Multiple forms of PKR present in the nuclei of acute leukemia cells represent an active kinase that is responsive to stress.

A number of cancers possess constitutive activity of the dsRNA-dependent kinase, PKR. Inhibition of PKR in these cancers leads to tumor cell death. We recently reported the increased presence of PKR phosphorylated on Thr451 (p-T451 PKR) in clinical samples from myelodysplastic syndrome (MDS) patients and acute leukemia cell lines. Whereas p-T451 PKR in low-risk patient samples or PTEN-positive acute leukemia cell lines was mostly cytoplasmic, in high-risk patient samples and acute leukemia cell lines deficient in PTEN, p-T451 PKR was mainly nuclear. As nuclear activity of PKR has not been previously characterized, we examined the status of nuclear PKR in acute leukemia cell lines. Using antibodies to N-terminus, C-terminus and the kinase domain in conjunction with a proteomics approach, we found that PKR exists in diverse molecular weight forms in the nucleus. Analysis of PKR transcripts by reverse transcriptase-PCR, and PKR-derived peptides by MS/MS revealed that these forms were the result of post-translational modifications (PTMs). Biochemical analysis demonstrated that nuclear PKR is an active kinase that can respond to stress. Given the association of PKR with PTEN and the Fanconi complex, these results indicate that PKR likely has other previously unrecognized roles in nuclear signaling that may contribute to leukemic development.

Tear proteomics in evaporative dry eye disease.

PURPOSE: To analyze tear protein variations in patients suffering from dry eye symptoms in the presence of tear film instability but without epithelial defects. METHODS: Five microlitres of non-stimulated tears from 60 patients, suffering from evaporative dry eye (EDE) with a break-up time (BUT) <10 s, and from 30 healthy subjects as control (no symptoms, BUT >10 s) were collected. Tear proteins were separated by mono and bi-dimensional SDS-PAGE electrophoresis and characterized by immunoblotting and enzymatic digestion. Digested peptides were analyzed by liquid chromatography coupled to electrospray ionization quadrupole-time of flight mass spectrometry followed by comparative data analysis into Swiss-Prot human protein database using Mascot. Statistical analysis were performed by applying a t-test for independent data and a Mann-Whitney test for unpaired data (P<0.05). RESULTS: In EDE patients vscontrols, a significant decrease in levels of lactoferrin (data in %+/-SD): 20.15+/-2.64 vs 24.56+/-3.46 (P=0.001), lipocalin-1: 14.98+/-2.70 vs 17.73+/-2.96 (P=0.0001), and lipophilin A-C: 2.89+/-1.06 vs 3.63+/-1.37 (P=0.006) was revealed, while a significant increase was observed for serum albumin: 9.45+/-1.87 vs 3.46+/-1.87 (P=0.0001). No changes for lysozyme and zinc alpha-2 glycoprotein (P=0.07 and 0.7, respectively) were shown. Proteomic analysis showed a downregulation of lipophilin A and C and lipocalin-1 in patients, which is suggested to be associated with post-translational modifications. CONCLUSIONS: Data show that tear protein changes anticipate the onset of more extensive clinical signs in early stage dry eye disease.

Requirement of tyrosine-phosphorylated Vav for morphological differentiation of all-trans-retinoic acid-treated HL-60 cells.

Our previous data demonstrated that cellular and nuclear tyrosine-phosphorylated Vav associate with phosphoinositide 3-kinase during all-trans-retinoic acid-dependent granulocytic differentiation of HL-60 cells. In this study, aimed to analyze the mechanism by which Vav is recruited and activated, we report that the Src homology 2 domain of Vav interacts with tyrosine-phosphorylated proteins in a differentiation-dependent manner. Two adaptor proteins, Cbl and SLP-76, were identified, showing a discrete distribution inside the cells, with Cbl absent from the nuclei and SLP-76 particularly abundant in the nuclear compartment. Of note, Vav interacts with the tyrosine kinase Syk, which is also present in the nuclear compartment and may phosphorylate Vav in vitro when cells differentiate. Inhibition of Syk activity by piceatannol prevents both in vitro and in vivo Vav tyrosine phosphorylation, its association with the regulatory subunit of phosphoinositide 3-kinase, and the nuclear modifications typically observed during granulocytic differentiation of this cell line. These findings suggest that tyrosine-phosphorylated Vav and its association with phosphoinositide 3-kinase play a crucial role in all-trans-retinoic acid-induced reorganization of the nucleoskeleton, which is responsible for the changes in nuclear morphology observed during granulocytic differentiation of HL-60 cells.

Inhibition of phosphoinositide 3-kinase impairs pre-commitment cell cycle traverse and prevents differentiation in erythroleukaemia cells.

During the early hours after exposure to differentiation inducing agents, Friend erythroleukaemia cells undergo alterations which commit them to cessation of growth and development of the characteristics of differentiation. Our current experiments have compared the expression and activity of phosphoinositide 3-kinase (PI 3-kinase) in control cells with cells undergoing differentiation which has been induced by dimethyl sulfoxide (DMSO). When the cultures were initiated with stationary phase cells and DMSO was added at the time of seeding, PI 3-kinase activity was stimulated in both treated and control cells during the first 3 h from seeding. This event appears to be a rate limiting step in commitment since pretreatment of cells with 10 microM LY294002 or down-regulation of p85 expression prior to adding DMSO completely prevents commitment to erythropoiesis. Accordingly, PI 3-kinase inhibition during the commitment period prevents DNA-binding of the transcription factor GATA-1, essential for erythroid differentiation. However, once cells are committed to differentiate, PI 3-kinase activity and expression dramatically decreases along with the differentiation programme, to become barely detectable after 96 h. Remarkably, LY294002 treatment leads to accumulation of cell in G1 phase and prevents DMSO-dependent cyclin D3 induction. Based on these data, we suggest that PI 3-kinase is rate limiting for the completion of the first round cycle of cell division required for initiation of erythrocytic differentiation. On the other hand, the late decrease of PI 3-kinase associated with the differentiation process seems to be part of the programmed shut off of genes not needed in mature erythrocytes.

Phosphatidylinositol 3-kinase translocation to the nucleus is induced by interleukin 1 and prevented by mutation of interleukin 1 receptor in human osteosarcoma Saos-2 cells.

Although interleukin 1 (IL-1) functions have been extensively characterized, the mechanisms by which IL-1 signals are transduced from the plasma membrane to the nucleus are less known. Recent evidence indicates that phosphatidylinositol 3-kinase (PI3-kinase) could be activated by a direct association with the activated IL-1 receptor. In this study we analyzed the effects of IL-1 on the intracellular distribution of PI3-kinase in wild-type Saos-2 human osteosarcoma cells, and in cell clones overexpressing type I IL-1 receptor (IL-1RI). PI3-kinase intracellular distribution displays two distinct patterns. In quiescent cells, PI3-kinase is distributed through the cytoplasm, although a portion is present in the nucleus; following stimulation with IL-1, PI3-kinase is redistributed, increasing in the nuclear compartment. Both immunoblotting and immunofluorescence data indicate that IL-1 causes a rapid and transient translocation of PI3-kinase from the cytoplasm to the nucleus. This phenomenon is prevented by PI3-kinase inhibitors, suggesting that the maintenance of PI3-kinase activity is essential for IL-1-induced translocation. Indeed, in cell clones stably transfected with Y479F receptor mutant, in which the binding of the enzyme to the activated receptor is blocked, IL-1-induced PI3-kinase translocation to the nucleus is completely prevented. These data suggest that PI3-kinase translocation to the nucleus upon IL-1R activation is an early event in IL-1 signaling mechanism, and may be involved in transcriptional activation.

Phosphatidylinositol 3-kinase is recruited to a specific site in the activated IL-1 receptor I.

Interleukin 1 (IL-1) delivers a stimulatory signal which increases the expression of a set of genes by modulating the transcription factor NF-kappaB. The IL-1 receptors are transmembrane glycoproteins which lack a catalytic domain. The C-terminal portion of the type I IL-1 receptor (IL-IRI) is essential for IL-1 signalling and for IL-1 dependent activation of NF-kappaB. This portion contains a putative phosphatidylinositol 3-kinase (PI 3-kinase) binding domain (Tyr-E-X-Met), which is highly conserved between the human, mouse and chicken sequences, as well as the related cytoplasmic domain of the Drosophila receptor Toll. This observation prompted us to investigate the role of PI 3-kinase in IL-1 signalling. Here we report evidence that PI 3-kinase is recruited by the activated IL-IRI, causing rapid and transient activation of PI 3-kinase. We also show that the receptor is tyrosine phosphorylated in response to IL-1. Expression of a receptor mutant lacking the putative binding site for p85 demonstrates that Tyr479 in the receptor cytoplasmic domain is essential for PI 3-kinase activation by IL-1. Our results indicate that PI 3-kinase is likely to be an important mediator of some IL-1 effects, providing docking sites for additional signalling molecules.

Nuclear but not cytoplasmic phospholipase C beta 1 inhibits differentiation of erythroleukemia cells.

A body of evidence has shown the existence of a nuclear phosphoinositide cycle in different cell types. The cycle is endowed with kinases as well as phosphatases and phospholipase C (PLC). Among the PLC isozymes, the beta family is characterized by a long COOH-terminal tail that contains a cluster of lysine residues responsible for nuclear localization. Indeed, PLC beta 1 is the major isoform that has been detected in the nucleus of several cells. This isoform is activated by insulin-like growth factor I, and when this isoform is lacking, as a result of gene ablation, the onset of DNA synthesis induced by this hormone is abolished. On the contrary, PLC beta 1 is down-regulated during the erythroid differentiation of Friend erythroleukemia cells. A key question is how PLC beta 1 signaling at the nucleus fits into the erythroid differentiation program of Friend erythroleukemia cells, and whether PLC beta 1 signaling activity is directly responsible for the maintenance of the undifferentiated state of erythroleukemia cells. Here we present evidence that nuclear PLC beta 1 but not the isoform located at the plasma membrane is directly involved in maintaining the undifferentiated state of Friend erythroleukemia cells. Indeed, when wild-type PLC beta 1 is overexpressed in these cells, differentiation in response to DMSO is inhibited in that the expression of beta-globin is almost completely abolished, whereas when a mutant lacking the ability to localize to the nucleus is expressed, the cells differentiate, and the expression of beta-globin is the same as in wild-type cells.

Essential role for nuclear phospholipase C beta1 in insulin-like growth factor I-induced mitogenesis.

The nucleus has been shown to be a site for the inositol lipid cycle that can be affected by treatment of quiescent cells with growth factors such as insulin-like growth factor I (IGF-I). Indeed, the exposure of Swiss 3T3 cells to IGF-I results in a rapid and transient increase in nuclear phospholipase C (PLC) beta1 activity. In addition, several other reports have shown the involvement of PLC beta1 in nuclear signaling in different cell types. Although the demonstration of phosphatidylinositol 4-phosphate and phosphatidylinositol 4,5-bisphosphate hydrolysis by nuclear PLC beta1 established the existence of nuclear PLC signaling, the significance of this autonomous pathway in the nucleus has yet to be thoroughly clarified. By inducing both the inhibition of PLC beta1 expression by antisense RNA and its overexpression, we show that this nuclear PLC is essential for the onset of DNA synthesis following IGF-I stimulation of quiescent Swiss 3T3 cells.

Cytoplasmic and nuclear localization sites of phosphatidylinositol 3-kinase in human osteosarcoma sensitive and multidrug-resistant Saos-2 cells.

The intracellular localization of phosphatidyl-inositol 3-kinase (PI 3-kinase) has been analyzed by western blotting, confocal, and electron microscopy immunocytochemistry in human osteosarcoma Saos-2 cells. By western blotting, the enzyme appears to be present in both the cytoplasmic and nuclear subfractions. By confocal microscope immunocytochemistry, the cytoplasmic fluorescence is localized in the perinuclear region and on a network of filaments, while a diffused signal is present in the nucleus, except for the nucleolar areas. Ultrastructural analyses on whole cells and on in situ matrix preparations reveal that nuclear PI 3-kinase is localized in interchromatin domains, in stable association with inner nuclear matrix components, while the enzyme diffused in the cytosol is partly associated with the cytoskeletal filaments. Quantitative evaluations indicate that, in a multidrug-resistant variant obtained by continuous exposure of Saos-2 cells to doxorubicin, the amount of nuclear and cytoplasmic PI 3-kinase is significantly lower than in the sensitive parental cell line. The nuclear localization of PI 3-kinase and its variation in multidrug-resistant cells, characterized by a reduced mitotic index, are consistent with the data on the existence of a nuclear inositol lipid cycle, which could also utilize 3-phosphorylated inositides to modulate signal transduction for the control of some key functional activities.

Transfected Saos-2 cells overexpressing phosphoinositidase C beta 1 isoform accumulate it within the nucleus.

The subcellular partitioning of the phosphoinositidase C (PIC) isoforms involved in signal transduction, with the selective localization of the PIC beta 1 isoform in the nucleus, represents a crucial aspect of the complex mechanism of cell response to agonists. In order to further elucidate this phenomenon, we utilized human osteosarcoma Saos-2 cells, transfected with the cDNA for rat PIC beta 1. In the cells overexpressing this isoform, immunocytochemical analyses at the electron microscope level reveal an increased synthesis at the cytoplasm and a significant accumulation within the nucleus of the protein. Interestingly, the sites of intranuclear localization are, as in wild type cells, the interchromatin domains. These results indicate that the transfected cells maintain the capability of accumulating the enzyme within the nucleus and can be considered a model for functional studies on the nuclear signal transduction also in response to specific agonists.

The intranuclear amount of phospholipase C beta 1 decreases following cell differentiation in Friend cells, whereas gamma 1 isoform is not affected.

The existence of a signal transduction system in the nucleus, based on polyphosphoinositide breakdown mediated by specific phosphoinositidases (PLC), has been widely documented. In different cell systems, nuclear PLCs can be modulated, in response to agonists, either by enhancing or by down-regulating their activity, thus leading to DNA replication or to cell differentiation. Friend cells, induced to erythroid differentiation by dimethyl sulfoxide (DMSO), show a down-regulation of PLC beta 1 isoform, as indicated by the reduction of the transcription of its mRNA and of the in vitro synthesis of its translation product. The intracellular localization and the amount of different PLC isoforms have been evaluated by electron microscope immunocytochemistry. In untreated Friend cells, PLC beta 1 and gamma 1 isoforms are both present within the nucleus, whereas mainly the gamma 1 isoform is detected in the cytoplasm. The small amount of cytoplasmic PLC beta 1 is probably representative only of the newly synthesized enzyme. Quantitative immunolabeling analyses demonstrate that erythroid differentiation is associated with a significant decrease of the PLC beta 1 amount in the nucleus and with an almost complete disappearance of that isoform in the cytoplasm, whereas the PLC gamma 1 isoform is unaffected. The two PLC isoforms, moreover, appear to be differently associated with the nuclear components, PLC beta 1 being steadily bound to the inner nuclear matrix, whereas PLC gamma 1 is almost completely soluble.

Tiazofurin induces a down-modulation of ICAM-1 expression on K562 target cells impairing NK adhesion and killing.

Tiazofurin treatment of K562 leukemia cells in vitro depletes the metabolites of the guanylate biosynthetic pathway, inducing erythroid differentiation, that, in turn, alters the phenotypic profile. As a consequence, K562 cells possibly modify their interaction with immune cells. Here we describe the binding and killing activity of peripheral blood NK cells against differentiating K562 cells and the correlation between their altered binding capacity and ICAM-1 expression levels in differentiating K562 cells. We found that decreased percentages of NK (and T) cells were bound to differentiating K562 cells generating a decreased cytotoxic activity. This corresponded to decreased expression of ICAM-1, as detected by FACS analysis and Western blot. Erythroid differentiation, binding and killing reduction, and ICAM-1 down-modulation were completely abrogated by guanosine treatment. Tiazofurin causes a decrease in lymphocyte recognition and binding to K562 target cells. This can be ascribed to the down-modulation of ICAM-1 expression on target cells, which, therefore, can escape killing, acquiring a selective survival advantage.

Immunocytochemical evaluation of protein kinase C translocation to the inner nuclear matrix in 3T3 mouse fibroblasts after IGF-I treatment.

The complex pathway which links the agonist-cell membrane receptor binding to the response at the genome level involves, among other elements, protein kinase C (PKC). Agonists acting at the cell membrane can affect an autonomous nuclear polyphosphoinositide signaling system inducing an activation of nuclear phosphoinositidase activity and a subsequent translocation of PKC to the nuclear region. The fine localization of PKC has been investigated by means of electron microscopy quantitative immunogold labeling in 3T3 mouse fibroblasts, mitogenically stimulated by IGF-I. The enzyme, which in untreated cells is present in the cytoplasm, except for the organelles, and in the nucleoplasm, after IGF-I treatment is reduced in the cytoplasm and almost doubled in the nucleus. The PKC isoform translocated to the nucleus is the alpha isozyme, which is found not only associated with the nuclear envelope but mainly with the interchromatin domains. By using in situ matrix preparations, PKC appears to be retained at the nuclear matrix level, both at the nuclear lamina and at the inner nuclear matrix, suggesting a direct involvement in the phosphorylation of nuclear proteins which are responsible for the regulation of DNA replication.

Immunocytochemical detection of the specific association of different PIC isoforms with cytoskeletal and nuclear matrix compartments in PC12 cells.

The increasing evidence of discrete roles of phosphoinositidase C (PIC) isoforms and the assessment of their localization in the cytoskeleton and in the nucleus support the involvement of particular isotypes of this enzyme in signal transduction at multiple levels. PC12 rat pheochromocytoma is one of the few cell lines expressing three immunologically distinct isoforms of PIC. We have analyzed the subcellular distribution of the PIC beta 1, gamma 1 and delta 1 isoforms using confocal and electron microscope immunocytochemistry. PIC beta 1 is mainly found in the nucleus and is associated with interchromatin domains. On the other hand, the PIC gamma 1 isoform is found in the nucleus and in the cytosol, while PIC delta 1 is exclusively cytoplasmic. Immunoblot and immunocytochemical experiments indicate that the various PIC isoforms are differently bound to structural cell compartments, such as cytoskeletal and nuclear matrix elements. In fact, PIC beta 1 and PIC gamma 1 isoforms are tightly associated with the nuclear matrix, while only about 50% of PIC gamma 1 is associated with the cytoskeleton after DNase I and high salt extractions. PIC gamma 1 is almost completely soluble under these conditions. These results further confirm the complexity of the inositide signal transduction mechanism, which involves several PIC isoforms, specifically localized in different cell compartments and support the existence of a membrane-unrelated inositol lipid-dependent signalling in the nuclear interior.

Discrete localization of different DNA topoisomerases in HeLa and K562 cell nuclei and subnuclear fractions.

Monoclonal antibodies raised against DNA topoisomerase I and against topoisomerase II alpha and beta isoforms, which have been previously demonstrated to be highly specific and capable of detecting cell cycle-related variations of the topoisomerase II isoforms (Negri et al., 1992, Exp. Cell Res. 200, 452-459), have been utilized for a fine subcellular localization. Immunocytochemistry by confocal and electron microscopy have been used for a topological and quantitative evaluation of the fine distribution of the different topoisomerases in HeLa and K562 cells. Topoisomerase I and topoisomerase II alpha are present both in the nucleoplasm and in the nucleolus, though at different relative ratios, while topoisomerase II beta is exclusively present at the nucleolar level. This is further confirmed by immunoblotting and immunocytochemical quantitative evaluations performed on purified nuclear matrix fractions obtained from K562 cells. In fact, the amount of topoisomerase I and topoisomerase II alpha present in the whole cell nuclei is partly lost in isolated nuclei but, while topoisomerase I is further significantly reduced in nuclear matrix preparations, the topoisomerase II alpha content is only slightly decreased. On the other hand, the great majority of topoisomerase II beta is retained in the nuclear matrix and can be detected exclusively in association with the nucleolar remnant. These results are consistent with specific functional roles hypothesized for the different topoisomerase types.

Interleukin 1 alpha stimulates nuclear phospholipase C in human osteosarcoma SaOS-2 cells.

Interleukin 1 (IL-1) is one of the most potent stimulators of bone resorption. However, the early biochemical events elicited by IL-1 receptor binding are not fully understood. Here we show that in human osteosarcoma SaOS-2 cells the treatment with IL-1 alpha is able to evoke a rapid and transient increase of nuclear phospholipase C (PLC) activity. A parallel decrease of nuclear phosphatidylinositol monophosphate and phosphatidylinositol bisphosphate is observed. All these events are strictly confined to the nuclear compartment without affecting the cytoplasmatic inositol lipid pool. In addition we show that by Western blot analysis with specific monoclonal antibodies the PLC gamma is located both in the cytoplasm and in the nucleus, while PLC beta appears exclusively localized in the nucleus. Moreover, the increase of PLC activity in response to IL-1 alpha is completely neutralized by monoclonal antibody against the beta-form. While confirming the existence of an autonomous nuclear phosphoinositide signaling system, our data clearly indicate that in SaOS-2 cells one of the earliest events following IL-1 alpha treatment is the breakdown of nuclear phosphatidylinositol monophosphate and phosphatidylinositol bisphosphate because of the activation of a specific nuclear PLC isoform.

Phosphoinositidase C isozymes in SaOS-2 cells: immunocytochemical detection in nuclear and cytoplasmic compartments.

SaOS-2 cell line presents osteoblastic characteristics which can be modulated by specific agonists involving also phosphoinositide breakdown. In order to determine whether SaOS-2 cells display a phosphoinositide signalling system not only at the cytosol-cell membrane level but also, as recently reported for other cell lines, at the nuclear level, a study has been performed to evaluate the phosphoinositidase C (PIC) activity and to localize different isoforms of PIC in nuclear and cytoplasmic compartments. By immunochemicals methods, and by confocal and electron microscope immunocytochemistry, both PIC beta 1 and gamma 1 have been detected in the nucleus, while only PIC gamma 1 was found in the cytoplasm. A specific association with the inner nuclear matrix has been demonstrated for PIC beta 1 and gamma 1; this latter resulted, on the other hand, in relationship with cytoskeletal filaments after high salt extraction. These findings suggest that these enzymes are not completely soluble but functionally related with cytoskeletal and nucleoskeletal structures.