The characterization of novel, dual ErbB-2/EGFR, tyrosine kinase inhibitors: potential therapy for cancer.

The type I receptor tyrosine kinases constitute a family of transmembrane proteins involved in various aspects of cell growth and survival and have been implicated in the initiation and progression of several types of human malignancies. The best characterized of these proteins are the epidermal growth factor receptor (EGFR) and ErbB-2 (HER-2/neu). We have developed potent quinazoline and pyrido-[3,4-d]-pyrimidine small molecules that are dual inhibitors of ErbB-2 and EGFR. The compounds demonstrate potent in vitro inhibition of the ErbB-2 and EGFR kinase domains with IC(50)s <80 nM. Growth of ErbB-2- and EGFR-expressing tumor cell lines is inhibited at concentrations <0.5 microM. Selectivity for tumor cell growth inhibition versus normal human fibroblast growth inhibition ranges from 10- to >75-fold. Tumor growth in mouse s.c. xenograft models of the BT474 and HN5 cell lines is inhibited in a dose-responsive manner using oral doses of 10 and 30 mg/kg twice per day. In addition, the tested compounds caused a reduction of ErbB-2 and EGFR autophosphorylation in tumor fragments from these xenograft models. These data indicate that these compounds have potential use as therapy in the broad population of cancer patients overexpressing ErbB-2 and/or EGFR.

Nitric oxide stimulates the phosphorylation of rap1b in human platelets and acts synergistically with iloprost.

Phosphorylation of rap 1b in human platelets correlates with both an upward shift of the protein on sodium dodecyl sulfate polyacrylamide gels and the translocation of the phosphorylated protein to the cytosolic fraction of platelets. We reported that this phenomenon occurs in platelets in response to agents that stimulate adenylate cyclase and thereby activate the cyclic AMP-dependent protein kinase. We now have evidence that phosphorylation of rap1b in platelets is also induced by nitric oxide generating compounds through stimulation of guanylate cyclase and activation of the cyclic GMP-dependent protein kinase. We observed time-dependent phosphorylation of rap1b and dose-dependent inhibition of collagen-stimulated aggregation in washed platelets incubated with S-nitroso serum albumin. In the presence of a combination of iloprost and 3-morpholinosydnonimine, when both PKA and PKG are activated, phosphorylation of rap1b increased synergistically to a level three times higher than the sum of their individual actions.

Glyceraldehyde-3-phosphate dehydrogenase is required for the transport of nitric oxide in platelets.

Nitric oxide (NO) or NO-generating compounds like sodium nitroprusside (SNP) increase cellular levels of cGMP and produce S-nitrosylation of glyceraldehyde-3-phosphate dehydrogenase [GAPDH; D-glyceraldehyde-3-phosphate:NAD+ oxidoreductase (phosphorylating), EC 1.2.1.12]. In search of a reagent that could discriminate between these two effects, we used the sesquiterpene antibiotic koningic acid, which binds to GAPDH at the Cys-149 of the active site. Koningic acid inhibited basal and sodium nitroprusside-stimulated NAD-dependent covalent modification of purified rabbit muscle GAPDH in a dose-dependent manner. Furthermore, we tested the effect of koningic acid on human platelets. Approximately 90% of GAPDH is present in the cytosol of human platelets, and the exposure of platelet cytosol to koningic acid inhibited GAPDH activity, while the soluble guanylyl cyclase (basal and sodium nitroprusside-stimulated) activity remained unaltered. Pretreatment of intact platelets with koningic acid slowed the rate of aggregation induced by a submaximal concentration of thrombin. In addition, the antibiotic also inhibited the cGMP increases triggered by SNP, S-nitroso-N-acetylpenicillamine (SNAP), and 3-morpholinosyndomidine (SIN-1) but failed to prevent an increase in cGMP caused by nitrosylated albumin. Under the same conditions, koningic acid also inhibited basal and SNP- SNAP-, and SIN-1-stimulated NAD-dependent modification of GAPDH and its enzymatic activity. These results suggest that the mechanism of delivery of NO from SNP, SNAP, and SIN-1 to platelets may require the active form of GAPDH. When NO is delivered by nitrosylated albumin, active GAPDH was not necessary.

Purification and cDNA sequence of an inducible nitric oxide synthase from a human tumor cell line.

A combination of cytokines induced the expression of nitric oxide synthase (NOS) in a human colorectal adenocarcinoma cell line, DLD-1. We have purified the enzyme and examined some of its biochemical properties. An antiserum to an inducible NOS from murine macrophages cross-reacted with the DLD-1 NOS. The purified human and murine enzymes displayed a similar lack of dependence on exogenous calcium and calmodulin for activity, which contrasts with the requirement for calcium and calmodulin of purified brain and endothelial isoforms of NOS. We have also isolated a cDNA for a cytokine-induced NOS from DLD-1 cells. Sequence analysis of this cDNA and NOS cDNAs from human liver, smooth muscle, and macrophages suggests that, at the genetic level, there is a single isoform of human-inducible NOS.

Nitric oxide-induced S-nitrosylation of glyceraldehyde-3-phosphate dehydrogenase inhibits enzymatic activity and increases endogenous ADP-ribosylation.

Using conditions that produced chronic inflammation in rat liver, we were able to find a correlation between induction of nitric oxide production and inhibition of glyceraldehyde-3-phosphate dehydrogenase (GAPDH; EC 1.2.1.12). This enzyme is a tetramer composed of identical M(r) 37,000 subunits. The tetramer contains 16 thiol groups, four of which are essential for enzymatic activity. Our information indicates that four thiol groups are S-nitrosylated by exposure to authentic nitric oxide (NO) gas. Furthermore, NO decreased GAPDH activity while increasing its auto-ADP-ribosylation. Reduced nicotinamide adenine dinucleotide and dithiothreitol are required for the S-nitrosylation of GAPDH caused by the NO-generating compound sodium nitroprusside. Our results suggests that a new and important action of nitric oxide on cells is the S-nitrosylation and inactivation of GAPDH. S-Nitrosylation of GAPDH may be a key covalent modification of multiple regulatory consequences in chronic liver inflammation.

Antisera specific for rap 1 proteins distinguish between processed and nonprocessed rap 1b.

Polyclonal antisera were generated against synthetic peptides corresponding to distinct regions of the rap 1 protein sequences. A "rap 1-common" antiserum, prepared against an 18-amino acid segment of the rap 1a protein near the proposed GTP-binding region, reacted with both rap 1a and rap 1b recombinant proteins expressed in Escherichia coli and with two low molecular weight GTP-binding proteins of 22 and 24 kDa in unstimulated human platelets. An antiserum raised against a carboxyl-terminal peptide of rap 1b containing the putative site of post-translational processing reacted strongly with bacterial-expressed recombinant rap 1b and with a 24-kDa GTP-binding protein in platelets, but not with recombinant rap 1a or a 22-kDa GTP-binding protein. The mobility on sodium dodecyl sulfate-polyacrylamide gel electrophoresis of this rap 1b immunoreactive protein coincided with that of bacterial-expressed rap 1b and not with the faster migrating form of rap 1b that incorporates radioactivity from [3H]mevalonic acid in the insect/baculovirus system. This suggests that our rap 1b-specific antiserum recognizes only one form of rap 1b, that which has not undergone carboxyl-terminal post-translational processing.

Inhibition of ras-induced germinal vesicle breakdown in Xenopus oocytes by rap-1B.

A cDNA clone (Krev-1) has recently been identified that possesses the ability to reverse the transformed phenotype when introduced into a K-ras-transformed NIH/3T3 cell line. The Krev-1 protein, also known as rap-1A, was found to share 50% homology with the ras proteins. The rap-1A protein has also been shown to block the interaction of ras with its GTPase activating protein in vitro, leading to speculation regarding its role in vivo. A closely related protein, rap-1B, has also been identified in platelets, human erythroleukemia cells, neutrophils, and aortic smooth muscle cells. Unlike rap-1A, rap-1B has been shown to be phosphorylated in platelets. Given the high degree of similarity between the amino acid sequences of rap-1A and rap-1B, we sought to investigate the effect of microinjected rap-1B on H-ras(Val12)-induced germinal vesicle breakdown in Xenopus laevis oocytes. In this assay system, equimolar concentrations of rap-1B were found to block germinal vesicle breakdown triggered by the oncogenic ras protein. However, in the presence of IGF-1, this inhibition was not observed. Moreover, rap-1B is readily phosphorylated in the oocytes.

RAP2B: a RAS-related GTP-binding protein from platelets.

A platelet cDNA expression library was screened with the monoclonal antibody M90, which recognizes a specific epitope on RAS-encoded p21 proteins (amino acids 107-130). DNA sequence analysis of one clone revealed that it encoded a partial amino acid sequence of a protein closely related to RAP2, which we have named RAP2B. A repeated screening of the platelet cDNA library with an internal Ava I fragment of the RAP2B cDNA allowed the isolation of a full-length cDNA for the RAP2B sequence. RAP2B is 90% identical to RAP2 at the amino acid level with the most variability at the carboxyl terminus of the protein. Oligonucleotides were synthesized to complete the amino acid sequence of the RAP2B protein and the entire sequence was expressed in Escherichia coli. Analysis of crude soluble extracts indicated that RAP2B was a Mr 22,000 protein that specifically bound GTP on blots. Moreover, incubation of similar extracts with the catalytic subunit of cAMP-dependent protein kinase did not cause phosphorylation of RAP2B, as had been observed for the closely homologous proteins, RAP1A and RAP1B. These results suggest that RAP2B, like the other RAP proteins, is a low molecular weight GTP-binding protein in human platelets.

Thrombolamban, the 22-kDa platelet substrate of cyclic AMP-dependent protein kinase, is immunologically homologous with the Ras family of GTP-binding proteins.

Platelet inhibition by agents that increase intracellular levels of cAMP is associated with cAMP-dependent phosphorylation of specific intracellular proteins, including a membrane-associated 22-kDa microsomal protein called thrombolamban. In view of recent studies suggesting that platelets also contain 22-kDa GTP-binding proteins that are homologous with ras-encoded p21 proteins, the present work was undertaken to examine the possibility that thrombolamban and the Ras-like proteins were the same. Platelet microsomes were labeled with [gamma-32P]ATP and the labeled proteins were examined by autoradiography of sodium dodecyl sulfate/polyacrylamide gels. On Western blots of both one-dimensional and two-dimensional gels, thrombolamban immunoreacted with M90, a monoclonal antibody that recognizes the GTP-binding domain of Ras p21 proteins, but not with Y13-259, a monoclonal antibody that recognizes another domain and is specific for Ras proteins. Overlay experiments with unlabeled platelet microsomes demonstrated numerous low molecular weight proteins that bound [alpha-32P]GTP, although none could be identified as thrombolamban. Finally, thrombolamban was immunoprecipitated by M90. These studies show that thrombolamban is a low molecular weight protein that is immunologically related to the Ras family of GTP-binding proteins.

A ras-related protein is phosphorylated and translocated by agonists that increase cAMP levels in human platelets.

The antigenicity of platelet proteins was assayed against various monoclonal antibodies (mAbs) that recognize specific epitopes of the ras-encoded p21 protein. mAb M90, which detects the region of p21 protein within amino acids 107-130 and inhibits its GTP-binding activity, strongly reacted with a 22-kDa protein present in the particulate fraction of human platelets. Other mAbs against ras-encoded proteins, including Y13-259, which efficiently detects ras proteins from a variety of organisms, did not recognize the platelet 22-kDa protein. Transfer of the platelet 22-kDa protein to nitrocellulose paper showed that the protein binds [alpha-32P]GTP. Moreover, preincubation of the transferred protein with mAb M90 drastically reduced its GTP-binding activity. Treatment of platelets with iloprost, a prostacyclin analog, caused (i) a time-dependent increase of a 24-kDa protein that is recognized by mAb M90 in particulate and cytosolic fractions and (ii) the gradual decrease of the 22-kDa protein from the particulate fraction. When platelets were labeled with 32P and then treated with iloprost, the 24-kDa protein was found to be phosphorylated. The 32P-labeled 24-kDa protein was specifically immunoprecipitated by mAb M90. These results suggest that appearance of the 24-kDa protein results from phosphorylation of the 22-kDa protein, which shifts its mobility to a higher molecular mass area.

Platelet cytosolic 44-kDa protein is a substrate of cholera toxin-induced ADP-ribosylation and is not recognized by antisera against the alpha subunit of the stimulatory guanine nucleotide-binding regulatory protein.

ADP-ribosylation induced by cholera toxin and pertussis toxin was studied in particulate and cytosolic fractions of human platelets. Platelets were disrupted by a cycle of freezing and thawing in the presence of a hyposmotic buffer containing protease inhibitors. In both fractions, the A subunit of cholera toxin ADP-ribosylates two proteins with molecular masses of 42 and 44 kDa, whereas pertussis toxin ADP-ribosylates a 41-kDa polypeptide. Two antisera against the alpha subunit of the stimulatory guanine nucleotide-binding regulatory protein recognize only the 42-kDa polypeptide. Cholera toxin-induced ADP-ribosylation of the 42- and 44-kDa proteins is reduced by pretreatment of platelets with iloprost, a prostacyclin analog. The 44-kDa protein, which is substrate of cholera toxin, could be extracted completely from the membrane and recovered in the cytosolic fraction when the cells were disrupted by Dounce homogenization and the pellet was extensively washed. A 44-kDa protein can also be labeled with 8-azidoguanosine 5'-[alpha-32P]triphosphate in the cytosol and membranes. These findings indicate that cholera and pertussis toxins produced covalent modifications of proteins present in particulate and cytosolic platelet fractions. Moreover, the 44-kDa protein might be an alpha subunit of a guanine nucleotide-binding regulatory protein that is not recognized by available antisera.

Specific binding of [alpha-32P]GTP to cytosolic and membrane-bound proteins of human platelets correlates with the activation of phospholipase C.

We have assessed the binding of [alpha-32P]GTP to platelet proteins from cytosolic and membrane fractions. Proteins were separated by NaDodSO4/PAGE and electrophoretically transferred to nitrocellulose. Incubation of the nitrocellulose blots with [alpha-32P]GTP indicated the presence of specific and distinct GTP-binding proteins in cytosol and membranes. Binding was prevented by 10-100 nM GTP and by 100 nM guanosine 5'-[gamma-thio]triphosphate (GTP[gamma S]) or GDP; binding was unaffected by 1 nM-1 microM ATP. One main GTP-binding protein (29.5 kDa) was detected in the membrane fraction, while three others (29, 27, and 21 kDa) were detected in the soluble fraction. Two cytosolic GTP-binding proteins (29 and 27 kDa) were degraded by trypsin; another cytosolic protein (21 kDa) and the membrane-bound protein (29.5 kDa) were resistant to the action of trypsin. Treatment of intact platelets with trypsin or thrombin, followed by lysis and fractionation, did not affect the binding of [alpha-32P]GTP to the membrane-bound protein. GTP[gamma S] still stimulated phospholipase C in permeabilized platelets already preincubated with trypsin. This suggests that trypsin-resistant GTP-binding proteins might regulate phospholipase C stimulated by GTP[gamma S].

Ionophore A23187 stimulates phosphorylation of the 40,000 dalton protein in human platelets without phospholipase C activation.

The Ca2+ ionophore A23187 (0.2-5 microM) stimulates the phosphorylation of the substrates of protein kinase C (40,000 dalton protein) and myosin light chain kinase (20,000 dalton protein) in the presence or absence of cyclooxygenase inhibitors. In the presence of cyclooxygenase inhibitors or millimolar Ca2+ there is no stimulation of phospholipase C by A23187. Fingerprints of the 32P-labeled 40,000 dalton protein isolated from platelets that have been stimulated with A23187, thrombin, phorbol 12,13-dibutyrate and 1,2-didecanoylglycerol were identical. Higher concentrations of A23187 (1-5 microM) induced the loss of polyphosphoinositides through phosphomonoesterase activity.

Treatment of human platelets with trypsin, thrombin, or collagen inhibits the pertussis toxin-induced ADP-ribosylation of a 41-kDa protein.

Permeabilization of human platelets with saponin (15-25 micrograms/ml) allows the determination of the ADP-ribosylation of a 41-kDa protein by pertussis toxin. The ADP-ribosylated protein is present in the particulate fraction. ADP-ribosylation of the 41-kDa protein increases for 20 min; it is not affected by indomethacin, prostacyclin, and 1,2-diacylglycerols but is inhibited by 1 mM Ca2+ and phorbol esters. Treatment of platelets with trypsin, thrombin, or collagen before saponin addition precludes subsequent pertussis toxin-induced ADP-ribosylation of the 41-kDa protein. The effect of trypsin or thrombin is blocked by soybean trypsin inhibitor and leupeptin. Trypsin proteolytically cleaves the ADP-ribosylated 41-kDa protein to an ADP-ribosylated fragment slightly smaller than 20 kDa. The results suggest that a modification of a guanine nucleotide-binding regulatory protein is associated with the actions of trypsin, thrombin, and collagen on platelet activation.

Adhesion of human platelets to collagen in the presence of prostacyclin, indomethacin and compound BW 755C.

Prostacyclin (1 ng to 2 micrograms per ml), which effectively inhibits platelet secretion and aggregation, does not affect adhesion of a proportion of platelets (10-38%) to collagen (50-100 micrograms/ml). Adhesion is not detectable by changes of light transmission (as measured in the optical aggregometer) and is not affected by inhibitors of cyclooxygenase and lipoxygenase enzymes such as indomethacin and compound BW 755C. This adhesion is independent of the collagen concentration (50-400 micrograms/ml) and the incubation time (5-20 min). This suggests that adhesion to collagen is related to a specific platelet population. Adhesion in the presence of prostacyclin, indomethacin and BW 755C occurs in parallel with the formation of a limited amount of phosphatidic acid. Under those conditions it is also possible to observe some phosphorylation of a 40,000 dalton protein which is a substrate for protein kinase C activity. Phosphorylation of the 20,000 dalton protein, or myosin light chain, is less evident. Chlorpromazine (25-100 micrograms/ml) inhibited the adhesion of platelets to collagen, but propanolol (0.5-4 microM) was inactive. The adhesion of platelets to collagen in these experiments parallels the formation of a fraction of phosphatidic acid and 40,000 dalton protein phosphorylation, which are independent of the increased levels of platelet cyclic-AMP induced by high concentrations of prostacyclin. It is also independent of the formation of cyclooxygenase or lipoxygenase products.

Collagen stimulates [3H]inositol trisphosphate formation in indomethacin-treated human platelets.

The present study investigates the pathway of metabolism of inositol phospholipids in human platelets exposed to collagen. Platelet activation by collagen was preceded by a lag phase usually lasting 10-20 s. Formation of [3H]inositol trisphosphate (IP3) was not observed during this period, but occurred in parallel with the onset of aggregation, release of ATP and phosphorylation of a 20 000 Da and a 40 000 Da protein. Indomethacin treatment partially inhibited all of these responses. Aggregation and ATP release, but not IP3 formation, were further inhibited in indomethacin-treated platelets loaded with the fluorescent Ca2+ indicator, quin2. Under these conditions there was no detectable mobilization of Ca2+. These results demonstrate that activation of platelets by collagen is associated with rapid hydrolysis of polyphosphoinositides by phospholipase C, thereby producing IP3. This observation is discussed in relation to IP3 as a possible Ca2+-mobilizing agent.

Exogenous sn-1,2-diacylglycerols containing saturated fatty acids function as bioregulators of protein kinase C in human platelets.

The ability of exogenous sn-1,2-diacylglycerols and analogs to function as bioregulators of protein kinase C in human platelets was investigated. The activation of protein kinase C in platelets is indicated by specific phosphorylation of a 40,000-dalton protein. Dihexanoylglycerol, dioctanoylglycerol (diC8), didecanoylglycerol, and sn-1-oleoyl-2-acetylglycerol were active in stimulating 40,000-dalton protein phosphorylation. Only a trace of phosphorylation was elicited by dibutyrylglycerol. Phosphorylation was not induced by analogs of diC8 in which an -H, -SH, or -Cl group replaced the free -OH, nor by monoacylglycerols or long chain diacylglycerols. Maximum phosphorylation was induced by dihexanoylglycerol, diC8, and didecanoylglycerol at concentrations from 5 to 20 microM and between 5 and 30 S after exposure of platelets to these diacylglycerols. Under conditions of maximal phosphorylation of the 40,000-dalton protein, these diacylglycerols did not induce phosphatidylinositol turnover, or platelet aggregation, or stimulate release of ATP or serotonin. A small degree of aggregation was evident with platelets isolated in the absence of prostacyclin, and release of serotonin was observed when 1 mM Ca2+ or submaximal concentrations of ionophore A23187 were included. These results are consistent with a model in which platelet activation requires the simultaneous formation of two intracellular signals, diacylglycerols and Ca2+. These diacylglycerols and diacylglycerol analogs provide useful tools to investigate the function of diacylglycerols as bioregulators in intact cells.