Progesterone inhibits human endothelial cell proliferation through a p53-dependent pathway.

Previous studies have shown that progesterone inhibits endothelial cell proliferation through a nuclear receptor-mediated mechanism. Here, we further demonstrate that progesterone at physiologic levels (5 - 500 nM) dose- and time-dependently inhibited DNA synthesis of cultured human umbilical vein endothelial cells (HUVEC). The mRNA and protein levels of p21, p27, and p53 in HUVEC were increased by progesterone. The formation of CDK2-p21 and CDK2-p27 were increased and the CDK2 activity was decreased in the progesterone-treated HUVEC. The progesterone-inhibited [3H]thymidine incorporation was completely blocked when the expressions of p21 and p27 were knocked-down together. Transfection of HUVEC with dominant negative p53 cDNA prevented the progesterone-induced increases in p21 and p27 promoter activity and protein level, decreases in thymidine incorporation, and capillary-like tube formation. Matrigel angiogenesis assay in mice demonstrated the antiangiogenic effect of progesterone in vivo. These findings demonstrate for the first time that progesterone inhibited endothelial cell proliferation through a p53-dependent pathway.

Adenovirus-mediated heme oxygenase-1 gene transfer inhibits the development of atherosclerosis in apolipoprotein E-deficient mice.

BACKGROUND: Increasing evidence supports the role of heme oxygenase-1 (HO-1) in cytoprotective response and iron homeostasis. The object of this study was to investigate whether adenovirus-mediated gene transfer of HO-1 in arteries reduces iron overload and inhibits lesion formation in apolipoprotein E (apoE)-deficient mice. METHODS AND RESULTS: Infection of rat aortic smooth muscle cells with adenovirus carrying the human HO-1 gene (Adv-HO-1) resulted in a high-level expression of HO-1 protein, which effectively reduced the hemin-induced iron overload in these cells. Adenovirus-mediated gene transfer in arteries in vivo was achieved by direct injection of Adv-HO-1 into the left ventricles of anesthetized animals. Transgene was expressed in the endothelium and aortic lesion of apoE-deficient mice after they had received recombinant adenovirus for 1 week and gradually decayed during the next 5 weeks. When young apoE-deficient mice (14 weeks old) received Adv-HO-1 (2.5 x 10(9) pfu) for 6 weeks, lesions that developed in the aortic root or aortic arch were significantly smaller than those in control littermates receiving empty viral vector. Furthermore, the iron deposition as well as tissue iron content was much less in aortic tissue of Adv-HO-1-treated mice. The inhibitory effect of HO-1 gene transfer on the progression of advanced lesions was also observed in older apoE-deficient mice (20 weeks old) receiving Adv-HO-1 intraventricularly. CONCLUSIONS: Overexpression of HO-1 in vascular cells facilitates iron metabolism and attenuates development of atherosclerosis in apoE-deficient mice.

Inhibition of angiotensin II induced endothelin-1 gene expression by 17-beta-oestradiol in rat cardiac fibroblasts.

OBJECTIVE: To examine whether 17-beta-oestradiol (E(2)) may alter angiotensin II (Ang II) induced cell proliferation and to identify the putative underlying signalling pathways in rat cardiac fibroblasts. DESIGN: Cultured rat cardiac fibroblasts were preincubated with E(2) then stimulated with Ang II. [(3)H]Thymidine incorporation and endothelin-1 (ET-1) gene expression were examined. The effect of E(2) on Ang II induced NADPH oxidase activity, reactive oxygen species (ROS) formation, and extracellular signal regulated kinase (ERK) phosphorylation were tested to elucidate the intracellular mechanism of E(2) in proliferation and ET-1 gene expression. RESULTS: Ang II increased DNA synthesis, which was inhibited with E(2) (1-100 nmol/l). E(2), but not 17-alpha-oestradiol, inhibited Ang II induced ET-1 gene expression as shown by northern blotting and promoter activity assay. This effect was prevented by co-incubation with the oestrogen receptor antagonist ICI 182,780 (1 micromol/l). E(2) also inhibited Ang II increased NADPH oxidase activity, ROS formation, ERK phosphorylation, and activator protein-1 mediated reporter activity. CONCLUSIONS: The results suggest that E(2) inhibits Ang II induced cell proliferation and ET-1 gene expression, partially by interfering with the ERK pathway through attenuation of ROS generation. Thus, this study provides important new insight regarding the molecular pathways that may contribute to the proposed beneficial effects of oestrogen on the cardiovascular system.

Studies on the interaction between ferritin and ceruloplasmin.

We showed previously that ceruloplasmin associates with the H chain of rat liver ferritin during iron loading into ferritin such that the iron oxidized by ceruloplasmin was deposited into ferritin [S.-H. Juan et al. (1997) Arch. Biochem. Biophys. 341, 280-286]. Three synthetic decapeptides derived from domains 2, 4, and 6 of ceruloplasmin, referred to CP-2, CP-4, and CP-6, were utilized to identify a possible binding site on ceruloplasmin for ferritin. Two of the peptides, CP-4 and CP-6, were found to inhibit iron loading into the recombinant ferritin H chain homopolymer (rH-Ft) by ceruloplasmin. The extent of inhibition of iron loading into ferritin by ceruloplasmin by CP-6, but not CP-4, varied with pH, whereas the inhibitory effect remained constant in increasing concentrations of NaCl. The addition of rH-Ft quenched the fluorescence emission of CP-4 and CP-6, but not CP-2. The quenching of fluorescence was used to estimate dissociation constants for the peptides. Iron loading into ferritin in Hepes buffer was not affected in the presence of these peptides. In addition, synthetic peptides corresponding to the BC loop of ferritin H and L chains were utilized to localize an interaction site on ferritin for ceruloplasmin. The BC loop of H chain but not L chain of ferritin stimulated the ferroxidase activity of ceruloplasmin. Only the BC loop of ferritin H chain decreased the amount of iron loading into ferritin by ceruloplasmin.

Iron and phosphate content of rat ferritin heteropolymers.

An attempt was made to relate the iron and phosphate content of ferritin to its subunit composition. Ferritins from various tissues were separated according to their subunit composition by anion exchange chromatography and according to their iron content by density-gradient centrifugation. Iron and phosphate contents were not related to subunit composition. Recombinant rat liver ferritin heteropolymers of different subunit composition (1, 4, 6, 10, 15, and 17 H chains per 24 mer) were maximally loaded with iron, using ceruloplasmin and phosphate. All loaded approximately the same amount of iron and phosphate (2250 and 380 atoms, respectively). The iron and phosphate content of all ferritin, including the maximally loaded recombinant ferritin heteropolymers, fit an equation we previously reported: [Fe] = 4404 - 5.61 [Pi] (D. deSilva et al., 1993, Arch. Biochem. Biophys. 303, 451-455). These results suggest that the amount of iron and apparently the space within the core of ferritin were not related to different subunit composition.

Mutational analysis of loading of iron into rat liver ferritin by ceruloplasmin.

Site-directed mutagenesis was used to investigate the loading of iron into rat liver ferritin by ceruloplasmin. Changes were made in the H chain to investigate the role of tyrosines involved in an inherent ferroxidase activity thought to be involved in the self-loading of iron into ferritin. Mutation Y34F affected the rate of iron loading by ceruloplasmin and incorporation of the oxidized iron into the core. Mutation Y29R (making it analogous to the L chain) had no effect on iron oxidation but slightly decreased core formation. A double mutation in the L chain, to open the alpha-helix bundle channel, and R25Y, making the protein more analogous to the H chain, increased the amount of iron incorporated into the core, again suggesting that this Tyr is involved in ligand exchange for core formation. Additional changes in the L chain involving the BC loop suggest that the entire BC loop is involved in the association of ferritin with ceruloplasmin, increasing its ferroxidase activity and the rate of iron loading into ferritin.

The effect of putative nucleation sites on the loading and stability of iron in ferritin.

The L chain of the iron storage protein ferritin contains more putative nucleation sites in the core (Glu53, 56, 57, 60, and 63) than does the H chain (Glu61, 64, and 67). Recombinant DNA techniques were used to investigate the role of these putative nucleation sites on iron loading by ceruloplasmin and on the stability of the iron core. Recombinant rat liver ferritin H chain homopolymer and the two mutants (E61A and E61A-E64A), containing three, two and one nucleation sites, respectively, loaded up to 2010 +/- 50, 2010 +/- 40, and 1950 +/- 40 atoms of iron per ferritin, respectively. However, the mutations resulted in a 50% decrease in the rate of iron loading by ceruloplasmin. The ferritin variants incorporated the same amount of phosphate after iron loading (410 +/- 20, 400 +/- 30, and 420 +/- 20 atoms per ferritin, respectively). The stability of the iron cores prior to phosphate incorporation, assessed by the rate of iron release by 10 mM EDTA and the paraquat cation radical, corresponded to numbers of proposed nucleation sites. The subsequent incorporation of phosphate seemed to stabilize the iron core and minimized the effect of numbers of putative nucleation sites in ferritin on the rate of iron release by EDTA and the paraquat cation radical. After incorporation of phosphate the ferritins behaved similarly to the native rat liver ferritin with respect to the rate of iron release by the paraquat cation radical.

Loading of iron into recombinant rat liver ferritin heteropolymers by ceruloplasmin.

We have reported previously that the heavy chain of ferritin is required for iron incorporation by ceruloplasmin (J.-H. Guo, M. Abedi, and S. D. Aust (1996) Arch. Biochem. Biophys. 335(1)). The purpose of this study was to determine how many heavy chains were required for ceruloplasmin to interact with ferritin such that iron loading occurred. The cDNA sequences encoding the heavy and light chains of rat liver ferritin were cloned into the baculovirus transfer vector pA-cUW51 under the control of polyhedrin and p10 promoters, respectively, which was then incorporated by homologous recombination into the infections Autographa californica nuclear polyhedrosis virus genome. Both ferritin chains were expressed and assembled into two heteropolymers following the infection of insect cells by recombinant virus, which were separated by DEAE-Sepharose chromatography. The percentage of heavy (H) and light (L) chains making up the two heteropolymers, determined by gel scanning following the resolution of chains on SDS-PAGE, were equivalent to 1 H and 23 L chains and 2 H and 22 L chains. The maximal extent of iron loading was observed using 1 mol of rat ceruloplasmin per mole of H chain in the two heteropolymers. The extent of iron incorporation decreased with additional ceruloplasmin. Iron incorporation into rat liver ferritin, found to contain 10 H chains, increased as the molar ratio of ceruloplasmin to ferritin increased to 4:1 and remained the same up to 8:1. Iron loading into horse spleen ferritin, found to have one H chain, appeared similar to that for recombinant ferritin, having only one H chain. Therefore, we propose that the optimal molar ratio of ceruloplasmin to ferritin depends upon the numbers of H chain making up the ferritin molecule for the maximal incorporation of iron into ferritin. These results also suggest that the iron loading channel is contained within a single H chain subunit.

Mutational analysis of the four alpha-helix bundle iron-loading channel of rat liver ferritin.

We previously reported that the heavy chain of ferritin was required for loading it with iron using ceruloplasmin as a ferroxidase [J.-H. Guo, M. Abedi, and S. D. Aust (1996) Arch. Biochem. Biophys. 335, 197-204]. Site-directed mutagenesis, K58E and G61H, on recombinant rat liver L chain ferritin (rL-Ft) was performed to construct a proposed iron-loading channel in the alpha-helix bundle similar to rat liver H chain ferritin (rH-Ft). Conversely, the channel in rH-Ft was closed by mutations E62K and H65G to form a K62 to E107 salt bridge, which is believed to exist in the L chain. Both variants were expressed in insect cells and were soluble and able to form multi-subunit homopolymers. The rH-Ft mutant homopolymer could not be loaded, whereas the rL-Ft mutant homopolymer could be loaded with iron by ceruloplasmin. However, we found that the initial rate of iron loading into the rL-Ft mutant homopolymer by ceruloplasmin was less than that into the rH-Ft homopolymer. When 500 atoms of iron per ferritin were used for loading, 98% was loaded into the rH-Ft homopolymer by ceruloplasmin in 15 min, but only 30% was loaded into the rL-Ft mutant homopolymer in the same time. Moreover, the ferroxidase activity of ceruloplasmin was enhanced in the presence of the rH-Ft and the rH-Ft mutant homopolymers, but not in the presence of the rL-Ft or the rL-Ft mutant homopolymers. These observations suggested that the four alpha-helix bundle channel of ferritin is required for iron loading, but an additional factor, i.e. , a site which stimulate the ferroxidase activity of ceruloplasmin, is also essential.

Suppression of cell growth by heavy chain ferritin.

While producing recombinant rat liver H and L chain ferritin homopolymers using the baculovirus expression system, we noticed that rat liver H chain ferritin, but not L chain ferritin, had a suppressive effect on the growth of Spodoptera frugiperda (Sf-21) cells. Suppression was observed immediately after infection with recombinant H chain ferritin baculovirus prepared from lysed infected cells. Immediate suppression was observed when purified with either recombinant H chain apoferritin or various holoferritins (loaded with 1,970 +/- 50 or 2,520 +/- 90 atoms of iron/ferritin) indicating that suppression was not due to sequestration of iron required for cell growth. Suppression by H chain ferritin was also observed upon attempting to express the protein in Escherichia coli. Strategies for expression of recombinant rat liver H and L ferritin homopolymers in both prokaryotic and eukaryotic expression systems were developed.