Fasciola hepatica hijacks host macrophage miRNA machinery to modulate early innate immune responses

Fasciola hepatica, a global worm parasite of humans and their livestock, regulates host innate immune responses within hours of infection. Host macrophages, essential to the first-line defence mechanisms, are quickly restricted in their ability to initiate a classic protective pro-inflammatory immune response. We found that macrophages from infected animals are enriched with parasite-derived micro(mi)RNAs. The most abundant of these miRNAs, fhe-miR-125b, is released by the parasite via exosomes and is homologous to a mammalian miRNA, hsa-miR-125b, that is known to regulate the activation of pro-inflammatory M1 macrophages. We show that the parasite fhe-miR-125b loads onto the mammalian Argonaut protein (Ago-2) within macrophages during infection and, therefore, propose that it mimics host miR-125b to negatively regulate the production of inflammatory cytokines. The hijacking of the miRNA machinery controlling innate cell function could be a fundamental mechanism by which worm parasites disarm the early immune responses of their host to ensure successful infection.


Results
Parasite-derived miRNAs are located within macrophages during infection with Fasciola hepatica. To investigate whether miRNAs produced by F. hepatica contributed to the inhibition of macrophage activation in early infection, we orally infected BALB/c mice with F. hepatica metacercariae and searched for parasite miRNAs in peritoneal macrophages at selected time-points (6 h, 12 h, 24 h, 3 days and 5 days after infection) using RNASeq (deposited in NCBI's GeneExpression Omnibus; accession number GSE145597; https:// www. ncbi. nlm. nih. gov/ geo/ query/ acc. cgi? acc= GSE14 5597). After first cleaning the datasets obtained, reads were mapped against the mouse database from miRBase (mmu). Then, to identify F. hepatica-specific miRNAs the remaining 15,796,193 reads were screened against the miRBase list of 30 F. hepatica miRNAs 31 . Using this approach, we identified 10 F. hepatica miRNAs that were present within peritoneal macrophages of infected mice (Table 1). F. hepatica miR-125b (fhe-miR-125b) was the most abundant miRNA observed in peritoneal macrophages during the first 24 h of infection, with expression peaking at approximately 12 h after infection. This is coincident with the migration of NEJs through the intestine and into the peritoneal cavity, and also with the observed early suppression of host innate immune responses, as described above 10,13,14 . This rapid and transient expression pattern of fhe-miR-125b in peritoneal macrophages was corroborated by RT-qPCR (Fig. 1a). Consistent with these observations, the expression of fhe-miR-125b was also observed in peritoneal cells isolated from sheep after an experimental infection with 150 metacercariae (Fig. 1b). These miRNAs were not detected in peritoneal macrophages taken from non-infected mice or sheep (Fig. 1a,b).
To verify that the primers were specifically detecting a miRNA of parasite origin, infectious metacercaraie were experimentally excysted in vitro and the RNA of the emerging newly excysted juvenile (NEJ) subjected to Table 1. Read numbers of known Fasciola hepatica miRNAs detected by miRNASeq in peritoneal macrophages harvested from infected mice over a time-course of 5 days. a Two biological samples were sequenced at each time point with read counts for each sample presented in columns below. b "-" no read counts detected.  -2  1  --1  ----fhe-miR-125b  --1  16  6  16  1  2  -1  -fhe-miR-36a  --1  --------fhe-miR-61  ---3  1  5  -3  ---fhe-miR-125a  ---1  --------fhe-miR-46 www.nature.com/scientificreports/ RT-qPCR. This confirmed that fhe-miR-125b is expressed by the NEJ parasites (Fig. 1c). RT-qPCR also demonstrated that miR-125b could be amplified from RNA isolated from Schistosoma mansoni, a trematode parasite like F. hepatica, but not from the free-living nematode Caenorhabditis elegans. This result concurs with reports demonstrating that the expression of a miR-125b is specific to trematode parasites 32 . Comparison of the expression of fhe-miR-125b across the different intra-mammalian life-stages of F. hepatica revealed strict temporal expression, supporting the idea that worm miRNAs govern developmental processes. In the context of a role in the immune modulation of innate immune responses after infection, it was notable that the expression of fhe-miR-125b was most abundant in NEJs, the stage that initiate infection. Several studies suggest that extracellular vesicles (EVs) are the main reservoir of exogenous miRNAs. In this study, EVs were harvested from in vitro cultured NEJs and shown to contain fhe-miR-125b (Fig. 1d). Since EVs released during helminth infections are primarily taken up by macrophages 26,27 this provides a mechanism by which fhe-miR-125b and other miRNAs are delivered into the macrophages.
Fasciola hepatica miR-125b targets innate immune signalling pathways. The secondary structure of fhe-miR-125b indicates a typical miRNA with the mature strand located on the 5′ arm of the precursor (Fig. 2a). Previous studies have documented human hsa-miR-125b as a miRNA that regulates the inflammatory response of macrophages by controlling the activation of M1 macrophages [33][34][35] . Using a phylogenetic approach, we compared the structure of fhe-miR-125b to its human homologue and to miR-125 sequences from the closely related trematodes, Schistosoma japonicum and mansoni (Fig. 2b). This analysis revealed that miR-125b in Fasciola is closely related to the human miRNA, but is distinct from miR-125b found in the Schistosome parasites. This may suggest fhe-miR-125b could target the same host genes as hsa-miR-125b and thus play a significant role in the control of host immune responses during early infection.
To address this possibility experimentally, a list of putative cellular targets for mature fhe-miR-125b, (CCC CUG AGA CUG AUA AUU GCUC) was initially generated using the online miRNA target prediction tool miRDB (https:// mirdb. org). Only those genes assigned a target score > 60 (a predictive score generated by MirTarget 36 ) were regarded as authentic and selected for further analysis. The resultant 511 genes (Supplementary Table 1) were mapped for protein-protein interactions using the innate immune data reference set 37 and provided 798 immune-related protein-protein interactions with an enrichment in the nucleus (Fig. 2c). This dataset was then filtered down to cell specific associations restricted to peritoneal macrophages. Surprisingly, from this analysis, only two gene targets emerged; Traf6 and Nlrc5. Pathway analysis of the interactions showed that Traf6-mediated events were over-represented (Fig. 2e). To further scrutinise if Traf6 could be the prime target candidate for fhe-miR-125b, two additional predictive tools were then used, miRanda and TargetScan (Supplementary Table 1). Supporting the miRDB initial analysis, TRAF6 was one of only 43 genes that were commonly predicted as targets for fhe-miR-125b by all three tools. Furthermore, the significant reduction in the expression of TRAF6 in peritoneal macrophages harvested from mice at the same timepoints that fhe-miR-125b is most abundant (Fig. 2f), provides experimental evidence to validate the predictive analysis. As a transcription factor that is central to (c) The expression of fhe-miR125b was measured in RNA extracted from the F. hepatica newly excysted juveniles (NEJ), 21-day old worms (immature) and adult worms and compared to expression levels in adult (male and female mixed population) Schistosoma mansoni and Caenorhabditis elegans. (d) RT-qPCR using primers specific for fhe-miR125b was carried out on RNA extracted from NEJ extracellular vesicles (EVs). In these experiments, levels of fhe-miR125b expression determined by RT-qPCR are represented as the starting quantity of genetic material prior to amplification (N0) as determined by LinRegPCR (v.11). The data is presented as the mean ± SD of triplicate biological samples. A central role for fhe-miR-125b in the regulation of these signalling pathways was also corroborated by expression profiling of macrophages harvested from the peritoneal cavity of mice 18 h after an oral infection with F. hepatica. Of the 610 genes that were identified to be significantly (p > 0.05) downregulated (> twofold change in expression), a large number of these were pro-inflammatory/innate cytokines and chemokines or transcription factors that regulate cytokine-signalling pathways (Supplementary Table 3). This is illustrated in the global KEGG analysis of these genes which created a profile suggesting involvement predominantly in MAPK, JAK-STAT, TNF and TLR signalling (Supplementary Fig. 1) and which was almost identical to that predicted for fhe-miR125b using GO tool DAVID above (Fig. 2).
To provide experimental proof of the ability of fhe-miR-125b to directly engage in the regulation of proinflammatory pathways in macrophages, primary murine macrophages were transfected with synthetic versions of both the F. hepatica and human miR-125b. While neither miRNA had any apparent effect on macrophages when transfected under basal culturing conditions, both transfections significantly inhibited the ability of macrophages to express TRAF6 in response to a challenge with bacterial LPS (Fig. 3a) and, as a consequence, the production of both IL-6 and TNF were significantly reduced (Fig. 3b,c). This outcome was specific to the miR-125b as transfection with a scrambled non-targeting miRNA had no effect. We propose, therefore, that the F. hepatica fhe-miR-125b, like its mammalian counterpart, functions to regulate the ability of macrophages to respond to pro-inflammatory signalling, thus preventing the differentiation to an M1 phenotype.
Fasciola hepatica miR-125b complexes with the mammalian Argonaute-2 protein in host macrophages to regulate mammalian gene expression. The mechanism(s) by which the exogenous worm miRNAs operate in mammalian immune cells to alter their function remains enigmatic. However, studies in mammalian cells have shown that in the first step miRNAs must form a complex with an Argonaute protein (Ago). This Ago-miRNA association forms the core of the miRNA-induced silencing complex (miRISC), which is then guided by the complexed miRNA to bind complementary (target) sequences in the messenger RNA (mRNA) 40 . Helminth Ago proteins have been detected within secreted EVs and may be involved in the regulation of mammalian target genes 26 . However, given the similarity of fhe-miR-125b to its mammalian homolog we surmised that helminth miRNAs delivered into immune cells could directly load onto mammalian Ago2. In this way worm miRNAs would compete with the mammalian miRNA machinery to influence host gene expression. To test this idea, murine Ago2 was isolated by immunoprecipitation from the peritoneal macrophages taken from infected mice at 6 h and 12 h post infection (IP; Supplementary Fig. 2) and the miRNA extracted and subjected to RT-qPCR. We used the detection of murine miR-21 as a control for the pulldown and to demonstrate the normalisation of input miRNA to the RT-qPCR. This miRNA has been characterised as a central regulator of macrophage activity and its expression is enhanced by inflammatory stimuli [41][42][43] . Therefore, the observation of this murine miRNA loaded onto the Ago-2 within peritoneal macrophages of infected mice at both 6 h and 12 h, but not in uninfected animals (Fig. 3d) supports the efficacy of the pulldowns and validates the normalisation of miRNA input to the RT-qPCR. In contrast, the parasite-derived mir-125 was loaded onto Ago-2 only at 12 h timepoint (a time when fhe-miR-125b expression was at its peak; Table 1) reflecting the changes to the loading of mir-125b as a result of the progression of parasite infection (Fig. 3d).

Discussion
On invasion of the host by the parasite F. hepatica protective pro-inflammatory cytokine expression in macrophages is quickly blocked. Thus, a major player in the innate cell immune system is inactivated at a stage when the parasite is most vulnerable and depends on its glycogen stores for energy to power its way from the intestine to the liver. Our study suggests that this trematode parasite, and perhaps others, exploit homologs of the immune-regulatory miRNAs of the host to hijack macrophage signalling pathways to disarm their function. The hairpin structure of fhe-miR125b, produced by RNAfold (http:// rna. tbi. univie. ac. at/ cgi-bin/ RNAWe bSuite/ RNAfo ld. cgi) (b) Phylogram of miR125 generated using T-Coffee (http:// tcoff ee. crg. cat), based on the hairpin structures. (c) Illustration of a Protein-Protein Interaction map for the 798 immune-related genes predicted from gene target analysis (created by Innate DB; https:// www. innat edb. com). Of these, the nucleus showed the highest number of interactions account for greater than 50% of all the interactions. (d) GO analysis of pathways represented by the predicted gene targets for fhe-miR-125b, produced using bioconductor (https:// www. bioco nduct or. org). (e) Mapping predicted gene targets to a mouse interactome produced 66 nodes (image created by Innate DB; https:// www. innat edb. com). Within these, Traf6 (red box) mapped to 35-40% of the nodes (Blue dots). All other nodes not connected to Traf6 are shown in yellow and red. A detailed map of these interactions can be viewed in supplementary Fig. 3. (f) The expression of TRAF6 was measured in RNA extracted from peritoneal macrophages harvested from uninfected mice (Uninf) or from mice at 6 h and 12 h after an oral infection with F. hepatica. The level of expression is represented as the starting quantity of genetic material prior to amplification (N0) as determined by LinRegPCR (v.11). The data is representative of three independent experiments and is presented as the mean ± SD of triplicate biological samples. The data was analyzed using oneway analysis of variance (ANOVA). www.nature.com/scientificreports/ Several parasitic worm miRNAs share homology with mammalian miRNAs. Thus, they have the capacity to regulate the expression of an array of host genes within macrophages and other innate cells to effectively evade expulsion and continue their growth and development to maturity. Our gene target analysis and in vitro studies pinpoint a specific mechanism whereby F. hepatica could interrupt the MAPK signalling pathway of innate immune responses via the regulation of TRAF6 by fhe-miR-125b. This mechanism is supported by the presence of a single miR-125b binding site on the 3′ UTR of TRAF6 44 . The identification of mammalian genes and immune pathways targeted by parasite miRNAs has major implications for our understanding of the interaction between parasite and host.
It is highly significant that miR-125b, the most predominant parasite miRNA found in host macrophages of mice infected with F. hepatica, is conserved amongst the trematode clade of helminths, including the blood flukes of the genus Schistosoma, as it points towards a common mechanism for the anti-inflammatory control amongst these important parasites of humans and their livestock. Indeed, recent studies have shown that the miR-125b orthologue is highly enriched in EVs released by S. mansoni and S. japonicum, representing 40% and 65% respectively of the total miRNA reads 27,45 . However, in contrast to the immune modulatory effect of fhe-miR-125b we have shown here, transfection of murine macrophages with sja-miR-125b resulted in the induction of pro-inflammatory immune responses as assessed by an increased production of TNF 27 . Predictive analysis of this parasite miRNA led Liu et al. 27 to suggest that the sja-miR-125b was targeting the expression of Pros1, a gene known to inhibit TLR-driven pro-inflammatory responses. Neither of the three predictive tools used in our study, predicted this gene as a target for the F. hepatica miR-125b, nor was it identified amongst the genes downregulated in macrophages isolated from infected mice. These contrasting observations could suggest that miR-125b plays different functions for different parasites. Unlike F. hepatica, a pro-inflammatory Th1-type immune response dominates during the first 3-5 weeks of infection with Schistosome parasites. As the parasites mature and females become fecund, parasite eggs trapped in tissues induce potent Th2 responses that suppress Th1 responses 46 . The sja-miR-125b was identified within EVs isolated from parasites collected from mice at 28 days post-infection 27 , a pre-patent time point, coincident with a strong Th1 pro-inflammatory immune response.
A single miRNA can influence the expression of a multitude of different genes. mRNA silencing efficiency is determined by the sequence of the seed region, free energy produced by the miRNA-target duplex and stability of the miRNA 5′ terminal and seed region 47 . While high specificity of the miRNA seed sequence has been largely reported as the crucial factor to miRNA target recognition, nucleotide pairings outside of the seed are also known to affect gene target recognition and overall thermodynamic stability of the miRNA-target duplex 48 .
These key concepts are integrated in predictive tools such as miRDB and MiRanda that scrutinise whole mature miRNA sequences. Examining the overall structure of miR-125 from various species, phylogenetic analysis indicates that fhe-miR-125b is closely related to its human counterpart and distal to that of miR-125b found in Schistosome. This may explain why Pros1 was not predicted to be a target for fhe-miR-125b and only regulated by sja-miR-125b. It would be interesting to determine which exact nucleotides outside the seed region may account for this specificity. miRNAs are not functional until they are loaded onto an Ago protein, a process that facilitates base-pairing to mRNA targets and ultimately translational repression and mRNA degradation. Here we provide proof that exogenous, worm-derived miRNA is delivered into macrophages, probably via exosomes, during infection and loads onto host Ago2. Coupled with our in vitro transfection studies that shows fhe-miR-125b can alter macrophage phenotype we can deduce that parasite miRNAs are functional within mammalian cells. We have to assume that this binding mimics that of host hsa-miR-125b and, therefore, operates in the same way, to regulate the pro-inflammatory response of macrophages [33][34][35] .
It has been reported that miRNAs that have homologs in mammalian hosts dominate the miRNA cargo of F. hepatica EVs. However, a number of miRNAs which are novel to the parasite have also been identified here and elsewhere 23 . Predictive analysis suggests that these too have the capacity to regulate immune-related mammalian genes 23,25 . As the loading of miRNAs to Ago protein is a very precise process with the requisite thermodynamics and base-match composition, it remains to be determined whether homology to a mammalian miRNA is essential for a parasite miRNA to associate with mammalian Ago and annex the mammalian miRNA machinery.
The ability of F. hepatica to invade, infect and produce progeny in a very broad range of species indicates a strong universal compatibility between the parasite and its mammalian host, particularly in terms of attuning the immune response. This study provides direct in vitro and in vivo evidence for the role of helminth-derived Figure 3. TRAF6 signaling and downstream production of IL-6 and TNF by macrophages is suppressed by fhe-miR-125b by binding the mammalian argonaute protein-2. Bone marrow derived macrophages were transfected with the parasite-derived fhe-miR-125b, the human miR-125b or a scrambled miRNA (20 nM) and stimulated 24 h later with bacterial lipopolysaccharide (LPS; 10 ng/ml) for 6 h. (a) The expression of TRAF6 was measured using RT-PCR and is presented the fold change in expression as compared to the LPS stimulated cells. The data is the mean ± SD of three independent experiments, each with biological triplicate samples. (b) The quantity of TNF and (c) IL-6 secreted by the macrophages into culture medium was measured by ELISA. The data shown is the mean ± SD of two independent experiments, each with biological triplicate samples. (d) Argonaute(Ago)-2 proteins were pulled down from extracted murine peritoneal macrophages pooled from five uninfected BALB/c mice (n = 3 pooled samples) and from five mice (n = 3 pooled samples) at 6 h and 12 h post-infection with F. hepatica. Ago-2-bound RNA was extracted and analysed using RT-qPCR with primers specific for fhe-miR-125b. The quantity of miRNA detected is presented as the starting quantity of genetic material prior to amplification (N0) as determined by LinRegPCR (v.11). Statistical significance was determined by one-way analysis of variance (ANOVA) with Tukey post hoc analysis. www.nature.com/scientificreports/ miRNAs in this complex parasite-host immune-interplay. Further elucidation of how mechanisms of immune modulation, involving nucleic acids, proteins and glycans, are employed by the parasite, especially during the early stage of infection, is critical for development of effective anti-helminth therapies to prevent infection and pathogenesis. Small RNA library preparation and bioinformatics analysis. The quality of the raw fastq files was assessed using FastQC and the adapters were removed using Cutadapt v1.13. The sequences with length smaller that 18 bp were discarded also using Cutadapt. The reads were then pre-processed with an in-house perl script that filters homopolymers, sequences that contain Ns and low quality reads. The mature miRNA sequences of Mus musculus were retrieved from miRbase and used to build the index for the alignment using bowtie. After the pre-processing, the reads were aligned against the mouse mature miRNAs. The reads that did not align against the mouse mature miRNAs were aligned against the F. hepatica known miRNAs. After the alignments, another in-house script was used to count the reads that were successfully mapped.

Methods
Preparation of parasite RNA. Fasciola hepatica metacercariae were sourced from Ridgeway Research Ltd, UK. The South Gloucester isolate was used to infect sheep which allowed the recovery of adult worms from the bile ducts. BALB/c mice aged 6-8 weeks were orally infected with 30 metacercariae of F. hepatica (Italian isolate, Ridgeway Research Ltd) to recover the 21 day old immature flukes. All animal experimental procedures were carried out at Queen's University Belfast, UK under license from the Home Office by the Animal (Scientific Procedures) Act 1986 (License No. PPL/2806) after ethical review by the Queen's University Belfast Animal Welfare and Ethical Review Body. The Italian isolate was also used for the excystment of newly excysted juveniles (NEJ) 49 and parasite culture for the recovery of extracellular vesicles (EVs) 50 . Total RNA from all parasites stages (NEJ, NEJ EV, immature and adult worms) was extracted using the miRNeasy Mini Kit (Qiagen) according to the manufacturer's instructions, eluted in 30 μl RNase-free water.
Caenorhabditis elegans were maintained in 35 mm diameter petri dishes containing Nematode Growth Medium agar, (3 g NaCl, 17 g agar, 2. Preparation of macrophage extracellular vesicles. RAW 264.7 macrophages were cultured to confluency in RPMI + 10% FBS (v/v) at 37 °C and 5% CO 2 . Cells were collected by scraping and cellular debris removed by sequential centrifugation steps of 2000×g for 20 min at 4 °C followed by 10,000×g for 30 min at 4 °C. To obtain the extracellular vesicles the supernatant was subjected to ultracentrifugation at 104,492×g for 2 h at 4 °C and the pelleted material re-suspended to maximum volume depending on the size of the centrifuge tube. This resuspension was then ultracentrifuged at 104,492×g overnight at 4 °C to separate the soluble proteins (now in the supernatant) and vesicles (now in the pellet). www.nature.com/scientificreports/ miRNA target prediction. To determine possible gene targets of selected F. hepatica miRNA, the miRNA sequence of fhe-miR-125b was initially analysed using custom prediction with a murine background on the online miRNA target prediction tool, miRDB (mirdb.org). Additionally, mouse mRNA 3′UTR sequences were analysed for fhe-miR-125b target pairing using miRNA target predictive tools miRanda (http:// www. micro rna. org/) and TargetScan (http:// www. targe tscan. org). miRNA:mRNA pairings with a target score of > 60 were considered in miRDB, target score > 155 and minimum free energy (MFE) < − 20 (Energy-Kcal/Mol) were considered in miRanda, and a seed sequence match of 7mer-m8 site in TargetScan. Only common genes determined by all predictive tools were considered for further analysis. The acquired list of gene targets was analysed using the Database for Annotation, Visualization and Integrated Discovery (DAVID) v6.7 (david.ncifcrf.gov) and significant pathways targeted by fhe-miR-125a were determined using the murine Kyoto Encyclopaedia of Genes and Genomes (KEGG) database 39 .
Microarray analysis. Microarray analyses were performed using the Illumina whole genome microarray platform (Illumina, USA) as previously described 51 . Quality control of data was performed using GenomeStudio (version 1.1.1, Illumina) by examining intensity histograms of hybridisation efficiency and noise. GeneSpring GX version 11 (Agilent Technologies, Foster City, USA) was used for all subsequent data analyses. All data were entered into GeneSpring GX 11 and normalised to the 75th percentile. Data for each replicate were normalised to respective uninfected controls and filtered for significance on the basis of detection score, a GenomeStudio generated measure of signal intensity relative to background. At least half of all hybridisations had to have a detection score > 0.949, (which equates to a confidence value of p ≤ 0.05) for a gene to be accepted. Gene expression values for commonly up-and down-regulated genes were averaged and a Students t-test was performed to identify significant changes in gene expression compared with controls (p ≤ 0.05). An additional cut-off of ± twofold change in expression was applied to identify changes in gene expression with likely biological significance.

Transfection of bone marrow derived murine macrophages. Bone marrow derived macrophages
(1 × 10 6 cells/500 µl) were transfected with synthetic miRNAs (20 nM; (IDT, USA) combined with Lipofectamine RNAiMAX (Life Technologies, USA) over a period of 24 h in accordance to manufacturer's protocol. Transfected cells were washed with sterile PBS before stimulation with E. coli lipopolysaccharide (LPS; 10 ng/ml) for 6 h at 37 °C and 5% CO 2 to induce an inflammatory response. Non-transfected cells, and cells treated with Lipofectamine RNAiMAX alone, without LPS stimulation, were also analysed to control for contamination and possible immune responses induced by transfection reagents. Non-transfected cells were stimulated with LPS to determine the baseline inflammatory response of cultured macrophages. After LPS stimulation, the culture media was collected from all treatment groups for cytokine assays, and the cells washed with sterile PBS and collected for RNA extraction.

RNA extraction from mammalian cells. Cells or extracellular vesicles were homogenised with the
RNAzol RT at 1 mL to 10 7 cells. Ultrapure DNAse/RNase free distilled water (dH 2 O) (Life Technologies, USA) was added to the homogenate, at 0.4 mL of dH 2 O per 1 mL of RNAzol RT used, and then centrifuged at 12,000×g for 15 min at 4 °C, to pellet DNA, proteins, and polysaccharides. Supernatant containing soluble RNA was collected and combined with 4-bromoanisole (BAN; Sigma-Aldrich, USA) at 0.5% of supernatant volume, and then centrifuged at 12,000×g for 10 min at 4 °C to pellet residual DNA, proteins, and polysaccharides. To precipitate RNA, the supernatant was collected and combined with isopropanol (100% v/v; Sigma-Aldrich, USA) and 5 μL glycogen (5 mg/mL; Life Technologies, USA), and incubated overnight at − 20 °C. The RNA precipitate was pelleted by centrifugation at 12,000×g for 10 min at 4 °C, washed with 500 µL ethanol (75% v/v; Sigma-Aldrich, USA) twice and re-solubilised in 25 µL dH 2 O. Concentrations and quality of isolated RNA was measured using the NanoDrop One/One UV-Vis spectrophotometer (Life Technologies).
Pull-down of argonaute-2 (Ago-2)-bound miRNA from peritoneal macrophages. Peritoneal macrophages were collected and pooled from five uninfected or infected mice for Ago2 pulldown using the MagCapture microRNA isolation kit specific for mouse Ago-2 (FUJIFILM Wako Pure Chemical, JPN) according to manufacturer's instructions. Briefly, cleared cell lysate from each pooled sample of 2 × 10 6 macrophages was incubated with AntiMouse Ago2 magnetic beads for 2 h at 4 °C with rotation. The beads were then washed and subjected to Proteinase K digestion. RNA was isolated from the digest using RNAzol RT and quantified using Qubit Fluorometer with a microRNA Assay kit (ThermoFisher Scientific). The miRNA samples were then normalised to a concentration of 0.5 ng/sample for RT-qPCR. This two-step normalisation of cell number and input miRNA ensured that all samples contained similar amounts of miRNA for comparable analysis by RT-qPCR.

RT-qPCR.
After extraction of RNA, all samples were normalised to 100 ng/μL prior to cDNA synthesis.
cDNA synthesis was carried out using the TaqMan MicroRNA Reverse Transcription Kit (Life Technologies, USA) according to manufacturer's instructions. Synthesised cDNA was further diluted 1:4 dH 2 O. For RT-qPCR, 1 µL of diluted cDNA was combined with 5 µL of SensiFAST Probe Hi-ROX Kit (Bioline, UK), 3.5 µL dH 2 O, and 0.5 µL of TaqMan Probe for gene expression or miRNA. Samples were prepared on a MicroAmp optical 96-well reaction plate as technical triplicates (Life Technologies, USA) and analysed using the QuantStudio 6 flex realtime PCR system (Life Technologies, USA). Data was analysed using LinRegPCR (v2017.1), and presented as absolute expression of genetic material prior to amplification (N0). The N0 value is the raw fluorescent value and an indication of the expression.