Expression of occludin in human rectal carcinoid tumours as a possible marker for glandular differentiation.

AIMS: To examine whether or not the tight junction-associated transmembrane protein occludin is expressed in rosette or gland-like structures in human rectal carcinoid tumours. The tight junction is crucial for the formation and maintenance of organized tubular structures in glandular epithelia. Previous studies have reported the presence of glandular structures in carcinoid tumours, though they are not believed to arise from glandular epithelium. METHODS AND RESULTS: The expression profiles of occludin in 40 carcinoid tumours were examined immunohistochemically, using an anti-occludin monoclonal antibody. In eight (20%) samples of typical carcinoid tumours, a small number of rosette-like tubular structures outlined by occludin were detected. CONCLUSIONS: Tight junction-associated molecules, including occludin, are thought to be one of the most characteristic structural markers of polarized glandular structures. The results of the present study provide supportive evidence that carcinoid tumour cells are capable of glandular differentiation.

Overexpression of granulocyte colony-stimulating factor induces severe osteopenia in developing mice that is partially prevented by a diet containing vitamin K2 (menatetrenone).

Mice transgenic for granulocyte colony-stimulating factor (G-CSF) exhibit severe osteopenia with an increase of osteoclast number and acceleration of bone resorption in adult mice. To examine the effect of G-CSF overexpression on developing bone, bone mineral density levels were examined from 4 weeks through 36 weeks after birth. Peak bone mass was observed at around 24 weeks of age irrespective of G-CSF expression. Apparent osteopenia was observed as early as 4 weeks of age without detectable developmental retardation in bone length and skeletal structure. Morphological examination confirmed a reduction of cancellous bone and cortical bone at this early stage of life, indicating that overexpression of G-CSF results in apparent osteopenia in developing mice, similar to that in adult animals. The effect of vitamin K2 (menatetrenone) (MK4) on bone phenotypes during development was then examined. Mice were fed chow containing either 0.05 mg MK-4 per 100 g or 20.0 mg MK-4 per 100 g for 12 weeks as the control and experimental diets, respectively. This treatment did not change bone length, irrespective of the type of mouse or diet. Peripheral quantitative computed tomography (pQCT) revealed an increase of in CT value bone of MK4-treated mice. Taken together, these results indicate that overexpression of G-CSF induces an apparent reduction of bone mass and results in osteopenia in developing mice. The bone reduction was partially restored by feeding the mice MK4, suggesting a choice for treatment on the osteopenia induced by G-CSF.

Neuropathy in diabetic mice overexpressing human aldose reductase and effects of aldose reductase inhibitor.

The present study was designed to examine the effect of aldose reductase (AR) overexpression on the development of diabetic neuropathy by using mice transgenic for human AR. At 8 weeks of age, transgenic mice (Tg) and non-transgenic littermates (Lm) were made diabetic with streptozotocin. After 8 weeks of untreated diabetes, plasma glucose levels and the reduction in body weight were similar between the groups of diabetic animals. Despite the comparable levels of hyperglycaemia, levels of sorbitol and fructose were significantly greater in the peripheral nerve of diabetic Tg than in diabetic Lm (both P < 0.01). Ouabain sensitive Na(+),K(+)-ATPase activity was similarly decreased in both diabetic Tg and Lm. Protein kinase C activity in the sciatic nerve membrane fraction was unaffected by diabetes in Lm, but was reduced by nearly 40% in the diabetic Tg. Although both groups of diabetic animals exhibited a significant decrease in tibial nerve motor nerve conduction velocity (MNCV), this decrease was significantly more severe (P < 0.01) in diabetic Tg than in diabetic Lm. Consistent with these findings, nerve fibre atrophy was significantly more severe in diabetic Tg than in diabetic Lm (P < 0.01). These findings implicate increased polyol pathway activity in the pathogenesis of diabetic neuropathy. In support of this hypothesis, treating diabetic Tg with an aldose reductase inhibitor (WAY121-509, 4 mg/kg/day) for 8 weeks significantly prevented the accumulation of sorbitol, the decrease in MNCV and the increased myelinated fibre atrophy in diabetic Tg.

Overexpression of the granulocyte colony-stimulating factor gene impairs bone morphogenetic protein responsiveness in mice.

Granulocyte colony-stimulating factor (G-CSF) is the major hematopoietic growth factor regulating the production and differentiation of neutrophils. We previously demonstrated that permanent overexpression of G-CSF in transgenic mice produces a dramatic enlargement of the bone cavity and reduction of bone mass. This phenotype was shown to be associated with an increase of osteoclast-mediated bone resorption. As a way of determining the role of G-CSF in bone formation in vivo, an ectopic bone was induced subcutaneously into G-CSF transgenic mice by bone morphogenetic protein (BMP)-2, a potent initiator of bone and cartilage from undifferentiated mesenchymal cells. A BMP-2/atelocollagen pellet containing recombinant human BMP-2 was implanted into a dorsal subfascial pocket. At one week after implantation, proliferation of mesenchymal cells around the implant was significantly decreased in transgenic mice compared with control mice. At three weeks, an ectopic bone containing bone marrow was formed both in transgenic and control mice. However, the ectopic bones of the transgenic mice were smaller and less consistent than those of control mice, and the calcium contents were reduced to 56.2% of those of controls. The ectopic bone in the G-CSF mice showed poor development of both lamellar and trabecular bone. Semiquantitative reverse transcription-polymerase chain reaction analysis of the ectopic bone at 3 weeks disclosed no significant differences in the mRNA levels of type I collagen, osteopontin, and osteocalcin between G-CSF mice and control mice. Immunohistochemical study in G-CSF mice showed reduced staining of osteocalcin in the bone matrix surrounding the reduced number of osteoblasts. The half-life of BMP in the implants was prolonged to 7 to 9 days in the G-CSF mice, whereas it was 5 days in the control mice. Collectively, the permanent expression of G-CSF may retard the differentiation process of osteoblasts by impairing the initial induction of mesenchymal cells, resulting in reduction of bone mass, suggesting that G-CSF regulates the bone metabolism by modulating both osteoclast and osteoblast function. Furthermore, it is suggested that G-CSF is a potent modulator of the BMP-2 signal pathway in vivo.

Occludin regulates actin cytoskeleton in endothelial cells.

Occludin is a major membrane component of tight junctions of endothelial cells, though the role of this molecule is not fully understood. RLE cells, derived from rat lung endothelial cells, express a negligible level of occludin with clear expression of E-cadherin and ZO-1 at cell junctions. Introduction of occludin by transfection induced clear junctional expression of occludin with few or no changes of expression of E-cadherin and ZO-1. The paracellular barrier function, as determined by transelectrical resistance and flux of non-ionic small molecules, was not detectably upregulated. When cells expressing occludin were cocultured with RLE cells null for occludin, clear junctional expression of occludin was observed irrespective of the expression of occludin on the apposing cells. Cortical actin was developed at the site of these occludin positive cell junctions. Treatment of cells with an actin depolymerizing agent, mycalolide B, abolished junctional expression of occludin together with E-cadherin and circumferential actin. ZO-1 showed relative resistance to this actin depolymerizing treatment and was maintained at the cell junctions, though fragmentation of immunoreactivity was detectable. Collectively, junctional expression of occludin was not associated with paracellular barrier function in this cell line. There was, however, a close correlation of occludin with the actin cytoskeleton, indicating a role of occludin as an important molecule in the regulation of the actin cytoskeleton in endothelial cells.

Loss of ErbB2 expression in pulmonary metastatic lesions in osteosarcoma.

OBJECTIVES: The c-erbB2 protooncogene is located on human chromosome 17 and encodes a 185-kilodalton transmembrane glycoprotein (ErbB2). Elevated ErbB2 expression or gene amplification has been shown to be associated with a poor prognosis in many cancers. Recently, it has been demonstrated that overexpression of the ErbB2 protein in osteosarcoma is associated with the presence of pulmonary metastasis and decreased survival. To further investigate the role of ErbB2 overexpression in pulmonary metastasis of osteosarcoma, its expression in the primary and metastatic lesions of the same osteosarcoma patients was compared. METHODS: We compared the expression levels of ErbB2 receptor protein between the biopsy samples and pulmonary metastatic lesions in each of 19 patients with osteosarcoma who had not presented with metastasis at diagnosis. All archival materials from patients were retrieved and stained with monoclonal antibody CB11 to detect ErbB2 protein. RESULTS: The rate of overexpression was significantly lower in the pulmonary metastatic tumors than in the biopsy samples (11 versus 42%; p = 0.03). Among 8 patients who had shown increased levels of ErbB2 in the biopsy samples, 7 exhibited complete absence of ErbB2 in the pulmonary metastatic lesions. Overall loss of ErbB2 expression was noted in 14 of 19 patients as the initial tumor became metastatic. CONCLUSIONS: It is suggested that overexpression of ErbB2 decreases within individual osteosarcomas as they become metastatic. Overexpression of ErbB2 may not play an important role in the development of pulmonary metastases of osteosarcoma. Further data are needed before ErbB2 can be used in making clinical decisions for osteosarcoma patients.

Occludin and claudin-1 concentrate in the midbody of immortalized mouse hepatocytes during cell division.

It has been believed that epithelial cells maintain tight junctions at all times, including during cell division, to provide a continuous epithelial seal. However, changes in localization of integral tight junction proteins during cell division have not been examined. In this study, using SV40-immortalized mouse hepatocytes transfected with human Cx32 cDNA, in which tight junction strands and the endogenous tight junction proteins occludin, claudin-1, ZO-1, and ZO-2 were induced, we examined changes in localization of the tight junction proteins at all stages of cell division. All tight junction proteins were present between mitotic cells and neighboring cells throughout cell division. In late telophase, the integral tight junction proteins occludin and claudin-1, but not the cytoplasmic proteins ZO-1 and ZO-2, were concentrated in the midbody between the daughter cells and were observed at cell borders between the daugher and neighboring cells. These results indicate that the integral tight junction proteins are regulated in a different manner from the cytoplasmic proteins ZO-1 and ZO-2 during cytokinesis.

Cx32 but not Cx26 is associated with tight junctions in primary cultures of rat hepatocytes.

On freeze-fracture replicas, gap junctions are frequently colocalized with tight junctions. In this study, to elucidate the relationship between gap- and tight-junction proteins, we investigated the localization of gap-junction proteins Cx32 and Cx26 and tight-junction proteins occludin, claudin-1, ZO-1, and ZO-2 in primary cultured rat hepatocytes, using confocal laser microscopy. In hepatocytes cultured in 2% DMSO and 10(-7) M glucagon medium, Cx32- but not Cx26-immunoreactive lines were observed on the most subapical plasma membrane at cell borders, while on the basolateral membrane both Cx32- and Cx26-positive spots were colocalized. Occludin-, claudin-1-, ZO-1-, and ZO-2-immunoreactive lines were also linearly observed on the most subapical plasma membrane and were colocalized with only Cx32-immunoreactive lines. In freeze-fracture analysis, many small gap-junction plaques were observed within a well-developed tight-junction strand network. The fence function of tight junctions in the cells, as examined by diffusion of labeled sphingomyelin, was well maintained. We also carried out Western blotting for Cx32 following immunoprecipitation with anti-occludin, anti-claudin-1, or anti-ZO-1 antibodies. Cx32 was detectable in all immunoprecipitates. These results suggest that Cx32 gap junctions, but not those with Cx26, are closely coordinated with the expression and function of tight junctions in hepatocytes and that Cx32 gap-junction formation may affect cell polarity through modification of tight-junction expression.

Uncoupling of gate and fence functions of MDCK cells by the actin-depolymerizing reagent mycalolide B.

The tight junction serves as a paracellular gate to seal the paracellular space of apposing cells and as a molecular fence to prevent diffusion of membrane proteins and lipids in epithelial cells. Although involvement of the actin cytoskeleton has been considered to be important in these two functions, it remains to be elucidated whether both functions are regulated in a coupled manner or differentially by actin. Treatment of highly polarized MDCK cells with mycalolide B (MB), a recently developed actin-depolymerizing reagent, induced a decrease of transepithelial resistance in a dose- and time-dependent manner with reversibility when the reagent was washed out. Changes in cytoskeletal actin, such as a reduction of cortical actin, irregularity of stress fibers, and punctated actin aggregates, were observed after MB treatment. However, the fence function, as studied by diffusion of apically labeled sphingomyelin/BSA complex, remained intact in the MB-treated MDCK cells. Localization of junctional molecules and apical marker proteins such as E-cadherin, ZO-1, and 114-kDa protein was shown to be unaffected. Furthermore, freeze-fracture study showed apparent tight junction strands. Collectively, MB treatment abolished the paracellular gate but not the fence function of MDCK cells, suggesting that cytoskeletal actin may play differential roles in the gate and fence functions of the tight junction.

Induction of tight junctions in human connexin 32 (hCx32)-transfected mouse hepatocytes: connexin 32 interacts with occludin.

Small gap junction plaques are associated with tight junction strands in some cell types including hepatocytes and it is thought that they may be closely related to tight junctions and the establishment of cell polarity. In order to examine roles of gap junctions in regulating expression and structure of tight junctions, we transfected human Cx32 cDNA into immortalized mouse hepatocytes (CHST8 cells) which lack endogenous Cx32 and Cx26. Immunocytochemistry revealed that endogenous integral tight junction protein occludin was strongly localized and was colocalized with Cx32 at cell borders in transfectants, whereas neither was detected in parental cells. In Northern blots, mRNAs encoding occludin and the other integral tight junction proteins, claudin-1 and -2, were induced in the transfectants compared to parental cells. In Western blots, occludin protein was increased in the transfectants compared to parental cells, and binding of occludin to Cx32 protein was demonstrated by immunoprecipitation. In freeze fracture of the transfectants, tight junction strands were more numerous and complex compared to parental cells, and small gap junction plaques appeared within induced tight junction strands. Nevertheless, no change in barrier function of tight junctions was observed. These results indicate that in hepatocytes, gap junction, and tight junction expression are closely coordinated, and that Cx32 may play a role in regulating occludin expression.

Glial cell line-derived neurotrophic factor induces barrier function of endothelial cells forming the blood-brain barrier.

Since a deep involvement of astrocytes, a kind of glial cells, in differentiation of the blood-brain barrier (BBB) has been suggested, we examined the relation of glial cell line-derived neurotrophic factor (GDNF) to the BBB. First, immunohistochemical examination of the cerebral cortex of rats revealed that glial cell line-derived neurotrophic factor receptor (GFRalpha1) was preferentially expressed on the cell membranes of capillary endothelial cells. Second, to elucidate the effects of GDNF on the BBB, capillary endothelial cells isolated from the porcine cerebral cortex were cultured and then changes in tight junction function of the endothelial cells were examined after addition of GDNF, in terms of transendothelial electrical resistance (TER) and permeability. GDNF at concentrations of 0.1 and 1 ng/ml significantly activated the barrier function of the endothelial cells in the presence of cAMP. Since GDNF is secreted from astrocytes sheathing capillary endothelial cells in the brain cortex, our results strongly suggest that GDNF enhances the barrier function of tight junctions of the BBB on the one hand, and also supports the survival of neurons on the other hand.

Disruption of circumferential actin filament causes disappearance of occludin from the cell borders of rat hepatocytes in primary culture without distinct changes of tight junction strands.

We investigated the relationship of actin filament organization to occludin and tight junction strands in primary cultured rat hepatocytes using an actin depolymerizing agent, mycalolide B. In control cultures, well-developed circumferential actin filaments and occludin immunoreactivity were observed on the most subapical plasma membrane of the cells, and tight junction strands formed well-developed networks in freeze-fracture replicas. In hepatocytes treated with 3 microM mycalolide B for 6 h, circumferential actin filaments and occludin immunoreactivity disappeared from the cell borders. However, there were no marked abnormalities of tight junction strands in freeze fracture replicas. Similar results were obtained from cells cultured in medium with 0.05 mM Ca2+ for 6 h. The close association of occludin with actin and the existence of intact tight junction strands that are virtually free of both occludin and actin suggest a physiological role of occludin, but not the other proteins forming the tight junction strands, in the linkage between actin cytoskeleton and tight junction.

Expansion of extrathymic T cells as well as granulocytes in the liver and other organs of granulocyte-colony stimulating factor transgenic mice: why they lost the ability of hybrid resistance.

When we attempted to characterize the immunological state in G-CSF transgenic mice, a large number of not only granulocytes but also lymphoid cells expanded in various immune organs. Such lymphoid cells were present at unusual sites of these organs, e.g., the parenchymal space in the liver. We then determined the phenotype of these lymphoid cells by immunofluorescence tests. It was demonstrated that CD3intIL-2Rbeta+ cells (i.e., extrathymic T cells), including the NK1.1+ subset of CD3int cells (i.e., NKT cells), increased in the liver and all other tested organs. These T cells as well as NK cells mediated NK and NK-like cytotoxicity, especially at youth. However, they were not able to mediate such cytotoxicity in the presence of granulocytes. This result might be associated with deficiency in the hybrid resistance previously ascribed to these mice. In other words, G-CSF transgenic mice had a large number of extrathymic T cells (including NKT cells) and NK cells that mediate hybrid resistance, but their function was suppressed by activated granulocytes. Indeed, these granulocytes showed an elevated level of Ca2+ influx upon stimulation. The present results suggest that, in parallel with overactivation of granulocytes, extrathymic T cells and NK cells are concomitantly activated in number but that their function is suppressed in G-CSF transgenic mice.

Occludin modulates organization of perijunctional circumferential actin in rat endothelial cells.

The tight junction is not a constitutional junctional apparatus in endothelial cells, but develops in a particular lineage of endothelia, such as the capillary endothelia in the brain and retina, and thus is considered to be pivotal for the maintenance of the blood-tissue barrier. Occludin is an integral membrane component of tight junctions, but the role of occludin in the endothelial cell function remains to be elucidated. We have cloned and transfected rat full-length occludin cDNA into a rat endothelial cell line (RLE) that expressed only a trace amount of occludin with no fine circumferential actin bundles at the cell border in native conditions. Occludin was expressed at the cell border of RLE cells, and circumferential fine actin bundles developed in close relation to the sites of occludin localization. Even under subconfluent culture conditions, fine circumferential actin bundles were formed at the sites where occludin-positive cell-cell contact was achieved. In immunoelectron microscopy, occludin was localized at distinct areas of the plasma membrane, always in association with the cytoplasmic actin filaments. On the other hand, actin bundles were not seen in occludin-negative juxtaposing plasma membranes. Collectively, these data strongly suggested a possible determinant function of occludin for the organization of actin in endothelial cells.

Bacterial lipopolysaccharide reduced intestinal barrier function and altered localization of 7H6 antigen in IEC-6 rat intestinal crypt cells.

The intestinal epithelial barrier restricts the passage of potentially toxic substances into the systemic circulation and is considered to be mostly mediated by tight junctions, though the mechanisms involved in the regulation of intestinal tight junctions are not yet fully understood. In the present study, we examined whether bacterial lipopolysaccharide (LPS) altered the barrier function of tight junction and localization of tight junctional proteins, ZO-1 and 7H6 antigen, in IEC-6 intestinal cells. Administration of LPS to the basolateral surface of IEC-6 cells disrupted the barrier function and caused the disappearance of 7H6 antigen from the cell border, whereas LPS administered to the apical surface altered neither the barrier function nor the localization of 7H6 antigen in IEC-6 cells. On the other hand, the localization of ZO-1 was not influenced by these treatments of LPS. These results suggest that the interaction of LPS with the basolateral surface of intestinal epithelial cells disrupts the barrier function and 7H6 antigen take part in the maintenance of the barrier function in IEC-6 cells.

Host F1 mice pretreated with granulocyte colony-stimulating factor accept parental bone marrow grafts in hybrid resistance system.

In the setting of hybrid resistance, parental C57BL/6 bone marrow (BM) grafts are vigorously rejected by lethally irradiated (C57BL/6xDBA/2) F1 mice. However, F1 mice pretreated by continuous administration of granulocyte colony-stimulating factor (G-CSF) with a miniosmotic pump before BM grafting developed day-8 splenic colonies of donor origin. This inhibitory effect on rejection was reversible because F1 mice regained the capacity to reject parental BM when the pump ceased functioning. The appearance of only a small number of colonies with the administration of G-CSF soon after BM grafting suggested the importance in producing this inhibitory effect of pre-exposure of host mice to G-CSF. Because G-CSF administration with a syngeneic combination did not influence the number of colonies, an altered distribution of grafted precursors was unlikely. The absence of a reduction in the number of NK1.1-positive cells in G-CSF-treated mice suggested functional impairment of natural killer cells, major effectors in hybrid resistance, but further study is necessary to elucidate the mechanism underlying this phenomenon. However, our results indicate the importance of G-CSF as a regulator in a certain type of immune response and raise the possibility of clinical application in transplantation medicine.

P21waf-1/cip-1/sdi-1 is expressed at G1 phase in primary culture of hepatocytes from old rats, presumably preventing the cells from entering the S phase of the cell cycle.

To elucidate whether p21waf-1/cip-1/sdi-1 expression is associated with loss of growth potential of hepatocytes of old rats, we determined p21waf-1/cip-1/sdi-1 expression of hepatocytes from old (30 months) rats during the cell cycle in primary culture. A high level of expression of p21waf-1/cip-1/sdi-1 was detected at the G1 phase in old-rat hepatocytes, but after the S phase in young-rat hepatocytes. Consistently, the incidence of the cells positive for p21waf-1/cip-1/sdi-1 in nuclei before entering the S phase was significantly higher in old-rat hepatocytes than in young-rat hepatocytes. These results account for the loss of growth potential of old-rat hepatocytes in vitro and the marked retardation of regeneration of liver in old rats in vivo.

Expression of p21(waf-1/cip-1) is significantly induced in the livers of LEC rats with chronic liver injury.

It is reported that hepatocytes isolated from LEC rats with chronic liver injury show reduced growth activity in primary culture. To elucidate the molecular basis of this phenomenon, we examined expression of p21(waf-1/ciP-1) and p27, cyclin-dependent kinase inhibitors, by northern blot analysis. The expression of p21(waf-1/cip-1 ) in the LEC rat liver was 3-fold higher than that of age-matched SD rat liver, while there was no significant difference in p27 expression level. Western blot analysis also revealed a significant increase in p21(waf-1/cip-1) in the nuclear matrix fraction of the LEC rat liver. Immunohistochemically, p21(waf-1/cip-1) was detected in the nuclei of normal LEC rat hepatocytes, but not in those of hepatocellular carcinoma cells, suggesting selective growth of neoplastic hepatocytes.

Selective inhibition of resistance to hemopoietic allografts but not rejection to a natural killer cell sensitive tumor in transgenic mice for granulocyte colony stimulating factor.

Transplanted allogeneic marrow grafts often fail to engraft in a lethally irradiated host. Resistance to hemopoietic allograft is a complexed phenomenon involving multiple components. To study the involvement of a hemopoietic cytokine, which was known to play a role for stem cell function, we established lines of mice that were transgenic for human granulocyte colony-stimulating factor (hG-CSF). Elevated and constitutive expression was found in sera (1,041 +/- 242 pg/ml) of these transgenic mice regardless of their sexes and ages. Strong neutrophilic granulocytosis correlated with the elevated G-CSF activity in transgenic mice but not in littermate controls, establishing a functional expression of this cytokine. In lethally irradiated mice transgenic for G-CSF, infusion of fully allogeneic marrow cells induced donor-derived spleen colony. Growth of hemopoietic allografts appeared to be similar to those of syngeneic marrow cells, which indicates inhibition of resistance for allogeneic marrow grafts. Because of a positive correlation, involvement of natural killer (NK) cells in resistance of transplanted allografts has been suggested. Inocula of NK-sensitive lymphoma cells were, however, vigorously rejected in the G-CSF-transgenic mice. This observation indicates that G-CSF may play a role in engraftment of transplanted allogeneic marrow grafts and may represent a component of mechanisms of hemopoietic resistance. Furthermore, this result may be an indication that alloresistance and NK cells use different mechanisms to resist each target.