Sanchis et al. Journal of Neuroinflammation (2020) 17:304 https://doi.org/10.1186/s12974-020-01969-0 RESEARCH Open Access A new mouse model to study restoration of interleukin-6 (IL-6) expression in a Cre- dependent manner: microglial IL-6 regulation of experimental autoimmune encephalomyelitis Paula Sanchis1†, Olaya Fernández-Gayol1,2†, Gemma Comes1, Kevin Aguilar1, Anna Escrig1, Mercedes Giralt1, Richard D. Palmiter3 and Juan Hidalgo1* Abstract Background: Interleukin-6 (IL-6) is a pleiotropic cytokine that controls numerous physiological processes both in basal and neuroinflammatory conditions, including the inflammatory response to experimental autoimmune encephalomyelitis (EAE). IL-6 is produced by multiple peripheral and central cells, and until now, the putative roles of IL-6 from different cell types have been evaluated through conditional cell-specific IL-6 knockout mice. Nevertheless, these mice probably undergo compensatory responses of IL-6 from other cells, which makes it difficult to assess the role of each source of IL-6. Methods: To give some insight into this problem, we have produced a novel mouse model: a conditional reversible IL-6 KO mouse (IL6-DIO-KO). By using double-inverted, open-reading-frame (DIO) technology, we created a mouse line with the loss of Il6 expression in all cells that can be restored by the action of Cre recombinase. Since microglia are one of the most important sources and targets of IL-6 into the central nervous system, we have recovered microglial Il6 expression in IL6-DIO-KO mice through breeding to Cx3cr1-CreER mice and subsequent injection of tamoxifen (TAM) when mice were 10–16 weeks old. Then, they were immunized with myelin oligodendrocyte glycoprotein 35-55 peptide (MOG35-55) 7 weeks after TAM treatment to induce EAE. Clinical symptoms and demyelination, CD3 infiltration, and gliosis in the spinal cord were evaluated. Results: IL6-DIO-KO mice were resistant to EAE, validating the new model. Restoration of microglial Il6 was sufficient to develop a mild version of EAE-related clinical symptoms and neuropathology. (Continued on next page) * Correspondence: Juan.Hidalgo@uab.es † Paula Sanchis and Olaya Fernández-Gayol contributed equally to this work. 1 Institute of Neurosciences and Department of Cellular Biology, Physiology and Immunology, Animal Physiology Unit, Faculty of Biosciences, Universitat Autònoma de Barcelona, 08193 Barcelona, Spain Full list of author information is available at the end of the article © The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, 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 included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Sanchis et al. Journal of Neuroinflammation (2020) 17:304 Page 2 of 17 (Continued from previous page) Conclusions: IL6-DIO-KO mouse is an excellent model to understand in detail the role of specific cellular sources of IL-6 within a recovery-of-function paradigm in EAE. Keywords: DIO technology, Conditional reversible IL-6 knockout, Experimental autoimmune encephalomyelitis, Microglial IL-6 Introduction described as one of the most important IL-6-producing Since their inception in the early 80s, genetically engi- brain cells in neuroinflammatory conditions and are neered mice have been a very powerful tool in biomed- clearly implicated in MS and EAE pathogenesis [20–23]. ical research [1]. With time, the technology has been Our results show that this new IL6-DIO-KO mouse refined to allow for conditional knock-out of genes in model is a useful tool to easily investigate the specific specific cell populations [2] or in a timed manner [3]. functions of IL-6 from each cellular source in EAE path- However, the reversibility of these knockouts to allow ology. In particular, in the case of recovery of microglial for recovery of function experiments has been limited IL-6, this was sufficient to initiate mild paralyzing symp- and somewhat flawed. Therefore, in this paper, we intro- toms related to EAE and slightly regulate the inflamma- duce a conditional (Cre-dependent) and reversible tory cascade of EAE. knockout for interleukin-6 (IL-6) using the double- inverted open reading frame (DIO) paradigm [4]. Our Material and methods gene of choice for this proof-of-concept knockout was Animals IL-6 since we had already successfully created a floxed Mice were kept under constant temperature and a 12-h mouse to achieve a conditional knockout [5]. The new light:12-h dark with food and water available ad libitum. mouse (IL6-DIO-KO) lacks IL-6 production in all cells, All experiments were approved by the Ethics Committee which can be turned back on by the action of Cre re- on Animal Experiments of the Universitat Autònoma de combinase. In this paper, we demonstrate the knockout Barcelona and the Generalitat de Catalunya (Refs. 3782 nature of this mouse in its initial form and test its ap- and 9684, respectively). plicability to specific research questions. We show the effect of selective recovery of IL-6 expression in micro- Generation of IL6-DIO-KO mice glia by crossing IL6-DIO-KO with Cx3cr1-CreER mice The conditional reversible IL-6 KO (IL6-DIO-KO) mice [6, 7] in the context of experimental autoimmune en- were developed using double-inverted open reading cephalomyelitis (EAE), the most popular mouse model frame (DIO) technology [4]. Exon 2 of the Il6 gene of multiple sclerosis (MS). (Il6DIOex2) was inverted and flanked by two pairs of mu- IL-6 is a pleiotropic and multifunctional cytokine that tually incompatible loxP sites (lox2272 and loxP) in op- mediates numerous physiological processes and is clas- posite orientations resulting in an inactive Il6 allele. The sically involved in acute and chronic inflammatory con- action of Cre recombinase inverts exon 2 thereby restor- ditions [8]. Indeed, its overproduction has been reported ing normal Il6 expression. in many pathological situations, including MS [9]. In the This mouse was generated as follows: a BAC clone nervous parenchyma of patients with MS, IL-6 is pre- from a C57Bl/6 J mouse containing the Il6 sequence was dominantly linked to the active CNS plaques and located obtained from Invitrogen (RP23-121 M2). A 5614 bp within resident glial cells concentrated in demyelinated fragment with engineered HpaI and PacI sites upstream areas [10]. IL-6 is also essential in the pathogenesis of of exon 2 and a 5241 bp fragment with engineered XhoI EAE, as evidenced by IL-6 KO mice being resistant to it and NotI sites downstream of exon 2 of the Il6 gene [11–14] and displaying altered markers of the disease were synthesized by PCR (using Q5 polymerase, NE Bio- such as reduced infiltration of T lymphocytes and mono- labs) (Fig. 1a, left). Exon 2 flanked by double lox sites cytes into the central nervous system (CNS), reduced ex- with PacI and SalI sites at the ends was synthesized by pression of adhesion molecules in endothelial cells, as Genscript and inserted into a targeting vector in the re- well as defective differentiation of Th17 lymphocytes, verse orientation. The three fragments were sequentially which have been proposed as key factors leading to EAE cloned into a targeting vector containing Pgk-DTa and resistance [11, 15]. In this context, and knowing that IL- HSV-TK genes for negative selection and frt-flanked 6 regulates EAE in a source-specific way, given the var- SvNeoR for positive selection. The final construct was ied results with conditional cell-specific IL-6 KO mice purified from a CsCl gradient, linearized with AscI and [16–19], we chose microglia as the cell population to electroporated into G4 hybrid (C57Bl/6 × Sv129) embry- test our reversible KO. Microglial cells have been onic stem (ES) cells (Fig. 1a, right). Correctly targeted Sanchis et al. Journal of Neuroinflammation (2020) 17:304 Page 3 of 17 Fig. 1 Generation of a conditional reversible KO mouse for IL-6. a Outline of the targeting strategy. Left: 5614 bp HpaI-PacI (pink; upstream) and a 5241 bp XhoI-NotI (orange; downstream) flanking Il6DIOex2 fragments. Right: the final product that consists of a PacI-SalI fragment with inverted exon 2 flanked by double lox sites, and three fragments containing Pgk-DTa and HSV-TK genes for negative selection and frt-flanked SvNeo gene for positive selection. b Southern blot for Il6 gene of the positive ES clones using the probe (left) revealed a ~11 kb band on the wild type allele and a ~8 kb band on the targeted allele, as expected (right); those clones with two bands were selected. c Left: The Il6 gene-modified sequence from IL6-DIO-KO mouse was sequenced by the Sanger method. The resulting sequence had 1361 bp and exon 1 (yellow), exon 2 (blue), lox2272 cassettes (green), loxP cassettes (purple), and FRT (red) were genetic structures observed in that sequence. Middle: PCR products for Il6 gene of WT, IL6-DIO-Het, and IL6-DIO-KO mice yielded the expected bands. Right: IL-6 protein in serum was undetectable by ELISA in IL6-DIO-KO mice after LPS administration. All results are represented as mean ± SEM; ★p ≤ 0.05 vs. LPS-C57BL/6 mice clones were identified by Southern blot; approximately DIO-KO mice), Il6WT/WT (WT mice) and Il6DIOex2/WT 5 μg of DNA from ES cells were digested with NcoI, (IL6-DIO-Het mice, with one functional Il6 allele). electrophoresed on a 0.7% agarose gel and transferred to a nylon membrane. Next, the membrane was hybridized Reactivation of microglia-derived Il6 expression to a unique probe of 621 bp located outside the 5’ arm IL6-DIO-KO mice were crossed with heterozygous (synthesized using the primers Fw: 5’-CAGCATCTCA Cx3cr1-CreER mice, produced by Drs. Jung and Prinz TCTGAGTTCCG-3’ and Rv: 5’-CTCACTGTTC and obtained from the European Mouse Mutant Archive ACAAAGCACAGG-3’) (Additional file 1) that would (EMMA) repository, where were backcrossed with give a ~11 kb band for the wild-type allele and a ~8 kb C57BL/6 mice more than 10 generations [7]. Then, the band for the targeted allele (Fig. 1b). Clones with the offspring Cre-positive animals from the first crossing two bands were reassessed using Neo as a probe. Four were selected and crossed again with homozygous IL6- positive clones (out of 81 analyzed) were correctly tar- DIO-KO mice to obtain mice homozygous for IL6-DIO- geted and had a single NeoR gene insert. Three of those KO and carrying Cx3cr1-CreER allele; these mice are were injected into C57Bl/6 blastocysts, which were im- designated IL6-DIO-ONCx3cr1 and should have Il6 ex- planted in receptive females. Those mice with a high pression only where Cx3cr1 gene is expressed, including percentage of chimerism were then crossed with Gt(RO- microglia. SA)26Sor-FLP recombinase to remove the Sv-NeoR gene [24]. Mice heterozygous for the Il6DIOex2 gene were Genotyping backcrossed >10 times to C57BL/6OlaHsd mice, and Mice were genotyped by PCR analysis of tail and liver non-littermate heterozygous mice for Il6DIOex2 gene DNA, extracted previously by boiling in 100 μl of 50 mM from the tenth crossing were crossed to each other to sodium hydroxide for 7 min, using three primers (Rev1: obtain three possible genotypes: Il6DIOex2/DIOex2 (IL6- 5’-GAGACTGTGAGAGAGGAGTGTG-3’; Rev2: 5’- Sanchis et al. Journal of Neuroinflammation (2020) 17:304 Page 4 of 17 CATCTTATCTGGGCTGACCCTAG-3’; Rev3: 5’-TCTC placing it on a gentleMACS Octo Dissociator for 30 min TGCTGGGATCTAGGGCC-3’) and the following reac- at 37 °C. After dissociation, debris and erythrocytes were tion conditions: 95 °C 2 min, followed by 30 cycles of removed and microglia were isolated using magnetic 95 °C 30 s, 52 °C 30 s, and 72 °C 30 s, and 72 °C 5 min. Fi- microbeads against Cluster of differentiation molecule nally, Cre primers and PCR conditions were adapted 11b (CD11b; Miltenyi 130-093-634) following the manu- from Sanz et al. [25]. facturer’s instructions. The purity of microglia in CD11b+ cell population was previously checked in [26]. Sanger DNA sequencing The CD11b-negative flow-through was named The modified fragment of Il6 gene from IL6-DIO-KO “remaining sample”. DNA from cells was extracted as mouse was sequenced by the Sanger method. Two PCR described above. products were amplified separately using two pairs of primers. The first pair of primers amplified from the Induction of EAE and clinical evaluation middle of the first exon to the end of the third intron The induction of EAE was carried out using adult mice. (1F: 5’-CCCACCAAGAACGATAGTCA-3’ and ΔR: 5’- All animals were immunized under isoflurane anesthesia ATGCCCAGCCTAATCTAGGT-3’ [5] while the second (oxygen flowmeter to 0.8 L/min and isoflurane vaporizer pair amplified from the first intron to the end of the sec- to 4 and 1.5% during induction and maintenance, re- ond intron (fw: 5’-CGATGCTAAACGACGTCACA- spectively) as described [19]. On day 0, mice were 3’and Rev3 primer). Then, the PCR fragments were puri- injected subcutaneously into the hind flanks with an fied by ExoSAP-IT Express PCR Cleanup Reagent (Ther- emulsion of 100 μL MOG35–55 (3 mg/ml; sequence: moFisher, 75001). The enzymatic cleanup method and MEVGWYRSPFSRVVHLYRNGK-carboxyl. Purity: sequencing were performed by Servei de Genòmica i >98%. Peptide synthetized upon request in the Peptide Bioinformàtica (IBB-UAB). The data were imported Synthesis Facility at the Universitat Pompeu Fabra) and firstly into Snapgene Viewer 4.3.4. Software (from 100 μL of complete Freund’s adjuvant (CFA; Sigma- Insightful Science; available at snapgene.com), in which Aldrich F5881) supplemented with 3 mg/ml Mycobacter- both sequences were overlapped creating a unique se- ium tuberculosis H37RA (BD Difco 231141). Control quence with 1361 bp. mice were immunized with an emulsion containing bo- vine serum albumin (BSA; 3 mg/ml; Sigma A9647) in- Tamoxifen treatment stead MOG35–55 peptide. All mice received an Cx3cr1-CreER mouse is a tamoxifen (TAM)-inducible intraperitoneal injection of 500 ng Bordetella pertussis model [7], then the IL6-DIO-ONCx3cr1 mice and their toxin (Native Antigen Company PT-TNL-50) on day 0 littermates were injected for five or eleven consecutive following MOG35-55/BSA injection and on day 2 post- days with TAM. Their littermates were also adminis- immunization. Bordetella pertussis toxin is widely used tered with TAM to control the effects of TAM in the to facilitate the induction of EAE since it destabilizes the analyzed parameters. TAM (Sigma 5648) was dissolved blood-brain barrier and facilitates the migration of T in ethanol at a concentration of 10 mg/100 μl by shaking cells into the CNS. and heating at 37 °C [26]; next, the solution was diluted After immunization, body weight and clinical score with sunflower oil to 10 mg/ml and one daily dose of were monitored daily. The progression of the disease 100 μl (1 mg) was injected intraperitoneally to each was assessed in each animal with a numerical scale fol- mouse. lowing these criteria: 0 = no clinical signs, 0.25 = slight loss of tail tonus, 0.5 = partial loss of tail tonus, 1 = par- Lipopolysaccharide injection alyzed tail, 2 = moderate hind limb paraparesis, 2.5 = se- Mice were injected intraperitoneally with 0.5 mg/Kg vere hind limb paraparesis, 3 = partial hind limb lipopolysaccharide (LPS O55:B5, Sigma L2880) and paralysis, 3.5 = hind limb paralysis, 4 = tetraplegia, and 5 blood was drawn from the tail 90 min later; it was centri- = death. When mice lost more than 10% weight, they fuged at 9.500xg for 10 min at 4 °C and serum was stored were injected intraperitoneally with 200 μl of saline buf- at −80 °C before testing. fer supplemented with 3.6% glucose. We applied the fol- lowing endpoint criteria: (i) self-mutilation; (ii) weight Microglia isolation loss greater than 25%; or (iii) tetraplegia for more than 2 IL6-DIO-KO and IL6-DIO-ONCx3cr1 mice were eutha- days. The time to disease onset and time to peak disease nized 1 month after 11 days of TAM treatment and (days to peak score) were defined by a clinical score ≥1. brains (excluding cerebellum) were quickly dissected The peak score (highest score) and median score were and chopped in small pieces. Then, the tissue was disso- also calculated individually. ciated both enzymatically and mechanically using the Mice were euthanized by decapitation and their spine adult brain dissociation kit (Miltenyi 130-107-677) and was cut between the ninth and tenth thoracic vertebrae. Sanchis et al. Journal of Neuroinflammation (2020) 17:304 Page 5 of 17 The upper part was processed for histological analysis (H2O2)) for 15 min and then blocked in 1% BSA in 0.05 (see “Histology” section) and the lower part was cut M Tris-buffered saline and 0.5% triton-X100 for 1 h. open to extract the spinal cord (segments T12-Co3), The incubation with primary antibodies was overnight at which was snap-frozen in liquid nitrogen and stored at 4 °C. The following day, tissues were incubated with a −80 °C for molecular analysis of the inflammatory re- biotin-conjugated secondary antibody Atom BA-1000 sponse. Blood was collected from the trunk and the (1/300) for 1 h and then, with horseradish peroxidase- serum was obtained as mentioned before and stored at coupled streptavidin (Vector SA-5004, 1/500-600). −80 °C. Immunoreactivity was visualized by adding 50 mg/ml 3, 3-diaminobenzidine-tetrahydrochloride and 0.033% IL-6 enzyme-linked immunosorbent assays (ELISA) H2O2 for 1–5 min, depending on the primary antibody. IL-6 levels in serum were measured using IL-6 ELISA fol- Samples for CD3 were counterstained with hematoxylin lowing manufacturer’s instructions (Bionova 860020192). solution (Sigma MHS16-500 ml). Serial images of the tissues from duplicates of each Real-time polymerase chain reaction (qPCR) animal were taken at ×10 using a Nikon Eclipse 90i RNA was extracted using Promega Maxwell RSC sim- microscope and Nikon Act-1 software. Quantification plyRNA kit following the manufacturer’s instructions. was performed with ImageJ software [27], using a color Then, 2 μg of RNA of each sample were treated with an deconvolution plugin [28] in CD3 IHC counterstained extra DNase I step (Qiagen). After removing DNA from with hematoxylin and LFB/hematoxylin staining to sep- the RNA samples, a two-step qPCR protocol was per- arate both colors and quantify only CD3+ cells and mye- formed using iScript cDNA synthesis kit (BioRad) and lin, respectively. In LFB staining, the percentage of iTaq Universal SYBR Green Supermix (Bio-Rad Labora- demyelination was evaluated in the white matter (WM). tories 1725124). Fold change expression was calculated For CD3 IHC, we quantified the integrated density (sum with an adaptation of delta-delta-Ct method using Glyc- of all the pixel intensities in the region of interest) of eraldehyde 3-phosphate dehydrogenase (Gapdh) as the CD3 per infiltrate. In IBA-1 immunostaining, manual reference gene and the EAE IL6-DIO-Het group as count of IBA-1+ cells in the white and gray (GM) matter calibrator. of the spinal cord was carried out in the first EAE ex- periment whereas the quantification of the integrated Histology density of white and gray matter, as well as the manual The upper part of the vertebral column with the spinal count of IBA-1+ cells in the white matter, were per- cord from sham and MOG35-55-immunized mice was formed in the second EAE experiment. In addition, we fixed with 4% PFA for 24 h. Afterwards, they were care- characterized IBA-1+ cells in terms of their morphology. fully dissected, being the spinal cord post-fixed for an- IBA-1+ cells with long and thin processes (ramified other 24 h in 4% PFA, subsequently embedded in morphology) were classified as resting IBA-1+ cells paraffin and cut sagittally into 8-μm-thick sections. Two whereas those with an enlarged body and small process non-consecutive slices from each animal were used for were considered reactive IBA-1+ cells. IBA-1+ cells with each histological analysis. For myelin evaluation, samples a round body were classified as fully activated cells. In were stained with 0.1% Luxol Fast Blue (LFB) overnight GFAP immunostaining, manual count of GFAP + cells at 56 °C. The following day, the tissue was differentiated were performed in the GM in the first experiment; how- with 0.05% lithium carbonate (Fluka, 62 K70) and 70% ever, this was not possible in the white matter given its ethanol, and counterstained with hematoxylin solution background was too high. The quantification of the inte- (Sigma MHS16). For IHC analysis, microglia/macro- grated density of white and gray matter, as well as man- phages, astrocytes and lymphocytes cluster of differenti- ual count of GFAP + cells in the GM gray matter, were ation 3+ (CD3+) were identified by antibodies against carried out in the second EAE experiment. Furthermore, ionized calcium-binding adaptor molecule-1 (IBA-1; those cells with thin and reduced processes (ramified Wako 019-19741, 1/1500), Glial fibrillary acidic protein morphology) were classified as resting GFAP+ cells (GFAP; Dako Z0334, 1/900), and CD3+ (Dako A0452, 1/ whereas those with thicker hypertrophic processes (star- 100), respectively. shaped morphology) were recorded as reactive GFAP+ Antigen retrieval was necessary in IBA-1 and CD3 cells. All results were normalized to the measured area. IHC. For the former, samples were treated with 10 mM sodium citrate 0.05% tween for 20 min at 96 °C, whereas Statistics for the latter, they were incubated with 1 g/l protease Data is shown as mean ± SEM, plus individual data type XIV (Sigma, P5147) for 8 min at 37 °C. Then, sam- points (except for clinical score), except for EAE sum- ples were incubated with endogenous peroxidase block- mary tables, where median and interquartile range (IQR) ing buffer (70% methanol and 3% hydrogen peroxide are used. Statistical calculations were carried out using Sanchis et al. Journal of Neuroinflammation (2020) 17:304 Page 6 of 17 the Statistical Package for Social Sciences software recognition target [32];) site is also present. The re- (SPSS, version 19) unless otherwise stated. The clinical sult of the sequencing showed the inverted exon 2 score was analyzed with a generalized linear model resulting in an exon that cannot be spliced due to (GzLM) using the area under the curve (AUC), which lack of splice acceptor and donor sites. had been previously calculated using GraphPad Prism. Three primers were designed for the routine genotyp- Body weight gain and incidence were analyzed by gener- ing of the normal and inverted exon 2 (see “Genotyping” alized estimated equations (GEE) and Fisher’s exact test, section). By using Rev3 and Rev2, a product of 358 bp respectively. The day of onset (with score ≥1) and time was amplified for the WT allele of WT and IL6-DIO- to peak (with score ≥1) were examined using Kaplan– Het mice, whereas using Rev3 and Rev1 primers, a 193 Meier survival analysis plus log-rang test, and the post- bp band was obtained for the KO allele of IL6-DIO-Het hoc pairwise comparisons were performed using the and IL6-DIO-KO mice (Fig. 1c, middle). Analysis of IL-6 Holm-Bonferroni correction (R packages “survival” and production in IL6-DIO-KO and non-littermate control “survminer”; RStudio (version 1.2.5033, with R version mice was carried out in serum collected 90 min after a 3.6.3)). Peak score and median score were analyzed by 0.5 mg/kg LPS injection. IL-6 protein levels were dra- Mann–Whitney U test to compare two groups or Krus- matically increased by LPS in C57BL/6OlaHsd mice, but kal–Wallis rank test and Dunn’s test (post hoc) to com- they remained absent in LPS-IL6-DIO-KO mice (Fig. 1c, pare three groups. Histological and gene expression right). results were analyzed by GzLM using genotype as the IL6-DIO-KO mice and littermate WT controls were main factor and a post hoc sequential Bonferroni test immunized with the MOG33-55 peptide, and the clin- was carried out when the main factor was significant. In ical course and body weight were monitored for 19 the case of comparisons between two groups, the Stu- days post-immunization (dpi). MOG-WT mice dent t test was used. Outlier detection was carried out showed the prototypical course of ascending paralysis using the interquartile range (IQR) rule computed from starting at 12 dpi (IQR 3.5) and peaking approxi- Tukey’s hinges, and values more than 3 IQR’s from the mately at 15 dpi (IQR 3.5) with a clinical score of 3 end of the box were considered potential outliers and re- (IQR 1.6) (Fig. 2a, right and 2b). They also showed a moved from the statistical analysis of histological and prominent body weight loss in parallel with the ap- gene expression results. Heatmaps to show IL6 levels (or pearance of clinical signs (Fig. 2a, left). MOG-IL6- gene expression) alongside AUC were created with the R DIO-KO mice were resistant to EAE since only 14% package “pheatmap” with no clustering and color breaks showed minor paralyzing symptoms. The day of dis- set with quantiles. Statistical significance was set at p ≤ ease onset, peak score, time to peak, and median score 0.05 for all analyses. were all statistically significant between genotypes (Fig. 2b and Additional file 3). As expected, IL-6 pro- Results tein was undetectable in serum samples from MOG- Generation and characterization of IL6-DIO-KO mice IL6-DIO-KO mice, but present in their controls (Fig. We created a new murine model with an inactive Il6 al- 2c). Furthermore, a heatmap of IL-6 levels and area lele (see “Generation of IL6-DIO-KO mice” section) that under the curve was performed, reinforcing that the can be specifically reactivated by the action of Cre re- idea that mice with more severe EAE-related symp- combinase (Fig. 1), using a strategy based on DIO tech- toms also show more serum IL-6 levels (Fig. 2d). nology [4, 29–31]. In the absence of Cre recombinase, The upper part of the spinal cord (segments C1–T11) homozygous mice are null for Il6 expression (exon 2 is from MOG-WT mice, but not those from MOG-IL6- inverted) and designated as IL6-DIO-KO. The action of DIO-KO mice, showed clear signs of focal demyelination Cre recombinase flips exon 2 so that it can be spliced at 19 dpi, in agreement with clinical signs. Infiltration of normally thereby restoring Il6 expression; these mice are CD3+ T cells was also prominent in the white matter of referred to as IL6-DIO-ON with a superscript indicating MOG-WT mice whereas it was practically absent in what Cre-driver line was used. MOG-IL6-DIO-KO mice (Fig. 3a and b, top). We also We sequenced the modified Il6DIOex2 gene of IL6- counted both the number of GFAP+ and IBA-1+ cells DIO-KO mice by the Sanger method. By using two along the spinal cord since astrocytes (GFAP+ cells) and pairs of primers, we obtained two sequences that microglia /macrophages (IBA-1+ cells) since they play were overlapped to create a single genetic map of an important role in EAE pathogenesis mediating the in- 1361 bp (Fig. 1c, left and Additional file 2). As ex- flammatory response. A more prominent accumulation pected, 2 pairs of lox cassettes are flanking exon 2 of of GFAP+ cells in the gray matter and IBA-1+ cells both the Il6 gene on each side (LoxP and Lox2272), two in the white and gray matter were seen in spinal cords heterologous loxP sites that cannot recombine with from MOG-WT mice, and to a lesser extent in those each other. In addition, one residual Frt (FLPase from MOG-IL6-DIO-KO mice (Fig. 3a and b, bottom). Sanchis et al. Journal of Neuroinflammation (2020) 17:304 Page 7 of 17 Fig. 2 IL6-DIO-KO mice are resistant to EAE. a Left: Weight loss of WT and IL6-DIO-KO mice after being immunized with MOG35–55 peptide. Right: Clinical evaluation of MOG35-55-immunized WT and IL6-DIO-KO mice. IL6-DIO-KO mice did not develop EAE-related symptoms. b Clinical evaluation of EAE. A significant effect on the disease onset, time to peak disease, peak score, and median score between WT and IL6-DIO-KO mice was observed. The incidence, disease onset, and time to peak were calculated using a clinical score ≥1. c Serum IL-6 protein levels were not detected in IL6-DIO-KO mice at 19 dpi. d Heatmap of IL-6 ELISA levels and the respective AUC for each animal in the experiment. Mice with higher disease scores also showed higher IL-6 levels. It is also clear that the two WT mice that did not develop clinical symptoms also did not have increased IL-6 levels. All results are represented as mean ± SEM; ★p ≤ 0.05 vs. MOG-WT mice IL6-DIO-ONCx3cr1 mice showed genetic rearrangement of later, extracted DNA from the liver (as a control) and Il6 specifically in microglia 1 month after TAM injections the brain, separating its cells in microglia (CD11b+ sam- We evaluated the ability to reactivate Il6 expression se- ple) and the remainder (CD11b- sample) (Fig. 4b, left). lectively in Cx3cr1-expressing cells and defined the dur- This control is necessary because in Cx3Cr1-CreER ation of TAM exposure for optimal inversion of exon 2 mice, CreER is driven by the Cx3cr1 promoter, which is specifically in Cx3cr1-expressing cells. First, we treated expressed in all mononuclear phagocytic cells, which in- 13- to 17-week-old mice with TAM for 5 days (1 mg/ clude macrophages [7]. Also, DNA inversion should be day) and 2 weeks afterwards injected them with LPS. We present longer in microglia than in Cx3cr1-expressing detected IL-6 in the serum of none of LPS-IL6-DIO-KO peripheral cells given microglia’s longevity. Therefore, mice, all of LPS-IL6-DIO-Hz animals, but only in two we showed a partial reversion of Il6 in IL6-DIO- out of six LPS-IL6-DIO-ONCx3cr1 animals, showing that ONCx3cr1 mice by PCR detection of the reversed band recombination was insufficient (data not shown). Since (463 bp) in microglia (CD11b+ sample) but not in a liver exon 2 inversion did not work with 5 days of TAM treat- sample (where presumably there are circulating Kupffer ment, 10- to 16-week-old mice were treated with TAM cells and/or mononuclear cells). As expected, the for 11 days, and with LPS 1 day after the TAM treatment inverted exon 2 band was observed in both tissues from ended. Serum IL-6 now increased modestly in all LPS- IL6-DIO-KO mice (Fig. 4b, right). injected IL6-DIO-ONCx3cr1 mice while all LPS-treated IL6-DIO-KO controls remained undetectable. LPS- Reactivation of Il6 expression in microglia partially treated IL6-DIO-Het mice produced high levels of IL-6, rescues EAE pathology as expected (Fig. 4a). This protocol was therefore estab- IL6-DIO-Het, IL6-DIO-ONCx3cr1, and IL6-DIO-KO fe- lished for subsequent use. males were used to physiologically confirm Il6 reversion To test the specificity of recombination in microglial in Cx3cr1-expressing cells. These mice were immunized cells, we reactivated Il6 expression in Cx3cr1-expressing with MOG35-55 peptide 7 weeks after being injected with cells by treating with TAM for 11 days and, 4 weeks a total of 11 mg of TAM. A group of BSA-injected Sanchis et al. Journal of Neuroinflammation (2020) 17:304 Page 8 of 17 Fig. 3 MOG35-55-immunized IL6-DIO-KO mice showed an absence of demyelination and CD3 infiltrates, and limited gliosis. a Representative images of LFB staining and CD3 infiltrates (top), astrogliosis and microgliosis (bottom) of the spinal cord from MOG-WT and MOG-IL6-DIO-KO mice. The contrast of the representative images was enhanced. The lack of IL-6 prevented the infiltration of inflammatory cells. The discontinuous line delimits white matter (WM) from gray matter (GM). Scale bar: 100 μm. b From top to bottom: Quantification of demyelination by LFB staining and CD3 immunostaining levels (top) and the total number of GFAP+ and IBA-1+ cells (bottom) in the spinal cord from MOG-WT and MOG-IL6- DIO-KO mice. All results are represented as mean ± SEM; ★p ≤ 0.05 vs. MOG-WT mice female and male IL6-DIO-Het mice (but not injected IL6-DIO-KO mice, and between MOG-IL6-DIO-Het previously with TAM either LPS) was also included as a and MOG-IL6-DIO-ONCx3cr1 mice. Furthermore, the se- control for illustrating the neuropathology of EAE. verity of the disease was also analyzed by the evaluation Following MOG35-55 immunization, MOG-IL6-DIO- of the AUC for each mouse from 21 to 27 dpi. We ob- Het mice showed the typical ascending paralysis course served differences between MOG-IL6-DIO-Het and of EAE, starting at 13 dpi (IQR 1.5) and peaking at 18 MOG-IL6-DIO-ONCx3cr1 mice (p < 0.001), and MOG- dpi (IQR 7) with a clinical score of 2.5 (IQR 0.5) (Fig. 5a, IL6-DIO-Het and MOG-IL6-DIO-KO mice (p < 0.001). middle and Fig. 5b). Moreover, a prominent body weight Regarding the day of disease onset, peak score, and me- loss occurred in these mice in parallel with the clinical dian score, MOG-IL6-DIO-Het mice were significantly signs (Fig. 5a, left). Consistent with our initial results, different from both MOG-IL6-DIO-ONCx3cr1 and MOG-IL6-DIO-KO mice were almost completely resist- MOG-IL6-DIO-KO mice (Fig. 5b and Additional file 4). ant to EAE (Fig. 5a, b). In contrast, mice with the recov- Immunization with BSA did not cause any paralyzing ery of Il6 only in microglia (MOG-IL6-DIO-ONCx3cr1) symptoms in IL6-DIO-Het mice. Furthermore, they showed a delayed progression of the clinical course, with maintained body weight for all days analyzed, as ex- paralyzed tail at 19 dpi (IQR 10) and peaking at 22 dpi pected (data not shown). (IQR 7.5), with a peak score of 1 (IQR 1.75) (Fig. 5a and Il6 mRNA levels were higher in the lower part of the b). As shown in Fig. 5b, the incidence of disease symp- spinal cord (segments T12-Co3) from MOG-IL6-DIO- toms (clinical score ≥1) was 100% in MOG-IL6-DIO- Het than MOG-IL6-DIO-ONCx3cr1 mice and were un- Het and 57% in MOG-IL6-DIO-ONCx3cr1 mice. No mice detectable in MOG-IL6-DIO-KO mice at 27 dpi, even died during the experiment (27 days) (Fig. 5b). though no differences were significant (Fig 6a, top left). Significant differences regarding EAE symptomatology However, Il6 levels were in line with the disease extent were observed between MOG-IL6-DIO-Het and MOG- (AUC) (Fig. 6b). In addition, to dissect the putative role Sanchis et al. Journal of Neuroinflammation (2020) 17:304 Page 9 of 17 Fig. 4 Reactivation of microglia-derived Il6 expression in IL6-DIO-ONCx3cr1 mice. a Serum IL-6 protein levels of LPS-injected IL6-DIO-Het, IL6-DIO- ONCx3cr1 and IL6-DIO-KO mice 1 day following 11 days of TAM treatment. b Left: IL6-DIO-ONCx3cr1 and IL6-DIO-KO mice were euthanized 1 month after TAM treatment and the brains were homogenized and incubated with CD11b+ antibody. Besides, the positive cell population was selected using columns with ferromagnetic spheres. Right: The reversion band (463 bp) was only seen in the brain CD11b+ fraction of IL6-DIO-ONCx3cr1 mice. All results are represented as mean ± SEM of microglial IL-6 in EAE dynamics, we also analyzed MOG35-55-immunization also caused an increase of the expression of several genes related to different sub- microgliosis (analyzed both by counting IBA-1+ cells sets of T cells (Rorc, which encodes RORyt protein, and and quantitating IBA-1 immunostaining levels) in the Foxp3) and Il17a expression. The expression of all these white and gray matter of the upper part of the spinal genes was higher in MOG-IL6-DIO-Het mice than cord (segments C1–T11) from IL6-DIO-Het mice com- MOG-IL6-DIO-KO and MOG-IL6-DIO-ONCx3cr1 mice, pared with those immunized with BSA. In the white but no differences were seen between MOG-IL6-DIO- matter, MOG-IL6-DIO-Het showed more microgliosis KO and MOG-IL6-DIO-ONCx3cr1 mice (Fig. 6a). than MOG-IL6-DIO-KO and MOG-IL6-DIO-ONCx3cr1 In accordance with the clinical signs, the upper part of mice (Fig. 8a and b, and Additional file 5, left top). Al- the spinal cord (segments C1–T11) from MOG- though the IBA-1 immunostaining levels in the white immunized IL6-DIO-Het mice, but not those immu- matter of MOG-IL6-DIO-ONCx3cr1 and MOG-IL6-DIO- nized with BSA, showed demyelination of the white mat- KO mice were not different, more IBA-1+ cells were ter. Moreover, a significant effect of IL-6 was observed, counted in MOG-IL6-DIO-ONCx3cr1 than MOG-IL6- since MOG-immunized IL6-DIO-KO mice were practic- DIO-KO mice. In the gray matter, MOG-IL6-DIO-Het ally equal to those of BSA-immunized IL6-DIO-Het showed more IBA-1 immunostaining levels than MOG- mice, while recovery of microglial IL-6 in MOG-IL6- IL6-DIO-KO mice, but not compared with MOG-IL6- DIO-ONCx3cr1 mice, reverted the demyelination (seem- DIO-ONCx3cr1 mice (Fig. 8a, left, and Additional file 5, ingly partially, but it was not significantly different from left bottom). These results were similar to those for de- MOG-immunized IL6-DIO-Het mice) (Fig. 7). myelination and suggesting that recovery of microglial Sanchis et al. Journal of Neuroinflammation (2020) 17:304 Page 10 of 17 Fig. 5 IL6-DIO-ONCx3cr1 showed mild EAE-related symptoms. a Clinical course of MOG-IL6-DIO-Het, MOG-IL6-DIO-ONCx3cr1 and MOG-IL6-DIO-KO mice. Body weight (left) and clinical signs (middle) changes were observed in IL6-DIO-Het and IL6-DIO-ONCx3cr1 mice whereas IL6-DIO-KO mice were almost completely resistant to EAE. Area under the curve (AUC; right) was calculated from EAE clinical course for each mouse between days 21 and 27. b Clinical evaluation of EAE. IL6-DIO-ONCx3cr1 and IL6-DIO-KO mice were significantly less affected than IL6-DIO-Het mice. The incidence, disease onset, and time to peak were calculated using a clinical score ≥1. Results in (a) are represented as mean ± SEM, while (b) shows median and IQR (except for incidence and mortality). NA (not applicable) shows the impossibility to calculate the median (given the criterion for disease). ★p ≤ 0.05 vs. MOG-IL6-DIO-Het mice IL-6 in MOG-immunized IL6-DIO-ONCx3cr1 mice might among MOG-IL6-DIO-KO and IL6-DIO-ONCx3cr1 mice partially revert the microgliosis in this spinal cord. (p = 0.098). A detailed study about the morphology of IBA-1+ cells In addition, we analyzed the astrogliosis by count- (microglia/macrophages) in the white matter was also car- ing GFAP+ cells and studying GFAP immunostaining ried out (Fig. 8a, right, and 8b, differentiating three kinds levels in the upper part of the spinal cord (segments of cells: ramified cells (resting; Fig. 8a, right, panel 1), re- C1–T11). In particular, no differences between treat- active cells (activated; Fig. 8a, right, panel 2), and round ments and genotypes were appreciated concerning cells (fully activated; Fig. 8a, right, panel 3). The 98.2 ± GFAP immunostaining levels in the white matter, 0.72 % of IBA-1+ cells from negative control (BSA- perhaps because of the myelin tracts also seemed to injected) IL6-DIO-Het mice displayed a ramified morph- be immunostained, thus increasing background levels ology. Similar results were seen in MOG-IL6-DIO-KO (Fig. 8a, middle and Additional file 5, right top). Re- since 96.7 ± 0.70% of IBA-1 cells displayed a ramified garding the gray matter, MOG35-55 immunization morphology whereas 3.3 ± 0.7% seemed to take reactive showed a trend to increase the GFAP immunostain- morphology. Among all cell types, the reactive IBA-1+ ing levels in IL6-DIO-Het compared with those with cells were the largest population (78.9 ± 1.70%) present in BSA immunization (p = 0.067), being this difference MOG-IL6-DIO-Het mice. The IL6-DIO-ONCx3cr1 mice, statistically significant when we counted the number however, displayed 64.9 ± 14.7% and 34.9 ± 14.5% of rami- of GFAP+ cells. Both GFAP immunostaining levels fied and reactive IBA-1+ cells, respectively. A treatment- and the number of GFAP+ cells were different dependent effect was observed in all three groups of IBA- among MOG-IL6-DIO-Het and MOG-IL6-DIO-KO 1 + cells. Regarding MOG-immunized mice, MOG-IL6- mice (Fig. 8c and Additional file 5, right bottom). DIO-Het mice showed more ramified and reactive IBA-1+ Although GFAP immunostaining levels in the gray cells compared with MOG-IL6-DIO-KO and IL6-DIO- matter were not statistically significant between ONCx3cr1 mice, and IL6-DIO-ONCx3cr1 mice showed more MOG-IL6-DIO-Het and MOG-IL6-DIO-ONCx3cr1 reactive, but not ramified IBA-1+ cells compared with mice, we counted a higher number of GFAP+ cells MOG-IL6-DIO-KO mice. Concerning round IBA-1+ cells, in MOG-IL6-DIO-Het than MOG-IL6-DIO-ONCx3cr1 a significant increase was appreciated in MOG-IL6-DIO- mice in this matter (Fig. 8c and Additional file 5, Het compared with MOG-IL6-DIO-KO mice whereas a right bottom). trend (p = 0.098) was shown in MOG-IL6-DIO-Het as to We also classified the number of astrocytes in the gray IL6-DIO-ONCx3cr1 mice. The same trend was observed matter of the spinal cord according to their ramified Sanchis et al. Journal of Neuroinflammation (2020) 17:304 Page 11 of 17 Fig. 6 Inflammation-related genes are down-regulated accordingly to disease severity. The changes of inflammation-related genes expression were studied using the upper part of the spinal cord (segments C1–T11) of three groups of mice euthanized at 27 dpi. a Gene expression of Il6, Il17a, Foxp3, and Rorc was lower in MOG-IL6-DIO-ONCx3cr1 and MOG-IL6-DIO-KO mice than in MOG-IL6-DIO-Het control mice. b Heatmap of Il6 expression and AUC of disease score. Il6 was only increased in mice with higher AUC. Results in (a) are represented as mean ± SEM; ★p ≤ 0.05 vs. MOG-IL6-DIO-Het mice (Fig. 8a, right, panel 4) or reactive (Fig. 8a, right, panel mice tended to increase the number of reactive GFAP+ 5) morphology. In IL6-DIO-Het mice immunized with cells in comparison with MOG-IL6-DIO-KO mice (p = BSA, 93.5 ± 1.15 % of GFAP+ cells displayed a resting 0.063; Fig. 8b, right). ramified morphology, while in those immunized with MOG35-55, 33.9 ± 4.76 % had a ramified morphology Discussion and 66.1 ± 4.76 % GFAP+ cells seemed to take a star- Traditional methods for creating reversible KO mice shaped morphology. In contrast, 85.0 ± 2.93% and 63.2 are based on strategies that require numerous cross- ± 10.30% of GFAP+ cells displayed a resting morphology ings or the continuous administration of drugs. The in the gray matter of MOG-IL6-DIO-KO and IL6-DIO- main methods are (1) cassette stop/trap flanked by ONCx3cr1 mice, respectively. MOG35-55-immunization loxP sites [33–35], (2) Tet-off and tet-on [36–38], decreased ramified GFAP+ cells, but increased reactive and (3) LOFT: LoxP-flippase (FLP) recognition target GFAP+ cells in IL6-DIO-Het mice compared with those (FRT) Trap [39]. with BSA-immunization. The number of ramified The stop-cassette strategy is based on creating a GFAP+ cells were lower in MOG-IL6-DIO-Het and knockout by introducing transcription termination ele- MOG-IL6-DIO-ONCx3cr1 than MOG-IL6-DIO-KO mice, ments that prevent the complete production of the tar- and a trend to increase the number of ramified GFAP+ get protein. Reversibility is achieved by driving Cre cells was also appreciated in MOG-IL6-DIO-ONCx3cr1 expression with tissue-specific promoters fused with the compared to MOG-IL6-DIO-Het mice (p = 0.081). estrogen receptor (ER) paired with tamoxifen injection Genotypic differences were also observed regarding the for temporal and spatial specificity. Alternatively, injec- number of GFAP+ cells with a reactive morphology, tion of a Cre-expressing virus can rescue in a similarly which was higher in MOG-IL6-DIO-Het than MOG- time- and spatial-specific manner. However, stop cas- IL6-DIO-KO and MOG-IL6-DIO-ONCx3cr1 mice. The settes can be leaky and trace expression must be recovery of microglial IL-6 in MOG-IL6-DIO-ONCx3cr1 assessed empirically. Sanchis et al. Journal of Neuroinflammation (2020) 17:304 Page 12 of 17 Fig. 7 Demyelination was mildly affected by the recovery of microglial IL-6 at 27 dpi. Representative images of LFB/H&E staining and percentage of demyelination in the white matter of the spinal cord from BSA-immunized IL6-DIO-Het and MOG35-55-immunized IL6-DIO-Het, IL6-DIO-ONCx3cr1, and IL6-DIO-KO mice. A prominent accumulation of cells was shown in MOG-IL6-DIO-Het mice, and to a lesser extent in MOG-IL6-DIO-ONCx3cr1 mice. The contrast of the representative images was enhanced. The discontinuous line delimits white matter (WM) from gray matter (GM). Scale bar: 100 μm. Results were normalized to the total area and are represented as mean ± SEM; ●p ≤ 0.05 vs. BSA-IL6-DIO-Het mice; ★p ≤ 0.05 vs. MOG-IL6-DIO-Het mice Tet-off and Tet-on use Escherichia coli’s tetO operon require extensive breeding and animals that are trans- to induce expression of the adjacent gene. In the “off” genic for multiple genes. system, the mouse becomes a KO upon (and only dur- The LOFT system uses two modified alleles for the ing) administration of tetracycline (or, in practice, its de- gene of interest, one of them floxed and the other with rivative doxycycline), which then binds to a tetracycline- an frt-flanked gene-trap cassette (that also has Neo and controlled transactivator (tTA) that has been engineered GFP) inserted in an intron. The latter will produce by to be constitutively expressed, and prevents the tran- default an inactive fusion protein with the N-terminal scription of the gene downstream of tetO. In the “on” domain of the target protein and Neo. When Cre is system, the transactivator is modified to have the reverse expressed, the animal will become a KO. If there is add- phenotype (rtTA), i.e., it only binds to tetO in the pres- itionally flippase activity, the phenotype will be reversed. ence of doxycycline. In this case, the animal is initially a This method requires a considerable number of cross- KO, and it will only produce the target protein upon ings to incorporate the two alleles and recombinases drug administration. As downsides, both strategies de- making it labor- and cost-intensive. In addition, given pend on continuous administration of the drug to main- that the allele with the gene-trap is null by default, the tain the animal’s genotype; the use of a stop cassette gene must be haplosufficient. with loxP sites in order to obtain tissue specificity has To overcome these pitfalls, we have generated a condi- the problem mentioned above; and furthermore, they tional and reversible KO mouse using the widely used in Sanchis et al. Journal of Neuroinflammation (2020) 17:304 Page 13 of 17 Fig. 8 Microglial IL-6 plays a minor role in the regulation of microgliosis and astrogliosis. a Representative images of IBA-1 (left) and GFAP (middle) immunostaining of BSA-immunized IL6-DIO-Het and MOG35-55-immunized IL6-DIO-Het, IL6-DIO-ONCx3cr1 and IL6-DIO-KO mice. High magnification (right) of (a) showing the different morphologies of IBA-1+ and GFAP+ cells: (1) ramified, resting IBA-1+ cells; (2) reactive IBA-1+ cells, with shorter and thicker cell processes (black arrows); (3) round IBA-1+ cells (dashed black arrow); (4) Ramified, resting GFAP+ cells; (5) reactive, hypertrophic GFAP+ cells. The contrast of the representative images was enhanced. The discontinuous line delimits white matter (WM) from gray matter (GM). Scale bar: 100 μm. b The total number of IBA-1+ cells and the number of IBA-1+ cells showing a resting phenotype (ramified), partially activated (reactive), and fully activated (round) counted in the white matter of the spinal cord from BSA-immunized WT and MOG35-55-immunized IL6-DIO- Het, IL6-DIO-ONCx3cr1, and IL6-DIO-KO mice. c The total number of GFAP+ cells and the number of GFAP+ cells showing a resting phenotype (ramified) and activated (reactive) counted in the gray matter of the spinal cord from BSA-immunized WT and MOG35-55-immunized IL6-DIO-Het, IL6-DIO-ONCx3cr1, and IL6-DIO-KO mice. All results were relativized per total area and are represented as mean ± SEM; ●p ≤ 0.05 vs. BSA-IL6-DIO- Het mice; ★p ≤ 0.05 vs. IL6-DIO-Het mice. ☆p ≤ 0.05 vs. IL6-DIO-KO mice opto- and chemogenetics double-inverted, open-reading- addition, WT mice immunized with MOG35-55 peptide frame paradigm [4, 29–31]. This technique consists of and orally treated during ongoing EAE with tocilizumab, flanking exons of interest by pairs of incompatible loxP a humanized monoclonal antibody that inhibits IL-6R- sites such that Cre recombinase flips the exons and then mediated signaling, showed a drastic reduction of EAE- locks them into the flipped orientation [4, 29–31]. Exon related symptoms [43]. It is therefore clear that IL-6 2 of the Il6 gene was inverted in IL6-DIO-KO mice signaling is involved in the pathogenesis of EAE. How- causing a depletion of IL-6 protein (undetectable after ever, given the functional complexity of IL-6, the priority inflammation challenges such as LPS and EAE), which is of many groups during the last decades has been to equivalent to results with the total IL-6 KO mice [12, study the multiple cellular sources of IL-6 to better 40]. In our EAE paradigm, almost all IL6-DIO-KO mice understand the contribution of each one to the EAE were resistant to MOG35-55-inducible EAE, except for pathogenesis. Until now, the roles of IL-6 in a specific one, which presented EAE-related symptoms even with- context have been described through the use of mice out showing serum IL-6 levels, as previously observed lacking IL-6 throughout the body or in specific cells with other IL-6 KO models [11, 12, 41]. (total and conditional IL-6 KO mice, respectively). The EAE-related symptoms of IL6-DIO-KO mice were Nevertheless, the interpretation of the results using the similar to those of WT mice immunized with MOG35-55 latter model is difficult because the data could be the re- peptide and administered intraperitoneally with anti-IL- sult of a compensatory response from IL-6 of other cells. 6 receptor monoclonal antibody (clone MR16-1) at the With our new conditional reversible IL-6 KO mouse, we same day of EAE induction, but different to those of now have a tool that allows for the reactivation of Il6 ex- WT mice immunized with MOG35-55 peptide and pression in cells expressing Cre recombinase. Thus, we treated with clone MR16-1 after disease onset [42]. In were able to recover microglial IL-6 in the reversible Sanchis et al. Journal of Neuroinflammation (2020) 17:304 Page 14 of 17 total IL-6 KO mouse through breeding to Cx3cr1-CreER immunosuppressant which could interfere with EAE mice and subsequent TAM administration. Our results evaluation [50]. suggest that recovery of microglial IL-6 gives rise to a The longevity and limited self-renewal are the main mild version of EAE, albeit with a limited role in the features of microglia which distinguishes them from regulation of the inflammatory cascade. In conclusion, other mononuclear phagocytic cells [51, 52]. In fact, we have verified that the new reversible total IL-6 KO is Goldmann and colleagues observed that inducible an excellent tool to study the role of specific cellular Cx3cr1CreER:R26-YFP mice, obtained by crossing sources of IL-6 within a recovery-of-function paradigm. Cx3Cr1CreER mice with R26-YFP mice, displayed YFP- Although EAE is considered a peripheral disease, labeled microglia at all times after TAM treatment microglia, besides CD4+ cells, have been described as whereas other mononuclear phagocytic cells lost the responsible for EAE initiation [23, 44]. In fact, in YFP-label for the first 4 weeks after treatment [6]. We MOG35-55-immunized C57BL/6 mice, CD4+ T cells also observed similar results using our inducible micro- infiltrated the brain before mice developed EAE- glial IL-6 KO mice, obtained by crossing Cx3Cr1CreER related symptoms, coinciding with activation of mice with Il6lox/lox mice, since microglia of these mice, CD11b+ microglia and production of IL-1β, TNF-α, but not other brain cells or liver (assuming this tissue as and IL-6 [45]. Furthermore, we have recently ob- representative of macrophages/monocytes), kept the served that microglia are one of the main brain genomic modification 4 weeks following TAM adminis- sources of IL-6 during EAE pathogenesis and micro- tration [26]. Therefore, we also performed the validation glial Il6-deficient females seemed to be less affected of IL6-DIO-ONCx3cr1 mice at the same time as above by MOG35-55 immunization than their controls [19]. mentioned. As expected, IL6-DIO-ONCx3cr1 mice kept In this paper, we have focused on studying the effect the reversion of the DIO allele in microglia, but not in of selective recovery of microglial IL-6 expression by other brain cells or liver, 4 weeks after being treated with crossing IL6-DIO-KO with Cx3cr1-CreER mice to TAM for 11 days. These data confirm that the contribu- show the applicability of this new system and to gain tion of mononuclear phagocytic cells to results obtained insight into the putative role(s) of microglial IL-6 in with IL6-DIO-ONCx3cr1 mice from 4 weeks following the EAE disease. TAM-treatment will be limited. In Cx3Cr1CreER mice, Cre is driven by the Cx3cr1 Reactivation of microglial Il6 expression in MOG35-55- promoter, which is broadly expressed in the mono- immunized IL6-DIO-ONCx3cr1 females developed some nuclear phagocytic cells and it is only expressed in of the prototypical paralyzing symptoms showing that microglia in the adult brain [7, 46, 47]. Furthermore, despite the fact that EAE is considered essentially a per- Cre recombinase protein of these mice is fused to the ipheral disease, microglial Il6 was sufficient to trigger a estrogen ligand-binding domain (CreER), thereby re- mild version of EAE-related symptoms. Results are in quiring TAM for its activation [6, 7]. Then, IL6-DIO- line with females lacking IL-6 in microglia, which tended ONCx3cr1 mice should revert Il6 expression in to reduce EAE score during the acute phase of the dis- Cx3cr1-expressing cells, including macrophages and ease in comparison with their controls [19]. Clinical microglia, after TAM treatment. Since IL-6 protein signs of IL6-DIO-ONCx3cr1 mice were lower than those levels in serum are practically undetected in basal of IL6-DIO-Het animals, reinforcing the idea that IL-6 conditions, mice were also injected with LPS follow- derived from other cellular sources is also involved in ing TAM treatment to increase IL-6 production. The the EAE-related symptomatology. Of note, scores (AUC) recovery of IL-6 occurred in IL6-DIO-ONCx3cr1 mice within the IL6-DIO-ONCx3cr1 group have a bimodal dis- administrated with TAM for 11 days and not for 5 tribution, which also does not fully correlate with Il6 ex- days. However, IL-6 production was lower in LPS- pression. This could be due to only partial recovery of IL6-DIO-ONCx3cr1 than LPS-IL6-DIO-Het mice, microglial Il6 or to this source not being as relevant to probably due to the fact that only CX3CR1+ cells in disease progression and just indicating a stochastic IL6-DIO-ONCx3cr1 mice would be responsible for IL-6 process. Indeed, T-cell- and astrocyte-derived Il6-defi- production following LPS administration while all cient mice showed ameliorated symptoms compared cells would produce IL-6 protein in IL6-DIO-Het with their controls after being immunized with mice (these mice have one WT allele and one DIO MOG35-55 [17–19]. However, deficiency of dendritic allele; in addition, TAM administration would cause cell-derived Il6 caused that naïve T cells could not dif- the Cre activation and subsequent Il6 reversion of the ferentiate into Th17 cells blocking EAE score in mice DIO allele in CX3CR1+ cells in these mice). We did [17]. In contrast, Il6 deficiency in other cells caused no not attempt longer TAM administrations given that (in neurons), contradictory (in B cells), or aggravated our current dose is already at the higher end of the (macrophage) effects in the regulation of EAE-related reported spectrum [48, 49] and the drug is a known symptoms [16, 17, 19]. Sanchis et al. Journal of Neuroinflammation (2020) 17:304 Page 15 of 17 The expression of mRNA encoding Il6 was higher in immunized with MOG35-55 [11, 12], and highlighting the lower part of the spinal cord (segments T12-Co3) once again a critical role of IL-6 in the prototypical from MOG-IL6-DIO-Het than MOG-IL6-DIO-ONCx3cr1 autoimmune disorder. Interestingly, the recovery of mice and it was absent in MOG-IL6-DIO-KO mice at 27 microglial IL-6 in MOG-IL6-DIO-ONCx3cr1 females dpi, but high variability precluded any statistically sig- seemed to revert partially the demyelination, in agree- nificant effect between genotypes. Moreover, we cannot ment with results obtained with microglial Il6-defi- rule out that differences would exist at other time cient females at 29 dpi, which showed a trend to points. EAE is a complex disease in which IL-6 undoubt- decrease the demyelination [19], pointing to in both edly plays an important role since it suppresses the dif- cases that IL-6 derived from microglia might regulate ferentiation of TGF-β-induced Foxp3+ Treg, and with the demyelination. As stated, another key pathological TGF-β promotes the differentiation of pathogenic TH17 feature of EAE is the inflammatory response of astro- cells from naive T cells [15]. In this paper, IL6-DIO-KO cytes and microglia/macrophages, therefore we quan- mice showed downregulation of Rorc expression, the tified the GFAP and IBA-1 immunostaining levels and crucial transcription factor of Th17, suggesting a reduc- classify IBA-1 and GFAP+ cells according to their tion of Th17 differentiation, and a similar trend regard- morphology. In general, the lack of IL-6 reduced the ing Il17 mRNA levels in the spinal cord compared to microgliosis and astrogliosis in the spinal cord of IL6-DIO-Het. Indeed, less percentage of Th17 cells infil- mice, displaying most of IBA-1+ and GFAP+ cells a trating the CNS was previously reported using resting morphology, in line with IL-6 KO [12]. The MOG35-55-immunized IL-6 KO mice [15]. Surprisingly, recovery of microglial IL-6, in agreement with the the recovery of microglial IL-6 did not modify Il17 ex- mild clinical signs of MOG-IL6-DIO-ONCx3cr1 mice, pression in comparison with the lack of total IL-6 in increased the number of reactive IBA-1+ cells in the mice at 27 dpi, as it was observed in IL-17 immunostain- white matter and a similar trend was observed regard- ing levels in microglial Il6 deficient mice at 27–29 dpi ing IBA-1 immunostaining levels in the gray matter. [19]. Besides, the expression of mRNA encoding Foxp3, Somewhat surprisingly, considering the relevance of the master transcription factor of Treg, was also down- IL-6 in inflammatory processes of EAE, is that the re- regulated by lack of total IL-6, in agreement with IL-6 covery of microglial IL-6 in MOG-IL6-DIO-ONCx3cr1 KO mice, whose cerebellum showed less percentage of mice did not modify the total number of GFAP+ cells Foxp3 cells per infiltrate than those of MOG35–55-immu- in the gray matter, although it reduced the ramified nized WT mice [12]. Interestingly, another work re- GFAP+ cells and tended to increase the reactive ported that IL-6 KO mice (crossed previously with the GFAP+ cells. Taking into account the results obtained Foxp3gfp.KI mice to monitor Treg cells) showed more with microglial Il6-deficient females at 29 dpi, in Foxp3+ Treg cells in the draining lymph nodes than which deficiency of microglial Il6 tended to reduce their controls [53], which suggests the necessity of fur- IBA-1 immunostaining levels in the white matter, but ther studies to determine these discrepancies. Foxp3 and it did not play a relevant role in the regulation and Rorc expression levels were practically equal among GFAP immunostaining levels [19], it seems that MOG-IL6-DIO-ONCx3cr1 and MOG-IL6-DIO-KO mice microglial Il6 may regulate the microgliosis, but it at 27 dpi, but distinct to MOG-IL6-DIO-Het, suggesting would have a limited function in the regulation of that other cellular sources of IL6, and not microglia, astrogliosis during the inflammatory response to EAE. might regulate these transcription factors at this time point. IL-6 has also been described as a regulatory pro- Conclusions tein of some inflammation-related genes at 15-20 dpi In summary, we have presented and validated a novel re- [12, 54]. In particular, microglial Il6 deficiency caused versible total IL-6 KO mouse, whose Il6 expression was changes in Gfap and Cd206 (alternatively activated mac- recovered specifically in microglia. The recovery of rophages/microglia (M2) phenotypic marker) expression microglial Il6 was sufficient to develop a mild version of in the lower part of the spinal cord (segments T12-Co3) EAE-related symptoms in females, although it seemed to from mice at 15 dpi [19]. Future studies are needed to have a minor role in the regulation of the inflammatory determine what occurs regarding inflammation-related cascade. Furthermore, the reversion of Il6 expression in markers in the MOG-IL6-DIO-ONCx3cr1 mice at peak different cells using this IL6-DIO-KO mouse model, disease. combined with cell-specific Cre transgenic lines, would The histological analysis of the upper part of the be key to elucidate the specific functions of IL-6 from spinal cord (segments C1–T11) from MOG-IL6-DIO- each cellular source and study cell-specific IL-6 targeting KO mice revealed a decrease of demyelination and strategies in EAE and a wide range of autoimmune and limited invasion of CD3+ in the white matter, as pre- neuroinflammatory diseases. Indeed, our group is cur- viously observed using other IL-6 KO models rently analyzing the role of IL-6 in a mouse model of Sanchis et al. Journal of Neuroinflammation (2020) 17:304 Page 16 of 17 Alzheimer’s disease and in a mouse model of progressive Authors’ contributions encephalopathy resembling Leigh syndrome using this Conceptualization: JH and RDP; methodology: PS, OFG, GC, KA, AE, MG; formal analysis and investigation: PS, OFG and JH; writing - —original draft novel strategy. Furthermore, the IL6-DIO-KO mouse preparation: PS and OFG; writing - —review and editing: JH and RDP; model will be available for the scientific community and funding acquisition: JH. All authors read and approved the final manuscript. will be distributed worldwide following material transfer Funding agreements with Universitat Autònoma de Barcelona. Fi- PS, OFG, KA, and AE, acknowledge the support of BES-2015-071959, nally, we also encourage other investigators to take up FPU2012-00365, FPU17/02065 and La Marató de TV3 20142210, fellowships this reversible and conditional knock-out strategy to respectively. This study was supported by Ministerio de Economía y Competi- tividad y Fondo Europeo de Desarrollo Regional SAF2014-56546-R and carry out rescue experiments in other genetic models. RTI2018-101105-B-I00 to J.H. Availability of data and materials Supplementary information The datasets used and/or analysed during the current study are available Supplementary information accompanies this paper at https://doi.org/10. from the corresponding author on reasonable request. 1186/s12974-020-01969-0. Ethics approval and consent to participate All experiments were approved by the Ethics Committee on Animal Additional file 1. Probe sequence. Primers Fw:5’-CAGCATCTCATCTGAG Experiments of the Universitat Autònoma de Barcelona and the Generalitat TTCCG-3’ and Rv:5’-CTCACTGTTCACAAAGCACAGG-3’ were used to design de Catalunya (Refs. 3782 and 9684, respectively). a unique probe of 621 bp. Additional file 2. DNA sequencing of the modified Il6 gene of the IL6- Consent for publication DIO-KO mice. The resulting sequence had 1361 bp and the localization, Not applicable. size and color code of the genetic structures (exons and protein binding zones) detected in this sequence are explained in the figure legend. The Competing interests loxP and lox2272 cassettes (purple and green, respectively) are in the op- Not applicable. posite orientation, then sequence between them (including the exon 2) will be reverted after Cre action. Author details Additional file 3. Kaplan-Meier analysis in the EAE experiment with IL6- 1 Institute of Neurosciences and Department of Cellular Biology, Physiology DIO-KO and WT mice. Mice were assigned status “sick” on the first day and Immunology, Animal Physiology Unit, Faculty of Biosciences, Universitat their score was ≥1 and the proportion of sick mice was represented for Autònoma de Barcelona, 08193 Barcelona, Spain. 2Current affiliation: each day. Mice that never showed scores above the threshold were cen- Department of Pediatrics, Division of Molecular Genetics, Columbia University sored. The time-course of disease for IL6-DIO-KO mice is significantly de- Irving Medical Center, New York 10032, USA. 3Department of Biochemistry, layed, with only one mouse fulfilling the disease criterion (no median). Genome Sciences, and Howard Hughes Medical Institute, University of Table shows number of mice, sick mice, median day and 95% confidence Washington, Seattle, WA 98195, USA. interval. NA (not applicable) indicates impossibility to calculate value. Additional file 4. Kaplan-Meier analysis in the EAE experiment with IL6- Received: 22 July 2020 Accepted: 28 September 2020 DIO-ONCx3cr1, IL6-DIO-KO and IL6-DIO-Het mice. Mice were assigned sta- tus “sick” on the first day their score was ≥1 and the proportion of sick mice was represented for each day. The time-course of disease for IL6- References DIO-ONCx3cr1 and IL6-DIO-KO mice is significantly delayed. IL6-DIO-KO 1. Palmiter RD, Brinster R. Transgenic mice. Cell. 1985;41:343–5. again showed almost complete resistance to EAE, with only one mouse 2. Sternberg N, Hamilton D. Bacteriophage P1 site-specific recombination I. actually developing the disease (no median). Tables show number of recombination between loxP sites. J Mol Biol. 1981;150:467–86. mice, sick mice, median day and 95% confidence interval; as well as post- 3. Metzger D, Clifford J, Chiba H, Chambon P. Conditional site-specific hoc pair-wise comparisons. NA (not applicable) indicates impossibility to recombination in mammalian cells using a ligand-dependent chimeric Cre calculate value. recombinase. Proc Natl Acad Sci U S A. 1995;92:6991–5. 4. Atasoy D, Aponte Y, Su HH, Sternson SM. A FLEX switch targets Additional file 5. Quantification of IBA-1 and GFAP immunostaining Channelrhodopsin-2 to multiple cell types for imaging and long-range levels shown in Fig. 8. Quantification was carried out in both white and circuit mapping. J Neurosci. 2008;28:7025–30. gray matter of BSA-immunized IL6-DIO-Het and MOG35-55-immunized IL6- 5. Quintana A, Erta M, Ferrer B, Comes G, Giralt M, Hidalgo J. Astrocyte-specific DIO-Het, IL6-DIO-ONCx3cr1 and IL6-DIO-KO mice. All results were relativ- deficiency of interleukin-6 and its receptor reveal specific roles in survival, ized per total area and are represented as mean ± SEM; ●p ≤ 0.05 vs. body weight and behavior. Brain Behav Immun. 2013;27:162–73. BSA-IL6-DIO-Het mice; ★p ≤ 0.05 vs. IL6-DIO-Het mice. 6. Goldmann T, Wieghofer P, Müller PF, Wolf Y, Varol D, Yona S, et al. A new type of microglia gene targeting shows TAK1 to be pivotal in CNS autoimmune inflammation. Nat Neurosci. 2013;16:1618–26. Abbreviations 7. Yona S, Kim KW, Wolf Y, Mildner A, Varol D, Breker M, et al. Fate mapping AUC: Area under the curve; BSA: Bovine serum albumin; CD11b: Cluster of reveals origins and dynamics of monocytes and tissue macrophages under differentiation molecule 11b; CNS: Central nervous system; DIO: Double- homeostasis. Immunity. 2013;38:79–91. inverted open reading frame; EAE: Experimental autoimmune 8. Erta M, Quintana A, Hidalgo J. Interleukin-6, a major cytokine in the central encephalomyelitis; GFAP: Glial fibrillary acidic protein; GM: Gray matter; IBA- nervous system. Int J Biol Sci. 2012;8:1254–66. 1: Ionized calcium-binding adaptor molecule-1; IL-6: Interleukin-6; LFB: Luxol 9. Frei K, Fredrikson S, Fontana A, Link H. Interleukin-6 is elevated in plasma in Fast Blue; LOFT: LoxP-flippase (FLP) recognition target (FRT) Trap; multiple sclerosis. J Neuroimmunol. 1991;31:147–53. LPS: Lipopolysaccharide; MOG35-55: Myelin oligodendrocyte glycoprotein 35- 10. Maimone D, Guazzi GC, Annunziata P. IL-6 detection in multiple sclerosis 55 peptide; MS: Multiple sclerosis; TAM: Tamoxifen; WM: White matter; brain. J Neurol Sci. 1997;146:59–65. YFP: Yellow fluorescent protein 11. Eugster HP, Frei K, Kopf M, Lassmann H, Fontana A. IL-6-deficient mice resist myelin oligodendrocyte glycoprotein-induced autoimmune encephalomyelitis. Eur J Immunol. 1998;28:2178–87. Acknowledgements 12. Giralt M, Ramos R, Quintana A, Ferrer B, Erta M, Castro-Freire M, et al. We thank Profs. Marco Prinz and Steffan Jung for providing Cx3cr1-CreER Induction of atypical EAE mediated by transgenic production of IL-6 in mice. astrocytes in the absence of systemic IL-6. Glia. 2013;61:587–600. Sanchis et al. Journal of Neuroinflammation (2020) 17:304 Page 17 of 17 13. Okuda Y, Sakoda S, Bernard CCA, Fujimura H, Saeki Y, Kishimoto T, et al. IL- 38. Shimizu E, Tang Y-P, Rampon C, Tsien JZ. NMDA receptor-dependent 6-deficient mice are resistant to the induction of experimental autoimmune synaptic reinforcement as a crucial process for memory consolidation. encephalomyelitis provoked by myelin oligodendrocyte glycoprotein. Int Science. 2000;290:1170–4. Immunol. 1998;10:703–8. 39. Chaiyachati BH, Kaundal R, Zhao J, Wu J, Flavell R, Chi T. LoxP-FRT trap 14. Samoilova EB, Horton JL, Hilliard B, Liu TT, Chen Y. IL-6-deficient mice are (LOFT): a simple and flexible system for conventional and reversible gene resistant to experimental autoimmune encephalomyelitis: roles of IL-6 in the targeting. BMC Biol. 2012;10:96. activation and differentiation of autoreactiveT cells. J Immunol. 1998;161: 40. Kopf M, Baumannt H, Freer G, Freudenberg M, Lamers M, Kishimoto T, et al. 6480–6. Impaired immune and acute-phase responses in interleukin-6-deficient 15. Bettelli E, Carrier Y, Gao W, Korn T, Strom TB, Oukka M, et al. Reciprocal mice. Nature. 1994;368:339–42. developmental pathways for the generation of pathogenic effector TH17 41. Mendel I, Katz A, Kozak N, Ben-Nun A, Revel M. Interleukin-6 functions in and regulatory T cells. Nature. 2006;441:235–8. autoimmune encephalomyelitis: a study in gene-targeted mice. Eur J 16. Molnarfi N, Schulze-Topphoff U, Weber MS, Patarroyo JC, Prod’homme T, Immunol. 1998;28:1727–37. Varrin-Doyer M, et al. MHC class II–dependent B cell APC function is 42. Serada S, Fujimoto M, Mihara M, Koike N, Ohsugi Y, Nomura S, et al. IL-6 required for induction of CNS autoimmunity independent of myelin-specific blockade inhibits the induction of myelin antigen-specific Th17 cells and antibodies. J Exp Med. 2013;210:2921–37. Th1 cells in experimental autoimmune encephalomyelitis. Proc Natl Acad 17. Heink S, Yogev N, Garbers C, Herwerth M, Aly L, Gasperi C, et al. Trans- Sci U S A. 2008;105:9041–6. presentation of IL-6 by dendritic cells is required for the priming of 43. Brod SA, Bauer VL. Ingested (oral) tocilizumab inhibits EAE. Cytokine. 2014; pathogenic TH17 cells. Nat Immunol. 2017;18:74–85. 68:86–93. 18. Erta M, Giralt M, Jiménez S, Molinero A, Comes G, Hidalgo J. Astrocytic IL-6 44. Ponomarev ED, Shriver LP, Maresz K, Dittel BN. Microglial cell activation and influences the clinical symptoms of EAE in mice. Brain Sci. 2016;6:E15. proliferation precedes the onset of CNS autoimmunity. J Neurosci Res. 2005; 19. Sanchis P, Fernández-gayol O, Comes G, Escrig A, Giralt M, Palmiter RD, 81:374–89. et al. Interleukin-6 derived from the central nervous system may influence 45. Murphy ÁC, Lalor SJ, Lynch MA, Mills KHG. Infiltration of Th1 and Th17 cells the pathogenesis of experimental autoimmune encephalomyelitis in a cell- and activation of microglia in the CNS during the course of experimental dependent manner. Cells. 2020;9. autoimmune encephalomyelitis. Brain Behav Immun. 2010;24:641–51. 20. Gosselin D, Skola D, Coufal NG, Holtman IR, Schlachetzki JCM, Sajti E, et al. 46. Jung S, Aliberti J, Graemmel P, Sunshine MJ, Kreutzberg GW, Sher A, et al. An environment-dependent transcriptional network specifies human Analysis of Fractalkine receptor CX3CR1 function by targeted deletion and microglia identity. Science. 2017;356. green fluorescent protein reporter gene insertion. Mol Cell Biol. 2000;20: 21. Hanisch UK. Microglia as a source and target of cytokines. Glia. 2002;40:140–55. 4106–14. 22. Marik C, Felts PA, Bauer J, Lassmann H, Smith KJ. Lesion genesis in a subset 47. Geissmann F, Jung S, Littman DR. Blood monocytes consist of two principal of patients with multiple sclerosis: a role for innate immunity? Brain. 2007; subsets with distinct migratory properties. Immunity. 2003;19:71–82. 130:2800–15. 48. Valny M, Honsa P, Kirdajova D, Kamenik Z, Anderova M. Tamoxifen in the 23. Ajami B, Bennett JL, Krieger C, McNagny KM, Rossi FM. Infiltrating mouse brain: implications for fate-mapping studies using the tamoxifen- monocytes trigger EAE progression, but do not contribute to the resident inducible cre-loxP system. Front Cell Neurosci. 2016;10:1–12. microglia pool. Nat Neurosci. 2011;14:1142–50. 49. Ellisor D, Zervas M. Tamoxifen dose response and conditional cell marking: 24. Farley F, Soriano P, Steffen LS, Dymecki SM. Widespread recombinase is there control? Mol Cell Neurosci. 2010;45:132–8. expression using FLPeR (flipper) mice. Genesis. 2000;28:106–10. 50. Bebo BF, Dehghani B, Foster S, Kurniawan A, Lopez FJ, Sherman LS. 25. Sanz E, Yang L, Su T, Morris DR, McKnight GS, Amieux PS. Cell-type-specific Treatment with selective estrogen receptor modulators regulates myelin isolation of ribosome-associated mRNA from complex tissues. Proc Natl specific T-cells and suppresses experimental autoimmune Acad Sci U S A. 2009;106:13939–44. encephalomyelitis. Glia. 2009;57:777–90. 26. Sanchis P, Fernández-Gayol O, Vizueta J, Comes G, Canal C, Escrig A, et al. 51. Ajami B, Bennett JL, Krieger C, Tetzlaff W, Rossi FMV. Local self-renewal can Microglial cell-derived interleukin-6 influences behavior and inflammatory sustain CNS microglia maintenance and function throughout adult life. Nat response in the brain following traumatic brain injury. Glia. 2020;68:999–1016. Neurosci. 2007;10:1538–43. 27. Schneider CA, Rasband WS, Eliceiri KW. NIH image of ImageJ: 25 years of 52. Ginhoux F, Greter M, Leboeuf M, Nandi S, See P, Gokhan S, et al. Fate Umage analysis. Nat Methods. 2012;9:671–5. mapping analysis reveals that adult microglia derive from primitive 28. Ruifrok AC, Johnston DA. Quantification of histochemical staining by color macrophages. Science. 2010;701:841–5. deconvolution. Anal Quant Cytol Histol. 2001;23:291–9. 53. Korn T, Bettelli E, Gao W, Awasthi A, Jäger A, Strom TB, et al. IL-21 initiates an alternative pathway to induce proinflammatory T H17 cells. Nature. 2007; 29. Cardin JA, Carlén M, Meletis K, Knoblich U, Zhang F, Deisseroth K, et al. 448:484–8. Driving fast-spiking cells induces gamma rhythm and controls sensory 54. Quintana A, Muller M, Frausto RF, Ramos R, Getts DR, Sanz E, et al. Site- responses. Nature. 2009;459:663–7. specific production of IL-6 in the central nervous system retargets and 30. Sohal VS, Zhang F, Yizhar O, Deisseroth K. Parvalbumin neurons and gamma enhances the inflammatory response in experimental autoimmune rhythms enhance cortical circuit performance. Nature. 2009;459:698–702. encephalomyelitis. J Immunol. 2009;183:2079–88. 31. Tsai H-C, Zhang F, Adamantidis A, Stuber GD, Bonci A, de Lecea L, et al. Phasic firing in dopaminergic neurons is sufficient for behavioral conditioning. Science. 2009;324:1080–4. Publisher’s Note 32. Turan S, Bode J. Site-specific recombinases: from tag-and-target- to tag-and- Springer Nature remains neutral with regard to jurisdictional claims in exchange-based genomic modifications. FASEB J. 2011;25:4088–107. published maps and institutional affiliations. 33. Samokhvalov IM, Thomson AM, Lalancette C, Liakhovitskaia A, Ure J, Medvinsky A. Multifunctional reversible knockout/reporter system enabling fully functional reconstitution of the AML1/Runx1 locus and rescue of hematopoiesis. Genesis. 2006;44:115–21. 34. Guy J, Gan J, Selfridge J, Cobb S, Bird A. Reversal of neurological defects in a mouse model of Rett syndrome. Science. 2007;315:1143–7. 35. Ventura A, Kirsch DG, McLaughlin ME, Tuveson DA, Grimm J, Lintault L, et al. Restoration of p53 function leads to tumour regression in vivo. Nature. 2007;445:661–5. 36. Zhou X, Vink M, Klaver B, Berkhout B, Das AT. Optimization of the Tet-on system for regulated gene expression through viral evolution. Gene Ther. 2006;13:1382–90. 37. Heyer J, Kwong LN, Lowe SW, Chin L. Non-germline genetically engineered mouse models for translational cancer research. Nat Rev Cancer. 2010;10: 470–80.