miR-511 Deficiency Protects Mice from Experimental Colitis by Reducing TLR3 and TLR4 Responses via WD Repeat and FYVE-Domain-Containing Protein 1

Antimicrobial responses play an important role in maintaining intestinal heath. Recently we reported that miR-511 may regulate TLR4 responses leading to enhanced intestinal inflammation. However, the exact mechanism remained unclear. In this study we investigated the effect of miR-511 deficiency on anti-microbial responses and DSS-induced intestinal inflammation. miR-511-deficient mice were protected from DSS-induced colitis as shown by significantly lower disease activity index, weight loss and histology scores in the miR-511-deficient group. Furthermore, reduced inflammatory cytokine responses were observed in colons of miR-511 deficient mice. In vitro studies with bone marrow-derived M2 macrophages showed reduced TLR3 and TLR4 responses in miR-511-deficient macrophages compared to WT macrophages. Subsequent RNA sequencing revealed Wdfy1 as the potential miR-511 target. WDFY1 deficiency is related to impaired TLR3/TLR4 immune responses and the expression was downregulated in miR-511-deficient macrophages and colons. Together, this study shows that miR-511 is involved in the regulation of intestinal inflammation through downstream regulation of TLR3 and TLR4 responses via Wdfy1.


Introduction
MicroRNA's (miRNAs) are short (18)(19)(20)(21)(22)(23)(24) nucleotides in length), endogenous, noncoding single-stranded RNA's that regulate gene expression by controlling stability and translation of protein-coding mRNA [1]. Post-transcriptional regulation by miRNAs affects many processes, including cell development, cell differentiation, tissue homeostasis, and activation of the immune system [2][3][4]. miR-511 is a miRNA embedded within intron 5 of the CD206 gene (MRC1), referred to as the macrophage mannose receptor [5,6]. After transcription and cleavage, the pre-miRNA is formed, which is transported to the cytoplasm where through Mice aged between 10 to 14 weeks were used for DSS colitis experiments. Male and female mice were housed in separate cages and they were distributed in a random manner between groups. Two DSS colitis experiments were conducted to assess clinical scores and for the isolation of intestinal cells. 2% (w/v) DSS (TdB Consultancy, Uppsala, Sweden) was added to the drinking water (Innovive, Thermo Fisher Scientific, Leiden, The Netherlands) and refreshed every day for 7 days. Bodyweights were recorded daily and mice were euthanized with CO 2 at the end of the experiment. Colon lengths and wet colon weights were recorded. Colon weight per cm was used as a disease parameter. Disease activity index was determined based on body weight loss score, where a score of 0 indicates less than 1% weight loss, 1 = 1-5% weight loss, 2 = 5-10%, 3 = 10-15% and the highest score of 4 is given to a weight loss of more than 15%, stool consistency score (0 = normal consistency, 1 = loose droppings, 2 = loose stool, colon filled with feces, 3 = loose stool, feces near caecum, 4 = empty bowel) and stool bleeding scores ranging from 0 which indicates no bleeding to a score of 4 for grossly/bloody stool [27]. Colons were longitudinally divided in two parts, where one part was used for histology and the other part was snap frozen to isolate RNA.

mRNA, cDNA and qPCR Analysis
Colonic tissues were homogenized in Tripure isolation reagent with stainless steel beads (Qiagen Benelux B.V, Venlo, The Netherlands) for 15 min using a tissue homogenizer (Qiagen Benelux B.V). mRNA was then isolated according to the manufacturer's protocols. Bioline ISOLATE II mini kit was used for RNA clean up in accordance to manufacturer's protocol.

Cytokine Measurements
In supernatant of BMDM stimulations, protein concentrations of TNF-α, IL-6, and IL-10 were measured with a mouse inflammation kit, BD Cytometric Bead Assay (CBA; BD Bioscience, Vianen, The Netherlands) according to manufacturer's instructions, with the exception that samples were diluted 3 to 4 times. Samples were acquired on the flow cytometer, LSR Fortessa.

RNA Sequencing of BMDMs
RNA quality was assured using the Agilent 2200 Tapestation, using only samples with a RIN score of ≥8. mRNA was isolated and converted into cDNA with the KAPA mRNA HyperPrep Kit (Roche Diagnostic Nederland BV, Almere, The Netherlands), whereupon the cDNA was prepared for sequencing on the HiSeq4000 at the Core Facility Genomics, Amsterdam UMC in a 50 bp single-ended fashion to a depth of 40 M per sample. The raw reads were checked for quality using FastQC (v0.11.15), trimmed for adapter sequences using Trimmomatic v0.32 and aligned to the mus musculus genome (GRCm38v93) using HISAT2 (v2.1.0). Counts were obtained using HTSeq using annotation fromEnsembl v94 [33]. Duplication rates were inspected using dupRadar v1.0.0. Genes with more than 2 count-per-million reads in 2 or more of the samples were kept. For normalization, the weighted trimmed mean of M-values (to the reference) is used (TMM (edgeR)) [34]. Genes were reannotated using BiomaRt and Ensembl (v100). The count data was transformed to log2-counts per million (logCPM) using voom, estimating the mean-variance relationship. Differential expression was assessed using a Bayes moderated t-test using the linear model framework from the limma package [35] using the Benjamini-Hochberg false discovery rate to correct for multiple testing of the resulting p-values. Homologene v68 was used to map the Entrez Gene IDS from mus musculus to human to be able to perform a gene set enrichment analysis against the human genesets from MSigDB v7.0 Cells 2022, 11, 58 6 of 17 (H,C1,C2,C3,C5,C6,C7) [36,37]. For this gene set enrichment analysis we used the CAM-ERA [38] function from the limma package, with inter.gene.corr = 0.01. Analysis was performed using Rv4.0.0 and Bioconductor v3.11 [39,40].

Immunohistochemistry
Immunohistochemistry was done to visualize WDFY1 protein in colon sections of WT and miR-511-deficient control and DSS induced mice and in human. Non-IBD control (colorectal carcinoma), IBD (Crohn's) inflamed and non-inflamed colon sections from male adult participants were used. The specimen collection was approved by the biobank review committee under biobank number 178#A201470 and written informed consent was obtained from all study participants. The paraffin embedded sections were deparaffinized in xylene and gradually rehydrated in (v/v) ethanol gradient to PBS followed by antigen retrieval in sodium citrate buffer, pH 6.0 for 20 min at 98 • C and 30 min blocking in 10% normal goat serum (R&D Systems, Abingdon, UK). The slides were incubated overnight at 4 • C with primary antibody (rabbit polyclonal anti-WDFY1, 1:400) in blocking buffer. The slides were blocked in 3% (v/v) hydrogen peroxidase (H 2 O 2 ) in PBS for 20 min followed by a 30 min incubation with secondary antibody (anti-rabbit polyclonal HRP, Brightvision BV, Almere, The Netherlands). Staining was visualized after a 4 min incubation with chromagen substrate diaminobenzidine liquid plus (DAB, Dako Netherlands BV). The sections were finally counter stained with hematoxylin and mounted. Images were processed using the Olympus BX51 microscope.

miR-511 Deficiency Affects Macrophage Responses to Microbial Stimuli
Previously we suggested that miR-511-3p affects macrophage TLR4 responses and ameliorates intestinal inflammation [16]. We further investigated the exact effect of miR-511 deficiency on intestinal inflammation using miR-511-deficient mice. As expected miR-511deficient mice lacked miR-511-3p in BMDMs and liver tissue (Figure 1a). Colons of miR-511deficient mice showed no infiltration of granulocytes/monocytes, ulceration, fibrosis etc. as shown by histology (Figure 1b). No differences were observed in relative mRNA expression of CD206, TNF-α, MCP-1, TLR4, and CXCL10 in colons of WT and miR-511-deficient mice ( Figure 1c). Interestingly, a slight but significant upregulation of IL-6 expression was seen in miR-511-deficient mice compared to WT (Figure 1c). Since we previously found that knockdown of miR-511-3p reduces macrophage lipopolysaccharide (LPS) responses we tested if macrophages from miR-511-deficient mice were also affected in their LPS responses. miR-511-3p is highly expressed in M2 derived macrophages [6] therefore, BMDM from miR-511-deficient and WT mice were cultured and M2 differentiated macrophages were stimulated with LPS ( Figure 1d). These macrophages showed a significant reduction in TNF-α and IL-6 response and no significant effect was seen on IL-10 response upon LPS stimulation compared to WT M2 macrophages (Figure 1d). In contrast, no significant effect was seen on TNF-α, IL-6 and IL10 mRNA levels of these macrophages (supplementary Figure S1a). Together, our data show that miR-511-deficient mice have healthy colons but macrophage function is affected.

miR-511-Deficient Mice Are Protected against DSS-Induced Colitis
To address the effect of miR-511 in intestinal inflammation, both WT and miR-511deficient mice received DSS for 7 days (Figure 2a). Following the induction of colitis, a significant reduction in body weight loss (Figure 2b), disease activity index (Figure 2e), and colitis histology index score (Figure 2f) was observed in miR-511-deficient mice compared to WT mice. However, no differences were observed between the groups in terms of spleen weight ( Figure 2c) and colon weight/length ratio (Figure 2d). Representative histology pictures of the colons of WT and miR-511-deficient mice under DSS conditions are shown in Figure 2g, which shows less inflammation in the miR-511-deficient mice comparatively. Also, a significant downregulation of colonic cytokine and chemokine levels (TNF-α, IL-6, MCP-1, and CXCL10) was observed in miR-511-deficient mice compared to WT mice ( Figure 2h). These data indicate that miR-511 deficiency ameliorates DSS-induced colitis in mice.

Changes in Colonic Monocyte and Maturing Monocyte Populations in miR-511-Deficient Mice in Inflamed Condition
To further investigate the effect of miR-511 deficiency on immune cell composition in the colon tissue, we next examined subpopulations of lamina propria mononuclear phagocytes and neutrophils in colons with and without inflammation. Using flow cytometry we distinguished CD11b+Ly6G+ neutrophils, CD11b+Ly6G-CD64-CD11c+ dendritic cells (DCs), CD11b+Ly6G-CD64+Ly6C+MHCIIa-monocytes, CD11b+Ly6G-CD64+Ly6C+MHCIIa+ maturing monocytes and CD11b+Ly6G-CD64+Ly6C-MHCIIa+ macrophages, through distinctive antibody panels as established earlier [41,42] (Figure 3a). As previously described [41,42], healthy colons compared to inflamed colons contained relatively more DCs and mature macrophages while in inflammatory conditions an increase in monocytes, maturing monocytes and neutrophils were seen (Figure 3b). In healthy colons, no significant differences were found in relative contribution of the analyzed populations between WT and miR-511-deficient mice (Figure 3b). This suggests that miR-511 deficiency does not affect resident myeloid populations in healthy state. Interestingly, in inflammatory conditions colonic mononuclear phagocyte populations were differently distributed in miR-511-deficient mice compared to WT control mice (Figure 3b). Comparatively, miR-511deficient mice showed lower percentage of monocytes and a higher percentage of maturing monocytes present in inflamed colons while no difference was seen for neutrophils, DCs and macrophage populations. The higher number of monocytes is most likely a reflection of differences in severity of colonic inflammation between miR-511-deficient and WT mice.    are expressed as mean and standard error of the mean. Statistical differences were tested by independent t-test, where a p-value of <0.05 was considered to be significant. * p-value < 0.05; ** p-value < 0.01 and *** p-value < 0.001.

miR-511 Deficiency Affects Wdfy1
In order to investigate potential miR-511 targets in macrophages, we screened for differentially expressed genes using RNA-seq. M2 macrophages were generated from WT and miR-511-deficient mice and groups with and without LPS stimulation were compared. To visualize the variation in expression between samples, a multidimensional scaling plot (MDS) was demonstrated, where it became apparent that miR-511 deficiency had little effect on macrophage gene expression (Figure 4a). As expected, LPS stimulation had a clear significant effect on macrophage gene expression in both WT and miR-511-deficient macrophages (Figure 4a). When comparing WT and miR-511-deficient macrophages only 11 genes were differentially expressed (Figure 4b,d) of which 3 genes were upregulated and 5 genes were downregulated in both groups (controls and LPS stimulated macrophages) whereas 3 genes were downregulated (growth differentiation factor 3, calcium/calmodulindependent protein kinase II beta and an unidentified gene 13339) in control groups alone as seen in the Venn diagram (Figure 4c). Furthermore, in the group of differential expressed genes, Wdfy1 was the only gene that was identified by TargetScan as a potential target of miR-511-3p in mice (fold change and BH-significance) (Figure 4e). WDFY1 is a protein involved in positive regulation of the TLR3 and TLR4 signaling pathways [20]. To investigate this, we next checked for colonic Wdfy1 gene expression and also observed a significant downregulation of Wdfy1 mRNA in colons of miR-511-deficient mice compared to WT (Figure 4f). Interestingly, DSS induced inflammatory conditions reduced Wdfy1 expression in both WT and miR-511-deficient mice (Figure 4f). Also, less Wdfy1 protein was detected in the miR-511 deficient mice compared to controls with no significant difference observed between the groups in DSS condition (supplementary Figure S2a). A representative immunohistochemistry picture is shown in supplementary Figure S2a. Moreover, WDFY1 protein was also detected in the colons of non-inflammatory bowel disease (IBD) controls and inflamed and non-inflamed colon section of IBD patients (supplementary Figure S2b). WDFY1 is localized to endosomal membrane as reported earlier [43], and interestingly, from our staining it seems that WDFY1 is present in both endosomal membrane and nucleus in mouse and human colonic crypts, although no clear distinction can be made in between the healthy and IBD colon sections. Together our data underscore that Wdfy1 is one of the targets of miR-511 involved in macrophage inflammatory responses.

Macrophages from miR-511-Deficient Mice Have Reduced Wdfy1 Expression, Which Affects Antimicrobial Responses
We further investigated Wdfy1 protein levels and anti-microbial responses in WT and miR-511-deficient macrophages. M2 bone marrow-derived macrophages were generated and as expected a significant reduction in Wdfy1 mRNA expression was observed in miR-511-deficient macrophages (Figure 5a). Surprisingly when determining Wdfy1 protein by western blotting, Wdfy1 protein was not detected in miR-511-deficient macrophages (Figure 5b). We had already found that LPS mediated cytokine responses were reduced in miR-511-deficient macrophages compared to WT macrophages (Figure 1d). To determine if responses via other pattern recognition receptors are also affected in these macrophages we next challenged macrophages with Polyinosinic:polycytidylic acid (Poly I:C) (TLR3 agonist) and Zymosan (dectin-1 and TLR2 agonists). Interestingly, Poly I:C induced TNF-α and IL-6 responses were reduced in miR-511-deficient macrophages compared to WT macrophages and miR-511 deficiency did not affect macrophage cytokine responses towards Zymosan (Figure 5c). Together these data suggest that miR-511 specifically regulates TLR3 and TLR4 responses via regulation of Wdfy1 but does not affect other pattern recognition receptors like TLR2 and dectin-1, which are involved in responses towards Zymosan [44,45].  Individual values are expressed as mean and standard error of mean. Statistical differences were tested by independent t-test, where a p-value of <0.05 was considered to be significant. ** p-value < 0.01.  Statistical differences were tested by independent t-test, where a p-value of <0.05 was considered to be significant. *** p-value < 0.001.

Discussion
miR-511 is embedded within intron region 5 of the CD206 gene (MRC1) [5,6]. Previously, we showed that CD206 deficient mice also lacked miR-511 and that this knockdown of miR-511 led to downregulation of TLR4 and reduced intestinal inflammation in these mice [16]. In this study, we confirm, as hypothesized before, that miR-511 deficiency protects against DSS colitis (Figure 2). Reduced colonic inflammation in miR-511 deficient mice coincided with a lower monocyte percentage and a higher maturing monocytes percentage in inflamed colons compared to WT (Figure 3b). Monocytes are recruited to sites of infection and inflammation where they mediate antimicrobial activity, thus playing a key role in defense against pathogens [46,47]. In our study, fewer monocyte populations were recruited in miR-511-deficient colons, which corroborates with reduced intestinal inflammation (Figure 3b).
This study is the first to identify Wdfy1 as a potential target for miR-511. Using RNAseq to identify potential targets on bone marrow-derived M2 macrophages, we found Wdfy1 to be one of the few differentially expressed genes between WT and miR-511-deficient macrophages (Figure 4). WDFY1 is a crucial adapter protein involved with positively regulating TLR 3 and 4 signaling pathways, by recruiting TIR-domain-containing adapterinducing interferon-β (TRIF) [20]. Furthermore, WDFY1 was also shown to play a role in autophagy [48,49] and Wdfy1 expression is also related to mitochondrial dysfunction in Alzheimer's disease of a mouse model [50]. For instance, an increase in WDFY1 levels followed by vascular endothelial growth factor-nontyrosine kinase receptor (VEGFC-NRP2) axis depletion induces cell death [48]. Although the function of WDFY1 still remains obscure, one study reported the role of Wdfy1 in innate immunity by comparing the immune response of WT and Wdfy1-deficient mice against TLR3 (Poly I:C) and TLR4 (LPS) agonists [21]. Wdfy1-deficient mice showed impaired inflammatory cytokines in BMDMs [21], which shows that mouse Wdfy1 is important for TLR3 and TLR4 signaling. In line with this, we first showed that the Wdfy1 gene was significantly downregulated in both colons of miR-511-deficient mice ( Figure 4f) and macrophages cultured from these mice (Figure 5a). Interestingly, Wdfy1 was not detected at protein level in these macrophages ( Figure 5b). Also, the mRNA expression of the gene in macrophages was low (Figure 5a) so it might indicate that miR-511 is protecting it against degradation and thus affecting stability of mRNA, post-translationally [51]. Besides reduced inflammatory responses upon TLR4 stimulation, miR-511-deficient macrophages also showed reduced TNF-α and IL-6 responses when stimulated with TLR3 agonist (Poly I:C) (Figure 5c). Upon Zymosan challenge, miR-511-deficient macrophages did not respond differently from WT macrophages, which suggests that signaling pathways downstream of Wdfy1 is affected while other pattern recognition remain intact (Figure 5c). Together this suggests that miR-511 deficiency causes a downregulation of Wdfy1 which affects TLR3 and TLR4 responses. However, to confirm a direct interaction between miR-511 and Wdfy1, experiments such as reporter gene assays or pull-down assays are needed.
We are the first to report a protective effect of miR-511 deficiency against DSS-induced inflammation in mice. Our data suggests miR-511 regulates Wdfy1 expression which subsequently affects TLR (3 and 4) responses. The distinct role of miR-511 in inflammation needs to be unraveled to understand disease pathogenesis and mechanisms in IBD better. Further studies are required to explore the function and interaction of miR-511 and WDFY1 in humans.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/10 .3390/cells11010058/s1, Figure S1: Relative mRNA expression of cytokines in BMDMs stimulated with LPS. (a) mRNA expression levels of TNF-α, IL-6 and IL-10 after 24 h of LPS stimulation. The levels of mRNA were normalized for reference genes. N = 5 WT and 6 miR-511-deficient mice. Individual values are expressed as mean and standard error of mean. Statistical differences were tested by independent t-test, where a p-value of <0.05 was considered to be significant. Figure S2: Detection of WDFY1 protein in colon sections of mouse and human IBD sections. (a) Detection of Wdfy1 by immunohistochemistry in colon tissue in WT and miR-511-deficient mice in control and DSS conditions, 10× magnification. (b) Detection of WDFY1 by immunohistochemistry in colon tissue sections in non-IBD healthy and IBD inflamed and non-inflamed colon sections, 10× magnification.