Type III interferons (IFNs) are the latest members of the IFN family. They play an important role in immune defense mechanisms, especially in antiviral responses at mucosal sites. Moreover, they control inflammatory reactions by modulating neutrophil and dendritic cell functions. Therefore, it is important to identify cellular mechanisms involved in the control of type III IFN expression. All IFN family members contain AU-rich elements (AREs) in the 3′-untranslated regions (3′-UTR) of their mRNAs that determine mRNA half-life and consequently the expressional level of these cytokines. mRNA stability is controlled by different proteins binding to these AREs leading to either stabilization or destabilization of the respective target mRNA. The KH-type splicing regulatory protein KSRP (also named KHSRP) is an important negative regulator of ARE-containing mRNAs. Here, we identify the interferon lambda 3 (IFNL3) mRNA as a new KSRP target by pull-down and immunoprecipitation experiments, as well as luciferase reporter gene assays. We characterize the KSRP-binding site in the IFNL3 3′-UTR and demonstrate that KSRP regulates the mRNA half-life of the IFNL3 transcript. In addition, we detect enhanced expression of IFNL3 mRNA in KSRP−/− mice, establishing a negative regulatory function of KSRP in type III IFN expression also in vivo. Besides KSRP the RNA-binding protein AUF1 (AU-rich element RNA-binding protein 1) also seems to be involved in the regulation of type III IFN mRNA expression.
The interferon (IFN) cytokine family consists of type I, II and III IFNs. All family members have important antiviral and antimicrobial activities in innate and adaptive immune responses. Type III IFNs, also known as interferon lambda (IFNL), were first discovered in 2003. In humans, four different type III IFNs exist, IFNL1 (IL-29), IFNL2 (IL-28A), IFNL3 (IL-28B) [1,2] and IFNL4 , whereas in mice only two members, IFNL2 (IL-28A) and IFNL3 (IL-28B) , have been identified so far. Various cell types express type III IFNs, including plasmacytoid dendritic cells, myeloid-lineage cells, T-regulatory cells, epithelial cells and hepatocytes [5–7]. Type III IFNs share many similarities with type I IFNs (IFN-α and IFN-β): They are primarily induced by the activation of pattern recognition receptors and mitochondrial antiviral signaling proteins (MAVS) upon viral or microbial infections [8–11]. Both types of IFNs activate the Janus kinase/signal transducers and the activator of the transcription pathway followed by the induction of IFN-stimulated genes. However, the distribution of the receptors for these two IFN types is quite different. The type I IFN receptor (IFNAR) is ubiquitously expressed, and this seems to be the reason for the strong and robust activation of immune responses and inflammatory reactions upon type I IFN secretion. Based on the restricted expression of the IFNL receptor on epithelial and myeloid-lineage cells, IFNL-mediated immune responses are primarily restricted to epithelial barriers  as well as neutrophil and dendritic cells (DCs) . In neutrophils, they inhibit the production of reactive oxygen species , release of neutrophil extracellular traps  and cell migration . DCs activated by type III IFNs favor polarization of T cell either towards a Th1 phenotype, at least in vitro , or promote FoxP3+ Treg proliferation .
The antiviral effects of type III IFNs have been studied intensely demonstrating their pivotal role in innate and adaptive immune responses. In addition, an aberrant expression of type III IFNs has been described in systemic lupus erythematosus (SLE)  or rheumatoid arthritis , suggesting an etiologic role in autoimmune diseases. A tight regulation of type III IFN expression seems therefore necessary. Regulatory processes on the transcriptional level are well characterized [19,20], but much less is known about post-transcriptional regulatory mechanisms of type III IFN expression .
Besides splicing, mRNA editing, translatability and localization, regulation of mRNA stability is an important post-transcriptional mechanism that modulates gene expression. Especially the expression of pro- and anti-inflammatory genes is regulated on the level of mRNA stability. Important sequence elements that determine transcript stability are AU-rich elements (AREs) located in the 3′-untranslated region (3′-UTR) of the mRNA [22,23]. Those AREs are recognized by different RNA-binding proteins (RBPs) that either stabilize (HuR/HuA) or destabilize [TTP, KSRP, AU-rich element RNA-binding protein 1 (AUF1)] the corresponding mRNAs . mRNA decay may also be induced by microRNAs (miRNAs) that bind directly to specific sequences in 3′-UTRs.
Type I, II and III IFNs, all contain AREs in the 3′-UTRs of their mRNAs that appear to regulate the expression of IFN transcripts . However, only little is known about the factors involved in the post-transcriptional regulation of those mRNAs. For both type I IFNs, IFN-α and IFN-β, it has been shown that the RBP KSRP (KH-type splicing regulatory protein, also named KHSRP) destabilizes their mRNAs in the murine system. Therefore, KSRP−/− mice are less susceptible to viral infections . The 3′-UTRs of the type III IFNs, IFNL2 and IFNL3, each contain three different ARE motifs that confer transcript instability, but almost nothing is known about the RBPs involved. In contrast, there are several reports about miRNA-mediated decay of type III IFN mRNAs, primarily in hepatitis C virus defense .
KSRP is a single-stranded nucleic acid-binding protein with multiple functions. It plays an important role in different steps of the post-transcriptional control of gene expression, e.g. regulation of mRNA splicing, stability and translatability, as well as miRNA maturation [27–29]. One of the best-characterized KSRP functions is its ability to mediate rapid decay of ARE-containing mRNAs . To this end, KSRP recruits enzymes involved in mRNA decay, such as the exosome or other nucleases. KSRP seems thus to be a central component of the ARE-mediated decay (AMD) . Many mRNAs of pro-inflammatory mediators that possess AREs are targets of KSRP-mediated mRNA decay and are therefore often inherently instable. For example, it has been demonstrated that KSRP decreases stability of the mRNAs for tumor necrosis factor-α (TNF-α) [31,32], interleukin-8 , type I IFNs  and inducible nitric oxide synthase  by binding to AREs in the 3′-UTR.
Control of mRNA stability seems to be an important mechanism in the regulation of IFN gene expression. As type I and type III IFNs exhibit remarkable similarities and because type I IFN mRNA stability is regulated by KSRP, we wondered whether this is also true for type III IFN mRNAs. Using pull-down and immunoprecipitation experiments and luciferase reporter gene assays in human cell lines as well as cells isolated from KSRP−/− mice, we investigated whether the RBP KSRP regulates the decay of type III IFN mRNAs.
Human epithelial colon carcinoma DLD-1 cells (ATCC, #CCL-221) were cultivated in DMEM with 10% inactivated fetal bovine serum (FCS) (PAN-Systems, Nürnberg, Germany), 2 mM l-glutamine, 100 U/ml penicillin and 100 µg/ml streptomycin (Sigma, Deisenhofen, Germany).
DLD-1 shKSRP cells
Both DLD-1 cells with diminished KSRP expression (shKSRP) and control cells were stably transfected by lentiviral transduction as described before . The pGIPZ lentiviral shRNA expression vector (pGIPZ-hmr-shKSRP) coding for miRNA-adapted shRNA (CGCTGGGAAGAGTATTACAAAAAATAGTGAAGCCACAGATGTATTTTTTGTAATACTCTTCCCAA) directed against human, murine and rat KSRP was obtained from ThermoFisher Scientific, Schwerte, Germany. To obtain a vector control, the pGIPZ-hmr-shKSRP plasmid was digested with Mlu I and Xho I and treated with Klenow enzyme resulting in pGIPZ CO.
Cloning of the 3′-UTR sequences of human TNF-α, IFNL3 and type I interferon mRNA
To obtain constructs for the generation of RNA probes for pull-down experiments and constructs for the transfection of DLD-1 cells, the 3′-UTR sequences of the human TNF-α and IFNL3 mRNA were cloned into the pCR-Script (Agilent Technologies, Waldbronn, Germany) and/or pCMVcontrolInglo vector. We designed the pCMVcontrolInglo vector for analyses of mRNA stability under the control of different 3′-UTRs. The vector consists of a CMV promoter derived from the pCDNA4/TO vector, part of exon 1, intron 1 and part of exon 2 from pRHCglo (a kind gift from Gopal Singh and Thomas A. Cooper ), the firefly luciferase gene and late SV40 poly(A) signal from the pGL3control vector (Supplementary Figure S2). As the pCMVcontrolInglo vector contains an intron, qRT-PCR (quantitative real-time reverse transcription polymerase chain reaction) primers can be designed which detect only the mature luciferase mRNA and no DNA contamination (see Supplementary Figure S2).
To obtain the plasmids pCR-TNFα-3UTR and pCR-TNFα-3UTRdelARE, the plasmids pGL3control_TNF3UTR and pGL3control_h-TNF3UTRAUdel (kind gifts from Dr W.F.C. Rigby ) were digested with Xba I and treated with Klenow enzyme. The blunt-ended fragments containing the 3′-UTR of the human TNF-α mRNA (with or without AREs) were isolated and ligated to pCR-Script digested with Eco RV and treated with calf intestinal alkaline phosphatase (CIAP). All enzymes were obtained from New England Biolabs, Frankfurt a.M., Germany.
To obtain the 3′-UTR sequence of IFNL3, PCR with genomic DNA isolated from DLD-1 cells was performed using the oligonucleotides IFNL3 sense ACCCTTCCGCCAGTCATGC and IFNL3 antisense AACAAGGATTTCAAAAAGTAG. All oligonucleotides were purchased from Sigma, Deisenhofen, Germany. The PCR fragment was purified, treated with polynucleotide kinase (PNK) and cloned into pCR-Script vector restricted with Eco RV and treated with CIAP, and into pCMVcontrolInglo vector restricted with Fse I and treated with Klenow enzyme and CIAP.
To obtain the 3′-UTR sequences of IFNL3 in pCMVcontrolInglo vector in reverse orientation, pCR-IFNL3-3UTR-A was restricted with Xba I and Ale I, and the fragment was treated with Klenow enzyme. The blunt-ended fragments were isolated and cloned into pCMVcontrolInglo digested with Fse I and treated with Klenow enzyme and CIAP.
To obtain the mutated 3′-UTR sequence of IFNL3 (IFNL3-Mut ARE), the overlapping oligonucleotides IFNL3-Mut sense ACCCTTCCGCCAGTCATGCAACCTGAGATTTTGCCGCTAAATTAGCCACTTGGCTTAGCCGCTTGTCACCCAGTCGCTGCCGCTGTATTTGTGTATGTAAATCCAACTCACCTCCAGGAA and IFNL3-Mut antisense AACAAGGATTTCAAAAAGTAGAAAAATAAACATTTTCCTGGAGGTGAGTTGGATTTACATACACAAATACAGCGGCAGCGACTGGGTGACAAGCGGCTAAGCCAAGTGGCTAATTTAGCG were annealed and treated with Klenow enzyme and PNK. The blunt-ended fragments were cloned into pCMVcontrolInglo digested with Fse I, treated with Klenow enzyme and CIAP, and into pCR-Script restricted with Eco RV and treated with CIAP.
To obtain the different deletion constructs of IFNL3 3′-UTR sequence (IFNL3 ARE1 and IFNL3 delARE1–3), corresponding oligonucleotide sequences were designed (Figure 4A) and cloned into pCMVcontrolInglo digested with Fse I, treated with Klenow enzyme and CIAP, and into pCR-Script restricted with Eco RV and treated with CIAP.
To obtain constructs for the generation of RNA probes for pull-down experiments, the 3′-UTR sequences of the human IFN-α and IFN-β mRNA were cloned into the pCR-Script vector (Agilent Technologies, Waldbronn, Germany). To obtain the 3′-UTR sequences of IFN-α and IFN-β, PCR with genomic DNA isolated from DLD-1 cells was performed using the oligonucleotides IFN-α sense CATCTGGTCCAACATGAAAACAATTC, IFN-α antisense GAGTAAATATAAGGAACATGTTTTATTACACTAACC; IFN-β sense AGATCTCCTAGCCTGTGCCTC, IFN-β antisense TGACTTTTGCACCAAAAATAATTTATTTTCC. All oligonucleotides were purchased from Sigma, Deisenhofen, Germany. The PCR fragments were purified, treated with PNK and cloned into pCR-Script vector restricted with Eco RV and treated with CIAP. The relevant DNA sequences of all plasmids were determined using the dideoxy chain termination method (StarSEQ, Mainz, Germany).
To obtain protein extracts for pull-down or immunoprecipitation assays, DLD-1 cells, DLD-1 pGIPZ CO and DLD-1 pGIPZ shKSRP cells, or spleen tissue derived from KSRP+/+ and KSRP−/− mice were lysed in RIPA buffer [50 mM Tris–HCl (pH 7.4), 150 mM NaCl, 2 mM EDTA, 10% glycerol, 1% NP40, 1× complete EDTA-free protease and phosphatase inhibitor cocktail (Roche Diagnostics, Mannheim, Germany)] using an ultrasonic homogenizer. Protein concentration was determined using the Protein Assay Dye Reagent (Bio-Rad, München, Germany) as described by the manufacturer.
In vitro transcription and pull-down assay
To generate biotinylated RNA probes, 1 µg of linearized pCR-TNFα-3UTR, pCR-TNFα-3UTRdelARE, pCR-IFNα-3UTR, pCR-IFNβ-3UTR, pCR-IFNL3-3UTR-A, pCR-IFNL3-3UTR-B pCR-IFNL3-3UTR-Mut, pCR-IFNL3-3UTR-ARE1 and pCR-IFNL3-3UTR-delARE1–3 plasmid were in vitro transcribed using the biotin RNA labeling mixture as described by the manufacturer (Epicentre, Madison, U.S.A.).The reaction was conducted at 37°C for 2 h. Digestion of the plasmid DNA was performed using DNase I (Roche Diagnostics, Mannheim, Germany) for 20 min at 37°C.
One-fifth of the biotinylated RNA transcripts were incubated with buffered aqueous streptavidin-agarose from Streptomyces avidinii (Sigma, Deisenhofen, Germany) for 6 h at 4°C in 100 µl binding buffer [10 mM HEPES (pH 7.9), 3 mM MgCl2, 5 mM EDTA, 2 mM DTT, 5% glycerol, 0.5% NP-40, 3 mg/ml heparin and 0.5 mg/ml baker's yeast transfer RNA; Sigma, Deisenhofen, Germany] supplemented with 40 mM KCl. Protein extracts from DLD-1 cells were precleared by incubating 5 mg of cell extract with buffered aqueous streptavidin-agarose for 2 h at 4°C. Streptavidin-agarose with biotinylated RNA transcripts was incubated with precleared protein extract and binding buffer with 40 mM KCl overnight at 4°C. Streptavidin-agarose was washed five times with binding buffer with 40 mM KCl. Proteins were eluted by incubating with elution buffer [10 mM HEPES (pH 7.9), 3 mM MgCl2, 5 mM EDTA, 2 mM DTT, 0.2% glycerol and 2 M KCl] for 30 min at 4°C.
To identify the eluted proteins, streptavidin-agarose was pelleted and a fifth of the supernatant fraction was separated on SDS–polyacrylamide gels and transferred onto the nitrocellulose membrane (GE Healthcare, Freiburg, Germany) by semi-dry electroblotting. The antibody against AUF1 was purchased from Upstate (Temecula, CA, U.S.A.). For detection of KSRP, a polyclonal anti-KSRP antibody, a kind gift from Dr Ching-Yi Chen (Department of Biochemistry and Molecular Genetics, University of Alabama, Birmingham, U.S.A.), was used. The antibody against HuR was purchased from Sigma (H1663, Deisenhofen, Germany). The anti-mouse IgG antibody was purchased from Sigma (A6782, Deisenhofen Germany). The antibody against GAPDH was obtained from Santa Cruz Biotechnology (32233, Dallas, U.S.A.). The immunoreactive proteins on the blot were visualized by the enhanced chemiluminescence detection system (ThermoFisher Scientific, Darmstadt, Germany). Densitometric analyses were performed with Canon 9000F Mark II Scanner and Quantity One software (Bio-Rad, Munich, Germany).
Immunoprecipitation PCR assay
For the determination of intracellular protein–RNA interaction, DLD-1 cells were transiently transfected by lipofection with pCMVcontrolInglo-IFNL3-3UTR-A or pCMVcontrolInglo-IFNL3-3UTR-ARE1 (negative control). Twenty-four hours after transfection, cells were lysed in a buffer containing 10 mM HEPES (pH 7.9), 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT, 0.1% NP40, 1× complete EDTA-free protease and phosphatase inhibitor cocktail (Roche Diagnostics, Mannheim, Germany) and 100 U/ml RNasin. Cell lysates (1.5 mg) were preincubated with 20 µl of protein A/G plus agarose beads (Santa Cruz Biotechnology, Heidelberg, Germany) in 500 µl of NT2 buffer [50 mM Tris–HCl (pH 7.4), 150 mM NaCl, 1 mM MgCl2, 0.05% NP-40, 1 mM DTT, 2 mM EDTA, 0.2% VRC, 100 U/ml RNasin and 1× complete EDTA-free protease and phosphatase inhibitor cocktail (Roche Diagnostics, Mannheim, Germany)] for 30 min at 4°C. After centrifugation at 1200×g for 5 min at 4°C, the supernatant was incubated with 50 µl of protein A/G plus agarose beads [precoated with specific KSRP antibody (NBP-1-18910, Novus Biological, Wiesbaden, Germany) or IgG antibody (#12-370; MerckMillipore, Darmstadt, Germany) as control, in 50 mM Tris (pH 7.4), 150 mM NaCl, 1 mM MgCl2, 0.05% NP-40, 0.5 µg/µl tRNA, 0.5 mg/ml heparin for 16 h at 4°C] for 3 h at 4°C. Subsequently, the beads were washed four times in 1 ml of NT2 buffer and then digested with 0.5 mg/ml Proteinase K in NT2 buffer containing 0.1% SDS for 15 min at 55°C. The amount of luciferase reporter gene mRNA bound by KSRP was determined by qRT-PCR using the pCMVcontrolInglo primers (Supplementary Figure S2). The same protocol was used with protein extracts derived from DLD-1 pGIPZ CO, DLD-1 pGIPZ shKSRP or splenic T cells isolated from KSRP wild-type (WT) and knockout mice. For detection of human and murine IFNL3, mRNA-specific primers listed in Table 1 were used.
|Luciferase mRNA encoded by pCMVcontrolInglo||CTCGTTTAGTGAACCGTCAG||TCTTCCATGGTGGCTTTACC||GCTCCGGACTCGGATCC|
|Luciferase mRNA encoded by pCMVcontrolInglo||CTCGTTTAGTGAACCGTCAG||TCTTCCATGGTGGCTTTACC||GCTCCGGACTCGGATCC|
All oligonucleotides and dual-labeled probes were purchased from Sigma, Deisenhofen, Germany.
To demonstrate immunoprecipitation of KSRP protein from human DLD-1 cells or murine spleen, the protocol described above was used without Proteinase K digestion. Proteins were separated on SDS–polyacrylamide gels and transferred onto the nitrocellulose membrane (GE Healthcare, Freiburg, Germany) by semi-dry electroblotting. For detection of KSRP, a polyclonal anti-KSRP antibody, a kind gift from Dr Ching-Yi Chen (Department of Biochemistry and Molecular Genetics, University of Alabama, Birmingham, U.S.A.), was used.
Transient transfection/reporter gene assay
To investigate the effect of type III IFNs 3′-UTR on luciferase activity, DLD-1 control and DLD-1 shKSRP cells were plated onto 24-well plates and transiently transfected by lipofection with pCMVcontrolInglo, pCMVcontrolInglo-IFNL3-3UTR-A, pCMVcontrolInglo-IFNL3-3UTR-B, pCMVcontrolInglo-IFNL3-3UTR-Mut, pCMVcontrolInglo-IFNL3-3UTR-ARE1, pCMVcontrolInglo-IFNL3-3UTR-delARE1–3 and pEF-1α-RL (encoding for a renilla luciferase, for normalization). Lipofection was performed with GeneJuice according to the manufacturer's recommendations. After overnight incubation, cells were lysed in 1× Passive Lysis Buffer (PLB), and firefly and renilla luciferase activities were determined. The light units of the firefly luciferase were normalized by those of the renilla luciferase after subtraction of extract background.
mRNA stability analyses
To examine the influence of KSRP on the stability of luciferase mRNA under the control of IFNL3 3′-UTR, shKSRP and control cells were transiently transfected with pCMVcontrolInglo-IFNL3-3UTR-A or pCMVcontrolInglo-IFNL3-3UTR-ARE1 for 24 h by lipofection with GeneJuice (Merck, Darmstadt, Germany) according to the manufacturer's recommendations and incubated with serum-free medium for another 16 h. Then, 25 µg/ml 6-dichloro-1-ribofuranosylbenzimidazole (DRB) (Sigma, Deisenhofen, Germany) was added, and RNAs were prepared 0, 2, 8 and 16 h thereafter. The amount of luciferase mRNA was determined by qRT-PCR by normalizing to 18S rRNA expression. The relative amount of luciferase mRNA at 0 h DRB was set at 100% in each cell type. To determine mRNA decay of endogenous IFNL3 transcript, shKSRP and control cells were incubated with serum-free medium for 16 h. Then, 25 µg/ml DRB (Sigma, Deisenhofen, Germany) was added, and RNAs were prepared 0, 2, 4, 8 and 16 h thereafter. The amount of IFNL3 mRNA was determined by qRT-PCR by normalizing to 18S rRNA expression. The relative amount of IFNL3 mRNA at 0 h DRB was set at 100% in each cell type. Curve fittings of the resulting DRB time curves were performed by non-linear regression using GraphPad Prism 7.04d.
Analysis of mRNA expression in cells and tissues by real-time reverse transcription polymerase chain reaction
To analyze mRNA expression, total RNA was prepared by homogenizing cells or tissue samples in GIT buffer . RNA was isolated and specific gene expression was quantified in a two-step real-time RT-PCR as previously described  with the oligonucleotides listed below.
All oligonucleotides and dual-labeled probes were purchased from Sigma, Deisenhofen, Germany.
Specific mRNA expression was normalized to human 18S rRNA or murine RNA Polymerase II (Pol2A). To calculate the relative mRNA expression, the method  was used.
All mice were housed in accordance with standard animal care requirements and were maintained under specified pathogen-free conditions on a 12/12-h light/dark cycle. Water and food were given ad libitum. The animal studies were approved by the ethical board and were performed in accordance with the German animal protection law and the guidelines for the use of experimental animals as stipulated by the Guide of Care and Use of Laboratory Animals of the National Institutes of Health. Mice were killed by i.p. injection of 700 µl Pentobarbital solution (1% Pentobarbital in PBS). KSRP+/− mice had a C57BL/6 background. Experimental KSRP−/− (KO) and KSRP+/+ (WT) animals were obtained by mating KSRP+/− animals . Genotyping of the animals was performed by PCR, using primers that span the region of the WT KSRP gene flanked by loxP sites which were deleted by the Cre recombinase . The following oligonucleotides were used for genotyping the KSRP locus: KSRP-wt-for GCGGGGAGAATGTGAAGG, KSRP-ko-for CTCCGCCTCCTCAGCTTG and KSRP-wt/ko-rev GAGGCCCCTGGTTGAAGG.
Induction of acute inflammation by LPS treatment
For induction of acute inflammation, KSRP−/− and KSRP+/+ mice were treated by i.p. injection of either 0.9% NaCl (solvent control) or 20 mg/kg LPS in 0.9% NaCl. After 6 h, mice were killed by i.p. injection of 700 µl Pentobarbital solution (1% Pentobarbital in PBS). Organs were collected and kept at −80°C.
Induction of collagen antibody-induced arthritis
Collagen antibody-induced arthritis (CAIA) was induced as described recently .
Isolation of spleen cells
To investigate the effects of KSRP on Ifnl3 mRNA expression in primary murine cells, the spleen from KSRP−/− or KSRP+/+ mice was squeezed through a cell strainer (40 µm). To lyse the erythrocytes, the cell suspension was treated with lysis buffer (155 mM NH4Cl, 10 nM KHCO3, 100 µM EDTA disodium, pH 7.4) for 1 min and washed afterwards. The cells were cultured for 2 h in 24-well plates with IMDM medium (supplemented with 5% FCS, 2 mM l-glutamine, 100 U/ml penicillin, 100 µg/ml streptomycin and 50 µM β-mercaptoethanol). For stimulation, 100 ng/ml LPS was applied to the medium. In addition, splenic T cells were isolated with CD4+ T Cell Isolation Kit mouse (Miltenyi Biotec, Bergisch Gladbach, Germany) as described by the manufacturer and were polyclonally stimulated with anti-CD3 (1 µg/ml) and anti-CD28 (2 µg/ml) specific antibodies (both from Affymetrix/eBiosciences, San Diego, CA, U.S.A.).
Data represent means + SEM. Statistical differences were determined by factorial analysis of variance followed by ‘Tukey's’ or ‘Dunnett's’ multiple comparison test. In the case of two means classical two-sided unpaired t-test analyses were used. To compare the mRNA stability, transient transfection analyses and mRNA expression in LPS- or CAIA-treated mice, two-way ANOVA analysis followed by the multiple comparisons test was performed. All statistical analyses were performed using GraphPad Prism 7.04 (GraphPad Software, San Diego, CA, U.S.A.).
KSRP modulates IFNL3 expression in epithelial cells
It is already known that the human IFNL2 and IFNL3 3′-UTRs contain AREs responsible for enhanced mRNA degradation , but almost nothing is known about RBPs that might be involved in this post-transcriptional regulation mechanism. For the murine system, such information is also missing. Stabilities of type I IFN mRNAs are regulated by the RBP KSRP. Because they exhibit some similarities with type III IFN mRNAs , we postulated that KSRP might also be involved in the control of mRNA stability of the latter. Epithelial cells are one important source of IFNL3. Therefore, we decided to investigate IFNL3 mRNA expression in human epithelial colon carcinoma DLD-1 cells with or without diminished KSRP expression. Cells with largely reduced KSRP expression due to transfection of a shRNA directed against KSRP (shKSRP) exhibited a more than twofold higher IFNL3 mRNA expression compared with control cells (control) as demonstrated by quantitative real-time RT-PCR (qRT-PCR) (Figure 1). This result hints towards a KSRP-mediated regulation of IFNL3 expression.
KSRP modulates IFNL3 expression in epithelial cells.
KSRP binds to IFNL3 3′-UTR
To investigate whether KSRP directly interacts with IFNL3 3′-UTR, we cloned the human IFNL3 3′-UTR (sequence information, see Figure 4A) in the pCR-Script vector. The IFNL3 3′-UTR was transcribed in vitro in the presence of biotinylated dNTPs. The transcripts were then incubated with extracts obtained from human epithelial colon carcinoma DLD-1 cells and a pull-down was performed using streptavidin-agarose. KSRP binding to IFNL3 3′-UTR was analyzed by Western blot experiments. As control, we used an in vitro transcript of a well-known KSRP target mRNA, the TNFA 3′-UTR. As expected, KSRP binding was detected to TNFA 3′-UTR containing the endogenous AREs (TNF-α), but not when these AREs were missing (TNF-α delARE) (Figure 2A, eluate). In parallel assays, we also detected KSRP binding to IFNL3 3′-UTR (Figure 2A, IFNL3 A, eluate). To verify the specificity of this interaction, we generated a control plasmid with the IFNL3 3′-UTR sequence in reverse orientation. No KSRP binding was detected in this negative control (Figure 2A, IFNL3 B, eluate). The amount of KSRP present in the protein extracts used for pull-down analyses that did not bind to streptavidin-agarose beads was quantified in the supernatants. GAPDH expression was determined to demonstrate that in all samples, comparable amounts of proteins were present (Figure 2A, supernatant). The amount of KSRP protein bound to each 3′-UTR was normalized to the KSRP content in the supernatant fraction. The results show that similar amounts of KSRP protein, ∼5% of KSRP protein present in DLD-1 cell extracts, bound to the IFNL3 3′-UTR A and TNFA 3′-UTR (Figure 2B). To identify similarities between AMD mechanisms in the control of type I and III IFN expression, we analyzed whether the mRNA-stabilizing RBP HuR is capable of interacting with the IFNL3 3′-UTR. We choose this RBP, because Herdy et al.  demonstrated direct binding of HuR to IFN-β 3′-UTR, a member of type I IFN family. Indeed, we reproduced his data and detected also binding of HuR to IFNA 3′-UTR (Supplementary Figure S1). However, in contrast with type I IFN, no binding of the stabilizing RBP HuR was detected to IFNL3 mRNA (Supplementary Figure S1).
KSRP binds to the 3′-UTR sequence of human IFNL3.
Next, we cloned the human IFNL3 3′-UTR in a luciferase reporter vector (pCMVcontrolInglo, Supplementary Figure S2) and transfected this construct into DLD-1 cells with or without diminished KSRP expression to analyze the biological effect of KSRP binding to the IFNL3 3′-UTR sequence. As control, we used a reporter gene without a 3′-UTR sequence (pCMVcontrolInglo). Compared with this control, cells transfected with the IFNL3 3′-UTR (IFNL3 A) reporter gene construct exhibited reduced luciferase activity demonstrating a negative regulatory effect of this sequence. This result confirmed the already published data that IFNL3 3′-UTR confers transcript instability . In shKSRP DLD-1 cells, the luciferase activity of the IFNL3 3′-UTR reporter gene construct almost returns to the level of the pCMVcontrolInglo control vector (Figure 3). IFNL3 3′-UTR in the reverse orientation (IFNL3 B) was used as control and, as expected, we detected no effect on luciferase expression compared with the control vector, neither in control cells nor in cells with diminished KSRP expression (Figure 3). These data indicate that KSRP is involved in the post-transcriptional regulation of type III IFN expression.
KSRP modifies IFNL3 3′-UTR reporter transcript expression.
KSRP binding to IFNL3 3′-UTR is located downstream of the classical AREs.
Characterization of KSRP-binding site in IFNL3 3′-UTR
KSRP mediates mRNA decay by binding to AREs in the 3′-UTR of its target mRNAs. Hence, we mutated the sequences of all three known AREs in the IFNL3 3′-UTR (IFNL3 MutARE, Figure 4A) to test whether ARE of IFNL3 serves as a KSRP-binding site. The 3′-UTR sequence with mutated AREs was cloned into pCR-Script vector to perform pull-down experiments with DLD-1 cell extracts. Surprisingly, the mutated IFNL3 3′-UTR (IFNL3 MutARE) still bound KSRP to a similar extent as the wildtype (WT) IFNL3 3′-UTR (IFNL3 A) (Figure 4B,C) and TNF-α 3′-UTR (TNF-α) used as a positive control. Almost no binding of KSRP to the negative control, TNF-α 3′-UTR without ARE sequences (TNF-α delARE), was observed. These results indicate that KSRP does not bind directly to yet described ARE sequences of IFNL3. To confirm that our nucleotide exchange is suitable to abrogate protein interaction, we analyzed ARE-binding capacity of other RBP. In these experiments, we detected binding of the RBP AUF1 isoform p40 and p42 to the WT IFNL3 3′-UTR (IFNL3 A), and this direct protein interaction was abolished when we used the IFNL3 3′-UTR with the mutated AREs (IFNL3 MutARE) (Figure 4D). These results indicate that the sequence chosen to mutate ARE is suitable to prevent RBP binding. Altogether, these findings suggest that KSRP does not bind to the three ARE sequences shown before  to be responsible for IFNL3 transcript instability.
To identify the KSRP-binding region, we dissected the IFNL3 3′-UTR. The first ARE contains, beside the core ATTTA, additional ARE sequences. To find out whether KSRP binds to these sequences, we generated a clone that contains only this first ARE sequence (IFNL3 ARE 1, Figure 4A). In pull-down analyses, we did not detect any KSRP binding to this fragment (Figure 4B,C). This result implicates that KSRP interacts with sequence elements downstream of the three characterized AREs. To test this hypothesis, we generated a clone where all three ARE sequences of the IFNL3 3′-UTR were deleted (IFN3 delARE1–3, Figure 4A). In pull-down analyses, we detected KSRP binding to this modified 3′-UTR sequence that was even stronger than the binding to the WT 3′-UTR (Figure 4B,C). Figure 4C presents a summary of the densitometric analysis of different pull-down experiments. Similar amounts of KSRP, ∼6% of KSRP protein present in the DLD-1 cell extract, bound to IFNL3 A and IFNL3 MutARE transcripts. No binding of KSRP to IFNL3 ARE1 or TNF-α delARE (negative control) was detected. The in vitro transcripts of IFNL3 delARE1–3 and TNF-α displayed the strongest interaction with KSRP. Altogether, these results demonstrate that KSRP does not interact with any of the classical ARE sequences. The binding site seems to be located in the 3′-region of the IFNL3 3′-UTR sequence which contains no classical AREs but Uracil-rich (U-rich) sequence.
Next, DLD-1 control or DLD-1 shKSRP cells were transfected either with luciferase reporter constructs containing the WT IFNL3 3′-UTR (IFNL3 A), the WT IFNL3 3′-UTR in reverse orientation (IFNL3 B) or the different mutated IFNL3 3′-UTR constructs (IFNL3 MutARE, IFNL3 ARE1, IFNL3 delARE1–3; Figure 5) or a control vector without 3′-UTR sequences (pCMVcontrolInglo). Twenty-four hours after transfection, cells were lysed and luciferase activity was determined. The presence of IFNL3 3′-UTR (IFNL3 A) reduced luciferase expression, and this effect was less pronounced in shKSRP cells. Mutation or deletion of the three well-known ARE sequences (IFNL3 MutARE, IFNL3 delARE1–3) partially restored luciferase activity in DLD-1 control cells (Figure 5). Those results indicate that these ARE sequences might be also important for regulation of IFNL3 expression. But instead of KSRP, most probably, other RBPs such as AUF1, which we have shown to bind to this region (Figure 4D), are presumably responsible for ARE-mediated regulation of luciferase expression detected in control cells. Most probably due to co-operate action of different RBPs such as AUF1 and KSRP, only rather small effects of KSRP, which still binds to these constructs (IFNL3 MutARE, IFNL3 delARE1–3), on luciferase expression are detectable in control cells. Nevertheless, the importance of KSRP for regulation of IFNL3 expression is demonstrated by the fact that in constructs which still display KSRP-binding activity (IFNL3 MutARE, IFNL3 delARE1–3; see Figure 4B), a significant increase in luciferase activities was determined in shKSRP cells compared with control cells. In cells transfected with a construct that lacks KSRP-binding site (IFNL3 ARE1), no such differences were observed (Figure 5). Altogether, these results imply that KSRP binding to sequences distinct to the three already characterized AREs is involved in the regulation of IFNL3 expression.
Luciferase expression of mutated IFNL3 3′-UTR constructs in the absence of KSRP.
Intracellular binding of KSRP to IFNL3 3′-UTR
To confirm the results of our pull-down in vitro analyses, we determined KSRP binding to the IFNL3 3′-UTR reporter gene construct in DLD-1 cells by immunoprecipitation experiments followed by qRT-PCR analysis. The human IFNL3 3′-UTR luciferase reporter vectors pCMVcontrolInglo-IFNL3-3UTR-A and pCMVcontrolInglo-IFNL3-ARE1 (negative control, no KSRP binding, see Figure 4B) were transfected into DLD-1 control cells without diminished KSRP expression. Twenty-four hours after transfection, cells were lysed and immunoprecipitation was performed with a polyclonal KSRP antibody (IP anti-KSRP). As negative control, an immunoprecipitation with an IgG control antibody was performed (IP anti-IgG). Western blot analyses confirmed successful immunoprecipitation of KSRP protein (Figure 6A). Firefly luciferase mRNA bound to immunoprecipitated KSRP protein was detected by qRT-PCR using specific primers that recognize only the spliced mRNA, but not the intron-containing plasmid DNA (see Supplementary Figure S2). We subtracted CT values of IgG control immunoprecipitations from CT values of immunoprecipitations performed with KSRP antibody to normalize for unspecific background signals. To quantify enrichment of luciferase reporter transcript bound by KSRP protein, we normalized the data of cells transfected with the IFNL3 3′-UTR (pCMVcontrolInglo-IFNL3-3UTR-A) to data of cells that expressed a reporter transcript with no KSRP-binding site (IFNL3 ARE1, Figure 4B). Compared with the negative control (IFNL3 ARE1), we detected a three- to fourfold increase in IFNL3 3′-UTR luciferase reporter gene mRNA (IFNL3 A) bound by KSRP (Figure 6B). This supports our hypothesis that KSRP is able to interact directly with IFNL3 3′-UTR in intact cells.
Intracellular binding of KSRP to IFNL3 3′-UTR.
To prove that KSRP binds to the endogenous IFNL3 mRNA, we performed immunoprecipitation qRT-PCR experiments as described above using DLD-1 control cells (DLD-1 pGIPZ CO) and DLD-1 shKSRP cells (DLD-1 pGIPZ shKSRP, negative control), with diminished KSRP protein expression. Compared with the negative control, we detected two times more IFNL3 mRNA bound by KSRP in DLD-1 control cells (Figure 6C). These data imply intracellular binding of KSRP to endogenous IFNL3 mRNA.
Binding of KSRP to IFNL3 3′-UTR alters mRNA decay
To verify that KSRP interaction with IFNL3 3′-UTR reduces mRNA stability, we determined the half-life of the luciferase IFNL3 3′-UTR mRNA in DLD-1 cells with diminished KSRP expression (shKSRP), compared with DLD-1 control cells with normal KSRP expression (control). IFNL3 3′-UTR reporter gene construct pCMVcontrolInglo-IFNL3-A was transfected in DLD-1 cells with or without diminished KSRP expression. Forty hours after transfection, DRB was added to inhibit RNA polymerase II-dependent transcription. After 2–16 h, RNA was isolated and luciferase mRNA was quantified by qRT-PCR. As shown in Figure 7A, the half-life of luciferase mRNA containing the IFNL3 3′-UTR was enhanced from 10.40 ± 1.65 h in control cells to 15.94 ± 2.97 h in shKSRP cells. In addition, luciferase mRNA levels in cells transfected with the pCMVcontrolInglo-IFNL3-A reporter gene construct before we added DRB (time point 0 in Figure 7A) were significantly higher in cells with reduced KSRP expression (Figure 7B). These results confirm the negative regulatory KSRP function on IFNL3 3′-UTR.
Absence of KSRP enhanced stability of IFNL3 3′-UTR.
As expected, no KSRP effect on mRNA half-life or initial mRNA content was observed when a reporter gene without KSRP-binding sites (pCMVcontrolInglo-IFNL3-ARE) was used (Figure 7C,D). Altogether, these data indicate that the modification of IFNL3 3′-UTR A reporter mRNA decay is mediated by a KSRP-specific effect.
In addition, we analyzed mRNA decay of endogenous IFNL3 mRNA in DLD-1 control and shKSRP cells. In accordance with the data described above, we observed prolonged half-life of IFNL3 mRNA in shKSRP cells (half-life = 9.68 ± 5.51 h) compared with control cells (half-life = 3.25 ± 0.74 h), indicating a negative regulatory effect of KSRP on IFNL3 expression (Figure 7E).
KSRP regulates Ifnl3 mRNA expression in vivo
To evaluate whether type III IFN expression is also regulated by KSRP in vivo, we performed experiments with cells derived from KSRP+/+ (WT) and KSRP−/− (KO) mice. Knockdown of KSRP protein in murine KSRP−/− cells (data not shown) and immunoprecipitation of murine KSRP protein was confirmed in Western blot experiments (Supplementary Figure S3). We determined by IP qRT-PCR experiments whether KSRP binds to endogenous Ifnl3 mRNA transcript. Immunoprecipitation was performed with protein lysates from splenic T cells obtained from KSRP+/+ (WT) mice and KSRP−/− (KO) mice, which served as negative controls in our experiments. mRNA bound to immunoprecipitated KSRP protein was determined by qRT-PCR. Compared with cells derived from KSRP−/− (KO) mice, we observed a threefold increase in Ifnl3 mRNA in cells from KSRP+/+ (WT) mice upon KSRP immunoprecipitation (Figure 8A). In addition, Ifnl3 mRNA expression was determined in total mRNA isolated from splenic T cells from mice of both genotypes to demonstrate Ifnl3 mRNA expression in both groups (input). Ifnl3 mRNA was detected in immunoprecipitates from cells of WT mice only, but not in immunoprecipitates obtained from cells of KSRP−/− mice (Figure 8B). These data may indicate that KSRP could be able to interact with endogenous IFNL3 mRNA.
KSRP binds to endogenous IFNL3 mRNA.
Analysis of type III interferon expression in different inflammation models
In spleen cells isolated from KSRP WT and KO mice, we observed a significant increase in Ifnl3 mRNA expression in cells from KSRP−/− mice (KO, Figure 9A). This result strengthens our hypothesis that KSRP also negatively regulates type III IFN expression in mice. It has been reported that IFNL can have protective functions in different inflammatory settings such as colitis, arthritis or allergic airway diseases [14,16,17]. Therefore, we were interested whether KSRP-mediated post-transcriptional regulation of type III IFN expression is important in an inflammatory setting in vivo. We started to investigate this question in a model of LPS-induced acute inflammation: we compared mRNA levels of type III IFNs in spleen tissue isolated from KSRP+/+ (WT) and KSRP−/− (KO) mice treated for 6 h with 20 mg/kg LPS. As expected, we detected increased expression of Ifnl3 mRNA in the solvent control group (0.9% NaCl) of KSRP−/− mice compared with control mice (Figure 9B). These results confirm the negative regulatory function of KSRP detected in the above-described experiments in spleen cells. The pro-inflammatory LPS stimulus in KSRP−/− mice decreased Ifnl3 mRNA expression to the level of WT mice (LPS, Figure 9B). Also, LPS-mediated decrease in Ifnl3 mRNA expression was detected in WT animals (Figure 9B and Supplementary Figure S4). Furthermore, for Ifnl2 mRNA expression, a similar KSRP-mediated regulation was observed as for Ifnl3 (Supplementary Figure S5).
Inactivation of the KSRP gene enhances Ifnl3 mRNA expression, whereas inflammatory conditions lead to a decrease in Ifnl3 mRNA expression in murine spleen.
Similar findings were obtained in CAIA, an established model of a chronic inflammatory disease. CAIA was induced in WT and KSRP−/− (KO) mice as described recently . Type III IFN expression was analyzed in spleens of WT and KSRP−/− mice treated with (CAIA) or without (control) monoclonal antibodies against collagen. Again, we detected increased expression of Ifnl3 mRNA in the solvent-treated control group of KSRP−/− mice. Similar to the results obtained in the LPS model also in CAIA-treated KSRP−/− mice, the expression of Ifnl3 mRNA was reduced compared with solvent-treated KSRP−/− mice (Figure 9C). Altogether, these results implicate that KSRP negatively regulates type III IFN expression.
Type III IFNs are important for antiviral immune responses. In addition, they have immune modulatory functions. A tight regulation of the expression of these cytokines is necessary to avoid detrimental consequences of an overwhelming immune response. Regulation of gene expression occurs on different levels. Besides transcriptional mechanisms, post-transcriptional modifications also determine the protein amount of type III IFNs . The importance of AREs for the regulation of type III IFN expression has been documented . Nevertheless, only very limited information exists which RBPs are involved in these processes. IFNL2 and IFNL3, both members of the type III IFN family, share high sequence similarity and both contain AREs in their 3′-UTR sequences that have been shown to mediate transcript instability . miRNA-mediated mechanisms of post-transcriptional control during hepatitis C virus (HCV) infection have been described especially for IFNL3. Upon HCV infection, the expression of two miRNAs (miR-208b and miR499a-5p) is increased. These two miRNAs target the 3′-UTR of IFNL3 and decrease IFNL3 transcript stability and thus reduce IFNL3 expression, and therefore, HCV clearance is impaired. A single-nucleotide polymorphism (G/T) in IFNL3 3′-UTR abolishes binding of those miRNAs and, consequently, increases IFNL3 transcript stability and improves elimination of HCV . This example shows that regulation of type III IFN mRNA stability is important in disease pathogenesis. The presence of AREs in the 3′-UTR implies the involvement of AMD mechanisms in the post-transcriptional control of IFNL3 expression. AMD is regulated by RBPs that recruit enzymes, such as poly(A)-specific ribonuclease or the exosome complex necessary for degradation of the mRNA, to their target sequences [31,41]. As transcriptional regulation of type III IFN expression shares some similarities with type I IFNs , we speculated that this might also be the case for post-transcriptional control involving AMD. For type I IFNs, Lin et al.  demonstrated that the RBP KSRP is important for the initiation of AMD. They demonstrated in immortalized murine embryonic fibroblasts (MEFs) isolated from KSRP−/− mice that mRNA half-life of IFN-α4 and IFN-β is increased upon poly I:C stimulation compared with MEF from WT mice. The prolonged mRNA half-life of type I IFNs resulted in increased protein expression, and this resulted in an improved herpes simplex virus-l resistance of KSRP−/− mice. For these reasons, we postulated that KSRP might also be involved in the post-transcriptional control of the closely related type III IFNs.
Indeed, we detected binding of KSRP to 3′-UTR of type III IFNs and demonstrated that knockdown of KSRP protein increases mRNA stability of IFNL3 reporter transcript as well as endogenous IFNL3 mRNA. However, KSRP did not interact with the AUUUA motives that have been described to be important for HCV-mediated regulation of IFNL3 mRNA. In general, AREs can be divided into class I, II and III elements. Class I AREs have multiple copies of AUUUA sequence elements surrounded by U-rich regions and are found in mRNAs coding for nuclear transcription factors like c-myc or c-fos. Class II AREs also contain multiple, often overlapping AUUUA motives. Such motives are typically found in 3′-UTR sequences of cytokines like TNF-α or IL-2. Class III AREs do not contain AUUUA sequences, and only U-rich stretches serve as RBP-binding sites . ARE1–3 of IFNL3 3′-UTR resembles class I and II ARE-binding sites, but according to our analyses these are not the binding sites of KSRP. Additionally, the IFNL3 3′-UTR contains a U-rich stretch which could be classified as class III element and this sequence is likely to be responsible for KSRP interaction with IFNL3 3′-UTR. The affinity of KSRP binding to ARE has been studied intensely. The RNA-binding domain of KSRP consists of four hnRNP K homology (KH) domains (KH1–4), with KH3 displaying the highest binding activity to ARE sequences. However, previous analyses of the sequence preference of the KH domains of KSRP have shown that the protein per se does not have a high binding specificity for AU-rich sequences [44,45]. That means KSRP recognizes and binds to other sequence elements as well. Therefore, it is quite likely that interaction with other sequence motives also initiates KSRP-mediated mRNA decay. This idea is supported by our findings that depletion of KSRP protein in DLD-1 cells enhances luciferase activity of INFL3 3′-UTR reporter gene constructs lacking type I and II ARE sequences compared with control cells.
Our data indicate that KSRP is one important player in regulation of IFNL3 mRNA decay, but we have strong evidence that other RBPs are also involved in post-transcriptional regulation of IFNL3 mRNA stability since shRNA-mediated reduction in KSRP protein expression does not fully restore luciferase activity in our IFNL3 3′-UTR reporter gene constructs compared with control constructs. Alternatively, this phenomenon might be explained by residual KSRP protein expression in shKSRP cells. Our data demonstrate that deletion or mutation of ARE sequences that are not involved in KSRP binding to IFNL3 3′-UTR also partially restore luciferase activity of IFNL3 reporter gene constructs, rather favor the hypothesis that besides KSRP other regulators of IFNL3 mRNA expression exist. Accordingly, the pCMVcontrolInglo-IFNL3-ARE1 construct, which contains one ARE, which, however, is not a binding site of KSRP, still shows mRNA decay. Most probably, transcript instability of INFL3 ARE1 is mediated by other RBP than KSRP. We detected binding of the p40 and p42 isoforms of AUF1 to ARE sequences in IFNL3 3′-UTR in our pull-down analyses. AUF1 is an intensely studied RBP that has been shown to promote rapid ARE-mRNA degradation . The binding site of AUF1 to IFNL3 3′-UTR is distinct from the one of KSRP, so both proteins may co-operate in post-transcriptional regulation of IFNL3 expression.
Compared with full-length IFNL3 3′-UTR (IFNL3 A), we detected enhanced binding of KSRP to the IFNL3 delARE1–3 construct in pull-down experiments. Also other RBPs, for example AUF1, seem to be involved in the post-transcriptional regulation of IFNL3 expression by interaction with ARE sequences. The absence of ARE sequences in the IFNL3 delARE1–3 construct most probably abrogates binding of those RBPs to IFNL3 3′-UTR, and this could facilitate binding of KSRP to its target sequence. This could be a possible explanation for the results observed in our pull-down analyses.
Owing to the high sequence similarity to IFNL3 (>90%), KSRP-mediated regulation of IFNL2 mRNA stability seems very likely. Indeed, preliminary data from our laboratory are in favor of this hypothesis (data not shown). KSRP and AUF1 are the first RBPs identified that regulate type III IFN transcript stability. In contrast with type I IFN , the mRNA-stabilizing RBP HuR does not bind to type III IFN 3′-UTR. These data reveal differences between the post-transcriptional regulation of type I and III IFN mRNA stability.
Data in spleen cells from KSRP−/− mice confirm the importance of KSRP in post-transcriptional control of IFNL3 mRNA expression in vivo. Our preliminary data indicate that in vivo Ifnl3 mRNA expression is reduced in acute or chronic inflammatory disease situations. These findings are in line with data derived from primary human bronchial epithelial cells from asthmatic subjects. In these cells, less IFNL3 was detected compared with normal subjects. The reduced IFNL3 expression seems to correlate with the severity of allergic asthma and allergic asthma exacerbation . This is in line with reports suggesting a protective anti-inflammatory function of IFNL based on its ability to reduce migration and ROS production of neutrophils. Therefore, it is tempting to speculate that decrease in IFNL3 expression upon inflammation would enable enhanced neutrophil activation and recruitment to inflammatory sites. Several reports describe consistently that the addition of recombinant type III IFNs ameliorates inflammation in rheumatoid arthritis  and colitis  models.
As stated above, we detected reduced amounts of Ifnl3 expression in inflammatory situations in KSRP WT as well as in KSRP knockout cells. The decline of Ifnl3 mRNA in KSRP−/− mice upon inflammatory conditions (LPS or CAIA treatment) may indicate that other destabilizing RBPs, like AUF1 or TTP, are involved in AMD of Ifnl3 mRNA under these conditions. This would fit with our data from pull-down and reporter gene experiments, which indicate that other RBPs are involved in the control of IFNL3 mRNA stability.
Concerning the biological role of type III IFNs, their antiviral effects have been extensively studied. However, evidence exists that they also modulate immune functions in inflammatory and autoimmune diseases. Some reports exist that type III IFNs are able to induce proliferation of T-regulatory cells  or to modify inflammatory processes in a murine model of allergic airway disease . Moreover, it has been demonstrated that administration of IFNL2 suppresses induction of collagen-induced arthritis by inhibition of neutrophil function and IL-1β production , clearly indicating an anti-inflammatory property of type III IFNs in a model of chronic inflammatory autoimmune disease. In the human autoimmune disease SLE, elevated levels of IFNL2 (IL-28) were detected in serum and in activated CD4+ T cells of patients compared with healthy control subjects , indicating rather a pro-inflammatory function of type III IFNs, at least in this disease.
In summary, our data demonstrate that the RBP KSRP modulates half-life of IFNL3 mRNA. We identified binding of KSRP to the 3′-UTR of IFNL3 and demonstrated functional relevance of this interaction. Moreover, we confirmed the importance of KSRP for IFNL2 and IFNL3 AMD also in vivo. The role of type III IFNs in inflammation remains to be elucidated as well as the importance of KSRP-mediated regulation of mRNA half-life in this situation.
AU-rich element RNA-binding protein 1
collagen antibody-induced arthritis
calf intestinal alkaline phosphatase
fetal bovine serum
hepatitis C virus
interferon lambda 3
KH-type splicing regulatory protein
murine embryonic fibroblasts
Passive Lysis Buffer
quantitative real-time reverse transcription polymerase chain reaction
shRNA with KSRP
systemic lupus erythematosus
tumor necrosis factor-α
L.S. contributed to the experimental design, data acquisition, data analysis and interpretation, and manuscript preparation. S.S. contributed to data acquisition. K.S., R.K. and F.G. contributed to animal experiments. J.W-M. and H.K. contributed to experimental design, data interpretation and manuscript preparation. A.P. contributed to the experimental design, data analysis and interpretation, and manuscript preparation. All authors reviewed the results and approved the final version of the manuscript.
This work was supported by grant 961-386261/1105 (to A.P.) from the Innovation Foundation of the State of Rhineland-Palatinate and the German Research Foundation (DFG) grant PA1933/2-3, PA1933/3-1 (both to A.P.) and KL1020/10-1 (to H.K.).
We thank Ms Ihrig-Biedert and Ms Göllner of the Department of Pharmacology of the University Medical Center of the Johannes Gutenberg-University Mainz for their excellent technical assistance. We thank Ms Closs and Mr Michel for their assistance during manuscript preparation.
The Authors declare that there are no competing interests associated with the manuscript.