C57BL/6JRj mice were obtained from Janvier (Le Genest Saint Isle, France) and housed at 22 ± 2°C under a 12 h light/dark cycle with ad libitum access to food and water. All animal procedures were conducted in accordance with the German federal animal welfare law, local ethical guidelines of the University Freiburg and have been approved by the animal experimentation committee of the University of Freiburg as well as the Regierungspräsidium Freiburg (G-13/57 [Tgfbr2-MG-KO], X-15/01A [primary microglia]).

The generation of Cx3cr1^CreERT2:R26-YFP:Tgfbr2^fl/fl mice has been described recently (10). Briefly, mice carrying loxP-site-flanked alleles of Tgfbr2 were crossed to the Cx3cr1^CreERT2 mouse line (16). Furthermore, the reporter mouse line B6.129 × 1-Gt(ROSA)26Sortm1(EYFP)Cos/J (17) was introduced to obtain Cx3cr1^CreERT2:R26-YFP:Tgfbr2^fl/fl mice. Cre recombinase activity was induced in vivo by treatment of 6- to 8-week-old mice with 8 mg tamoxifen (TAM, T5648, Sigma-Aldrich, Germany) solved in 200 µl corn oil (C8267, Sigma) injected intraperitoneally (two time points 48 h apart). Littermates carrying the respective loxP-flanked alleles but lacking expression of Cre recombinase (+/+TAM) or not receiving tamoxifen (cre/+OIL) were used as controls. In vitro recombination was achieved using OH-TAM (H7904, Sigma-Aldrich, Germany) at a final concentration of 1 µM to 25 cm² glia culture flasks (single brain cultures) at least 3 days before harvesting microglia. Ethanol (EtOH) was used as a solvent control for in vitro experiments.

Primary microglia cultures were generated as described by Spittau et al. (12). Vessels and meninges were removed from brains of P0/P1 C57BL/6JRj mice (Janvier) and brains were washed and collected in ice-cold Hank’s Buffered Salt Solution (Gibco, Germany). After enzymatic dissociation with Trypsin-EDTA (Gibco, Germany) for 15 min at 37°C, an equal volume of fetal calf serum (FCS, Gibco, Germany) and DNase (Roche, Mannheim, Germany) at a final concentration of 0.05 mg/ml was added. Cells were dissociated using wide- and narrow-bored polished Pasteur pipettes and further centrifuged and resuspended in DMEM/F12 medium (Gibco, Germany) containing 10% FCS and 1% penicillin/streptomycin (Invitrogen). Dissociated cells from two to three brains were plated on poly-d-lysine-coated (Sigma-Aldrich, Schnelldorf, Germany) 75 cm² culture flasks. Cells were kept in a 5% CO₂/95% humidified atmosphere at 37°C. After 10–14 days in culture, microglia were shaken off (250–300 rpm for 1 h) from adherent astrocytes and plated according to the experimental designs. Treatment with recombinant human TGFβ1 (Peprotech, Hamburg, Germany) was performed at a concentration of 5 ng/ml. For inhibition of microglial TGFβ signaling, a TGFβ receptor type I inhibitor (#616454, Calbiochem, Merck, Darmstadt, Germany) at a final concentration of 500 mM was used. For inhibition of protein biosynthesis, cycloheximide (CHX) (C7698, Sigma-Aldrich, MO, USA) was used at a concentration of 50 µg/ml.

BV2 cells were cultured as recently described by Zhou et al. (18). Briefly, cells were maintained in DMEM/F12 culture medium (Thermo Fisher Scientific) supplemented with 10% FCS and 1% penicillin/streptomycin (Invitrogen) and kept in at 37°C in 5% CO₂/95% humidified atmosphere. Treatments with 5 ng/ml TGFβ1 (Peprotech, Hamburg, Germany) for ChIP experiments were performed under serum-free conditions.

RNA was isolated from primary microglia using TRIzol reagent (Invitrogen, Karlsruhe, Germany) according to the manufacturer’s instructions. RNA concentration and quality were determined using the NanoDrop 2000 (Thermo Scientific, Germany). 1 µg total RNA from each sample was reverse transcribed to cDNA using Protoscript^® II First Strand cDNA Synthesis Kit (#E6560S, New England Biolabs, Frankfurt, Germany) according to the manufacturer’s instructions.

Quantitative RT-PCR was performed using the CFX Connect™ System (Bio-Rad, München, Germany) in combination with the SYBR Green GoTaq^® qPCR Kit (A6002, Promega, Madison, WI, USA). 5 µl of cDNA template was used in 20 µl reaction mixture. Results were analyzed using the CFX Connect™ System (Bio-Rad, München, Germany) Software and the comparative CT method. All data are expressed as 2^−ΔΔCT for the gene of interest normalized to the housekeeping gene Gapdh and presented as fold change relative to controls. The following primers have been used throughout this study: Olfml3 for 5′-CACCTTGTGGAGTACATGGAAC-3′, Olfml3 rev 5′-CTACCTCCCTTTCAAGACGGT-3′ [NM_133859.2], Gapdh for 5′-GGCATTGCTCTCAATGACAA-3′, Gapdh rev 5′-ATGTAGGCCATGAGGTCCAC-3′ [NM_001289726], Olfml3-SBE1 for 5′-TGACAGCTCTAACAGGGCCTA-3′, Olfml3-SBE1 rev 5′-ACTCTGACCCCTTGAAAAGGC-3′ [Chr. 3; 103739520-103739508], Olfml3-SBE2 for 5′-CCCATCTCTGGTGTCTCTCAC-3′, Olfml3-SBE2 rev 5′-TAGTTAAGGCTTCTGGCGACT-3′ [Chr. 3; 103738043-103738055].

Chromatin immunoprecipitation was performed as recently described (10). The enzymatic ChIP Kit (#9003, Cell Signaling Technology, Inc., Danvers, MA, USA) was used according to the manufacturer’s instructions. After plating of 9 × 10⁶ BV2 cells in 6 × 75 cm² cell culture flasks in DMEM/F12 medium containing 10% FCS and 1% penicillin/streptomycin, BV2 cells were kept 2 h in serum-free medium. Cells were incubated in serum-free medium containing 5 ng/ml TGFβ1. Proteins were cross-linked to the DNA by adding 1% formaldehyde to the cells for 10 min at room temperature (RT). The cells were harvested by scraping them into PBS + Protease inhibitor cocktail. After nuclei preparation with Buffer A and B from the kit, the nuclei from each treatment were separated into four IP samples per treatment. Digestion of chromatin was initiated by adding 0.25 µl Micrococcal Nuclease (#10011, Cell Signaling) for 20 min at 37°C and the nuclei were lysed with three sets of 20 s pulses using a Bioruptor™ (Diagenode, Liège, Belgium) sonicator. Chromatin concentration of every sample was measured using NanoDrop 2000/2000c (Thermo Fisher Scientific, Waltham, MA, USA) and 10 µg of digested, cross-linked chromatin of every sample was used for further steps. A 2% input control was taken aside before starting the IP. A Histone H3 antibody (Cell Signaling, #4620, 10 µl) as a positive control, normal rabbit IgG (Cell Signaling, #2729, 2 µl) as a negative control and Smad2 (Cell Signaling, #5339, 1:100) and Smad4 (Cell Signaling, #38454, 1:100) antibodies were used for IP. The Chromatin was incubated overnight at 4°C. After incubating the chromatin with protein G magnetic beads for 2 h and washing the chromatin in magnetic separation racks (Cell Signaling, #7017), elution of the chromatin was performed. The protein cross-link was reversed by using 2 µl Proteinase K (Cell Signaling, #10012) for 2 h at 65°C. Finally, DNA purification in spin columns was performed and the amount of DNA was quantified by using qPCR. Data are expressed as 2^−ΔCT for the Smad-binding element (SBE) of interest normalized to the rabbit IgG control. PCR products were visualized using agarose gel electrophoresis and staining with GelRed (Genaxxon Bioscience). Images were captured using a Biometra (Göttingen, Germany) gel documentation station.

Anesthetized 6-month-old C57BL/6 mice were transcardially perfused using PBS followed by 4% paraformaldehyde (PFA). Afterward, brains were extracted and postfixed in 4% PFA overnight. Free-floating 50 µm thick vibratome (Leica, Wetzlar, Germany) sections were stained overnight with anti-Iba1 (1:500, Wako Chemicals, Japan) and anti-Olfml3 (1:100, sc-243668, Santa Cruz Biotechnology Inc.). Alexa Fluor-488, as well as Alexa Fluor-568-conjugated secondary antibodies (1:200, Cell Signaling Technology) was used for 2 h at RT. Finally, nuclei were stained using 4′6-diamidino-2-phenylindole (Dapi, Roche) for 5 min and sections were mounted on objective slides and covered with Fluoromount G mounting medium. Immunocytochemistry was performed using PFA-fixed primary microglia on glass coverslips as mentioned above. FITC-coupled tomatolectin (Sigma-Aldrich) was used to label microglia and anti-Olfml3 (Santa Cruz Biotechnology Inc.) to detect Olfml3 expression. Localization of Olfml3 in the endoplasmic reticulum (ER) and the Golgi apparatus was confirmed using the Organelle Localization IF Antibody Sampler Kit (Cell Signaling technologies, #8653). Fluorescence-coupled secondary antibodies (1:200, Cell Signaling Technologies) were used for 1 h at RT. Coverslips were washed three times with PBS for 3 min each and nuclei were counterstained using DAPI (Roche). After a final washing (3×), coverslips were mounted on objective slides using Fluoromount G mounting medium (SouthernBiotech). Fluorescence images were captured using the Leica TCS SP8 confocal laser scanning microscope (Leica, Wetzlar, Germany) and the LAS AF image analysis software.

Data are given as mean ± SEM. Statistical differences between two groups were determined using Student’s t-test. Multiple-group analysis was performed using one-way ANOVA followed by Bonferroni’s multiple comparison post-test. P-values ≤0.05 were considered as statistically significant. All statistical analyses were performed using the GraphPad Prism6 software (GraphPad Software Inc., La Jolla, CA, USA).

The Olfml3 gene is located on chromosome 3 and consists of three exons. Figure 1A shows that a 180 bp Exon 1 is separated from Exon 2 (286 bp) by Intron 1, which has a total length of 584 bp. A 288 bp Intron 2 links Exon2 with Exon 3, the latter of which has a length of 4,999 bp and includes a long 5′UTR. The protein structure of Olfml3 as depicted in Figure 1A, demonstrates the presence of 21 aa signal peptide followed by a coiled-coil domain. The central motif is the olfactomedin-like domain which harbors two N-linked glycosylation sites. We further used immunocytochemistry to detect the intracellular localization of Olfml3 in primary microglia. As shown in Figure 1B, Olfml3 immunoreactivity was found in the perinuclear cytoplasm and in microglial processes. FITC-coupled tomatolectin was used to label the cell membrane of primary microglia. Figure 1C displays the presence of Olfml3 speckles in microglial processes suggesting Olfml3 immunoreactivity in exocytotic vesicles. The perinuclear staining pattern of Olfml3 suggests an ER localization (Figure 1D). To confirm this subcellular localization, protein disulfide isomerase (PDI) and receptor binding cancer antigen expressed on SiSo cells (RCAS1) were used as markers for the ER and the Golgi apparatus, respectively. As shown in Figure 1E, Olfml3 colocalized with PDI indicating presence in the ER. Moreover, Olfml3 further displayed accumulation in the Golgi apparatus as evidenced by colocalization with RCAS1 (Figure 1F). Together, these data demonstrate that primary microglia show robust expression of the secreted glycoprotein Olfml3.

It has been described that Olfml3 belongs to a set of genes, which have been demonstrated to be microglia-specific and, which are not expressed by other macrophage populations (5, 6, 8, 9). Since these data were obtained from RNAseq and/or gene expression studies, we used immunohistochemistry to determine Olfml3 expression in cortical microglia of adult mice. As shown in Figure 2, Olfml3 immunoreactivity could be detected in the molecular layer and the external granule cell layer of the frontal cortex (Figure 2A). Iba1⁺ cortical microglia (Figure 2B) were also positive for Olfml3 (Figure 2C). High magnification images reveal that strong Olfml3 immunoreactivity was detectable in the cytoplasm of Iba1⁺ microglia (Figures 2D–F). Interestingly, microglial processes only displayed weak Olfml3 signals and Olfml3 speckles could be found in the periphery of microglia suggesting an extracellular localization and accumulation (Figure 2F). Iba1⁺ meningeal macrophages can be found in close proximity to the pial surface and, thus, Olfml3 immunoreactivty in this non-microglial cell was analyzed. As depicted in Figures 2G–I, Iba1⁺ meningeal macrophages (white arrows) showed no or very faint Olfml3 immunoreactivity whereas Iba1⁺ cortical microglia (white asterisks) show strong Olfml3 expression. These data clearly show that Olfml3 is predominantly expressed by Iba1⁺ microglia and further suggest that Olfml3 is secreted and accumulated in the extracellular space.

It has recently been described that TGFβ1 is necessary to induce the expression of microglia-specific genes such as Olfml3 (11). Moreover, we have demonstrated that microglial TGFβ signaling at P7 precedes the upregulation of microglia-specific genes in vivo (10). To address whether the Olfml3 expression in microglia is dependent on TGFβ signaling, we used microglia from Cx3cr1^CreERT2:Tgfbr2^fl/fl mice. As displayed in Figure 3A, 2-month-old Cx3cr1^CreERT2:Tgfbr2^fl/fl mice were used to generate the tamoxifen-induced microglia-specific deletion of ligand-binding receptor Tgfbr2 in vivo. Four weeks after recombination, microglial RNA was isolated and used to detect Olfml3 expression. Figure 3B shows that silencing of TGFβ signaling in adult microglia did not impair Olfml3 expression. Moreover, normal Olfml3 expression was observed in cortical Iba1⁺ microglia after deletion of Tgfbr2 in vivo (Figure 3D) when compared with wild-type microglia (Figure 3C). Using postnatal microglia cultures from Cx3cr1^CreERT2:Tgfbr2^fl/fl mice and tamoxifen-induced recombination in vitro (Figure 3E), we observed that Olfml3 expression was significantly reduced in Tgfbr2-deficient microglia (Figure 3F). In summary, the presented data indicate that TGFβ signaling is essential to induce Olfml3 expression in immature postnatal microglia but is dispensable to maintain Olfml3 expression in mature adult microglia. The results further suggest that Olfml3 might be a direct TGFβ1 target gene.

To address whether Olfml3 is direct TGFβ1 target gene, primary microglia were treated with recombinant human TGFβ1 (5 ng/ml) for 2, 6, and 24 h. As depicted in Figure 4A, significant upregulation of Olfml3 expression was detected after 2 and 6 h. Although Olfml3 expression was increased after 24 h of TGFβ1 treatment, no significant changes were detectable. Next, immunocytochemistry was employed to monitor increases in Olfml3 expression in primary microglia. Whereas a distinct but weak immunoreactivity for Olfml3 could be observed in untreated microglia (Figure 4B) strong increases in Olfml3 signals were detectable after treatment of primary microglia with TGFβ1 for 24 h (Figure 4C). Using CHX-mediated inhibition of protein synthesis revealed that TGFβ1-induced Olfml3 expression was independent on de novo synthesis of proteins (Figure 4D) indicating a direct upregulation of Olfml3 transcription by activated TGFβ signaling. In silico analysis of the Olfml3 promoter region revealed the presence of two putative SBEs upstream of the transcriptional start site. Both SBEs contain the typical palindromic CAGAC DNA-binding sequence for Smads (Figure 4E). BV2 cells were treated with TGFβ1 for 2 h and ChIP was performed using Smad2- as well as Smad4-specific antibodies. Figures 4F,G show that Smad2 interacted with both putative SBEs whereas Smad4 did not. Precipitation of Histone3 was used as the positive control for all performed ChIP experiments. These data clearly demonstrate that Olfml3 is a direct TGFβ1/Smad2 target in microglia and further show that Smad4 is not interacting with SBEs of the Olfml3 promoter.