Direct observation and quantification of macrophage chemoattraction to the growth factor CSF-1

Sarah E. Webb 1 Jeffrey W. Pollard 0 Gareth E. Jones 1 0 Departments of Developmental & Molecular Biology and Obstetrics & Gynecology, Albert Einstein College of Medicine , Bronx, NY 10461, USA 1 The Randall Institute, King's College London , 26-29 Drury Lane, London WC2B 5RL, UK *Author for correspondence SUMMARY The cloned mouse macrophage cell line, BAC1.25F, resembles primary macrophages in its dependence on colony stimulating factor-1 (CSF-1) for both viability and proliferation. Re-addition of CSF-1 to cytokine-deprived cells, which are rounded with diffusely organised F-actin, stimulates rapid cell spreading and cell polarisation. Using the Dunn chemotaxis chamber the movement of stimulated macrophages was monitored over a 2 hour period. Cells restimulated with 1.32 nM human recombinant CSF-1 migrated at a mean rate of 7.71 m m per hour, but showed no directional preferences. In a linear concentration gradient of CSF-1, cytokine-deprived cells were again stimulated to migrate and the mean rate of cell motility, at 6.88 m m per hour, was not significantly different from that measured in an isotropic environment of CSF-1. However, there was a strong preference for the cells to orientate so When tissue is injured or becomes locally infected, an inflammatory reaction ensues, leading to increased migration of leucocytes into the affected area. Movement of cells into sites of inflammation depends initially upon the interactions between leucocytes and endothelia (Springer, 1994), and is then largely directed by the process of chemotaxis (Downey, 1994). Mediated by the co-ordinated activation of the cytoskeleton and associated adhesive integrins (Bienvenu et al., 1992), leucocytes will migrate within an extracellular matrix towards a high point of a concentration gradient of diffusible chemoattractant. Leucocytes are extremely sensitive to variations in the concentrations of chemotactic factors, and may detect a difference of as little as 0.1% across a cell diameter (Zigmond et al., 1981). A variety of products present at the site of inflammation can act as chemoattractants including bacterial products such as formylmethionyl peptides, C5a, platelet activating factor and leukotriene B4 (see Murphy, 1994, for review). The cytokine, colony-stimulating factor-1 (CSF-1, also known as M-CSF) is a sialoglycoprotein normally produced by endothelia which also affects cell motility. CSF-1 has been shown to stimulate monocyte migration (Wang et al., 1988) and to act that their long axes aligned with the CSF-1 gradient and they migrated preferentially towards the source of CSF-1. Migrating cells contained abundant F-actin within the leading lamellae as judged by confocal imaging of fluorescent phalloidin, but the actin was not arranged into stress fibre-like structures. These data support the proposition that CSF-1 is both a chemokinetic and chemotactic agent for macrophages. Tumour necrosis factor (TNF-a ) failed to stimulate cell migration and thus was neither chemokinetic nor a chemotactic agent. However, cells exposed to a dual concentration gradient of both TNF-a and CSF-1 did migrate successfully, although the chemotactic response to CSF-1 was abolished. as a chemoattractive agent for receptor-expressing myeloid progenitor cells (Pierce et al., 1990). CSF-1 is best known as a regulator of the proliferation, differentiation and survival of cells of the mononuclear phagocyte lineage (Stanley and Heard, 1977). The effects of CSF-1 are mediated through its binding to a single class of high affinity cell surface transmembrane tyrosine kinase receptor expressed by mature monocytes and macrophages which is encoded by the c-fms proto-oncogene (Sherr, 1990). CSF-1 also plays a role outside haemopoiesis where it has been shown to regulate the invasion and proliferation of mammary epithelial cells in culture (Sapi et al., 1995; Filderman et al., 1992). It thus seems possible that locally produced CSF-1 at sites of inflammation or tumorogenesis may play a significant role in the selective recruitment of cells (Scholl et al., 1993; Tang et al., 1992). Another cytokine which is produced largely by activated macrophages (Chen and Mueller, 1990) is tumour necrosis factor, TNF-a . This is also associated with many of the tissue responses to inflammation and injury; it stimulates neutrophil release from the bone marrow and activates these cells, which induces their adhesion to endothelial cell surfaces and enhances their phagocytic activity (Gamble et al., 1985). In addition, TNF-a has been demonstrated to stimulate a chemotactic response in both monocytes and neutrophils (Wang et al., 1987), though its effect on neutrophils has been disputed recently with the report that normal human neutrophils, which undergo chemotaxis to f-Met-Leu-Phe, do not respond to TNFa and this cytokine actually inhibits neutrophil chemotaxis to f-Met-Leu-Phe (Vollmer et al., 1992). Several assays exist for measuring chemotaxis in vitro (reviewed by Wilkinson et al., 1982), and for differentiating between this stimulated directional migration and either simple random migration, which occurs in the absence of any known chemical stimulant and is non-directional, and chemokinesis, whereby the rate of migration is stimulated but the direction of migration is random (Wilkinson et al., 1984). These assays include those, such as the Boyden chamber, which analyse the migratory behaviour of a large population of cells in the presence or absence of potential locomotory modifiers. After an appropriate length of incubation the experiment is terminated, for example by fixing the cells, and changes in the distribution of cells with respect to the putative chemoattractant are determined. Unfortunately, however, it is impossible to view the cells whilst they are responding and so cell behaviour and evaluation of chemotaxis can only be deduced from the final distribution of the cells. The results obtained are therefore less convincing than visual assays which involve observing the locomotion of a small number of cells directly, for example by time-lapse photography, and there is considerable uncertainty over the assignment of chemotaxis to results based upon these forms of experiments (Wilkinson, 1990a). The chemotaxis chamber devised by Zigmond (1977) represented a considerable advance in overcoming these problems, but even this device has limitations, particularly in regard to quantifying and maintaining a linear gradient of putative chemoattractant over prolonged time intervals, limiting its usefulness for slowly migrating cells. In order to test whether CSF-1 was truly a candidate chemoattractant for macrophages, we therefore took advantage of a newly developed direct-viewing chamber (Zicha et al., 1991) which overcomes the deficits inherent in both the Boyden chamber and the Zigmond chamber. Using this novel apparatus, we report here for the first time the motile responses elicited by activation of a receptor tyrosine kin (...truncated)