William Allen

Distinct PI(3)Ks mediate mitogenic signalling and cell migration in macrophages

lass IA phosphoinositide-3-kinases (PI(3)Ks) have been implicated in many cellular responses downstream of tyrosine kinases. Such cellular responses include protection from apoptosis, proliferation, cytoskeletal rearrangements and cell migration. These PI(3)Ks are heterodimers that consist of a catalytic subunit of relative molecular mass 110,000 (Mr 110K) (p110 α, β or δ) in complex with an adaptor molecule (p85α, p85β, p55γ or their splice variants) that contains Src-homology-2 (SH2) domains. Whereas p110α and p110β are ubiquitously expressed, p110δ has a restricted tissue distribution and is abundant in leukocytes. Approaches that have been used to implicate PI(3)K activity in responses to signalling through tyrosine-kinase-type receptors do not distinguish between the p110 isoforms and it is unknown whether the isoforms are functionally redundant or have distinct signalling roles. Here, by introducing antibodies specific to p110α, β or δ into macrophages, we identify specific PI(3)K isoforms as key regulators of the different PI(3)K-dependent responses downstream of the tyrosine-kinase receptor for colony-stimulating factor-1 (CSF-1) (Fig. 1a). CSF1-induced DNA synthesis is inhibited by antibodies to p110α but not to p110β or p110δ. CSF1-induced actin reorganization and migration are unaffected by anti-p110α antibodies but inhibited by antibodies against p110β or p110δ. We used western blotting and immunoprecipitation techniques to show that antibodies raised against p110α, β or δ peptides were specific to each isoform (see Methods). Antibodies against the carboxy termini of p110s selectively neutralized the lipid-kinase activity of the specific p110 in vitro, whereas neither preimmune IgG nor an antibody against the Ras-binding domain of p110δ (δRBD antibody) blocked lipid-kinase activity of any p110 PI(3)K (Fig. 1b). The murine macrophage cell line Bac1.2F5 depends on CSF-1 for survival and proliferation, and CSF-1 also stimulates actin reorganization and cell motility. Bac1.2F5 cells express p110α, β and δ, which all become recruited to the activated CSF-1 receptor (data not shown). The PI(3)K inhibitor LY294002, which inhibits all p110 isoforms to the same extent, inhibited CSF1-stimulated DNA synthesis with a half-maximal inhibitory concentration similar to that reported for class IA PI(3)Ks in other cells (that is, 5–10 μM; ref. 5). LY294002 also blocked CSF1-induced actin reorganization and cell migration (data not shown). We tested the effect on CSF1-stimulated DNA synthesis of microinjecting p110-specific antibodies; only antibodies to p110α were inhibitory (Fig. 1). We used two independently derived antibodies to p110α (Ab1 and Ab2; see Methods) in these experiments. Ab1 reduced DNA synthesis to the level seen in CSF1-starved cells. Preblocking of this antibody with the peptide against which it was raised relieved this inhibition (Table 1). Ab2 was more potent than Ab1 in inhibiting DNA synthesis (Table 11) and also appeared to reduce cell survival over the 24-h incubation time (data not shown). We next investigated the effect of anti-p110 antibodies on CSF1induced changes to the actin cytoskeleton. None of these antibodies affected the actin cytoskeleton in cells starved of CSF-1 (Fig. 1c) CSF1-stimulated cytoskeletal changes were not altered by preimmune (data not shown) or anti-p110α (Fig. 1c) antibodies. In contrast, antibodies against p110β reduced lamellipodium formation and membrane ruffling. Anti-p110δ antibodies completely inhibited CSF1-induced actin changes, although filopodia were observed on some cells (Fig. 1c; similar results were obtained with δRBD antibody). Preblocking of the p110β and δ antibodies with cognate peptide before injection prevented their effects on cell morphology (data not shown). Given the distinct effect of each anti-p110 antibody on CSF1induced actin-cytoskeletal reorganization, we tested their impact on cell migration. We exposed cells injected with antibodies to a linear concentration gradient of CSF-1 in a Dunn chemotaxis chamber, and filmed them by time-lapse microscopy. Cells injected with anti-p110α Ab1 (data not shown) or Ab2 (Fig. 1d) migrated towards the source of CSF-1 and moved at a similar speed to control cells injected with preimmune antibodies. Cells injected with anti-p110β antibodies showed a small but significant reduction in migration speed, frequently changed their direction of locomotion and failed to migrate towards the source of CSF-1 (Fig. 1d). Cells injected with antibodies to p110δ were most severely blocked in their migration, almost resembling CSF1-starved cells (Fig. 1d). Again, preblocking of the antibodies against p110β or p110δ with their cognate peptide fully reversed the inhibition (data not shown). These results indicate that p110α, but not p110β or p110δ, is necessary for CSF1-induced mitogenic signalling in Bac1.2F5 cells. Conversely, p110β and p110δ but not p110α are required for CSF1mediated regulation of the actin cytoskeleton and cell migration. Microinjection of antibodies to p110a has been shown to block C

Chemotaxis of macrophages is abolished in the Wiskott-Aldrich syndrome

Wiskott‐Aldrich syndrome (WAS) is a rare disease characterized by microthrombocytopenia, eczema and immune deficiency. In this study a direct‐viewing chemotaxis chamber was used to analyse chemotactic responses of WAS neutrophils and macrophages in stable linear concentration gradients. In five patients with classic WAS, chemotaxis of macrophages but not of neutrophils was found to be abolished, whereas the speed of random motility of both cell types was found to be indistinguishable from control cells. This supports the existence of an essential functional link, previously suggested by biochemical studies, between Cdc42, WAS protein (WASp) and the actin cytoskeleton in primary human macrophages. Moreover, these data suggest that Cdc42‐WASp‐mediated filopodial extension is a requirement for chemotaxis but not for chemokinesis in these cells. Abnormal directional cell motility of macrophages and related antigen‐presenting cells may play a significant part in the immune deficiency and eczema of WAS.

The Rho GTPases in Macrophage Motility and Chemotaxis

The GTP-binding proteins, Rho, Rac and Cdc42 are known to regulate actin organisation. Rho induces the assembly of contractile actin-based microfilaments such as stress fibres, Rac regulates the formation of membrane ruffles and lamellipodia, and Cdc42 activation is necessary for the formation of filopodia. In addition, all three proteins can also regulate the assembly of integrin-containing focal adhesion complexes. The orchestration of these distinct cytoskeletal changes is thought to form the basis of the coordination of cell motility and we have investigated the roles of Rho family proteins in migration using a model system. We have found that in the macrophage cell line Bac1, the cytokine CSF-1 rapidly induces actin reorganisation: it stimulates the formation of filopodia, lamellipodia and membrane ruffles, as well as the appearance of fine actin cables within the cell. We have shown that Cdc42, Rac and Rho regulate the CSF-1 induced formation of these distinct actin filament-based structures. Using a cell tracking procedure we found that both Rho and Rac were required for CSF-1 stimulated cell translocation. In contrast, inhibition of Cdc42 does not prevent macrophages migrating in response to CSF-1, but does prevent recognition of a CSF-1 concentration gradient, so that cells now migrate randomly rather than up the gradient of this chemotactic cytokine. This implies that Cdc42, and thus probably filopodia, are required for gradient sensing and cell polarisation in macrophages.

Glucose-Potentiated Chemotaxis in Human Vascular Smooth Muscle Is Dependent on Cross-Talk Between the PI3K and MAPK Signaling Pathways

Atheroma formation involves the movement of vascular smooth muscle cells (VSMC) into the subendothelial space. The aim of this study was to determine the involvement of PI3K and MAPK pathways and the importance of cross-talk between these pathways, in glucose-potentiated VSMC chemotaxis to serum factors. VSMC chemotaxis occurred in a serum gradient in 25 mmol/L glucose (but not in 5 mmol/L glucose) in association with increased phosphorylation (activation) of Akt and ERK1/2 in PI3K and MAPK pathways, respectively. Inhibitors of these pathways blocked chemotaxis, as did an mTOR inhibitor. VSMC expressed all class IA PI3K isoforms, but microinjection experiments demonstrated that only the p110&bgr; isoform was involved in chemotaxis. ERK1/2 phosphorylation was reduced not only by MAPK pathway inhibitors but also by PI3K and mTOR inhibitors; when PI3K was inhibited, ERK phosphorylation could be induced by microinjected activated Akt, indicating important cross-talk between the PI3K and ERK1/2 pathways. Glucose-potentiated phosphorylation of molecules in the p38 and JNK MAPK pathways inhibited these pathways but did not affect chemotaxis. The statin, mevinolin, blocked chemotaxis through its effects on the MAPK pathway. Mevinolin-inhibited chemotaxis was restored by farnesylpyrophosphate but not by geranylgeranylpyrophosphate; in the absence of mevinolin, inhibition of farnesyltransferase reduced ERK phosphorylation and blocked chemotaxis, indicating a role for the Ras family of GTPases (MAPK pathway) under these conditions. In conclusion, glucose sensitizes VSMC to serum, inducing chemotaxis via pathways involving p110&bgr;-PI3K, Akt, mTOR, and ERK1/2 MAPK. Cross-talk between the PI3K and MAPK pathways is necessary for VSMC chemotaxis under these conditions.