J Mol Neurosci (2009) 38:2–11 DOI 10.1007/s12031-008-9151-x
Discoidin Domain Receptor 1, a Tyrosine Kinase Receptor, is Upregulated in an Experimental Model of Remyelination and During Oligodendrocyte Differentiation In Vitro Neus Franco-Pons & Jordi Tomàs & Bárbara Roig & Carme Auladell & Lourdes Martorell & Elisabet Vilella
Received: 22 January 2008 / Accepted: 16 September 2008 / Published online: 4 October 2008 # Humana Press 2008
Abstract The discoidin domain receptor (DDR1) is highly expressed in oligodendrocytes during the neurodevelopmental myelination process and is genetically associated to schizophrenia. In this study, we aimed to further assess the involvement of DDR1 in both remyelination and oligodendrocyte differentiation. In the mouse model of demyelination–remyelination induced by oral administration of cuprizone, in situ hybridization showed an upregulation of the DDR1 gene in three different white matter areas (corpus callosum, dorsal fornix, and external capsule) during the remyelination period. Moreover, real time reverse transcriptase polymerase chain reaction showed that the increase in DDR1 messenger RNA (mRNA) was strongly correlated with the number of
DDR1-positive cells in the corpus callosum (Spearman coefficient=0.987, P=0.013). Cells positive for DDR1 mRNA were also positive for oligodendrocyte markers (OLIG2, carnosine, and APC) but not for markers of oligodendrocyte precursors (NG2), myelin markers (CNPase), microglia (CD11b), or reactive glia (GFAP). Differentiation of a human oligodendroglial cell line, HOG16, was associated with an increase in mRNA expression of DDR1 and several myelin proteins (MBP and MOBP) but not other proteins (APC and CNPase). Here, we demonstrate that DDR1 is upregulated in vitro and in vivo when oligodendrocyte myelinating machinery is activated. Further studies are needed to identify the specific molecular pathway. Keywords Cuprizone . CNPase . DDR1 . HOG16 cells . Myelin . OLIG2 . Schizophrenia
N. Franco-Pons : J. Tomàs : B. Roig : L. Martorell : E. Vilella (*) Unitat de Psiquiatria i Psicologia Mèdica, Facultat de Medicina i Ciències de la Salut, Universitat Rovira i Virgili, C/Sant Llorenç 21, 43201 Reus, Spain e-mail:
[email protected] N. Franco-Pons : J. Tomàs : B. Roig : L. Martorell : E. Vilella Departament de Formació i Investigació, Hospital Psiquiàtric Universitari Institut Pere Mata, Ctra. de l’Institut Pere Mata s/n, 43206 Reus, Spain C. Auladell Departament de Biologia Cel·lular, Facultat de Biologia, Universitat de Barcelona, Avda. Diagonal 645, 08034 Barcelona, Spain
Introduction We recently reported a genetic association between the discoidin domain receptor 1 (DDR1) gene and schizophrenia in a large case-control study (Roig et al. 2007). Converging evidences suggest that altered oligodendrocyte function and myelin impairment may be involved in schizophrenia pathogenesis (Davis et al. 2003; McInnes and Lauriat 2006). Moreover, several white matter disorders are associated with schizophrenia-like psychoses at a rate garter than chance (Walterfang et al. 2006). Chronic administration of quetiapine, an antipsychotic drug, to C57BL/6 mice prevented both cortical demyelination and concomitant spatial working memory impairment induced by the neurotoxin cuprizone (Xiao et al. 2008). Oral
J Mol Neurosci (2009) 38:2–11
administration of cuprizone, which is toxic to oligodendrocytes, causes demyelination followed by spontaneous remyelination episodes during exposure and definitive remyelination following cuprizone withdrawal if the exposure does not exceed 12 weeks (Matsushima and Morell 2001; Stidworthy et al. 2003). An oral dose of 0.2% of cuprizone given to C57BL/6J mice for up to 6 weeks causes demyelination without hepatic toxicity (Hiremath et al. 1998) and moderate behavioral deficits (Liebetanz and Merkler, 2006; Franco-Pons et al. 2007). In the mouse brain, DDR1 is expressed in neuroepithelial cells during early developmental stages (Zerlin et al. 1993; Sanchez et al. 1994). During late postnatal development, DDR1 is mainly expressed in oligodendrocytes, and its spatial–temporal pattern overlaps the myelination process. Mice oligodendrocytes in the adult brain and human oligodendroglial cell lines also express DDR1 (FrancoPons et al. 2006). DDR1 is a tyrosine kinase receptor with a discoidin domain in its extracellular region which binds collagen. Upon collagen binding, receptor dimers begin a slow and prolonged process of auto-phosphorylation (Vogel et al. 1997; Abdulhussein et al. 2004). Activation of DDR1 induces a variety of cell functions that include migration (Jonsson and Anderson 2001; Kamohara et al. 2001; Hou et al. 2002), axon elongation of the cerebellar granule cells (Bhatt et al. 2000), branching regulation of the mammary gland (Vogel et al. 2001), kidney branching tubulogenesis (Wang et al. 2005), maturation of dendritic cells (Matsuyama et al. 2003), and modulation of collagen fibrillogenesis (Agarwal et al. 2007). In the central nervous system (CNS), axons are wrapped by concentric lipid-rich membranes to isolate them and improve electrical signaling. The multilayered membrane structure surrounding axons, known as myelin, is formed by prolongation of the oligodendocyte processes. Overall, myelination is a complex process that involves proliferation, migration, and differentiation of oligodendrocyte progenitor cells (Baumann and Pham-Dinh 2001). Some types of insults to the CNS lead to axon demyelination and result in disorders such as multiple sclerosis. Remyelination occurs when the CNS attempts to recover from myelin loss (Chang et al. 2002). One of the most important steps in remyelination is the activation of oligodendrocyte precursor cells (OPC), which progress through several phenotypic stages from immature to mature oligodendrocytes (Baumann and Pham-Dinh 2001). With the rationale that DDR1 is expressed in myelinating oligodendrocytes and associated to schizophrenia and that schizophrenia, in turn, is associated to myelin impairment, we directed the present study to evaluate the expression of DDR1 in: (1) an in vivo model of remyelination caused by oral administration of cuprizone and (2) an in vitro model of oligodendrocyte differentiation.
3
Materials and Methods Animals and Cuprizone Administration We reproduced the demyelination–remyelination mouse model using a diet supplemented with 0.2% cuprizone, as previously described (Franco-Pons et al. 2007; Hiremath et al. 1998). Briefly, C57BL/6J male mice (n=72) were divided into control groups fed with a regular diet and treatment groups that received a diet supplemented with 0.2% cuprizone for 3, 4, 5, or 6 weeks. In the recovery group, animals were fed the cuprizone-containing diet for 6 weeks, followed by 6 weeks on a regular diet. Figure 1A shows a schematic drawing of the experimental design. The mice were cared for in accordance with the guidelines in the European Community’s Council Directive of 24 November 1986 (86/609/EEC) for the humane treatment of laboratory animals. All mice were deeply anesthetized by intraperitoneal injection of ketamine (80 mg/kg) and xylazine (10 mg/kg) dissolved in 0.9% saline and transcardially perfused with 4% paraformaldehyde. The brains were surgically removed and either embedded in paraffin for histochemical studies or cryopreserved for in situ hybridization and immunohistochemical studies. Serial coronal paraffin-embedded (7 μm) and cryopreserved sections (30 μm) corresponding to bregma 0.62 mm to bregma −0.46 mm of the mouse brain atlas (Paxinos and Franklin 2001) were obtained and processed for immunohistochemistry and in situ hybridization. Using a binocular magnifier, the corpus callosum was dissected from the surgically removed brains and processed for RNA isolation. We selected this brain region because it is one of the most widely studied regions in this model (Hiremath et al. 1998; Matsushima and Morell 2001). Luxol fast blue staining and real time reverse transcriptase polymerase chain reaction (RT-PCR) were used to prove the de- and remyelination processes, as has been previously described (Hiremath et al. 1998; Morell et al. 1998; Franco-Pons et al. 2007). Cell Lines and Cell Culture The human oligodendroglial cell line HOG16 was accessed from Eucellbank (Department of Cellular Biology, University of Barcelona, Barcelona, Spain) with the permission of Dr. G. Dawson (University of Chicago, Chicago IL, USA). The cell line was cultured in growth medium (GM) containing high glucose Dulbecco’s Modified Eagle medium (HGDMEM, Gibco, Invitrogen Ltd., Paisley, UK) supplemented with 10% fetal bovine serum (FBS, HyClone, Utah, USA), 50 U/ml penicillin, and 50 μg/ml streptomycin (Gibco, Invitrogen Ltd., Paisley, UK) in a humidified 5% CO2 incubator at 37°C. Cells were free of mycoplasms, as tested
4
A R 6
Group
5 4 3
0.2% Cuprizone Diet Regular Diet
0 -1
0
1
2
3
4
5
6
7
8
9
10
11
12
Experimental time (weeks)
B
C
Luxol fast blue 0
5
3
6
4
12
Relative mRNA expression
Figure 1 Demyelination– remyelination mouse model generated by administering 0.2% oral cuprizone. A Schematic drawing of the experimental design; B luxol fast blue myelin stain in coronal slices of the corpus callosum at 0, 3, 4, 5, and 6 weeks of cuprizone administration and following 6 weeks recovery with a regular diet (12). Scale bar, 100 μm. C Relative mRNA expression, as measured by real time RT-PCR of cell markers in the corpus callosum. Quantification of macrophage (MHCI), astrocyte (GFAP), axon (paranodin) and oligodendrocyte (MAG) markers. For technical details, see the “Materials and Methods”
J Mol Neurosci (2009) 38:2–11
8
M HCI GFAP
6
Paranodin M AG
4
2
0 0
1
2
3
4
5
6 9 10 11 12
Time (weeks)
with a commercial kit (EZ-PCR Mycoplasma Test Kit, Biological Industries, Kibbutz Beit Haemek, Israel). To induce differentiation, HOG16 cells were subjected to differentiation medium (DM) containing 0.05% FBS, 30 nM triiodothyronine (T3), 30 nM selenium, 0.5 μg/ml insulin (all from Sigma), 50 μg/ml transferrin (US Biological, Swampscott, MA, USA), and the antibiotics used above according to Buntinx, 2003. Cells that grew in multi-well plates with GM and reached 50% confluency were divided into two groups. Half of the wells continued with GM, and in the other half, the medium was substituted with DM. After a 48-h incubation, the medium was removed, and the cells were dissolved in lysis buffer. The experiment was performed in triplicate wells and during three different weeks. Parallel experiments were carried out in a Lab-Tek II Chamber Slide System (Nunc, Roskilde, DK). After a 48-h incubation, the medium was removed, and the cells were fixed.
with PBS, the cells were incubated for 30 min with goat anti-rabbit FITC-labeled secondary antibody; washing the cells three times with PBS eliminated excess secondary antibody.
Immunocytochemistry
First-strand cDNA was synthesized from 1 μg of total RNA using random hexamers and Superscript II RNAse H-Reverse Transcriptase (Invitrogen, Barcelona, Spain). Messenger RNA (mRNA) expression was quantified by RT-PCR using an ABI PRISM 5700 Sequence Detector System (Applied Biosystems, Madrid, Spain) combined with a dual-label fluorogenic detection system (Applied Biosystems, Madrid, Spain) based on a 5′ nuclease assay.
The cells were washed in phosphate-buffered saline (PBS), fixed in 2% paraformaldehyde for 30 min, permeabilized with 0.05% Triton X-100 for 30 min, washed again three times in PBS, and left to incubate overnight at 4°C in PBS with anti-DDR1 antibody (sc532, dilution 1:100; Santa Cruz Biotechnology, Madrid, Spain). After three washes
RNA Isolation The corpus callosum from each of the cuprizone-exposed and control mice (22 animals in total) was sonicated with a blender in RNAase-free lysis buffer (Applied Biosystems, Madrid, Spain). Samples were kept for 1 h at 4°C. Cell cultures were stopped by removing the medium and adding RNAase-free lysis buffer (Applied Biosystems, Madrid, Spain). Samples were kept for 1 h at 4°C. In both cases, total RNA was isolated using an ABI PRISM 6100 Nucleic Acid PrepStation (Applied Biosystems, Madrid, Spain). RNA concentration was estimated by spectrophotometry. Real Time Quantitative RT-PCR Analysis
J Mol Neurosci (2009) 38:2–11
G3PDH mRNA was used as an endogenous control. TaqMan primers and probes for mouse DDR1, MAG, paranodin, glial fibrillary acidic protein (GFAP), and major histocompatibility complex class I (MHCI) cDNAs were obtained from validated and predesigned Assays-onDemand (Applied Biosystems, Madrid, Spain) and used in real time PCR amplifications to detect the expression of myelin (MAG), reactive glia and astrocytes (GFAP), microglia and macrophages (MHCI), and neurons (paranodin). TaqMan primers and probes for human DDR1, APC, CNPase, MBP, and MOBP cDNAs were obtained from validated and predesigned Assays-on-Demand (Applied Biosystems, Madrid, Spain). The reactions were performed in triplicate using 5 μl of cDNA in a 25 μl reaction volume. mRNA expression for each sample was calculated using the comparative cycle threshold (Ct) method using 2−ΔΔCt, according to the manufacturer’s instructions (Applied Biosystems, Madrid, Spain). Quantification of specific cDNAs was performed relative to a “calibrator” control sample serving as reference. The 2−ΔΔCt for this “calibrator” control sample was arbitrarily set to 1.
5
1:2000; Dako, Barcelona, Spain), CD11b (dilution 1:300; AbD Serotec, BioNova Cientifica, Madrid, Spain), NG2 (dilution 1:300; Chemicon, Barcelona, Spain), OLIG 2 (dilution 1:100; IBL, Hamburg, Germany), carnosine (dilution 1:300; a kind gift from Dr. M.L. Margolis of the University of Maryland, MD, USA), APC (dilution 1:100; Calbiochem, Barcelona, Spain) and CNPase (dilution 1:100; Sternberger monoclonals, Madrid, Spain) were used to phenotype reactive glia (GFAP), microglia (CD11b), OPCs (NG2, OLIG2), mature oligodendrocytes (carnosine and APC), and myelin (CNPase), respectively. Following overnight incubation, the slides were washed in PBS and incubated with appropriated secondary biotinylated antibody for 1 h at room temperature. After additional washes, secondary antibodies were detected using the avidin–biotin complex reaction (ABC Elite Kit; Vector Laboratories, Barcelona, Spain) and visualized with a solution of 0.025% diaminobenzidine tetrahydrochloride in phosphate buffer (pH 7.4) and 0.005% H2O2. The sections were rinsed in PBS and mounted with Mowiol. Image Analysis and Cell Counting
In Situ Hybridization In total, 34 animals were assigned to the treated and control groups. In situ hybridization and preparation of digoxigenin-labeled riboprobes were performed as previously described in detail (Franco-Pons et al. 2006). Briefly, the cDNA template that was used (a kind gift from Dr. M.E. Hatten of the Rockefeller University, New York, NY, USA) was complementary to an approximately 900 bp mRNA fragment encoding a portion of the tyrosine kinase domain of the mouse DDR1 gene. The construct was linearized to prepare sense or antisense probes, and RNA probes were transcribed with an in vitro transcription kit (Ambion, Madrid, Spain) to incorporate digoxigenin-UTP (Roche Diagnostics, Barcelona, Spain). The digoxigenin-labeled riboprobes were hybridized to 30 μm cryosections of the excised brains. Digoxigenin was detected with an alkaline phosphatase-conjugated rabbit anti-digoxigenin antibody (Roche Diagnostics, Barcelona, Spain), followed by reaction with nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate substrate (Roche Diagnostics, Barcelona, Spain). Double Labeling Immunohistochemistry and In Situ Hybridization In situ-hybridized sections from control animals and those exposed to 0.2% cuprizone for 3, 4, 5, or 6 weeks were washed in 0.2% Triton ×100 in 0.1 M phosphate buffered saline (PBS-T) and then incubated overnight at 4°C with primary antibody diluted in PBS-T containing 1% fetal bovine serum solution. Antibodies against GFAP (dilution
Representative sections were analyzed under a light or fluorescence microscope (Nikon Eclipse 600, Barcelona, Spain) equipped with a camera and filters that allow excitation at 495/10 nm (FITC), and labeled structures were digitally photographed using objectives of several magnifications (×2 to ×40, ×100). Low magnification was used for overviews of myelin expression, and high magnification was used primarily for double in situ hybridizationimmunohistochemistry photographs and immunochemistry where necessary. The amount of cells displaying DDR1 in specific brain regions of treated animals were quantified using AnalySISTM software (Soft Imaging System, Münster, Germany) in accordance with the manufacturer’s instructions. Results are expressed as cell count/mm2 (mean ± SD of three independent counts). Fluorescence pictures were taken after 40 s of fluorescent exposure in accordance with the software manager’s instructions. Statistical Analysis A Spearman correlation coefficient was used to measure association between the number of cells expressing DDR1 (labeled by in situ hybridization) and the number of DDR1 mRNA copies (measured by Real time RT-PCR) in the corpus callosum. Comparisons of the number of mRNA copies in cells grown in GM or DM were done by analysis of variation (ANOVA). Statistical significance was set at P
