SEARCH

SEARCH BY CITATION

Keywords:

  • ATP receptor;
  • plasminogen;
  • TNF-α, IL-6

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PURINE AND PYRIMIDINE RECEPTORS EXPRESSED IN MICROGLIA
  5. CHEMOTAXIS AFTER MEMBRANE RUFFLING
  6. RELEASE MECHANISM OF PLASMINOGEN AND CYTOKINES BY ATP STIMULATION
  7. CELL DEATH
  8. THE FATE OF NEURON: “TO BE OR NOT TO BE”
  9. REFERENCES

Microglial activation by purines and pyrimidines is reviewed, with emphasis on the actions of adenosine 5′-triphosphate (ATP) on chemotaxis or releases of plasminogen and cytokines from microglia. ATP activates microglia, causing morphological changes with membrane ruffling. Activated microglia exhibit chemotaxis to ATP. Microglia stimulated by a low concentration of ATP (∼30–50 μM) rapidly release plasminogen (within 5–10 min), which may protect neurons. Microglia stimulated by a higher concentration of ATP release tumor necrosis factor-α (TNF-α), 2–3 h after the stimulation and interleukin-6 (IL-6), 6 h after the stimulation. It is reported that TNF-α stimulation causes an increase in the expression of IL-6 receptor mRNA and expression in neuronal cells (März et al. 1996. Brain Res 706:71–79). After binding with gp130, the IL-6 receptor matures and can accept IL-6 molecules. It is speculated that neurons may require several hours to prepare for the full reception of IL-6, which induces a more efficient protective effect by IL-6 after stimulation with TNF-α. After neurons are ready to accept IL-6 fully, microglia release IL-6 to neurons. Stronger and longer stimulation by ATP may change the function of microglia and cause cell death. The conditions evoking the heavy stimulation would result from serious injury. Activated microglia act as scavenger cells that induce apoptosis in damaged neurons by releasing toxic factors, including NO, and removing dead cells, their remnants, or dangerous debris by phagocytosis. These actions lead to a suitable environment for tissue repair and neural regeneration. The fate of neurons may therefore be regulated in part by ATP through the activation of microglia. GLIA 40:156–163, 2002. © 2002 Wiley-Liss, Inc.

INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PURINE AND PYRIMIDINE RECEPTORS EXPRESSED IN MICROGLIA
  5. CHEMOTAXIS AFTER MEMBRANE RUFFLING
  6. RELEASE MECHANISM OF PLASMINOGEN AND CYTOKINES BY ATP STIMULATION
  7. CELL DEATH
  8. THE FATE OF NEURON: “TO BE OR NOT TO BE”
  9. REFERENCES

The most characteristic feature of microglia is their rapid activation in response to acute pathological events including apoptosis, neurodegeneration, and inflammation in the central nervous system (CNS). Activated microglia mainly act as scavenger cells, i.e., inducing apoptosis in damaged cells by releasing toxic factors and removing dead cells, their remnants, or dangerous debris by phagocytosis. They also perform various functions in tissue repair and neural regeneration by releasing trophic factors such as nerve growth factor and brain-derived neurotrophic factor, cytokines, and plasminogen (Nakajima and Kohsaka, 1993). It has been reported that plasminogen is a neuroprotective substance (Nakajima and Kohsaka, 1993). It is also known that interleukin-6 (IL-6) regulates neuronal survival and differentiation in the CNS (Hama et al., 1991; Yamada and Hatanaka, 1994; Umegaki et al., 1996). More recently, cytokines were implicated as mediators and inhibitors of diverse forms of neurodegeneration (Allan and Rothwell, 2001). Thus, microglia appear to have a dual regulatory function in maintaining or facilitating tissue homeostasis in the CNS. These properties of microglia can be modulated by cytokines and neurotransmitters including ATP via their specific receptors in the CNS (Kreutzberg, 1996).

Microglia possess functional receptors for purines and pyrimidines, i.e., GTP-binding protein (G-protein)-coupled P2 receptors such as P2Y2, P2Y12-like receptor (Honda et al., 2001), and ionotropic P2 receptors such as P2X7 (Kreutzberg, 1996; Ferrari et al., 1996; Verkhratsky and Kettenmann, 1996). ATP evokes currents in rat microglia (Nörenberg et al., 1994) and increased intracellular calcium ([Ca2+]i) in mouse and human microglia (Walz et al., 1993; Toescu et al., 1998; Möller et al., 2000). ATP induces the release of interleukin-1β (IL-1β) (Ferrari et al., 1996, 1997b) and IL-6 (Shigemoto-Mogami et al., 2001) from mice microglia, as well as the release of plasminogen (Inoue et al., 1998b) and TNF-α (Hide et al., 2000; Morigiwa et al., 2000) from rat microglia. ATP activates nuclear factor of activated T cells (NFAT) (Ferrari et al., 1999b), which modulates the early inflammatory gene expression, and transcriptional activator NF-κB, which controls cytokine expression and apoptosis (Ferrari et al., 1997c) through P2X7. ATP also stimulates the phosphorylation of mitogen-activated protein (MAP) kinase through P2Y receptors in several cells (Neary and Zhu, 1994; Soltoff et al., 1998; Gao et al., 1999; Neary et al., 1999), including microglia (Hide et al., 2000; Honda et al., 2001; Shigemoto-Mogami et al., 2001).

These data suggest that ATP plays very important roles in a variety of functions in microglia in the inflammation or injury of the CNS, mainly through evoking the release of cytokines and related substances. This article examines and discusses the mechanism and the physiological meaning of the release of these substances from activated microglia by ATP stimulation.

PURINE AND PYRIMIDINE RECEPTORS EXPRESSED IN MICROGLIA

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PURINE AND PYRIMIDINE RECEPTORS EXPRESSED IN MICROGLIA
  5. CHEMOTAXIS AFTER MEMBRANE RUFFLING
  6. RELEASE MECHANISM OF PLASMINOGEN AND CYTOKINES BY ATP STIMULATION
  7. CELL DEATH
  8. THE FATE OF NEURON: “TO BE OR NOT TO BE”
  9. REFERENCES

ATP receptor (P2 receptor or purinoceptor) is the generic name for receptors that are activated by purines (e.g., ATP, ADP) and pyrimidines (e.g., UTP, UDP). It is now accepted that ATP receptors can be divided into two families: P2X and P2Y. The P2X family (P2X1-P2X7) is a ligand-gated type of ion channel. The P2Y family (P2Y1, P2Y2, P2Y4, P2Y6, P2Y11, P2Y12, P2Y13) is a member of the superfamily of G-protein-coupled receptors. Pharmacological and physiological findings, as well as immunohistochemical evidence, suggest that microglia express protein molecules of P2X7, P2Y2, and P2Y12-like receptors (Walz et al., 1993; Neary and Zhu, 1994; Nörenberg et al., 1994; Kreutzberg, 1996; Verkhratsky and Kettenmann, 1996; Ferrari et al., 1996; Soltoff et al., 1998; Toescu et al., 1998; Gao et al., 1999; Neary et al., 1999; Hide et al., 2000; Möller et al., 2000; Honda et al., 2001; Shigemoto-Mogami et al., 2001). However, these findings do not exclude the possibility that other receptors are expressed in microglia.

CHEMOTAXIS AFTER MEMBRANE RUFFLING

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PURINE AND PYRIMIDINE RECEPTORS EXPRESSED IN MICROGLIA
  5. CHEMOTAXIS AFTER MEMBRANE RUFFLING
  6. RELEASE MECHANISM OF PLASMINOGEN AND CYTOKINES BY ATP STIMULATION
  7. CELL DEATH
  8. THE FATE OF NEURON: “TO BE OR NOT TO BE”
  9. REFERENCES

The initial microglial responses that occur after brain injury and in various neurological diseases are characterized by microglial accumulation in the affected sites of the brain, which results from the migration and proliferation of these cells. The early-phase signal responsible for this accumulation is likely to be transduced by rapidly diffusible factors. Honda et al. (2001) examined the possibility that ATP released from injured neurons and nerve terminals affects the cell motility in rat primary cultured microglia. These investigators found that extracellular ATP and ADP induced membrane ruffling and markedly enhanced chemokinesis in a Boyden chamber assay. Further analyses using the Dunn chemotaxis chamber assay, which permits direct observation of the cell movement, showed that both ATP and ADP induced chemotaxis of microglia. The elimination of extracellular calcium or treatment with pyridoxalphosphate-6-azophenyl-2′,4′-disulfonic acid, suramin, or adenosine-3′-phosphate-5′-phosphosulfate did not inhibit ATP- or ADP-induced membrane ruffling, whereas AR-C69931MX, a P2Y12 receptor blocker (Hollopeter et al., 2001), or pertussis toxin (PTX) treatments clearly inhibited the ruffling. As an intracellular signaling molecule underlying these phenomena, the small G-protein Rac was activated by ATP and ADP stimulation, and its activation was also inhibited by pretreatment with PTX. These findings suggest that the membrane ruffling and chemotaxis of microglia induced by ATP or ADP are mediated by G(i/o)-coupled P2Y receptors (P2Y12 and/or P2Y12-like receptors).

RELEASE MECHANISM OF PLASMINOGEN AND CYTOKINES BY ATP STIMULATION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PURINE AND PYRIMIDINE RECEPTORS EXPRESSED IN MICROGLIA
  5. CHEMOTAXIS AFTER MEMBRANE RUFFLING
  6. RELEASE MECHANISM OF PLASMINOGEN AND CYTOKINES BY ATP STIMULATION
  7. CELL DEATH
  8. THE FATE OF NEURON: “TO BE OR NOT TO BE”
  9. REFERENCES

Plasminogen

ATP stimulated the release of plasminogen in a concentration-dependent manner from 10 μM to 100 μM, with a peak response at 5–10 min after the stimulation (Inoue et al., 1998). A 1-h pretreatment with BAPTA-AM (200 μM), which is metabolized in the cytosol to BAPTA (an intracellular Ca2+ chelator), completely inhibited the plasminogen release evoked by ATP (100 μM). The Ca2+ ionophore A23187 induced plasminogen release in a concentration-dependent manner (0.3–10 μM). ATP induced a transient increase in the [Ca2+]i in a concentration-dependent manner, which was very similar to the ATP-evoked plasminogen release. A second application of ATP (100 μM) stimulated an increase in [Ca2+]i similar to that of the first application (21 out of 21 cells). The ATP-evoked increase in [Ca2+]i was totally dependent on extracellular Ca2+. 2-Methylthio-ATP was effective (7 out of 7 cells), but α,β-methylene ATP was ineffective (7 out of 7 cells) at inducing an increase in [Ca2+]i. Suramin (100 μM) was shown not to inhibit the ATP-evoked increase in [Ca2+]i (20 out of 20 cells). 2′- and 3′-O-(4-Benzoylbenzoyl)adenosine 5′-triphosphate (BzATP), a selective agonist of P2X7 receptors, evoked a long-lasting increase in [Ca2+]i even at 1 μM, a concentration at which ATP did not evoke the increase. A 1-h pretreatment with adenosine 5′-triphosphate-2′, 3′-dialdehyde (oxidized ATP, 100 μM), a selective antagonist of P2X7 receptors, blocked the increase in [Ca2+]i induced by ATP (10 and 100 μM). These data suggest that ATP may transit information from neurons to microglia, resulting in an increase in [Ca2+]i via the ionotropic P2X7 receptor, which stimulates the release of plasminogen from the microglia. It has been found that UTP also stimulates plasminogen release from a subpopulation of microglia (about 20% of total cells), presumably through store-operated Ca2+ entry (SOC) activated by ATP stimulation of G protein-coupled receptors, since the release evoked by UTP was also dependent on extracellular Ca2+ (K. Inoue and S. Kohsaka, unpublished data).

Tumor Necrosis Factor-α

Hide et al. (2000) reported that ATP potently stimulates TNF-α release beginning at 2 h after the stimulation with TNF-α mRNA expression in primary cultures of rat brain microglia. In that report, the TNF-α release was maximally elicited by 1 mM ATP and BzATP, suggesting the involvement of P2X7. This TNF-α release was Ca2+ dependent and was correlated with a sustained Ca2+ influx. The release was inhibited by PD98059, an inhibitor of MAP kinase kinase 1 (MEK1) that activates extracellular signal-regulated protein kinase (ERK), and SB203580, an inhibitor of p38 MAP kinase. However, both ERK and p38 were rapidly activated by ATP, even in the absence of extracellular Ca2+. These results indicate that extracellular ATP triggers TNF-α release in rat microglia via the P2 receptor, probably P2X7, by a mechanism that is dependent on both the sustained Ca2+ influx and ERK/p38 cascade and is regulated independently of Ca2+ influx.

Iinterleukin-6

We found that ATP evokes the release of IL-6 at 24 h in a concentration-dependent manner (10–1,000 μM) in MG-5 (Ohsawa et al., 1997; Shigemoto-Mogami et al., 2001) and that the release was observed from 6 h after stimulation with ATP. Neither ADP (1,000 μM), UTP (1,000 μM) nor adenosine (1,000 μM) stimulated the release of IL-6. There is a possibility that ATP might evoke the release of IL-6 secondarily by releasing TNF-α and IL-1β because TNF-α and IL-1β are reported to stimulate IL-6 production in other glial cells (Norris et al., 1994). In MG-5, a significant amount of TNF-α appeared at 1 h and peaked at 3 h after the stimulation by 1,000 μM ATP, but TNF-α (≤10,000 pg/ml) never evoked the release of IL-6 in MG-5. This result was in accord with a previous report that TNF-α failed to stimulate IL-6 production in microglia (Sawada et al., 1992). Moreover, the release of IL-1β, a potent inducer of IL-6 gene (Lee et al., 1993), was not evoked by 1,000 μM ATP in MG-5 (K. Inoue et al., unpublished data).

ATP induced an approximately 7-fold increase in the expression of IL-6 mRNA, which was inhibited by 1 mM suramin to 5.6 ± 6.9% of ATP alone, indicating that ATP stimulates the de novo synthesis of IL-6 via a P2 receptor–mediated pathway and the subsequent production of IL-6. Previous studies with IL-1β-stimulated synoviocytes indicated that p38 activation leads to IL-6 gene expression through enhancement of IL-6 mRNA stability (Miyazawa et al., 1998). However, in our experiment, p38 appeared to act on the initial ATP signal transduction pathway affecting IL-6 gene expression, since no significant decrease of IL-6 mRNA expression was observed when SB203580, an inhibitor of p38, was added 3 or 5 h after the ATP application. The finding that the ATP-evoked phosphorylation of p38 was rapid, i.e., peaking at 1 min and almost disappearing at 60 min after ATP stimulation, may also exclude the possibility that p38 activation leads to increased IL-6 expression via its effects on IL-6 mRNA stability.

ATP could activate two distinct MAP kinases, i.e., ERK1/2 and p38, in MG-5. ADP activated ERK1/2 strongly but p38 only slightly. ATP stimulated IL-6 release but ADP did not. The ATP-stimulated IL-6 release was inhibited by SB203580 but not by an inhibitor of ERK1/2. These results strongly suggest that p38 but not ERK1/2 MAP kinase is responsible for the IL-6 release.

Phosphorylation by ATP of p38 was dependent on the extracellular Ca2+. ATP produced a phospholipase C (PLC)-dependent transient Ca2+ release by inositol-1,4,5-trisphosphate (InsP3), which was followed by sustained Ca2+ entry via both SOC and P2X7 receptors. Several groups have reported that P2X7 receptors have a central role in the production of cytokines in microglia (Ferrari et al., 1996, 1997b; Hide et al., 2000). In fact, BzATP evoked sustained Ca2+ entry via P2X7 receptors, leading to the phosphorylation of p38 in MG-5. However, BzATP induced only a very small amount of IL-6 production in the cells. Brilliant Blue G (≤10 μM), a specific P2X7 antagonist (Jiang et al., 2000), did not inhibit the release of IL-6 induced by ATP from MG-5. As opposed to the release of TNF-α from microglia, P2Y, rather than P2X7 receptors, seem to have a major role in the IL-6 production by the cells. This idea is supported by the most recent observation that ATP may evoke IL-6 production in bacterial endotoxin lipopolysaccharide (LPS)-primed P2X7-deficient mice (P2X7R−/−) (Solle et al., 2001). However, the fact that high ATP concentrations are required for the IL-6 production may still support the idea that P2X7, rather than P2Y receptors, might be involved in the responses. Very recently, it has been reported that ATP acting on P2Y receptors stimulated the release of IL-12 in human dendritic cells (Wilkin et al., 2001). The ATP concentrations required for this IL-12 induction were similar to those required for IL-6 production in MG-5. Thus, there seems to be much variety among the subclass of P2Y receptors in their sensitivity to ATP. The deduced P2Y receptors in MG-5 would therefore require high ATP concentrations for their activation. In addition, when cells were incubated with either NEM, an inhibitor of some G-proteins (Rhee et al., 2000), or U73122, the ATP-evoked IL-6 production was abolished. Thus, activation of PLC-linked P2Y receptors and their receptor-mediated signals including p38 would trigger the IL-6 synthesis in MG-5.

The activation of p38 is not sufficient for the IL-6 induction because BzATP activated p38, but it did not evoke the release of IL-6. This result would predict the existence of additional signals for the IL-6 production. A Ca2+-dependent PKC may be an additional signal, since the ATP-evoked IL-6 production was attenuated by Gö6976, an inhibitor of Ca2+-dependent protein kinase C (PKC). The P2Y receptor responsible for these responses was insensitive to PTX and was linked to phospholipase C. Some transcriptional factors, such as NF-κB p65 (RelA) (Ferrari et al., 1997c), Jun, and Fos (Neary et al., 1996), are known to be activated by ATP. Berghe et al. (1998) described the involvement of MAP kinase pathways in NF-κB transactivation, which leads to the induction of IL-6 gene expression. Although such transcriptional factors may work as the downstream signals of either p38 or Ca2+-dependent PKC, the detailed mechanism underlying such a cooperative regulation of IL-6 production remains to be clarified.

Finally, we tried to determine the subclass of P2Y receptors involved in the ATP-evoked responses. Reverse transcription-polymerase chain reaction (RT-PCR) analysis showed that P2Y1, P2Y2, and P2Y6, but not P2Y4 mRNAs are present in the cells. However, the pharmacological profile for recombinant P2Y1, P2Y2, and P2Y6 receptors is inconsistent with our findings on the IL-6 gene expression by P2 agonists, as neither ADP (a potent agonist to P2Y1) nor UTP (an activator of P2Y2 and P2Y6 receptors) stimulated significant IL-6 gene expression in MG-5. This finding suggests that the properties of the P2Y receptors coupled to IL-6 gene expression in MG-5 differ from those of the recombinant P2Y receptors or that MG-5 express novel ATP-preferring receptors coupled to IL-6 production.

Although the definitive classification of P2Y receptors for the ATP-evoked IL-6 release remains to be determined, it is suggested that ATP, acting on the PTX-insensitive P2Y receptors linked to PLC, stimulates the synthesis of IL-6 through pathways involving p38 and Ca2+-dependent PKC. This is a novel pathway for the induction of cytokines in microglia.

Interleukin-1β

The LPS-dependent release of IL-1β from mouse microglial cells is an inefficient process, as it is slow and leads to the accumulation of only a modest amount of extracellular cytokine. This evidence suggested that a second stimulus is needed to elicit the IL-1β secretion from microglia. Di Virgilio's group has examined the mechanism of IL-1β secretion from microglia. This group first reported that extracellular ATP triggers IL-1β release from LPS-treated macrophages or microglia by activating the purinergic P2Z receptor (Ferrari et al., 1996, 1997b). They confirmed that ATP is a powerful stimulus for IL-1β release from LPS-treated macrophages or microglia and showed that IL-1β release is not necessarily associated with cell death, as it occurs at lower ATP concentrations and much earlier than the leakage of cytoplasmic markers. Sanz and Di Virgilio (2000) examined the kinetics and mechanism of ATP-dependent IL-1β release from microglial cells. The addition of extracellular ATP to LPS-primed microglia caused a burst of release of a large amount of processed IL-1β. ATP had no effect on the accumulation of intracellular pro-IL-1β in the absence of LPS. In LPS-treated cells, ATP slightly increased the synthesis of pro-IL-1β. The optimal ATP concentration for IL-1 β secretion was between 3–5 mM, but significant release could be observed at concentrations as low as 1 mM. At all ATP concentrations, the IL-1β release could be inhibited by increasing the extracellular K+ concentration. The ATP-dependent IL-1β release was also inhibited by 90% and 60% by the caspase inhibitors YVAD and DEVD, respectively. Accordingly, in ATP-stimulated microglia, the p20 proteolytic fragment derived from activation of the IL-1β-converting enzyme could be detected by immunoblotting. Sanz and Di Virgilio concluded that ATP triggers accelerated maturation and the release of intracellularly accumulated IL-1β by activating the IL-1β-converting enzyme/caspase 1 in mouse microglia.

CELL DEATH

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PURINE AND PYRIMIDINE RECEPTORS EXPRESSED IN MICROGLIA
  5. CHEMOTAXIS AFTER MEMBRANE RUFFLING
  6. RELEASE MECHANISM OF PLASMINOGEN AND CYTOKINES BY ATP STIMULATION
  7. CELL DEATH
  8. THE FATE OF NEURON: “TO BE OR NOT TO BE”
  9. REFERENCES

Surprenant et al. (1996) reported cloning of P2X7 that is responsible for the ATP-dependent lysis of macrophages (Blanchard et al., 1995) through the formation of membrane pores permeable to large molecules. Other members of P2X receptors are permeable only to small cations. P2X7 is homologous to other P2X receptors but has a unique carboxyl-terminal domain that is required for the lytic actions of ATP. Surprenant et al. (1995) concluded that the P2X7 receptor is a bifunctional molecule that could function in both fast synaptic transmission and the ATP-mediated lysis of antigen-presenting cells. Ferrari et al. (1997a) reported that the pore-forming P2Z/P2X7 receptor exists in N9 and N13 cells, as shown by a specific polyclonal antibody, and that microglial cells expressing the P2/P2X7 receptor are sensitive to ATP-mediated cytotoxicity, while clones selected for the lack of this receptor are resistant. Transfection of HEK293 cells with P2X7 (but not P2X2) receptor cDNA confers susceptibility to ATP-mediated cytotoxicity. On the basis of morphological and biochemical analysis, these workers also suggested that ATP-dependent cell death in microglial cells occurs by apoptosis. Finally, microglial cells release ATP via a nonlytic mechanism when activated by bacterial endotoxin, suggesting the operation of a purinergic autocrine/paracrine loop. It is therefore suggested that microglia die through a strong activation of P2X7 by ATP (Ferrari et al., 1997a, 1999a). How strong? The answer was obtained by Virginio et al. (1999), who examined the kinetics of cell lysis, the uptake of the propidium dye YO-PRO-1, and changes in permeability in human embryonic kidney (HEK293) cells expressing the rat P2X7 receptor and found that ATP acting on P2X7 receptors opens a channel permeable to small cations, creates an access pathway for the entry of larger molecular weight dyes, and causes cell death. BzATP (30 μM) induced membrane blebbing within 30–40 s of sustained application. Electrophysiological measurements of the current reversal potentials with intracellular sodium and extracellular cations of different sizes showed that the ionic channel progressively dilated during 10–20 s to a diameter of >1 nm (10 A). With short agonist applications (3–5 s), pore dilation and YO-PRO-1 uptake were reversible and repeatable. They concluded that maximum P2X7 receptor activation causes an exponential dilatation of the ion channel with a time constant of 25 s to a final diameter of 3–5 nm from an initial minimum pore diameter of 0.8 nm.

What kind of mechanism is involved in the apoptosis in microglia through the activation of P2X7? Ferrari et al. (1999a) examined the role of caspases, a family of proteases implicated in apoptosis and cytokine secretion, to obtain the answer to this question. They observed that extracellular ATP induced the activation of multiple caspases including caspase-1, -3, and -8, and subsequent cleavage of the caspase substrates PARP and lamin B. Using caspase inhibitors, it was found that caspases were specifically involved in ATP-induced apoptotic damage, such as chromatin condensation and DNA fragmentation. These investigators suggested that the activation of caspases by P2X7 is required for apoptotic cell death by ATP. An important regulator implicated in the control of apoptosis is the transcriptional activator NF-κB. They also examined the relationship between NF-κB and caspase activation by ATP and found that exposure of microglial cells to ATP resulted in potent NF-κB activation through P2X7 (Ferrari et al., 1997c). NF-κB activation required reactive oxygen intermediates and proteases of the caspase family as evidenced by the fact that it was abolished by antioxidants and specific protease inhibitors. ATP stimulation resulted only in the appearance of a p65 homodimer, although exposure to LPS induced prototypical NF-κB p50 homodimers and p65 (RelA)/p50 heterodimers. This result suggests that ATP may control the expression of a subset of NF-κB target genes distinct from those activated by classical pro-inflammatory mediators.

Microglia may be able to kill other cells through the release of nitric oxide (NO) (Sikora et al., 1999). Ohtani et al. (2000) examined the effect of ATP on the expression of inducible nitric oxide synthase (iNOS) mRNA and the production of nitric oxide in cultured rat microglia. These investigators found that ATP induced iNOS mRNA dose-dependently (100–1,000 μM). The induction began within 1 h after the addition of ATP (100 μM) with peak expression occurring at 6 h. The release of NO in the culture medium was significantly increased by the treatment with ATP (100 and 1,000 μM) for 12 and 24 h. They suggested that ATP is a potential mediator that induces iNOS mRNA expression and NO production in microglia.

THE FATE OF NEURON: “TO BE OR NOT TO BE”

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PURINE AND PYRIMIDINE RECEPTORS EXPRESSED IN MICROGLIA
  5. CHEMOTAXIS AFTER MEMBRANE RUFFLING
  6. RELEASE MECHANISM OF PLASMINOGEN AND CYTOKINES BY ATP STIMULATION
  7. CELL DEATH
  8. THE FATE OF NEURON: “TO BE OR NOT TO BE”
  9. REFERENCES

Figure 1 summarizes the release mechanisms revealed so far. It is notable that plasminogen release is evoked only 5–10 min after the ATP stimulation and that the concentration of ATP is rather low (estimated EC50 is 30 μM). This means that the release of plasminogen is not via de novo synthesis and may be an urgent response to rescue the neuron. It takes at least 2–3 h after the stimulation of ATP to evoke the release of TNF-α. Both the sustained Ca2+ influx and TNF-α release were maximally elicited by 1 mM ATP and BzATP, concentrations higher than those evoking plasminogen release. This is in agreement with previous reports showing that higher concentrations are needed to stimulate P2X7. One might then conclude that only P2X7 stimulation is necessary to evoke the release. However, both ERK and p38 were rapidly activated by ATP, even in the absence of extracellular Ca2+, although the ATP-evoked TNF-α release was inhibited by both PD98059 and SB203580 (Hide et al., 2000). These results indicate that extracellular ATP triggers TNF-α release in rat microglia via P2X7 and via another signaling pathway activating the ERK/p38 cascade independently of Ca2+ influx. In MG-5 cells, ATP caused the mRNA expression and release of IL-6 in a concentration-dependent manner from 6 h after the stimulation. When MG-5 cells were treated with 1 mM ATP, a significant amount of TNF-α was also released. The time course of the ATP-evoked TNF-α release was much faster than that of IL-6, which raises the possibility that the released TNF-α may stimulate the release of IL-6, thereby accounting for its slower time-course. However, in our experiments, TNF-α never produced the release of IL-6 in MG-5, even at concentrations of ≤1,000 pg/ml. It has been reported that plasminogen is a neuroprotective substance (Nakajima and Kohsaka, 1993). It is also known that IL-6 regulates neuronal survival and differentiation in the CNS (Hama et al., 1991; Yamada and Hatanaka, 1994; Umegaki et al., 1996). TNF-α is generally considered cytotoxic and to be involved in the pathology of many neurodegenerative diseases (Brosnan et al., 1988; Fillit et al., 1991). Recently, however, there is increasing evidence that TNF-α also has a neuroprotective role (Cheng et al., 1994; Barger et al., 1995). It has been reported that mice lacking TNFR1 showed enhanced brain injury under excitatory and ischemic conditions (Bruce et al, 1996). In a human neuronal cell line, the inhibition of TNFR2 by antisense oligonucleotides increased hypoxic injury and β-amyloid toxicity (Shen et al., 1997). TNF-α itself performs multiple functions in neurodegeneration and neuroprotection and the transcription factor NF-κB has been suggested to be a key molecule for deciding cell survival in response to TNF-α (Van Antwerp et al., 1998).

thumbnail image

Figure 1. Release mechanisms of plasminogen, tumor necrosis factor-α (TNF-α,) and interleukin-6 (IL-6). The three curves at the bottom are rough time-courses of the release of these substances. The word “stimulation” indicates the starting point of ATP stimulation. SOC: store-operated Ca2+ entry.

Download figure to PowerPoint

It takes 5–10 min for plasminogen release, 2–3 h for TNF-α release, and 6 h for IL-6 release. The physiological meaning of the release of these substances may be suggested by the different time lags for secretion. One possibility for the physiological significance is shown in Figure 2. Activated microglia exhibit chemotaxis to ATP (Honda et al., 2001). The microglia are first activated by a low concentration of ATP (∼30–50 μM) and release plasminogen rapidly, which may protect neurons. After the microglia are exposed to a higher concentration of ATP, the microglia release TNF-α 2–3 h after stimulation. At this point, the microglia do not release IL-6. For comparison, we examined and found that LPS stimulated both the release of TNF-α and that of IL-6 from MG-5 with the same time-course, i.e., both releases peaked at 2 h after the stimulation and remained for more than 12 h (Inoue and Shigemoto, unpublished data). The difference between the ATP-evoked release and the LPS-evoked one is remarkable. The former appears to be a well-organized physiological response and the latter is an abnormal and toxic response. A question remains concerning the reason for the time difference between TNF-α and IL-6 release by ATP. One possible explanation can be hypothesized from the fact that the neuroprotective effect of IL-6 was enhanced by pretreatment with TNF-α in the serum-deprived death of PC12 cells (Inoue and Shigemoto, unpublished data). These data suggest that neurons stimulated by TNF-α may require several hours to prepare for the reception of IL-6 to induce a more efficient protective effect. This idea is supported by the evidence that TNF-α stimulation causes an increase in IL-6 receptor mRNA and expression in neuronal cells (März et al., 1996). We obtained a similar result in PC12 cells in which TNF-α stimulated IL-6 mRNA (Inoue and Shigemoto, unpublished data). IL-6 receptor binds with ubiquitous gp130, presumably enabling the receptor to bind IL-6 molecules completely.

thumbnail image

Figure 2. Hypothesis: the physiological meaning of the time difference in the release of these substances from microglia.

Download figure to PowerPoint

Stronger and longer stimulation of microglia by ATP may cause cell death. Conditions evoking the heavy stimulation would result from serious injury. Neurons under such a condition would sustain unrecoverable damage. Activated microglia act as scavenger cells, i.e., inducing apoptosis in damaged neurons by releasing toxic factors including NO and removing dead cells, their remnants or dangerous debris by phagocytosis. Such actions promote a suitable environment for tissue repair and neural regeneration. The survival of neurons must be regulated in part by ATP through the activation of microglia.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PURINE AND PYRIMIDINE RECEPTORS EXPRESSED IN MICROGLIA
  5. CHEMOTAXIS AFTER MEMBRANE RUFFLING
  6. RELEASE MECHANISM OF PLASMINOGEN AND CYTOKINES BY ATP STIMULATION
  7. CELL DEATH
  8. THE FATE OF NEURON: “TO BE OR NOT TO BE”
  9. REFERENCES