[1] 7w Transforming growth factor-b signalling in brain disorders

[1] 122w Transforming growth factor-b (TGF-b) has been characterized as an injury-related factor, based on the observation that it is strongly upregulated in many acute or chronic central nervous system disorders. TGF-b is generally thought to be neuroprotective and several mechanisms have been proposed to explain this beneficial action. For instance, TGF-b protects neurons against the potentiating effect of tissue-type plasminogen activator on NMDA receptor-mediated excitotoxicity, by up-regulating type-1 plasminogen activator inhibitor expression in astrocytes. TGF-b has also anti-apoptotic properties, through a recruitment of a mitogen-activated protein kinase pathway and a concomitant activation of anti-apoptotic members of the Bcl-2 family. These multiple mechanisms might reflect the pleiotropic nature of TGF-b, reinforcing the potential therapeutic value of this cytokine in several central nervous system disorders.

[1] 191w The transforming growth factor-beta (TGF-b) was discovered in the early 1980s for its transforming properties [1,2]. It soon appeared that this pleiotropic cytokine orchestrates many critical physiological processes (embryogenesis, immune response, extracellular matrix metabolism, cell cycle, . . .), while dysregulations in its signalling pathways are involved in several pathological conditions [3][4][5]. Once released from a latent secreted complex [6], TGF-b isoforms elicit their cell type-specific responses through the ligandinduced formation of a heteromeric receptor complex between the serine/threonine kinases TbR-I and TbR-II: the type II receptor binds TGF-b by its own, then recruits the type I receptor, allowing the transphosphorylation-mediated activation of TbR-I. The subsequent activation of the signalling mother against decapentaplegic peptide (Smad) transcription factor cascade thus regulates the transcription of key target genes [7][8][9]. Historically, Smads were thought to be the main substrates and the key signal transducers of activated TGF-b receptors. However, an ever increasing complexity in the canonical TGF-b intracellular cascade emerges, with the discovery of cross-talks with other signalling pathways, such as mitogen-activated protein kinases (MAPK) or Wnt and the description of multiple regulatory factors, including Ski and SnoN or processes, such as ubiquitination [10][11][12][13][14].

[1] 162w Pre-, co-or post-treatment with TGF-b1, and to a lesser extent TGF-b3, have been shown to protect cultured rat cortical or chick telencephalic neurons against the toxicity of glutamate [61,64], the excitatory amino acid massively released during ischemic processes and responsible for excitotoxic necrosis. In initial studies, TGF-b has been shown to exert opposite actions in slowly-triggered versus rapidlytriggered excitotoxicity. Indeed, in in vitro paradigms of rapidly-triggered/acute necrosis (short application of high doses of excitotoxins, i.e. agonists of the glutamatergic ionotropic receptors), TGF-b1 protects hippocampal and cerebellar neurons against N-methyl-D-aspartate (NMDA) and kainate toxicity [65][66][67] and cortical neurons against NMDA [68]. Inversely, the toxicity induced by a moderate and prolonged stimulation of non-NMDA glutamatergic ionotropic receptors is exacerbated by TGF-b1 in cultured hippocampal neurons [67]. It has also been suggested that TGF-b could be toxic to neurons by itself, through a potentiation of NMDA receptor-dependent neurotransmission, by acting directly on neurons [69] or indirectly via an alteration of glutamate metabolism in astrocytes [70,71].

[2] 203w As for us, we have demonstrated, in primary murine cortical cultures, that TGF-b1 exerts a selective neuroprotective activity against NMDA receptor-mediated excitotoxicity, with no effect against AMPA or kainate toxicity. This neuroprotective activity requires the obligatory induction of the synthesis and release of the type 1 plasminogen activator inhibitor (PAI-1) upon activation of TGF-b receptors and through a recruitment of Smad3 at the astrocytic level [72][73][74] (Fig. 1). PAI-1 is the main physiological inhibitor of the serine protease tPA, the only approved acute thrombolytic therapy for ischemic stroke [40]. In contrast to this beneficial vascular effect, animal studies have demonstrated that within the cerebral parenchyma, tPA exacerbates ischemic and excitotoxic damages to the brain [75][76][77][78]. As a correlate, PAI-1 deficient mice are more sensitive to experimental paradigms of acute brain injuries than wild-type mice [79]. In vitro, depolarized neurons release tPA that cleaves the NR1 subunit of NMDA receptors, leading to increased NMDA-evoked calcium influx and enhanced neuronal death [78,80]. This effect is blocked by the addition of recombinant PAI-1. Altogether, our data suggest that TGF-b protects neurons against excitotoxicity, by inducing the synthesis of PAI-1 in astrocytes, which in turn, prevents the deleterious proteolysis of NMDA receptors induced by tPA (Fig. 1).

[3] 115w In vitro, the three isoforms of TGF-b accelerate apoptosis of immature cerebellar granule neurons not maintained under depolarizing conditions [81]. In contrast, a pretreatment with TGF-b1 limits apoptotic cell death of primary cultured rat hippocampal neurons induced by trophic factor removal [65] or staurosporin [62,82]. Of note, in the same studies, TGF-b1 was, respectively, inefficient or less effective when co-applied during the challenge, suggesting the requirement of a protein neosynthesis for this cytokine to act as an anti-apoptotic agent. In contrast to these reports, the same doses of TGF-b1 (co-administration) fail to alter serum deprivation-or staurosporin-induced apoptosis of mouse cortical neurons [72]. Whether these discrepancies reflect species and/or neuronal type specificities remains to be investigated.

[4] 156w The anti-apoptotic activity of TGF-b has been related to its ability to stabilize Ca 2+ homeostasis and to increase the expression of anti-apoptotic proteins, such as Bcl-2 and Bcl-xL [65]. More recently, in cultured hippocampal neurons during staurosporin exposure, TGF-b1 has been shown to reduce the extent of caspase-3 activation, a major executioner of apoptosis [82]. In vivo, adenovirus-mediated over-expression of TGF-b1 reduces ischemic brain damages and neurological deficits, an effect associated with a reduction of the expression of the pro-apoptotic protein Bad and of the activation of caspase-3 [83] (Fig. 2). These authors have suggested that the effect of TGF-b1 on Bad expression and phosphorylation is linked to the activation of extracellular-signal related kinases-1 and 2 (Erk1/2) in neurons [83]. In addition, they have recently proposed that the anti-apoptotic effect of TGF-b1 in cultured hippocampal neurons depends on both PI3 kinase/Akt and Erk1/2 pathways and is a process driven by the activation of NF-k-B [84].

[1] 133w There is already compelling evidence that a variety of cytokines contribute to the progression of both acute and chronic neurodegenerative disorders. However, their mechanisms of action remain generally unclear. Among these cytokines, the role and the mechanisms of action of TGF-b1 has been extensively investigated, in particular in experimental stroke. It has thus been demonstrated that TGF-b1 could protect neurons against both apoptosis (by a mechanism associated with the MAPK pathway and members of the Bcl-2 family) and excitotoxicity (through the astrocytic Smad-dependent expression of PAI-1). Though far less documented, TGF-b1 might also prevent amyloid-b neurotoxicity. It thus appears that a better knowledge of the effect and mechanisms of action of this cytokine might offer interesting targets for the development of pharmacological drugs for the treatment of both acute and chronic CNS pathologies.

[1] 36w Like other members of the TGF-b superfamily, including bone morphogenetic proteins (BMPs), activins or Glial cell line-derived neurotrophic factor and related proteins [15][16][17], TGF-b has received much attention as a potential therapeutic target for brain disorders.

[2] 138w In animals and humans, from early embryonic stages, all three mammalian isoforms of TGF-b (TGF-b 1-3 ) are expressed (mRNA and protein) in the central nervous system (CNS) [18][19][20][21]. Interestingly, in the healthy adult brain, TGF-b 2 and 3 isoforms account for almost all the TGF-b immunoreactivity, while TGF-b1 is virtually absent apart from within the meninges and choroid plexus [19,20,22], though TGF-b1 mRNA is constitutively expressed in some brain regions, such as the hippocampus, cortex and hypothalamus [23][24][25]. Similarly, the cognate TGF-b receptors (TbR-I and TbR-II) are widely expressed in the CNS, both during development and in the adulthood. For instance, both TbR-I and TbR-II mRNAs can be found in the cortex and hippocampus [24,[26][27][28][29][30]. There is no detailed brain map of Smad expression, but, for instance, Smad3 mRNA is detected in the hippocampus and cortex [31].

[3] 74w At the cellular level, in vivo and in vitro evidence suggest that all CNS cell types can be a source of and respond to TGF-b. Several neuronal populations, prominently large neurons (cortex, hippocampus, brainstem and spinal cord), express both TGF-b receptors, isoforms À2 and À3 of TGFb and smad3 [20,24,[31][32][33][34]. Similarly, throughout the CNS, TGF-b receptors and isoforms À2 and À3 of TGF-b have been found in glial cells, including astrocytes and microglia [20,24,[33][34][35][36].

[4] 151w Ischemic stroke, the third leading cause of death in industrialised countries, represents a major economic burden to society and often leads to devastating invalidating handicaps in patients [37]. The vascular occlusion induces an important reduction of cerebral blood flow, thereby cutting off energy supply to levels potentially lethal to brain cells. Tissue infarction is irremediable in the direct vicinity of the occluded vessel (ischemic core), while perturbations in the surrounding area (penumbra) are potentially reversible. Several patho-physiological mechanisms contribute to this progression of injury, including hemodynamic modifications, electrophysiological perturbations, neurotransmission alterations and altered gene expression, ultimately leading to cell death through two main ways, namely excitotoxicity and apoptosis [38,39]. To date, two therapeutic approaches have been employed for the treatment of stroke: neuroprotection, which has failed in clinical trials, and thrombolysis, by using tissue-type plasminogen activator (tPA), the only strategy recommended by the National Institute of Neurological Disorders and Stroke [40].

[5] 216w TGF-b has been defined as an injury-related peptide, based on the observation that though it is virtually absent from the normal brain, plethora of reports evidence an increased expression of TGF-b1 mRNA and protein following experimental hypoxia [41,42], global [43][44][45][46][47] or focal [48][49][50][51][52][53] ischemia. TGF-b2 expression is also upregulated [24,46,51], while controversial data have been published regarding TGF-b3 and TGF-b receptors [24,46,51]. There is also no consensual answer regarding the cellular sources of TGF-b1 in the ischemic brain, since astrocytes [46,51,54], neurons [51,54], activated microglia/ infiltrating macrophages [43,49,51,54] and endothelial cells [51,54] have all been reported to secrete this cytokine under ischemic conditions. As a clinical correlate, TGF-b1 has been reported to be up-regulated in the brain [54,55] and in the cerebro-spinal fluid [55], but unchanged or even decreased in the serum [56,57] of stroke patients. As suggested by Krupinski et al. [54], we have demonstrated in the baboon brain, by combining magnetic resonance imagery, positon emission tomography, histology and semi-quantitative reverse transcription-polymerase chain reaction analysis, that ischemia-induced up-regulation of TGF-b1 expression occurs in an area of moderate hypometabolism, characterized by a mixed histological profile, corresponding to the so-called penumbra [58]. This suggests that the expression of TGF-b may represent an endogenous adaptative response of the brain that could be targeted for a neuroprotective therapeutic strategy.

[6] 166w The effect of TGF-b on the extent of brain damages has been investigated soon after the demonstration of an endogenous production of this cytokine in response to ischemia. The intra-carotid injection of TGF-b1 prior to [59] but not after [60] the administration of an autologous thrombus in rabbits reduces by around 50% the lesion volume, an effect independent on an improvement of cerebral blood flow. Accordingly, intra-cerebroventricular injection of TGF-b1 before the induction of focal ischemia in mice [61] or global ischemia in rats [62] leads to a moderate reduction of brain lesions. Moreover, adenovirusmediated over-expression of TGF-b1 is also beneficial following transient ischemia in mice [63]. In addition to these studies addressing the impact of exogenously applied TGF-b, we have generated a soluble receptor to acutely block the action of TGF-b produced in response to the ischemic insult. This TGF-b antagonist markedly exacerbates transient focal ischemic damages in rats, demonstrating that the endogenous TGF-b production serves a potent intrinsic protective response of the brain [53].

[7] 49w While in vivo studies concourse to evidence a beneficial activity of exogenous and endogenous TGF-b in ischemic processes, there is no consensus regarding the potential mechanism involved in this neuroprotection. The differences in the proposed mechanisms in the literature might in fact reflect the pleiotropic actions of this cytokine.

[8] 223w Alzheimer's disease (AD) is characterized by a progressive cognitive impairment leading to dementia. Two neuropathological features characterize the AD brain: amyloid plaques and neurofibrillary tangles [85][86][87]. Plaques are mostly formed of extracellular deposits of amyloid-b peptide (Ab), which is derived from the processing of a transmembrane protein, the amyloid precursor protein (APP) [85,86]. Neurofibrillary tangles correspond to intracellular accumulation of fibrils called paired helical filaments. These are composed of hyper-and abnormally-phosphorylated tau protein aggregates [87]. Although these two hallmarks of Alzheimer's disease are extensively studied, it is not clear whether they are the causes or markers of AD. Familial Alzheimer's disease (FAD) cases are linked to three genes: the amyloid precursor protein gene, the presenelin 1 (PS1) gene and the presenilin 2 (PS2) gene [85]. Most mutations associated with these genes lead to an increased production of Ab peptide and an early onset of the symptoms. Because late onset sporadic AD displays identical characteristics to FAD, all suggest common pathogenic pathways for both forms of AD. However, it is not possible to exclude that non-genetic factors could also influence the amyloid plaque and tangle formation, and thus could play important roles in the genesis or progression of AD. Most of the therapeutic proposals, which are based on the treatment of AD-caused disorders slow the progression of the disease but do not halt it.

[9] 372w Several data show that TGF-b expression is decreased in the plasma [88,89], but increased in the cerebrospinal fluid (CSF) and intrathecal compartment [90,91] but also within the brain parenchyma of AD patients [92][93][94][95]. Similarly, both receptors are up-regulated in the brain [96]. TGF-b immunoreactivity is found in both amyloid plaques and tangles [92], with TGF-b1 isoform mainly found in plaques and the type 2 isoform in glial cells [93,94]. These data have also been observed in a transgenic mouse model, which over-expresses the human amyloid precursor protein carrying FAD mutations (Tg2576). These mice show an astroglial overproduction of TGF-b around amyloid deposits [97]. The consequences of this increased expression of TGF-b in the brain of AD patients have been studied in another transgenic mouse model, over-expressing TGF-b1 in astrocytes. These mice develop vascular Ab deposits at 16 month old, suggesting that increased levels of TGF-b are correlated with the degree of cerebral amyloid angiopathy observed in AD brains [98]. In the same study, it has been observed that double transgenic mice which over-express both human APP and TGF-b1 exhibit earlier Ab deposits around the vessels than single APP transgenic mice, suggesting that TGF-b1 is able to increase APP metabolism or processing [98]. TGF-b has also been shown to enhance the formation of cerebral amyloid deposits in the rat when co-injected with Ab in the brain [99], and to increase the number of Ab plaque like-deposits in hippocampal slices [100]. Accordingly, several in vitro studies have reported that TGF-b up-regulates the expression of APP, either in rodent [101,102] or human [102][103][104] astrocytes. Two mechanisms have been proposed to explain this phenomenon. The first implies a stabilization of APP mRNA [103] and the second a stimulation of the APP promoter [102,104]. We have shown that this over-expression of APP in astrocytes is mediated by a region located in the 5 0 UTR of the promoter of APP between À71 and À54 [105]. Importantly, the positive effect of TGF-b on APP transcription in astrocytes leads to an accumulation of Ab [102]. In addition to this TGF-b signalling-mediated transcriptional effect on Ab accumulation, recent data suggest that TGF-b and amyloid-b peptide can interact and favour the fibrillogenesis of amyloid-b peptide in a TGF-b receptor-independent mechanism [106].

[10] 106w Only few studies have addressed the effect of TGF-b against Ab-induced neuronal death, but they generally report a beneficial activity. In rat hippocampal neurons, TGF-b1 exerts a beneficial effect against both Ab 1-40 and Ab [25][26][27][28][29][30][31][32][33][34][35] , possibly through a preservation of the mitochondrial potential [107]. This observation has been confirmed by the team of Flanders [108][109][110], which has suggested that TGF-b could protect neurons against Ab-induced neuronal death through an up-regulation of the anti-apoptotic genes bcl-xL and bcl-2 [110]. In addition, it has been suggested that TGF-b could reduce neuronal loss, by promoting the clearance of parenchymal amyloid-b peptide through the activation of microglia [111].