Cellular and Molecular Life Sciences (2021) 78:2963–2985 https://doi.org/10.1007/s00018-020-03695-5 Cellular and Molecular Life Sciences ORIGINAL ARTICLE Human RNase3 immune modulation by catalytic‑dependent and independent modes in a macrophage‑cell line infection model Lu Lu1,3 · RanLei Wei2 · Guillem Prats‑Ejarque1 · Maria Goetz1 · Gang Wang2 · Marc Torrent1 · Ester Boix1 Received: 25 February 2020 / Revised: 21 September 2020 / Accepted: 31 October 2020 / Published online: 23 November 2020 © The Author(s) 2020 Abstract The human RNase3 is a member of the RNaseA superfamily involved in host immunity. RNase3 is expressed by leukocytes and shows broad-spectrum antimicrobial activity. Together with a direct antimicrobial action, RNase3 exhibits immu- nomodulatory properties. Here, we have analysed the transcriptome of macrophages exposed to the wild-type protein and a catalytic-defective mutant (RNase3-H15A). The analysis of differently expressed genes (DEGs) in treated THP1-derived macrophages highlighted a common pro-inflammatory “core-response” independent of the protein ribonucleolytic activity. Network analysis identified the epidermal growth factor receptor (EGFR) as the main central regulatory protein. Expression of selected DEGs and MAPK phosphorylation were inhibited by an anti-EGFR antibody. Structural analysis suggested that RNase3 activates the EGFR pathway by direct interaction with the receptor. Besides, we identified a subset of DEGs related to the protein ribonucleolytic activity, characteristic of virus infection response. Transcriptome analysis revealed an early pro-inflammatory response, not associated to the protein catalytic activity, followed by a late activation in a ribonucleolytic- dependent manner. Next, we demonstrated that overexpression of macrophage endogenous RNase3 protects the cells against infection by Mycobacterium aurum and the human respiratory syncytial virus. Comparison of cell infection profiles in the presence of Erlotinib, an EGFR inhibitor, revealed that the receptor activation is required for the antibacterial but not for the antiviral protein action. Moreover, the DEGs related and unrelated to the protein catalytic activity are associated to the immune response to bacterial and viral infection, respectively. We conclude that RNase3 modulates the macrophage defence against infection in both catalytic-dependent and independent manners. Keywords Ribonucleases · Host defence · EGFR · Infection · Mycobacterium aurum · Respiratory syncytial virus · Transcriptome Introduction RanLei Wei, Guillem Prats-Ejarque contributed equally to this Human RNase3, also known as the Eosinophil Cationic work. Protein (ECP), is a member of the ribonuclease A super- family. It is a small and highly cationic protein that partici- Electronic supplementary material The online version of this article (https​://doi.org/10.1007/s0001​8-020-03695​-5) contains pates in the immune defence response. RNase3 is mainly supplementary material, which is available to authorized users. expressed in eosinophils, the mature protein is stored in the eosinophil secondary granules and is secreted upon infec- * Ester Boix tion and inflammation [1–3]. Together with its main produc- Ester.Boix@uab.es tion source, the protein is also reported to be expressed in 1 Department of Biochemistry and Molecular Biology, other leukocytes, such as neutrophils and macrophages [4]. Faculty of Biosciences, Universitat Autonoma de Barcelona, The importance of the selective expression and secretion Cerdanyola del Vallès, Spain of RNase3 has been emphasized due to its association with 2 Center of Precision Medicine and Precision Medicine Key multiple diseases, such as asthma, intestinal tract inflam- Laboratory of Sichuan Province, West China Hospital, mation or autoimmune disorders [1, 2, 4, 5]. The protein is Sichuan University, Chengdu, China routinely used as a clinical diagnostic marker of eosinophil 3 Present Address: College of Animal Science and Technology, activation during inflammatory processes [2]. Secretion of Sichuan Agricultural University, Chengdu, Sichuan, China 13 Vol.:(0123456789) 2964 L. Lu et al. RNase3 is also induced by damaged epithelia and the pro- revealed that RNase3 can inhibit both bacterial and viral tein can participate in tissue healing and remodelling [4]. infection through the modulation of the macrophage immune In addition, RNase3 expression is induced by infection and response in both ribonucleolytic-dependent and independ- might contribute to the protection of biological fluids. The ent ways. protein exhibits antimicrobial activity against a wide range of microorganisms, such as bacteria, yeast, viruses, and parasites [3, 6–10]. Abundance of surface-exposed cationic Materials and methods and hydrophobic residues can mediate the protein bind- ing and subsequent destabilization of bacterial membranes Cell culture through a carpet-like mechanism characteristic of many host defence antimicrobial peptides (AMPs) [10–12]. Although Human THP1 cells (NCTC #88081201) were maintained the protein triggers the pathogen death mainly through a or passaged in 25 or 75 cm2 tissue culture flasks (BD Bio- direct mechanical action at the cell envelope, the targeting sciences) using RPMI-1640 (Lonza, BE12-702F) medium of intracellular components, such as nucleic acids cannot be with 10% heat-inactivated foetal bovine serum (FBS) at disregarded. Indeed, RNase3 internalization is observed in 37 °C, humidified 5% C­ O2 conditions. THP1 cells used in treated yeast cells and protozoa [8, 13]. Besides, RNase3 can this study were controlled below passage 25. THP1 cells enter the macrophage and eradicate the intracellular dwelling were treated with 50 nM phorbol myristate acetate (PMA, bacteria [14]. On the other hand, the protein antiviral activity Sigma-Aldrich, P8139) for 48 h to induce differentiation into on single stranded RNA viruses has been directly correlated macrophage-like cells and allowed to rest for 24 h before to its ribonucleolytic action [9]. Together with the RNase3 further treatment. direct antimicrobial activity, a series of evidences illustrated that RNase3 also plays an immunomodulatory role during host defence [2, 4]. Early studies showed that RNase3 can Recombinant protein expression and purification activate rat mast cell and induce histamine release, mediat- ing the cross talk between mast cells and eosinophils [15]. RNase3 and RNase3-H15A recombinant proteins were RNase3 displays remodelling activity partly mediated by produced as previously reported [8]. Briefly, E. coli inducing the expression of epithelial insulin-like growth BL21(DE3)/pET11c cells were induced by 1 mM Isopropyl factor 1 (IGF1) [5] and the fibroblast chemotaxis can be β-d-1-thiogalactopyranoside (IPTG, St. Louis, MO, USA) enhanced by RNase3 release into the injured tissue site [2, and the inclusion bodies enriched pellet was resuspended in 16]. Recently, we observed that RNase3 expression in mac- 80 mL of 10 mM Tris–HCl pH 8.5, 2 mM EDTA and left rophages is regulated by mycobacteria infection. Moreover, incubating 30 min with 40 µg/mL of lysozyme prior to soni- we demonstrated the protein ability to enter macrophage cation. Following, the sample was centrifuged at 30.000 × g cells and induce autophagy, contributing thereby to the for 30 min at 4 °C and the pellet was resuspended in 25 mL eradication of the intracellular infection [14]. However, how of the same buffer with 1% triton X-100 and 1 M urea and RNase3 regulates the macrophage function and whether the was left stirring at room temperature for 30 min and then protein immunomodulatory properties are dependent on its centrifuged 30 min at 22.000 × g. Following, 200 mL of ribonucleolytic activity remains to be clarified. 10 mM Tris–HCl pH 8.5, 2 mM EDTA was added to the In the present study, we have investigated how RNase3 pellet, and then the sample was centrifuged at 22.000 × g for modulates the transcriptome profile of the macrophage cell. 30 min (4 °C). The resulting pellet solubilized in 6 M guani- Towards this end, we used the human monocytic cell line dine hydrochloride and rapidly 80-fold diluted in the refold- THP1 derived to macrophage and applied the next-gener- ing buffer was left in gentle stirring for 48–72 h at 4 °C. The ation RNAseq methodology to analyse the cell response folded protein was then concentrated, dialyzed against the to protein treatment. In addition, we explored the poten- chromatography buffer and purified first by cation chroma- tial contribution of the protein catalytic activity on mac- tography using a Resource S (GE Healthcare Life Sciences) rophages using a mutant enzymatically defective mutant column and then by reverse phase chromatography on a (RNase3-H15A). Notably, this is the first systematic study Vydac C4 (ThermoFisher Scientific) column. Sample purity to investigate the biological function of human RNase3 in (> 99%) was checked by SDS-PAGE and MS spectrometry. macrophages by the whole transcriptome analysis. Comple- Complementary, recombinant proteins corresponding to mentarily, we built an RNase3 overexpression THP1 cell RNase3 wild-type (rRNase3-97Arg, R97) and the arginine to line by CRISPRa to validate the transcriptome results and threonine variant (rRNase3-97Thr, T97), expressed in insect evaluated the direct contribution of the endogenous protein cells using the pFASTBAC baculovirus expression system, in the eradication of intracellular infection by M. aurum and were purchased from Diagnostics Development company. the human respiratory syncytial virus (RSV). The results 13 Human RNase3 immune modulation by catalytic-dependent and independent modes in a… 2965 Both protein variants had been purified and quantified as and non-coding genes were filtered out using biomaRT [20] previously described [17]. before passing to the Bioconductor package DESeq2 [21]. The resulting P values were adjusted using Benjamini and Treatment of THP1 cells derived to macrophages Hochberg’s approach for controlling the false discovery rate with RNase3 and RNase3‑H15A (FDR). Genes with an adjusted P value (Q value) < 0.01 & log2FC absolute value > 1 found by DESeq2 were assigned THP1 cells were differentiated into macrophage by 50 nM of as differentially expressed genes (DEGs). PMA treatment as previously described [14]. The cells were GO enrichment and KEGG pathway enrichment analysis then washed three times with prewarmed PBS and replaced of the differential expression of genes across the samples with fresh RPMI + 10%FBS medium. THP1-derived mac- was carried out using the clusterProfiler R package [22]. rophages were then treated with 10 μM of the recombinant Protein–protein interactions between DEGs were analysed proteins, wild-type (wt) RNase3 or the RNase3-H15A using NetworkAnalyst which integrates the experimentally mutant. Previous work confirmed that the selected protein validated interactions database, InnateDB database [23]. The concentration ensured an effective antimicrobial activity network was visualized using Cytoscape software (https​:// and was non-toxic to the macrophage cells [14]. After 4 and cytos​cape.org). 12 h of treatment, the cells were washed three times with pre-warmed PBS and collected for further RNA extraction. Molecular modelling Three biological repeats were carried out for each experi- ment. Alternatively, prior to treatment with the protein, cells HADDOCK2.2 (Utrecht Bioinformatics Center, University were pre-incubated with a monoclonal antibody against the of Utrecht) was applied to perform the modelling of the pro- extracellular domain of the EGFR, anti-EGFR Ab Cetuxi- tein complexes and predict the associated free energies [24, mab, Abcam) at 10 μg/mL for 1 h at 4 °C, as previously 25]. described [18]. Construction of an RNase3 overexpression THP1 cell RNA isolation and sequencing line Total RNA was extracted using the mirVanaTM miRNA We applied CRISPRa to activate the endogenous expression Isolation Kit (Ambion, Life Technologies, AM1560) as of RNase3 in THP1 [26]. Five distinct sgRNAs targeting the described by the manufacturer. RNA purity was deter- region from 100 to 500 bp relative to the transcription start mined by spectrophotometry and RNA integrity (RIN) sites (TSS) of the RNase3 gene were designed using Cas9 was analysed using Agilent 2100 Bioanalyzer (Table S1). Activator Tool (https​://sam.genom​e-engin​eerin​g.org/datab​ Following RNA extraction, total RNAs were submitted to ase/) [26]; the sgRNA sequence is listed in Table S2. The the CNAG-CRG Sequencing Service (Centre for Genomic pLenti239G plasmid, used to co-express the dCas9-VP64 Regulation, Barcelona) for cDNA library preparation,polyA fusion protein and EGFP, was constructed by inserting the enrichment and NGS sequencing. Sequencing libraries were T2A-EGFP cassette from the LentiCRISPRv2-GFP-puro prepared according to protocols provided by Illumina. 50 bp- (gifted by Manuel Kaulich) in place of the T2A-BSD cas- long single-end sequencing was carried out in an Illumina sette of the lenti-dCAS-VP64-Blast (Addgene61425, gifted HiSeq2500 sequencer with a depth of > 20 million reads by Manuel Kaulich). The pLenti239R, used to co-express per sample. Raw sequence reads have been deposited in the the gRNA, cherry red fluorescence protein, and puromycin NCBI Sequence Read Archive (SRA) under accession num- resistance gene, was constructed by displacing the Cas9- ber PRJNA574982. T2A-EGFP of LentiCRISPRv2-GFP-puro (gifted by Manuel Kaulich) with Cherry red fluorescent gene (gifted by Marcos Transcriptome analysis Gil García, UAB). The sgRNA sequences were inserted in pLenti239R at downstream of the U6 promoter using the FastQC was used to carry out the quality assessment of BbsI cloning sites. The correct construction of the plasmids reads, assessing the distribution of phred quality scores was validated by Sanger sequencing. and mean percentage GC content across each read. Reads Following, HEK293T cells were used for lentiviral pro- were aligned to the latest human genome assembly from the duction. HEK293T cells were maintained in DMEM + 10% Genome Reference Consortium (GRCh38) using HISAT2 FBS complete medium in a 5% C ­ O 2 humidified incu- [19]. Aligned reads were stored in the SAM file format. bator at 37 °C. Cells were co-transfected with psPAX2 StringTie was used to assemble the alignments into tran- (Addgene#35002, gifted by Marina Rodriguez Muñoz) scripts and estimate the expression levels of all genes and packaging plasmid, pMD2.G (Addgene#12259, gifted transcripts [19]. Low expression (sum count less than 10) by Marina Rodriguez Muñoz) envelope plasmid, and 13 2966 L. Lu et al. pLenti-239G (encoding dCAS9 and eGFP) or pLenti-239R antibody (Abcam, ab207429), or anti-p44/42 MAPK (Cell (encoding sgRNA and Cherry red fluorescent marker) using Signaling, 9106S) overnight at 4 °C. After washing, the calcium phosphate precipitation protocol [27]. The trans- membranes were treated for 1 h at room temperature fection medium was replaced with fresh medium follow- (RT) with horseradish peroxidase (HRP)-conjugated goat ing overnight incubation. Supernatants were collected 24, anti-rabbit IgG (Sigma Aldrich, 12-348) or HRP-conju- 48 and 72 h after transfection, centrifuged to remove cell gated horse anti-mouse IgG (Cell Signaling, 7076P2), debris, filtered using 0.45 μm filters and concentrated using respectively. Finally, the membranes were exposed to an the PEG6000 precipitation method [28]. Viral pellets were enhanced chemiluminescent detection system (Supersig- resuspended in PBS, aliquoted and stored at – 80 °C until nal West Pico Chemiluminescent Substrate, ThermoFisher use. THP1 monocytes were infected with 20 μL concentrated Scientific, 32209) for detection. As a control, GAPDH was lentivirus in the presence of 8 μg/mL polybrene and incu- detected with chicken anti-GAPDH antibodies (Millipore). bated overnight. Next day, the cells were replaced with fresh medium and cultured for 72 h. Fluorescence positive mono- cytes were checked by fluorescence microscopy and then Macrophage infection by mycobacteria and CFU sorted by Cell sorter. The fluorescence positive cells were assay evaluated by suspending the cells in PBS and fixing them with 2% paraformaldehyde for 10 min prior to flow cytom- Infection of THP1 cells derived to macrophage by myco- eter. The fluorescence positive cells were sorted into single bacteria was performed as previously described [14]. cells by Cell sorter BD FACSJazz. M. aurum was purchased from the UK National Col- lection of Type Cultures (NCTC). Cells cultures of M. aurum (NCTC, 10437 were grown in Middlebrook (MB) Real‑time qPCR assays 7H9 broth (BD Biosciences, 271310) enriched with 10% (v/v) albumin/dextrose/catalase (ADC; BD Biosciences, Following the treatment of THP1-derived macrophages 212352) containing 0.05% Tween 80, and in MB7H10 (BD with recombinant RNase3 proteins the transcriptional Biosciences, 262710) with 10% (v/v) oleic acid/albumin/ expression profiles of selected genes were measured by dextrose/catalase (OADC; BD Biosciences, 212240) for RT-qPCR with GAPDH as the internal control gene. semi-solid agar growth at 37 °C. Stock cultures of log- Recombinant EGF (ThermoFisher) was used as a posi- phase cells were maintained in glycerol (25% final concen- tive control in some assays. Primers used for RT-qPCR tration) at − 80 °C. The bacteria were vortexed and soni- validation are listed in Table S2. The same total RNA was cated using ultrasound sonication bath to obtain a single reverse transcribed into cDNA using iScriptTM cDNA cell suspension, and then the bacterial concentration was Synthesis Kit (Bio-Rad, 1708891). RT-qPCR assays were determined by measuring the optical density (OD) of the performed in 20 μL using the iTaq Universal SYBR Green culture at 600 nm (OD = 109 CFU/mL). Mid-log phase M. Supermix (Bio-Rad, 1725121) according to the manufac- aurum cells, harvested in RPMI-1640 complete medium, turer’s instructions. The reactions were conducted using were co-cultured with macrophages at a multiplicity of CFX96 Real-Time PCR detection system (Bio-Rad, Her- infection (MOI) of 10:1 and were incubated at 37 °C for cules, CA, USA) under the following conditions: 95 °C 3 h, then were washed three times with PBS and replaced for 2 min; 40 cycles of 95 °C for 15 s and 60 °C for 30 s; with fresh media supplied with 50 μg/mL gentamycin melting curve generation (60–95 °C). The relative expres- (Apollo Scientific, BIG0124) to remove extracellular sion ratios were calculated using the 2 − ΔΔCT method mycobacteria during further treatment. [29, 30]. Colony forming units (CFU) counting assay was applied to compare the infectivity and living rate of M. aurum toward THP1-derived macrophages with or without Western blot RNase3 overexpression. 2 × 105 THP1 cells were seeded in 24-well plates per well and induced to macrophages by For the western blot assays to detect RNase3, cells were 50 nM of PMA treatment. Next, log-phase cultures of M. harvested and lysed with RIPA buffer. Samples to analyse aurum ­(OD600 ~ 1) were diluted to 2 × 106 CFUs/mL and pMAPK were extracted with TRIzol™ (ThermoFisher) were used to infect macrophages derived from wild-type and resuspended in 1 M of urea, 2% of SDS and 5 mM macrophage (WT) or RNase3-overexpression macrophage EDTA. The samples were separated by 10% SDS-PAGE. cells (OX) at a multiplicity of infection of 10:1 at 37 °C Then, the samples were transferred to polyvinylidene dif- for 3 h. Then the cells were washed 3 times with PBS and luoride membranes, blocked with either 5% non-fat milk replaced with fresh media supplied with 50 μg/mL genta- or BSA in TBST, and incubated with anti-RNase3 primary mycin to remove extracellular mycobacteria, referenced as 13 Human RNase3 immune modulation by catalytic-dependent and independent modes in a… 2967 the “0 h post-infection” time point. At 0 h, 24 h, 48 h, and (GAPDH) control was used for relative quantification of the 72 h, the cells were washed, collected, lysed with distilled intracellular virus. The primers and probe used are listed in water, and plated in 10 mm petri dish containing MB7H10/ Table S2. OADC/agar. The CFU were counted after 2 weeks. Five independent experiments were conducted for this assay. Cell viability assay Macrophage infection by RSV virus and viral Cell viability was measured using MTT assay as previously quantification reported [35]. After each time point of post of infection, cells were incubated with 100 μL of MTT reagent. The media Human respiratory syncytial virus (RSV, ATCC, VR-1540) were then removed and 200 μL of DMSO was added per stock was ordered from ATCC. Hela cells were used to pro- well. The absorbance of the formazan was determined at duce RSV under biosafety level II conditions [31]. Briefly, 570 nm in an ELISA reader. Hela cells are plated in 75 cm2 culture flask and incubated at 37 °C in DMEM + 10% FBS until they are approximately 50% confluent. The cells were then washed and infected Results with RSV stock under a MOI of 0.1. After 3 h infection, the cells were washed and replaced with fresh medium Comparative transcriptome analysis (DMEM + 10%FBS) and incubated for 4 days at 37 °C , 5% of THP1‑derived macrophages treated ­CO2. The cells and the virus suspension were collected when with wild‑type RNase3 and catalytic‑defective the cytopathology appeared, with scraping and vortexing of RNase3‑H15A mutant the cells to release more viral particles. The virus suspen- sion was centrifuged for 10 min at 1800 g to remove the To characterize the immunomodulatory properties of cell debris. The virus suspension without cell debris were RNase3 on macrophages we incubated THP1-derived cells concentrated using Amicon Ultra-15 centrifugal filters with in the presence of the protein and analysed the cell transcrip- 100 kDa cut-off (Millipore, UFC910024). The produced tome at 4 h and 12 h incubation time. Complementarily, to viruses were titrated by measuring the median tissue culture explore the contribution of the protein enzymatic activity to infectious dose (TCID50) in HEK293T cells [32]. its immune-regulatory action, we compared the macrophage- Before RSV infection, THP1 cells were induced to mac- induced response by wild-type (wt) RNase3 with a protein rophage by 50 nM of PMA treatment for 48 h. Cells were mutant variant (RNase3-H15A) devoid of ribonucleolytic washed three time with prewarmed PBS and replaced with activity. Our previous work indicated that the H15A substi- fresh RPMI + 10% FBS medium for 24 h incubation. After tution fully abolishes the RNase3 catalytic activity without that, macrophages were washed and incubated with RSV, any alteration of the 3D protein structure [8]. RNase activity mixing every 15 min for the first 2 h. All virus treatment contributes to the protein antifungal and antiviral properties tests were performed using RSV at a MOI of 1 TCID50/cell. [8, 36]. However, some antimicrobial properties of RNase3 RSV was detected by RT-qPCR [33]. After the indi- were reported to be independent on its catalytic action [12, cated post-infection time, the extracellular RSV virus was 37, 38]. Besides, we recently observed that both wild-type collected by the PEG6000 precipitation methodology [28, RNase3 and H15A mutant can induce the autophagy path- 34] and intracellular RSV virus were collected by lysing way and mediate the eradication of intracellular mycobacte- the macrophage cells with the lysis buffer from mirVa- ria within macrophages [14]. Therefore, we decided to carry naTM miRNA Isolation Kit (Ambion, Life Technologies, out a whole transcriptome analysis to identify the distinct AM1560). Total RNA from RSV-infected macrophage cell pathways associated to the protein action. cells as well as stock virus was extracted using mirVanaTM For each condition, three biological replicates were pre- miRNA Isolation Kit according to the manufacturer’s pared for control, wt RNase3 and RNase3-H15A treatment instructions. cDNA was synthesized using iScriptTM cDNA of THP1 macrophage-derived cells. A total of 511 million Synthesis Kit (Bio-Rad, 170-8891). The synthesis was per- reads (mReads) were obtained, providing about 28 mReads formed using random hexamers, starting with 1 μg of total per sample. Approximately, 96% of the reads were uniquely cell RNA. The RT-qPCR was performed using ddPCR ™ mapped to the Homo sapiens genome (Table S3). Expres- Supermix for Probes (Bio-Rad, 1863024). Samples with a sion of 14801 protein coding genes were detected in at least cycle threshold value of more than 40 were recorded as neg- one individual (filtered by sum count more than 10) (Addi- ative. A standard curve was prepared using serially diluted tional file 1). The Principal Components Analysis (PCA) RNA extracts from a known quantity and used to quantify plot shows tight clustering within the control group and the RSV as TCID50/mL. In parallel with the RSV probe assays, protein treated groups (either RNase3 or RNase3-H15A an endogenous glyceraldehyde-3-phosphate dehydrogenase 13 2968 L. Lu et al. treatment) for mRNA expression (Figure S1). Addition- found 30 (8 up-regulated and 22 down-regulated) and 143 ally, RNase3- and RNase3-H15A-treated samples clustered (10 up-regulated and 133 down-regulated) DEGs at 4 h separately from each other, indicating temporal alterations in and 12 h post of treatment, respectively. All the identified gene expression driven by different protein treatment. DEGs are listed in Additional file 2. Following, we used DESeq2 to identify Differently Expressed Genes (DEGs) for the following conditions: RNase3 modulates the macrophage global innate Genes with an adjusted P value < 0.01 & log2FC absolute immune response in a ribonuclease independent value, (i) RNase3 versus control, (ii) RNase3-H15A versus manner control, and (iii) RNase3-H15A versus RNase3. We identi- fied 4930 (2563 up-regulated and 2367 down-regulated) Given the high overlapping group of DEGs between RNase3 and 4933 (2584 up-regulated and 2349 down-regulated) and RNase3-H15A-treated cells either at 4 h or 12 h (Fig. 1b, DEGs at 4 h and 12 h, respectively, between RNase3- c), we sought to identify the genes with similar response treated cells and untreated control cells. When compar- profiles. Profiles with overlap between RNase3 and RNase3- ing RNase3-H15A-treated cells to untreated control cells H15A-treated cells, comprising a total of 3818 genes, were at 4 h and 12 h, respectively, 4869 (2516 up-regulated identified to conform the core response of macrophage mod- and 2353 down-regulated), and 4986 DEGs were identi- ulated by both RNase3 and RNase3-H15A. Core response fied (2585 up-regulated gene and 2401 down-regulated) genes not related to the protein ribonucleolytic activity are (Fig. 1a). Next, we identified DEGs by comparing 4 h and listed in Additional file 3 (1972 and 1846 are up-regulated 12 h timing for each treatment, 18 (10 up regulated and and down-regulated, respectively; see Fig. 1b, c). 8 down regulated), 351 (223 up-regulated and 128 down- KEGG pathway enrichment of the up-regulated DEGs regulated) and 198 (81 up-regulated and 117 down-regu- indicated that both RNase3 and RNase3-H15A treatment lated) DEGs were found in control, RNase3, and RNase3- triggered the cells overall immune response. In total, 58 H15A treatment as a function of time, respectively. The KEGG pathways were significantly enriched (padj < 0.01) small amount of DEGs between 4 and 12 h found in con- (see Fig. 2a and Additional file 4). The top ten listed path- trol samples indicated that the biology status of the cells ways included TNF signal pathway, Cytokine–cytokine has been stable during all the experiment. A direct com- receptor interaction, NF-kappa B signal pathway, Chemokine parison between RNase3 and RNase3-H15A-treated cells signalling pathway, Toll-like receptor signal pathway and Fig. 1  DEseq2 analysis Count of differential expressing genes (DEGs). a Genes with an adjusted P value < 0.01 & log2FC absolute value > 1 found by DESeq were assigned as differentially expressed (DE) genes. The common DEGs responded to RNase3 and RNase3-H15A treatment of THP1-derived macrophages at 4 h and 12 h was identified by overlapping. Venn plot of the DEGs from paired comparison: b common up-regulated and c common down-regulated DEGs 13 Human RNase3 immune modulation by catalytic-dependent and independent modes in a… 2969 Fig. 2  KEGG pathway enrich map of the common response DEGs to RNase3 and RNase3- H15A treatment of THP1- derived macrophages. KEGG pathways enriched by a com- mon up-regulated DEGs and b common down-regulated DEGs MAPK signalling pathway. To examine more closely the cor- transcription factors that mediate the inflammatory response: relations between the genes of the core response to RNase3 STAT1, responding to cell stress injuries [39]; NF-kB, which and RNase3-H15A treatment, we incorporated a network- controls pro-inflammatory cytokine production and cell sur- based approach using the NetworkAnalyst3.0 with Innat- vival [40], SRC, a regulatory protein kinase [41] and EGR1, eDB database [23]. Ranked by connectivity and betweenness activated during tissue remodelling. Therefore, we can con- centrality, EGR1, FN1, EGFR, SRC, JUN, STAT1, NFkB1 clude from the network analysis results that EGFR is the and CDKN1A were identified as the primary hub and bot- main hub of the core response induced by both wt RNase3 tleneck genes of macrophages induced upon wt RNase3 and and the RNase3-H15A mutant. RNase3-H15A exposure (Fig. 3, Additional file 5). Among Next, we explored the down-regulated DEGs in THP1- them, the Epithelial Growth Factor Receptor (EGFR) is one derived macrophages upon exposure of both wt and the cat- of the main up-regulated connector, together with EGR1, alytic-defective RNase3. We identified 1846 DEGs counts SRC, JUN, STAT1 and NFkB1. Moreover, EGFR shares associated to RNase3/RNase3-H15A common response direct or indirect interactions with all the other central hub (Fig. 1c), 13 pathways were significantly enriched (Fig. 2b; genes: EGR1, SRC, STAT1, NFkB1, and JUN and SRC Additional file 4). Down-regulated pathways include cell transcription regulator. EGFR activation positively regulates cycle, DNA replication, homologous recombination and 13 2970 L. Lu et al. Fig. 3  Protein–protein interaction (PPI) analysis of common response cated with lines, the circle size is proportional to interaction degree. DEGs. The common DEGs were studied using the NetworkAn- Only the genes with more than 25 interactions were shown. The col- alyst3.0 online tool. The interaction between each protein pair is indi- our bar indicated the log2FC of each gene mismatch repair among others, which are indicators of a cel- the EGFR pathway [42]. Moreover, PCNA is involved in lular growth inhibition and duplication arrest response. To the direct response to EGFR related to DNA damage [43]. note, CDK2, KIAA0101, BRCA1, CUL1, PCNA and CDK1 Following, we studied the macrophage cell response were considered the hub and bottleneck genes by network upon protein exposure as a function of time. Considering analysis (Fig. 3, Additional file 5). As illustrated by their that cytokine–cytokine receptor interaction is within the top key connectivity we can highlight CDK2 and KIAA0101, common enriched pathway upon wt RNase3 and RNase3- followed by BRCA1, CUL1 and PCNA, and last CDK1. Inter- H15A treatment, with more than 78 genes significantly up- estingly, BRCA1, CDK1 and CDK2 are also associated to regulated (Fig. 2a, Additional file 4), we decided to compare 13 Human RNase3 immune modulation by catalytic-dependent and independent modes in a… 2971 their dynamic changes at 4 h and 12 h time points. Com- indicating that the pro-inflammatory effect triggered by pared to control, RNase3 or RNase3-H15A triggered a fast RNase3 is mainly a short-term effect and is reduced upon inflammatory response of the macrophage cell, as witnessed time. Notwithstanding, a small group of genes associated to by a significant up-regulation of 20 hallmark inflammatory inflammation were upregulated after a 12 h exposure time, genes, such as TNFα, IL1β, CD40, CXCL8 and CCL4L2. such as CCL2, CXCL10 and CXCL11. Notably, together Moreover, 19 hallmark TNFα-signalling via NFkB genes with pro-inflammatory cytokines, we can identify up-regu- were also significantly upregulated, such as CXCL1, CXCL2, lated chemokines related to leukocyte recruitment, such as CXCL3, IL7R, LIF, CCL3L1, CCL4 and TNFRSF12A (see CCL1 and CCL2 for monocytes, CCL20 for lymphocytes, Additional file 6). Likewise, we identified a group of sig- CCL21 for activated T cells, CXCL3 for neutrophils and nificantly down-regulated genes, such as CXCR4, IL31RA, tissue remodelling, such as ICAM1, VCAM1, MMP9 and IL1RAP, PRLR, CCL28, IL12RB1, CXCR2, CXCR1, GDF11, TGFβ. Overall, the observed immune-metabolic response CCL23, CX3CR1, CCR2, BMP8B, and CSF3R. According is characteristic of a M1 macrophage activation type [44]. to the timeline, the gene counts of most pro-inflammatory To explore further the involvement of the EGFR receptor genes decreased by 12 h in comparison to a 4 h (Fig. 4), in the signalling pathway activated by RNase3, we assayed Fig. 4  Heat-map of DEGs enriched in cytokine–cytokine receptor Colour bar indicates the Log2FC for each gene. Red stars indicate interaction pathways. Pairwise comparison for control, RNase3 and correlation with the protein catalytic activity RNase3-H15A at 4 h and 12 h expression profiles was performed. 13 2972 L. Lu et al. the gene expression profile of selected top DEGs by qPCR of THP1-derived macrophages with Cetuximab, we con- at 4 and 24 h, using EGF as a positive control (Figure S2). firmed the inhibition of the protein effect (Figure S3). Next, Moreover, we evaluated the inhibitory action of a monoclo- we confirmed the induction of the downstream pathway nal antibody anti-EGFR (Cetuximab) on the gene expression associated to the activation of EGFR by detection of MAP levels. The antibody recognizes the extracellular domain of kinase (MAPK/ERK1/2) phosphorylation by western blot the receptor and blocks its activation. Following treatment analysis. We observed an increase of MAPK phosphoryla- tion as a function of time upon THP1-derived macrophages incubation with RNase 3. Besides, pre-treatment of cells with the anti-EGFR Ab, inhibited the kinase phosphoryla- tion (Fig. 5). RNase3 modulates the macrophage antiviral pathway in a catalytic‑dependent manner Although RNase3 and RNase3-H15A elicited an overall similar transcriptional response of macrophage, a direct comparison between macrophage cells treated with the wt Fig. 5  Western blot (WB) detection of pMAPK by RNase3 treatment and mutant proteins identified significant changes related to with or without anti-EGFR Ab (Cetuximab). 1 × 106 THP1 cells were plated in a 6-well culture plate and induced into macrophage by PMA the RNase catalytic activity. Transcriptome analysis of wt treatment. a the phosphorylation of MAPK was detected by com- versus H15A mutant revealed 30 and 143 DEGs at 4 and paring 10 μM of RNase3 treatment for 5 min, 15 min and 30 min; b 12 h, respectively (Fig. 1a). Within the DEGs paired set, we comparison of the phosphorylation of MAPK of the cells treated with found only six genes activated at both 4 and 12 h time points RNase3 (RN3) for 15 min at 10 μM with and without Cetuximab pre- treatment. Ratio of pMAPK to GAPDH signal normalized to control (Fig. 6a): CXCL10, IFIT1, LAMTOR1, SOCS3, OASL, and reference value is indicated at the bottom of each lane IFIT2. Interestingly, several of these genes are activated by Fig. 6  Analysis of RNase3 induced response of THP1- derived macrophages associated to catalytic activity. a Overlap of the DEGs identified by directly comparing the number of upregulated genes in RNase3 vs. RNase3-H15A treatment at 4 h and 12 h, respectively; b Protein–protein interaction (PPI) Network analysis of ribonuclease dependent protein interaction of DEGs identi- fied by comparing RNase3 and RNase3-H15A at 12 h. The interaction between each protein pair is indicated with lines, the circle size is proportional to interaction degree. The colour bar indicated the log2FC of each gene 13 Human RNase3 immune modulation by catalytic-dependent and independent modes in a… 2973 the interferon pathway, such as the 2′5′ oligoadenylate syn- the DEGs at 12 h, identified diverse biological processes but thase like protein (OASL), the CXCL10 chemokine and the in all the cases related to the cell defence response to virus interferon induced RNA binding proteins (IFIT1 and IFIT2) (Table 2, Additional file 7). In addition, the comparative [45–47]. Furthermore, SOCS3 activation is usually observed heatmap of DEGs highlighted that significant changes of wt following infection and plays a tissue protecting role against versus H15A mutant are mostly prominent at 12 h. Overall, inflammation side effects [48]. the results indicate that the late immune-regulatory action On the other hand, when all the DEGs were analysed (12 h) induced by RNase3 dependent on the protein cata- by connectivity and betweenness centrality, we observed 31 lytic activity would complement the first short-term response seed genes connected by 44 edges. The network analysis (4 h), mostly mediated by EGFR activation. Results suggest clearly pointed out to STAT1 and ISG15 as the hub genes that both RNase3-activated pathways, catalytic-dependent (Fig. 6b). ISG15 expression is regulated by STAT1 tran- and independent, might work by complementing each other. scription factor, which is characteristic of the macrophage activation by the interferon pathway and adaptation to an Characterization of an RNase3 overexpression THP1 antiviral state. The connectivity map outlined the contribu- cell line tion of other interferon induced regulatory proteins, such as IFIT1-3 and the CXCL10 chemokine. By applying the To corroborate our transcriptomic results and evaluate the KEGG pathway enrichment analysis, we mostly obtained at potential contribution in situ of the macrophage endog- 12 h signalling pathways related to infection sensing and in enous RNase3, we decided to obtain a THP1 cell line that particular associated to antiviral response (see Table 1 and overexpresses the protein. As quantified in our previous Additional file 6). In contrast, no significant pathway was work, basal levels of RNase3 in the monocytic THP1 cell found when applying the DEGs at 4 h. In any case, accord- line are detectable but considerably low (estimated about ing to GO term annotation analysis “response to virus” is the 1/500 respect to the GAPDH control gene) [14]. There- single significantly enriched at 4 h. Moreover, analysis of fore, the overexpression of RNase3 in THP1 cells could Table 1  KEGG pathway Description Gene ratio Bg ratio P adjust Count enrichment analysis of DEGs comparing RNase3-H15A Influenza A 16/82 170/7925 1.60E-09 16 treatment with RNase3 at 12 h Hepatitis C 13/82 155/7925 3.40E-07 13 Viral protein interaction with cytokine and 10/82 100/7925 2.88E-06 10 cytokine receptor NOD-like receptor signalling pathway 12/82 181/7925 9.50E-06 12 Cytokine-cytokine receptor interaction 14/82 294/7925 4.10E-05 14 Chemokine signalling pathway 11/82 189/7925 6.41E-05 11 Cytosolic DNA-sensing pathway 7/82 63/7925 6.41E-05 7 Toll-like receptor signalling pathway 8/82 104/7925 0.000177 8 Measles 8/82 138/7925 0.001219 8 RIG-I-like receptor signalling pathway 5/82 70/7925 0.009897 5 Significant enriched pathway list was filtered using P adjust < 0.01 Table 2  Top ten significant Description Gene ratio Bg ratio P adjust Count GO terms enriched comparing RNase3-H15A treatment with Defence response to virus 39/130 224/17,653 7.32E-40 39 RNase3 at 12 h Response to virus 42/130 310/17,653 8.42E-39 42 Negative regulation of viral process 21/130 95/17,653 9.30E-23 21 Negative regulation of viral life cycle 19/130 79/17,653 2.78E-21 19 Negative regulation of viral genome replication 16/130 55/17,653 3.21E-19 16 Regulation of multi-organism process 28/130 391/17,653 2.00E-17 28 Negative regulation of multi-organism process 21/130 172/17,653 2.00E-17 21 Regulation of viral life cycle 19/130 137/17,653 9.75E-17 19 Response to interferon-gamma 21/130 191/17,653 1.43E-16 21 Regulation of viral process 21/130 196/17,653 2.21E-16 21 13 2974 L. Lu et al. serve to analyse the protein action within the macrophage- confirmed that overexpression of RNase3 did not alter the derived cells. CRIPSRa methodology was applied to over- macrophage cell viability, as evaluated by the MTT assay. express RNase3 in THP1 cells. The plasmids encoding However, the overexpression of RNase3 is significantly CRISPRa components were delivered to THP1 cells by slowing down the duplication rate. We estimate a reduc- lentiviral infection. An empty lentiviral plasmid was used tion of about 20% between 1st and 2nd day of incubation for the control reference cells. The GFP and Cherry red (results not shown), that might be attributed to the pro- both positive cells successfully transduced by lenti239G tein downregulation of genes associated to cell cycle (see (encoding dCas9-VP64-GFP) and lenti239R (encod- Figs. 2 and 3). ing sgRNA-Cherry red) were sorted by FACS (Fig. 7a). To validate the transcriptome results obtained by After induction of THP1 to macrophage, the expression NGS sequencing, we analysed the expression profile of of RNase3 was quantified by qPCR, observing an increase the RNase3 overexpression THP1 cell line. A total of 18 from about 1.5 to threefold in all the analysed cell lines genes were selected among the top DEGs from libraries (Figure S4). The best performance cell line (OX5) was of wild-type versus control and catalytically defective selected for final characterization. RNase3 levels in rela- mutant versus wild-type for further qPCR analysis. The tion to wild-type THP1 cells were quantified by both qPCR results confirmed similar expression levels induced by and WB, confirming a significant increase of RNase3 both endogenous and recombinant proteins (Fig. 7d). both at the gene and protein level (Fig. 7b, c). Besides, we Moreover, we compared the gene expression profile of the RNase3 overexpression THP1 cell line (OX) in the Fig. 7  Overexpression of RNase3 in human THP1 derived macrophages. a FACS selection of both GFP (indicating the successful transduction of len- ti239G-dCas9-VP64) and RFP (indicating the successful trans- duction of lenti239R-sgRNA) positive cells; b Comparison of the transcriptional expression of RNase3 gene in wild-type THP1 macrophages (WT) and RNase3 overexpression THP1 macrophages (OX) by qPCR; c Comparison of the expres- sion of RNase3 protein in WT and OX THP1 macrophages by WB. d Comparison of 18 gene expression levels between tran- scriptome analysis by RNAseq of THP1 macrophages treated with recombinant RNase3 and RNase3 OX THP1 macrophage cells analysed by qPCR. 18 DEGs identified from transcrip- tome sequencing were validated by qPCR, using WT and OX THP1-induced macrophages 13 Human RNase3 immune modulation by catalytic-dependent and independent modes in a… 2975 absence and presence of the anti-EGFR antibody Cetuxi- particular, Hung and co-workers confirmed that residues mab and confirmed the Ab inhibition of genes associated Q93 and Y94 were essential for the receptor binding to to the EGFR pathway (Figure S5). RNase5. Here, in our model for RNase3, we observed the putative binding of R97 to the receptor extracellular Analysis of the putative RNase3‑EGFR interaction domain. Interestingly, RNase3 presents a single nucleo- by molecular modelling tide polymorphism (SNP) at position 97 with two alter- nate substitutions (R/T) [51] and molecular modelling Considering that our whole transcriptome analysis pointed predicted slight differences in the affinity to the receptor out to a direct activation of EGFR by RNase3, we decided of R97 and T97 variants (Table 3). to explore further the protein-receptor-binding process. Interestingly, it was recently demonstrated that RNase5/ Overexpression of endogenous RNase3 inhibits Ang, another member from the RNase A superfamily, is the macrophage intracellular infection by both M. a direct ligand of EGFR [18]. The authors identified a aurum and RSV C-terminal region in RNase5 needed for EGFR interac- tion, with significant homology to EGF binding region Next, we evaluated the efficacy of the macrophage endog- to the receptor. Notably, we observed how the key resi- enous RNase3 against infection. We selected two models of dues shared by RNase5 and EGF (C81, C92, Q93, Y94) macrophage intracellular infection (M. aurum and RSV). are mostly conserved in RNase3 (Fig. 8a and Additional Both intracellular pathogens have been previously proven to file 8). In particular, the QYRD sequence was spotted to be effectively eradicated within the macrophages by recom- participate in EGF interaction with the receptor by site- binant RNase3 addition [9, 14]. directed mutagenesis [49] and molecular dynamics [50]. First, we infected both wild-type (WT) and RNase3-over- Considering that the region of RNase5 involved in the expression (OX) THP1 cells derived to macrophages with M. interaction with EGFR is mostly conserved in RNase3, aurum and monitored the cell lines for 3 days. Intracellular we decided to explore the potential protein interaction infection was quantified by Colony Forming Unit counting to the receptor. We applied HADDOCK software to pre- (CFUs) in lysed cells after removal of extracellular myco- dict the complex structures of RNase3 with the EGFR bacteria. No significant differences in M. aurum initial CFUs extracellular domain (Fig. 8b). The crystal structure of between WT and OX cell lines were observed at the initial the EGFR-EGF complex was used as a model PDB (ID: time point (Fig. 9a). However, M. aurum can easily prolifer- 3NJP). A high binding affinity of RNase3 to EGFR was ate in wild-type macrophage as indicated by an exponential estimated (see Table 3). Notably, we obtained for RNase3 CFU count increase as a function of time. On the contrary, similar or even slightly higher binding energies, as calcu- mycobacteria cannot proliferate in the RNase3-overexpres- lated for EGF. A close inspection of the predicted com- sion cell line. A significant growth inhibition of M. aurum plex revealed that the interaction of RNase3 with EGFR in RNase3 overexpression macrophages was detected at 24 h might be mediated by residues regions 1–4, 19–28, post-infection compared to WT macrophage. Noteworthy, 33–35, 86–87 and 94–100. In particular, we can iden- the growth inhibition was even more prominent as the infec- tify the key role of N95, R97, Y98 and D100 of RNase3, tion progressed (48 h and 72 h). The present data corrobo- which can interact within the 417–470 region of the EGF rate our previous results showing that recombinant RNase3 receptor (see Additional file 8 for a full list of predicted can efficiently mediate the eradication of macrophage intra- protein–protein interactions). The molecular modelling cellular infection by mycobacteria [14]. also indicates that the RNase active site residues (H15, Likewise, both WT and OX THP1 cells derived to mac- K38 and H128) would not be involved in the receptor rophages were infected with the RSV virus and the cell binding, in agreement with our transcriptomic results. lines were monitored for 3 days. Probe-qPCR was applied Besides, the main region identified in RNase3 overlaps to quantify the RSVs both extracellularly and intracellularly with the RNase5 and EGF counterparts (Fig. 8a). How- using an internal RSV quantification standard curve built ever, we also observed significant differences in the main by serially dilutions from a RSV stock (Figure S6). Starting interacting region from EGF and RNases, where residues from an equivalent RSV initial titer in both WT and OX cell 45–51 in EGF are not conserved in RNases sequence. lines, we observed the increase as a function of time of RSV This suggests that the RNases mechanism is not fully at intracellular and extracellular levels (Fig. 9b and Fig. S7). equivalent to EGF functioning and might indicate a dis- Intracellular and extracellular RSV reached a maximum peak tinct target within the EGFR family. Notwithstanding, we at 24 h and 48 h post of infection, respectively. Importantly, do observed conservation of main key interacting resi- we found that overexpression of RNase3 in macrophage dues, specifically within the CQYRD segment (residues can significantly inhibit the RSV duplication at 24–72 h poi 96–100 and 92–96 in RNase3 and EGF, respectively). In (Fig. 9b and Fig. S7). In addition, overexpression of RNase3 13 2976 L. Lu et al. Fig. 8  Comparison of the interaction of EGFR with RNase3 (R97 and and RNase3-T97. The crystal complex of EGFR with EGF (PDB ID: T97), RNase5 and EGF by molecular modelling. a Primary sequence 3NJP) is depicted. EGFR is coloured in grey, the protein ligands were alignment of EGF, RNase3 and RNase5. The common interacting coloured in pink, the interacting residues from EGFR and the ligands region is underlined. b Representative binding of EGFR with EGF, were coloured in blue and red, respectively RNase5, and RNase3 single nucleotide polymorphisms: RNase3-R97 reduced the macrophage cell death rate caused by M. aurum Overall, we confirmed that the OX cell line expresses sig- and RSV infection (Fig. 9c). nificantly higher level of RNase3 than WT cells at all the Moreover, the transcriptional expression levels of time points (Fig. 9d). In addition, we explored whether RNase3 were quantified in both cell lines (WT and OX) in the cell infection by M. aurum and RSV could regulate the the absence and presence of M. aurum and RSV infection. macrophage endogenous RNase3 expression. The results 13 Human RNase3 immune modulation by catalytic-dependent and independent modes in a… 2977 Table 3  Predicted free energies for EGFR-protein modelled com- infection. Furthermore, the OX cell line kept the expression plexes levels of RNase3 stable during all the 3 days’ experiment in ΔG (kcal/mol) Kd (M) at 25 °C both M. aurum and RSV infection studies. EGF (docking) − 10.3 2.9·10–8 Blockage of the EGFR receptor by Erlotinib inhibits EGF (crystal) − 15.8 2.5·10–12 the RNase3 antibacterial activity. RNase5 − 12.4 8.2·10–10 RNase3 (R97) − 12.0 1.5·10–9 We next assessed the effect of adding an inhibitor of EGFR RNase3 (T97) − 10.2 3.4·10–8 activity on the observed anti-infective action of RNase3. The binding affinities ΔG (kcal/mol) and dissociation constants Kd Erlotinib inhibits EGFR activation by blocking the recep- (M) were predicted using HADDOCK2.2 tor tyrosine kinase [52]. The RNase3 overexpression mac- rophages were treated with Erlotinib for 24 h before infection of either M. aurum or RSV. As illustrated in Fig. 10a, upon indicate that short-term M. aurum infection is downregulat- addition of 10 μM of Erlotinib, we observed a significant ing RNase3 whether a long-term exposure (48–72 h) induced increase of M. aurum growth within the macrophages. On the protein expression. The data corroborate down- and up- the contrary, this scenario was not observed in RSV infec- regulation profile previously reported upon mycobacteria tion even after addition of 100 μM of Erlotinib (Fig. 10b). infection [14]. On the contrary, the expression of RNase3 Moreover, we quantified and compared the fold change of a in WT macrophage was not significantly altered by RSV set of 18 selected genes in RNase3 overexpressed cell line Fig. 9  RNase3 inhibits bacterial and viral infection within mac- to measure the cell viability, 0 h WT group was used for normaliza- rophages. Overexpression of RNase3 in macrophage inhibits M. tion (100%); d Comparison of the relative transcriptional expression aurum and RSV proliferation. a the number of M. aurum inside of of RNase3 gene by qPCR using GAPDH as a reference. Cell viabil- macrophage was counted by CFU assay and compared with wild- ity and RNase3 expression was measured in WT, OX, WT infected type THP1-induced macrophages (WT) and RNase3 overexpres- with M. aurum (WT + M. aurum) or RSV (WT + RSV), and OX sion THP1-induced macrophages (OX) for up to 3 days; b RSV was cells infected with M. aurum (OX + M. aurum) or RSV (OX + RSV) quantified by probe-qPCR, the intracellular RSV was normalized THP1-macrophage-derived cells; significance is indicated as using GAPDH gene in WT and OX cells; c MTT assay was applied *P < 0.05 and **P < 0.01 13 2978 L. Lu et al. in the presence and absence of Erlotinib treatment. The selected genes had previously been quantified as DEGs by both recombinant RNase3 addition and endogenous protein overexpression (see Fig. 7d). Besides, by the side-by-side comparison of the overall transcriptome between wild-type and catalytic defective RNase3, we identified the DEGs dependent or independent of the protein catalytic activity (see Additional file 1). Here, following Erlotinib treatment we observed how gene expression is uniquely altered for the genes not dependent on the protein catalytic activity (Fig. 10c). That is, addition of 10 μM of Erlotinib signifi- cantly inhibited RNase3′s induction of the expression of genes such as CXCL10, CEBPB, TNF, BCL3, IL1, NFKIBA, SRC, TLR4, ATP6V1H, GABARAP. On the contrary, MDA5/ IFIH1, ISG15 and RIG1 that are dependent on RNase3′s catalytic activity, did not response to Erlotinib treatment. The results corroborated the presence of two distinct type of pathways that are regulated by RNase3 and are associated to either catalytic or non-catalytic-dependent mechanisms of action (see Fig. 11 for a schematic overview). Discussion Antimicrobial peptides (AMPs) are important components of natural defence against a wide range of pathogens [53]. AMPs, originally reported to work by a direct action at the microbial cell wall, were later ascribed a diversity of immune modulatory properties that can contribute to the infection eradication [54, 55]. Human antimicrobial RNases from the RNaseA superfamily can be regarded as multifac- eted AMPs that combine a microbial membrane destabiliza- tion action with immune regulatory properties [4, 56, 57]. In our recent work, we have reported how RNase3 induction of macrophage autophagy mediated the eradication of M. aurum intracellular infection [14]. Here, we have applied the NGS RNA sequencing methodology to explore the immune-regulatory mechanism of action of RNase3 within macrophages. The comparative transcriptomic profile of macrophages treated with wild-type and catalytic-defective RNase3 has enabled us to identify the regulation pathways Fig. 10  EGFR inhibitor, Erlotinib, blocks the induction of RNase3 antimycobacterial activity. Before infection, the EGFR inhibitor, related and unrelated to the protein ribonucleolytic activ- Erlotinib, was used to treat RNase3 overexpression macrophages for ity. Overall, we observed that treatment with both RNase3 24 h at 0 μM, 10 μM and 100 μM. At the indicated post-infection and RNase3-H15A triggered a common immune response. time, the intracellular M. aurum (a) and RSV (b) was quantified by The shared transcriptome profile pattern outlined an up-reg- CFU assay and probe-qPCR assay, respectively; the median tissue culture infectious dose (TCID50) was calculated and the intracellu- ulation core response characteristic of a macrophage pro- lar RSV was normalized using GAPDH gene. c The relative expres- inflammation condition [44]. Furthermore, downregulation sion of genes was quantified by qPCR with or without Erlotinib core response indicated that the protein addition promoted treatment (10 μM, 24 h), and the fold change was calculated by com- the cell growth arrest and duplication inhibition (Fig. 2). paring Erlotinib treatment with control. Significance is indicated as *P < 0.05 and **P < 0.01 Protein–protein network analysis identified EGFR as the main hub gene together with a group of five other genes (JUN, NFkB1, STAT1, EGR1, SRC), which are on their turn directly or indirectly interacting with EGFR (Fig. 3). 13 Human RNase3 immune modulation by catalytic-dependent and independent modes in a… 2979 Fig. 11  Schematic illustra- tion of the proposed molecular mechanism of RNase3 modula- tion in human macrophage. The genes associated to RNase catalytic activity were labelled in red. EGFR epidermal growth factor receptor, IFIT interferon induced protein with tetratricopeptide repeats, EGR1 early growth response 1, JUN jun proto-oncogene or AP-1 transcription factor subunit, NFkB nuclear factor Kappa B, transcription regulator, STAT​ signal transducer and activator of transcription, IRF inter- feron regulatory factor, VCAM vascular cell adhesion molecule, MMP matrix metallopeptidase, TGF transforming growth fac- tor, TNF tumour necrosis factor, TRAF TNF receptor-associated factor, IL interleukin, CCL C–C motif chemokine ligand, CXCL C-X-C motif chemokine ligand, ISG15 interferon induced 15 kDa protein, OAS 2′-5′-oli- goadenylate synthetase, IFN interferon, PCNA proliferat- ing cell nuclear antigen, CDK cyclin-dependent kinase, BRCA1 Breast Cancer gene 1, DNA repair associated gene Furthermore, inspection of central hub down-regulated sharing with RNase3 a common three-dimensional structure genes also revealed an association to EGFR receptor activa- fold, was taken as a reference to model a putative RNase3- tion. Potential involvement of EGFR activation in RNase3 receptor complex using the EGF-receptor structure as a com- signalling pathway was corroborated by the blockage of mon reference for both homologous RNases (Fig. 8). The the expression level of selected DEGs using Cetuximab, a results predicted a strong binding affinity of RNase3 to the monoclonal anti-EGFR Ab (Figure S3). Besides, we also EGFR extracellular domain. Molecular modelling suggested confirmed the activation of MAPK phosphorylation upon no involvement of the protein catalytic site in the recep- cell treatment with RNase3 and its inhibition in the presence tor interaction, as indicated by our transcriptome results. of Cetuximab (Fig. 5). Interestingly, EGFR, the epidermal Likewise, EGFR activation by RNase5/Ang was not depend- growth factor receptor, is not only a key membrane receptor ent on the protein catalytic activity [18]. Interestingly, we involved in cell survival and tissue remodelling [58, 59], but found a high sequence homology between both RNases at it can also mediate the macrophage activation during bacte- the conserved key target sequence previously identified at rial [60] and virus infection [61]. Specifically, EGFR signal- RNase5 and EGF C-terminus [18] (Fig. 8a). In particular, ling is critical for pro-inflammatory cytokine and chemokine structural alignment of RNase5 with the EGF molecule in production [60, 62, 63]. Therefore, our present results sug- complex with the extracellular domain of the receptor [65] gest that RNase3 mostly modulates the macrophage response highlighted four conserved residues: two cysteines (C81 via direct EGFR activation. Taking into consideration the and C92), a glutamine (Q93) and a tyrosine (Y94) [64]. very recent report of human RNase5 direct binding to EGFR C81 and C92 are conserved in most of the RNaseA fam- [64], we envisaged here a structural analysis of the putative ily members and participate in the disulphide bonding that interaction of RNase3 to the receptor. Human RNase5/Ang, connects the protein N and C ends and stabilizes the overall 13 2980 L. Lu et al. three-dimensional structure. On their turn, Q93 and Y94 being arginine the predominant substitution [75, 76]. The counterparts in EGF were reported to bind to the receptor importance of this residue was highlighted by Hung and co- [66] and were confirmed by site-directed mutagenesis to par- workers, who demonstrated that human RNase1 homologue, ticipate in RNase5 -EGFR interaction. Pairwise sequence where position 97 is occupied by an Ala, does not activate alignment indicated that Y94 is conserved in all RNase the EGFR [18]. Additional variability is observed within the members. On the other hand, we found partially conserved last two residues of the CQYRD region. Arginine at posi- substitutions for RNase5 Q93. In particular, in RNase3 the tion 99, present at EGF and RNase5 is substituted by an Ala glutamine is substituted by a threonine (T97). Interest- in some RNases, like RNase2 and RNase3 [75]. Notably, ingly, RNase3 displays a polymorphism at this position (R/ A99 in human RNase3 is conserved in all primates [75]. T97), which has been correlated to the protein properties. In addition, our modelling study identified another residue In particular, R97 shows an enhanced cytotoxicity on sev- key for RNase3 interaction to the receptor, D100. An aspar- eral tested eukaryote cell lines [67]. On the contrary, the tic residue at position 100 is also key for EGF binding but presence of a T at this position creates an N-glycosylation is not present in RNase5. The residue is conserved in all site and native N-glycosylated forms display a reduced cyto- RNase3 primates but absent in RNase2, its closest homo- toxicity and antimicrobial activity [17, 68, 69]. Particular logue counterpart. Therefore, structural analysis revealed attention to R/T97 SNP in RNase3 was drawn due to the subtle differences that might significantly alter the recep- specific distribution within the population [70]. Presence of tor recognition pattern. We can speculate that the observed an R at 97 position has been linked to an enhanced severity variability within the interacting sequence is indicating the of malaria and schistosomiasis side-effects [71, 72], such targeting of distinct receptor subtypes within the EGFR as hepatic fibrosis or neurological disorders, that might be family. In any case, our results are in agreement with the associated to an overabundance of the secreted RNase3 at characterization of RNase5-EGFR interaction, that discarded the infectious focus [73, 74]. The protein overproduction at the contribution of the protein catalytic activity in the recep- the infected tissue could be detrimental to the host health, as tor activation [64]. In addition, the present results highlight also observed in allergic asthma and other chronic inflam- the induction by both wild-type RNase3 and RNase3-H15A matory diseases. Evolutionary studies indicated that R is the of pro-inflammatory cytokines. Complementarily, we also amino acid present in RNase3 lineage ancestor and is shared identified the expression of tissue remodelling proteins and by most of the other primate homologues [75]. chemokines, such as FN1, VCAM1, MMPs and TGFβ, in The current molecular modelling results of the protein- a catalytic-independent manner. The present data would EGFR complex suggest the involvement of residue 97 in reinforce the RNase3 role in tissue remodelling processes EGFR activation. To explore the potential contribution of reported by in vitro wound healing models [2, 16, 77–79]. the R/T97 SNP in the protein action, we compared here the Although transcriptome analysis indicates that the major- recombinant non-glycosylated protein (coding for the wild- ity of the macrophage response is common to both RNase3/ type R97) with the two glycosylated variants (R/T97). Over- RNase-H15A treatment, we also identified a group of genes all, we observe a similar expression profile for all recombi- specifically associated to the protein ribonucleolytic activ- nant proteins. However, we identified a significant increase ity (see Fig. 6, Additional file 1). Overall, the main bulk of in the expression levels of the pro-inflammatory IL1A and identified genes are related to interferon induced pathways, NFKB1A markers as a function of the protein glycosylation such as CXCL10, a leukocyte chemoattractant during viral degree, from the non-glycosylated form expressed in E. coli infection, IFIT1 and 2, two antiviral RNA-binding proteins (RN3) to the T97 variant that incorporates an additional gly- [46] and OASL [80, 81]; a 2′-5′ oligoadenylate synthetase cosylation site. In addition, we also confirmed that pre-treat- like gene which can bind double-stranded RNA and displays ment of macrophage cells with the anti-EGFR Ab inhibits antiviral activity through mediating RIG-I activation [82, the induced gene expression profile for all the tested recom- 83]. Accordingly, the top-rated KEGG pathways are related binant proteins (Figure S8). Therefore, the qPCR data also to host response to virus infection (Table 2). Among them, suggest that the protein region around residue 97 is involved the RIG-I-like receptors (RLRs) play a major role in patho- in the activation of EGFR related pathways. However, fur- gen sensing of RNA virus infection to initiate and modu- ther studies will be required for a deep understanding of the late the antiviral immunity [84]. Importantly, not only viral structural underling basis of the distinct biological proper- RNA ligands but also processed self RNA can be detected ties of the natural protein variants. On the other hand, the by RLRs in the cytoplasm to trigger innate immunity and analysis of the CQYRD region within the overall vertebrate inflammation and induce gene expression against infec- family context reveals a significant variability at position 97. tion [84, 85]. Our network analysis identified ISG15 as a A glutamine is present in RNase5/Ang and ancestral family hub gene dependent on RNase3 catalytic activity (Fig. 6b). members, shared by most fish but not amphibian RNases, ISG15 is regarded as the hallmark gene of RIG-I like recep- whereas a high variability is found in mammalian RNases, tor-signalling pathway, which is mainly activated by viral 13 Human RNase3 immune modulation by catalytic-dependent and independent modes in a… 2981 RNA and DNA. In addition, a specific processing of the host Interestingly, a very recently work describes the participa- non-coding RNA and release of dsRNA can also activate tion of the human RNase2, a close relative of RNase3 shar- the cell intrinsic immune response through RIG-I, MDA5 ing a 70% sequence identity, in the detection of pathogen or OASL sensors [85]. In addition, we observe activation RNA mediated by TLR8. The authors suggest that RNase2 of STAT1, which might be mediated by the release of spe- cleavage products in synergy with RNaseT2 would act as cific RNA products, such as dsRNA, by RNase3. Interest- direct ligands of TLR8 [99]. ingly, STAT1 is activated by both the IFN and EGFR sig- To complement our comparative transcriptome analy- nalling pathways. Indeed, the initial increase of IFNα and sis, we decided to analyse in situ the potential role of the IFNβ could be induced by the NF-kB pathway, which was RNase3 endogenously expressed by macrophages. Charac- observed to be activated by an RNase independent manner. terization of a THP1-macrophage-derived cell line that over- Following, RNase3 activity might complementary activate expresses RNase3 corroborated the transcriptomic results. STAT1 and induce the further release of chemokines. Based The observed gene expression pattern induced by RNase3 is on our data, we can hypothesize that the IFN pathway might characteristic of both EGFR- and IFN-associated pathways, be reinforcing the short-term signalling paths induced by which can participate in the macrophage response to bacte- a direct targeting of EGFR by RNase3 (see Fig. 11 for an rial and viral infection, respectively [60, 98, 100]. Moreover, illustrative scheme). our results revealed that native RNase3 expressed within Complementarily, the RNA products can work as inter- the macrophages can mediate the eradication of M. aurum cellular signalling molecules to mediate a prompt host or RSV infection. Our previous work using recombinant response to infection [86]. A complex interplay between RNases highlighted the protein antimicrobial activity against the host ncRNA and the RNA virus might determine the both extracellular and macrophage intracellular dwell- final outcome during infection [87]. We can speculate that ing mycobacteria [6, 14]. The present results indicate that RNase3 antiviral immune response may be mediated by the overexpression of endogenous RNase3 within macrophages generation of specific RNA cleavage products. Our tran- can inhibit both M. aurum and RSV intracellular prolifera- scriptome results also highlighted that most of the significant tion (Fig. 9). In addition, we observed how the mycobac- DEGs related to RNase3 catalytic activity were visualized teria infection is modulating the expression of RNase3, as at the late (12 h) exposure time rather than at the early (4 h) reported for other host defence peptides [101–103]. Last, time point (Figs. 4 and 6). Moreover, among the activated with the final aim to confirm RNase3 activation of the EGFR pathways associated to the protein ribonucleolytic we found pathway, we treated the THP1 cell lines with Erlotinib, an the activation of endosomal receptors, such as the TLRs 7–9, inhibitor of the EGFR receptor. The receptor blockage only which are associated to endosomal detection of both foreign altered the expression profile of genes unrelated to RNase3 and host signalling nucleic acid molecules [86]. On its turn, catalytic activity (Fig. 10). More importantly, the experi- TLR8 activated by self RNAs might reinforce the NFkB mental results indicated that the receptor is required for pathway (see Fig. 11). In addition, the TLR9 signalling path- RNase3 antibacterial but not for antiviral activity. The pre- way is reinforced by a direct interaction and phosphorylation sent data are in agreement with previous reports on RNase3 by the EGFR [88]. catalytic activity direct contribution on the protein antiviral We find in the literature reference to other members of [9] but not antibacterial action [14, 37, 38]. the RNaseA superfamily related to interferon-mediated RNase3 has been reported to contribute against per- response and endosomal TLR activation [89–91]. RNase7 sistent intracellular pathogens, such as the tuberculosis was recently reported to activate the TLR9 signalling path- bacilli or the HIV virus, that frequently coexist and threaten way [92, 93]. Earlier, Schein and co-workers related BS- immune-depressed patients [98, 101]. Therefore, it is cru- RNase activity to a direct interaction with IFN and the cleav- cial to understand the protein mechanism of action against age of dsRNA [94, 95]. Besides, IFN is reported to activate macrophage intracellular infections. Unfortunately, RNase3 the expression of RNase L, another RNase that belongs to a pro-inflammatory action following infection might also have protein family totally unrelated to the RNaseA superfamily. a detrimental effect on the host tissues. Nevertheless, the RNase L is also induced by IFN and can release small RNA protein action at the infection focus might turn out benefi- products that reinforce the IFN-mediated response [96]. cial by promoting the tissue remodelling and healing [4, 16, Both RNase L and RNase3 induce the OAS gene, which 78]. Besides, RNase3 induction of leukocyte recruitment is related to dsRNA and antiviral response pathways. The should reinforce the role of other blood cell type. In par- present results suggest that both protein families might coop- ticular, RNase3 is abundantly secreted by eosinophils during erate to fight infection by shared convergent mechanisms of inflammation and infection. Upon eosinophil degranulation, RNA processing. Indeed, recent work highlighted the impor- pro-inflammatory cytokines and chemokines are released tance of coordinated signalling pathways that mediate the to the infected tissue. On its turn, the eosinophil secretory macrophage defence mechanism against infection [97, 98]. proteins can be engulfed by macrophages and participate 13 2982 L. Lu et al. in the eradication of intracellular infection. Therefore, the PID2019-106123GB-I00 to EB co-financed by FEDER funds and protein would mediate a positive feedback and will ensure an Fundació La Marató de TV3 (ref. 20180310). efficient host response. Further work is in progress to fully comprehend RNase3 signalling role. A better understanding Compliance with ethical standards of the regulatory pathways that mediate the host response Conflict of interest The authors declare no conflicts of interest. processes induced by RNase3 should facilitate the design of alternative anti-infective drugs. The present work underlines Open Access This article is licensed under a Creative Commons Attri- once again the therapeutic potentiality of our own defence bution 4.0 International License, which permits use, sharing, adapta- molecules. tion, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are Conclusions included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in Comparative transcriptome profile analysis of macrophages the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will treated with wild-type RNase3 and the catalytic-defective need to obtain permission directly from the copyright holder. To view a mutant (RNase3-H15A) revealed that the protein triggers copy of this licence, visit http://creat​iveco​mmons​.org/licen​ses/by/4.0/. an early pro-inflammatory response in a ribonuclease-inde- pendent manner. Moreover, protein–protein network analysis of comparative gene expression profiles indicated that the References overall cell response is triggered by a direct activation of the EGFR. Interestingly, addition of an anti-EGFR antibody 1. Acharya KR, Ackerman SJ (2014) Eosinophil granule proteins: inhibits the induced expression of EGFR-associated genes form and function. J Biol Chem 289:17406–17415. https​://doi. org/10.1074/jbc.R113.54621​8 and the phosphorylation of MAPK upon RNase3 treatment. 2. Venge P, Byström J, Carlson M et al (1999) Eosinophil cati- By structural analysis, we have identified the protein region onic protein (ECP): molecular and biological properties and potentially involved in the receptor binding. Complemen- the use of ECP as a marker of eosinophil activation in dis- tarily, comparative transcriptome analysis suggested that ease. Clin Exp Allergy 29:1172–1186. https​://doi.org/10.104 6/j.1365-2222.1999.00542​.x RNase3 catalytic activity would participate in the activation 3. Bystrom J, Amin K, Bishop-Bailey D (2011) Analysing of specific pathways associated to antiviral host defence. In the eosinophil cationic protein—a clue to the function of addition, the specific blockage of EGFR by Erlotinib indi- the eosinophil granulocyte. Respir Res 12:10. https​: //doi. cates that the receptor-associated pathways participate in org/10.1186/1465-9921-12-10 4. Lu L, Li J, Moussaoui M, Boix E (2018) Immune modulation by the protein antibacterial but not antiviral actions. Last, we human secreted RNases at the extracellular space. Front Immunol demonstrated that endogenous overexpression of RNase3 in 9:1–20. https​://doi.org/10.3389/FIMMU​.2018.01012​ macrophages can inhibit M. aurum and RSV intracellular 5. Chihara J, Urayama O, Tsuda A et al (1996) Eosinophil cationic proliferation, which advances novel strategies in the design protein induces insulin-like growth factor I receptor expression on bronchial epithelial cells. Int Arch Allergy Immunol 111:43– of alternative anti-infective drugs. 45. https​://doi.org/10.1159/00023​7414 6. Pulido D, Torrent M, Andreu D et al (2013) Two human host Acknowledgements The authors wish to thank Francisco Cortes defense ribonucleases against mycobacteria, the eosinophil cati- (SCAC, UAB), Laura Tusell and Maria Rodriguez Muñoz (Dpt. of onic protein (RNase 3) and RNase 7. Antimicrob Agents Chem- Cell Biology, Physiology and Immunology, UAB) for helping with len- other 57:3797–3805. https​://doi.org/10.1128/AAC.00428​-13 tiviral production. We also thank Manuel Kaulich (Goethe University 7. Torrent M, Badia M, Moussaoui M et al (2010) Compari- Medical School) and Marcos Gil Garcia (IBB, UAB) for gifting gene son of human RNase 3 and RNase 7 bactericidal action at the and plasmids and Yundong Peng (Max Planck Institute for Heart and Gram-negative and Gram-positive bacterial cell wall. FEBS J Lung Research) for offering help with CRISPR gene editing experi- 277:1713–1725. https:​ //doi.org/10.1111/j.1742-4658.2010.07595​ ment. LL and GPE are recipients of a CSC and PIF-UAB predoctoral .x fellowships, respectively. MT was awarded a Ramon y Cajal research 8. Salazar VA, Arranz-Trullén J, Navarro S et al (2016) Explor- contract (MINECO). ing the mechanisms of action of human secretory RNase 3 and RNase 7 against Candida albicans. Microbiologyopen 5:830– Author contribution LL and EB designed the experimental work. LL, 845. https​://doi.org/10.1002/mbo3.373 MG and GPE performed the laboratory experimental work; LL, RLW, 9. Domachowske JB, Dyer KD, Adams AG et al (1998) Eosinophil GW and MT analysed the transcriptomic data; LL and EB prepared the cationic protein/RNase 3 is another RNase A-family ribonuclease first draft and edited the manuscript. LL, MT and EB revised the final with direct antiviral activity. Nucleic Acids Res 26:3358–3363. paper version. All authors approved the final version of the manuscript. https​://doi.org/10.1093/nar/26.14.3358 10. Boix E, Salazar VA, Torrent M et al (2012) Structural determi- Funding Experimental work was supported by the Ministerio de nants of the eosinophil cationic protein antimicrobial activity. Economía y Competitividad, (MINECO): SAF2015-66007P and Biol Chem 393:801–815. https:​ //doi.org/10.1515/hsz-2012-0160 11. Torrent M, Cuyás E, Carreras E et al (2007) Topogra- phy studies on the membrane interaction mechanism of the 13 Human RNase3 immune modulation by catalytic-dependent and independent modes in a… 2983 eosinophil cationic protein. Biochemistry 46:720–733. https​:// 28. Kutner RH, Zhang X-Y, Reiser J (2009) Production, concen- doi.org/10.1021/bi061​190e tration and titration of pseudotyped HIV-1-based lentiviral 12. Carreras E, Boix E, Rosenberg HF et al (2003) Both aromatic vectors. Nat Protoc 4:495–505. https​://doi.org/10.1038/nprot​ and cationic residues contribute to the membrane-lytic and bac- .2009.22 tericidal activity of eosinophil cationic protein. Biochemistry 29. Livak KJ, Schmittgen TD (2001) Analysis of relative gene 42:6636–6644. https​://doi.org/10.1021/bi027​3011 expression data using real-time quantitative PCR and the 2− 13. Abengozar MA, Torrent M, Fernandez-Reyes M, et al (2012) ΔΔCT method. Methods 25:402–408. https​://doi.org/10.1006/ Deconvolution of eosinophil cationic protein (ECP) for cell- METH.2001.1262 penetrating and microbicidal activities on Leishmania. FEBS J, 30. Zhang L, Jie H, Xiao Y et al (2019) Genomic identification and p. 279 expression analysis of the cathelicidin gene family of the forest 14. Lu L, Arranz-Trullén J, Prats-Ejarque G et al (2019) Human musk deer. Animals (Basels). https:​ //doi.org/10.3390/ani908​ 0481​ antimicrobial RNases inhibit intracellular bacterial growth and 31. Vissers M, Habets MN, Ahout IML et al (2013) An in vitro induce autophagy in mycobacteria-infected macrophages. Front model to study immune responses of human peripheral blood Immunol. https​://doi.org/10.3389/fimmu​.2019.01500​ mononuclear cells to human respiratory syncytial virus infection. 15. Zheutlin LM, Ackerman SJ, Gleich GJ, Thomas LL (1984) J Vis Exp. https​://doi.org/10.3791/50766​ Stimulation of basophil and rat mast cell histamine release 32. Sun Y, Jain D, Koziol-White CJ et al (2015) Immunostimulatory by eosinophil granule-derived cationic proteins. J Immunol defective viral genomes from respiratory syncytial virus promote 133:2180–2185 a strong innate antiviral response during infection in mice and 16. Zagai U, Lundahl J, Klominek J et al (2009) Eosinophil cati- humans. PLOS Pathog 11:e1005122. https​://doi.org/10.1371/ onic protein stimulates migration of human lung fibroblasts journ​al.ppat.10051​22 in vitro. Scand J Immunol 69:381–386. https​://doi.org/10.111 33. Dewhurst-Maridor G, Simonet V, Bornand J et al (2004) Devel- 1/j.1365-3083.2009.02233​.x opment of a quantitative TaqMan RT-PCR for respiratory syncy- 17. Trulson A, Byström J, Engström Å et al (2007) The functional tial virus. J Virol Methods 120:41–49. https​://doi.org/10.1016/J. heterogeneity of eosinophil cationic protein is determined JVIRO​MET.2004.03.017 by a gene polymorphism and post-translational modifica- 34. Sun Y, López CB (2016) Respiratory syncytial virus infection in tions. Clin Exp Allergy 37:208–218. https​: //doi.org/10.111 mice and detection of viral genomes in the lung using RTqPCR. 1/j.1365-2222.2007.02644​.x Bio Protoc 6:e1819. https​://doi.org/10.21769​/BioPr​otoc.1819 18. Wang Y-N, Lee H-H, Chou C-K et al (2018) Angiogenin/ribo- 35. Ferrari M, Fornasiero MC, Isetta AM (1990) MTT colorimet- nuclease 5 Is an EGFR ligand and a serum biomarker for erlo- ric assay for testing macrophage cytotoxic activity in vitro. J tinib sensitivity in pancreatic cancer. Cancer Cell 33:752-769. Immunol Methods 131:165–172. https​://doi.org/10.1016/0022- e8. https​://doi.org/10.1016/j.ccell​.2018.02.012 1759(90)90187​-Z 19. Pertea M, Kim D, Pertea GM et al (2016) Transcript-level 36. Domachowske JB, Dyer KD, Bonville CA, Rosenberg HF expression analysis of RNA-seq experiments with HISAT, (1998) Recombinant human eosinophil-derived neurotoxin/ StringTie and Ballgown. Nat Protoc 11:1650–1667. https​:// RNase 2 functions as an effective antiviral agent against res- doi.org/10.1038/nprot​.2016.095 piratory syncytial virus. J Infect Dis 177:1458–1464. https​://doi. 20. Durinck S, Spellman PT, Birney E, Huber W (2009) Mapping org/10.1086/51532​2 identifiers for the integration of genomic datasets with the R/ 37. Rosenberg HF (1995) Recombinant human eosinophil cationic Bioconductor package biomaRt. Nat Protoc 4:1184–1191. https​ protein. Ribonuclease activity is not essential for cytotoxicity. J ://doi.org/10.1038/nprot​.2009.97 Biol Chem 270:7876–7881 21. Love MI, Huber W, Anders S (2014) Moderated estima- 38. Torrent M, Pulido D, Nogués MV, Boix E (2012) Exploring tion of fold change and dispersion for RNA-seq data with new biological functions of amyloids: bacteria cell agglutina- DESeq2. Genome Biol 15:550. https​://doi.org/10.1186/s1305​ tion mediated by host protein aggregation. PLoS Pathog 8:2012– 9-014-0550-8 2014. https​://doi.org/10.1371/journ​al.ppat.10030​05 22. Yu G, Wang L-G, Han Y, He Q-Y (2012) clusterProfiler: an R 39. Han W, Carpenter RL, Cao X, Lo H-W (2013) STAT1 gene package for comparing biological themes among gene clusters. expression is enhanced by nuclear EGFR and HER2 via coop- OMICS 16:284–287. https​://doi.org/10.1089/omi.2011.0118 eration with STAT3. Mol Carcinog 52:959–969. https​://doi. 23. Zhou G, Soufan O, Ewald J et al (2019) NetworkAnalyst 3.0: org/10.1002/mc.21936​ a visual analytics platform for comprehensive gene expres- 40. Guo G, Gong K, Wohlfeld B et al (2015) Ligand-independ- sion profiling and meta-analysis. Nucleic Acids Res. https​:// ent EGFR signaling. Cancer Res 75:3436–3441. https​://doi. doi.org/10.1093/nar/gkz24​0 org/10.1158/0008-5472.CAN-15-0989 24. van Zundert GCP, Rodrigues JPGLM, Trellet M et al (2016) 41. El-Hashim AZ, Khajah MA, Renno WM et al (2017) Src- The HADDOCK2.2 web server: user-friendly integrative mod- dependent EGFR transactivation regulates lung inflammation via eling of biomolecular complexes. J Mol Biol 428:720–725. downstream signaling involving ERK1/2, PI3Kδ/Akt and NFκB https​://doi.org/10.1016/J.JMB.2015.09.014 induction in a murine asthma model. Sci Rep 7:9919. https:​ //doi. 25. Dominguez C, Boelens R, Bonvin AMJJ (2003) HADDOCK: A org/10.1038/s4159​8-017-09349​-0 protein-protein docking approach based on biochemical or bio- 42. Kumaraswamy E, Wendt KL, Augustine LA et al (2015) BRCA1 physical information. J Am Chem Soc 125:1731–1737. https​:// regulation of epidermal growth factor receptor (EGFR) expres- doi.org/10.1021/ja026​939x sion in human breast cancer cells involves microRNA-146a and 26. Konermann S, Brigham MD, Trevino AE et al (2015) Genome- is critical for its tumor suppressor function. Oncogene 34:4333– scale transcriptional activation by an engineered CRISPR-Cas9 4346. https​://doi.org/10.1038/onc.2014.363 complex. Nature 517:583–588. https​://doi.org/10.1038/natur​ 43. Wang SC, Nakajima Y, Yu YL et al (2006) Tyrosine phospho- e1413​6 rylation controls PCNA function through protein stability. Nat 27. Al Yacoub N, Romanowska M, Haritonova N, Foerster J (2007) Cell Biol 8:1359–1368. https​://doi.org/10.1038/ncb15​01 Optimized production and concentration of lentiviral vectors 44. Murray PJ, Allen JE, Biswas SK et al (2014) Macrophage activa- containing large inserts. J Gene Med. https​://doi.org/10.1002/ tion and polarization: nomenclature and experimental guidelines. jgm.1052 Immunity 41:14–20 13 2984 L. Lu et al. 45. Der SD, Zhou A, Williams BRG, Silverman RH (1998) Identifi- 63. Rayego-Mateos S, Rodrigues-Diez R, Morgado-Pascual JL et al cation of genes differentially regulated by interferon alpha, beta, (2018) Role of epidermal growth factor receptor (EGFR) and its or gamma using oligonucleotide arrays. Proc Natl Acad Sci U S ligands in kidney inflammation and damage. Mediators Inflamm A 95:15623–15628. https​://doi.org/10.1073/pnas.95.26.15623​ 2018:1–22. https​://doi.org/10.1155/2018/87394​73 46. Fensterl V, Sen GC (2015) Interferon-induced Ifit proteins: their 64. Wang Y-N, Lee H-H, Hung M-C (2018) A novel ligand-recep- role in viral pathogenesis. J Virol 89:2462–2468. https​://doi. tor relationship between families of ribonucleases and receptor org/10.1128/JVI.02744​-14 tyrosine kinases. J Biomed Sci 25:83. https​://doi.org/10.1186/ 47. Leisching G, Wiid I, Baker B (2017) The association of OASL s1292​9-018-0484-7 and type I interferons in the pathogenesis and survival of intra- 65. Ogiso H, Ishitani R, Nureki O et al (2002) Crystal structure of the cellular replicating bacterial species. Front Cell Infect Microbiol. complex of human epidermal growth factor and receptor extra- https​://doi.org/10.3389/fcimb​.2017.00196​ cellular domains. Cell 110:775–787. https​://doi.org/10.1016/ 48. Carow B, Rottenberg ME (2014) SOCS3, a major regula- S0092​-8674(02)00963​-7 tor of infection and inflammation. Front Immunol. https​://doi. 66. Li S, Schmitz KR, Jeffrey PD et al (2005) Structural basis for org/10.3389/fimmu​.2014.00058​ inhibition of the epidermal growth factor receptor by cetuximab. 49. Mehrabi M, Mansouri K, Soleymani B et al (2017) Develop- Cancer Cell 7:301–311. https:​ //doi.org/10.1016/j.ccr.2005.03.003 ment of a human epidermal growth factor derivative with 67. Rubin J, Zagai U, Blom K et al (2009) The coding ECP EGFR-blocking and depleted biological activities: a compara- 434(G>C) gene polymorphism determines the cytotoxicity of tive in vitro study using EGFR-positive breast cancer cells. Int ECP but has minor effects on fibroblast-mediated gel contraction J Biol Macromol 103:275–285. https​://doi.org/10.1016/j.ijbio​ and no effect on RNase activity. J Immunol 183:445–451. https​ mac.2017.05.035 ://doi.org/10.4049/jimmu​nol.08039​12 50. Sanders JM, Wampole ME, Thakur ML, Wickstrom E (2013) 68. Blom K, Rubin J, Halfvarson J et al (2012) Eosinophil asso- Molecular determinants of epidermal growth factor binding: a ciated genes in the inflammatory bowel disease 4 region: molecular dynamics study. PLoS ONE 8:e54136. https​://doi. Correlation to inflammatory bowel disease revealed. World J org/10.1371/journ​al.pone.00541​36 Gastroenterol 18:6409–6419. https​://doi.org/10.3748/wjg.v18. 51. Zhang J, Rosenberg HF (2000) Sequence in variation at two i44.6409 eosinophil-associated ribonuclease loci in humans. Genetics 69. Salazar VA, Rubin J, Moussaoui M et al (2014) Protein post- 156:1949–1958 translational modification in host defense: the antimicrobial 52. Welch SA, Moore MJ (2007) Erlotinib: success of a molecu- mechanism of action of human eosinophil cationic protein larly targeted agent for the treatment of advanced pancreatic native forms. FEBS J 281:5432–5446. https​://doi.org/10.1111/ cancer. Future Oncol 3:247–254. https​://doi.org/10.2217/14796​ febs.13082​ 694.3.3.247 70. Diop G, Derbois C, Loucoubar C et al (2018) Genetic variants 53. Zasloff M (2002) Antimicrobial peptides of multicellular organ- of RNASE3 (ECP) and susceptibility to severe malaria in Sen- isms. Nature 415:389–395 egalese population. Malar J 17:61. https:​ //doi.org/10.1186/s1293​ 54. Hancock REW, Sahl H-G (2006) Antimicrobial and host-defense 6-018-2205-9 peptides as new anti-infective therapeutic strategies. Nat Biotech- 71. Adu B, Dodoo D, Adukpo S et al (2011) Polymorphisms in nol 24:1551–1557. https​://doi.org/10.1038/nbt12​67 the RNASE3 gene are associated with susceptibility to cer- 55. Yeung ATY, Gellatly SL, Hancock REW (2011) Multifunctional ebral malaria in Ghanaian children. PLoS ONE. https​://doi. cationic host defence peptides and their clinical applications. org/10.1371/journ​al.pone.00294​65 Cell Mol Life Sci 68:2161–2176. https​://doi.org/10.1007/s0001​ 72. Eriksson J, Reimert CM, Kabatereine NB et al (2007) The 8-011-0710-x 434(G>C) polymorphism within the coding sequence of Eosino- 56. Boix E, Nogués MV (2007) Mammalian antimicrobial proteins phil Cationic Protein (ECP) correlates with the natural course of and peptides: overview on the RNase A superfamily members Schistosoma mansoni infection. Int J Parasitol 37:1359–1366. involved in innate host defence. Mol Biosyst 3:317. https​://doi. https​://doi.org/10.1016/j.ijpar​a.2007.04.001 org/10.1039/b6175​27a 73. Blanchard C, Rothenberg ME (2009) Chapter 3 Biology of the 57. Rosenberg HF (2008) RNase A ribonucleases and host defense: Eosinophil, 1st ed. Elsevier Inc. an evolving story. J Leukoc Biol 83:1079–1087. https​://doi. 74. Rosenberg HF, Dyer KD, Foster PS (2013) Eosinophils: changing org/10.1189/jlb.11077​25 perspectives in health and disease. Nat Rev Immunol 13:9–22. 58. Herbst RS (2004) Review of epidermal growth factor recep- https​://doi.org/10.1038/nri33​41 tor biology. Int J Radiat Oncol 59:S21–S26. https ​ : //doi. 75. Zhang J, Rosenberg HF, Nei M (1998) Positive Darwinian selec- org/10.1016/J.IJROB​P.2003.11.041 tion after gene duplication in primate ribonuclease genes. Proc 59. Sigismund S, Avanzato D, Lanzetti L (2018) Emerging func- Natl Acad Sci U S A 95:3708–3713 tions of the EGFR in cancer. Mol Oncol 12:3–20. https​://doi. 76. Prats-Ejarque G, Lu L, Salazar VA et al (2019) Evolutionary org/10.1002/1878-0261.12155​@10.1002/(ISSN)1878-0261. trends in RNA base selectivity within the RNase A superfam- REVIE​WS ily. Front Pharmacol 10:1170. https​://doi.org/10.3389/fphar​ 60. Hardbower DM, Singh K, Asim M et al (2016) EGFR regulates .2019.01170​ macrophage activation and function in bacterial infection. J Clin 77. Altman L, Ayars G, Baker C, Luchtel D (1993) Cytokines and Invest 126:3296–3312. https​://doi.org/10.1172/JCI83​585 eosinophil-derived cationic proteins upregulate intercellular 61. Kalinowski A, Galen BT, Ueki IF et al (2018) Respiratory syn- adhesion molecule-1 on human nasal epithelial cells. J Allergy cytial virus activates epidermal growth factor receptor to sup- Clin Immunol 92:527–536. https ​ : //doi.org/10.1016/0091- press interferon regulatory factor 1-dependent interferon-lambda 6749(93)90077​-S and antiviral defense in airway epithelium. Mucosal Immunol 78. Zagai U, Sköld CM, Trulson A et al (2004) The effect of eosin- 11:958–967. https​://doi.org/10.1038/mi.2017.120 ophils on collagen gel contraction and implications for tissue 62. Minutti CM, Drube S, Blair N et al (2017) Epidermal growth fac- remodelling. Clin Exp Immunol 135:427–433. https​://doi.org/1 tor receptor expression licenses type-2 helper T cells to function 0.1111/j.1365-2249.2004.02396​.x in a T cell receptor-independent fashion. Immunity 47:710-722. 79. Zagai U, Dadfar E, Lundahl J et al (2007) Eosinophil cationic e6. https​://doi.org/10.1016/J.IMMUN​I.2017.09.013 protein stimulates TGF-β1 release by human lung fibroblasts 13 Human RNase3 immune modulation by catalytic-dependent and independent modes in a… 2985 in vitro. Inflammation 30:153–160. https:​ //doi.org/10.1007/s1075​ 93. Kopfnagel V, Wagenknecht S, Brand L et al (2017) RNase 7 3-007-9032-4 downregulates TH2 cytokine production by activated human 80. Sidahmed AME, León AJ, Bosinger SE et al (2012) CXCL10 T-cells. Allergy. https​://doi.org/10.1111/all.13173​ contributes to p38-mediated apoptosis in primary T lympho- 94. Schein CH, Haugg M, Benner SA (1990) Interferon-gamma cytes in vitro. Cytokine 59:433–441. https​://doi.org/10.1016/j. activates the cleavage of double-stranded RNA by bovine cyto.2012.05.002 seminal ribonuclease. FEBS Lett 270:229–232. https​://doi. 81. Abbas YM, Pichlmair A, Górna MW et al (2013) Structural org/10.1016/0014-5793(90)81275​-S basis for viral 5′-PPP-RNA recognition by human IFIT proteins. 95. Schein CH, Hauggt M (1995) Deletions at the C-terminus of Nature 494:60–64. https​://doi.org/10.1038/natur​e1178​3 interferon y reduce RNA binding and activation of double- 82. Zhu J, Zhang Y, Ghosh A et al (2014) Antiviral activity of human stranded-RNA cleavage by bovine seminal ribonuclease. Bio- OASL protein is mediated by enhancing signaling of the RIG-I chem J 307:123–127. https​://doi.org/10.1042/bj307​0123 RNA sensor. Immunity 40:936–948. https​://doi.org/10.1016/J. 96. Chakrabarti A, Jha BK, Silverman RH (2011) New insights into IMMUN​I.2014.05.007 the role of RNase L in innate immunity. J Interferon Cytokine 83. Kristiansen H, Gad HH, Eskildsen-Larsen S et al (2011) The Res 31:49–57. https​://doi.org/10.1089/jir.2010.0120 oligoadenylate synthetase family: an ancient protein family with 97. Wang X, Iyer A, Lyons AB et al (2019) Emerging roles for multiple antiviral activities. J Interferon Cytokine Res 31:41–47. G-protein coupled receptors in development and activation of https​://doi.org/10.1089/jir.2010.0107 macrophages. Front Immunol. https​://doi.org/10.3389/fimmu​ 84. Loo Y-M, Gale M (2011) Immune signaling by RIG-I-like .2019.02031​ Receptors. Immunity 34:680–692. https​://doi.org/10.1016/j. 98. Chai Q, Lu Z, Liu CH (2019) Host defense mechanisms against immun​i.2011.05.003 Mycobacterium tuberculosis. Cell Mol Life Sci 77:1859–1878 85. Zhao Y, Ye X, Dunker W et al (2018) RIG-I like receptor sensing 99. Ostendorf T, Zillinger T, Andryka K et al (2020) Immune sensing of host RNAs facilitates the cell-intrinsic immune response to of synthetic, bacterial, and protozoan RNA by Toll-like recep- KSHV infection. Nat Commun 9:4841. https​://doi.org/10.1038/ tor 8 requires coordinated processing by RNase T2 and RNase s4146​7-018-07314​-7 2. Immunity 52:591-605.e6. https​://doi.org/10.1016/j.immun​ 86. Tsatsaronis JA, Franch-Arroyo S, Resch U, Charpentier E (2018) i.2020.03.009 Extracellular vesicle RNA: a universal mediator of microbial 100. Chattopadhyay S, Veleeparambil M, Poddar D et al (2015) EGFR communication? Trends Microbiol 26:401–410. https ​://doi. kinase activity is required for TLR4 signaling and the septic org/10.1016/J.TIM.2018.02.009 shock response. EMBO Rep 16:1535–1547 87. Damas ND, Fossat N, Scheel TKH (2019) Functional interplay 101. Arranz-Trullén J, Lu L, Pulido D et al (2017) Host antimicrobial between RNA viruses and non-coding RNA in mammals. Non- peptides: the promise of new treatment strategies against tuber- coding RNA 5:7. https​://doi.org/10.3390/ncrna​50100​07 culosis. Front Immunol 8:1499. https​://doi.org/10.3389/fimmu​ 88. Veleeparambil M, Poddar D, Abdulkhalek S et al (2018) Con- .2017.01499​ stitutively bound EGFR–mediated tyrosine phosphorylation of 102. Rivas-Santiago B, Hernandez-Pando R, Carranza C et al (2008) TLR9 is required for its ability to signal. J Immunol 200:2809– Expression of cathelicidin LL-37 during Mycobacterium tuber- 2818. https​://doi.org/10.4049/jimmu​nol.17006​91 culosis infection in human alveolar macrophages, monocytes, 89. Rademacher F, Dreyer S, Kopfnagel V et al (2019) The Anti- neutrophils, and epithelial cells. Infect Immun 76:935–941. https​ microbial and Immunomodulatory Function of RNase 7 in ://doi.org/10.1128/IAI.01218​-07 Skin. Front Immunol 10:1–11. https​://doi.org/10.3389/fimmu​ 103. Rivas-Santiago CE, Rivas-Santiago B, León DA et al (2011) .2019.02553​ Induction of β-defensins by l-isoleucine as novel immunotherapy 90. Schein CH (2001) Producing soluble recombinant RNases in experimental murine tuberculosis. Clin Exp Immunol 164:80– and assays to measure their interaction with interferon-gamma 89. https​://doi.org/10.1111/j.1365-2249.2010.04313​.x in vitro. Methods Mol Biol 160:113–137 91. Schein CH (1997) From housekeeper to microsurgeon: the diag- Publisher’s Note Springer Nature remains neutral with regard to nostic and therapeutic potential of ribonucleases. Nat Biotechnol jurisdictional claims in published maps and institutional affiliations. 15:529–536. https​://doi.org/10.1038/nbt06​97-529 92. Kopfnagel V, Wagenknecht S, Harder J et al (2018) RNase 7 strongly promotes TLR9-mediated DNA sensing by human plas- macytoid dendritic cells. J Invest Dermatol 138:872–881. https​ ://doi.org/10.1016/j.jid.2017.09.052 13