard curve of the cytokines. Establishment of experimental colitis. All animal experiments were performed in accordance with the National Institute of Health's Guide for the Care and Use of Laboratory Animals and were approved by the Animal Care and Use Committee of Nanjing Medical University (Nanjing, China).
Female BALB/c mice (6-8 weeks old) were obtained from the Laboratory Animal Center of Nanjing University (Nanjing, China). All animals received care according to Chinese legal requirements. TNBS-induced acute colitis in BALB/c mice was established as previously reported 41 . Briefly, 2.5 mg TNBS (Sigma-Aldrich, St. Louis, MO, USA) in 100 μ L 50% ethanol was slowly administered into the colon lumen via a catheter inserted 4 cm into the colon through the anus. Mice were kept in a vertical position for 30 s. Control mice received 50% ethanol in saline alone. Mice were monitored daily for bleeding, body weight and stool consistency. Mice were sacrificed three days after colitis induction. Disease activity was evaluated according to the DAI 42 . Histological scores of H&E sections were graded from 0 to 4 according to previous reports 41,42 . Treatment. Polyetherimide (PEI, 25 kDa; Sigma)/pre-miRNA or pre-scramble complexes were prepared by mixing an aqueous PEI solution (4 mg/ml) with an equal volume of miRNA precursor or control solutions (2 mg/ml). A total volume of 100 μ l of this solution was administered to mice via rectal enema. TNBS-treated mice were divided into three groups: PEI/pre-miR-19b treatment, PEI/pre-scramble treatment, and PEI alone. Twelve hours after TNBS injection, 5 mg pre-miRNA or pre-scramble/kg body weight was administered 43 . Twenty-four hours after intracolonic administration with the Cy3-labeled miRNA precursor (GenePharma, Shanghai, China)/PEI complex, frozen colon sections were collected for immunofluorescence staining to analyse miRNA precursor localization. Following treatment, colon lengths were measured, and histopathological analysis was performed blindly. The DAI was used to evaluate intestinal inflammation severity based on previously published literature 42 . The histological score was evaluated based on H&E staining, as previously described 41,42 . Additionally, intestinal tissues samples were evaluated by performing Western blot, mRNA expression analysis, ELISA and miRNA quantification, as described above.
Statistical analysis. All images are representative of at least three independent experiments. Each qRT-PCR, luciferase reporter assay and cell viability assay was performed in triplicate, and each experiment was repeated several times. The results are presented as the mean ± SEM. P-values were evaluated with the two-tailed Student's t-test using SPSS Statistics v20.0 (IBM), and p < 0.05 was considered significant.