Osteoblast-like cell adhesion to bone sialoprotein peptides.

A number of studies have suggested that biomimetic peptides can be used in the design of a new generation of prosthetic implants to promote the successful biointegration of the implant materials. In the current study, the in vitro bioactivities of several peptides representing RGD (Arg-Gly-Asp)-containing sequences of bone sialoprotein (BSP) toward an osteoblast-like cell line (MC3T3-E1) were examined to provide insight into the molecular basis of BSP's interaction with bone cells. BSP residues 283-288, 281-290, 278-293 and 278-302 were coated on polystyrene surfaces in 96-well non-tissue (untreated) culture plates, and their osteoblast adhesive properties compared to intact BSP and fibronectin as positive controls. BSP peptides 278-302 and 278-293 were found to be the most potent in their adhesive activity, increasing the number of adherent cells to 350% of control levels at an added concentration of 1 microM. Since these two peptides were equivalent in potency, it is suggested that the region 294-302 beyond the RGD domain is not necessary for cell binding. In comparison, peptides 283-288 and 281-290 were only active at concentrations greater than 200 microM. 50-70% of the peptide-stimulated adhesion was inhibited by the pretreatment of cell suspensions with solution phase RGD, suggesting that a portion of the peptides' adhesive effects was specific and integrin-mediated, although other non-RGD flanking regions were probably also involved in the mechanism of adhesion. Importantly, a modified BSP peptide, in which an aspartic acid residue at position 288 of the RGD sequence was replaced by a glutamic acid residue to form RGE, was completely inactive as a cell adhesion stimulus at concentrations up to 200 microM. Thus, despite the potential role of non-RGD flanking regions, an intact RGD tripeptide was essential for all of the adhesive activity of the BSP peptides.

Thermal and chemical modification of titanium-aluminum-vanadium implant materials: effects on surface properties, glycoprotein adsorption, and MG63 cell attachment.

The microstructure, chemical composition and wettability of thermally and chemically modified Ti-6Al-4V alloy disks were characterized and correlated with the degree of radiolabeled fibronectin-alloy surface adsorption and subsequent adhesion of osteoblast-like cells. Heating either in pure oxygen or atmosphere (atm) resulted in an enrichment of Al and V within the surface oxide. Heating (oxygen/atm) and peroxide treatment both followed by butanol treatment resulted in a reduction in content of V, but not in Al. Heating (oxygen/atm) or peroxide treatment resulted in a thicker oxide layer and a more hydrophilic surface when compared with passivated controls. Post-treatment with butanol, however, resulted in less hydrophilic surfaces than heating or peroxide treatment alone. The greatest increases in the adsorption of radiolabeled fibronectin following treatment were observed with peroxide/butanol-treated samples followed by peroxide/butanol and heat/butanol, although binding was only increased by 20-40% compared to untreated controls. These experiments with radiolabeled fibronectin indicate that enhanced adsorption of the glycoprotein was more highly correlated with changes in chemical composition, reflected in a reduction in V content and decrease in the V/Al ratio, than with changes in wettability. Despite promoting only a modest elevation in fibronectin adsorption, the treatment of disks with heat or heat/butanol induced a several-fold increase in the attachment of MG63 cells promoted by a nonadhesive concentration of fibronectin that was used to coat the pretreated disks compared to uncoated disks. Therefore, results obtained with these modifications of surface properties indicate that an increase in the absolute content of Al and/or V (heat), and/or in the Al/V ratio (with little change in hydrophilicity; heat+butanol) is correlated with an increase in the fibronectin-promoted adhesion of an osteoblast-like cell line. It would also appear that the thermal treatment-induced enhancement of cell adhesion in the presence of this integrin-binding protein is due to its increased biological activity, rather than a mass effect alone, that appear to be associated with changes in chemical composition of the metallic surface. Future studies will investigate the influence of the surface chemical composition of various implantable alloys on protein adsorption and receptor-mediated cell adhesion. In addition, by altering the properties of bound osteogenic protein enhancing exposure to cell integrin binding domains, it may be possible to develop implant surfaces which enhance the attachment, adhesion and developmental response of osteoblast precursors leading to accelerated osseointegration.

Activators of protein kinase A decrease the levels of free arachidonic acid in osteoblasts via stimulation of phosphatidylcholine and phosphatidylethanolamine synthesis.

In order to examine the role of protein kinase A (PKA) in the regulation of arachidonic acid availability, the interaction between cAMP agonists and the G protein activator AIF4- in their effects on phospholipid metabolism were measured in MC3T3-E1 osteoblasts. We show that forskolin and 8-brcAMP, activators of PKA, amplify the AIF4(-)-induced stimulation of phosphatidylinositol-specific phospholipase C (phosphatidylinositol inositolphosphohydrolase; EC 3.1.4.3), measured by the formation of [3H]inositol phosphates in prelabeled cells. However, the AIF4(-)-stimulated production of 1,2-diacylglycerols and the release of [3H]arachidonic acid ([3H]AA) were inhibited 50-75% by forskolin and 8-bromocAMP. Furthermore, pretreatment with PKA activators prevented much of the AIF4(-)-induced loss of [3H]AA from phosphatidylcholine and phosphatidylethanolamine in prelabeled osteoblasts. In addition, in the absence of AIF4-, forskolin was found to stimulate the incorporation of [3H]AA and [32P]orthophosphoric acid selectively into these two major phospholipids and selectively increased their mass. The effects of forskolin and 8-BrcAMP on the levels of free [3H]AA were completely reversed by pretreatment with the PKA inhibitor H-89. Therefore, our findings suggest that the activation of cAMP-dependent protein kinase can reduce the availability of free arachidonic acid for prostaglandin synthesis in osteoblast cells by stimulating its reesterification via phospholipid resynthesis.

Protein kinases A and C positively regulate G protein-dependent activation of phosphatidylinositol-specific phospholipase C by tumor necrosis factor-alpha in MC3T3-E1 osteoblasts.

The role(s) of protein kinases in the regulation of G protein-dependent activation of phosphatidylinositol-specific phospholipase C by tumor necrosis factor-alpha was investigated in the osteoblast cell line MC3T3-E1. We have previously reported the stimulatory effects of tumor necrosis factor-alpha and A1F4-, an activator of G proteins, on this phospholipase pathway documented by a decrease in mass of PI and release of diacylglycerol. In this study, we further explored the mechanism(s) by which the tumor necrosis factor or A1F4(-)-promoted breakdown of phosphatidylinositol and the polyphosphoinositides by phospholipase C is regulated. Tumor necrosis factor-alpha was found to elicit a 4-5-fold increase in the formation of [3H]inositol-1,4-phosphate and [3H]inositol-1,4,5-phosphate; and a 36% increase in [3H]inositol-1-phosphate within 5 min in prelabeled cells. [3H]inositol-4-phosphate, a metabolite of [3H]inositol-1,4-phosphate and [3H]inositol-1,4,5-phosphate, was found to be the predominant phosphoinositol product of tumor necrosis factor-alpha and A1F4(-)-activated phospholipase C hydrolysis after 30 min. In addition, the preincubation of cells with pertussis toxin decreased the tumor necrosis factor-induced release of inositol phosphates by 53%. Inhibitors of protein kinase C, including Et-18-OMe and H-7, dramatically decreased the formation of [3H]inositol phosphates stimulated by either tumor necrosis factor-alpha or A1F4- by 90-100% but did not affect basal formation. The activation of cAMP-dependent protein kinase, or protein kinase A, by the treatment of cells with forskolin or 8-BrcAMP augmented basal, tumor necrosis factor-alpha and A1F4(-)-induced [3H]inositol phosphate formation. Therefore, we report that protein kinases can regulate tumor necrosis factor-alpha-initiated signalling at the cell surface in osteoblasts through effects on the coupling between receptor, G-protein and phosphatidylinositol-specific phospholipase C.

Role of protein kinase A in collagenase-1 gene regulation by prostaglandin E1: studies in a rabbit synoviocyte cell line, HIG-82.

Gene expression of the matrix-degrading enzyme collagenase-1 in rabbit synoviocytes and human fibroblasts is down-regulated by prostaglandin E1 (PGE1) through a cyclic adenosine monophosphate (cAMP)-dependent pathway. In the current study, we examined the role of protein kinase A (PKA) in the PGE1-mediated effect on collagenase-1 gene expression. Collagenase-1 gene expression was rapidly induced several-fold above control both by a phorbol ester, 12-o-tetradecanoyl phorbol 13 acetate, and interleukin-1 beta (IL-1 beta) in HIG-82 synoviocytes. Treatment with PGE1 and forskolin increased PKA activity in the HIG-82 cells within 15 minutes of adding the stimulating agents. Two inhibitors of PKA, the isoquinoline-sulfonamide derivative, H-89 and a cAMP analog, RpcAMP, blocked the ability of PGE1 to down-regulate collagenase-1 gene expression. However, if PGE1 was added from 6 h to 30 minutes before the PKA inhibitor H-89, collagenase-1 gene expression was inhibited. Constitutive PKA activity was increased in HIG-82 synoviocytes stably transfected with an expression vector pCMV.C alpha that caused the HIG-82 cells to overexpress an active catalytic subunit of PKA. Cells stably transfected with an inactive, mutated C-alpha-variant showed no change in PKA activity. Collagenase-1 mRNA levels in TPA-stimulated cells were reduced to baseline levels in the pCMV.C alpha but not in the mutated C-alpha-transfected cells. These data show the importance of PKA in regulating collagenase-1 gene expression in a synoviocyte cell line.

Protein kinase C-independent activation of a novel nonspecific phospholipase C pathway by phorbol myristate acetate releases arachidonic acid for prostaglandin synthesis in MC3T3-E1 osteoblasts.

The effects of phorbol myristate acetate, an activator of protein kinase C, on the release of [3H]arachidonic acid and prostaglandin synthesis were studied in an osteoblast cell line (MC3T3-E1). Phorbol myristate acetate (20 uM) liberated 16 and 55% of the [3H]arachidonate in prelabeled phosphatidylinositol and phosphatidylethanolamine, respectively; and evoked a 19-fold stimulation in the synthesis of prostaglandin E2. Phorbol myristate acetate doubled the cellular mass of 1,2-diacylglycerol and stimulated the liberation of [3H]arachidonate from the diacylglycerol pool in prelabeled cells. The diacylglycerol lipase inhibitor RHC 80267 blocked 75-80% of the phorbol ester-promoted (total) cellular liberation of [3H]arachidonic acid and production of prostaglandin E2. In comparison, the release of [3H]arachidonate from phosphatidylethanolamine (but not phosphatidylinositol) was only partially antagonized (to the same degree) by the PLA2 inhibitor p-bromophenacylbromide and the protein kinase C inhibitor Et-18-OMe, PMA-induced formation of diacylglycerol or synthesis of PGE2 was not affected by the prior inhibition of protein kinase C. Therefore, we have shown a novel pathway for the liberation of arachidonic acid in osteoblasts involving the nonspecific hydrolysis of phosphatidylinositol and phosphatidylethanolamine by phospholipase C followed by the deesterification of diacylglycerol. This pathway can be activated by a phorbol ester through a protein kinase C-independent mechanism.

Prostaglandin E1 inhibits collagenase gene expression in rabbit synoviocytes and human fibroblasts.

Cartilage breakdown, as seen in inflammatory and degenerative joint diseases, can be mediated by proteolytic enzymes, such as the metalloproteinase collagenase, the only enzyme able to digest collagen at neutral pH. In vitro collagenase gene expression can be stimulated by the phorbol ester tumor promoter 12-O-tetradecanoyl-phorbol-13-acetate. We have investigated the effect of prostaglandin E1 (PGE1) on 12-O-tetradecanoyl-phorbol-13-acetate-stimulated collagenase mRNA levels in the rabbit synoviocyte cell line HIG-82. PGE1, but not PGE2 or PGF2 alpha, was able to selectively reduce collagenase mRNA levels in a dose-dependent fashion. PGE1 markedly increased intracellular levels of cAMP, while PGE2 and PGF2 alpha had little or no effect on cAMP production in the HIG-82 synoviocytes. Agents known to increase intracellular cAMP levels, such as the adenyl cyclase activator forskolin and the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (IBMX), mimicked the effect of PGE1, on collagenase mRNA levels. PGE1, forskolin, and IBMX also decreased collagenase mRNA levels in human skin fibroblasts, demonstrating that this observation was not unique to the HIG-82 cell line. Transient transfection experiments carried out in HIG-82 cells using a 1.2-kilobase portion of the 5'-flanking region of the human collagenase gene linked to the reporter gene luciferase demonstrated that PGE1, forskolin, and IBMX exert their inhibitory effect on the promoter region of the collagenase gene.

Tumor necrosis factor-alpha stimulates phosphatidylinositol breakdown by phospholipase C to coordinately increase the levels of diacylglycerol, free arachidonic acid and prostaglandins in an osteoblast (MC3T3-E1) cell line.

The effects of (human recombinant) tumor necrosis factor-alpha on phosphatidylinositol breakdown, release of 1,2-diacylglycerols, mobilization of arachidonate from diacylglycerol and prostaglandin synthesis were examined in a model osteoblast cell line (MC3T3-E1). Tumor necrosis factor-alpha (10 nM) caused a specific (30%) decrease in the mass of phosphatidylinositol (and no other phospholipids) within 30 min of exposure. Tumor necrosis factor-alpha doubled the rate of incorporation of [32P]orthophosphoric acid into phosphatidylinositol, indicating that the turnover of inositol phosphate was enhanced, and increased the content of diacylglycerol in parallel with phosphatidylinositol breakdown. The cytokine (10-50 nM; 4 h) also promoted a specific release of 24-34% of the [3H]arachidonate from prelabeled phosphatidylinositol, a release of 80% of the 3H-fatty acid from the diacylglycerol pool, and a 30-fold increase in the synthesis of prostaglandin E2. The tumor necrosis factor-alpha induced liberation of [3H]arachidonate from diacylglycerol, cellular arachidonate release and the synthesis of prostaglandin E2 were each blocked by an inhibitor of diacylglycerol lipase, the compound RHC 80267 (30 microM). Therefore, we conclude that, in the MC3T3-E1 cell line, tumor necrosis factor-alpha activates a phosphatidylinositol-specific phospholipase C (phosphatidylinositol inositolphosphohydrolase; EC 3.1.4.3) to release diacylglycerol, and increases the metabolism of diacylglycerol to liberate arachidonate for prostaglandin synthesis.

A23187 and protein kinase C activators stimulate phosphatidylinositol metabolism and prostaglandin synthesis in a human lung cancer cell line.

Activation of cell phospholipase, release of arachidonic acid and stimulation of prostaglandin synthesis were studied in a newly described human tumor cell line (Lu-65). In the Lu-65 tumor cell line, the calcium ionophore A23187 (2 microM) caused a 100% increase in the release of 3H-arachidonic acid and a 7-fold increase in the synthesis of prostaglandin E2. 1-oleoyl, -2-acetyl-glycerol (100 microM) increased arachidonate release and prostaglandin E2 synthesis by 100%. A23187 and the protein kinase C activators, 1,2-dioctanoyl-glycerol and 1-oleoyl, -2-acetyl-glycerol, decreased the specific radioactivity of 3H-arachidonate in phosphatidylinositol by 37% and 57%, respectively. The effects of A23187 were blocked in Ca2+-free media or in the presence of the phospholipase A2 inhibitor, p-bromophenacyl bromide, while those of 1-oleoyl, -2-acetyl-glycerol were not. The data provide evidence in a human tumor cell line for calcium/phospholipase A2-dependent and independent pathways for arachidonic acid release, both of which preferentially hydrolyze phosphatidylinositol.

Effects of hypophysectomy and administration of growth and thyroid hormones on the hydroperoxide-induced calcium release process and glutathione levels in rat liver mitochondria.

In order to investigate the effects of growth and thyroid hormones on Ca2+ transport in liver mitochondria, Ca2+ release and other accompanying changes induced by t-butylhydroperoxide were measured in mitochondria from hypophysectomized and hormone-injected rats. Mitochondria from normal and hypophysectomized rats showed similar rates of Ca2+ uptake (40 nmol.min.-1 mg protein-1) and ruthenium red-insensitive release (3 nmol.min.-1 mg protein-1). However, the t-butylhydroperoxide (0.5 mM)-induced release of 90% of the added Ca2+ required 1027 +/- 98 and 560 +/- 35 s in the hypophysectomized and normal groups, respectively, and the difference was independent of Ca2+. The release was accompanied by a loss of membrane potential, large amplitude swelling, the oxidation of NAD(P)H and stimulation of respiration. At conditions of equivalent release rates, the rate and extent of swelling as well as the stimulation of respiration were lower in mitochondria from hypophysectomized rats than those in the normal group. These results were confirmed by electron microscopy and provided evidence for a dissociation between the release of Ca2+ and the increase of membrane permeability. The administration of bovine growth hormone and/or 3,5,3'-triiodo-L-thyronine to hypophysectomized rats decreased the Ca2+ release times by different degrees and thyroid hormone was more effective than growth hormone. Hypophysectomy doubled the GSH content and hormone injections decreased it and the Ca2+ release times in parallel. Indeed, a high degree of correlation (r = 0.96) was obtained between mitochondrial GSH and the release times from groups of differing hormone status. Differences in the groups in the time required for oxidation of 80-90% of GSH were correlated with the differences in the times for release of 90% Ca2+. Therefore, these results demonstrated that growth and thyroid hormones can alter both the hydroperoxide-induced Ca2+ release and the metabolism of GSH in liver mitochondria, suggesting that the two processes are related. It is proposed that the effects of these hormones on Ca2+ transport may result from the promotion of its efflux from mitochondria and its mobilization into the cytosol.

The relationship between high-affinity noncatalytic binding of snake venom phospholipases A2 to brain synaptic plasma membranes and their central lethal potencies.

The basic phospholipase A2 from Naja nigricollis (African spitting cobra) snake venom is enzymatically less active but more toxic than the acidic phospholipase A2 from Naja naja atra (Taiwan cobra) snake venom, following injection into the right lateral ventricle of the brain of rats. When radiolabeled with 125I, these phospholipases A2 retained enzymatic activities and lethal potencies. Both enzymes bound with high affinity and specificity to brain synaptic plasma membrane preparations in vitro even in the absence of calcium, suggesting a non-catalytic binding. The acidic enzyme, in a calcium-free medium, had two binding components with Kd values of 1 X 10(-10) and 2.75 X 10(-8) M and Bmax values of 6 X 10(-13) and 3.4 X 10(-11) mol/mg, respectively. Multiple specific and nonspecific binding components were observed for each phospholipase A2; saturability for all of the binding sites was conclusively demonstrated only for the N. naja atra phospholipase A2 in a calcium-free medium (Bmax = 3.4 X 10(-11) mol/mg). The levels of specific and total binding were 150 pmol/mg and 450 pmol/mg, respectively, for the comparatively toxic enzyme and 15 pmol/mg and 35 pmol/mg, respectively, for the comparatively nontoxic enzyme at a concentration of 2.5 X 10(-8) M. These levels of binding (both total and specific) were directly correlated with the intraventricular lethal potencies of the phospholipases A2 (0.5 and 5.0 micrograms/rat for the N. nigricollis and N. naja atra phospholipases A2, respectively), suggesting a possible relationship between binding and lethal potency. Carbamylation of lysines reduced the levels of binding and the lethal potencies of both enzymes to a greater extent than their enzymatic activities. Pretreatment with high temperature, proteinases, phospholipases A2 or C suggested that radiolabeled phospholipase A2 binds to phospholipids rather than proteins. However, only the N. naja atra phospholipase A2 manifested a strict dependence on a divalent cation (Ca2+ or Sr2+) for most of its binding. The N. nigricollis enzyme demonstrated a much lower rate of dissociation from synaptic plasma membranes than did N. naja atra phospholipase A2, suggesting that hydrophobic interactions are more important in the binding of the more toxic enzyme as compared to the less toxic enzyme. It is proposed that differences in the extent of high-affinity noncatalytic binding to membrane phospholipids may be at least partly responsible for the marked difference in central toxicities of these two phospholipases A2.

Dissociation of pharmacological and enzymatic activities of snake venom phospholipases A2 by modification of carboxylate groups.

The carboxylate groups in an acidic and in a basic phospholipase A2 (PLA2) enzyme, purified, respectively, from Naja naja atra and Naja nigricollis snake venoms, were modified with carbodiimide and semicarbazide. The derivatives modified at pH 3.5 and pH 5.5 had less than 1% (N. nigricollis) or 2% (N. n. atra) residual enzymatic activity, whereas 12-16% enzymatic activity remained following modification at pH 5.5 in the presence of Ca2+. In marked contrast, these derivatives retained variable, but significantly greater, levels of lethal potency, hemolytic and anticoagulant activities, and abilities to block indirectly and directly induced contractions of the diaphragm. By this modification of aspartic and glutamic acid residues we have, for the first time, obtained derivatives of PLA2 which selectively retain greater pharmacological activity relative to enzymatic activity. Previously, we had found that modification of lysine and arginine residues produced derivatives which retain enzymatic activity but show a loss of pharmacological properties. These findings suggest that some pharmacological effects of snake venom PLA2 enzymes are due to a non-enzymatic action, suggesting two distinct but perhaps overlapping active sites.

Effect of methylation of histidine-48 on some enzymatic and pharmacological activities of snake venom phospholipases A2.

The effects on some pharmacological and enzymatic properties were determined following methylation of histidine at the enzymatic active site of the basic relatively toxic Naja nigricollis and the acidic relatively non-toxic Naja naja atra phospholipases A2. Following methylation a very low residual enzymatic activity (0.4-1% of control) was accompanied by a parallel loss in intraventricular lethality, anticoagulant potency, direct hemolytic action and ability to block directly and indirectly evoked contractions of the mouse phrenic nerve-diaphragm preparation. Since methylation does not impair the enzyme's ability to bind monomeric or micellar substrates or Ca2+, the results suggest that the pharmacologically active region of the molecule is different from the micellular substrate binding site but strongly influenced by the invariant histidine-48 located at the enzymatic active site.

Ethoxyformylation and guanidination of snake venom phospholipases A2: effects on enzymatic activity, lethality and some pharmacological properties.

Lysine residues in the basic and relatively toxic N. nigricollis phospholipase A2 and in the acidic and relatively nontoxic N. n. atra phospholipase A2 were modified by acylation with ethoxyformic anhydride (in the presence or absence of the substrate dihexanoyl lecithin) or guanidination with O-methylisourea. Ethoxyformylation gave rise to some protein fractions in which enzymatic activity was preserved to a greater degree than intraventricular lethality. Guanidination had little effect on the isoelectric point or catalytic activity of either enzyme or on the lethal potency of the N. n. atra enzyme. However, the intraventricular lethality of the N. nigricollis enzyme was decreased much more than was its intravenous lethality, direct hemolytic potency, anticoagulant activity or cardiotoxic action on rat atrium. These results are compared to those previously obtained when the lysines in these two enzymes were carbamylated with potassium cyanate, a procedure which markedly decreased the isoelectric point of the enzymes. It is concluded that charge alone does not account for differences in toxicity. The data also indicate that there are at least two distinct active sites in both enzymes, one being primarily responsible for enzymatic activity and the other(s) associated with lethal and pharmacological effects of the protein. Modification of lysines affects the latter site(s), while having little or no effect on enzymatic activity.