ADAM12 is expressed by astrocytes during experimental demyelination

BR A IN RE S EA RCH 1 3 26 ( 20 1 0 ) 1 –1 4

available at www.sciencedirect.com

www.elsevier.com/locate/brainres

Research Report

ADAM12 is expressed by astrocytes during experimental demyelination Fabian Baertlinga,1 , Maria Kokozidoud,1 , Thomas Pufeb , Tim Clarnera , Reinhard Windoffer c , Christoph J. Wruckb , Lars-Ove Brandenburgb , Cordian Beyer a , Markus Kippa,⁎ a

Institute of Neuroanatomy, Faculty of Medicine, RWTH Aachen University, Wendlingweg 2, D-52074 Aachen, Germany Institute of Anatomy and Cell Biology, Faculty of Medicine, RWTH Aachen University, Wendlingweg 2, D-52074 Aachen, Germany c Institute of Molecular and Cellular Anatomy, Faculty of Medicine, RWTH Aachen University, Wendlingweg 2, D-52074 Aachen, Germany d Department of Vascular Surgery, Faculty of Medicine, RWTH Aachen University, Pauwelsstrasse 30, D-52074 Aachen, Germany b

A R T I C LE I N FO

AB S T R A C T

Article history:

A disintegrin and metalloproteinase (ADAM) 12 represents a member of a large family of similarly

Accepted 14 February 2010

structured multi-domain proteins. In the central nervous system (CNS), ADAM12 has been

Available online 20 February 2010

suggested to play a role in brain development, glioblastoma cell proliferation, and in experimental autoimmune encephalomyelitis. Furthermore, ADAM12 was reported to be almost exclusively

Keywords:

expressed by oligodendrocytes and could, therefore, be considered as suitable marker for this cell

ADAM12

type. In the present study, we investigated ADAM12 expression in the healthy and pathologically

Astrocyte

altered murine CNS. As pathological paradigm, we used the cuprizone demyelination model in

Neuron

which myelin loss during multiple sclerosis is imitated. Besides APC+ oligodendrocytes, SMI311+

Oligodendrocyte

neurons and GFAP+ astrocytes express ADAM12 in the adult mouse brain. ADAM12 expression was

Cuprizone

further analyzed in vitro. After the induction of demyelination, we observed that activated

Demyelination

astrocytes are the main source of ADAM12 in brain regions affected by oligodendrocyte loss. Exposure of astrocytes in vitro to either lipopolysaccharides (LPS), tumor necrosis factor α (TNFα), glutamate, or hydrogen peroxide revealed a highly stimulus-specific regulation of ADAM12 expression which was not seen in microglial BV2 cells. It appears that LPS- and TNFα-induced ADAM12 expression is mediated via the classic NFκB pathway. In summary, we demonstrated that ADAM12 is not a suitable marker for oligodendrocytes. Our results further suggest that ADAM12 might be implicated in the course of distinct CNS diseases such as demyelinating disorders. © 2010 Elsevier B.V. All rights reserved.

⁎ Corresponding author. Wendlingweg 2, D-52074 Aachen, Germany, RWTH Aachen, Institute of Neuroanatomy, Germany. Fax: +49 241 80 82 472. E-mail address: [email protected] (M. Kipp). Abbreviations: ADAM, a disintegrin and metalloproteinase; APC, adenomatous polyposis coli; AT, annealing temperature; bFGF, basic fibroblast growth factor; BP, base pairs; BSA, bovine serum albumine; CC, corpus callosum; CNS, central nervous system; Cx, cortex; ctrl, control; cup, cuprizone; DMEM, Dulbecco's modified Eagle's medium; EAE, experimental autoimmune encephalomyelitis; FCS, fetal calf serum; GFAP, glial fibrillary acidic protein; Glut, glutamate; H2O2, hydrogen peroxide; HBEGF, heparin-binding epidermal growth factor; HPRT, hypoxanthine guanine phosphoribosyltransferase; IL1ß, interleukin 1ß; IHC, immunohistochemistry; IGF, insulin-like growth factor; IGFBP, insulin-like growth factor binding protein; LPS, lipopolysaccharides; MS, multiple sclerosis; NFκB, nuclear factor kappa-lightchain-enhancer of activated B cells; PBS, phosphate-buffered saline; PLP, proteolipoprotein; rt, real-time; SDS-PAGE, sodium dodecylsulfate polyacrylamide gel electrophoresis; SEM, standard error of the mean; sq, semi-quantitative; TNFα, tumor necrosis factor α 1 Contributed equally to this work. 0006-8993/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2010.02.049

2

1.

B RA IN RE S EA RCH 1 32 6 (2 0 1 0 ) 1 –1 4

Introduction

The “a disintegrin and metalloproteinase” (ADAM) family consists of numerous similarly structured multi-domain proteins involved in many physiological and pathological processes. Altered expression and functions of specific ADAMs are implicated in the pathophysiology of several diseases including rheumatoid arthritis (Valdes et al., 2004), Alzheimer's disease (Malinin et al., 2005; Tanabe et al., 2007), cardiac hypertrophy (Asakura et al., 2002), asthma (Chiba et al., 2009), and several types of cancer (Mochizuki and Okada, 2007). They are found in various species and are expressed in many organs and tissues (Wolfsberg et al., 1995) such as the skeletal muscle (Yagami-Hiromasa et al., 1995), the placenta (Gilpin et al., 1998), and the central nervous system (CNS) (Bernstein et al., 2004). The prototypical ADAM protein contains an extracellular pro-metalloproteinase, a disintegrin-like, a cysteine-rich and an epidermal growth factor-like domain, followed by a transmembrane- and a cytoplasmic tail domain (Kveiborg et al., 2008). Functions of different ADAMs include cell adhesion (Najy et al., 2008), cell fusion (YagamiHiromasa et al., 1995), proteolysis (Moss et al., 2007), and facilitating of cell proliferation (Kodama et al., 2004) and migration (Estrella et al., 2009) allowing them to take part in different biological processes. These include fertilization (Pasten-Hidalgo et al., 2008), angiogenesis (Mahoney et al., 2009) and, as so far most elaborately investigated, the development and progression of cancer (Mochizuki and Okada, 2007). In the CNS, ADAMs have been suggested to play a role during brain development (Lin et al., 2008) and to contribute to axon extension (Fambrough et al., 1996) and neurogenesis (Pan and Rubin, 1997). ADAM12 possesses extracellular metalloproteinase and cell-binding properties as well as intracellular signaling capacities (Kveiborg et al., 2008). A variety of functions have been suggested for ADAM12, including support of releasing growth factors such as heparin-binding epidermal growth factor (HBEGF) (Asakura et al., 2002; Kodama et al., 2004) and insulin-like growth factor (IGF) 1 (Loechel et al., 2000) as well as interaction with cell surface integrins. ADAM12 is, therefore, likely to mediate cell adhesion, differentiation, proliferation, and migration (Kveiborg et al., 2008). Furthermore, it was suggested that ADAM12 takes part in myogenesis (Gilpin et al., 1998), bone growth (Kveiborg et al., 2006), and fetal development (Cowans and Spencer, 2007). Besides such physiological functions, ADAM12 seems to be involved in pathological processes such as cardiac hypertrophy (Asakura et al., 2002), osteoarthritis (Valdes et al., 2004), and cancers of various tissue origins (Kveiborg et al., 2008). So far, little is known about the function of ADAM12 within the CNS. It has been suggested that ADAM12 is necessary for proper brain development and regionalization, since it is expressed in highly restricted regions of the neuroepithelium (Lin et al., 2008). An increase of ADAM12 expression was found in human glioblastomas where it correlates with the proliferative activity of tumor cells (Kodama et al., 2004). Among brain resident cells, ADAM12 is supposed to be almost exclusively expressed by oligodendroglia under physiological conditions and is, therefore, considered to be a suitable oligodendrocyte

marker (Bernstein et al., 2004). During experimental autoimmune encephalomyelitis (EAE), an animal model widely used to investigate pathological mechanisms of human multiple sclerosis (MS), ADAM12 expression is up-regulated in the mouse spinal cord due to infiltrating T-cells (Toft-Hansen et al., 2004). In order to further investigate the role of ADAM12 within the intact and pathologically altered rodent brain, we analyzed the expression of ADAM12 in C57BL6 mice. The cuprizone demyelination model was used as a model for MS, since feeding of cuprizone induces highly reproducible demyelination of different brain areas (Kipp et al., 2009). The cellular source of ADAM12 in the intact and demyelinated brain was investigated by consecutive slice immunohistochemistry (IHC) and immunofluorescence double labeling. Primary cell culture experiments were additionally performed to gain insight into stimulus-specific regulation of ADAM12 expression.

2.

Results

2.1. ADAM12 is expressed by oligodendrocytes and cortical neurons It was recently reported that ADAM12 is expressed in the rodent brain. With the exception of very few immunopositive pyramidal neurons in the developing rat brain, ADAM12 was exclusively localized to oligodendrocytes (Bernstein et al., 2004). In order to confirm these findings in the murine CNS, we performed immunofluorescence double labeling with an antiADAM12 antibody and antibodies for respective cell markers in young male C57BL6 mice. SMI311 was selected to provide a specific marker for neurons. In contrast to markers for individual non-phosphorylated neurofilaments that identify different subsets of neurons and are, therefore, especially suitable for defining anatomic and functional differences in normal and pathologically altered neurons, SMI311 is a general marker for adult neurons and differentiating neuronal precursors (Ulfig et al., 1998). An anti-adenomatous polyposis coli (APC) antibody was selected to visualize late stage oligodendrocyte cells (Groebe et al., 2009; Norkute et al., 2009; Pott et al., 2009). Colocalization of ADAM12 and 95% of APC+ oligodendrocytes was confirmed within the corpus callosum (CC) (Fig. 1) and the cerebral cortex (not shown). ADAM12 staining in oligodendrocytes was restricted to the cell membrane (arrow in Fig. 1I) and the cytoplasm sparing the cell nucleus. Unexpectedly, immunofluorescence double labeling for ADAM12 and SMI311 revealed that ADAM12 is also expressed by numerous SMI311+ neurons in the telencephalic cortex of the adult mouse brain (Fig. 2). ADAM12 staining in neurons was mainly confined to the neuronal cell bodies sparing dendritic branches (arrows in Fig. 2I). Staining intensity was similar to that observed in APC+ cells. Since only few ADAM12+ neurons were noted in the study conducted by Bernstein et al., we performed Western Blot analysis to confirm the specificity of the anti-ADAM12 antibody. Proteins were isolated from the CC and the cortex and applied in different concentrations (20 µg, 10 µg and 5 µg) for sodium dodecylsulfate polyacrylamide gel electrophoresis

BR A IN RE S EA RCH 1 3 26 ( 20 1 0 ) 1 –1 4

3

Fig. 1 – Immunofluorescence double labeling of ADAM12 and APC+ late stage oligodendrocyte cell bodies in the intact CC. Cell nuclei are visualized by Hoechst 33342. Images are taken from the lateral radiation of the CC which is designated by the rectangle in (A). Arrowheads point at positive cells for the respective markers. Note that ADAM12 is expressed by almost all APC+ oligodendrocytes and mainly found near the cell membrane as indicated by the arrow in the merged image at higher magnification (I) and, furthermore, within the cytoplasm. For abbreviations see text. Scale bars: 30 µm (B–E), 3 µm (F–I).

(SDS-PAGE). After incubation with the anti-ADAM12 antibody, only one strong band was detected at 50 kDa for both investigated brain regions and for every protein concentration applied, even for the maximum one, thus, indicating highly specific antibody binding. Furthermore, Western Blot analysis confirmed the expression of relevant amounts of ADAM12 protein in the CC and the cortex and demonstrated that similar quantities of ADAM12 are expressed in both regions (Fig. 3).

2.2. ADAM12 is additionally expressed by astrocytes but not microglial cells In order to investigate whether other cell types than neurons and oligodendrocytes might express ADAM12 in the rodent brain, we performed ADAM12 gene expression analysis by means of semi-quantitative (sq) RT-PCR in a set of cell cultures. As shown in Fig. 4A, ADAM12 was expressed in relevant amounts in oligodendroglial (primary oligodendrocytes and OLN93 cells) and neuronal (primary cortical neurons) cultures. In addition, ADAM12 mRNA was strongly expressed in primary astrocyte cell cultures (Fig. 4B) under regular conditions. In contrast, the microglial cell line BV2 neither expressed ADAM12 mRNA under regular conditions nor after stimulation with distinct neuropathological toxins

such as lipopolysaccharides (LPS), tumor necrosis factor α (TNFα), glutamate (Glut), or hydrogen peroxide (H2O2) (Fig. 4C).

2.3.

Astrocytes also express ADAM12 in vivo

In a next step, we intended to confirm astrocytic ADAM12 expression in vivo. Since immunofluorescence double labeling using anti-glial fibrillary acidic protein (GFAP) and antiADAM12 antibodies produced results that were not entirely reliable, we used consecutive slices (each 5 µm thick) for GFAP and ADAM12 staining. Reasonable numbers of GFAP+ cells were found in the CC (Fig. 5B). However, consecutive ADAM12 stained slices only sporadically showed an ADAM12+ cell with typical astrocytic morphology yet plenty of ADAM12+ cells displaying the characteristic chain-like alignment of late stage oligodendrocytes (Fig. 5C). Thus, we concluded that ADAM12 is also expressed by GFAP+ astrocytes under regular conditions in vivo, however, only by few.

2.4. ADAM12 expression by astrocytes is increased in brain regions affected by cuprizone-induced demyelination In spinal cords of EAE mice, ADAM12 expression is upregulated due to infiltrating T-cells (Toft-Hansen et al., 2004). We analyzed ADAM12 expression using cuprizone-induced

4

B RA IN RE S EA RCH 1 32 6 (2 0 1 0 ) 1 –1 4

Fig. 2 – Immunofluorescence double labeling of ADAM12 and SMI311+ mature neuronal cell bodies and dendrites in the intact cortex. Cell nuclei are visualized by Hoechst 33342. Images are taken from the central part of the cortex which is designated by the rectangle in (A). Arrowheads point at positive cells for the respective markers. Note that ADAM12 is expressed by numerous SMI311+ cells and is localized to the neuronal cell bodies sparing dendritic branches as indicated by the arrows in the merged image at higher magnification (I). For abbreviations see text. Scale bars: 30 µm (B–E), 3 µm (F–I).

demyelination (Kipp et al., 2009), another MS animal model. During cuprizone exposure, demyelination occurs despite an intact blood–brain-barrier and the absence of T-cell invasion (Emerson et al., 2001; Mana et al., 2009). Besides oligodendrocyte loss and subsequent myelin break-down, the cuprizone model is characterized by astrocyte and microglia cell activation and invasion (Kipp et al., 2009). Young male mice were fed cuprizone for five weeks to induce acute demyelination within the CC and the telencephalic grey matter. Cuprizone ingestion led to profound and significant demyelination of the CC and the cortical grey matter as demonstrated by anti-proteolipoprotein

(PLP) IHC (Fig. 6A/B). Demyelination was accompanied by massive oligodendrocyte loss as demonstrated by anti-APC IHC (Fig. 6C–F). Furthermore, five weeks of cuprizone exposure led to a highly significant decrease of PLP mRNA levels to nearly 20% compared to controls in the CC and 25% in the cortex (p < 0.001, Fig. 7A). ADAM12 mRNA levels were not significantly affected in the CC and cortex after cuprizone exposure (Fig. 7B), despite the depletion of ADAM12+ oligodendrocytes. In the cuprizone demyelination model, the number of astrocytes significantly increases between weeks four and five (Kipp et al., 2009; Matsushima and Morell, 2001). After five weeks of cuprizone-induced demyelination, about 60% of astrocytes expressed ADAM12 in the CC (Fig. 8C/D) as well as in the cortex (not shown).

2.5. Expression of ADAM12 by astrocytes in vitro is regulated in a pathogen-specific manner

Fig. 3 – Western Blot analysis of ADAM12 expression in the CC and the cortex (Cx) is shown. ADAM12 protein is expressed in relevant amounts and to a similar extent in both regions. β-Actin served as internal control. For further abbreviations see text.

Astrocytes are major regulators of intra-cerebral inflammatory processes (Karakaya et al., 2007; Kipp et al., 2008; Williams et al., 2007). Therefore, we investigated whether cortical astrocytes in vitro respond differently to the application of neurotoxins with respect to their capacity of ADAM12 expression. We applied LPS which provoke sepsis-induced neuronal damage (Henry et al., 2008) or bacteria-induced

BR A IN RE S EA RCH 1 3 26 ( 20 1 0 ) 1 –1 4

5

Fig. 4 – ADAM12 mRNA expression determined by sqRT-PCR in primary oligodendrocytes, OLN93 cells, as well as primary neuronal (A) and primary astrocyte cultures (B). (C) shows ADAM12 mRNA expression in control and stimulated BV2 cells. Note that not only primary oligodendrocytes, OLN93 cells, and primary cortical neurons express relevant amounts of ADAM12 mRNA under control conditions but also primary cortical astrocytes. ADAM12 mRNA expression cannot be detected in BV2 cells under control conditions (ctrl) or after stimulation with neuropathological toxins (LPS, TNFα, Glut or H2O2 for 24 h). Expression of the housekeeping gene HPRT served as internal control. For further abbreviations see text.

Fig. 5 – IHC of two consecutive slices (5 µm thickness) using anti-ADAM12 and anti-GFAP antibodies in the intact CC. Images are taken from the medio-lateral radiation of the CC as demonstrated in (A). The arrowheads point at an astrocyte positive for GFAP and ADAM12. The arrow in (C) indicates ADAM12+ late stage oligodendrocytes displaying the characteristic chain-like alignment within the CC. Note that only few GFAP+/ADAM12+ cells can be found in the intact CC. Further note that ADAM12+ cells with typical astrocytic morphology can be found within the CC (D). For abbreviations see text. Scale bars: 50 μm (B/C), 10 µm (D).

6

B RA IN RE S EA RCH 1 32 6 (2 0 1 0 ) 1 –1 4

Fig. 6 – Effect of five weeks cuprizone exposure on myelination and the number of oligodendrocyte cells in the mouse brain. (A) shows anti-PLP staining of a control animal (ctrl) and the investigated telencephalic cortex (star) and the CC (arrowhead). (B) shows anti-PLP staining of an animal treated with cuprizone for five weeks (5wCup). (C–F) show anti-APC stainings of control and cuprizone-treated animals in the cortex (C–D) and CC (E–F, rectangles outline the CC). Note the decline of PLP immunoreactivity and reduction of oligodendrocytes after cuprizone exposure. For further abbreviations see text. Scale bars: 1000 µm (A/B), 100 µm (C/D), 50 μm (E/F).

inflammatory processes (O'Reilly et al., 2007; Palsson-McDermott and O'Neill, 2004), the cytokine TNFα which is implicated in several pathological events in the brain (Gosselin and Rivest, 2007; Huang et al., 2005), H2O2 which mimics oxidative stress (Watt et al., 2004), and Glut known for its excitotoxic potential and relation to neurological disorders (Pitt et al., 2003; Werner et al., 2000). Astroglial cultures were exposed to the above factors for 24 h at non-toxic concentrations for astrocytes (Braun et al., 2009). Subsequently, ADAM12 mRNA expression was analyzed

by real-time (rt) RT-PCR. As shown in Fig. 9, LPS and TNFα exposure promoted ADAM12 expression by nearly 400% (p < 0.001) and almost 200% (p < 0.05) compared to controls. Glut or H2O2 exposure significantly decreased ADAM12 expression levels by approximately 50% compared to controls (p < 0.01). In order to gain insight into the mechanism of ADAM12 mRNA upregulation by LPS and TNFα, we additionally exposed cortical astrocytes to curcumin which inhibits transcription factor NFκB (Bengmark, 2006; Thangapazham et al., 2006). We found that the

Fig. 7 – Comparative analysis of PLP and ADAM12 gene expression in the corpus callosum (CC) and cortex (Cx) of mice after five weeks of cuprizone treatment (5wCup). Note that there is a strong decrease in PLP transcript levels (A) but ADAM12 mRNA levels remain unaffected (B). Values were normalized against the housekeeping gene HPRT and expressed as percentage (%) of controls (set to 100%). Data represent means ± SEM. For further abbreviations see text. *p < 0.001 control vs. cuprizone-treated.

BR A IN RE S EA RCH 1 3 26 ( 20 1 0 ) 1 –1 4

7

Fig. 8 – ADAM12 and GFAP IHC (A–F) staining in consecutive slices (5 µm thick) of untreated control animals (ctrl) and animals treated with cuprizone for five weeks (5wCup) and ADAM12/GFAP immunofluorescence double labeling (G–I). Images are taken from the medio-lateral radiation of CC (for details see Fig. 5A). Arrowheads indicate cells which are positive for both markers. Boxes in (C) and (D) indicate the cells shown at higher magnification in (E) and (F), respectively. Note that only few GFAP+/ ADAM12+ cells can be found in the intact CC whereas numerous GFAP+/ADAM12+ cells can be found in the demyelinated CC. Further note that multiple ADAM12+ cells with typical astrocytic morphology can be found within the injured CC (arrows in D). For further abbreviations see text. Scale bars: 50 μm (A–D), 10 μm (E/F), 10 µm (G–I).

application of curcumin 30 min prior to stimulation inhibits LPS- and TNFα-induced ADAM12 up-regulation.

3.

Discussion

The metalloproteinase ADAM12 which has been detected in human and rat brain tissues as well as in cultured brain cells appears to be expressed by oligodendrocytes and pyramidal neurons (Bernstein et al., 2004). In this study, we confirmed the presence of ADAM12 in late stage APC+ oligodendrocytes and, in

addition, we clearly showed that SMI311+ neurons within the telencephalic cortex express ADAM12. The expression of ADAM12 in oligodendrocytes and neurons was further confirmed in vitro: primary cultured rat oligodendrocytes, the OLN93 cell line and primary cultured cortical neurons revealed the presence of ADAM12 mRNA. Furthermore, ADAM12 was detectable in astrocytes in vitro and in vivo. In contrast, BV2 cells which have microglial properties did not express ADAM12. Thus, we showed that ADAM12 is not exclusively expressed by oligodendrocytes but rather seems to be widely distributed within the adult CNS in glial and non-glial cell types. Bernstein et al.

8

B RA IN RE S EA RCH 1 32 6 (2 0 1 0 ) 1 –1 4

Fig. 9 – Effect of treatment with neuropathological toxins (LPS, TNFα, Glut or H2O2 for 24 h) on ADAM12 expression in cortical astrocytes. Note that expression of ADAM12 is differentially regulated under the given stimuli. Administration of NFκB inhibitor curcumin (Cum) inhibits LPS- and TNFα-induced up-regulation of ADAM12 expression. Values were normalized against the housekeeping gene HPRT and expressed as % of controls (set to 100%). Data represent means ± SEM. For further abbreviations see text. *p < 0.05 **p < 0.01 ***p < 0.001 control vs. treatment.

reported that about 75% of all grey matter but only 25% of white matter oligodendrocytes express this enzyme. Rat tissue was used in their study. In our study, almost all APC+ oligodendrocytes were positive for ADAM12 regardless of whether they were located in the grey or white matter. This difference might be due to the fact that different marker proteins were used to visualize oligodendrocytes. Anti-APC staining, used in our study, labels pre-myelinating and myelinating oligodendrocytes (Mela and Goldman, 2009), whereas anti-galactocerebroside staining used by Bernstein et al. predominantly labels the myelinating cell population. In another ongoing study, we were able to show that ADAM12 is already expressed very early during oligodendrocyte development in vitro. Furthermore, ADAM12 expression might be different in rats and mice. It should be noted at this point that since the original description of APC as an oligodendrocyte marker (Bhat et al., 1996) there has been some uncertainty in the literature concerning its reliability as marker for oligodendrocyte cell bodies. Lee et al., for example, have recently shown that activated hippocampal rat astroglia also expresses APC after the induction of an excitotoxic lesion by kainic acid injection (Lee et al., 2009). However, we suggest that APC is a reliable oligodendrocyte cell marker in the cuprizone demyelination model since (i) the number of APC+ cells is dramatically reduced after cuprizone exposure, (ii) other brain regions showing strong astrocytosis and microglia invasion/proliferation and oligodendrocyte depletion after cuprizone exposure, such as the lateral part of CC, the adjacent cortex, or the basal ganglia system did not display increased numbers of APC+ cells (Groebe et al., 2009; Kipp et al., 2009; Pott et al., 2009) and (iii) there are no APC+ cells after five weeks of cuprizone exposure with characteristic astrocytic morphology. We further analyzed ADAM12 expression under pathological conditions. In order to analyze the regulation of ADAM12 in vitro, cultured cells were stimulated with a set of toxins which are implicated in neuroinflammatory and neurodegenerative processes. LPS provoke sepsis-induced neuronal damage (Henry et al., 2008) or bacteria-induced inflammatory processes (O'Reilly et al., 2007; Palsson-McDermott and

O'Neill, 2004), TNFα is implicated in several pathological events in the brain (Gosselin and Rivest, 2007; Huang et al., 2005), H2O2 mimics oxidative stress (Watt et al., 2004), and Glut is known for its excitotoxic effects in the brain (Pitt et al., 2003; Werner et al., 2000). ADAM12 was neither found in resting nor in activated BV2 microglial cells, although the applied dosage is sufficient to activate BV2 cells which was confirmed by means of chemokine expression analysis. In contrast, we observed that the expression of ADAM12 in cultured astrocytes is regulated in a toxin-specific manner. In vivo, an increased number of ADAM12+ astrocytes was observed after the induction of demyelination by cuprizone feeding. Astrocytes fulfill important functions in the brain ranging from directing neuronal differentiation to regeneration. Moreover, this cell type governs intra-cerebral inflammatory processes and is required to protect neurons under pathological conditions (Kajta et al., 2006; Kipp et al., 2007; Kipp et al., 2008; Pawlak et al., 2005). A particular function of astrocytes is their involvement in regional energy metabolism which becomes important under degenerative challenges (Horvat et al., 2006). For example, proliferation of astrocytes is a well-known phenomenon in active demyelinating lesions of MS patients (Lassmann, 2008; Noseworthy et al., 2000). Here, astrocytes might play an active role during degeneration and demyelination by promoting inflammation, damage of oligodendrocytes and axons as well as glial scarring. Yet, they might also be beneficial by creating a permissive environment for remyelination and oligodendrocyte precursor migration, proliferation and differentiation (Williams et al., 2007). Astrocytes secrete a multitude of harmful cytokines such as TNFα and interleukin 1β (IL1ß) (Kipp et al., 2007; Kipp et al., 2008) and pro-inflammatory prostaglandins (Johann et al., 2008). On the other hand, astrocytes produce important growth factors such as IGF1 and basic fibroblast growth factor (bFGF) which are known to be involved in remyelination and prevention of oligodendrocyte apoptosis (Komoly et al., 1992; Mason et al., 2000; Mason et al., 2003). It is apparent that astrocytes possess a dual role in MS disease progression which appears to be destructive at the onset and during acute focal demyelination

BR A IN RE S EA RCH 1 3 26 ( 20 1 0 ) 1 –1 4

by providing TNFα or IL1ß (Aktas et al., 2007; Ye et al., 2007) but might later become supportive due to the expression of IGF1 and bFGF (Barnett and Prineas, 2004; Morell et al., 1998). The role of ADAM12 secreted by astroglia during neurotoxic processes can only be assessed on a highly speculative basis, since no detailed information is available. However, in the above context, it might be assumed that astrocytes can affect the function of neighboring cells by secreting ADAM12. We have shown that the pro-inflammatory factors LPS and TNFα provoke and excitotoxic Glut and pro-oxidative H2O2 reduce the expression of ADAM12 in vitro. This indicates that ADAM12 is not merely up-regulated whenever astrocytes are activated but suggests that astrocytic ADAM12 expression under pathological conditions is more complex. Inhibition of transcription factor NFκB by curcumin attenuated LPS- and TNFα-induced ADAM12 up-regulation. Interfering with ADAM12 expression and its functions is a current topic with regard to its involvement in numerous diseases (Jacobsen and Wewer, 2009). Our results suggest that NFκB can be taken into account as a novel target for the inhibition of ADAM12 expression. MS is a chronic, inflammatory and demyelinating CNS disease and appears in several courses with symptoms either appearing in discrete episodes or slowly accumulating over time. Between such episodes, symptoms may disappear completely, although neurological problems often persist. MS is believed to follow a primary T-cell-controlled autoimmune reaction against myelin-related proteins. The cuprizone mouse model mimics certain facets of MS regarding de- and remyelination processes that are independent of contribution of the immune system. ADAM12 modulates the release of growth factors involved in MS pathomechanisms such as IGF1 (Loechel et al., 2000) and HBEGF (Asakura et al., 2002; Kodama et al., 2004). IGF1 is necessary for proper myelination and can protect mature oligodendrocytes from a pathological insult (D'Ercole et al., 1996; Garcia-Segura et al., 1996; Mason et al., 2000). Elevated IGF1 expression by hypertrophic astrocytes within the CC can be seen during acute cuprizone-induced demyelination (Acs et al., 2009; Komoly et al., 1992; Mason et al., 2000). IGF binding proteins (IGFBPs) can form biologically inactive complexes with IGF1 and interfere with its functions (Mukherjee et al., 2008; Rechler and Clemmons, 1998). ADAM12 is capable of cleaving IGFBP3 and IGFBP5 (Loechel et al., 2000) which are both expressed in the CNS (Honda et al., 2009; Iwadate et al., 2003). Hence, ADAM12 might release IGF1 from the IGFBP/IGF1 complex and, thus, promote remyelination during acute demyelination. The activity of HBEGF is closely linked to the enzymatic action of ADAM12 (Asakura et al., 2002). HBEGF has been reported to stimulate the proliferation of CNS astrocytes and multipotent progenitors and has been described as useful candidate molecule for brain repair strategies (Kornblum et al., 1999). Furthermore, HBEGF has been suggested to play a role in de- and remyelination processes. In mice that underwent toxic focal demyelination by lysolecithin injection, treatment with HBEGF by intranasal administration was associated with a significant increase of progenitor cell migration from the subventricular zone to the lesion site and a higher number of oligodendrocytes in the demyelinated regions (Cantarella et al., 2008). ADAM12 releases and thus converts HBEGF from its membrane-bound form, proHBEGF, to its soluble form, mature HBEGF, by means of ectodomain-shedding

9

(Asakura et al., 2002; Nishi and Klagsbrun, 2004). This process has been shown to be relevant in CNS disease (Kodama et al., 2004). Thus, ADAM12-mediated HBEGF shedding could also take place in the pathological state of demyelination. Endogenous ADAM12-released HBEGF might contribute to remyelination by promoting proliferation of oligodendrocyte progenitor cells and their migration to demyelinated brain regions. At present, however, it remains elusive which functions ADAM12 takes over during demyelination. Further studies are needed to understand the particular function of ADAM12 in the healthy and demyelinated brain. Other possible roles of ADAM12 in CNS disease derive from its cell-binding and intracellular signaling capacity and concern cell differentiation, proliferation, and migration (Kveiborg et al., 2008). It might be conceivable that ADAM12 regulates astroglial proliferation and migration to injured sites in the brain. Furthermore, it has been demonstrated that ADAM12 interacts with amyloid-beta and mediates the neurotoxic effects of this Alzheimer's disease-related protein (Harold et al., 2007; Malinin et al., 2005) and, therefore, might be deleterious. In summary, we have demonstrated that ADAM12 is more widely expressed by brain cells as hitherto assumed and regulated by distinct neuropathological factors particularly in astroglia. ADAM12 appears to be implicated in CNS diseases such as demyelinating disorders. Further studies have to show the function of ADAM12 under neuropathological conditions.

4.

Experimental procedures

4.1.

Animals and induction of demyelination

C57BL6 male mice (Harlan Winkelmann, Germany) were bred and maintained in a pathogen-free environment. Animals underwent routine cage maintenance once a week and microbiological monitoring according to the Federation of European Laboratory Animal Science Associations recommendations. Food and water were available ad libitum. Research and animal care procedures were approved by the Review Board for the Care of Animal Subjects of the district government (NordrheinWestfalen, Germany). Demyelination was induced by feeding 8-week-old (19–21 g) male mice a diet containing 0.2% cuprizone (bis-cyclohexanone oxaldihydrazone; Sigma-Aldrich Inc., USA) mixed into a ground standard rodent chow for five weeks.

4.2.

Tissue preparation

Tissue preparation was performed as previously described (Acs et al., 2009; Groebe et al., 2009; Norkute et al., 2009; Pott et al., 2009). For histological and immunohistochemical studies, mice were intracardially perfused with 4% paraformaldehyde containing picric acid after five weeks of cuprizone treatment. After overnight post-fixation, brains were dissected, embedded, and then coronary sectioned into 5 μm slices from intersections 185 and 195 according to the mouse brain atlas by Sidman et al. (http://www.hms.harvard.edu/research/brain/atlas.html). Sections were placed on silane-coated slides. For gene expression analysis, mice were killed by rapid decapitation. Brains were quickly removed, and the entire CC and cortex separately dissected using a stereo-microscopic approach. Tissues were

10

B RA IN RE S EA RCH 1 32 6 (2 0 1 0 ) 1 –1 4

immediately frozen in liquid nitrogen and kept at −80 °C until used.

4.3.

Immunohistochemistry (IHC)

For IHC, sections were rehydrated and unmasked by citrate buffer and heating, blocked with phosphate-buffered saline (PBS) containing 2% normal horse serum, and incubated overnight with the primary antibody diluted in blocking solution. Anti-PLP antibody (mouse IgG; Serotec, Germany; 1:500) was used to detect intact myelin, anti-APC antibody (mouse IgG; Calbiochem, Germany; 1:100) was used to detect late stage oligodendrocyte cell bodies, whereas anti-GFAP antibody (rabbit IgG; Encore, USA; 1:1000) was used to visualize astrocytes. For the localization of ADAM12 protein, antiADAM12 antibody (rabbit IgG; Biomol, Germany; 1:100) was used. After washing, sections were incubated with biotinylated secondary antibodies (Vector Laboratories, Burlingame, UK) for 1 h, followed by incubation with peroxidase-coupled avidin– biotin-complex (ABC kit, Vector Laboratories). The immuneprecipitated product was visualized with the ACE reaction. Stained and processed sections were documented with a Nikon Digital Sight DS 2Mv camera. Cell counting was performed as previously described (Groebe et al., 2009; Norkute et al., 2009; Pott et al., 2009).

4.4.

Immunofluorescence double labeling

For immunofluorescence double labeling, sections were rehydrated, unmasked by citrate buffer and heating, blocked with PBS containing 2% heat-inactivated fetal calf serum (FCS) (Gibco, Germany) and 1% bovine serum albumin (BSA), and incubated for 90 min with the indicated combination of primary antibodies diluted in blocking solution. AntiADAM12 (rabbit IgG; Biomol, Germany; 1:100) was either combined with anti-APC (mouse IgG; Calbiochem, Germany; 1:100) for detection of oligodendrocytes or SMI311 (mouse IgG/ IgM; Convance, USA; 1:100) for detection of neuronal cells and dendrites or anti-GFAP for detection of astrocytes (mouse IgG; Santa Cruz Biotechnology; 1:10). After washing, sections were incubated with a combination of fluorescent anti-mouse secondary antibodies (Alexa Fluor 488 donkey IgG; Invitrogen, Germany; 1:500) and fluorescent anti-rabbit secondary antibodies (Alexa Fluor 568 goat IgG; Invitrogen, Germany; 1:500) both diluted in blocking solution. Sections were then incubated with Hoechst 33342 (Invitrogen, Germany; 1:1000) diluted in PBS for the staining of cell nuclei. In order to exclude unspecific binding of the fluorescent secondary antibodies to primary antibodies, appropriate negative controls were performed by first incubating sections with the primary antibodies of murine origin (anti-APC or SMI311) and subsequently incubating these sections with fluorescent antirabbit secondary antibody. Sections exposed to anti-ADAM12 antibody of rabbit origin were incubated with fluorescent antimouse secondary antibody. Unspecific secondary antibody binding to the tissue itself was excluded by performing negative controls by incubating sections with each of the fluorescent secondary antibodies alone. Stained and processed sections were documented with the microscope working station Zeiss LSM 7 Duo.

4.5.

Cell culturing and treatment

Primary astrocyte cell cultures were prepared from one- to three-day-old Balb/c mice (Harlan Winkelmann GmbH, Germany) as previously described (Braun et al., 2009; Pawlak et al., 2005). Cerebral cortices were quickly dissected, meninges removed and dispersed in TrypLE Express (Invitrogen, Germany), filtered through a 50 μm nylon mesh, and centrifuged at 300 × g for 5 min. Cells were re-suspended in Dulbecco's modified Eagle's medium (DMEM) (Invitrogen, Germany) containing 20% heat-inactivated FCS (Invitrogen, Germany), streptomycin (0.5%; Invitrogen, Germany), and fungizone (0.1%; Invitrogen, Germany). Finally, cells were seeded on 100 mm plates coated with poly-ornithine (Sigma, Germany). Astrocytes were maintained and grown at 37 °C in a humidified atmosphere containing 5% CO2. Medium was replaced every other day. Upon reaching confluency, cells were trypsinized and replated at lower density. This procedure was repeated twice. The resulting final astroglia culture is characterized by >95% homogeneity of GFAP+ cells and no microglial or neuronal contamination (Pawlak et al., 2005). For gene expression and cell viability analysis, astrocyte cultures were treated with LPS (Invitrogen, Germany; 100 ng/ml), TNFα (Invitrogen, Germany; 100 ng/ml), Glut (Sigma, Germany; 4 µM) or H2O2 (Roth, Germany; 200 µM) for 24 h. The NFκB inhibitor curcumin (Biomol, Germany; 5 µM solved in dimethyl sulfoxide) was applied 1 h prior cell stimulation. Proper controls were conducted to exclude that inhibitor or solvent application alone affects ADAM12 or housekeeping reference gene expression in primary astrocyte cultures. Primary neuronal cultures were prepared from gestational day 15/16 Balb/c mouse embryos and cultured as previously described (Ivanova and Beyer, 2003; Kajta et al., 2006). Briefly, pregnant females were anesthetized with CO2 vapor, killed by cervical dislocation and subjected to caesarean section in order to dissect fetal brains. Cerebral cortices were dissected, minced into small pieces, digested with trypsin (0.1% for 15 min at room temperature; Sigma, Germany), triturated in the presence of 10% heat-inactivated FCS and DNAse I (170U per ml; Sigma, Germany) and, finally, centrifuged for five min at 100 ×g. Cells were then suspended in Neurobasal medium (Invitrogen, Germany) supplemented with B27 (Invitrogen, Germany) and plated at a density of 1.5 × 105 cells per cm2 onto 24 well plates (Greiner, Germany) coated with poly-ornithine. This procedure typically yields cultures that contain >90% neurons and