Neuroendocrine cells of nasal mucosa are a cellular source of brain-derived neurotrophic factor.

There is growing evidence for extensive interaction between sensory neurons, immune and mucosal epithelial cells during airway inflammation and hyperreactivity. This neuro-immune cross-talk (neurogenic inflammation) involves different groups of mediators, which include the neurotrophin family (nerve growth factor, brain-derived neurotrophic factor (BDNF) and neurotrophin-3 and -4). Neurotrophins modulate airway inflammation by enhancing sensory nerve excitability and production of neuropeptides, and by interaction with different immune cell types. In the present study, it was questioned whether airway epithelial cells express BDNF, and if proinflammatory cytokines (tumour necrosis factor-alpha, interleukin-1beta and interferon-gamma) and a glucocorticoid (budesonide) affect this expression. Primary cultures of nasal epithelial cells were used. It was found that BDNF was stored in chromogranin A-containing secretory granules of specialised epithelial cells, i.e. neuroendocrine cells, and was secreted in a polarised manner. Apical secretion appears to be constitutive, whereas basolateral secretion is markedly enhanced upon stimulation with cytokines. This enhanced basolateral secretion was not due to enhanced synthesis and was not affected by inhibitors of the processing enzymes, such as furin and the metalloproteinases involved in the maturation of BDNF, but was considerably diminished by budesonide. Therefore, airway mucosa might contribute to neurogenic inflammation through increased secretion of brain-derived neurotrophic factor by neuroendocrine cells under inflammatory conditions.

BDNF and DPP-IV in polyps and middle turbinates epithelial cells.

HYPOTHESIS: Neuropeptides released from sensory nerves may contribute to airway inflammation, particularly if their metabolism is impaired through defective inactivation and/or increased production. In the airways, neuropeptides are subjected to degradation by enzymes such as dipeptidyl peptidase (DPP-IV), and are upregulated by neurotrophins such as brain derived neurotrophic factor (BDNF). We therefore assessed in primary human nasal epithelial cells the expression of DPP-IV and BDNF under basal and inflammatory conditions. METHODS: Human epithelial cells were isolated from nasal polyps and middle turbinates, and grown on collagen-coated polycarbonate filters with an air liquid-interface. After three weeks of culture, constitutive cellular expression of DPP-IV and BDNF was assessed by measuring its activity and by ELISA, respectively. To mimick in vivo inflammatory conditions, cells were exposed apically and basolaterally to 50 ng/ml of TNFalpha, IL-1beta, and IFN-gamma for two days. DPP-IV activity and BDNF protein expression were measured in cell lysates and in the apical and basolateral media. RESULTS: Constitutive DPP-IV activity was similar in polyps and turbinates cells. In contrast, polyps epithelial cells expressed higher amounts of BDNF compared to turbinates derived cells. On the other hand, TNFalpha, IL-1beta, and IFN-gamma did not affect DPP-IV activity but significantly increased the cellular expression and the basolateral secretion of BDNF. CONCLUSIONS: Our data show for the first time that primary human airway epithelial cells produced DPP-IV and BDNF under basal conditions. Furthermore, the production and secretion of BDNF were markedly increased in response to pro-inflammatory cytokines, confirming the involvement of BDNF in airway inflammation.

Nasal polyps and middle turbinates epithelial cells sensitivity to amphotericin B.

HYPOTHESIS: Intranasal application of the antimycotic agent amphotericin B (AmphoB) has been proposed as an effective treatment of chronic rhinosinusitis (CRS) with polyps. AmphoB is a sterol-binding agent known to modify cell membrane structure. The cytotoxic effects of AmphoB were studied on primary human nasal epithelial cells in vitro. METHODS: Human epithelial cells were isolated from nasal polyps and middle turbinates of patients suffering from CRS, and grown on collagen-coated polycarbonate filters with an air liquid-interface. After 15 days of culture, cells were exposed apically to 50 microM AmphoB during 4 h daily for 5 days. Some cells were treated during 4 weeks. The bioelectric properties of cells were then studied in Ussing chambers. Integrity of the cell monolayers was assessed by measurement of the transepithelial resistance (R) and immunofluorescent localization of the tight junction protein occludin. RESULTS: Disruption of the epithelial monolayer integrity was observed in all of the nasal polyps cell cultures, as demonstrated by a 60% drop in R. Immunofluorescence microscopy showed significant loss in cell number and disruption in the distribution of occludin. Turbinate cell cultures elicited no change in R and expression of occludin after AmphoB treatment. However, the transepithelial potential, the basal short-circuit current and the amiloride-sensitive current were reduced by 70%. CONCLUSIONS: AmphoB was cytotoxic for nasal polyp epithelial cells with disruption of the epithelium integrity and loss of tight junctions. In contrast, integrity of turbinate epithelial cells was conserved despite alterations in transepithelial ion transport. These observations may explain the beneficial effect of intranasal application of AmphoB on CRS observed in clinical trials.

N-acetylcysteine augments adenovirus-mediated gene expression in human endothelial cells by enhancing transgene transcription and virus entry.

BACKGROUND: It has previously been shown that oxidants reduce the efficiency of adenoviral transduction in human umbilical vein endothelial cells (HUVECs). In this study, the effect of the antioxidant N-acetylcysteine (NAC) in adenovirus-mediated gene transfer has been investigated. METHODS: HUVECs were pretreated or not with NAC, and infected with E1E3-deleted adenovirus (Ad) containing the LacZ gene expressed from the RSV-LTR promoter/enhancer in the presence and absence of NAC. Transgene expression was assessed at the protein level (histochemical staining, measurement of beta-Gal activity, and western blot), mRNA level (real-time RT-PCR) and gene level (nuclear run on) 24 h and 48 h after infection. Adenoviral DNA was quantitated by real-time PCR, and cell surface expression of Coxsackie/adenovirus receptors (CAR) was determined by FACS analysis. RESULTS: Pretreatment of cells with NAC prior to Ad infection enhanced beta-Gal activity by two-fold due to an increase in viral DNA, which was related to increased CAR expression. When NAC was present only during the post-infection period, a five-fold increase in beta-Gal activity and LacZ gene transcriptional activity was observed. When NAC was present during both the pretreatment and the post-infection period, beta-Gal activity was further enhanced, by 15-fold. Augmentation of beta-Gal activity was paralleled by an increase in beta-Gal protein and mRNA levels. NAC did not affect the half-life of LacZ mRNA. CONCLUSION: Pretreatment with NAC prior to Ad infection enhances virus entry, while treatment with NAC post-infection increases transgene transcription. This strategy permits the use of lower adenoviral loads and thus might be helpful for gene therapy of vascular diseases.

Effects of first generation E1E3-deleted and second generation E1E3E4-deleted/modified adenovirus vectors on human endothelial cell death.

Adenoviral vectors are promising tools for pulmonary vascular gene transfer. In first generation vectors, the viral E4 region is preserved (E4+ Ad), but E4 is deleted in second generation vectors (E4- Ad). These vectors were compared for their toxicity in human endothelial cells in terms of apoptosis and necrosis. Infection with E4+ Ad vectors reduced whereas E4- Ad vectors enhanced apoptosis under normal culture conditions. Furthermore, E4+ Ad protected against apoptosis induced by growth factor deprivation, while E4- Ad enhanced apoptosis triggered by ceramide. Ad vectors containing different E4 open reading frames, alone or in different combinations, showed similar effects to E4- Ad, leaving the viral genes that might be responsible for reducing apoptosis unidentified at the present time. As previously observed with E4+ Ad devoid of transgene, E4+ Ad carrying beta-galactosidase or green fluorescent protein under the control of either the RSV or CMV promoter also reduced apoptosis triggered by growth factor deprivation. In contrast, E4+ Ad containing a CFTR expression cassette did not reduce apoptosis, and E4- Ad with CFFR showed increased toxicity. We conclude that Ad vectors may have important effects on the control of apoptosis in transfected cells, depending on the residual expression of viral genes. This effect can be complicated by the action of transgene expression on cell survival.

Reactive oxygen metabolites increase mitochondrial calcium in endothelial cells: implication of the Ca2+/Na+ exchanger.

In endothelial cells, a bolus of hydrogen peroxide (H2O2) or oxygen metabolites generated by hypoxanthine-xanthine oxidase (HX-XO) increased the mitochondrial calcium concentration [Ca2+]m. Both agents caused a biphasic increase in [Ca2+]m which was preceded by a rise in cytosolic free calcium concentration [Ca2+]c (18 and 6 seconds for H2O2 and HX-XO, respectively). The peak and plateau elevations of [Ca2+] were consistently higher in the mitochondrial matrix than in the cytosol. In Ca2+-free/EGTA medium, the plateau phase of elevated [Ca2+] evoked by H2O2 due to capacitative Ca2+ influx was abolished in the cytosol, but was maintained in the mitochondria. In contrast to H2O2 and HX-XO, ATP which binds the P2Y purinoceptors induced an increase in [Ca2+]m that was similar to that of [Ca2+]c. When cells were first stimulated with inositol 1,4, 5-trisphosphate-generating agonists or the Ca2+-ATPase inhibitor cyclopiazonic acid (CPA), subsequent addition of H2O2 did not affect [Ca2+]c, but still caused an elevation of [Ca2+]m. Moreover, the specific inhibitor of the mitochondrial Ca2+/Na+ exchanger, 7-chloro-3,5-dihydro-5-phenyl-1H-4.1-benzothiazepine-2-on (CGP37157), did not potentiate the effects of H2O2 and HX-XO on [Ca2+]m, while causing a marked increase in the peak [Ca2+]m and a significant attenuation of the rate of [Ca2+]m efflux upon addition of histamine or CPA. In permeabilized cells, H2O2 mimicked the effects of CGP37157 causing an increase in the basal level of matrix free Ca2+ and decreased efflux. Dissipation of the electrochemical proton gradient by carbonylcyanide p-(trifluoromethoxy) phenylhydrazone (FCCP), and blocade of the Ca2+ uptake by ruthenium red prevented [Ca2+]m increases evoked by H2O2. These results demonstrate that the H2O2-induced elevation in [Ca2+]m results from a transfer of Ca2+ secondary to increased [Ca2+]c, and an inhibition of the Ca2+/Na+ electroneutral exchanger of the mitochondria.

Clusterin gene expression mediates resistance to apoptotic cell death induced by heat shock and oxidative stress.

Clusterin is a widely expressed, well conserved, secreted glycoprotein, which is highly induced in tissues regressing as a consequence of apoptotic cell death in vivo. It has recently been shown that clusterin expression is only confined to surviving cells following the induction of apoptosis in vitro, suggesting that it is involved in cell survival rather than death. In the hypothesis that clusterin may be implicated in cellular responses to stress, clusterin gene expression was analyzed in the A431 human epidermoid cancer cell line following heat shock and oxidative stress. Our results show that both a transient heat shock (20 min at 42 degrees C) and various oxidative stresses, including hydrogen peroxide, superoxide anion, hyperoxia and UVA exposure, induce a strong increase in clusterin mRNA levels as assessed by northern blot. Nuclear run-on analysis suggests that transcriptional activation is involved in inducing clusterin mRNA in response to heat shock. Using pulse-chase analysis of control and heat shocked cells, it is shown that clusterin mRNA is translated and secreted, thus resulting in increased extracellular levels of the protein following heat shock. To investigate the function of clusterin in response to these stresses, clusterin anti-sense transfectants that stably express virtually no clusterin at the mRNA and protein level were generated in A431 cells. These anti-sense transfectants are shown to be highly sensitive to apoptotic cell death induced by heat shock or oxidative stress compared with wild-type A431 cells or control transfectants. Taken together, our results show that clusterin gene expression is induced in response to heat shock and oxidative stress in human A431 cells, and confers cellular protection against heat shock and oxidative stress.

Modulation of the DNA binding activity of transcription factors CREP, NFkappaB and HSF by H2O2 and TNF alpha. Differences between in vivo and in vitro effects.

Human endothelial cells exposed to H2O2 showed reduced CREP DNA binding activity, enhanced HSF activation, and no induction of NFkappaB binding activity. Interestingly, H2O2 was able to induce NFkappaB subunit p65 translocation in the nucleus. In contrast, cells exposed to TNF alpha showed enhanced CREP binding activity, activation of NFkappaB and no induction of HSE-HSF complex. Addition of H2O2, diamide and iodoacetic acid to the binding reaction mixture markedly reduced the DNA binding ability of the three transcription factors. Thus free sulfhydryls were important in DNA binding activity of CREP, NFkappaB and HSF, and the lack of induction of NFkappaB by H2O2 in intact cells was likely caused by oxidation on a thiol, and not by a deficiency in the activation pathway.

Hyperoxia, unlike phorbol ester, induces glutathione peroxidase through a protein kinase C-independent mechanism.

Human selenium-dependent glutathione peroxidase (GP) is implicated as a mechanism of resistance against oxygen free radicals. The 5' flanking sequence upstream from the coding region of GP contained an oxygen-responsive element termed ORE1 that is responsive to hypoxia, as well as several copies of the activator protein-1 (AP-1)- and AP-1-like-binding sites. In this study, we sought to define the molecular events that lead to GP gene transcription in response to hyperoxia in human umbilical-vein endothelial cells, and asked whether such induction is mimicked and sustained by activation of protein kinase C (PKC) by phorbol esters. Treatment of cells with 100 nM phorbol 12,13-dibutyrate (PdBu) induced a delayed (24-48 h) but significant (2-fold) increase in steady-state GP mRNA levels. Steady-state GP mRNA levels also rose after exposure to 95% O2, again after considerable delay (48-72 h). For both PdBu and oxygen, induction was transcriptionally regulated, as demonstrated by nuclear run-on experiments. The simulations by PdBu and oxygen were additive. In contrast with PdBu, hyperoxia did not stimulate translocation of PKC from the cytosol to the particulate fraction, although the specific activity of both cytosolic and particulate-associated PKC was increased 2-fold in cells exposed to 95% O2 for 5 days. In addition, gel mobility-shift assays using double-stranded tumour-promoting-agent-responsive element (TRE) and nuclear extracts derived from phorbol- and oxygen-treated cells revealed that PdBu, but not hyperoxia, increased AP-1 DNA-binding activity. On the other hand, the up-regulation of GP expression by oxygen could not be accounted for by the ORE1 core sequence, since no specific protein-DNA binding activity could be detected using nuclear extracts from hyperoxic cells and ORE1. Taken together, these results suggest that there may be different molecular mechanisms controlling GP expression. After exposure to PdBu, GP undergoes transcriptional activation via a process that can be readily explained by a classic AP-1 interaction with the TRE sites in the GP promoter. During hyperoxia, GP also undergoes transcriptional activity, but via a process that appears to involve neither TRE nor ORE1.

Differential expression and regulation of hsp70 and hsp90 by phorbol esters and heat shock.

Human peripheral blood monocytes (PBM) produce superoxide anions (O2-.) by a process involving electron transfer from NADPH to O2, catalyzed by the respiratory burst enzyme NADPH oxidase. We have previously shown that phagocytosis, while activating NADPH oxidase, induced in PBM the synthesis of heat shock (HS) proteins (HSP). The present study was undertaken to establish whether this increase in HSP expression was related to O2-. and/or to classical second messengers such as protein kinase C (PKC). Thus, the effects of the PKC activator phorbol 12-myristate 13-acetate (PMA) were compared with those of heat shock on the expression, in PBM, of the major HSP, hsp70 and hsp90, using biometabolic labeling, Western and Northern blotting, and gel mobility shift assays. PMA induced the accumulation of mRNA and an increased expression of hsp90 and, to a lesser extent, hsp/hsc70 (hsc is the cognate, constitutive form). This induction was also observed in PBM from patients with chronic granulomatous disease, a genetic defect in NADPH oxidase, and was abolished by the PKC inhibitors staurosporine and H-7. PMA did not cause activation of the HS factor, and the PMA-induced overexpression expression of HSP was not blocked by the transcriptional inhibitor actinomycin D. HSP-specific mRNA stability was increased after PMA exposure as compared with heat shock. These results suggest that O2-. is not involved in the PMA-mediated induction of hsp70 and hsp90 and that, in contrast to HS, PMA increases the expression of HSP as a result of PKC-induced mRNA stabilization rather than of transcriptional activation of HS genes.

Variable glutathione levels and expression of antioxidant enzymes in human endothelial cells.

We have investigated the relationship between intracellular glutathione levels and the inducibility of the mRNAs encoding the major antioxidant enzymes Cu,Zn superoxide dismutase (Cu,Zn SOD), catalase (CAT), glutathione peroxidase (GP), and the stress protein heme oxygenase (HO) following exposure of human umbilical vein endothelial cells (HUVEC) to either hypoxanthine-xanthine oxidase or 95% O2. Treatment of HUVEC with 2 and 200 microM buthionine sulfoximine (BSO) for 16 h reduced total glutathione (GSH) levels by 51 and 95%, respectively, whereas treatment with 100 microM diethylmaleate (DEM) for 24 h increased the cellular GSH content by 58%. None of these treatments affected the responsiveness of HUVEC to a subsequent oxidant challenge, in terms of antioxidant enzymes activities and mRNA levels. On the contrary, HO mRNA was significantly induced by both BSO and DEM, as well as by hyperoxia, albeit to a different extent. We conclude that intracellular redox changes do not appear to regulate the expression of the mRNAs encoding Cu,Zn SOD, CAT, and GP. Furthermore, factors other than endogenous thiols may play a role in the control of HO mRNA expression.

Differential protective effects of O-phenanthroline and catalase on H2O2-induced DNA damage and inhibition of protein synthesis in endothelial cells.

The respective roles of H2O2 and .OH radicals was assessed from the protective effects of catalase and the iron chelator o-phenanthroline on 1) the inhibition of protein synthesis, and 2) DNA damage and the related events (activation of the DNA repairing enzyme poly(ADP)ribose polymerase with the associated depletion of NAD and ATP stores) in cultured endothelial cells exposed to the enzyme reaction hypoxanthine-xanthine oxidase (HX-XO) or pure H2O2. Catalase added in the extracellular phase completely prevented all of these oxidant-induced changes. O-phenanthroline afforded a complete protective effect against DNA strand breakage and the associated activation of the enzyme poly(ADP)ribose polymerase. By contrast, iron chelation was only partially effective in maintaining the cellular NAD and ATP contents, as well as the protein synthetic activity. In addition, the ATP depletion following oxidant injury was much more profound than NAD depletion. These results indicate that: 1) .OH radical was most likely the ultimate O2 species responsible for DNA damage and activation of poly(ADP)ribose polymerase; 2) both H2O2 and .OH radicals were involved in the other cytotoxic effects (inhibition of protein synthesis and reduction of NAD and ATP stores); and 3) NAD and ATP depletion did not result solely from activation of poly(ADP)ribose polymerase, but other mechanisms are likely to be involved. These observations are also compatible with the existence of a compartmentalized intracellular iron pool.

Differential effects of hyperoxia and hydrogen peroxide on DNA damage, polyadenosine diphosphate-ribose polymerase activity, and nicotinamide adenine dinucleotide and adenosine triphosphate contents in cultured endothelial cells and fibroblasts.

The effects of oxidative stress on DNA damage and associated reactions, increased polyadenosine diphosphate-ribose polymerase (PARP) activity and decreased nicotinamide adenine dinucleotide (NAD) and adenosine triphosphate (ATP) contents, have been tested in primary cultures of porcine aortic endothelial cells. The cells were treated with 50-500 microM H2O2 for 20 min or 100 microM paraquat for 3 days or were exposed to 95% O2 for 2 and 5 days. The administration of 250-500 microM H2O2 resulted in a marked increase in PARP activity and a profound depletion of ATP and NAD. Although hyperoxia had no effect on PARP activity and reduced only slightly the ATP and NAD stores, it markedly reduced the ability of endothelial cells to increase PARP activity upon exposure to DNase. Paraquat had a similar effect. Human dermal fibroblasts were also exposed to 50-500 microM H2O2 for 20 min or 95% O2 for 5 days. Their response to H2O2 differed from that of endothelial cells by their ability to maintain the ATP content at a normal level. Fibroblasts were also insensitive to the effect of hyperoxia. These results suggest that the oxidant-related DNA damage is a function of the type of oxidative stress used and may be cell-specific.

Hypoxanthine-xanthine oxidase-related defect in polypeptide chain initiation by endothelium.

After exposure to O2 intermediates generated by the hypoxanthine-xanthine oxidase system (HX-XO), the rate of [3H]phenylalanine incorporation into total proteins in cultured endothelial cells was markedly reduced. This reduction, which was prevented by catalase, could not be explained by 1) changes in amino acid pools, 2) increased rate of degradation of newly synthesized proteins, 3) impaired poly(A)+ RNA synthesis and efficiency, 4) decreased rate of amino acylation. On the other hand, the increase in the monoribosome-to-polyribosome ratio suggested that translation was affected at the level of chain initiation. Further analysis indicated that 40S initiation complex formation was normal, whereas the assembly of 80S initiation complex was inhibited. Results from reconstitution experiments showed that both normal and treated ribosomes could support normal protein synthesis in the presence of normal initiation factors (IFs). In contrast, IFs from HX-XO lysates did not support normal protein synthesis with ribosomes from either source. Thus, the effect of XO treatment on protein synthesis appears to be an initiation defect related to decreased IF activity and/or availability.

Comparative study on the selenium- and N-acetylcysteine-related effects on the toxic action of hyperoxia, paraquat and the enzyme reaction hypoxanthine-xanthine oxidase in cultured endothelial cells.

The potential protective effect of N-acetylcysteine against various types of oxidative stress (exposure to hyperoxia, treatment with paraquat, incubation in the presence of the hypoxanthine-xanthine oxidase system) was tested in primary cultures of porcine aortic endothelial cells. It was compared to that of selenomethionine (Se-Met), known to increase glutathione peroxidase activity, when given either alone or in combination with N-acetylcysteine. LDH release, 3H-thymidine (TdR) incorporation into DNA and DNA content were measured to assess the cytotoxic effect of the conditions tested. Total and oxidized glutathione content was also determined. Whereas Se-Met had a partial protective effect on all the conditions but paraquat treatment, N-acetylcysteine administration had no effect on the hyperoxia induced changes and significantly worsened the cytotoxic action of paraquat. On the other hand, LDH release following an incubation in the presence of the hypoxanthine-xanthine oxidase was significantly reduced after N-acetylcysteine treatment. No major change in total nor in oxidized glutathione followed N-acetylcysteine treatment in control and experimental conditions. A dose-dependent protective effect of N-acetylcysteine was obtained when this agent was given concomitantly with the xanthine oxidase system. These data suggest that in cultured endothelial cells a N-acetylcysteine-related protective effect, if present, is most likely to result from the direct scavenging action of N-acetylcysteine.

Protein synthesis in hyperoxic endothelial cells: evidence for translational defect.

To study the effects of hyperoxia on protein synthesis in primary cultures of porcine aortic endothelial cells, we exposed confluent cells to different O2 concentrations for various durations. Exposure to 95% O2 for 5 days resulted in a 71% inhibition of [3H]phenylalanine incorporation into total proteins. When compared with control cells, we observed no changes in 1) the pool size of free cytoplasmic phenylalanine and of phenylalanine attached to transfer RNA (tRNA), 2) the rate of protein degradation, and 3) the rate of charging of tRNA with phenylalanine. We found that under hyperoxic conditions 1) the incorporation of [3H]-uridine into total and polyadenylated RNA was increased, 2) the efficiency of extracted messenger RNA to direct protein synthesis in a reticulocyte lysate was maintained, 3) the proportion of polymeric to monomeric ribosomes was slightly increased, and 4) the rate of elongation, as measured by the ribosomal transit time, was decreased. Thus the reduction in protein synthesis in hyperoxic cells appears to result primarily from defects at the translational level in polypeptide chain elongation.

Differential effects of hyperoxia and hydrogen peroxide on thymidine kinase and adenosine kinase activities of cultured endothelial cells.

To compare the respective sensitivity of two nucleoside kinases, adenosine kinase and thymidine kinase, to oxidative stress, we measured these enzyme activities in cultured aortic endothelial cells exposed for 48 h to various O2 concentrations, and in cell extracts treated with H2O2 or the enzyme system hypoxanthine-xanthine oxidase. Adenosine kinase activity was not significantly influenced by the exposure to hyperoxia, nor by treatment with the enzyme system hypoxanthine-xanthine oxidase or with H2O2. On the other hand, there was a dose-dependent inhibitory effect on thymidine kinase activity resulting from exposure to various O2 concentrations or from treatment with various amounts of xanthine oxidase. Incubation of cell extracts in the presence of H2O2 also resulted in a significant reduction of thymidine kinase activity. These results indicate that thymidine kinase exhibits a selective sensitivity to the toxic effect of O2 concentrations and of O2 intermediates such as H2O2.