- Top of page
- Materials and methods
- Conflict of interest
Astrocytes comprise the major glial cell population within the CNS, which is involved in multiple brain functions in physiological conditions, including neuronal development, synaptic activity and homeostatic control of the extracellular environment (Swanson et al., 2004; Darlington, 2005). They also actively participate in the processes triggered by brain injuries, aimed at limiting and repairing brain damage (Emily et al., 2004; Takuma et al., 2004). Indeed, after any degenerative injury or insult, astrocytes are activated in a process known as reactive astrogliosis (Eng et al., 1992; Suryadevara et al., 2003; Sriram et al., 2004). The function of reactive astrocytes is controversial, in that both beneficial and detrimental properties are postulated. In this context, although moderate activation of astrocytes may be crucial in the recovery of the injured CNS by secretion of neurotrophic factors, a rapid, severe and prolonged activity of these cells, as observed in chronic neurodegenerative diseases, is believed to augment or initiate a massive inflammatory response leading to neuronal death (Tani et al., 1996).
Reactive astrogliosis is generally characterized by cell proliferation, morphological changes, such as hypertrophy, emission of branches in existing astrocytes and increased expression of glial fibrillary acidic protein (Takamiya et al., 1988; Narita et al., 2004). A combination of studies, performed in vivo and in vitro by several groups using different CNS injury models, has convincingly implicated a number of cytokines in the generation or modulation of these processes (John et al., 2003). Among these, interleukin-1β (IL-1β), which is known to be expressed, together with its receptors, in astrocytes, stands out as being the major neuroinflammatory cytokine responsible for astrogliosis (Woiciechowsky et al., 2004; Hailer et al., 2005). However, the specific signalling mechanism by which IL-1β regulates this process and in particular cell proliferation, is not yet fully elucidated.
Our recent findings showed that nitric oxide (NO) modulated cell division in astrocytoma cell via activation of Ca2+/calmodulin (CAM) and the extracellular signal-regulated protein kinases (ERK1/2) (Meini et al., 2006). In another series of experiments, we showed that the same NO/Ca2+ cascade is part of the signalling pathway subserving the pyrogenic proinflammatory function of IL-1β (Meini et al., 2000; Palmi and Meini, 2002), raising the likelihood that this signalling mediates the cytokine-induced cell division. In the present study, we investigated the effect of IL-1β on astrocyte proliferation and the involvement of the NO/Ca2+ signalling in ERK activation and cell division. This in turn may shed light on the role of IL-1β on astrocyte-associated neuroinflammation.
- Top of page
- Materials and methods
- Conflict of interest
Proinflammatory cytokines, including IL-1β, are intimately involved in the generation or modulation of astrogliosis (John et al., 2004). Activated astrocytes produce IL-1β, which in turn, by autocrine and paracrine processes, causes further activation and proliferation of astrocytes (Giulian and Lachman, 1985). To study the mechanism underlying IL-1β-induced astrocyte proliferation, we used a cellular model of human astrocytoma U-373MG cells. Despite its malignant origin, the following considerations led us to use this cell line. First, previous investigations have shown that astrocytoma cells are a very suitable cell culture system for studying molecular signal transduction pathways (Mollace et al., 1993; Lieb et al., 1996); second, our and others' studies have shown that signal transduction pathways in this cell line are similar to that observed in primary astrocytes (Meini et al., 2000; Pahan et al., 2000) and third, with regard to the mitogenic properties of IL-1β, this cell line reacts comparably to primary astrocytes (Giulian and Lachman, 1985; Bertoglio et al., 1987; Kasahara et al., 1990; Yong, 1992).
Using this model, we found that low concentrations of IL-1β increased cell proliferation in a dose-dependent manner. However, further increase in IL-1β levels resulted in progressive reversal of this effect, indicating that the amount of the cytokine stimulation determined the specificity of the proliferative response. Furthermore, low IL-1β concentrations induced a dose-dependent activation of ERK, which was also progressively downregulated by further increase in the cytokine levels. The strict parallelism between the ERK and the proliferative responses as well as data showing that inhibition of MEK prevented ERK activation and antagonized the mitogenic effect of IL-1β strongly indicated that ERK was part of the mechanism underlying IL-1β-induced cell proliferation.
In attempting to explain the biphasic effect of IL-1β on cell division, we may assume that this cytokine activates distinct and potentially conflicting signals whose balance determine the final proliferative cell fate. Thus, besides the mitogenic p42/44 MAPK pathway, IL-1β might activate different, antiproliferative signalling pathways, which could involve initiation of apoptosis or, alternatively, a cytostatic effect. For instance, IL-1β has been shown to regulate the Fas-FLIP pathway, which in turn can be switched from survival to death signal depending on IL-1β concentration. Indeed, whereas low IL-1β levels activate FLIP leading to increased cell division, high cytokine concentrations downregulate this pathway, directing Fas to death signals (Maedler et al, 2006). With this in mind, it is likely that the Fas-FLIP and ERK pathways, depending on IL-1β level, determine cooperative or otherwise conflicting signals leading to potentiation or reduction of cell division. Alternatively, since IL-1β has been proposed to produce peroxynitrite (Keira et al., 2002), it is likely that increasing IL-1β levels determine elevation of this toxic compound with progressive downregulation of cell proliferation. In line with these hypothesis, we found that reduction in cell division and increase in apoptosis induced by the high IL-1β concentrations were strictly correlated and varied to a similar extent.
A dual role of IL-1β in the context of CNS inflammation has been described. Indeed, data showing IL-1β to contribute to neurodegeneration in CNS inflammatory diseases (Trapp et al., 1998), or otherwise promote remyelination in demyelinating diseases (Mason et al., 2001), lead to the hypothesis that the beneficial vs detrimental effects of IL-1β are probably influenced by a multitude of factors, including the concentrations of cytokine achieved at the site of inflammation.
For an insight into the mechanism that regulates the mitogenic response of IL-1β, we investigated intracellular messengers that could be the likely mediators of the ERK response. Our data showed that pretreatment with L-NAME or the iNOS-specific inhibitor, 1400W, along with an increase in cell proliferation, prevented ERK activation and that L-NAME was more potent than 1400W in eliciting these effects. These data indicated that NO mediated the IL-1β-induced ERK response and that the constitutive and the inducible forms of NOS were involved.
Despite controversy, the role of NO in cell growth regulation is well established. Indeed, whereas growth inhibition appears to be the major effect of NO (Ciani et al., 2004), a significant number of reports describe a proliferative effect (Kim et al., 2003; Bal-Price et al., 2006). This paradoxical behaviour may be reconciled by our and others' observations showing the specificity of the proliferative response being related to the NO levels. Thus, whereas low doses upregulate cell growth, excessive NO levels determine an opposite effect (Meini et al., 2006).
Involvement of the constitutive form of NOS in eliciting ERK response is supported by published data showing that this enzyme is expressed in astroglial cells and proinflammatory cytokines, including IL-1β, upregulating transcription of the constitutive NOS gene (Ma et al., 1994; Czapski et al., 2007). Our data showed that L-NAME, in the presence but not in the absence, of IL-1β, reduced the proliferative response to less than control values. In view of our hypothesis of IL-1β activation of potentially conflicting signals, we may hypothesize that inhibition of the mitogenic ERK pathway by L-NAME would shift the balance in favour of antiproliferative signals resulting in downregulated cell division.
Many effects of NO in different tissues are elicited via activation of soluble guanylate cyclase and cGMP generation. Although in the majority of these actions the precise signalling pathways involved are still primarily unknown, NO-mediated activation of a G kinase is generally accepted as part of the overall mechanism (Fiscus, 2002).
Data showing that the selective inhibitor of guanylate cyclase, ODQ, antagonized ERK activation as well as cell proliferation induced by IL-1β, suggested that a cGMP-dependent pathway is involved in the mechanism underlying these responses. Nevertheless, even though significant, ODQ was a relatively poor inhibitor, indicating that cGMP-independent pathways also participated and possibly were mainly involved in IL-1β effects. Many groups have reported direct interaction of NO with cellular and extracellular proteins, nitrosylation as well as production of NO-derived products, such as peroxynitrite (see Moncada and Bolanos, 2006). Therefore, it is likely that the cGMP-independent part of the NO responses were accounted for by direct interaction of this molecule with specific target proteins relevant for ERK activation. Like NO, Ca2+ is an important signal-transducing molecule that plays a remarkable role in controlling a wide range of cellular functions including cell growth (Villereal and Byron, 1992; Berridge, 1993). Our findings that inhibition of IP3- and RY-sensitive receptors, antagonized IL-1β-induced ERK activation as well as cell proliferation, suggested that Ca2+ mobilization from IP3 and RY stores was part of the mechanism underlying these responses. Ca2+ involvement in IL-1β functions has been previously demonstrated by our in vivo and in vitro studies showing that the pyrogenic/proinflammatory effect of IL-1β was associated with Ca2+ release from endoplasmic reticulum via type I IL-1β receptors and NO production (Palmi et al., 1995, 1996; Palmi and Meini, 2002; Meini et al., 2003). Recently, we have also shown that the amplitude of Ca2+ signalling, via modulation of the strength of ERK activation, regulates proliferation of different cell lines including astrocytes (Meini et al., 2006).
These data are supported by studies showing that removal of Ca2+ by BAPTA/EGTA (Yang et al., 2000) or inhibition of bradykinin receptors responsible for Ca2+ release from IP3-sensitive stores (Yang et al, 2001), significantly attenuate [3H]thymidine incorporation and p42/p44 MAPK activation in IL-1β-treated canine tracheal smooth muscle cells.
Much attention has been focused in recent years on ascertaining the mechanism by which Ca2+ regulates ERK activity. Our findings that W7 prevented IL-1β-induced ERK activation along with cell proliferation indicated that CaM was involved. Molecules which have been described as potential CaM regulators of the ERK–MAPK pathway include the Ras GTP exchange factors, Ras GRF and Ras GRP (Farnsworth et al, 1995; Ebinu et al, 1998). With regards to the present work, it is relevant to point out that Ras/Raf signalling network has been convincingly implicated in promoting proliferative responses as well as differentiation and survival signals through ERK-dependent transactivation of cyclin D (Agell et al., 2002).
Data demonstrating Ca2+ and NO involvement in the mitogenic effect of IL-1β, were further supported by direct measurement of intracellular levels of these two messengers showing that NO and Ca2+ increased transiently and time dependently in response to IL-1β and that the NO response preceded by 15 min that of Ca2+.
All together, these data demonstrated that in human astrocytoma cells, the NO/Ca2+ signalling pathway mediated the mitogenic response of IL-1β via ERK activation and that in the sequence of events leading to the final response, the NO signalling was upstream to that of Ca2+.
Cell proliferation is a hallmark of astrogliosis and IL-1β is clearly the main proinflammatory cytokine involved in this process. Therefore, it is conceivable that the NO/Ca2+ signalling may be part of the mechanism regulating IL-1β-induced astrogliosis. Furthermore, because the role of glial activation in the context of CNS inflammation is controversial, it is likely that the amplitude of the NO/Ca2+ signalling might represent a switch that controls the beneficial vs detrimental effector function of astrogliosis. It is therefore tempting to speculate that modulation of this signalling have important implications in therapeutic approaches to chronic neuroinflammatory disorders.