TGF-alpha-driven tumor growth is inhibited by an EGF receptor tyrosine kinase inhibitor.

The simultaneous presence of the EGFR and its ligand TGF-alpha in human tumor tissues suggests that autocrine TGF-alpha stimulation drives tumor growth. Here we show that autocrine TGF-alpha stimulation does cause increased tumor growth in vivo, an effect that was proven to be mediated via EGFR activation, and that this TGF-alpha/EGFR autocrine loop was accessible to an EGFR specific tyrosine kinase inhibitor. Clones of the EGFR expressing glioma cell line U-1242 MG were transfected with TGF-alpha cDNA using a tetracycline-inhibitory system for gene expression. TGF-alpha expression was inhibited by the presence of tetracycline, and subcutaneous tumors forming from cell lines injected into nude mice could be inhibited by feeding mice tetracycline. We confirmed that TGF-alpha mRNA and protein were present in these tumors and that, subsequently, the endogenous EGFR was activated. Tumor growth could be inhibited by an EGFR specific tyrosine kinase inhibitor of the type 4-(3-chloroanilino)-6,7-dimethoxy-quinazoline, administered daily by intraperitoneal injection, thereby interrupting the autocrine loop.     

Ligand-independent oncogenic transformation by the EGF receptor requires kinase domain catalytic activity

The retroviral oncogene S3-v-erbB is a transduced, truncated form of the avian EGF (ErbB-1) receptor. Infection of avian fibroblasts with a retroviral vector expressing S3-v-ErbB results in ligand-independent cell transformation, which is accompanied by the assembly of a transformation-specific phosphoprotein signaling complex and anchorage-independent cell growth. It previously had been reported, using lysine-721 mutants (K721), that kinase domain function was required for ErbB-mediated cell transformation. However, since these initial reports, several studies using aspartate-813 mutants (D813) have demonstrated the ability of kinase-impaired ErbB receptors to induce mitogenic signal transduction pathways and cell transformation in a ligand-dependent manner. To determine the necessity of ErbB receptor kinase domain catalytic activity in ligand-independent cell transformation, we created S3-v-ErbB-K(-), a kinase-impaired oncoprotein constructed by replacing aspartate-813 with alanine (D813A). Subcellular routing as well as cell surface membrane and nuclear localization of the S3-v-ErbB-K(-) mutant receptor were unaffected by impairment of kinase activity. In contrast, avian fibroblasts expressing S3-v-ErbB-K(-) do not form the characteristic transformation-specific phosphoprotein complex, or induce soft agar colony growth in vitro. These results suggest that in contrast to ligand-dependent oncogenic signaling, ligand-independent cell transformation by a constitutively activated mutant form of the EGF receptor requires receptor kinase catalytic activity. In addition, these results demonstrate that phosphorylation and assembly of downstream signaling complexes require tyrosine phosphorylation events that are directly mediated by oncogenic forms of the EGF receptor.     

Insulin receptor serine kinase activation by casein kinase 2 and a membrane tyrosine kinase

The insulin receptor (IR) tyrosine kinase can apparently directly phosphorylate and activate one or more serine kinases. The identities of such serine kinases and their modes of activation are still unclear. We have described a serine kinase (here designated insulin receptor serine (IRS) kinase) from rat liver membranes that co-purifies with IR on wheat germ agglutinin-agarose. The kinase was activated after phosphorylation of the membrane glycoproteins by casein kinase-1, casein kinase-2, or casein kinase-3 (Biochem Biophys Res Commun 171:75–83, 1990). In this study, IRS kinase was further characterized. The presence of vanadate or phosphotyrosine in reaction mixtures was required for activation to be observed. Phosphoserine and phosphothreonine are only about 25% as effective as phosphotyrosine, whereas sodium fluoride and molybdate were ineffective in supporting activation. Vanadate and phosphotyrosine support IRS kinase activation by apparently inhibiting phosphotyrosine protein phosphatases present among the membrane glycoproteins. IR -subunit, myelin basic protein, and microtubule-associated protein-2 are good substrates for IRS kinase. The kinase prefers Mn²⁺ (K_a=1.3 mM) as a metal cofactor. Mg²⁺ (K_a=3.3 mM) is only 30% as effective as Mn²⁺. The kinase activity is stimulated by basic polypeptides, with greater than 30-fold activation achieved with polylysine and protamine. Our results suggest that both serine/threonine and tyrosine phosphorylation are required for activation of IRS kinase. Serine phosphorylation is catalyzed by one of the casein kinases, whereas tyrosine phosphorylation is catalyzed by a membrane tyrosine kinase, possibly IR tyrosine kinase. (Mol Cell Biochem121: 167–174, 1993)     

Determining the binding mode and binding affinity constant of tyrosine kinase inhibitor PD153035 to DNA using optical tweezers

Accurately predicting binding affinity constant (K_A) is highly required to determine the binding energetics of the driving forces in drug–DNA interactions. Recently, PD153035, brominated anilinoquinazoline, has been reported to be not only a highly selective inhibitor of epidermal growth factor receptor but also a DNA intercalator. Here, we use a dual-trap optical tweezers to determining K_A for PD153035, where K_A is determined from the changes in B-form contour length (L) of PD153035–DNA complex. Here, L is fitted using a modified wormlike chain model. We found that a noticeable increment in L in 1 mM sodium cacodylate was exhibited. Furthermore, our results showed that K_A = 1.18(±0.09) × 10⁴ (1/M) at 23 ± 0.5 °C and the minimum distance between adjacent bound PD153035 ≈ 11 bp. We anticipate that by using this approach we can determine the complete thermodynamic profiles due to the presence of DNA intercalators.     

Effect of cations on the tyrosine kinase activity of the insulin receptor: Inhibition by fluoride is magnesium dependent

We have recently reported that fluoride interacts directly with the insulin receptor, which causes inhibition of its phosphotransferase activity. The inhibitory effect of fluoride on phosphotransferase activity is not due to the formation of complexes with aluminium and occurs in the absence of alterations to the binding of ATP or insulin. In this report we substantiate that the tyrosine kinase activity of insulin receptors partially purified from rat skeletal muscle shows a strict requirement of Mg²⁺ ions (Ka near 11 mM). This effect of Mg²⁺ was inhibited in a competitive manner by Mn²⁺, which is compatible with competition of both divalent ions for binding sites. The inhibition of tyrosine kinase activity caused by fluoride was dependent on the concentration of Mg²⁺ in the medium and no inhibitory effect was detected at low concentrations of Mg²⁺. Moreover, the addition of increasing concentrations of Mn²⁺ in the presence of a constant high concentr rease in the inhibitory effect of fluoride. These results indicate that the Mg-insulin receptor complex is the major fluoride-susceptible form. Based on the characteristics of the inhibition of tyrosine kinase shown by fluoride it might be proposed that its action is exerted by the formation of multi-ionic MgF complexes analogous to Pi, which bind to the insulin receptor kinase.     

Cytoplasmic signalling by the c-Abl tyrosine kinase in normal and cancer cells

c-Abl is a non-receptor tyrosine kinase which is localized both in the nucleus and cytoplasm, and is involved in the regulation of cell growth, survival and morphogenesis. Although c-Abl nuclear function has been extensively studied, recent data also indicate an important role in cytoplasmic signalling through mitogenic and adhesive receptors. Here, we review the mechanisms by which growth factors promote cytoplasmic c-Abl activation and signalling and its function in the induction of DNA synthesis, changes in cell morphology and receptor endocytosis. The importance of de-regulated c-Abl cytoplasmic signalling in solid tumours is also discussed.     

Disruption of a putative SH3 domain and the proline-rich motifs in the 53-kDa substrate of the insulin receptor kinase does not alter its subcellular localization or ability to serve as a substrate

The recently identified 53-kDa substrate of the insulin receptor family was further characterized in several retroviral-generated stable cell lines overexpressing the wild type and various mutant forms of the protein. To facilitate the study of its subcellular localization in NIH3T3 cells overexpressing insulin receptor, a myc epitope-tag was added to the carboxy terminus of the 53-kDa protein. Like the endogenous protein in Chinese hamster ovary cells, the expressed myc-tagged 53-kDa protein was found partially in the particulate fraction and was tyrosine phosphorylated in insulin-stimulated cells. Immunofluorescence studies showed for the first time that a fraction of the 53-kDa protein was localized to the plasma membrane. Confocal microscopy of cells double-labeled with antibodies to the insulin receptor and the myc epitope showed the two proteins co-localize at the plasma membrane at the level of light microscopy. Further analyses of the protein sequence of the 53-kDa substrate revealed the presence of a putative SH3 domain and two proline-rich regions, putative binding sites for SH3 and WW domains. Disruption of these three motifs by the introduction of previously characterized point mutations did not affect the membrane localization of the 53-kDa protein, its ability to serve as substrate of the insulin receptor, or its colocalization with the insulin receptor, suggesting these domains are not important in the subcellular targeting of the protein and instead may function in the interaction with subsequent signaling proteins. J. Cell. Biochem. 68:139–150, 1998. © 1998 Wiley-Liss, Inc.     

Termination of tyrphostin AG1478 application results in different recovery of EGF receptor tyrosine residues 1045 and 1173 phosphorylation in A431 cells

Tyrphostin AG1478 is known as a specific and reversible inhibitor of TK (tyrosine kinase) activity of the EGFR [EGF (epidermal growth factor) receptor]. It is attractive as an anticancer agent for cancers with elevated EGFR TK levels. However, post‐application effects of AG1478 are not well studied. We have analysed EGFR phosphorylation after termination of AG1478 application using human epidermoid carcinoma A431 cells. It was found that AG1478 inhibitory action is fast, but not fully reversible: removal of tyrphostin resulted in incomplete restoration of the overall EGFR phosphorylation. Analysing the state of two individual autophosphorylation sites of internalized EGFR, Tyr¹⁰⁴⁵ and Tyr¹¹⁷³, we demonstrated that phosphorylation of Tyr¹¹⁷³ involved in stimulation of the MAPK (mitogen‐activated protein kinase) cascade was restored much more efficiently than that in position 1045, which binds the ubiquitin ligase c‐Cbl and is necessary for targeting the receptor for lysosomal degradation. c‐Cbl association with EGFR abolished by AG1478 was not reestablished after tyrphostin cessation. As a consequence, ubiquitination‐dependent EGFR delivery to lysosomes was blocked, while phosphorylation of ERK1/2 (extracellular‐signal‐regulated kinase 1/2) was even increased. Thus, after termination of AG1478, the intracellular level of the inhibitor can be reached at which mitogenic signalling will be restored, whereas the EGFR negative regulation due to lysosomal degradation will not.     

Characterization and development of a peptide substrate-based phosphate transfer assay for the human vascular endothelial growth factor receptor-2 tyrosine kinase

Vascular endothelial growth factor (VEGF), a critical regulator in angiogenesis, exerts its angiogenic effect via binding to its receptor, VEGF receptor-2 tyrosine kinase (VEGFR2) or kinase insert domain-containing receptor (Kdr), on the surface of endothelial cells. Kdr-mediated signaling plays an important role in the proliferation, migration, differentiation, and survival of endothelial cells. Therefore, the inhibition of this signaling pathway represents a promising therapeutic approach for the discovery of novel anticancer agents by destabilizing the progression of solid tumors via abrogating tumor-induced angiogenesis. To explore Kdr as an anticancer target and further characterize the enzyme, we purified a cytoplasmic domain of human Kdr (Kdr-CD) and characterized its autophosphorylation activity. We also designed and synthesized peptides containing amino acid sequences corresponding to the autophosphorylation sites of Kdr and developed a simple, robust, high-throughput assay for measuring the phosphate transfer activity of the enzyme. This assay was validated by the experiments showing that the phosphate transfer activity of the purified Kdr-CD required Mg2+ or Mn2+ and preactivation by adenosine 5'-triphosphate (ATP) and was inhibited by known Kdr inhibitors. Using this assay, we examined effects of Mg2+ and Mn2+ on the enzyme activity; optimized the concentrations of Kdr-CD, peptide and ATP substrates, and metal ions in the assay; and determined the kinetic properties of the enzyme for the peptide and ATP as well as IC50 values of two known Kdr inhibitors. Thus, the results of these studies have validated the utilities of this assay for biochemical characterizations of the enzyme and its inhibitors. This approach of designing peptides corresponding to the autophosphorylation sites of Kdr as substrates for the enzyme has general practical implications to other kinases.     

Inhibition of the human P2X7 receptor by a novel protein tyrosine kinase antagonist

A panel of 18 protein tyrosine kinase antagonists were tested for their inhibitory effect on human P2X₇ receptor-mediated ⁸⁶Rb⁺ (K⁺) efflux. The most potent compound (compound P), a phthalazinamine derivative and an inhibitor of vascular endothelial growth factor receptor kinase, blocked ATP-induced ⁸⁶Rb⁺-efflux in human B-lymphocytes and erythrocytes by 76% and 66%, respectively. This inhibition was dose-dependent in both cell types with an IC₅₀ of ∼5 μM. Kinetic analysis showed compound P was a non-competitive inhibitor of P2X₇. This compound also inhibited ATP-induced ethidium⁺ influx into B-lymphocytes and P2X₇-transfected-HEK-293 cells, as well as ATP-induced ⁸⁶Rb⁺-efflux from canine erythrocytes. Externally, but not internally, applied compound P impaired ATP-induced inward currents in P2X₇-transfected-HEK-293 cells. This study demonstrates that a novel protein tyrosine kinase antagonist directly impairs native and recombinant human P2X₇ receptors. The data suggests that antagonists which target ATP-binding sites of kinases may potentially block the P2X₇ receptor.     

Tyrphostin AG556 increases the activity of large conductance Ca2+‐activated K+ channels by inhibiting epidermal growth factor receptor tyrosine kinase

The present study was designed to investigate whether large conductance Ca²⁺‐activated K⁺ (BK) channels were regulated by epidermal growth factor (EGF) receptor (EGFR) tyrosine kinase. BK current and channel tyrosine phosphorylation level were measured in BK‐HEK 293 cells expressing both functional α‐subunits and the auxiliary β1‐subunits using electrophysiology, immunoprecipitation and Western blotting approaches, respectively, and the function of rat cerebral basilar arteries was determined with a wire myography system. We found that BK current in BK‐HEK 293 cells was increased by the broad spectrum protein tyrosine kinase (PTK) inhibitor genistein and the selective EGFR tyrosine kinase inhibitor AG556, one of the known tyrphostin. The effect of genistein or AG556 was antagonized by the protein tyrosine phosphatase (PTP) inhibitor orthovanadate. On the other hand, orthovanadate or EGF decreased BK current, and the effect was counteracted by AG556. The tyrosine phosphorylation level of BK channels (α‐ and β1‐subunits) was increased by EGF and orthovanadate, while decreased by genistein and AG556, and the reduced tyrosine phosphorylation of BK channels by genistein or AG556 was reversed by orthovanadate. Interestingly, AG556 induced a remarkable enhancement of BK current in rat cerebral artery smooth muscle cells and relaxation of pre‐contracted rat cerebral basilar arteries with denuded endothelium, and these effects were antagonized by the BK channel blocker paxilline or orthovanadate. These results demonstrate that tyrosine phosphorylation of BK channels by EGFR kinase decreases the channel activity, and inhibition of EGFR kinase by AG556 enhances the channel activity and dilates rat cerebral basilar arteries.     

Nordihydroguaiaretic acid (NDGA), an inhibitor of the HER2 and IGF-1 receptor tyrosine kinases, blocks the growth of HER2-overexpressing human breast cancer cells

We have reported that nordihydroguaiaretic acid (NDGA) inhibits the tyrosine kinase activities of the IGF-1 receptor (IGF-1R) and the HER2 receptor in breast cancer cells. Herein, we studied the effects of NDGA on the growth of estrogen receptor (ER) positive MCF-7 cells engineered to overexpress HER2 (MCF-7/HER2-18). These cells are an in vitro model of HER2-driven, ER positive, tamoxifen resistant breast cancer. NDGA was equally effective at inhibiting the growth of both parental MCF-7 and MCF-7/HER2-18 cells. Half maximal effects for both cell lines were in the 10-15 microM range. The growth inhibitory effects of NDGA were associated with an S phase arrest in the cell cycle and the induction of apoptosis. NDGA inhibited both IGF-1R and HER2 kinase activities in these breast cancer cells. In contrast, Gefitinib, an epidermal growth factor receptor inhibitor but not an IGF-1R inhibitor, was more effective in MCF-7/HER2-18 cells than in the parental MCF-7 cells and IGF binding protein-3 (IGFBP-3) was more effective against MCF-7 cells compared to MCF-7/HER2-18. MCF-7/HER2-18 cells are known to be resistant to the effects of the estrogen receptor inhibitor, tamoxifen. Interestingly, NDGA not only inhibited the growth of MCF-7/HER2-18 on its own, but it also demonstrated additive growth inhibitory effects when combined with tamoxifen. These studies suggest that NDGA may have therapeutic benefits in HER2-positive, tamoxifen resistant, breast cancers in humans.     

The Drosophila learning and memory gene linotte encodes a putative receptor tyrosine kinase homologous to the human RYK gene product

The linotte mutant was isolated on the basis of its learning and memory deficit. Interestingly, linotte individuals carrying a null mutation are viable, indicating that the linotte gene is not required for vital functions. We show here that the linotte gene encodes a putative receptor tyrosine kinase, homologous to the human protein RYK. These products are unique among receptor tyrosine kinases, since they possess a short extra cellular domain, and a modified intracellular catalytic domain. In particular, the subdomains directly involved in ATP binding and phosphotransfer reaction display remarkable variations. These results suggest that linotte is part of a novel signal transduction cascade involved in learning and memory.     

Intertwined dimeric structure for the SH3 domain of the c-Src tyrosine kinase induced by polyethylene glycol binding

Here we report the first crystal structure of the SH3 domain of the cellular Src tyrosine kinase (c-Src-SH3) domain on its own. In the crystal two molecules of c-Src-SH3 exchange their -RT loops generating an intertwined dimer, in which the two SH3 units, preserving the binding site configuration, are oriented to allow simultaneous binding of two ligand molecules. The dimerization of c-Src-SH3 is induced, both in the crystal and in solution, by the binding of a PEG molecule at the dimer interface, indicating that this type of conformations are energetically close to the native structure. These results have important implications respect to in vivo oligomerization and amyloid aggregation.     

Fas ligand-induced apoptosis is regulated by nitric oxide through the inhibition of fas receptor clustering and the nitrosylation of protein kinase Cepsilon

Apoptosis induced by the death-inducing ligand FasL (CD95L) is a major mechanism of cell death. Trophoblast cells express the Fas receptor yet survive in an environment that is rich in the ligand. We report that basal nitric oxide (NO) production is responsible for the resistance of trophoblasts to FasL-induced apoptosis. In this study we demonstrate that basal NO production resulted in the inhibition of receptor clustering following ligand binding. In addition NO also protected cells through the selective nitrosylation, and inhibition, of protein kinase Cε (PKCε) but not PKCα. In the absence of NO production PKCε interacted with, and phosphorylated, the anti-apoptotic protein cFLIP. The interaction is predominantly with the short form of cFLIP and its phosphorylation reduces its recruitment to the death-inducing signaling complex (DISC) that is formed following binding of a death-inducing ligand to its receptor. Inhibition of cFLIP recruitment to the DISC leads to increased activation of caspase 8 and subsequently to apoptosis. Inhibition of PKCε using siRNA significantly reversed the sensitivity to apoptosis induced by inhibition of NO synthesis suggesting that NO-mediated inhibition of PKCε plays an important role in the regulation of Fas-induced apoptosis.     

Effect of the tyrosine kinase inhibitor lapatinib on CUB-domain containing protein (CDCP1)-mediated breast cancer cell survival and migration

The surface receptor CUB domain-containing protein 1 (CDCP1) is highly expressed in several adenocarcinomas and speculated to participate in anchorage-independent cell survival and cell motility. Tyrosine kinase phosphorylation seems to be crucial for intracellular signaling of CDCP1. Lapatinib, a tyrosine kinase inhibitor (TKI), is approved for treatment of HER-2/neu overexpressing metastatic breast cancer and functions by preventing autophosphorylation following HER-2/neu receptor activation. This study aimed to investigate the effect of CDCP1 expression on anchorage-independent growth and cell motility of breast cancer cells. Moreover, studies were performed to examine if lapatinib provided any beneficial effect on HER-2/neu^(+)/−/CDCP1⁺ breast cancer cell lines. In our studies, we affirmed that CDCP1 prevents cells from undergoing apoptosis when cultured in the absence of cell–substratum anchorage and that migratory and invasive properties of these cells were decreased when CDCP1 was down-regulated. However, only HER-2/neu⁺, but not HER-2/neu^(+)/− cells showed decreased proliferation and invasion and an enhanced level of apoptosis towards loss of anchorage when treated with lapatinib. Therefore, we conclude that CDCP1 might be involved in regulating adhesion and motility of breast cancer cells but that lapatinib has no effect on tyrosine kinases regulating CDCP1. Nonetheless, other TKIs might offer therapeutic approaches for CDCP1-targeted breast cancer therapy and should be studied considering this aspect.     

Acute ischemic cardiac dysfunction is attenuated via gene transfer of a peptide inhibitor of the beta-adrenergic receptor kinase (betaARK1)

Acute myocardial ischemia is a critical adverse effect potentially occurring during cardiac procedures. A peptide inhibitor of the beta-adrenergic receptor kinase (betaARK1), betaARKct, has been successful in rescuing chronic myocardial ischemia. The present study focused on the effects of adenoviral-mediated betaARKct (Adv-betaARKct) delivery on left ventricle (LV) dysfunction induced by acute coronary occlusion. Rabbits received intracoronary delivery of phosphate-buffered saline (PBS) (n=9) or 5x10(11) viral particles of betaARKct (n=8). A loose prolene 5-0 Potz-loop suture was placed around the circumflex coronary artery (LCx) with both ends buried under the skin. Four days later, the suture was retrieved and pulled to occlude the LCx. Ischemia was confirmed by immediate ECG changes. LV function was continuously recorded for 45 min. Contractility (LVdP/dtmax), relaxation (LVdP/dtmin) and end diastolic pressure (EDP) were less impaired in the betaARKct group as compared to PBS (P<0.05, two-way ANOVA). betaAR density was higher in the ischemic area of the LV in the betaARKct group (betaARKct: 71.9+/-4.6 fmol/mg protein, PBS: 54.5+/-4.0 fmol/mg protein, P<0.05). Adenylyl cyclase activity was also improved basally and in response to betaAR stimulation. betaARK1 activation was less in the betaARKct group (P<0.05). Therefore, inhibition of myocardial betaARK1 may represent a new strategy to prevent LV dysfunction induced by acute coronary ischemia.     

Expression of a novel receptor tyrosine kinase gene and a paired-like homeobox gene provides evidence of differences in patterning at the oral and aboral ends of hydra

Axial patterning of the aboral end of the hydra body column was examined using expression data from two genes. One, shin guard, is a novel receptor protein-tyrosine kinase gene expressed in the ectoderm of the peduncle, the end of the body column adjacent to the basal disk. The other gene, manacle, is a paired-like homeobox gene expressed in differentiating basal disk ectoderm. During regeneration of the aboral end, expression of manacle precedes that of shin guard. This result is consistent with a requirement for induction of peduncle tissue by basal disk tissue. Our data contrast with data on regeneration of the oral end. During oral end regeneration, markers for tissue of the tentacles, which lie below the extreme oral end (the hypostome), are detected first. Later, markers for the hypostome itself appear at the regenerating tip, with tentacle markers displaced to the region below. Additional evidence that tissue can form basal disk without passing through a stage as peduncle tissue comes from LiCl-induced formation of patches of ectopic basal disk tissue. While manacle is ectopically expressed during formation of basal disk patches, shin guard is not. The genes examined also provide new information on development of the aboral end in buds. Although adult hydra are radially symmetrical, expression of both genes in the bud's aboral end is initially asymmetrical, appearing first on the side of the bud closest to the parent's basal disk. The asymmetry can be explained by differences in positional information in the body column tissue that evaginates to form a bud. As predicted by this hypothesis, grafts reversing the orientation of evaginating body column tissue also reverse the orientation of asymmetrical gene expression.     

The production of macrophage inflammatory protein-2 induced by soluble intercellular adhesion molecule-1 in mouse astrocytes is mediated by src tyrosine kinases and p42/44 mitogen-activated protein kinase

Severe traumatic brain injury stimulates the release of soluble intercellular adhesion molecule-1 (sICAM-1) into CSF. Studies in cultured mouse astrocytes suggest that sICAM-1 induces the production of macrophage inflammatory protein-2 (MIP-2). In the present study, we investigated the underlying mechanisms for MIP-2 induction. sICAM-1 induced MIP-2 in astrocytes lacking membrane-bound ICAM-1, indicating that its action is due to heterophilic binding to an undescribed receptor rather than homophilic binding to surface ICAM-1. Signal transduction may be mediated by src tyrosine kinases, as the src tyrosine kinase inhibitors herbimycin A and PP2 abolished MIP-2 induction by sICAM-1. Phosphorylation of p42/44 mitogen-activated protein kinase (MAPK), but not of p38 MAPK, occurred further downstream, as evidenced by western blot analysis combined with the use of herbimycin A and specific MAPK inhibitors. By contrast, induction of MIP-2 by tumour necrosis factor-alpha (TNF-alpha) involved both p42/44 MAPK and p38 MAPK. Following stimulation with either sICAM-1 or TNF-alpha, astrocyte supernatants promoted chemotaxis of human neutrophils and incubation of these supernatants with anti-MIP-2 antibodies more efficiently suppressed the migration induced by sICAM-1 than by TNF-alpha. These results show that sICAM-1 induces the production of biologically active MIP-2 in astrocytes by heterophilic binding to an undefined receptor and activation of src tyrosine kinases and p42/44 MAPK.