Amplified Ras-MAPK signal states correlate with accelerated EGFR internalization, cytostasis and delayed HER2 tumor onset in Fer-deficient model systems.

The non-receptor tyrosine kinase Fer belongs to a distinct subfamily of F-BAR domain containing kinases implicated in vesicular trafficking and signaling downstream of adhesion and growth factor receptors. Targeted inactivation of the fer gene in a transgenic mouse model of HER2(+), breast cancer was associated with delayed tumor onset and reduced proliferative rates in tumor cells. Fer deficiency was associated with increased rates of epidermal growth factor (EGF)-induced epidermal growth factor receptor (EGFR) internalization and amplified Ras-Raf-Mek-Erk (Ras-MAPK) signaling in primary mammary tumor epithelial cells, as well as increased cytotoxic and anti-proliferative sensitivity to the dual EGFR/HER2 inhibitor Lapatinib (LPN). These observations suggest a model in which accelerated ligand-induced EGFR internalization in Fer-deficient cells hypersensitizes the Ras-MAPK pathway to EGF, resulting in MAPK signal amplification to levels that induce cytostasis, rather than proliferation. Thus, Ras-MAPK cytostatic signaling delays HER2 tumor initiation and increases LPN cytotoxicity in Fer-deficient model systems. Taken together, these data suggest that targeting Fer alone, or in combination with LPN, may be of therapeutic benefit in HER2(+) breast cancer.

Vascular defects in gain-of-function fps/fes transgenic mice correlate with PDGF- and VEGF-induced activation of mutant Fps/Fes kinase in endothelial cells.

BACKGROUND: Fps/Fes is a cytoplasmic tyrosine kinase that is abundantly expressed in the myeloid, endothelial, epithelial, neuronal and platelet lineages. Genetic manipulation in mice has uncovered potential roles for this kinase in hematopoiesis, innate immunity, inflammation and angiogenesis. OBJECTIVE: We have utilized a genetic approach to explore the role of Fps/Fes in angiogenesis. METHODS: A hypervascular line of mice generated by expression of a 'gain-of-function' human fps/fes transgene (fps(MF)) encoding a myristoylated variant of Fps (MFps) was used in these studies. The hypervascular phenotype of this line was extensively characterized by intravital microscopy and biochemical approaches. RESULTS: fps(MF) mice exhibited 1.6-1.7-fold increases in vascularity which was attributable to increases in the number of secondary vessels. Vessels were larger, exhibited varicosities and disorganized patterning, and were found to have defects in histamine-induced permeability. Biochemical characterization of endothelial cell (EC) lines derived from fps(MF) mice revealed that MFps was hypersensitive to activation by vascular endothelial growth factor (VEGF) and platelet-derived growth factor (PDGF). CONCLUSIONS: MFps mediates enhanced sensitization to VEGF and PDGF signaling in ECs. We propose that this hypersensitization contributes to excessive angiogenic signaling and that this underlies the observed hypervascular phenotype of fps(MF) mice. These phenotypes recapitulate important aspects of the vascular defects observed in both VEGF and angiopoietin-1 transgenic mice. The fps/fes proto-oncogene product therefore represents a novel player in the regulation of angiogenesis, and the fps(MF) line of mice constitutes a unique new murine model for the study of this process.

Fps/Fes and Fer non-receptor protein-tyrosine kinases regulate collagen- and ADP-induced platelet aggregation.

Fps/Fes and Fer proto-oncoproteins are structurally related non-receptor protein-tyrosine kinases implicated in signaling downstream from cytokines, growth factors and immune receptors. We show that Fps/Fes and Fer are expressed in human and mouse platelets, and are activated following stimulation with collagen and collagen-related peptide (CRP), suggesting a role in GPVI receptor signaling. Fer was also activated following stimulation with thrombin and a protease-activated receptor4 (PAR4)-activating peptide, suggesting a role in signaling downstream from the G protein-coupled PAR4. There were no detectable perturbations in CRP-induced activation of Syk, PLCgamma2, cortactin, Erk, Jnk, Akt or p38 in platelets from mice lacking Fps/Fes, Fer, or both kinases. Platelets lacking Fps/Fes, from a targeted fps/fes null strain of mice, showed increased rates and amplitudes of collagen-induced aggregation, relative to wild-type platelets. P-Selectin expression was also elevated on the surface of Fps/Fes-null platelets in response to CRP. Fer-deficient platelets, from mice targeted with a kinase-inactivating mutation, disaggregated more rapidly than wild-type platelets in response to ADP. This report provides the first evidence that Fps/Fes and Fer are expressed in platelets and become activated downstream from the GPVI collagen receptor, and that Fer is activated downstream from a G-protein coupled receptor. Furthermore, using targeted mouse models we show that deficiency in Fps/Fes or Fer resulted in disregulated platelet aggregation and disaggregation, demonstrating a role for these kinases in regulating platelet functions.

Hemostatic and hematological abnormalities in gain-of-function fps/fes transgenic mice are associated with the angiogenic phenotype.

The Fps/Fes tyrosine kinase has been implicated in the regulation of hematopoiesis and inflammation. Mice expressing an activated variant of Fps/Fes (MFps) encoded by a gain-of-function mutant transgenic fps/fes allele (fps(MF)) exhibited hematological phenotypes, which suggested that Fps/Fes can direct hematopoietic lineage output. These mice also displayed marked hypervascularity and multifocal-hemangiomas which implicated this kinase in the regulation of angiogenesis. Here we explored the potential involvement of Fps/Fes in the regulation of hemostasis through effects on blood cells and the vascular endothelium. Hematological parameters of fps(MF) mice were characterized by peripheral blood analysis, histology, and transmission electron microscopy. Hemostasis parameters and platelet functions were assessed by flow cytometry and measurements of activated partial thromboplastin time, prothrombin time, thrombin clot time, platelet aggregation, bleeding times and in vitro fibrinolytic assays. Hematological and morphological analyses showed that fps(MF) mice displayed mild thrombocytopenia, anemia, red cell abnormalities and numerous hemostatic defects, including hypofibrinogenemia, hyper-fibrinolysis, impaired whole blood aggregation and a mild bleeding diathesis. fps(MF) mice displayed a complex array of hemostatic perturbations which are reminiscent of hemostatic disorders such as disseminated intravascular coagulation (DIC) and of hemangioma-associated pathologies such as Kasabach-Merritt phenomenon (KMS). These studies suggest that Fps/Fes influences both angiogenic and hemostatic function through regulatory effects on the endothelium.

Characterization of the interaction of recombinant apolipoprotein(a) with modified fibrinogen surfaces and fibrin clots.

Elevated levels of lipoprotein(a) [Lp(a)] in plasma are a significant risk factor for the development of atherosclerotic disease, a property which may arise from the ability of this lipoprotein to inhibit fibrinolysis. In the present study we have quantitated the binding of recombinant forms of apolipoprotein(a) [17K and 12K r-apo(a); containing 8 and 3 copies, respectively, of the major repeat kringle sequence (kringle IV type 2)] to modified fibrinogen surfaces. Iodinated 17K and 12K r-apo(a) bound to immobilized thrombin-modified fibrinogen (i.e., fibrin) surfaces with similar affinities (Kd approximately 1.2-1.6 microM). The total concentration of binding sites (Bmax) present on the fibrin surface was approximately 4-fold greater for the 12K than for the 17K (Bmax values of 0.81 +/- 0.09 nM, and 0.20 +/- 0.01 nM respectively), suggesting that the total binding capacity on fibrin surfaces is reduced for larger apolipoprotein(a) (apo(a)) species. Interestingly, binding of apo(a) to intact fibrin was not detected as assessed by measurement of intrinsic fluorescence of free apo(a) present in the supernatants of sedimented fibrin clots. In other experiments, the total concentration apo(a) binding sites available on plasmin-modified fibrinogen surfaces was shown to be 13.5-fold higher than the number of sites available on unmodified fibrin surfaces (Bmax values of 2.7 +/- 0.3 nM and 0.20 +/- 0.01 nM respectively) while the affinity of apo(a) for these surfaces was similar. The increase in Bmax was correlated with plasmin-mediated exposure of C-terminal lysines since treatment of plasmin-modified fibrinogen surfaces with carboxypeptidase B produced a significant decrease in total binding signal as detected by ELISA (enzyme linked immunosorbent assay). Taken together, these data suggest that apo(a) binds to fibrin with poor affinity (low microM) and that the total concentration of apo(a) binding sites available on modified-fibrinogen surfaces is affected by both apo(a) isoform size and by the increased availability of C-terminal lysines on plasmin-degraded fibrinogen surfaces. However, the low affinity of apo(a) for fibrin indicates that Lp(a) may inhibit fibrinolysis through a mechanism distinct from binding to fibrin, such as binding to plasminogen.

Expression and characterization of apolipoprotein(a) kringle IV types 1, 2 and 10 in mammalian cells.

We have designed expression constructs containing sequences corresponding to apolipoprotein(a) kringle IV types 1, 2 and 10 and used these constructs to transfect human embryonic kidney cells. We have also expressed a mutant form of kringle IV type 2 in which the N-linked glycosylation site has been removed by replacement of an asparagine residue with an alanine. Immunoprecipitation analysis of [35S]Cys-labeled transfected cell culture supernatants resulted in the observation of two bands for kringle IV type 1 (M(r) approximately 30,000 and 26,000), two bands for kringle IV type 2 (M(r) approximately 25,000 and 22,000), two bands for kringle IV type 10 (M(r) approximately 27,000 and 23,000) and one band for the glycosylation mutant (M(r) approximately 22,000). In all cases, observed molecular weights greatly exceeded those predicted from amino acid sequence, suggesting the presence of both N- and O-linked glycans. None of the recombinant single kringles were observed to bind to fibrinogen as determined by ELISA or by co-immunoprecipitation in the case of kringle IV type 10 and only kringle IV type 10 was able to bind to lysine--Sepharose. These data suggest that apo(a) binding to fibrinogen/fibrin may require motif(s) in addition to apo(a) kringle IV type 10.

Antifibrinolytic effect of recombinant apolipoprotein(a) in vitro is primarily due to attenuation of tPA-mediated Glu-plasminogen activation.

The effect of a 17-kringle form of recombinant apo(a) [r-apo(a)] on in vitro fibrin clot lysis was studied. In these assays, fibrin clots were formed in the wells of microtiter plates, and lysis of the clots was monitored by measurement of the turbidity at 405 nm. The results indicate that r-apo(a) produces a dose-dependent antifibrinolytic effect in clots formed using either purified components or barium-adsorbed plasma. This effect was found to be independent of clot structure, since lysis of clots formed using both high and low concentrations of thrombin was prolonged by r-apo(a) to the same extent. The two components of the antifibrinolytic effect of r-apo(a) were determined to be (i) attenuation of tPA-mediated plasminogen activation (the major component) and (ii) inhibition of plasmin degradation of fibrin, although r-apo(a) did not directly attenuate plasmin activity, as measured by S-2251 hydrolysis. r-Apo(a) interfered most substantially with tPA-mediated activation of Glu-plasminogen and less substantially with tPA-mediated Lys-plasminogen activation and urokinase-mediated activation of plasminogen. In summary, we have demonstrated that apo(a) is able to attenuate fibrin clot lysis in vitro, primarily as a consequence of the interference by apo(a) with tPA-mediated Glu-plasminogen activation. These studies illuminate possible mechanisms by which Lp(a) may contribute to the development of vascular disease in vivo.

Binding of recombinant apolipoprotein(a) to extracellular matrix proteins.

Elevated levels of lipoprotein(a), which consists of apolipoprotein(a) [apo(a)] covalently linked to a low-density lipoprotein-like moiety, is an independent risk factor for the development of atherosclerosis. We show that a recombinant form of apo(a) [r-apo(a)] binds strongly to fibronectin and fibrinogen, weakly to laminin, and not at all to von Willebrand factor, vitronectin, or collagen type IV. In contrast to the binding of plasminogen to fibronectin, r-apo(a) binding does not appear to be mediated by lysine-dependent interactions, based on the inability of epsilon-aminocaproic acid concentrations up to 0.2 mol/L to significantly decrease r-apo(a) binding to fibronectin. Plasminogen competed weakly for the binding of r-apo(a) to fibronectin, whereas r-apo(a) completely abolished plasminogen binding. The 29- and 38-kd heparin-binding thermolysin fragments of fibronectin, previously identified as the lipoprotein(a) binding domains, were digested with trypsin, and a peptide that retained the ability to bind r-apo(a) was isolated; the sequence of the peptide (AVTTIPAPTDLK) corresponds to the amino terminus of the 29- and 38-kd domains. A synthetic peptide with this sequence was able to compete effectively with fibronectin for r-apo(a) binding.

Apolipoprotein(a) attenuates endogenous fibrinolysis in the rabbit jugular vein thrombosis model in vivo.

BACKGROUND: In many case-control as well as epidemiological studies, increased lipoprotein(a) [Lp(a)] levels are considered to constitute an independent risk factor for premature coronary artery and cerebrovascular disease. Lp(a) resembles an LDL particle with an additional linked protein [apolipoprotein(a), apo(a)], whose molecular structure has been demonstrated to be homologous to the fibrinolytic proenzyme plasminogen. Because of the high similarity between plasminogen and apo(a), apo(a) may potentially interfere in the fibrinolytic system by competing with plasminogen for fibrin binding sites. In vitro studies have demonstrated that Lp(a) indeed competes with plasminogen binding to fibrin and inhibits tissue plasminogen activator (TPA)-mediated activation of plasminogen. No direct in vivo studies to test this hypothesis have been performed. METHODS AND RESULTS: To test this hypothesis, we studied the effect of a recombinant form of apo(a) on endogenous and TPA-mediated thrombolysis in an in vivo model of experimental venous thrombosis. Thrombi containing either 16 microg r-apo(a), 8 microg r-apo(a), or vehicle (HEPES-buffered saline, control) were formed in the jugular veins of a rabbit and showed significantly reduced endogenous thrombolysis after 60 minutes in a dose-dependent fashion, ID 2.7+/-0.9% and 4.6+/-1.8%, respectively, versus 7.4+/-1.6% of that of the control. High concentrations of incorporated apo(a) significantly reduced TPA-induced thrombolysis (12.2+/-2.5% versus 22.2+/-2.6% in the control thrombi), but no effect of lower concentrations of incorporated r-apo(a) was demonstrated on the exogenous TPA-induced thrombolysis. CONCLUSIONS: The present study demonstrates the attenuation of endogenous fibrinolysis by apo(a) in an in vivo model of experimental venous thrombosis, lending support to the proposed mechanism of impaired fibrinolysis by which Lp(a) may contribute to atherothrombotic disorders.

Apolipoprotein(a) enhances platelet responses to the thrombin receptor-activating peptide SFLLRN.

Elevated levels of lipoprotein(a) [Lp(a)] are correlated with an increased risk of atherosclerotic disease. We examined the effect of recombinant apolipoprotein(a) [r-apo(a)] and Lp(a) on responses of washed human platelets, prelabeled in the dense granules with [14C]serotonin and suspended in Tyrode's solution, to ADP and the thrombin receptor-activating peptide SFLLRN. No effect of the 17 kringle (K), 12K, or 6K r-apo(a) derivatives (at concentrations of 0.35 and 0.7 micromol/L) or Lp(a) (up to 0.1 micromol/L) on primary ADP-induced platelet aggregation was observed. In contrast, weak platelet responses stimulated by 7.5 micromol/L SFLLRN were significantly enhanced by the r-apo(a) derivatives; eg, 0.7 micromol/L 17K r-apo(a) increased aggregation from 15+/-4% to 58+/-6%, release of [14C]serotonin from 9+/-3% to 36+/-6%, and formation of thromboxane A2, measured as its stable metabolite thromboxane B2, from 7+/-1 to 29+/-5 ng/10(9) platelets (n=3; P<0.04 to 0.015). Significant enhancement of aggregation and release of granule contents was observed at a concentration of 17K r-apo(a) as low as 0.175 micromol/L. Purified Lp(a) (0.25 to 0.1 micromol/L) also enhanced SFLLRN-induced aggregation and release in a dose-dependent manner. Although plasminogen (0.7 and 1.5 micromol/L) and low density lipoprotein (0.025 to 0.1 micromol/L) both exhibited potentiating effects on SFLLRN-mediated platelet aggregation, the magnitude of the responses was less than that observed with either the r-apo(a) derivatives or Lp(a). The enhanced responses of platelets via the protease-activated receptor- thrombin receptor in the presence of Lp(a) may contribute to the increased risk of thromboembolic complications of atherosclerosis associated with this lipoprotein.

The solution phase interaction between apolipoprotein(a) and plasminogen inhibits the binding of plasminogen to a plasmin-modified fibrinogen surface.

In the present study, we assessed the binding of recombinant forms of apolipoprotein(a) [r-apo(a)] to plasminogen. Apo(a)-plasminogen interactions were demonstrated to be lysine-dependent, as they were abolished by the addition of epsilon-aminocaproic acid. Binding of r-apo(a) and plasma-derived Lp(a) to Glu-plasminogen was assessed in solution using a mutant form of recombinant plasminogen [Plg(S741C)] labeled at the active site with 5'-(iodoacetamido)fluorescein. High-affinity binding of apo(a) to plasminogen was observed with the 17-kringle r-apo(a) (Kd = 20.1 +/- 3.3 nM) as well as with plasma-derived Lp(a) (Kd = 5.58 +/- 0.08 nM). Binding studies using various truncated and mutant forms of r-apo(a) demonstrated that sequences within apo(a) kringle IV types 2-9 and the strong lysine binding site (LBS) in apo(a) kringle IV type 10 are not required for high-affinity binding to plasminogen. In all cases, the binding stoichiometry for the apo(a)-plasminogen interaction was determined to be 1:1. Binding data obtained using a 17-kringle r-apo(a) derivative lacking the protease-like domain (17KDeltaP; Kd = 3158 +/- 138 nM) indicate that sequences within the protease-like domain of apo(a) mediate its interaction with LBS in plasminogen. We determined that r-apo(a) and plasminogen bind to distinct sites on plasmin-modified fibrinogen with the concentration of plasminogen binding sites exceeding the concentration of r-apo(a) sites by a factor of 10. Furthermore, r-apo(a) is capable of inhibiting the binding of plasminogen to plasmin-modified fibrinogen surfaces, an effect which we show is attributable to the formation of a solution phase apo(a)/plasminogen complex which exhibits a greatly reduced affinity for plasminogen binding sites on plasmin-modified fibrinogen. The results of this study provide new insights into the mechanism by which apo(a) and Lp(a) may inhibit fibrinolysis, thus contributing to the atherothrombotic risk associated with this lipoprotein.