Betalains increase vitexin-2-O-xyloside cytotoxicity in CaCo-2 cancer cells.

Vitexin-2-O-xyloside (XVX) from Beta vulgaris var. cicla L. (BVc) seeds, betaxanthin (R1) and betacyanin (R2) fractions from Beta vulgaris var. rubra L. (BVr) roots were combined and tested for cytotoxicity in CaCo-2 colon cancer cells. XVX was the most cytotoxic molecule, but the combination of XVX with R1 and R2 significantly prolonged its cytotoxicity. Cytotoxicity was mediated by the intrinsic apoptotic pathway, as shown by an increase in Bcl2-like protein 4, cleaved Poly ADP-Ribosyl Polymerase 1 and cleaved Caspase 3 levels with a parallel decrease in anti-apoptotic protein B-cell leukemia/lymphoma 2 levels. R1 and R2, used alone or in combination, reduced oxidative stress triggered by H2O2 in CaCo-2 cells. Betalains dampened cyclooxygenase-2 and interleukin-8 mRNA expression after lipopolysaccharide induction in CaCo-2, showing an anti-inflammatory action. Our results support the use of a cocktail of R1, R2 and XVX as a chemopreventive tool against colon cancer.

Release and toxicity comparison between industrial- and sunscreen-derived nano-ZnO particles.

Many consumer products containing ZnO have raised concern for safety in regard to environmental impact and the public health. Widely used sunscreens for protecting against UV and avoiding sunburns represent a great exposure to nano-ZnO, one of the ingredients commonly applied in sunscreens. Applying nanoproducts on beaches may release nanoparticles unintentionally into the ocean. Despite the accumulation of such nanoproducts in the ocean harming or being detrimental to critical marine organisms, few studies have investigated the release and potential toxicity of nanoparticles extracted from products and compared them with those from industrial-type nanoparticles. Results show that the cytotoxicity of both industrial- and sunscreen-derived nano-ZnO to the marine diatom algae, Thalassiosira pseudonana, increased as exposure increases over time, as measured by growth inhibition (%) of the algae at a constant concentration of nano-ZnO (10 mg/L). The extent of toxicity appeared to be higher from industrial-type nano-ZnO compared with sunscreen-extracted nano-ZnO, though the extent becomes similar when concentrations increase to 50 mg/L. On the other hand, at a fixed exposure time of 48 h, the cytotoxicity increases as concentrations increase with the higher toxicity shown from the industrial-type compared with sunscreen-induced nano-ZnO. Results indicate that while industrial-type nano-ZnO shows higher toxicity than sunscreen-derived nano-ZnO, the release and extent of toxicity from nano-ZnO extracted from sunscreen are not trivial and should be monitored for the development of safe manufacturing of nanomaterials-induced products.

Distinct proteomic profiles characterise non-erosive from erosive reflux disease.

BACKGROUND: Erosive reflux disease (ERD) and non-erosive reflux disease (NERD) are often regarded as part of the spectrum of the same disease. AIM: To elucidate molecular features that characterise NERD and ERD at the protein level. METHODS: A total of 56 consecutive subjects were enrolled: 10 healthy subjects, 24 with NERD and 22 with ERD. Eight specimens were taken from macroscopically normal mucosa at 5 cm of gastro-oesophageal junction. Four were processed for the proteins extraction and four for evaluation using haematoxylin-eosin and immunohistochemistry. We used shotgun proteomics to identify tentative disease molecular features for ERD or NERD. Candidate distinctive proteins were verified using immunohistochemistry. RESULTS: Shotgun proteomics analysis revealed 33 differentially expressed proteins in NERD vs. ERD samples, involved in cellular proliferation, keratinisation, stress responses and sugar metabolism. Based on a gene ontology meta-analysis, seven of them were further analysed using Western blotting (WB) and four also using immunohistochemistry. We identified novel candidate disease molecular features for GERD and few distinctive proteins to discriminate NERD and ERD. In particular, Transitional Endoplasmic Reticulum ATPase (TER ATPase), GAPDH, Alpha 1 Acid Glycoprotein 1, Annexin A1, Calmodulin and 14-3-3 proteins were confirmed at WB analysis. CONCLUSIONS: Non-erosive reflux disease and ERD are distinct disease entities at the protein level. This study proposes an array of candidate biomarkers possibly useful to discriminate between NERD and ERD.

Selective cyclooxygenase-2 silencing mediated by engineered E. coli and RNA interference induces anti-tumour effects in human colon cancer cells.

BACKGROUND: Cyclooxygenase-2 (COX-2) overexpression is strongly associated with colorectal tumourigenesis. It has been demonstrated that the chronic use of non-steroidal anti-inflammatory drugs (COX inhibitors) partially protects patients from colorectal cancer (CRC) development and progression but induces severe cardiovascular side effects. New strategies for selective COX-2 blockade are required. METHODS: We developed an improved technique, based on RNA interference (RNAi), to gain a selective COX-2 silencing in CRC cells by a tumour-dependent expression of anti-COX-2 short-hairpin RNA (shCOX-2). Anti-COX-2 shRNA-expressing vectors were delivered in CRC cells (in vitro) and in colon tissues (ex vivo) using engineered Escherichia coli strains, capable of invading tumour cells (InvColi). RESULTS: A highly tumour-dependent shCOX-2 expression and a significant COX-2 silencing were observed in CRC cells following InvColi strain infection. Cyclooxygenase-2 silencing was associated with a strong reduction in both proliferative and invasive behaviour of tumour cells. We also demonstrated a pivotal role of COX-2 overexpression for the survival of CRC cells after bacterial infection. Moreover, COX-2 silencing was achieved ex vivo by infecting colon tissue samples with InvColi strains, leading to anti-inflammatory and anti-tumour effects. CONCLUSION: Our RNAi/InvColi-mediated approach offers a promising tool for a highly selective COX-2 blockade in vitro and in vivo.

Potential role of the cannabinoid receptor CB in the pathogenesis of erosive and non-erosive gastro-oesophageal reflux disease.

BACKGROUND: Cannabinoid (CB) receptors have been located in brain areas involved in the triggering of TLESRs as well as in the nodose ganglion from which vagal afferents emanate. The distribution of CB(1) receptors has been investigated in the human gastrointestinal mucosa, as expression of inflammatory process. AIM: To evaluate the CB(1) expression in oesophageal mucosa. METHODS: A total of 87 consecutive subjects were enrolled: 10 controls, 39 NERD and 38 erosive oesophagitis. Eight specimens were taken from macroscopically normal mucosa. Five were processed by haematoxylin-eosin, MIB1/CB(1) evaluation and three for the RNA and proteins extraction. RESULTS: The mean MIB1-LI value was 31% and 22% in NERD and ERD patients, respectively, compared to 68% in the healthy subjects. Mean CB(1)mRNA/GUSB mRNA value of the controls was 0.66, while in GERD patients, it was 0.28. In NERD and ERD, the mean values of CB(1)/GUSB were 0.38 and 0.17, respectively, with highly significant differences between the NERD vs. ERD groups. Semi-quantitative analysis of CB(1) expression, performed with WB, shows in NERD patients a higher CB(1) receptor expression than ERD patients. CONCLUSIONS: With this study, we showed for the first time the presence of CB(1) receptors in the human oesophageal epithelium.

RNA interference as a key to knockdown overexpressed cyclooxygenase-2 gene in tumour cells.

Silencing those genes that are overexpressed in cancer and contribute to the survival and progression of tumour cells is the aim of several researches. Cyclooxygenase-2 (COX-2) is one of the most intensively studied genes since it is overexpressed in most tumours, mainly in colon cancer. The use of specific COX-2 inhibitors to treat colon cancer has generated great enthusiasm. Yet, the side effects of some inhibitors emerging during long-term treatment have caused much concern. Genes silencing by RNA interference (RNAi) has led to new directions in the field of experimental oncology. In this study, we detected sequences directed against COX-2 mRNA, that potently downregulate COX-2 gene expression and inhibit phorbol 12-myristate 13-acetate-induced angiogenesis in vitro in a specific, nontoxic manner. Moreover, we found that the insertion of a specific cassette carrying anti-COX-2 short hairpin RNA sequence into a viral vector (pSUPER.retro) greatly increased silencing potency in a colon cancer cell line (HT29) without activating any interferon response. Phenotypically, COX-2 deficient HT29 cells showed a significant impairment of their in vitro malignant behaviour. Thus, the retroviral approach enhancing COX-2 knockdown, mediated by RNAi, proved to be an useful tool to better understand the role of COX-2 in colon cancer. Furthermore, the higher infection efficiency we observed in tumour cells, if compared to normal endothelial cells, may disclose the possibility to specifically treat tumour cells without impairing endothelial COX-2 activity.

Characterization of beta-keratins and associated proteins in adult and regenerating epidermis of lizards.

Reptilian epidermis contains two types of keratin, soft (alpha) and hard (beta). The biosynthesis and molecular weight of beta-keratin during differentiation of lizard epidermis have been studied by autoradiography, immunocytochemistry and immunoblotting. Tritiated proline is mainly incorporated into differentiating and maturing beta-keratin cells with a pattern similar to that observed after immunostaining with a chicken beta-keratin antibody. While the antibody labels a mature form of beta-keratin incorporated in large filaments, the autoradiographic analysis shows that beta-keratin is produced within the first 30 min in ribosomes, and is later packed into large filaments. Also the dermis incorporates high amount of proline for the synthesis of collagen. The skin was separated into epidermis and dermis, which were analyzed separately by protein extraction and electrophoresis. In the epidermal extract proline-labeled proteic bands at 10, 15, 18-20, 42-45, 52-56, 85-90 and 120 kDa appear at 1, 3 and 5 h post-injection. The comparison with the dermal extract shows only the 85-90 and 120 kDa bands, which correspond to collagen. Probably the glycine-rich sequences of collagen present also in beta-keratins are weakly recognized by the beta-1 antibody. Immunoblotting with the beta-keratin antibody identifies proteic bands according to the isolation method. After-saline or urea-thiol extraction bands at 10-15, 18-20, 40, 55 and 62 kDa appear. After extraction and carboxymethylation, weak bands at 10-15, 18-20 and 30-32 kDa are present in some preparations, while in others also bands at 55 and 62 kDa are present. It appears that the lowermost bands at 10-20 kDa are simple beta-keratins, while those at 42-56 kDa are complex or polymeric forms of beta-keratins. The smallest beta-keratins (10-20 kDa) may be early synthesized proteins that are polymerized into larger beta-keratins which are then packed to form larger filaments. Some proline-labeled bands differ from those produced after injection of tritiated histidine. The latter treatment does not show 10-20 kDa labeled proteins, but tends to show bands at 27, 30-33, 40-42 and 50-62 kDa. Histidine-labeled proteins mainly localize in keratohyalin-like granules and dark keratin bundles of clear-oberhautchen layers of lizard epidermis, and their composition is probably different from that of beta-keratin.

Putative histidin-rich proteins in the epidermis of lizards.

In the stratum granulosum of mammalian epidermis, histidin-rich proteins (filaggrins) determine keratin clumping and matrix formation into terminal keratinocytes of the stratum corneum. The nature of matrix, interkeratin proteins in the epidermis of nonmammalian vertebrates, and in particular in that of reptilian, mammalian progenitors are unknown. The present biochemical study is the first to address this problem. During a specific period of the renewal phase of the epidermis of lizards and during epidermal regeneration, keratohyalin-like granules are formed, at which time they take up tritiated histidine. The latter also accumulate in cells of the alpha-keratin layer (soft keratin). This pattern of histidine incorporation resembles that seen in keratohyalin granules of the stratum granulosum of mammalian epidermis. After injection of tritiated histidine, we have analysed the distribution of the radioactivity by histoautoradiography and electrophoretic gel autoradiography of epidermal proteins. Extraction and electrophoretic separation of interfilamentous matrix proteins from regenerating epidermis 3-48 hours post-injection reveals the appearance of protein bands at 65-70, 55-58, 40-43, 30-33, 25-27, and 20-22 kDa. Much weaker bands were seen at 100, 140-160, and 200 kDa. A weak band at 20-22 kDa or no bands at all are seen in the normal epidermis in resting phase and in the dermis. In regenerating epidermis at 22 and 48 hours post-injection, little variation in bands is detectable, but low molecular weight bands tend to increase slightly, suggesting metabolic turnover. Using anti-filaggrin antibodies against rat, human, or mouse filaggrins, some cross-reactivity was seen with more reactive bands at 40-42 and 33 kDa, but it was reduced or absent at 140, 95-100, 65-70, 50-55, and 25 kDa. This suggests that different intermediate degradative proteins of lizard epidermis may share some epitopes with mammalian filaggrins and are different from keratins with molecular weight ranging from 40 to 65-68 kDa. The immunocytochemical observation confirms that a weak filaggrin-like immunoreactivity characterizes differentiating alpha-keratogenic layers in normal and regenerating tail. A weak filaggrin labeling is discernable in small keratohyalin-like granules but is absent from the larger granules and from mature keratinocytes. The present results indicate, for the first time, that histidine-rich proteins are involved in the process of alpha-keratinization in reptilian epidermis. The cationic, interkeratin matrix proteins implicated may be fundamentally similar in both theropsid-derived and sauropsid amniotes.

Impact of IGF-I/IGF-IR circuit on the angiogenetic properties of Ewing's sarcoma cells.

The insulin-like growth factor receptor I (IGF-I)-mediated circuit is a major autocrine loop for Ewing's sarcoma (ES) cells, and plays a role the pathogenesis and malignancy of this tumor. IGF-I receptor (IGF-IR) has emerged as a good therapeutic site for ES patients. In this study, we analyzed the impact of strategies targeting the IGF-IR on the regulation of VEGFs, which are of fundamental importance in angiogenesis, and TGFbeta, CTGF and Cyr61, which are factors primarily involved in skeletal growth control and angiogenesis. IGF-I increases expression of VEGF-A, TGFbeta, CTGF and Cyr61 mRNA. However, only the modulation of VEGF-A expression appears to be mediated by IGF-IR. Functional assays on endothelial cells indicate a strict correlation between survival and proliferation of HUVECs and VEGF-A levels, confirming a major role for this factor in angiogenesis. Blockage of IGF-IR functions by neutralizing antibody or antisense strategies significantly reduced the expression and secretion of VEGF-A by ES cells, and supernatants of treated cells were unable to sustain the survival and proliferation of HUVECs. Analysis of the signaling mechanisms involved in constitutive or IGF-induced expression and secretion of VEGF-A indicated that PI3-K and MAPK signaling pathways are both required for VEGF expression and production in ES cells. Selective inhibitors LY294002 or PD98059 were highly effective in reducing the ability of ES cells to produce VEGF-A and stimulate survival and proliferation of HUVECs. Taken together, these findings add a new activity to the IGF-I repertoire in ES and highlight how disruption of IGF-IR functions may constitute an effective tool for the control of neovascularization in this tumor.

Glyceraldehyde-3-phosphate dehydrogenase is responsible for intranuclear localization of some oligonucleotides.

Nuclear accumulation of ODNs has been associated with their binding to a series of nuclear proteins. These interactions could be responsible for the sequence-independent effects of ODNs as well as for their sequence-specific interactions and their intracellular distribution. Investigation of interaction of ODNs with these proteins may shed light on the mechanisms of cellular uptake and nuclear accumulation of oligonucleotides.

The Rossmann fold of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is a nuclear docking site for antisense oligonucleotides containing a TAAAT motif.

The subcellular localisation of oligodeoxynucleotides (ODN) is a major limitation for their use against nuclear targets. In this study we demonstrate that an antisense ODN directed against cytosolic phospholipase A(2) (cPLA2) mRNA is efficiently taken up and accumulates in the nuclei of endothelial cells (HUVEC), human monocytes and HeLa cells. Gel shift experiments and incubation of cells with oligonucleotide derivatives show that the anti-cPLA2 oligo binds a 37 kDa protein in nuclear extracts. The TAAAT sequence was identified as the major binding motif for the nuclear protein in competition experiments with mutated ODNs. Modification of the AAA triplet resulted in an ODN which failed to localise in the nucleus. Moreover, inserting a TAAAT motif into an ODN localising in the cytosol did not modify its localisation. The 37 kDa protein was purified and identified after peptide sequencing as glyceraldehyde-3-phosphate dehydrogenase (GAPDH). It was shown by confocal microscopy that GAPDH co-localises with anti-cPLA2 ODN in the nucleus and commercial GAPDH effectively binds the oligo. Competition experiments with increasing concentration of NAD(+) co-factor indicate that the GAPDH Rossmann fold is a docking site for antisense oligonucleotides containing a TAAAT motif.

Colocalization prostacyclin (PGI2) synthase--caveolin-1 in endothelial cells and new roles for PGI2 in angiogenesis.

In vascular cells, prostacyclin (PGI2) synthase (PGI2s) has been localized in the endoplasmic reticulum of endothelial cells and in the nuclear and plasma membrane of smooth muscle cells. In human umbilical vein endothelial (HUVE) cells, we detected the enzyme in abundant cytoplasmic vesicles apparently originating from the plasma membrane and similar to those stained by gold-albumin, which interacts with a caveolar receptor. This prompted us to try a direct confocal microscopy approach aimed at colocalizing gold-albumin, caveolin-1, and PGI2 synthase. Moreover, the staining of HUVE cells with an anti-BiP7Grp78 antibody (a marker of endoplasmic reticulum) shows a perinuclear localization, sharply separated from PGI2 synthase localization. The results indicate that more than 80% of the enzyme resides in cellular sites costaining with caveolin-1 antibody and gold-albumin. This evidence was confirmed by the demonstration that PGI2 synthase and caveolin-1 coimmunoprecipitate in HUVE cell lysates and that they are associated to detergent-insoluble membrane domains in the same low-density fractions of a sucrose gradient. In addition, depletion of cellular cholesterol by mevalonate and methyl-beta-cyclodextrin leads to the shift of PGI2 synthase and caveolin-1 to higher density fractions of the gradient. Biochemical evidence about colocalization was supported by the use of a fusion protein glutathione S-transferase (GST)/caveolin-1, which retained either PGI2s purified from ram seminal vesicles or PGI2s present in HUVE cell lysates. Binding of PGI2s to caveolin "scaffolding domain" and to C-terminal region was deduced by using full-length GST--Cav-1, GST--Cav 61--101, and GST C- and N-terminal fusion proteins. A double approach based on the usage of filipin as a specific caveolae-disrupting agent and antisense oligonucleotides targeting PGI2 synthase mRNA suggests that the production of PGI2 in caveolae is likely to be connected to the regulation of angiogenesis, at least in vitro.

Knockdown of caveolin-1 by antisense oligonucleotides impairs angiogenesis in vitro and in vivo.

Knock-out of the gene coding for caveolin-1, the main organizer of caveolae, has not yet been performed. We devised a strategy to knock-down caveolin-1 gene expression using antisense oligodeoxynucleotides (ODNs). Seven ODNs, covering different regions of caveolin-1 mRNA, were screened by Western blot analysis of caveolin-1 levels. The most active and specific was found to reduce caveolin-1 protein levels by 70% at 1 microM concentration and its action, as demonstrated by a marked reduction (about 50%) in caveolin-1 mRNA levels, was due to a true antisense mechanism. In HUVEC treated with the active ODN, caveolae were undetectable by confocal and electron microscopy, while their number was not affected when cells were treated with a scrambled ODN. Using the fibrin gel 3 D angiogenesis test we established that the active (but not the scrambled) ODN strongly suppressed capillary-like tube formation. Moreover, an antisense tailored against chicken caveolin-1 mRNA, when tested using the chorio-allantoic membrane technique, dramatically reduced vessel formation at doses (10-20 microg) under which control ODNs were ineffective and devoid of toxicity. Thus, it is likely that caveolin-1 down regulation, followed by caveolae disruption, impairs angiogenesis in vitro and in vivo.

A 38 kDa nuclear protein is involved in the retention of an antisense oligonucleotide directed against cytosolic phospholipase A2.

Recent studies suggest that antisense phosphorothioate oligonucleotides (APO) are useful tools not only to impair gene expression, but also to modify the splicing of pre-mRNA, as the classical view that they act by suppressing the translation of mature mRNA has been challenged by several examples showing their nuclear site of action. In this work we show that an APO directed against cytosolic phospholipase A2 (cPLA2) mRNA localises in the nucleus and interacts with a specific nuclear protein.

Evidence that photodynamic stress kills Zellweger fibroblasts by a nonapoptotic mechanism.

Zellweger fibroblasts, which are devoid of peroxisomes and fail to synthesize plasmalogens, are very sensitive to the killing effect triggered by UV-activated 12-(1-pyrene) dodecanoic acid (P12). Although in some studied performed, it is assumed that reactive oxygen species (ROS) may damage plasma membrane causing necrosis, other studies suggest that ROS are involved in apoptotic cell death induced by a wide variety of stimuli. Analysing the P12 dose-response in Zellweger fibroblasts, we observed that at high doses (1-2 microM), more than 75% of the cells died after 24 h. This behaviour suggested that, at high doses, P12 kills the cells by unspecific lytic mechanisms or by necrosis, while at low doses (0.1-0.5 microM), an apoptotic mechanism could be involved. Cytofluorimetric analysis of Zellweger fibroblasts-treated with activated P12 (0.5 microM) did not show morphological modifications typical of apoptotic cell death. This was supported by comparative staining of fibroblast nuclei, DNA gel electrophoresis and identification of poly(ADP-ribose) polymerase (PARP) cleavage and Bcl-2 expression, assayed by Western blots. Thus, our results, while confirming the importance of plasmalogens in the protection against ROS, establish that apoptosis is not involved in photodynamic death induced by activated P12. Therefore, we can expect that in gene transfer experiments, the rescue of Zellweger cells will be dependent only on the correction of peroxisomal biogenesis.

Hepatic steatosis in artificially fed marine teleosts.

In a comparison of livers in fish (Sparus auratus and Dicentrarchus labrax) feeding on natural sources of food with livers of artificially fed animals, a much higher C18:1/C22:6 ratio was observed in the latter. Staining livers with oil red O showed extensive steatosis in artificially fed fish, but not in those naturally fed. Juvenile artificially fed fish showed a more extensive steatosis and a higher mortality rate. In steatotic fish fed a natural diet for 2 months, the liver exhibited extensive regeneration and only a few steatotic areas remained. Marine teleosts do not appear to have a proliferative response of peroxisomes and this is likely to contribute to liver lipid accumulation and subsequent steatosis. It is suggested that an excess of C18:1 (or other mono-unsaturated fatty acids), coupled with a lack of adaptive peroxisomal proliferation, is the primary cause of lipid droplet formation leading to hepatic steatosis.

Prostacyclin (PGI2) synthase is a constitutively expressed enzyme in human endothelial cells.

Biogenesis of prostanoids is under the control of some polypeptide growth factors. Cytosolic phospholipase A2, a form specific for arachidonic acid containing phospholipids, is activated by a translocation mechanism regulated by growth factors, while prostaglandin H synthase isoforms are induced de novo in several cell types. No information is available as far as PGI2 synthase is concerned. Human umbilical vein endothelial cells were cultured under conditions favoring proliferation or differentiation or capillary-like network formation in the presence of collagen gels. Basic fibroblast growth factor (bFGF 0.5-4 ng/ml) was used as a mitogen, interleukin-1 alpha (IL-1 alpha 10-60 UI/ml) as a differentiating agent, and prostacyclin (PGI2) biosynthesis was evaluated. Under the first condition, basal PGI2 production was unaffected while, in the presence of IL-1 alpha, a marked stimulation of PGI2 synthesis was observed. It is known that IL-1 alpha is a potent inducer of PGH synthase, while it is not known whether PGI2 synthase is also induced. Two lines of evidence indicate that PGI2 synthase is a constitutively expressed not inducible enzyme: (a) proliferating nonproducing cells when added with PGH2 produce an amount of PGI2 not different from the amount produced by cells stimulated with IL-1 alpha; (b) under this condition PGI2 synthase was immunodetectable either by immunofluorescence detected by confocal microscopy or by ELISA and, on microsomes isolated from endothelial cells, by Western blotting. It is concluded that the limiting step in the conversion arachidonate-PGI2 is represented solely by the level of PGH synthase. These results strongly suggest, but do not prove, the constitutive nature of the enzyme. The final demonstration requires the availability of a probe to detect mRNA level, a trial we are carrying out at the moment.

Possible involvement of prostaglandin H synthase-1 induction in the differentiation of U-937 cells.

The promyelocytic human cell line U937, cultured in the presence of TPA and/or vit. D3, differentiates to monocytes and to macrophage-like cells. A potent stimulus for differentiation is represented also by colony stimulating factor-1 (CSF-1). Since this factor is a strong inducer of PGH synthase in human monocytes, we have investigated whether this event may be connected to the differentiation of U937. We have found that TPA, in the presence of serum, increased the production of thromboxane B2 (TXB2) 4-5 fold, while DMSO, which induced differentiation to neutrophils, was not active. Here we report studies indicating that the effect of protein and RNA synthesis inhibitors on prostanoid production, in cells incubated in the presence of CSF-1 (or FCS), can be correlated with an inductive event carried out by the growth factor, as demonstrated by the use of Western and Northern blotting procedures. However, while in human monocytes PGH-s and its mRNA are absent in controls and are expressed at high levels in CSF-1 stimulated cells, in U937 cells exposed to TPA, PGH-s mRNA was clearly detected by Northern blots, but its translation product was expressed at low level, and cells generated low amounts of TXA2 (13% of maximal production). After incubation with CSF-1 (or FCS) mRNA levels were only slightly modified, but large amounts of TXA2 accumulated in the medium. We have interpreted these findings by suggesting that CSF-1 is capable not only of regulating the expression of the gene encoding PGHs, but also of acing translationally to regulate the expression of its mature mRNA.

Post-transcriptional regulation of thromboxane A2 synthase in U937 cells.

Translation of mRNAs is a process usually tightly coupled to transcription of genes. However, there are examples of mRNA species which accumulate without being translated. Some mRNAs present in oocytes and ferritin mRNA are the most studied models. Studying the biogenesis of thromboxane A2 (TXA2) in the promonocytic line U937, we have noted that in proliferating cells high levels of TXA2 synthase mRNA are detectable by Northern blot, whereas no TXA2 could be recovered in the medium. This has been explained on the basis of Western blot experiments: TXA2 synthase was not detectable in proliferating cells, while a band of about 55 kd appears after treatment with the differentiating agent TPA. Immunofluorescence detection by confocal microscopy was in agreement with Immunoblot results. Thus, in U937 cells, TPA behaves as a regulator of translation of TXA2 synthase mRNA. We have further observed that the induced enzyme in U937 cells has many characteristics in common with the human monocytic enzyme: a long half life (> 24 hrs), a marked stability during catalysis and similar Km and Vmax values. Thus, U937 cells are a good model to study the mechanism by which a mRNA is efficiently translated only after differentiation has been triggered.