The effect of interleukin-1beta (IL-1beta) and tumor necrosis factor-alpha (TNF-alpha) on granulocyte macrophage-colony stimulating factor (GM-CSF) production by neuronal precursor cells

AIMS: To determine whether granulocyte macrophage-colony stimulating factor (GM-CSF) production by neuronal precursor (NT2) cells can be regulated by IL-1beta and TNF-alpha. BACKGROUND: We have previously demonstrated GM-CSF expression by neurons of the developing human brain, as well as by NT2 cells. IL-1beta and TNF-alpha upregulate GM-CSF production in glial cells, but GM-CSF regulation in neurons is as yet undefined. We hypothesized that IL-1beta and TNF-alpha would increase GM-CSF mRNA and protein production in NT2 cells. METHODS: The effect of IL-1beta and TNF-alpha on GM-CSF production was assessed by dose response (0 to 2,000 U/ml), and time course (0 to 48 hours incubation) experiments. GM-CSF mRNA and protein production were assessed by quantitative RT-PCR and by ELISA. The effect of these cytokines on cell turnover was determined by BrdU incorporation. RESULTS: IL-1beta increased GM-CSF mRNA and protein expression by NT2 cells. This effect was time and dose dependent, and the effective dose ranging from (20-200 U/ml). TNF-alpha increased GM-CSF mRNA expression to a lesser extent than did IL-1beta (maximal stimulation at 200 U/ml), and a minimal increase in net protein accumulation was noted. Neither cytokine increased NT2 cell turnover. CONCLUSIONS: IL-1beta and TNF-alpha both increase GM-CSF mRNA expression by NT2 cells, but only IL-1beta increases net GM-CSF protein accumulation.     

Site-specific phosphorylation of neurofilament-L is mediated by calcium/calmodulin-dependent protein kinase II in the apical dendrites during long-term potentiation

Neurofilament-L (NF-L), one subunit of the neuronal intermediate filaments, is a major element of neuronal cytoskeletons. The dynamics of NF-L are regulated by phosphorylation of its head domain. The phosphorylation sites of the NF-L head domain by protein kinase A, protein kinase C, and Rho-associated kinase have been previously identified, and those by calcium/calmodulin-dependent protein kinase II (CaMKII) were identified in this study. A series of site- and phosphorylation state-specific antibodies against NF-L was prepared to investigate NF-L phosphorylation in neuronal systems. Long-term potentiation (LTP) is a cellular model of neuronal plasticity that is thought to involve the phosphorylation of various proteins. NF-L is considered a possible substrate for phosphorylation. During LTP stimulation of mouse hippocampal slices, the series of antibodies demonstrated the increase in the phosphorylation level of Ser(57) in NF-L and the visualization of the localized distribution of Ser(57) phosphorylation in a subpopulation of apical dendrites of the pyramidal neurons. Furthermore, Ser(57) phosphorylation during LTP is suggested to be mediated by CaMKII. Here we show that NF-L is phosphorylated by CaMKII in a subpopulation of apical dendrites during LTP, indicating that Ser(57) is a novel phosphorylation site of NF-L in vivo related to the neuronal signal transduction.     

Interleukin-1beta stimulates macrophage inflammatory protein-1alpha and -1beta expression in human neuronal cells (NT2-N)

Chemokines are important mediators in immune responses and inflammatory processes of neuroimmunologic and infectious diseases. Although chemokines are expressed predominantly by cells of the immune system, neurons also express chemokines and chemokine receptors. We report herein that human neuronal cells (NT2-N) produce macrophage inflammatory protein-1alpha and -1beta (MIP-1alpha and MIP-1beta), which could be enhanced by interleukin (IL)-1beta at both mRNA and protein levels. The addition of supernatants from human peripheral blood monocyte-derived macrophage (MDM) cultures induced MIP-1beta mRNA expression in NT2-N cells. Anti-IL-1beta antibody removed most, but not all, of the MDM culture supernatant-induced MIP-1beta mRNA expression in NT2-N cells, suggesting that IL-1beta in the MDM culture supernatants is a major factor in the induction of MIP-1beta expression. Investigation of the mechanism(s) responsible for IL-1beta-induced MIP-1alpha and -1beta expression demonstrated that IL-1beta activated nuclear factor kappa B (NF-kappaB) promoter-directed luciferase activity in NT2-N cells. Caffeic acid phenethyl ester, a potent and specific inhibitor of activation of NF-kappaB, not only blocked IL-1beta-induced activation of the NF-kappaB promoter but also decreased IL-1beta-induced MIP-1alpha and -1beta expression in NT2-N cells. These data suggest that NF-kappaB is at least partially involved in the IL-1beta-mediated action on MIP-1alpha and -1beta in NT2-N cells. IL-1beta-mediated up-regulation of beta-chemokine expression may have important implications in the immunopathogenesis of inflammatory diseases in the CNS.     

Interleukin-1beta activates specific populations of enteric neurons and enteric glia in the guinea pig ileum and colon

Fos expression was used to assess whether the proinflammatory cytokine interleukin-1beta (IL-1beta) activated specific, chemically coded neuronal populations in isolated preparations of guinea pig ileum and colon. Whether the effects of IL-1beta were mediated through a prostaglandin pathway and whether IL-1beta induced the expression of cyclooxygenase (COX)-2 was also examined. Single- and double-labeling immunohistochemistry was used after treatment of isolated tissues with IL-1beta (0.1-10 ng/ml). IL-1beta induced Fos expression in enteric neurons and also in enteric glia in the ileum and colon. For enteric neurons, activation was concentration-dependent and sensitive to indomethacin, in both the myenteric and submucosal plexuses in both regions of the gut. The maximum proportion of activated neurons differed between the ileal (approximately 15%) and colonic (approximately 42%) myenteric and ileal (approximately 60%) and colonic (approximately 75%) submucosal plexuses. The majority of neurons activated in the myenteric plexus of the ileum expressed nitric oxide synthase (NOS) or enkephalin immunoreactivity. In the colon, activated myenteric neurons expressed NOS. In the submucosal plexus of both regions of the gut, the majority of activated neurons were vasoactive intestinal polypeptide (VIP) immunoreactive. After treatment with IL-1beta, COX-2 immunoreactivity was detected in the wall of the gut in both neurons and nonneuronal cells. In conclusion, we have found that the proinflammatory cytokine IL-1beta specifically activates certain neurochemically defined neural pathways and that these changes may lead to disturbances in motility observed in the inflamed bowel.     

Neurofilament deficiency in quail caused by nonsense mutation in neurofilament-L gene

The existence of a neurofilament-deficient mutant of Japanese quail was recently documented (Yamasaki, H., C. Itakura, and M. Mizutani. 1991. Acta Neuropathol. 82:427-434), but the genetic events leading to the neurofilament deficiency have yet to be determined. Our molecular biological analyses revealed that the expression of neurofilament-L (NF- L) gene was specifically repressed in neurons of this mutant. To search for mutation(s) responsible for the shutdown of this gene expression, we cloned and sequenced the NF-L genes in the wild-type and mutant quails. It is eventually found that the NF-L gene in the mutant includes a nonsense mutation at the deduced amino acid residue 114, indicating that the mutant is incapable of producing even a trace amount of polymerization-competent NF-L protein at any situation. The identification of this nonsense mutation provides us with a solid basis on which molecular mechanisms underlying the alteration in the neuronal cytoskeletal architecture in the mutant should be interpreted.     

Toxin B of Clostridium difficile activates human VIP submucosal neurons,in part via an IL-1beta-dependent pathway

This study investigated whether toxin B of Clostridium difficile can activate human submucosal neurons and the involved pathways. Isolated segments of human colon were placed in organ culture for 3 h in the presence of toxin B or IL-1beta. Whole mounts of internal submucosal plexus were stained with antibodies against c-Fos, neuron-specific enolase (NSE), vasoactive intestinal polypeptide (VIP), and substance P (SP). The membrane potential (Vm) response of submucosal neurons to local application of toxin B and IL-1beta was determined by a multisite optical recording technique. Toxin B (0.1 to 10 ng/ml) increased the proportion of c-Fos-positive neurons dose dependently compared with the control. In the presence of toxin B (10 ng/ml), most c-Fos-positive neurons were immunoreactive for VIP (79.8 +/- 22.5%) but only 19.4 +/- 14.0% for SP. Toxin B induced a rapid rise in IL-1beta mRNA level and a sixfold increase in IL-1beta protein in supernatant after 3 h of incubation. c-Fos expression induced by toxin B was reduced dose dependently by IL-1 receptor antagonist (0.1-10 ng/ml). IL-1beta significantly increased c-Fos expression in submucosal neurons compared with the control (34.2 +/- 10.1 vs. 5.1 +/- 1.3% of NSE neurons). Microejection of toxin B had no effect on the Vm of enteric neurons. Evidence of a direct excitatory effect of IL-1beta on Vm was detected in a minority of enteric neurons. Therefore, toxin B of C. difficile activates VIP-positive submucosal neurons, at least in part, via an indirect IL-1beta-dependent pathway.     

IL-1beta regulation of BDNF expression in rat cultured hypothalamic neurons depends on the presence of glial cells

In the present study, we have shown that IL-1beta increased BDNF mRNA expression in hypothalamic neuron-enriched cultures whereas it reduced this expression in mixed cultures, i.e. containing astrocytes and neurons. Because functional relationships between stress and immunity signals are well documented we investigated the possible interaction between BDNF and IL-1beta in hypothalamic neurons. Notably, we investigated whether IL-1beta affected BDNF expression in vitro either on hypothalamic mixed cultures or on neuron-enriched cultures. We found that the response to IL-1beta was stimulatory when directly examined in neurons but was inhibitory when astrocytes were present in the cultures. Since it has been documented that astrocytes release PGE2 in response to IL-1beta, we examined the effect of indomethacin (a PGE2 synthesis inhibitor) on mixed or neuron-enriched cultures treated with IL-1beta. Indomethacin blocked both stimulatory and inhibitory IL-1beta effects on BDNF mRNA expression whereas picrotoxin (a GABA(A) blocker) or MK-801 (a NMDA receptor blocker) had no effect on BDNF mRNA levels. About 3 and 6h treatments of cells with exogenous PGE2 reproduced the effects of IL-1beta on neuron-enriched or on mixed cultures suggesting that PGE2 was involved in BDNF mRNA regulation. Analysis of PGE2 receptors mRNA expression revealed that the PGE2 receptor pattern was changed when neuron-enriched cultures were treated with conditioned medium produced by astrocytes treated with IL-1beta. Thus, EP3 mRNA levels were increased while EP1 and EP4 messengers were unchanged. This increased expression of the inhibitory prostaglandin receptor under astrocyte influence can explain the inhibition of BDNF mRNA levels observed in mixed cultures following IL-1beta or PGE2 treatment. Finally, we demonstrated by immunocytochemistry that EP3 receptors had a neuronal localization in the hypothalamic cultures. Taken together, these data contribute to underline an emerging physiological concept postulating that a same molecule may have opposite effects as a function of the cellular context.     

Expression of low-molecular-weight neurofilament (NF-L) mRNA during postnatal development of the mouse brain

A regional Northern blot analysis demonstrated that the highest levels of NF-L mRNA in the adult mouse brain are present in brain stem followed by mid-brain, with lower levels found in neocortex, cerebellum, and hippocampus. The study was extended to the cellular level over the time course of postnatal development using in situ hybridization. This developmental analysis revealed that the expression of NF-L mRNA closely follows the differentiation pattern of many large neurons during postnatal neurogenesis. Neurons which differentiate early such as Purkinje, mitral, pyramidal, and large neurons of brain stem and thalamic nuclei, expressed high levels of NF-L mRNA at postnatal day 1. Early expression of NF-L mRNA may be required for the maintenance of the extensive neurofilament protein networks that are detected within the axons of larger neurons. Smaller neurons which differentiate later, such as dentate gyrus granule cells, small pyramidal and granule cells of the neocortex, and granule cells of the cerebellum, exhibit a delayed expression of NF-L mRNA.To whom to address reprint requests.     

Peripherin-mediated death of motor neurons rescued by overexpression of neurofilament NF-H proteins

In previous studies, we showed that overexpression of peripherin, a neuronal intermediate filament (IF) protein, in mice deficient for neurofilament light (NF-L) subunits induced a progressive adult-onset degeneration of spinal motor neurons characterized by the presence of IF inclusion bodies reminiscent of axonal spheroids found in amyotrophic lateral sclerosis (ALS). In contrast, the overexpression of human neurofilament heavy (NF-H) proteins provoked the formation of massive perikaryal IF protein accumulations with no loss of motor neurons. To further investigate the toxic properties of IF protein inclusions, we generated NF-L null mice that co-express both peripherin and NF-H transgenes. The axonal count in L5 ventral roots from 6 and 8-month-old transgenic mice showed that NF-H overexpression rescued the peripherin-mediated degeneration of motor neurons. Our analysis suggests that the protective effect of extra NF-H proteins is related to the sequestration of peripherin into the perikaryon of motor neurons, thereby abolishing the development of axonal IF inclusions that might block transport. These findings illustrate the importance of IF protein stoichiometry in formation, localization and toxicity of neuronal inclusion bodies.     

Proteinase-activated receptor-2 induction by neuroinflammation prevents neuronal death during HIV infection

Proteinase-activated receptors (PARs), a newly discovered subgroup of G-protein coupled receptors, are widely expressed by neural cells, but their roles in the nervous system remain uncertain. In this study, we report that PAR-2 was up-regulated on neurons in conjunction with neuroinflammation in brain tissue from patients with HIV-1-associated dementia. The inflammatory cytokines TNF-alpha and IL-1beta were also increased in HIV-1-associated dementia brains compared with patients without dementia (p < 0.05), but these same cytokines induced PAR-2 expression on neurons. Enhanced PAR-2 expression and subsequent activation prevented neuronal cell death and induction of the tumor suppressor, p53, caused by the HIV-encoded protein, Tat (p < 0.01). Intrastriatal implantation of a PAR-2 peptide agonist also inhibited Tat-induced neurotoxicity in a mouse model of HIV neuropathogenesis (p < 0.05). Moreover, PAR-2 null animals showed more severe neuroinflammation and neuronal loss caused by Tat neurotoxicity (p < 0.05). TNF-alpha protected wild-type neurons from Tat-related neurotoxicity, but in PAR-2-deficient neurons, the same concentrations of TNF-alpha were cytotoxic (p < 0.001). Thus, neuroinflammation can exert protective effects by which it induces PAR-2 expression with the ensuing abrogation of neuronal death.     

Loss of SHP-1 phosphatase alters cytokine expression in the mouse hindbrain following cochlear ablation

Inflammatory cytokines in the central nervous system are largely modulated by glial cells and influence neuronal responses to CNS injury. The protein tyrosine phosphatase SHP-1, an intracellular regulator of many cytokine signaling pathways, has been implicated in mediating the activation of glia. There is a direct correlation between abnormally activated microglia and neuron loss within the SHP-1 deficient motheaten (me/me) mouse auditory brainstem after afferent injury. In order to determine whether loss of SHP-1 creates an aberrant cytokine environment driving the abnormal activation of me/me microglia, the expression of interleukin-4 (IL-4), interleukin-10 (IL-10), interleukin-1beta (IL-1beta), tumor necrosis factor-alpha (TNF-alpha), and interferon-gamma (IFN-gamma) was examined by enzyme-linked immunosorbent assay (ELISA). Normal uninjured me/me mice showed lower IL-10 but higher IL-1beta levels compared to wild-type. Following unilateral cochlear ablation, there is decreased expression of IL-4 and IL-10 in me/me brains compared to wild-type, but IL-1beta is significantly increased. These findings indicate that decreases in anti-inflammatory cytokines, in combination with increased expression of the pro-inflammatory cytokine IL-1beta, may initiate a robust inflammatory reaction within the me/me brain contributing to the neuronal degeneration in the deafferented me/me auditory brainstem. SHP-1 may therefore play a role in limiting CNS inflammation following injury and disease.     

Proteinase-activated receptor-2 exerts protective and pathogenic cell type-specific effects in Alzheimer's disease

The proteinase-activated receptors (PARs) are a novel family of G protein-coupled receptors, and their effects in neurodegenerative diseases remain uncertain. Alzheimer's disease (AD) is a neurodegenerative disorder defined by misfolded protein accumulation with concurrent neuroinflammation and neuronal death. We report suppression of proteinase-activated receptor-2 (PAR2) expression in neurons of brains from AD patients, whereas PAR2 expression was increased in proximate glial cells, together with up-regulation of proinflammatory cytokines and chemokines and reduced IL-4 expression (p < 0.05). Glial PAR2 activation increased expression of formyl peptide receptor-2 (p < 0.01), a cognate receptor for a fibrillar 42-aa form of beta-amyloid (Abeta(1-42)), enhanced microglia-mediated proinflammatory responses, and suppressed astrocytic IL-4 expression, resulting in neuronal death (p < 0.05). Conversely, neuronal PAR2 activation protected human neurons against the toxic effects of Abeta(1-42) (p < 0.05), a key component of AD neuropathogenesis. Amyloid precursor protein-transgenic mice, displayed glial fibrillary acidic protein and IL-4 induction (p < 0.05) in the absence of proinflammatory gene up-regulation and neuronal injury, whereas PAR2 was up-regulated at this early stage of disease progression. PAR2-deficient mice, after hippocampal Abeta(1-42) implantation, exhibited enhanced IL-4 induction and less neuroinflammation (p < 0.05), together with improved neurobehavioral outcomes (p < 0.05). Thus, PAR2 exerted protective properties in neurons, but its activation in glia was pathogenic with secretion of neurotoxic factors and suppression of astrocytic anti-inflammatory mechanisms contributing to Abeta(1-42)-mediated neurodegeneration.     

Neurofilament-light increases the cell surface expression of the N-methyl-D-aspartate receptor and prevents its ubiquitination

NMDA (N-methyl-D-aspartate) subtype of glutamate receptors are core components of dendritic spine postsynaptic densities (PSDs), in which they are anchored via their carboxy-terminal tails to cytoskeletal proteins. In this study, we examined the role of the neuronal intermediate filament protein, neurofilament-light (NF-L), also a component of the PSD, in the regulation of NMDA receptor (NMDAR) expression and function in a heterologous system. Coexpression of NF-L with NR1 or NR2B subunits of the NMDAR in HEK293 (human embryonic kidney 293) cells did not result in surface expression as measured by surface biotinylation and cell ELISAs, whereas the combined expression of the three elements resulted in a 20% increase in the surface abundance of NR1, along with a concomitant increase in NMDAR-mediated cytotoxicity. Investigating the origin of this increase, we found that the NR1 subunits are ubiquitinated in HEK293 cells, and that the coexpression of NF-L antagonizes this process. These results suggest a possible means of stabilization of NR1 via its association with NF-L.     

Specific Human Astrocyte Subtype Revealed by Affinity Purified GFAP+1 Antibody; Unpurified Serum Cross-Reacts with Neurofilament-L in Alzheimer

The human GFAP splice variants GFAPΔ164 and GFAPΔexon6 both result in a GFAP protein isoform with a unique out-of-frame carboxy-terminus that can be detected by the GFAP⁺¹ antibody. We previously reported that GFAP⁺¹ was expressed in astrocytes and in degenerating neurons in Alzheimer''s disease brains. In this study we aimed at further investigating the neuronal GFAP⁺¹ expression and we started by affinity purifying the GFAP⁺¹ antibody. The purified antibody resulted in a loss of neuronal GFAP⁺¹ signal, although other antibodies directed against the amino- and carboxy-terminus of GFAPα still revealed GFAP-immunopositive neurons, as described before. With an in-depth analysis of a western blot, followed by mass spectrometry we discovered that the previously detected neuronal GFAP⁺¹ expression was due to cross-reactivity of the antibody with neurofilament-L (NF-L). This was confirmed by double-label fluorescent immunohistochemistry and western blotting with the unpurified GFAP⁺¹ antibody and an antibody against NF-L. Our data imply that NF-L can accumulate in some tangle-like structures in Alzheimer brains. More importantly, the purified GFAP⁺¹ antibody clearly revealed a specific subtype of astrocytes in the adult human brain. These large astrocytes are present throughout the brain, e.g., along the subventricular zone, in the hippocampus, in the striatum and in the spinal cord of controls, Alzheimer, and Parkinson patients. The presence of a specific GFAP-isoform suggests a specialized function of these astrocytes.     

Cellular distribution and developmental expression of AMP-activated protein kinase isoforms in mouse central nervous system

The mammalian AMP-activated protein kinase is a heterotrimeric serine/threonine protein kinase with multiple isoforms for each subunit (alpha, beta, and gamma) and is activated under conditions of metabolic stress. It is widely expressed in many tissues, including the brain, although its expression pattern throughout the CNS is unknown. We show that brain mRNA levels for the alpha2 and beta2 subunits were increased between embryonic days 10 and 14, whereas expression of alpha1, beta1, and gamma1 subunits was consistent at all ages examined. Immunostaining revealed a mainly neuronal distribution of all isoforms. The alpha2 catalytic subunit was highly expressed in neurons and activated astrocytes, whereas the alpha1 catalytic subunit showed low expression in neuropil. The gamma1 noncatalytic subunit was highly expressed by neurons, but not by astrocytes. Expression of the beta1 and beta2 noncatalytic subunits varied, but some neurons, such as granule cells of olfactory bulb, did not express detectable levels of either beta isoform. Preferential nuclear localization of the alpha2, beta1, and gamma1 subunits suggests new functions of the AMP-activated protein kinase, and the different expression patterns and cellular localization between the two catalytic subunits alpha1 and alpha2 point to different physiological roles.