Nogo receptor is involved in the adhesion of dendritic cells to myelin

Whole human blood was obtained by venous puncture into EDTA tubes with informed, written consent from 9 healthy donors with approval from the local institutional review board of Innsbruck Medical University. Myeloid DCs were generated according to established standard procedures [25, 26]. Firstly, peripheral blood mononuclear cells (PBMCs) were isolated from the blood by density gradient centrifugation using Ficoll™-based lymphocyte separation medium (PAA, Pasching, Austria). PBMCs were washed with 0.9% saline solution (Fresenius Kabi, Graz, Austria) and seeded at a density of 3.3 × 10⁶ cells/ml in a 6-well plate in serum-free medium (Lonza x-vivo chemically-defined medium, Cologne, Germany). After two hours of incubation at 37°C, with 5% CO₂, monocytes selectively adhered to the cell culture-treated plastic. At this stage, all non-adherent cell populations were washed away by rinsing three times with RPMI1640 medium (Gibco, Invitrogen, Carlsbad, CA, USA). After the washing steps, adherent monocytes were cultured for 8 days in serum-free medium supplemented with 1% penicillin streptomycin (PenStrep, Invitrogen, Carlsbad, CA, USA), 800 U/ml granulocyte/monocyte colony stimulating factor (GM-CSF, Novartis, Leukomax, Basel, Switzerland) and 40 ng/ml interleukin-4 (human recombinant IL-4, Invitrogen). Every two days, cells were fed with fresh medium, PenStrep, GM-CSF and IL-4. By day 6, the human monocytes had differentiated into loosely adherent immature dendritic cells (immDCs). Addition of a defined maturation cocktail for the last two days of culture resulted in generation of mature DCs (matDCs). Maturation cocktail (MC) consisted of interleukin 1β (2 ng/ml, Invitrogen), IL-6 (10 ng/ml, Invitrogen), tumour necrosis factor-α (TNF-α, 10 ng/ml, Invitrogen) and prostaglandin E2 (PGE2, 1 μg/ml, Sigma-Aldrich, St. Louis, MO, USA). On day 8 of culture, immature and mature DCs were harvested for flow cytometric analysis, RNA extraction, and adhesion assay.

T cells were isolated from human PBMCs using a commercially available magnetic cell separator T cell depletion kit (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany) to produce ex vivo T cells (Tex). Cells were cultured in serum-free medium in the presence of anti-CD3 antibody (T cont), and in the presence or absence of T cell activator phytohaemagglutinin (T PHA+ and T PHA-, respectively) for 2 days. Epstein Barr virus-transformed B lymphocytes were used as a B cell line (BCL). Monocytes were isolated from PBMCs by magnetic cell separator monocyte depletion kit (Miltenyi Biotec GmbH) to produce ex vivo monocytes. They were then maintained in serum-free medium for 7 days, and given either interferon gamma (IFN-γ, 100 ng/ml, Invitrogen) or lipopolysaccharide (LPS, 100 ng/ml, Sigma-Aldrich) for the last two days of culture.

Wild type male C57BL/6J mice were obtained from Jackson Laboratories and housed in the animal house of Innsbruck Medical University. Ngr1/2 double knockout mice (Ngr1/2-/-) were generated by crossing Ngr1-/- mice [27] with Ngr2-/- mice as previously described [10]. Bone marrow derived myeloid DCs were prepared according to established standard procedures as described by Lutz et al. [28]. Mice were sacrificed by cervical dislocation and the tibiae and femurs were removed. The bones were cleaned of all muscle tissue and sterilised with 70% ethanol. The bone marrow was flushed out with cold RPMI1640 containing 10% foetal calf serum (FCS, Sigma-Aldrich) and β-mercaptoethanol (β-ME, 50 μM, Sigma-Aldrich). The marrow was separated into a single cell suspension by repeated pipetting and passed through a nylon mesh to remove bone and debris. Contaminating erythrocytes were removed by lysis on ice using erythrocyte lysis buffer (containing 0.15 M ammonium chloride, 10 mM potassium bicarbonate, 0.1 mM EDTA (all from Roth, Karlsruhe, Germany), with pH adjusted to 7.0-7.2) and cells were counted. 20 × 10⁶ bone marrow precursor cells were seeded in RPMI containing 10% FCS, 50 μM β-ME, and 20 ng/ml GM-CSF (ImmunoTools, Friesoythe, Germany) in 75 cm³ flasks. After two days, flasks were gently swirled and 75% of medium was removed. The same volume of fresh medium was added back, containing 40 ng/ml GM-CSF. On day 4, the culture is made up of firmly attached stromal cells covered in clusters of loosely attached DCs, and non-adherent granulocytes. The granulocytes were washed away and DCs were subcultured at a concentration of 1 × 10⁶ cells/ml/well in a 24-well plate, with 20 ng/ml GM-CSF. On day 6, cells were fed by removal of 75% of the medium and adding back the same volume containing GM-CSF. For the generation of mature DCs, maturation cocktail containing a final concentration of 2 ng/ml IL-1β (Invitrogen), 10 ng/ml IL-6 (ImmunoTools), 10 ng/ml TNF-α (ImmunoTools), and 1 μg/ml PGE-2 (Sigma-Aldrich) was added for the last two days of culture. Cells were harvested for flow cytometric analysis, RNA extraction, and adhesion assay on day 8.

Cells (a minimum of 10⁶) were washed with PBS and homogenised in 1 ml TRIzol reagent (Invitrogen). RNA was extracted as per the manufacturer's protocol, dissolved in diethylpyrocarbonate (DEPC)-treated water and the concentration was determined using a spectrophotometer (NanoDrop 1000, peqlab, Polling, Austria). 1 μg of RNA from each sample was reverse transcribed with the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Carlsbad, CA, USA) using random primers. The protocol was followed as described in the kit. cDNA was diluted 1:4 and immediately used for RT qPCR. Levels of Nogo receptor component mRNAs in human and mouse DCs were determined using TaqMan RT qPCR Assays (Applied Biosystems). Assays used for human mRNA detection were: NgR1 (Hs00368533_m1), NgR2 (Hs00604888_m1), LINGO-1 (Hs01072978_m1), TROY (Hs00218634_m1) and p75^NTR (Hs00609977_m1). Assays used for mouse mRNA detection were as follows: NgR1 (Mm00452228_m1), NgR2 (Mm01336368_g1), LINGO-1 (Mm01173306_m1), TROY (Mm00443506_m1), and p75^NTR (Mm00446296_m1). 18 s rRNA was measured in each sample as an endogenous control in order to control for varying cDNA concentrations and human or mouse brain cDNA were used as a positive control for all assays (commercially available human foetal brain RNA was used from Clontech Laboratories, Inc., Mountain View, CA, USA). Assays were performed as described by the manufacturer, with a final assay volume of 25 μl. Experiments were performed in duplicate wells and all assays were first screened for detection of genomic DNA. Data were collected using the 7300 Real-Time PCR System (Applied Biosystems) and analysed by the comparative Ct method, where; ΔCt = target Ct - endogenous Ct; and ΔΔCt = ΔCt_matDC - ΔCt_immDC; relative mRNA expression = 2^-ΔΔCt. Immature DCs were assigned as the calibrator for all relative quantifications, except where otherwise stated.

Human brain and human immature and mature DCs were lysed in buffer containing 150 mM NaCl, 1% Triton X-100, 10% glycerol, 50 mM Hepes pH 7.40, and protease inhibitor cocktail (Roche Applied Sciences, Mannheim, Germany). Protein concentration of all lysates was determined with the bicinchoninic acid protein assay (BCA, Sigma-Aldrich). 22 μg of human brain, and 70 μg each of immDC and matDC protein were denatured and loaded onto a 10% Bis/Tris gel (Invitrogen). After separation, protein was blotted onto a Hybond membrane (Amersham, GE Healthcare, Buckinghamshire, UK) and probed with NgR11-A antibody (diluted 1:3000, Alpha Diagnostic, San Antonio, TX, USA). Detection was performed with horseradish peroxidase-conjugated secondary antibodies and enhanced chemiluminescence detection on film (Amersham, GE Healthcare). To confirm antibody specificity, NgR11-A was first blocked with the immunising peptide (NgR11-P, Alpha Diagnostic) at 5 times the weight of NgR11-A used. In order to probe the same membranes for actin, they were stripped at 60°C with buffer containing 2% SDS, 100 mM beta-mercaptoethanol and 62.5 mM Tris-HCl, pH 6.8, then washed, blocked and incubated with anti-actin monoclonal antibody (diluted 1:20000, BD Biosciences).

Myelin was isolated from central nervous system tissue by the density gradient centrifugation method, as described previously [29]. Briefly, a segment of human brain, or whole mouse CNS was shock frozen in liquid nitrogen and stored at -80°C until needed. Tissue was thawed on ice and cut into smaller pieces and homogenised in a 0.32 M sucrose solution. The homogenized tissue was washed three times in 0.32 M sucrose before being layered over a 0.85 M sucrose solution. After centrifugation at 26000 × g for 60 minutes at 4°C, myelin was contained in the interphase between the high and low sucrose solutions. The myelin was subjected to osmotic shock by stirring with distilled water for 30 minutes at 4°C, before being washed and ultracentrifuged again with 0.32 M sucrose. Finally, the myelin was washed three times with, and resuspended in distilled water. The protein concentration of the myelin extract was determined with BCA protein assay.

Human myelin derived MBP was purified from normal human brain according to the procedure of Eylar et al. [30]. SDS-PAGE and Western blot with a monoclonal antibody to MBP 130-137 (Millipore GmbH, Vienna, Austria) was used to confirm the purity of the MBP preparation.

A DNA fragment encoding the mouse Nogo-66 loop was amplified from the mouse Nogo-A clone IRAVp968A04133D (ImaGene, Berlin, Germany) with the primers mM_RTN4-66-s (5'-CTA CCA TGG GCA GGA TAT ATA AGG GTG TGA TCC-3') and mM_RTN4-66-as (5'-GCT TGC GGC ACC CTT CAG GGA ATC AAC TAA ATC-3'). The fragment was digested with NcoI and ligated into pET28a(+) vector using the NcoI and a blunted NotI site. The sequence was verified by sequencing at LGC Genomics (Berlin).

E. coli Rosetta were transformed with Nogo-66 pET28 and induced in 1 litre of LB culture medium with 1 mM IPTG at an OD600 of 0.6 and a temperature of 30°C. After 3 hours the bacteria were harvested by centrifugation at 4000 × g for 5 minutes. The pellet was resuspended in PBS substituted with 1% Triton X-100 and protease inhibitors and sonicated. After 30 minutes of incubation on ice the lysate was centrifuged and the pellet dissolved in 8 M urea. Recombinant Nogo-66-His was purified with TALON^®Cobalt resin (Clontech Laboratories) under denaturing conditions and eluted with elution buffer (50 mM Hepes pH 4, 300 mM imidazol, 150 mM NaCl). Finally, the eluate was dialysed against DMEM adjusted to pH 4. Protein concentration was determined by BCA assay (Thermo Scientific, Rockford, IL, USA).

MAG-Fc was produced as described previously [31]. Briefly, conditioned medium of transiently transfected CHO-K1 cells was harvested and recombinant protein was purified using Protein A/G Agarose (Thermo Scientific, Rockford, IL, USA). Purity and concentration were confirmed by comparing band intensity on SDS-PAGE to BSA standard.

Adhesion assays for human DCs were conducted in 96-well plates. All cell types and conditions were assayed in triplicate. The following substrates were all used at concentrations of 100 and 10 μg/ml for adhesion assay: human myelin, human MBP, His-tagged mouse Nogo-66 (Nogo-66-His, amino acids 1025-1090). 50 μl of each substrate was added to the 96-well plate and incubated for 4 hours at 37°C. MAG-Fc was added at 10 μg/ml and was coated on human IgG to aid clustering of the protein. First, 15 μg/ml human IgG (Sigma-Aldrich) was added to wells in 50 mM bicarbonate buffer (pH 9) and incubated over night at 4°C. The next day, wells were washed and MAG-Fc was added along with the other adhesion substrates to their respective wells and the plate was incubated for 4 hours at 37°C, 5% CO₂. Excess substrate was removed and all wells were washed once with medium before addition of cells. Human monocyte-derived DCs were prepared as described above and harvested on day 8. Cells were collected, counted and plated at 200000 cells/ml in the 96-well plate in serum-free medium. Cells were allowed to adhere for 30 minutes at 37°C, 5% CO₂. Non-adherent cells were gently removed and wells were washed three times with medium. In order to detect and count CD11b⁺ DCs, cells were fixed and fluorescently stained as follows. Cells were fixed with 4% paraformaldehyde (PFA) for 30 minutes at room temperature. After the PFA was washed away, nonspecific antibody binding was blocked by addition of 20 μg/ml human IgG in PBS containing 5% normal goat serum (NGS, Invitrogen) and 1% bovine serum albumin (BSA, Sigma-Aldrich) for 1 hour at room temperature. In order to detect adherent DCs, CD11b antibody (BD Biosciences) diluted 1:100 in 1% NGS, 1% BSA was added to the cells and incubated over night at 4°C with gentle shaking. The following day, the cells were washed three times with PBS and visualised at 10 × magnification with a fluorescent microscope (Leica Microsystems, Cambridge, UK). Four digital photos were taken per well and cells were counted using the particle analysis tool from ImageJ [32].

Adhesion of mouse DCs was assayed in 96-well plates and based on the method described by Kueng et al [33]. 10 μg/ml mouse myelin was added to the respective wells and incubated for 4 hours at 37°C. Myelin was removed and wells were washed once with medium before addition of cells. Mouse bone marrow-derived DCs were prepared as described above and harvested on day 8. Cells were collected, counted and plated at 200000 cells/ml in triplicate in the 96-well plate in RPMI 1640. Cells were allowed to adhere for 30 minutes at 37°C, 5% CO₂. Non-adherent cells were gently removed and wells were washed three times with medium. In order to quantify the number of remaining adherent cells, they were stained with cresyl violet and absorbance was measured as follows. Cells were fixed with 4% PFA for 30 minutes at room temperature. Cells were stained with 0.04% cresyl violet (Sigma-Aldrich) in 20% methanol for 30 minutes. The dye was then extracted with 0.1 M citric acid in 50% ethanol for 30 minutes on a rotating shaker. Absorbance of each well was measured at 570 nm using the DTX 880 Multimode Detector with Multimode Analysis Software (Beckman Coulter, Krefeld, Germany).

All statistical analyses were carried out using GraphPad Prism 5 software (GraphPad Software Inc., San Diego, CA, USA). For flow cytometry data, an unpaired, two-tailed student's t test was used to compare the means of each marker in immature versus mature DCs. Microarray data was analysed using TIGR MeV_4_5 (Multiple Experiment Viewer), a Java tool for genomic data analysis http://​mev-tm4.​sourceforge.​net [34] which measures significance of microarray (SAM). Multi-class SAM was used to identify significant cytokines based on differential expression between the four groups at a false discovery rate (FDR, proportion of genes likely to have been identified by chance as being significant) of 0%. To determine significance of RT qPCR data, ΔCt values were compared using the Wilcoxon matched-pairs signed rank test, as per Yuan et al. [35]. Human DC adhesion to myelin was measured using a two-way repeated measure ANOVA. The association of human RT qPCR ΔCt values with adhesion was calculated using Spearman correlation. Mouse immature WT vs. Ngr1/2-/- and mature WT vs. Ngr1/2-/- adhesion to myelin were analysed using the Wilcoxon matched-pairs signed rank test.

As our aim was to expand on current knowledge of the role of NgRs in non-CNS cells, we began the study as a screen for NgR1 expression in human peripheral immune cells. Expression of NgR1 mRNA was measured in a panel of human immune cells (un-stimulated and stimulated T cells, B cell line, monocytes, immature and mature DCs) using TaqMan real time quantitative PCR (RT qPCR). Expression of NgR1 mRNA was five times higher in immature DCs compared to all other immune cells tested (Figure 1A). We thus concentrated on further examining NgR expression in DCs. Human monocyte derived myeloid DCs were generated as described and first characterised by flow cytometry (Figure 1B). More than 95% of the cells expressed CD11b, indicating high purity of monocyte-derived DCs. The DC phenotype was confirmed by high CD11 c expression (Figure 1B). Only 2.8 ± 0.6% of untreated DCs expressed CD83. The percentage of cells expressing CD83 increased significantly to 90.0 ± 2.2% after addition of maturation cocktail, indicating successful DC maturation. Significant increases in expression of human leukocyte antigen (HLA), CD86, CCR7 and a decrease in CD1a provide further evidence of successful maturation (Figure 1B). Having verified the in vitro generation of myeloid DCs, we went on to examine the regulation of NgR expression in these cells.

NgR1 expression was increased in comparison to the monocytes from which the immDCs were generated (Figure 1A). It is then down-regulated upon maturation. The increased transcription was confirmed by higher protein expression of NgR1 in human immature DCs, as determined by western blot (Figure 1C). This regulation of NgR1 expression between immature and mature DCs prompted us to also measure the expression of NgR1's co-receptors, as well as of NgR2. NgR2 mRNA is down-regulated in the same manner upon maturation of human DCs (Figure 1D). Expression of NgR1 co-receptors LINGO-1, TROY and p75^NTR was not down-regulated upon maturation in the same way as NgR1.

Due to the fact that we later used mouse DCs in our functional analysis of Ngr1/2 knockout, we also characterised mouse bone marrow derived DCs and analysed expression of NgR1, NgR2 and co-receptors. The characterisation of WT mouse DC is shown in Figure 2A. Like human DC, more than 95% of mouse in vitro-generated DCs express CD11b, suggesting a high purity of monocyte-derived DCs, whereas CD14 expression was low. There is a trend towards higher CD86 and lower CD11 c upon maturation, in concurrence with Lutz et al. [28].

In contrast to human DCs, mouse NgR1 expression does not change significantly upon maturation; however there is a trend towards up-regulation in mature DCs (Figure 2B). NgR2 expression on the other hand, is significantly down-regulated upon addition of maturation cocktail, however not to the same extent as in human mature DCs. LINGO-1 and p75^NTR are expressed at similar levels as NgR1 in mouse immature and mature DCs but also do not follow the same expression pattern observed in human DCs. Discrepancies in expression of NgRs between human and mouse DCs are also reflected in expression of DC cell surface markers, suggesting we are not dealing with directly comparable cell types. Furthermore, in both the human and mouse systems, we did not observe NgR1's co-receptors being regulated in the same way as NgR1. However, as it has previously been demonstrated that NgR1 can function without the full complement of identified co-receptors [13, 14, 36], we went on to determine the functional role of NgR1 and NgR2 in human and mouse DCs.

NgR1/2 have been found to mediate the inhibition of cellular adhesion to myelin. Thus, in order to determine the possible function of NgR1/2 down-regulation in human mature DCs, adhesion of immature and mature DCs to a myelin substrate was quantified. Immature and mature DCs were plated on human myelin and adhesion was then calculated as the fold change in adhesion compared to on plastic control. Human mature DCs were found to adhere significantly more to human CNS myelin compared to immature DCs (Figure 3A). As mentioned above, human mature DCs down-regulate NgR1 and NgR2 (Figure 1D). Figure 3B depicts the correlation between NgR1/2 mRNA expression (graphed as the raw value for both, 1/ΔCt) in all DCs and their adhesion to myelin, showing that increased NgR1/2 expression correlates significantly with decreased adhesion to myelin.

In order to identify which protein fraction of myelin is responsible for promoting adhesion of matDCs, we isolated various components of myelin that do or do not interact with NgRs, and measured adhesion of the cells in comparison to plastic control. His-tagged Nogo-66 recombinant peptide was plated at the same concentrations as myelin and used as a positive control for NgR1. As a positive control for NgR2, MAG-Fc was used. As MAG-Fc needs to be clustered in order to function correctly, the plate was first coated with IgG and then MAG-Fc was added. Myelin basic protein (MBP) was isolated from human brain white matter and used as a negative control as it is not known to bind or activate NgR1 or NgR2. Nogo-66-His demonstrated the same effect on adhesion of immature and mature DC as seen with myelin (Figure 3C). That is, mature DCs (which express less NgR1/2) were found to adhere to a much higher extent to Nogo-66-His than immature DCs. Adhesion of mature DCs to MBP and MAG-Fc was found to remain at a background level and not reach statistical significance. To ensure that the adhesion observed with Nogo-66-His was not due to side effects of the His-tag or bacterial contamination, we performed a control adhesion assay with two His-tagged and bacterially expressed peptides that do not bind NgR1/2. Neither NiR-His (amino acids 1 - 172 of rat Nogo-A) nor the 66-amino acid loop domain of RTN1 had any effect on adhesion of DCs (data not shown), thus suggesting that increased adhesion of matDCs to Nogo-66-His is indeed specific. This confirms that Nogo contained in the myelin preparation mediates promotion of matDC adhesion. Generally, matDCs adhere better to any substrate, however, the increase is most significant for myelin and Nogo-66-His. This indicates that Nogo might mediate this effect (probably due to the loss of NgR1).

Having found that Nogo-66 promotes adhesion of human matDC, we wanted to further clarify if it is indeed the loss of NgR1 expression that is the functional cause for increased adhesion of matDCs to myelin. To this end, we took DCs from NgR1/2 double knockout (Ngr1/2-/-) mice and compared them to wild type (WT) before measuring how they adhere to a myelin substrate. The double knockout mice were used rather than the single knockouts in order to exclude an effect resulting from the possible compensatory up-regulation of NgR2 in Ngr1-/- DCs.

DCs generated in vitro from WT and Ngr1/2-/- mice show similar phenotypes (Figure 2A and 2C). Furthermore, 32 cytokines released from WT and Ngr1/2-/- DCs were compared in cell culture supernatants using a glass chip protein array system (Figure 2D). We found no significant changes in secreted cytokines from mouse WT and Ngr1/2-/- DCs, thus indicating that the deletion of NgR1 and NgR2 had no influence on the differentiation and phenotype of DC. However, in the adhesion assay we did observe that mature DCs from Ngr1/2-/- mice adhere significantly more to myelin than mature DCs from WT mice (p = 0.02, Figure 3D). This indicates that a lack of NgR1/2 in mouse mature DCs promotes their adhesion to a myelin substrate.