P2U agonists induce chemotaxis and actin polymerization in human neutrophils and differentiated HL60 cells.

Human neutrophils or HL60 cells express P2U receptors and respond to micromolar concentrations of ATP, adenosine 5'-O-(thiotriphosphate) (ATPgammaS), or UTP with immediate increases in intracellular Ca2+ through activation of phosphoinositide phospholipase C (Cowen, D. S., Lazarus, H. M., Shurin, S. B., Stoll, S. E., and Dubyak, G. R. (1989) J. Clin. Invest. 83, 1651-1660). P2U agonists reportedly induce limited enzyme secretion and enhance the respiratory burst in response to chemotactic factors. We demonstrate here that P2U agonists are chemotactic for neutrophils or differentiated HL60 cells. Rhodamine phalloidin staining indicates that ATPgammaS treatment induces actin polymerization and shape changes similar to those seen when these cells are treated with chemotactic peptide fMet-Leu-Phe. Although undifferentiated HL60 cells fail to mount a rise in Ca2+ when challenged with fMet-Leu-Phe, they increase Ca2+ in response to P2U agonists. However, functional expression of phospholipase C-coupled receptors is not sufficient for chemotaxis since HL60 cell migration in response to these agonists or to fMet-Leu-Phe occurs only after exposure to differentiating agents such as BT2cAMP. In addition to the well known G protein-linked receptors for lipid or peptide chemotactic factors, neutrophils apparently also can utilize G protein-linked purino/pyrimidino receptors to recognize nucleotides as chemoattractants. High concentrations of ATP and UTP generated at sites of platelet aggregation and tissue injury could thus be important mediators of inflammation.


Human neutrophils or HL60 cells express P 2U receptors and respond to micromolar concentrations of ATP, adenosine 5-O-(thiotriphosphate) (ATP␥S), or UTP with immediate increases in intracellular Ca
. P 2U agonists reportedly induce limited enzyme secretion and enhance the respiratory burst in response to chemotactic factors. We demonstrate here that P 2U agonists are chemotactic for neutrophils or differentiated HL60 cells. Rhodamine phalloidin staining indicates that ATP␥S treatment induces actin polymerization and shape changes similar to those seen when these cells are treated with chemotactic peptide fMet-Leu-Phe. Although undifferentiated HL60 cells fail to mount a rise in Ca 2؉ when challenged with fMet-Leu-Phe, they increase Ca 2؉ in response to P 2U agonists. However, functional expression of phospholipase C-coupled receptors is not sufficient for chemotaxis since HL60 cell migration in response to these agonists or to fMet-Leu-Phe occurs only after exposure to differentiating agents such as BT 2 cAMP. In addition to the well known G protein-linked receptors for lipid or peptide chemotactic factors, neutrophils apparently also can utilize G protein-linked purino/pyrimidino receptors to recognize nucleotides as chemoattractants. High concentrations of ATP and UTP generated at sites of platelet aggregation and tissue injury could thus be important mediators of inflammation.
Extracellular ATP and other purine nucleotides bind to P 2 receptors, which are classified pharmacologically into five distinct subgroups of P 2T , P 2U , P 2X , P 2Y , and P 2Z (2)(3)(4). P 2T and P 2Z receptors have a restricted cellular distribution, whereas receptors for the other three subgroups are widely distributed. The physiological role of P 2X and P 2Y receptors has been studied extensively in nervous and cardiovascular systems. The typical order of potency for P 2X receptors is AMP-CPP 1 Ͼ ATP ϭ 2-methylthio-ATP, whereas that for P 2Y receptors is 2-methylthio-ATP Ͼ ADP␤S Ͼ ATP Ͼ AMP-PCP. Interestingly, P 2U receptors recognize not only purine nucleotides but also UTP as demonstrated recently by expression of cloned murine and human P 2U cDNA in human leukemia or astrocytoma cell lines (4,5). Receptors of this subclass are present on many cells including those of the immune system and the vasculature, but not much is known about their physiological functions (6). Recent studies with isolated cell systems demonstrate that P 2 receptors of the U, X, and Y subclasses are members of the G protein-linked superfamily and are coupled to various effectors including adenylate cyclase or phospholipases (1,(7)(8)(9)(10)(11).
Human neutrophils or promyelocytic HL60 cells in various stages of differentiation respond to ATP, ATP␥S, or UTP with an immediate increase in intracellular Ca 2ϩ . This response is mediated through the activation of a pertussis toxin-sensitive G protein, which in turn activates the phosphoinositide phospholipase C (PI-PLC) cascade (1,8,10,11). The magnitude and the time course of these two early biochemical responses are very similar to those observed with the chemotactic factor fMet-Leu-Phe (12). In contrast to fMet-Leu-Phe, P 2U agonists induce only a limited secretory response and are poor activators of O 2 . formation in neutrophils or differentiated HL60 cells and their effects on chemotaxis are unknown (8,13). Since high concentrations of ATP and UTP are present at sites of platelet aggregation (2) and neutrophils are the initial cell type found at sites of tissue injury, we wanted to determine whether these nucleotides could be important for the recruitment of neutrophils in the absence of other known peptide and lipid chemoattractants. Our results demonstrate that activation of P 2U receptors on human peripheral blood neutrophils induces a typical chemotactic response, accompanied by actin polymerization and changes in cell shape characteristic of those seen during chemotaxis in response to fMet-Leu-Phe. By comparing the effects of fMet-Leu-Phe and P 2U receptor agonists on differentiating HL60 cells, we show that Ca 2ϩ mobilization is an insufficient signal for chemotaxis, since undifferentiated HL60 cells fail to migrate in response to ATP␥S in spite of a robust Ca 2ϩ response.

EXPERIMENTAL PROCEDURES
Cells-Human peripheral neutrophils were isolated from fresh venous blood from healthy volunteers as described previously (12). HL60 cells were cultured in supplemented RPMI 1640 medium (Life Technologies, Inc.) under normal culture conditions and differentiated with 0.5 mM db-cAMP for 1-4 days as reported (12).
Ca 2ϩ Mobilization-Neutrophils or HL60 cells were labeled with the fluorescent Ca 2ϩ indicator Fura-2/AM (Molecular Probes, Eugene, OR), and changes in fluorescence were monitored in a Perkin-Elmer LS-5B spectrofluorimeter as described (12). For desensitization studies, the second agonist was added 5 min after addition of the first stimulus and fluorescence was monitored for an additional 3-5 min.
Chemotaxis-Neutrophils or HL60 cells were suspended at 2 ϫ 10 6 cells/ml of Hepes-buffered Hanks' balanced salt solution, and chemotaxis was assayed in modified bipartite Boyden chemotaxis chambers as described previously (12). Briefly, P 2U agonists or the synthetic peptide * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
fMet-Leu-Phe was added to the bottom compartment, and cells were placed in the upper compartment and allowed to migrate through a polycarbonate filter at 37°C for 90 min. In some experiments, agonists also were present in the upper compartment to distinguish directed from random migration (14). Filters were fixed and stained, and cells that had migrated through the filter were counted under ϫ 100 magnification in 10 different fields of each filter.
Fluorescence Staining of F-actin-Neutrophils or differentiated HL60 cells were prepared as for chemotaxis but exposed to agonists in polypropylene tubes for 5 min in a shaking water bath kept at 37°C. The cells were then fixed in the presence of 3.7% formaldehyde, permeabilized with 0.1 mg/ml of lysophosphatidylcholine, and stained with 0.16 M rhodamine-labeled phalloidin for 15 min at 37°C (15). Cytospin preparations from these cells were viewed under oil in a Leitz Diaplan microscope (ϫ 63) equipped with a mercury lamp and photographed with a Wild camera.

Pharmacological Characterization of P2 Receptors on Hu-
man Neutrophils-Fluorescent probes such as Fura-2 provide a convenient assay system to monitor changes in intracellular Ca 2ϩ and to characterize surface receptors that are coupled to the PI-PLC pathway (12). As reported previously (1), the tracings in Fig. 1 demonstrate the presence of P 2U receptors on human neutrophils, since ATP␥S, a nonhydrolyzable analog of ATP (Fig. 1A) and UTP (Fig. 1B) caused an immediate, concentration-dependent rise in intracellular Ca 2ϩ . In contrast, the most potent agonists for P 2Y receptors (ADP␤S, Fig. 1C) or P 2X receptors (AMP-CPP; Fig. 1D) failed to elicit any changes in intracellular Ca 2ϩ in human neutrophils. Similarly, weaker agonists for these receptors like AMP-PNP (Fig. 1C) or AMP-PCP (Fig. 1D), also were inactive in this assay, further supporting the absence of P 2Y or P 2X receptor-mediated activation of neutrophils. Desensitization studies showed that neither of these inactive nucleotides interfered with subsequent effects of ATP␥S. In contrast, ATP␥S (Fig. 1E) desensitized neutrophils against a second challenge with the same agonist (trace 1) or UTP (trace 2). The same results were obtained when UTP was used as the first stimulus (Fig. 1F), indicating that ATP␥S and UTP shared the same receptors. Ca 2ϩ mobilization in response to the chemotactic peptide fMet-Leu-Phe (Fig. 1G, second ar-row) remained intact in cells that were first exposed to UTP or ATP␥S (trace 1 or 2, respectively). Fig. 1H shows Fura-2 tracings for sequential treatment with fMet-Leu-Phe to demonstrate that the receptor for this chemoattractant could be desensitized by its own agonist.
ATP␥S Is Chemotactic for Human Neutrophils-Because ATP␥S elicited such a robust increase in intracellular Ca 2ϩ but failed to stimulate O 2 . production (13), we wanted to determine whether human neutrophils recognized P 2U agonists as chemotactic factors. To avoid degradation of ATP during the long incubation periods required in chemotaxis assays, most experiments were carried out using ATP␥S. The data in Fig. 2 (solid bars) demonstrate that ATP␥S induced chemotaxis at concentrations ranging from 10 to 100 M. The chemotactic response to 100 M ATP␥S (58 Ϯ 16 cells/field, n ϭ 5), the highest concentration tested, was approximately 2 ⁄3 of that observed with 1 nM fMet-Leu-Phe (87 Ϯ 12 cells/field, n ϭ 9). That the observed migration in response to ATP␥S was indeed directional chemotaxis and not simply increased random motility was demonstrated by the diminished migration seen when the concentration gradient was reduced by adding ATP␥S to the upper compartment of the chemotaxis chambers (hatched bars). For comparison, chemotaxis in response to 1 or 10 nM fMet-Leu-Phe was reduced by 65 and 37%, respectively, when equal concentrations of the peptide were present in the top compartment of the chambers (data not shown).
Relationship between Ca 2ϩ Mobilization and Chemotaxis of P 2U Agonists in Human Neutrophils-To determine whether increases in intracellular Ca 2ϩ were predictive of chemotactic activity, we compared the effects of a series of purinergic agonists and UTP on these two assays in aliquots from the same neutrophil preparations. The Fura-2 data were averaged from the peak excursions (distance in millimeters) of the fluorescence signal from the base line observed immediately after the addition of stimulus. Chemotaxis data were standardized relative to the responses observed using 1 nM fMet-Leu-Phe. As shown in Fig. 3, UTP was as effective as ATP␥S in eliciting chemotaxis (bottom panel) in human neutrophils, indicating that P 2U receptors were functionally linked not only to Ca 2ϩ mobilization (top panel) but also to an important physiological response of neutrophils. Although ATP and UTP raised intracellular Ca 2ϩ levels more than ATP␥S, migration in response to  1, 2, or 3, respectively). Panels C and D, desensitization to ATP␥S (second arrow) following the initial addition (first arrow) of ADP␤S or AMP-PNP (C, trace 1 or 2, respectively); AMP-CPP or AMP-PCP (D, trace 1 or 2, respectively). Panels E and F, desensitization to ATP␥S or UTP (second arrow for trace 1 or 2, respectively) following the initial addition (first arrow) of ATP␥S (E) or UTP (F). Panels G and H, desensitization to fMet-Leu-Phe (second arrow) following the initial addition of UTP or ATP␥S (G, trace 1 or 2, respectively) or fMet-Leu-Phe (H). Nucleotides were used at 100 M and fMet-Leu-Phe at 100 nM concentration. ATP was only about 50% of that seen with these two P 2U agonists. Chemotaxis in response to AMP-PNP reached only 10% of that stimulated by 1 nM fMet-Leu-Phe, and small Fura-2 responses to this agonist were observed in some but not all donors. Interestingly, ADP␤S elicited a significant chemotactic response, although this agent did not elevate Ca 2ϩ perceptibly in the same neutrophil preparations. Concentrations below 100 M for the latter P 2Y agonists were completely ineffective in this assay (data not shown). Adenosine did not elevate intracellular Ca 2ϩ or induce chemotaxis at the concentrations tested (10 -100 M), and the same was true for AMP (data not shown). Although P 2X agonists were not tested in the chemotaxis assay, they did not induce any changes in cell shape normally seen with chemotactic factors (see below). P 2U Receptors on HL60 Cells Induce Ca 2ϩ Mobilization, but Differentiation Is Required for Stimulation of Chemotaxis-Undifferentiated HL60 myeloid precursor cells are known to express P 2U receptors that stimulate PI-PLC and Ca 2ϩ mobilization (1). In contrast, fMet-Leu-Phe receptor-mediated changes in intracellular Ca 2ϩ and functional responses such as chemotaxis, enzyme secretion, and O 2 . production were detectable only after HL60 cells were forced to differentiate toward a neutrophil phenotype with db-cAMP (8,16). To determine whether coupling of P 2U receptors to PI-PLC was sufficient to transmit a chemotactic signal, we followed chemotaxis and Ca 2ϩ mobilization in response to ATP␥S or fMet-Leu-Phe during HL60 differentiation. The data in Fig. 4 exemplify the changes in Fura-2 fluorescence in response to ATP␥S (panel A) or fMet-Leu-Phe (panel B) at daily intervals after differentiating HL60 cells with 0.5 mM db-cAMP. Chemotaxis data obtained from the same set of cells (Fig. 5) clearly demonstrated that Ca 2ϩ mobilization per se was insufficient for chemotaxis, since ATP␥S did not induce this response in undifferentiated cells (0 h). Significant migration in response to either ATP␥S or fMet-Leu-Phe was detectable by 24 h and reached a maximum 48 h after the addition of db-cAMP. Results using plateletactivating factor as a chemoattractant followed essentially the same time course as those using fMet-Leu-Phe (data not shown).

ATP␥S Induces Actin Polymerization and Shape Change in Human Neutrophils and Differentiated HL60 Cells-Neutro-
phils readily change their shape from a round, resting form to a triangular shape containing a broad leading edge and a narrower trailing edge when exposed to fMet-Leu-Phe in suspension (14). These changes are accompanied by increased actin polymerization near the cell periphery as visualized with rhodamine-labeled phalloidin (15). Such changes from round (Fig. 6) to irregular shapes as well as increases in phalloidin staining also were apparent when neutrophils or db-cAMPdifferentiated HL60 cells were exposed to ATP␥S. These morphological changes were similar to those observed when the cells were treated with fMet-Leu-Phe, indicating that ATP␥S could function as a typical chemotactic factor. DISCUSSION Our observations that ATP␥S and UTP caused a complete cross-desensitization of the Ca 2ϩ mobilization response of human neutrophils support recent reports that P 2U receptors recognize both of these agonists (4,5). Only the P 2U sublass of P 2 receptors appeared to be coupled to Ca 2ϩ mobilization, since potent agonists for P 2Y receptors (ADP␤S) or P 2X receptors (AMP-CPP) were ineffective by themselves and did not affect subsequent changes in intracellular Ca 2ϩ by ATP␥S, UTP, or ATP (data shown only for ATP␥S). Similar P 2U receptors also were identified on myeloid progenitor cells but not on mature lymphocytes or lymphocytic leukemia cells (1).
In spite of the robust Ca 2ϩ mobilization elicited by P 2U agonists, the functional significance of P 2U receptors for neutrophil activation was not obvious. Occupancy of P 2U receptors caused only a partial secretory response and did not enhance the respiratory burst unless fMet-Leu-Phe was also added (Ref. 13; our data not shown). Our data demonstrate that high micromolar concentrations of P 2U agonists elicited a chemotactic response in neutrophils or differentiated HL60 cells that approached that observed with the classical chemoattractant fMet-Leu-Phe. Similarly high concentrations of agonists were typically required for activation of P 2 receptors in other systems, including the vasculature, neurotransmission, and cardiac function (2). Although most currently known peptide or lipid chemoattractants are active at nanomolar concentrations, ATP and UTP are present in cells at millimolar concentrations. Leakage of these nucleotides from damaged cells or degranulating platelets could lead to sufficiently high accumulations to be physiologically relevant.
Classical chemotactic factors, including lipids (leukotriene B 4 , platelet-activating factor), synthetic peptides (fMet-Leu-Phe), the complement fragmentation products (C5a), and cytokines (interleukin-8) are coupled to the PI-PLC pathway to mobilize intracellular Ca 2ϩ (12,17,18). In the present studies, the correlation between Ca 2ϩ mobilization and chemotaxis breaks down for ADP␤S, a potent P 2Y agonist, where chemotaxis of neutrophils is observed in the complete absence of Ca 2ϩ mobilization. Whether this is an example where chemotaxis is dissociated from a G protein/PI-PLC/Ca 2ϩ pathway, as recently described for transforming growth factor ␤, cannot be ascertained from our experiments (15). Since P 2Y receptors typically stimulate the formation of Ca 2ϩ -mobilizing phosphoinositides in other cells, we speculate that migration in response to ADP␤S more likely results from conversion of ADP␤S to another agonist such as ATP␥S during the prolonged incubation necessary for chemotaxis assays. Similarly, we attribute the decreased neutrophil chemotaxis in response to ATP as compared with ATP␥S to the instability of ATP during the chemotaxis assay. These two purine triphosphates were equipotent in the Ca 2ϩ mobilization assay, while UTP was more potent than either of these. However, the absence of PI-PLC coupling does not completely rule out the possibility that human neutrophils express other P2 receptors.
Interestingly, myeloid progenitor cells such as undifferentiated HL60 cells express P 2U receptors that appeared to be fully coupled to the PI-PLC pathway and subsequent Ca 2ϩ mobilization (1). In spite of this, undifferentiated HL60 cells failed to migrate in response to ATP␥S, indicating that Ca 2ϩ mobilization was not sufficient for chemotaxis. When HL60 cells were FIG. 6. Fluorescent micrographs of human neutrophils or differentiated HL60 cells in the absence or presence of ATP␥S or fMet-Leu-Phe. Neutrophils (PMN, left panels) or db-cAMP-differentiated HL60 cells (right panels) were treated in suspension with buffer, 100 M ATP␥S, or 1 M fMet-Leu-Phe (panels A, B, or C, respectively) for 10 min, fixed, permeabilized, and stained with rhodamine-phalloidin as described under "Experimental Procedures." induced to differentiate toward a neutrophil-like phenotype, they began to express receptors for fMet-Leu-Phe and other chemoattractants (platelet-activating factor), and chemotaxis in response to fMet-Leu-Phe, ATP␥S, or platelet-activating factor was detectable 24 h after adding db-cAMP as a differentiating agent. As these cells acquired the proficiency to migrate in response to fMet-Leu-Phe, they apparently also gained the capability of coupling their P 2U receptors to a migratory response. Interestingly, the functional coupling of phospholipase D to P 2U or fMet-Leu-Phe receptors during HL60 differentiation followed a time course that is very similar to the one we observed for the appearance of chemotaxis mediated by these receptors (11). Phospholipase D may thus provide the amplification signal necessary for the expression of fully functional receptors, since Ca 2ϩ mobilization was obviously insufficient to generate a chemotactic response.
Directed or random migration involves the rearrangement of the cytoskeleton, and changes in actin polymerization are typically observed for all chemoattractants described so far (14,15). In our studies, ATP␥S induced changes in cell shape in neutrophils or differentiated HL60 cells that were very similar to those observed with the classical chemotactic factor fMet-Leu-Phe. Moreover, patterns of actin polymerization as visualized with rhodamine-labeled phalloidin staining confirmed the similarities between fMet-Leu-Phe and ATP␥S elicited changes. Interestingly, G protein-linked receptors for cAMP are coupled to the PI-PLC pathway in Dictyostelium where extracellular cAMP induces changes in cell shape and chemotaxis (14). In contrast to Dictyostelium, cAMP does not activate the PI-PLC pathway in neutrophils and can actually inhibit chemotaxis of these cells (19).
To our knowledge, our results provide the first evidence that neutrophils can migrate in response to increasing concentration gradients of nucleotides such as P 2U agonists. Recent reports indicate that human P 2U receptors share more sequence homology with G protein-coupled receptors for peptides than for cAMP or adenosine (5). G protein-linked receptors that recognize such diverse signals as P 2U agonists, lipid chemoattractants, or peptide chemoattractants thus appear to utilize similar biochemical pathways leading to neutrophil chemotaxis.