- Email: [email protected]
microglial interactions: control of microglial activation
Methods 29 (2003) 345–350 www.elsevier.com/locate/ymeth
Use of cocultured cell systems to elucidate chemokine-dependent neuronal/microglial interactions: control of microglial activation Violetta Zujovic* and Veronique Taupin CNS Research Department, Sanofi-Synthelabo, 92225 Bagneux Cedex, France Accepted 10 December 2002
Abstract In order to understand processes involved in central nervous system inflammatory diseases, a critical appreciation of mechanisms involved in the control of immune function in the brain is needed. Microglial cells are watchful eyes for unusual events and detecting the presence of pathogens but are also alert to signals emanating from damaged neurons. Fractalkine (CX3CL1) is a chemokine which is expressed predominantly in the central nervous system, being localized on neurons, while its receptor, CX3CR1, is found on microglial cells. We have developed a strategy to investigate the role of this chemokine in neuronal–microglia interactions. Because fractalkine is expressed both as a soluble and as a membrane-attached protein, we have established various protocols involving different levels of cell-to-cell communication. Three experimental systems were instituted, including (1) a conditioned medium transfer system in which no cell–cell communication or contact is possible, (2) a transwell system that permits cell-contact-independent communication through diffusible soluble factors only, and (3) a coculture system allowing cell-to-cell communication via direct microglial–neuronal contacts. Using these in vitro cocultured systems, we have investigated the role of a soluble and/or cellassociated chemokine, such as fractalkine, in order to obtain insights into the role of glia–neuron interactions in cerebral inflammation. Ó 2003 Elsevier Science (USA). All rights reserved. Keywords: Fractalkine; CX3CR1; Neurons; Microglia; Coculture; LPS; Neurotoxicity
1. Introduction While the central nervous system (CNS) has long been considered an immune-privileged organ, it is now well recognized that it is under constant immune surveillance. In a healthy brain, few signs of immune activity are detected. However, numerous autoimmune or neurodegenerative diseases are associated with ‘‘inappropriate’’ inflammation in the CNS [1]. Microglial cells play an important role in mechanisms of brain immune surveillance [2]. In the normal brain, quiescent microglia present little immune-like activity but they can differentiate into efficient macrophages * Corresponding author. Present address: Department of Pharmacology and Therapeutics, Box 100267, University of Florida College of Medicine, Gainesville, FL 32610-0267, USA. Fax: +352-392-9696. E-mail address: [email protected]fl.edu (V. Zujovic).
expressing major histocompatibility complex molecules and producing a variety of inflammatory cytokines. Various signals can trigger microglial activation, including pathogens and damaged or dead brain cells. In healthy brain, diverse mechanisms also maintain microglial cells in a dormant state. Several studies demonstrate that neurons inhibit microglial immune functions through the production of neuropeptides, neurotransmitters, and adhesion molecules [3–6]. Chemokines are small secreted proteins that play a major role in the immune system. Most chemokines are expressed at low levels in normal brain yet are upregulated during inflammatory processes in the CNS [7]. Numerous studies suggest a role for these molecules in microglial cell recruitment [2] but their function in the control of microglial activation is still unclear. Only two chemokine/chemokine receptor systems are present constitutively in the normal brain: stromal cell-derived
1046-2023/03/$ - see front matter Ó 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S1046-2023(02)00358-4
346
V. Zujovic, V. Taupin / Methods 29 (2003) 345–350
factor 1, or CXCL12, and fractalkine (CX3CL1). The majority of chemokines are localized on astrocytes or microglial cells. However, fractalkine is expressed principally by neurons while its receptor is found on microglial cells [8–10]. This chemokine exhibits another particular property. It exists as either a membranebound protein or a soluble chemokine [11,12]. The expression of soluble forms of fractalkine by neurons is increased in vitro after an excitotoxic insult [13] and in vivo after facial nerve axotomy [8]. Recently, fractalkine and its receptor, CX3CR1, have also been shown to be upregulated following ischemic brain injury in rats [14].
2. Strategies In order to elucidate the role of fractalkine in neuronal glial interactions, we have established different protocols, including neuronal microglial cocultures. Three experimental systems were developed, including (1) a conditioned medium transfer system in which no cell–cell communication or contact is possible, (2) a transwell system that permits cell-contact-independent communication through diffusible soluble factors only,
and (3) a coculture system allowing cell-to-cell communication via direct microglial–neuronal contacts. The transmembrane versus soluble form paradigm is a common feature in the cytokine superfamily. Tumor necrosis factor a (TNFa) family members and two chemokines (CX3CL1, CXCL16) exhibit structurally distinct membrane-attached and shed/soluble forms that display different biological activities. The functional significance of transmembrane versus soluble forms of fractalkine is an important factor needing further clarification, particularly in regards to their role in the CNS. Thus various coculture strategies were developed to study the role of fractalkine in microglial activation and its consequences, namely neuronal cell death (Fig. 1). These approaches can be extended to studies of other molecules with similar structural and functional properties. 2.1. Conditioned medium transfer The conditioned medium transfer paradigm (Fig. 1A) consists of culturing microglial cells and neurons separately. The conditioned medium of microglial cells, prestimulated with various agents, is then transferred to
Fig. 1. Schematic representation of the different types of cocultures. (A) Conditioned medium transfer, (B) transwell coculture, (C) direct coculture.
V. Zujovic, V. Taupin / Methods 29 (2003) 345–350
the neuronal cell culture. Since activated microglial cells secrete soluble factors, the effect of microglial conditioned medium on neuronal survival can be investigated in order to study the influence of microglial cell-derived soluble factors on this and other responses of neuronal cells. We were interested in determining the role of soluble forms of fractalkine on microglial production of neurotoxic factors. However, using this approach, we were unable to detect any neurotoxicity after the addition of the microglial cell conditioned medium. 2.2. Transwell coculture This paradigm (Fig. 1B) consisted in culturing microglial cells in a transwell system that was placed above the neuronal layer. The neurons and microglial cells share the same medium but no direct cell–cell interactions are possible due to the physical separation of the cells by a polycarbonate membrane. The pore size of the transwell (0.4 lm) allows no cell migration through the membrane. Microglial cells were then activated. In the presence of inserts containing microglia on top of the neuronal layer, we observed the effects of diffusible factors coming from both the neurons and the microglia. Compared to the conditioned medium paradigm, the transwell system adds an additional dimension to the system since neurons can respond to microglial activation notably by releasing soluble factors. Among these factors fractalkine is a candidate, since after different stress episodes neurons have been shown to release the soluble form of fractalkine [9,14].
347
2.3. Direct coculture The use of a direct coculture system was necessary since fractalkine can also exist as a transmembrane protein (Fig. 1C). This approach allows one to study cell–cell interactions and to compare the effects of both membrane-bound and soluble forms of fractalkine. Because we were interested in elucidating the effect of endogenous expressed fractalkine, we pretreated the neurons for 3 h with an anti-fractalkine antibody before the microglial cells were added. Then, microglial cells were activated. The effect of endogenous fractalkine on microglial activation and neuronal death was evaluated. 2.4. Quantitation of neuronal cell death Various methods are available to quantify neuronal death. The most common approaches are: (1) measurement of cell activity evaluated by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay, (2) detection of fragmented or condensed nuclei representing neuronal cells undergoing apoptosis (using the Hoechst method), or (3) measurement of lactate dehydrogenase activity. Since some of the microglial cells die during activation, these markers are not specific enough to quantify only neuronal cell death. There are a number of neuronal markers that are commercially available, notably antibodies raised against specific neuronal proteins, including microtubule-associated protein-2 (MAP2) and neuronal nuclear protein NeuN [20]. Fig. 2 shows the immunostaining of neurons we obtained with these
Fig. 2. Specific neuronal immunostaining. Cultures were fixed in 4% paraformaldehyde and immunostained with anti-MAP-2 antibody (1:400 dilution) (A) or mouse anti-NeuN antibody (1:1000 dilution) (B).
348
V. Zujovic, V. Taupin / Methods 29 (2003) 345–350
two antibodies. The anti-NeuN antibody provided better definition and separation between the cells compared to the neuronal staining with the anti-MAP-2 antibody. Since it was desirable to develop an automatic counting system, the anti-NeuN antibody staining was more suitable for this purpose. Microglial cells can be activated by diverse factors such as bacterial endotoxin lipopolysaccharides (LPS) [15], human immunodeficiency virus envelope protein (gp120) [16], and toxic proteins from the CNS (e.g., bamyloid) [17]. Among them, LPS is the most potent activator of microglial cells. After LPS treatment, microglial cells produce various neurotoxic factors such as TNFa [1,18] and free radicals [19]. Therefore, one consequence of microglial activation is neuronal death. Preliminary experiments were conducted in order to determine in each different coculture system the concentration of LPS to use (Figs. 3 and 4). Neuronal death was observed only in the transwell and the direct co-
culture system. Fig. 3 shows a representative experiment of the effect of LPS on neuronal survival in the transwell coculture system. It is noted that LPS had no effect on neuronal survival in the absence of microglia (mentioned as ‘‘without transwell’’ in Fig. 3). A preliminary experiment was also conducted in order to establish the concentration of LPS to use in the direct coculture condition (Fig. 4). In the direct coculture paradigm, significantly lower concentrations of LPS produced neuronal cell death. Thus, the concentration of LPS utilized in the experiment was dependent on the specific coculture paradigm used.
3. Specific methods 3.1. Animals Pregnant female Sprague–Dawley rats (CERJ, France) were bred until embryos were 18 days of age (E18). Hippocampi of E18 embryos were isolated for primary neuronal culture. Cerebral cortices of newborn rats were used to establish microglial cultures. All dissections were performed in LeibovitzÕs L15 medium (Gibco Life Technologies, Cergy Pontoise, France). 3.2. Hippocampal neuronal culture
Fig. 3. Effect of LPS treatment on neuronal survival in the transwell paradigm. Microglial cells were plated in the transwell chambers and the chambers were placed on top of the wells containing the neuronal culture. Different concentrations of LPS were applied to neurons cultured alone or in transwell coculture with microglial cells. After 24 h of LPS treatment, neuronal survival was quantified using NeuN immunostaining. This is a representative experiment.
Fig. 4. Effect of LPS treatment on neuronal survival in the direct coculture paradigm. Isolated microglial cells were seeded into wells containing neurons from 5-day-old culture; 3 h after plating, microglial cells were treated for 24 h with different concentrations of LPS. Neuronal survival was then quantified using NeuN immunostaining. This is a representative experiment.
After the brain was isolated, meninges were removed and hippocampi dissected out. All the hippocampi were transferred in L15 culture medium. After centrifugation (150g, 4 min), hippocampal tissue was incubated in 20 ml of phosphate-buffered saline (PBS) (Gibco Life Technologies) containing glucose (0.6%), penicillin (50 U/ml), streptomycin (50 lg/ml), 0.5 mg/ml trypsin, and 0.1 mg/ml DNase (Sigma Chemical Co., Saint Quentin Fallavier, France). The tissue was digested for 10 min in a 37 °C water bath. The tube was gently shaken every 2 min. Subsequently, 0.1 mg/ml trypsin inhibitor was added, and after a centrifugation at 150g for 3 min, the tissue was rinsed with PBS containing glucose, antibiotics, and 0.1% bovine serum albumin. Tissue pieces were dissociated mechanically by gentle trituration through a flame-polished Pasteur pipette. After centrifugation, the pellets were suspended in LeibovitzÕs L15 medium (Gibco Life Technologies) supplemented with nutrients, insulin (10 lg/ml), transferrin (20 lg/ml), putrescein (0.1 mM), progesterone (20 nM), and sodium selenite (30 nM); all reagents were from Sigma Chemical Co. Cells were grown at 37 °C in 5% (v/v) CO2 for 5 days in 96-well Nunc dishes (5 104 cells/well) or 24-well Nunc dishes (105 cells/well) coated with poly-D -lysine (5 lg/ ml) (Sigma Chemical Co.). Under these culture conditions, neurons represented 98% of the cell population [21].
V. Zujovic, V. Taupin / Methods 29 (2003) 345–350
3.3. Microglial cells Once the brain was extracted, the meninges were removed and whole cortices were isolated in LeibovitzÕs L15 medium (Gibco Life Technologies). The tissue was then dissociated mechanically through a nylon membrane (cell strainer with pore size of 70 lm; Falcon, Pittsburgh, PA, USA). After centrifugation, the pellet was suspended in DulbeccoÕs modified EagleÕs medium (Gibco), 10% heat-inactivated fetal calf serum (Myoclone; Gibco), and 1% gentamicin (Gibco Life Technologies). The homogenate was then plated into 75-cm2 TPP flasks (ATGC; Noisy le Grand, France). Cells were grown at 37 °C in 5% CO2 , and medium was changed at day 5 of the culture. After 2 weeks of culture, flasks were gently shaken and the supernatant containing the microglia was collected. After centrifugation, cells were suspended in hippocampal neuronal medium (see above) and used as described further. 3.4. Conditioned medium transfer Microglial cells were plated into 96-well dishes (105 cells/well) and were treated after a 3-h adherence period. The cells were stimulated with or without LPS (10 ng/ ml) in the presence or absence of different concentrations of fractalkine. Cell-free supernatants were collected after 24 h of treatment and were added to 5day-old neurons. After 24 h of conditioned medium treatment, neuronal survival was quantified (as described below). 3.5. Transwell Neurons (105 cells/well) were plated in 24-well dishes. Freshly isolated microglia were plated (5 104 cells/ well) in the transwell chambers (Costar, France) and these inserts were placed on top of the wells containing the neuronal culture. The polycarbonate membrane in the transwell had a 0.4-lm pore size that prevents both cell–cell contact and cell migration but allows the diffusion of soluble factors. Microglial cells were then treated with different concentrations of LPS. After 24 h of LPS treatment, neuronal survival was quantified (as described below). The following experiments were run with a concentration of LPS of 0.1 and 1 lg/ml with or without different concentrations of fractalkine. 3.6. Direct coculture Isolated microglial cells were resuspended in hippocampal neuronal medium and seeded (2:5 104 cells/ well) into wells containing neurons from 5-day-old cultures in a 96-well plate. Three hours after plating, microglial cells were treated for 24 h with different concentrations of LPS. Neuronal survival was then
349
quantified. In order to establish if fractalkine had a direct effect on neurotoxicity induced by LPS-activated microglial cells, the cells were treated with LPS (1 ng/ml) in the presence or absence of different concentrations of fractalkine. In order to determine the role of endogenous neuronal fractalkine, neurons were pretreated for 3 h with anti-fractalkine antibody (10 lg/ml) (TEBU, Le Perray en Yvelines, France) before microglial cells were added. Three hours after plating, microglial cells were treated with or without LPS (from 0.1 to 10 ng/ml). After 24 h of LPS treatment, neuronal survival was quantified. 3.7. Quantification of neuronal survival The effect of microglial stimulation on the neuronal cultures was estimated by immunostaining for NeuN, a neuron-specific nuclear protein. Cultures were fixed in 4% paraformaldehyde at room temperature for 15 min and rinsed three times for 5 min with PBS containing 0.3% Triton X-100. Culture were immunostained with mouse anti-NeuN antibody (Chemicon International, Temecula, CA, USA) at a dilution of 1:1000 at room temperature for 2 h [20]. After being washed, cells were incubated for 30 min at room temperature with a Cy3-conjugated anti-mouse IgG secondary antibody (Sigma Chemical Co.). The same protocol was used for MAP-2 immunostaining. AntiMAP-2 antibody was used at 1:400 (Roche Diagnostics, Meylan, France). Neuronal survival was quantified by counting NeuN-positive cells. Counting analysis of cells was performed by using an automatic analysis system (Trakcell) designed by Biocom (Les Ulis, France).
4. Concluding remarks It is well established that microglial cells can be toxic to neurons through cell–cell contact-dependent mechanisms [22]. These observations were confirmed in our study with direct microglial neuronal coculture in which neuronal death was significant in the presence of low concentrations of LPS. In the conditioned medium transfer system, soluble factors released by microglial cells after treatment of microglia with similar concentrations of LPS are not sufficient to induce a neurotoxic cascade. Even in the transwell system, high concentrations of LPS were necessary to induce neurotoxicity. At these high LPS concentrations modulation of microglial activation, by other factors, is difficult. The fact that we did not observe any modulatory response in these conditions could reflect the need for various soluble signals, as has been shown in previous studies, e.g. interferon-c plus LPS or a cocktail of cytokines [23]. It should also be noted that the type of neuronal culture could have an
350
V. Zujovic, V. Taupin / Methods 29 (2003) 345–350
influence on the sensitivity of the cells to LPS treatment. Hippocampal, cortical, or mesencephalic neurons have different properties that may influence their responses to microglial-derived toxic products. The stage of neuronal maturation may also be of importance (embryonic versus neonatal-derived neuronal cultures). It is worthwhile to point out that LPS has no direct effect on neuronal cell death. However, other microglial cell activators such as b-amyloid can be directly neurotoxic [24]. In this scenario, the conditioned medium transfer paradigm would be the more appropriate method for studying the influence of other factors on bamyloid-stimulated microglial-dependent neurotoxicity. We have previously shown that soluble fractalkine is able to inhibit TNFa production by LPS-activated microglia [10]. The second step was to establish the effect of fractalkine on microglia-induced neurotoxicity. The conditioned medium transfer and transwell systems were not convenient since in the former approach we observed no neurotoxicity and in the latter paradigm the concentrations of LPS were too high; the concentration of LPS sufficient to produce neurotoxicity decreases with increased cell-to-cell communication. The direct coculture system provided us an ideal experimental paradigm to study this chemokine present in the medium as soluble and as neuronal membrane attached. Exogenous fractalkine treatment did not provide any neuroprotective effects. Since neurons are the principal source of fractalkine in the brain, we included the strategy of blocking endogenously synthesized fractalkine using a specific anti-fractalkine antibody. With this approach, we observed an increase in neuronal death, suggesting a role for fractalkine in the control of microglial activation and the consequent neuronal cell death. The involvement of constitutive fractalkine in the control of cerebral inflammatory responses was also emphasized by our results from an in vivo model of brain inflammation [25]. These strategies were developed to study the in vitro role of a specific chemokine: fractalkine. Each strategy provides unique opportunities for better understanding of the function of each form of the molecule. With the conditioned medium transfer system, the direct action of the soluble molecule on microglial secretion of toxic factors is observed. The transwell system adds another dimension to the study since the neurons can respond to microglial activation by secreting soluble factors. The coculture system permits observations of the actions of both soluble and membrane-attached forms. This mul-
tifaceted approach can be extended to any molecule expressed as either soluble or transmembrane protein. References [1] M. Spranger, A. Fontana, Neuroscientist 2 (1996) 293–299. [2] F. Aloisi, Glia 36 (2001) 165–179. [3] R. Delgado, A. Carlin, L. Airaghi, M.T. Demitri, L. Meda, D. Galimberti, P. Baron, J.M. Lipton, A. Catania, J. Leukocyte Biol. 63 (1998) 740–745. [4] W.K. Kim, Y. Kan, D. Ganea, R.P. Hart, J. Gozes, G.M. Jonakait, J. Neurosci. 2 (2000) 3622–3630. [5] L.M. Phillips, L.A. Lampson, J. Neuropathol. Exp. Neurol. 59 (2000) 177–187. [6] R.C. Chang, P. Hudson, L. Hadden, J.-S. Hong, Brain Res. 853 (2000) 236–244. [7] J. Hesselgesser, R.J. Horuk, Neurovirology 5 (1999) 13–26. [8] J.K. Harrison, Y. Jiang, S. Chen, Y. Xia, D. Maciejewski, R.K. McNamara, W.J. Streit, M.N. Salafranca, S. Adhikari, D.A. Thompson, P. Botti, K.B. Bacon, L. Feng, Proc. Natl. Acad. Sci. USA 95 (1998) 10896–10901. [9] A. Nishiyori, M. Minami, Y. Ohtani, S. Takami, J. Yamamoto, N. Kawaguchi, T. Kume, A. Akaike, M. Satoh, FEBS Lett. 429 (1998) 167–172. [10] V. Zujovic, J. Benavides, X. Vige, C. Carter, V. Taupin, Glia 29 (2000) 305–315. [11] J.F. Bazan, K.B. Bacon, G. Hardiman, W. Wang, K. Soo, D. Rossi, D.R. Greaves, Z. Zlotnik, T.J. Schall, Nature 385 (1997) 640–644. [12] Y. Pan, C. Lloyd, H. Zhou, S. Dolich, J. Deeds, J.A. Gonzalo, J. Vath, M. Gosselin, J. Ma, B. Dussault, E. Woolf, G. Alperin, J. Culpepper, J.C. Gutierrez-Ramos, D. Gearing, Nature 387 (1997) 611–617. [13] G.A. Chapman, K. Moores, D. Harrison, C.A. Campbell, B.R. Stewart, P.J. Strijbos, J. Neurosci. 20 (2000) RC87. [14] G. Tarozzo, M. Campanella, M. Ghiani, A. Bulfone, M. Beltramo, Eur. J. Neurosci. Res. 15 (2002) 1663–1668. [15] D.M. Bronstein, I. Perez-Otano, V. Sun, S.B. Mullis Sawin, J. Chan, G.C. Wu, P.M. Hudson, L.Y. Kong, J.S. Hong, M.K. McMillian, Brain Res. 704 (1995) 112–116. [16] J.E. Merrill, Y. Koyanagi, J. Zack, L. Thomas, F. Martin, I.S. Chen, J. Virol. 66 (1992) 2217–2225. [17] L. Meda, M.A. Cassatella, G.I. Szendrei, L. Otvos Jr., P. Baron, M. Villalba, D. Ferrari, F. Rossi, Nature 374 (1995) 647–650. [18] E.N. Benveniste, J. Mol. Med. 75 (1997) 165–173. [19] R.B. Banati, J. Gehrmann, P. Schubert, G.W. Kreutzberg, Glia 7 (1993) 111–118. [20] R.J. Mullen, C.R. Buck, A.M. Smith, Development 116 (1992) 201–211. [21] M.B. Emerit, J. Segovia, H. Alho, M.J. Mastrangelo Jr., B.C. Wise, J. Neurochem. 52 (1989) 952–961. [22] J.M. Frade, Y.A. Barde, Neuron 20 (1998) 35–41. [23] J.M. Pocock, A.C. Liddle, Prog. Brain Res. 132 (2001) 555–565. [24] T. Harkany, I. Abraham, C. Konya, C. Nyakas, M. Zarandi, B. Penke, P.G. Luiten, Rev. Neurosci. 11 (2000) 329–382. [25] V. Zujovic, N. Schussler, D. Jourdain, D. Duverger, V. Taupin, J. Neuroimmunol. 115 (2001) 135–143.
Use of cocultured cell systems to elucidate chemokine-dependent neuronal/microglial interactions: control of microglial activation Violetta Zujovic* and Veronique Taupin CNS Research Department, Sanofi-Synthelabo, 92225 Bagneux Cedex, France Accepted 10 December 2002
Abstract In order to understand processes involved in central nervous system inflammatory diseases, a critical appreciation of mechanisms involved in the control of immune function in the brain is needed. Microglial cells are watchful eyes for unusual events and detecting the presence of pathogens but are also alert to signals emanating from damaged neurons. Fractalkine (CX3CL1) is a chemokine which is expressed predominantly in the central nervous system, being localized on neurons, while its receptor, CX3CR1, is found on microglial cells. We have developed a strategy to investigate the role of this chemokine in neuronal–microglia interactions. Because fractalkine is expressed both as a soluble and as a membrane-attached protein, we have established various protocols involving different levels of cell-to-cell communication. Three experimental systems were instituted, including (1) a conditioned medium transfer system in which no cell–cell communication or contact is possible, (2) a transwell system that permits cell-contact-independent communication through diffusible soluble factors only, and (3) a coculture system allowing cell-to-cell communication via direct microglial–neuronal contacts. Using these in vitro cocultured systems, we have investigated the role of a soluble and/or cellassociated chemokine, such as fractalkine, in order to obtain insights into the role of glia–neuron interactions in cerebral inflammation. Ó 2003 Elsevier Science (USA). All rights reserved. Keywords: Fractalkine; CX3CR1; Neurons; Microglia; Coculture; LPS; Neurotoxicity
1. Introduction While the central nervous system (CNS) has long been considered an immune-privileged organ, it is now well recognized that it is under constant immune surveillance. In a healthy brain, few signs of immune activity are detected. However, numerous autoimmune or neurodegenerative diseases are associated with ‘‘inappropriate’’ inflammation in the CNS [1]. Microglial cells play an important role in mechanisms of brain immune surveillance [2]. In the normal brain, quiescent microglia present little immune-like activity but they can differentiate into efficient macrophages * Corresponding author. Present address: Department of Pharmacology and Therapeutics, Box 100267, University of Florida College of Medicine, Gainesville, FL 32610-0267, USA. Fax: +352-392-9696. E-mail address: [email protected]fl.edu (V. Zujovic).
expressing major histocompatibility complex molecules and producing a variety of inflammatory cytokines. Various signals can trigger microglial activation, including pathogens and damaged or dead brain cells. In healthy brain, diverse mechanisms also maintain microglial cells in a dormant state. Several studies demonstrate that neurons inhibit microglial immune functions through the production of neuropeptides, neurotransmitters, and adhesion molecules [3–6]. Chemokines are small secreted proteins that play a major role in the immune system. Most chemokines are expressed at low levels in normal brain yet are upregulated during inflammatory processes in the CNS [7]. Numerous studies suggest a role for these molecules in microglial cell recruitment [2] but their function in the control of microglial activation is still unclear. Only two chemokine/chemokine receptor systems are present constitutively in the normal brain: stromal cell-derived
1046-2023/03/$ - see front matter Ó 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S1046-2023(02)00358-4
346
V. Zujovic, V. Taupin / Methods 29 (2003) 345–350
factor 1, or CXCL12, and fractalkine (CX3CL1). The majority of chemokines are localized on astrocytes or microglial cells. However, fractalkine is expressed principally by neurons while its receptor is found on microglial cells [8–10]. This chemokine exhibits another particular property. It exists as either a membranebound protein or a soluble chemokine [11,12]. The expression of soluble forms of fractalkine by neurons is increased in vitro after an excitotoxic insult [13] and in vivo after facial nerve axotomy [8]. Recently, fractalkine and its receptor, CX3CR1, have also been shown to be upregulated following ischemic brain injury in rats [14].
2. Strategies In order to elucidate the role of fractalkine in neuronal glial interactions, we have established different protocols, including neuronal microglial cocultures. Three experimental systems were developed, including (1) a conditioned medium transfer system in which no cell–cell communication or contact is possible, (2) a transwell system that permits cell-contact-independent communication through diffusible soluble factors only,
and (3) a coculture system allowing cell-to-cell communication via direct microglial–neuronal contacts. The transmembrane versus soluble form paradigm is a common feature in the cytokine superfamily. Tumor necrosis factor a (TNFa) family members and two chemokines (CX3CL1, CXCL16) exhibit structurally distinct membrane-attached and shed/soluble forms that display different biological activities. The functional significance of transmembrane versus soluble forms of fractalkine is an important factor needing further clarification, particularly in regards to their role in the CNS. Thus various coculture strategies were developed to study the role of fractalkine in microglial activation and its consequences, namely neuronal cell death (Fig. 1). These approaches can be extended to studies of other molecules with similar structural and functional properties. 2.1. Conditioned medium transfer The conditioned medium transfer paradigm (Fig. 1A) consists of culturing microglial cells and neurons separately. The conditioned medium of microglial cells, prestimulated with various agents, is then transferred to
Fig. 1. Schematic representation of the different types of cocultures. (A) Conditioned medium transfer, (B) transwell coculture, (C) direct coculture.
V. Zujovic, V. Taupin / Methods 29 (2003) 345–350
the neuronal cell culture. Since activated microglial cells secrete soluble factors, the effect of microglial conditioned medium on neuronal survival can be investigated in order to study the influence of microglial cell-derived soluble factors on this and other responses of neuronal cells. We were interested in determining the role of soluble forms of fractalkine on microglial production of neurotoxic factors. However, using this approach, we were unable to detect any neurotoxicity after the addition of the microglial cell conditioned medium. 2.2. Transwell coculture This paradigm (Fig. 1B) consisted in culturing microglial cells in a transwell system that was placed above the neuronal layer. The neurons and microglial cells share the same medium but no direct cell–cell interactions are possible due to the physical separation of the cells by a polycarbonate membrane. The pore size of the transwell (0.4 lm) allows no cell migration through the membrane. Microglial cells were then activated. In the presence of inserts containing microglia on top of the neuronal layer, we observed the effects of diffusible factors coming from both the neurons and the microglia. Compared to the conditioned medium paradigm, the transwell system adds an additional dimension to the system since neurons can respond to microglial activation notably by releasing soluble factors. Among these factors fractalkine is a candidate, since after different stress episodes neurons have been shown to release the soluble form of fractalkine [9,14].
347
2.3. Direct coculture The use of a direct coculture system was necessary since fractalkine can also exist as a transmembrane protein (Fig. 1C). This approach allows one to study cell–cell interactions and to compare the effects of both membrane-bound and soluble forms of fractalkine. Because we were interested in elucidating the effect of endogenous expressed fractalkine, we pretreated the neurons for 3 h with an anti-fractalkine antibody before the microglial cells were added. Then, microglial cells were activated. The effect of endogenous fractalkine on microglial activation and neuronal death was evaluated. 2.4. Quantitation of neuronal cell death Various methods are available to quantify neuronal death. The most common approaches are: (1) measurement of cell activity evaluated by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay, (2) detection of fragmented or condensed nuclei representing neuronal cells undergoing apoptosis (using the Hoechst method), or (3) measurement of lactate dehydrogenase activity. Since some of the microglial cells die during activation, these markers are not specific enough to quantify only neuronal cell death. There are a number of neuronal markers that are commercially available, notably antibodies raised against specific neuronal proteins, including microtubule-associated protein-2 (MAP2) and neuronal nuclear protein NeuN [20]. Fig. 2 shows the immunostaining of neurons we obtained with these
Fig. 2. Specific neuronal immunostaining. Cultures were fixed in 4% paraformaldehyde and immunostained with anti-MAP-2 antibody (1:400 dilution) (A) or mouse anti-NeuN antibody (1:1000 dilution) (B).
348
V. Zujovic, V. Taupin / Methods 29 (2003) 345–350
two antibodies. The anti-NeuN antibody provided better definition and separation between the cells compared to the neuronal staining with the anti-MAP-2 antibody. Since it was desirable to develop an automatic counting system, the anti-NeuN antibody staining was more suitable for this purpose. Microglial cells can be activated by diverse factors such as bacterial endotoxin lipopolysaccharides (LPS) [15], human immunodeficiency virus envelope protein (gp120) [16], and toxic proteins from the CNS (e.g., bamyloid) [17]. Among them, LPS is the most potent activator of microglial cells. After LPS treatment, microglial cells produce various neurotoxic factors such as TNFa [1,18] and free radicals [19]. Therefore, one consequence of microglial activation is neuronal death. Preliminary experiments were conducted in order to determine in each different coculture system the concentration of LPS to use (Figs. 3 and 4). Neuronal death was observed only in the transwell and the direct co-
culture system. Fig. 3 shows a representative experiment of the effect of LPS on neuronal survival in the transwell coculture system. It is noted that LPS had no effect on neuronal survival in the absence of microglia (mentioned as ‘‘without transwell’’ in Fig. 3). A preliminary experiment was also conducted in order to establish the concentration of LPS to use in the direct coculture condition (Fig. 4). In the direct coculture paradigm, significantly lower concentrations of LPS produced neuronal cell death. Thus, the concentration of LPS utilized in the experiment was dependent on the specific coculture paradigm used.
3. Specific methods 3.1. Animals Pregnant female Sprague–Dawley rats (CERJ, France) were bred until embryos were 18 days of age (E18). Hippocampi of E18 embryos were isolated for primary neuronal culture. Cerebral cortices of newborn rats were used to establish microglial cultures. All dissections were performed in LeibovitzÕs L15 medium (Gibco Life Technologies, Cergy Pontoise, France). 3.2. Hippocampal neuronal culture
Fig. 3. Effect of LPS treatment on neuronal survival in the transwell paradigm. Microglial cells were plated in the transwell chambers and the chambers were placed on top of the wells containing the neuronal culture. Different concentrations of LPS were applied to neurons cultured alone or in transwell coculture with microglial cells. After 24 h of LPS treatment, neuronal survival was quantified using NeuN immunostaining. This is a representative experiment.
Fig. 4. Effect of LPS treatment on neuronal survival in the direct coculture paradigm. Isolated microglial cells were seeded into wells containing neurons from 5-day-old culture; 3 h after plating, microglial cells were treated for 24 h with different concentrations of LPS. Neuronal survival was then quantified using NeuN immunostaining. This is a representative experiment.
After the brain was isolated, meninges were removed and hippocampi dissected out. All the hippocampi were transferred in L15 culture medium. After centrifugation (150g, 4 min), hippocampal tissue was incubated in 20 ml of phosphate-buffered saline (PBS) (Gibco Life Technologies) containing glucose (0.6%), penicillin (50 U/ml), streptomycin (50 lg/ml), 0.5 mg/ml trypsin, and 0.1 mg/ml DNase (Sigma Chemical Co., Saint Quentin Fallavier, France). The tissue was digested for 10 min in a 37 °C water bath. The tube was gently shaken every 2 min. Subsequently, 0.1 mg/ml trypsin inhibitor was added, and after a centrifugation at 150g for 3 min, the tissue was rinsed with PBS containing glucose, antibiotics, and 0.1% bovine serum albumin. Tissue pieces were dissociated mechanically by gentle trituration through a flame-polished Pasteur pipette. After centrifugation, the pellets were suspended in LeibovitzÕs L15 medium (Gibco Life Technologies) supplemented with nutrients, insulin (10 lg/ml), transferrin (20 lg/ml), putrescein (0.1 mM), progesterone (20 nM), and sodium selenite (30 nM); all reagents were from Sigma Chemical Co. Cells were grown at 37 °C in 5% (v/v) CO2 for 5 days in 96-well Nunc dishes (5 104 cells/well) or 24-well Nunc dishes (105 cells/well) coated with poly-D -lysine (5 lg/ ml) (Sigma Chemical Co.). Under these culture conditions, neurons represented 98% of the cell population [21].
V. Zujovic, V. Taupin / Methods 29 (2003) 345–350
3.3. Microglial cells Once the brain was extracted, the meninges were removed and whole cortices were isolated in LeibovitzÕs L15 medium (Gibco Life Technologies). The tissue was then dissociated mechanically through a nylon membrane (cell strainer with pore size of 70 lm; Falcon, Pittsburgh, PA, USA). After centrifugation, the pellet was suspended in DulbeccoÕs modified EagleÕs medium (Gibco), 10% heat-inactivated fetal calf serum (Myoclone; Gibco), and 1% gentamicin (Gibco Life Technologies). The homogenate was then plated into 75-cm2 TPP flasks (ATGC; Noisy le Grand, France). Cells were grown at 37 °C in 5% CO2 , and medium was changed at day 5 of the culture. After 2 weeks of culture, flasks were gently shaken and the supernatant containing the microglia was collected. After centrifugation, cells were suspended in hippocampal neuronal medium (see above) and used as described further. 3.4. Conditioned medium transfer Microglial cells were plated into 96-well dishes (105 cells/well) and were treated after a 3-h adherence period. The cells were stimulated with or without LPS (10 ng/ ml) in the presence or absence of different concentrations of fractalkine. Cell-free supernatants were collected after 24 h of treatment and were added to 5day-old neurons. After 24 h of conditioned medium treatment, neuronal survival was quantified (as described below). 3.5. Transwell Neurons (105 cells/well) were plated in 24-well dishes. Freshly isolated microglia were plated (5 104 cells/ well) in the transwell chambers (Costar, France) and these inserts were placed on top of the wells containing the neuronal culture. The polycarbonate membrane in the transwell had a 0.4-lm pore size that prevents both cell–cell contact and cell migration but allows the diffusion of soluble factors. Microglial cells were then treated with different concentrations of LPS. After 24 h of LPS treatment, neuronal survival was quantified (as described below). The following experiments were run with a concentration of LPS of 0.1 and 1 lg/ml with or without different concentrations of fractalkine. 3.6. Direct coculture Isolated microglial cells were resuspended in hippocampal neuronal medium and seeded (2:5 104 cells/ well) into wells containing neurons from 5-day-old cultures in a 96-well plate. Three hours after plating, microglial cells were treated for 24 h with different concentrations of LPS. Neuronal survival was then
349
quantified. In order to establish if fractalkine had a direct effect on neurotoxicity induced by LPS-activated microglial cells, the cells were treated with LPS (1 ng/ml) in the presence or absence of different concentrations of fractalkine. In order to determine the role of endogenous neuronal fractalkine, neurons were pretreated for 3 h with anti-fractalkine antibody (10 lg/ml) (TEBU, Le Perray en Yvelines, France) before microglial cells were added. Three hours after plating, microglial cells were treated with or without LPS (from 0.1 to 10 ng/ml). After 24 h of LPS treatment, neuronal survival was quantified. 3.7. Quantification of neuronal survival The effect of microglial stimulation on the neuronal cultures was estimated by immunostaining for NeuN, a neuron-specific nuclear protein. Cultures were fixed in 4% paraformaldehyde at room temperature for 15 min and rinsed three times for 5 min with PBS containing 0.3% Triton X-100. Culture were immunostained with mouse anti-NeuN antibody (Chemicon International, Temecula, CA, USA) at a dilution of 1:1000 at room temperature for 2 h [20]. After being washed, cells were incubated for 30 min at room temperature with a Cy3-conjugated anti-mouse IgG secondary antibody (Sigma Chemical Co.). The same protocol was used for MAP-2 immunostaining. AntiMAP-2 antibody was used at 1:400 (Roche Diagnostics, Meylan, France). Neuronal survival was quantified by counting NeuN-positive cells. Counting analysis of cells was performed by using an automatic analysis system (Trakcell) designed by Biocom (Les Ulis, France).
4. Concluding remarks It is well established that microglial cells can be toxic to neurons through cell–cell contact-dependent mechanisms [22]. These observations were confirmed in our study with direct microglial neuronal coculture in which neuronal death was significant in the presence of low concentrations of LPS. In the conditioned medium transfer system, soluble factors released by microglial cells after treatment of microglia with similar concentrations of LPS are not sufficient to induce a neurotoxic cascade. Even in the transwell system, high concentrations of LPS were necessary to induce neurotoxicity. At these high LPS concentrations modulation of microglial activation, by other factors, is difficult. The fact that we did not observe any modulatory response in these conditions could reflect the need for various soluble signals, as has been shown in previous studies, e.g. interferon-c plus LPS or a cocktail of cytokines [23]. It should also be noted that the type of neuronal culture could have an
350
V. Zujovic, V. Taupin / Methods 29 (2003) 345–350
influence on the sensitivity of the cells to LPS treatment. Hippocampal, cortical, or mesencephalic neurons have different properties that may influence their responses to microglial-derived toxic products. The stage of neuronal maturation may also be of importance (embryonic versus neonatal-derived neuronal cultures). It is worthwhile to point out that LPS has no direct effect on neuronal cell death. However, other microglial cell activators such as b-amyloid can be directly neurotoxic [24]. In this scenario, the conditioned medium transfer paradigm would be the more appropriate method for studying the influence of other factors on bamyloid-stimulated microglial-dependent neurotoxicity. We have previously shown that soluble fractalkine is able to inhibit TNFa production by LPS-activated microglia [10]. The second step was to establish the effect of fractalkine on microglia-induced neurotoxicity. The conditioned medium transfer and transwell systems were not convenient since in the former approach we observed no neurotoxicity and in the latter paradigm the concentrations of LPS were too high; the concentration of LPS sufficient to produce neurotoxicity decreases with increased cell-to-cell communication. The direct coculture system provided us an ideal experimental paradigm to study this chemokine present in the medium as soluble and as neuronal membrane attached. Exogenous fractalkine treatment did not provide any neuroprotective effects. Since neurons are the principal source of fractalkine in the brain, we included the strategy of blocking endogenously synthesized fractalkine using a specific anti-fractalkine antibody. With this approach, we observed an increase in neuronal death, suggesting a role for fractalkine in the control of microglial activation and the consequent neuronal cell death. The involvement of constitutive fractalkine in the control of cerebral inflammatory responses was also emphasized by our results from an in vivo model of brain inflammation [25]. These strategies were developed to study the in vitro role of a specific chemokine: fractalkine. Each strategy provides unique opportunities for better understanding of the function of each form of the molecule. With the conditioned medium transfer system, the direct action of the soluble molecule on microglial secretion of toxic factors is observed. The transwell system adds another dimension to the study since the neurons can respond to microglial activation by secreting soluble factors. The coculture system permits observations of the actions of both soluble and membrane-attached forms. This mul-
tifaceted approach can be extended to any molecule expressed as either soluble or transmembrane protein. References [1] M. Spranger, A. Fontana, Neuroscientist 2 (1996) 293–299. [2] F. Aloisi, Glia 36 (2001) 165–179. [3] R. Delgado, A. Carlin, L. Airaghi, M.T. Demitri, L. Meda, D. Galimberti, P. Baron, J.M. Lipton, A. Catania, J. Leukocyte Biol. 63 (1998) 740–745. [4] W.K. Kim, Y. Kan, D. Ganea, R.P. Hart, J. Gozes, G.M. Jonakait, J. Neurosci. 2 (2000) 3622–3630. [5] L.M. Phillips, L.A. Lampson, J. Neuropathol. Exp. Neurol. 59 (2000) 177–187. [6] R.C. Chang, P. Hudson, L. Hadden, J.-S. Hong, Brain Res. 853 (2000) 236–244. [7] J. Hesselgesser, R.J. Horuk, Neurovirology 5 (1999) 13–26. [8] J.K. Harrison, Y. Jiang, S. Chen, Y. Xia, D. Maciejewski, R.K. McNamara, W.J. Streit, M.N. Salafranca, S. Adhikari, D.A. Thompson, P. Botti, K.B. Bacon, L. Feng, Proc. Natl. Acad. Sci. USA 95 (1998) 10896–10901. [9] A. Nishiyori, M. Minami, Y. Ohtani, S. Takami, J. Yamamoto, N. Kawaguchi, T. Kume, A. Akaike, M. Satoh, FEBS Lett. 429 (1998) 167–172. [10] V. Zujovic, J. Benavides, X. Vige, C. Carter, V. Taupin, Glia 29 (2000) 305–315. [11] J.F. Bazan, K.B. Bacon, G. Hardiman, W. Wang, K. Soo, D. Rossi, D.R. Greaves, Z. Zlotnik, T.J. Schall, Nature 385 (1997) 640–644. [12] Y. Pan, C. Lloyd, H. Zhou, S. Dolich, J. Deeds, J.A. Gonzalo, J. Vath, M. Gosselin, J. Ma, B. Dussault, E. Woolf, G. Alperin, J. Culpepper, J.C. Gutierrez-Ramos, D. Gearing, Nature 387 (1997) 611–617. [13] G.A. Chapman, K. Moores, D. Harrison, C.A. Campbell, B.R. Stewart, P.J. Strijbos, J. Neurosci. 20 (2000) RC87. [14] G. Tarozzo, M. Campanella, M. Ghiani, A. Bulfone, M. Beltramo, Eur. J. Neurosci. Res. 15 (2002) 1663–1668. [15] D.M. Bronstein, I. Perez-Otano, V. Sun, S.B. Mullis Sawin, J. Chan, G.C. Wu, P.M. Hudson, L.Y. Kong, J.S. Hong, M.K. McMillian, Brain Res. 704 (1995) 112–116. [16] J.E. Merrill, Y. Koyanagi, J. Zack, L. Thomas, F. Martin, I.S. Chen, J. Virol. 66 (1992) 2217–2225. [17] L. Meda, M.A. Cassatella, G.I. Szendrei, L. Otvos Jr., P. Baron, M. Villalba, D. Ferrari, F. Rossi, Nature 374 (1995) 647–650. [18] E.N. Benveniste, J. Mol. Med. 75 (1997) 165–173. [19] R.B. Banati, J. Gehrmann, P. Schubert, G.W. Kreutzberg, Glia 7 (1993) 111–118. [20] R.J. Mullen, C.R. Buck, A.M. Smith, Development 116 (1992) 201–211. [21] M.B. Emerit, J. Segovia, H. Alho, M.J. Mastrangelo Jr., B.C. Wise, J. Neurochem. 52 (1989) 952–961. [22] J.M. Frade, Y.A. Barde, Neuron 20 (1998) 35–41. [23] J.M. Pocock, A.C. Liddle, Prog. Brain Res. 132 (2001) 555–565. [24] T. Harkany, I. Abraham, C. Konya, C. Nyakas, M. Zarandi, B. Penke, P.G. Luiten, Rev. Neurosci. 11 (2000) 329–382. [25] V. Zujovic, N. Schussler, D. Jourdain, D. Duverger, V. Taupin, J. Neuroimmunol. 115 (2001) 135–143.