Rat mucosal mast cells express P2 purinoceptors, occupation of which mobilizes cytosolic Ca2+ and activates a potassium conductance. The primary function of this P2 system in mast cell biology remains unknown. Here, we show that extracellular ADP causes morphological changes in rat bone marrow-cultured mast cells (BMMC) typical of those occurring in cells stimulated by chemotaxins, and that the nucleotides ADP, ATP, and UTP are effective chemoattractants for rat BMMC. ADP was also a chemotaxin for murine J774 monocytes. The nucleotide selectivity and pertussis toxin sensitivity of the rat BMMC migratory response suggest the involvement of P2U receptors. Poorly hydrolyzable derivatives of ADP and ATP were effective chemotaxins, obviating a role for adenosine receptors. Buffering of external Ca2+ at 100 nM or reduction of the electrical gradient driving Ca2+ entry (by elevating external K+) blocked ADP-driven chemotaxis, suggesting a role for Ca2+ influx in this process. Anaphylatoxin C5a was a potent chemotaxin (EC50 ≈0.5 nM) for J774 monocytes, but it was inactive on rat BMMC in the presence or absence of laminin. Ca2+ removal or elevated [K+] had modest effects on C5a-driven chemotaxis of J774 cells, implicating markedly different requirements for Ca2+ signaling in C5a- vs ADP-mediated chemotaxis. This is supported by the observation that depletion of Ca2+ stores with thapsigargin completely blocked migration induced by ADP but not C5a. These findings suggest that adenine nucleotides liberated from parasite-infested tissue could participate in the recruitment of mast cells by intestinal mucosa.
In humans, mast cell populations expand dramatically in tissues affected by various inflammatory syndromes, such as scleroderma (1) and interstitial cystitis (2). During helminth infections of rodents, the mast cell population of intestinal mucosa increases by 10- to 50-fold (3, 4). Such accumulation may depend upon influx of immature mast cells from blood, as well as the division of preexisting mucosal mast cells. The physical proximity of mucosal mast cells and enteric neurons in rat intestines (5) could also depend upon the migration of mast cells toward neuronal secretory products, among which are adenine nucleotides. Recently, several mast cell chemoattractants (chemotaxins) have been identified, but of these, only Ag and TGF-β1 are known to direct migration of rat mast cells (6, 7). Here, we sought to determine whether adenine nucleotides stimulate directed migration of rat mast cells.
The physiological effects of extracellular adenine nucleotides are mediated by P1 and P2 purinoceptors (8). P1 receptors bind adenosine, which is generated from adenine nucleotides by cell surface ectonucleotidases. P2 receptors bind purine and pyrimidine nucleotides, and these receptors are subdivided according to agonist selectivity and mechanisms of signal transduction (8, 9, 10). Two major functional groups of P2 receptors are the P2X receptors, which form ligand-gated ion channels and pores, and the G protein-coupled P2U and P2Y receptors. Mast cells express adenosine receptors (11, 12), the P2X7 receptor (formerly known as P2Z) for ATP4− (13), and P2Y and/or P2U receptors for ADP and ATP (14, 15, 16). Although adenine nucleotides modulate chemotactic responses to other agonists (e.g., see Ref. 17), they are not generally recognized as chemotaxins per se. However, it was recently shown that P2U-selective agonists elicit chemotaxis of human neutrophils (18), and ATP, but not ADP, stimulates eosinophil chemotaxis (19).
In vitro studies of mast cell chemotaxis have focused on cultured murine and human mast cells and cell lines, with two exceptions (6, 7). They have shown that certain angiogenic factors (20), chemokines (21, 22), complement components C1q, C3a, and C5a (23, 24, 25), as well as hematopoietic growth factors (22, 26, 27, 28) bear chemotactic and/or chemokinetic activity for mast cells. Cultured murine mast cells also undergo haptotaxis toward laminin (29), a protein which is routinely added to chemotaxis assays to enhance adhesion and migration of mast cells. Virtually nothing is known regarding the downstream signals that drive mast cell chemotaxis.
Rat mucosal mast cells, including rat basophilic leukemia (RBL)4 cells and rat bone marrow-cultured mast cells (BMMC), express a P2 receptor-activated signal transduction pathway leading to Ca2+ release (15) and activation of a Gi-linked K+ conductance (14, 16). Occupation of this P2 receptor dramatically potentiates Ag-driven secretion of granule components from rat BMMC (14), and it also promotes Ag-independent release of cellular ATP (15). The physiological role of these effects remains unresolved, and the essential function of this purinoceptor-activated pathway is unclear. The potent chemoattractant C5a as well as extracellular adenine nucleotides initiate Ca2+ release (30, 31) and activate the same Gi-linked K+ conductance in murine J774 monocytes (30). Thus, it is possible that ADP and C5a share chemotactic activity for J774 cells, and that ADP is a chemotaxin for rat mast cells.
Here we show that ADP, ATP, and UTP, acting at low micromolar concentrations, cause directed migration of rat BMMC. The magnitude of response is comparable to that elicited by C5a in J774 cells. Ca2+ influx appears to contribute to the chemotactic response of BMMC to ADP, unlike the response of J774 cells to C5a. These findings point to a possible role for adenine or uridine nucleotides leaked from damaged cells or secreted by enteric neurons in the recruitment of mast cells by parasite-infested tissue.
Materials and Methods
ADP, ATP, UTP, BSA, NaHCO3, N-methyl-d-glucamine (NMDG), thapsigargin, quinidine, laminin, and human recombinant C5a were from Sigma (St. Louis, MO). Adenosine 5′-O-thiodiphosphate (ADPβS) and adenylyl-[β,γ-methylene]-diphosphate (AMPpCp) were from Boehringer Mannheim (Indianapolis, IN); 2-methlylthio-ATP was from Research Biochemicals International (Natick, MA); Pertussis holotoxin was from List Biological Laboratories (Campbell, CA). Tissue culture media, FBS, horse serum, and antibiotics were from Life Technologies (Rockville, MD). Thapsigargin, quinidine, and wortmannin were dissolved in methylsulfoxide and diluted 500-fold into aqueous medium to achieve the final working dilutions. At this dilution, vehicle alone had no effect on chemotaxis of either cell type.
Rat BMMC were cultured from bone marrow of Fisher 344 rats, as described previously (14), except that the growth medium consisted of the following mixture: 65% RPMI 1640 containing 20% horse serum and 50 μM 2-ME, 25% culture supernatant from COS-1 cells transfected 3 days prior with plasmid containing gene for IL-3 (32), 10% culture supernatant from COS-1 cells transfected 3 days prior with gene for soluble form of human stem cell factor (Genetics Institute, Cambridge, MA). Cells were trypsinized and seeded at 3–5 × 106 in 10-cm petri dishes (tissue culture grade) 3–5 days before the experiment. Rat peritoneal mast cells were isolated from retired breeder Sprague Dawley rats, as previously described (33), and used for chemotaxis experiments on the day of isolation. A stock culture of murine J774A.1 monocytes (34) from American Type Culture Collection (Rockville, MD) was maintained in spinner culture (3–8 × 105/ml) in DMEM with 5% heat-inactivated FBS and 100 U/ml penicillin and 100 μg/ml streptomycin. Three days before use, cells were removed from spinner culture and seeded at 2.0–3.5 × 106/10 ml in 10-cm petri dishes (bacteriological grade); the medium was replaced the day before the experiment.
J774 cells were eluted with PBS containing 0.2 g/L EDTA, rinsed with appropriate buffer, and suspended at a concentration of 2.6 × 106/ml in this buffer. Monolayers of adherent rat BMMC were rinsed three times with PBS containing 0.2 g/L EDTA, incubated in this buffer for 3–5 min at 37°C, and eluted with a pasteur pipette. Cells were pelleted by centrifugation, rinsed, and resuspended at a concentration of 2.6 × 106/ml in appropriate buffer. Modified Boyden chamber assays (35) for cell migration across 5-μm polycarbonate filters (Neuroprobe, Cabin John, MD) were performed using 48-well plexiglass chambers (Neuroprobe). The lower wells were filled with 27 μl of medium containing BSA at 2 mg/ml, with or without chemoattractants. After addition of 35 μl of cell suspension to the top half of each well, the assembly was incubated at 37°C in an atmosphere containing 5% CO2. After 3 h, the polycarbonate filter was removed, cells adhering to its upper surface were wiped off with Gey’s balanced salt solution (GBSS)-wetted tissue paper, and the filter stained with Diff-Quick (Baxter, McGaw Park, IL). Filters were air dried on the surface of a 7.5 × 5-cm glass microscope slide and examined under oil immersion at a magnification of 1000-fold. Two randomly selected fields were counted for each well, triplicate wells averaged, and the results of three or more such experiments conducted on different days were averaged and expressed as the number of migrated cells per high power field of view, <#/HPF>, ±SEM. Control experiments showed that mast cells and monocytes that had migrated through remained attached to the filter, as no cells were detected in the fluid phase of the lower chamber following agitation with a pipetman and examination by light microscopy.
Chemotaxis experiments were conducted with Medium 199 (Life Technologies), to which 2 mg/ml BSA was added. For experiments in which extracellular [K+] or [Ca2+]i was varied systematically, we used GBSS containing 2 mg/ml BSA; GBSS supported chemotaxis of both cell types to a similar extent as Medium 199. Standard GBSS contained, in mM: 120 NaCl, 4.9 KCl, 27 NaHCO3, 0.22 KH2PO4, 0.84 Na2HPO4, 1.53 CaCl2, 1 MgCl2, 0.28 MgSO4, 5.6 glucose, 2 mg/ml−1 BSA (pH 7.3–7.4). For potassium substitution experiments, equimolar replacement of Na+ with K+ in standard GBSS (5.2 mM K+) yielded a buffer containing 153 mM K+; these two buffers were mixed in proportion to yield solutions of 51, 101, and 126 mM K+. For the NMDG substitution experiments, NaCl was replaced with 120 mM NMDG to yield a buffer containing 28.6 mM Na+, and this solution was mixed with standard GBSS in proportion to yield buffers of 50.5, 70.9, and 111.9 mM Na+. The low calcium GBSS contained 1 mM EGTA and 0.625 mM CaCl2, yielding free Ca2+ ∼100 nM. For the calcium depletion experiments, control GBSS contained 1 mM EGTA and 2.53 mM total Ca2+ (∼1.53 mM free Ca2+).
Except where stated otherwise, chemotaxis assays were conducted using standard polycarbonate filters. Where noted, laminin-coated filters were used. These were prepared by soaking filters for 4 h at 37°C in GBSS containing 10 μg/ml laminin from a mouse sarcoma, rinsing briefly in GBSS containing 2 mg/ml−1 BSA and blotting the filters dry before use.
The viability of cells treated with the described reagents was determined with trypan blue exclusion. After the assay, cells remaining in the upper compartment were resuspended with a pipetman, a 10-μl aliquot diluted 10-fold with PBS containing 0.22% trypan blue, and the cells counted with a hemacytometer. For rat BMMC, viabilities so measured on n days, with ≥6 wells counted per day, were: buffer alone, 98 ± 2% (n = 5); 20 μM ADP, 98.5% (n = 2); 100 μM ADP, 98.6 (n = 1); 20 μM ADPβS, 100 (n = 1); 20 μM UTP, 98.4 (n = 1); 20 ADP/0.5 μM thapsigargin, 96.0 (n = 2); 200 ng/ml pertussis toxin (12–23 h), 98.6 + 1.5% (n = 3); pertussis toxin overnight plus 20 μM ADP during assay, 98% (n = 1). The postassay viability of J774 cells was not determined. The preassay viabilities of untreated and pertussis toxin-treated BMMC and J774 cells were consistently ≥99.5%. The recovery of rat BMMC from pertussis toxin-treated cultures was 93 ± 13% (n = 5) of that for untreated cultures, which, given the method of isolation (see above), could reflect decreased adhesiveness of the pertussis-treated cells, among other factors.
Scanning electron microscopy
Rat BMMC were seeded onto 13-mm round thermanox cover slips in the wells of a 24-well plate (each well contained 1.5 × 105 cells in 1 ml). Two days later, cell monolayers were rinsed twice with standard GBSS at 37°C, and then incubated for 25 min at 37°C in the same buffer with or without 20 μM ADP. Cells were fixed for 3 h with 2.5% glutaraldehyde (in 0.15 M sucrose + 0.15 M Na cacodylate (pH 7.2)), postfixed for 1 h with 1% OsO4 in 0.15 M Na cacodylate (pH 7.2), and then dehydrated in a graded series of ethanol-water mixtures. Cells were then rinsed (20 min) in hexamethyldisilazane-ethanol (1:1), neat hexamethyldisilazane, and after a final rinse in the same, the hexamethyldisilazane was evaporated overnight in the fume hood. Cells were sputter-coated with gold and examined at 10 kV in a JEOL (Peabody, MA) JSM 35 scanning electron microscope.
Morphological changes in rat BMMC treated with ADP
Rat BMMC used in these experiments were loosely adherent to tissue culture-treated plastic, even in the absence of exogenously added laminin or other proteins of the extracellular matrix. Typically, they had two to four pseudopod-like structures extending away from a round cell body (Fig. 1, top panel). After treatment with 20 μM ADP for 25 min at 37°C, BMMC had flattened dramatically against the substrate and lost many of their surface microvilli in the process (Fig. 1, bottom panel). The outline of flattened cells was usually asymmetric rather than circular, although definite polarization into an apparent leading and trailing edge was not common. Using Nomarski DIC optics, rapid ruffling at the cell periphery was observed within seconds of ADP addition to BMMC at 34°C, followed within minutes by generalized cell flattening. Such events were not observed at room temperature. Using light microscopy, similar changes were noted in J774 cells within seconds of exposure to extracellular ADP, even at room temperature (data not shown). In addition to spreading out, in response to ADP, the J774 cells showed very dynamic surface ruffling activity.
P2 receptor-mediated migration
As shown in Fig. 2, A and B, both rat BMMC and murine J774 monocytes migrated across unipore polycarbonate filters toward a source of extracellular ADP. The response of rat BMMC was greater and saturated at a lower concentration of ADP in the bottom well (EC50 ∼ 3 μM) than did the monocyte response. In 13 experiments conducted with rat BMMC in Medium 199, 20–25 μM ADP stimulated the migration of 42.1 ± 4.6 cells/HPF (range 22–74), compared with a spontaneous migration 8.7 ± 2.3 cells/HPF (range 0–20). On a day-by-day basis, the ratio of ADP-stimulated to spontaneous migration averaged 5.2 ± 2.8, range 2 to infinity (zero background). For 12 experiments in GBSS, the spontaneous and ADP-elicited migration was 7.0 ± 1.9 and 31.7 ± 5.2 cells/HPF, respectively. The ratio of ADP-stimulated to spontaneous migration was on average 8.4 ± 1.1, range 1.7 to infinity. The level of spontaneous migration varied substantially between experiments, but it showed no correlation with the magnitude of responses to ADP.
Because surface ectonucleotidases can convert ADP to adenosine and mast cells express adenosine receptors (11, 12), we tested the possible requirement for ADP hydrolysis in the chemotactic response. The poorly hydrolyzable derivative of ADP, ADPβS, had similar efficacy to ADP, suggesting that conversion of ADP to adenosine is not required for the chemotactic response. In paired experiments conducted on three days, 20 μM ADPβS and 20 μM ADP stimulated migration of 28.2 ± 8.6 and 32.5 ± 5.8 cells/HPF, respectively. Thus, P2 rather than P1, receptors appear to mediate the chemotactic response to ADP. Fig. 3 A shows results of experiments using laminin-coated filters, which further support this conclusion. The poorly hydrolyzable ATP analogue, AMPpCp, as well as ADPβS and ADP, each stimulated a large increase over spontaneous levels of BMMC migration.
Neither the P2T receptor for ADP nor the P2Z (P2×7) receptor for ATP4− appear to mediate the chemotactic response to ADP, given that ADP and ATP were similarly effective, and active at much lower concentrations than those required to ligate the P2×7 receptor with ATP4− (reviewed in Refs. 37, 38). In one experiment conducted in triplicate with rat BMMC, the number of cells migrated was 34 ± 9.2 and 25 ± 6.0 for 25 μM ADP and ATP, respectively. At 100 μM nucleotide, the respective values for ADP and ATP were 44 ± 10.2 and 27 ± 1.0. This suggests that ADP is at least as effective as ATP, unlike the case with P2×7 receptors for ATP4− (13, 39). As indicated in Fig. 3 B, none of the nucleotides that served as chemotaxins for rat BMMC exhibited appreciable chemotactic activity for RBL-2H3 cells, a transformed mucosal mast cell line from rats (40, 41). Under the present assay conditions, neither did ADP (20 and 100 μM) act as a chemoattractant for rat peritoneal mast cells.
A shown in Fig. 2,C, UTP was an effective chemoattractant for rat BMMC, although it was much less active on J774 cells, either on laminin-coated or standard filters (Fig. 2 D). Interestingly, UTP was much more potent than was ADP. The migratory response saturated over a range of concentrations 10-fold lower than those over which the response to ADP saturated, and the concentration of UTP which produced 50% of the maximal response was ∼0.5 μM, or 6-fold lower than the EC50 for ADP-stimulated migration. The magnitude of the maximal response was similar to that elicited by ADP; in three paired experiments, the number of migrated BMMC was 32.5 ± 4.6 vs 32.5 ± 5.8 for 20 μM UTP and ADP, respectively. This provides further evidence that adenosine (P1) receptors do not mediate the chemotactic response or rat BMMC to nucleotides. A much greater potency of UTP over ADP typifies P2U receptor ligand-binding specificity (37), and the effects of pertussis toxin described in a later section argue against a role for P2X receptors. Whether P2Y receptors might also contribute to the chemotactic response is unresolved.
Anaphalytoxin C5a is well-established as a chemoattractant for neutrophils and monocytes, and activated mouse serum containing C5a has been shown to mediate migration of J774 cells (42). As shown in Fig. 2 F, we found that over a broad range of concentrations, human recombinant C5a caused directed migration of J774 cells. The concentration-response curve was biphasic, peaking at 1–10 nM C5a. Biphasic concentration-response curves are typical of many chemoattractants, including those active on mast cells; for concentration-response curves of other mast chemoattractants see references (6, 7, 20, 22, 23, 24, 25, 26, 27). C5a was active on J774 monocytes at much lower concentrations than was ADP, producing 50% of the maximal response when present in the lower chamber at a concentration of ∼0.5 nM. This is consistent with the relatively higher affinity of the C5a receptor for its ligand (Kd ∼1–2 nM for human C5a receptor (43, 44)) than that of P2 receptors for adenine nucleotides (37, 38).
Two recent studies showed that purified C5a is chemotactic for human mast cells (24, 25). These studies employed either laminin- or fibronectin-coated filters, and Hartmann et al. (24) observed no migratory response to C5a in the absence of laminin. In contrast to these results, we found that C5a, over the range of concentrations 0.1–1000 nM, did not stimulate significant migration of rat BMMC through uncoated or laminin-coated filters (Fig. 2 E).
Chemotactic and chemokinetic components of cell migration
The stimulated migration shown in Fig. 2 could result either from an increase in random motion of cells (chemokinesis) or from a selective increase in directed migration (chemotaxis). To determine the relative contributions of chemotaxis and chemokinesis to the observed response, we performed a set of experiments where, at maximally effective concentrations, chemoattractants were added either to the lower compartment alone or to both upper and lower compartments of the chamber assembly. In the latter case, because no gradient of chemoattractant is present across the filter, cells that migrate to the lower surface of the filter must do so by an increase in random, rather than directed, motion. As shown in Fig. 4, chemokinesis accounted for 50% or less of the migratory response of both cell types to ADP. Migration due to chemokinesis was on average 50 ± 10% (n = 9) of that due to chemotaxis for rat BMMC and 27 ± 11 (n = 4) for J774 cells. In contrast, a major portion of the migratory response of J774 cells to C5a was due to chemokinesis; the number of cells migrating due to chemokinesis being 70 ± 7% (n = 11) of the value for chemotaxis. The concentration dependence of C5a-induced chemokinesis is shown in Fig. 2 F for the two chemokinesis experiments in which we varied C5a concentration. Apparently, J774 cells have similar chemokinetic and chemotactic sensitivities toward C5a.
Effects of extracellular calcium on chemotaxis toward ADP and C5a
Fig. 5 demonstrates the effects of depleting extracellular Ca2+ on C5a and ADP-directed chemotaxis. In these experiments, 1 mM EGTA was used in the upper and lower chambers to buffer external free Ca2+ at ∼100 nM, similar to the intracellular concentration of free Ca2+, [Ca2+]i, in resting cells. This treatment profoundly inhibited chemotaxis of J774 cells and rat BMMC toward ADP. The modest spontaneous migration of both cell types (no attractant) was also abrogated. In rat BMMC, ADP-induced migration in low Ca2+ buffer was merely 1.3 ± 0.7% (n = 6) of control buffer, and in J774 cells, the migration in 100 nM external Ca2+ was 12.4 ± 4.7% (n = 6) of that in 1.53 mM Ca2+. These differences are significant at the 0.5% level. In marked contrast, Ca2+ chelation had only a modest effect on C5a-mediated chemotaxis of J774 cells, the migration in low calcium buffer being 72.8 ± 1.4% (n = 5) of that in buffer containing a physiological level of calcium; in fact, this difference was insignificant. This suggests that influx of external Ca2+ does not make an essential contribution to C5a-induced migration. The inhibition of ADP-mediated chemotaxis was not a toxic effect of EGTA, as the control buffer (1.53 mM free Ca2+) also contained 1 mM EGTA, and such toxicity was not manifest in the response to C5a.
The dependence of ADP-driven chemotaxis on extracellular Ca2+ might reflect a requirement for Ca2+ influx per se, or a crucial role for Ca2+ ions in protein conformation or other surface features important in cell motility. In principle, one can discriminate between these possibilities by inhibiting influx of external Ca2+ with a Ca2+ channel antagonist like cadmium. In one experiment, we tested the effect of 2 mM extracellular Cd2+ and found that that it totally blocked chemotaxis of rat BMMC and J774 cells toward ADP; it also completely inhibited C5a-induced migration of J774 cells. One problem in interpreting this experiment is that, apart from its long-term cellular toxicity, external Cd2+ may inhibit chemotaxis by displacing Ca2+ ions from critical binding sites on the cell surface rather than by blocking Ca2+ influx. Below, we describe the results of a test that does not require removal of Ca2+ ions or the addition of Ca2+ competitors.
Inhibition of ADP-directed chemotaxis by membrane depolarization
To address the role of extracellular Ca2+ indicated above, we tested the effect of membrane potential on ADP-mediated chemotaxis of rat BMMC and J774 cells. Mast cells and monocytes are electrically inexcitable cells, in which membrane depolarization inhibits Ca2+ influx by reducing the electrical driving force on Ca2+ entry (45, 46, 47). As the predominant ionic conductances present in rat BMMC and J774 cells are K+ conductances (14, 48), increase in extracellular [K+] is expected to effectively depolarize the plasma membrane and inhibit calcium entry. Fig. 6 shows that elevation of extracellular [K+] inhibited both ADP- and C5a-mediated chemotaxis in a concentration-dependent manner. Consistent with its lower sensitivity to depletion of extracellular calcium, C5a-mediated chemotaxis was also less sensitive to elevation of [K+] than was chemotaxis toward ADP. These findings suggest that potassium elevation inhibits chemotaxis by depolarizing the membrane.
To further test this interpretation, we performed a set of experiments wherein extracellular sodium was replaced systematically with the nonpermeant monovalent cation, NMDG. Unlike potassium, NMDG is not expected to depolarize the cell membrane. Thus, if it is a decrease in extracellular sodium rather than the increase in external potassium concentration that inhibits directed migration, NMDG should mimic the effects of potassium elevation. In principle, decreases in external sodium could inhibit chemotaxis by reducing capacitative Ca2+ influx via altered mitochondrial Ca2+ transport (49) or though osmotic disequilibrium consequent to altered sodium transport. As shown in the inset to Fig. 6, NMDG mimicked the inhibition of C5a- but not ADP-mediated chemotaxis by potassium substitution. In these experiments, NMDG was substituted for sodium at the levels for which potassium substitution yielded 50% inhibition of ADP (98 mM NMDG) or C5a-directed chemotaxis (120 mM NMDG). There was little difference between the effect of potassium and NMDG subsitution on C5a-mediated chemotaxis, suggesting that, in this case, sodium depletion per se could account for the reduced chemotactic response in high potassium buffers. Thus, the results with C5a are inconclusive. On the other hand, the inhibition of ADP-directed migration cannot be explained by sodium depletion alone. These observations support the idea that a sufficiently negative membrane potential and receptor-coupled Ca2+ influx are critical for the chemotactic response of rat BMMC and J774 cells to ADP.
Effects of Ca2+ store depletion on chemotaxis
We examined the possible role of intracellular Ca2+ release in chemotaxis by depleting internal Ca2+ stores with the thapsigargin. By inhibiting the endoplasmic reticulum Ca2+ pump (50), thapsigargin blocks reuptake of stored Ca2+ ions, thereby preventing cyclical release or partial release and refilling of internal Ca2+ stores consequent to stimulation of calcium-mobilizing receptors. Addition of thapsigargin to both upper and lower wells of the chamber, at concentrations known to completely deplete intracellular Ca2+ stores in the related RBL-2H3 cells (L. Zhang and M. A. McCloskey, unpublished data), had differential effects on cell migration toward C5a and ADP. Migration of rat BMMC toward ADP was abrogated, whereas the inhibition of C5a-mediated chemotaxis of J774 cells was incomplete. In six experiments with BMMC stimulated with 20 μM ADP in the lower chamber, the <#/HPF> was reduced from 49.2 ± 6.8 to 0.3 ± 0.3 by treatment with 0.5 μM thapsigargin (p < 0.005); spontaneous migration in these experiments averaged 11.8 ± 2.5 cells per high power field. In contrast, in four experiments with 1 nM C5a, the average number of J774 cells per high power field was reduced only from 33.2 ± 4.6 to 14 ± 10.9 (p < 0.005); the background migration was 6 ± 3.5 cells/HPF. These findings suggest that partial release and refilling of internal Ca2+ stores is not absolutely essential for the migratory response of J774 cells to C5a, but that ADP-mediated chemotaxis of rat BMMC depends absolutely upon the existence of a refillable Ca2+ store.
Trimeric G proteins in P2 receptor-mediated chemotaxis
Fig. 7 shows that pretreatment of rat BMMC or J774 cells for 12–23 h with pertussis toxin (200 ng/ml) inhibited the chemotactic responses to ADP and C5a. Inhibition of C5a-mediated migration was complete, in three experiments with 1 nM C5a, the <#/HPF> being reduced from 36.5 ± 0.5 to 1.5 ± 0.5 (p < 0.005); in these experiments, the spontaneous migration was 0.5 ± 0.5 cells per HPF. In the same experiments, pertussis toxin essentially prevented the more modest migration of J774 cells toward 200 μM ADP, the <#/HPF> being reduced from 12.5 ± 4.5 to 1.5 ± 0.5. Pertussis toxin strongly inhibited ADP-stimulated migration of rat BMMC. In eight experiments with an average spontaneous migration of 7.5 ± 2.8 cells per HPF, pretreatment with the toxin reduced the <#/HPF> from 36.1 ± 4.6 to 11.8 ± 4.9 (p < 0.005). Pertussis toxin inhibited to a similar extent both UTP- and ADPβS-mediated chemotaxis. The migratory response to UTP was reduced from 32.5 ± 6.4 to 6.1 ± 4.0 cells/HPF, and the response to ADPβS was reduced from 28.2 ± 8.6 to 4.8 + 3.5 cells/HPF, respectively, in three experiments where the spontaneous migration was 3.7 ± 3.3 cells/HPF.
Pertussis toxin ADP ribosylates the trimeric G proteins transducin, Go and Gi1–3, uncoupling them from cognate receptors. Of these proteins, only Gi2 and Gi3 appear to be expressed in rat BMMC (Y.-X. Qian, Master’s thesis, Iowa State University, 1992) or the related RBL-2H3 cell line (51), and neither transducin nor Go has been found in hematopoietic cells. These observations suggest that the P2 receptor(s) mediating chemotaxis of J774 cells toward ADP is coupled to a Gi-related protein, and all the intracellular signals required for chemotaxis are initiated by this species. Incomplete block of ADP-mediated chemotaxis of rat BMMC suggests that in addition to Gi, another G protein could participate in the chemotactic response of rat BMMC to ADP.
The present work demonstrates that the nucleotides ADP, ATP, and UTP serve as directional cues for mast cell migration. Stimulation of mast cells with adenine nucleotides was previously shown to regulate pore formation (39) via a P2Z (or P2×7) receptor, to mobilize intracellular Ca2+ (15), and to activate K+ channels (14, 16) via P2Y and/or P2U receptors. The pertussis toxin-sensitivity of mast cell chemotaxis coupled with the 6- to 10-fold greater potency of UTP over ADP supports the involvement of a P2U receptor (10), although an additional role for P2Y receptors cannot be excluded at this point. Interestingly, P2U receptors were recently shown to mediate chemotaxis of human neutrophils (18).
Orida et al. (7) were the first to demonstrate in vitro chemotaxis of mast cells toward a defined substance, viz., Ag. Among four variants of the rat basophilic leukemia cell line tested, just one (926a) showed appreciable chemotactic activity equivalent to ∼6 cells/HPF in our system. The RBL-2H3 variant they tested was barely active, with a migratory index equivalent to ∼0.5 cells/HPF. In our study, P2 agonists also failed to induce substantial migration of RBL-2H3 cells, (Fig. 3 B), even though these cells express P2Y and/or P2U receptors that initiate Ca2+ release and activate K+ channels through a Ca2+-independent mechanism (14, 15, 16). A perhaps related physiological difference between this mucosal mast cell line and nontransformed mucosal mast cells is that ligation of P2 receptors does not reliably potentiate Ag-driven secretion from RBL-2H3 cells like it does from rat BMMC (14). Rat peritoneal mast cells express calcium-mobilizing P2 receptors (15), but these cells did not migrate toward extracellular ADP (20 and 100 μM). This probably does not reflect a loss of chemotactic ability during mast cell maturation/differentiation, as TGF-β1 was shown to be a chemotaxin for rat peritoneal mast cells (6).
In contrast to the situation with macrophages and neutrophils, little is known regarding the intracellular events that mediate mast cell chemotaxis. Studies with the tyrosine kinase inhibitor genistein and experiments with murine mast cells bearing a defective c-kit kinase have implicated tyrosine kinase activity in the migration of mast cells toward stem cell factor (27) and certain angiogenic factors (20). Pertussis toxin also inhibits the migratory response of a human mast cell line (HMC-1) to complement components C3a and C5a by 85% and 100%, respectively, implicating a Gi-dependent mechanism (24, 25). We observed complete inhibition by pertussis toxin of C5a- and ADP-driven chemotaxis in J774 cells, whereas the response of rat BMMC to ADP was blocked by ∼70%. This difference could mean that critical chemotactic signals are transmitted from the relevant purinoceptor(s) to G proteins of multiple families in mast cells, but that Gi-related species carry all the essential information for stimulated migration of J774 cells. In this context, it is intriguing that the ADP-evoked Ca2+ signal in J774 cells (30), but not RBL-2H3 mucosal mast cells (M. McCloskey, unpublished observations), is completely inhibited by pertussis toxin.
Evidence presented here indicates that in addition to its Gi dependence, ADP-stimulated migration of rat BMMC and J774 cells depends upon influx of extracellular Ca2+ ions. This is in sharp contrast to C5a-driven chemotaxis of J774 cells, which was rather insensitive to extracellular Ca2+ under the present conditions. Extracellular Ca2+ is also necessary for chemotaxis of neutrophils over vitronectin or fibronectin substrates (52). Although the exact function of external Ca2+ in neutrophil migration is unclear, it is required for observation of intracellular Ca2+ transients, which, via the activation of calcineurin, mediate detachment of the neutrophil’s trailing edge (53). Whether this mechanism operates in mast cells and what role Ca2+ influx or changes in [Ca2+]i might play in the mast cell system, remain to be determined.
It is relevant to compare the effectiveness of ADP with that of other mast cell chemoattractants. Complement fragment C3a has been described as the most effective chemoattractant yet found for human mast cells, eliciting at optimal concentrations a migratory response 2.6- to 6-fold greater than the spontaneous migration rate for the HMC-1 human mast cell line (24, 25). By this measure, the efficacy of ADP as a chemotaxin for rat BMMC was similar to that of C3a acting on HMC-1 cells, inducing average responses 5.2- and 8.4-fold greater than the spontaneous migration in Medium 199 and GBSS, respectively.
Active at subpicomolar levels, TGF-β1 is the most potent chemoattractant yet described for mast cells. Under conditions similar to those used here, TGF-β1 induces a greater migratory response of murine mast cells than does stem cell factor, IL-3, or laminin (6). At optimally effective concentrations, it elicits the migration of twice as many mast cells per unit area of filter as did 20 μM ADP. Given the smaller pore size (5 vs 8 μm) and cell-loading density used in the present work (0.9 vs 2 × 104/well), ADP might be at least as effective as TGF-β1 if assayed under identical conditions. By these criteria, although clearly less potent than TGF-β1 and C3a, ADP is a very effective attractant for rat BMMC.
Leakage of cytosol from cells injured during a parasite infestation is expected to generate a steep concentration gradient of extracellular adenine and uridine nucleotides, with local concentrations well within the effective range shown in Fig. 2. It is conceivable that nucleotides so released participate in recruitment of immature mast cells by the mucosa, whereupon the mast cells divide under the influence of local stimuli, such as IL-3, stem cell factor, or Ag. Alternatively, the true significance of purinoceptor-mediated chemotaxis of mast cells could lie in the establishment of intimate neuron-mast cell contacts observed in rat intestines and other anatomical sites.
We thank Ms. Kristin Brandner for technical assistance with some experiments and Dr. George Ehring for critical comments on the manuscript.
This work was supported by National Institutes of Health Grant GM48144.
Abbreviations used in this paper: RBL, rat basophilic leukemia; ADPβS, adenosine 5′-O-thiodiphosphate, AMPpCp, adenylyl-[β,γ-methylene]-diphosphate; BMMC, bone marrow-cultured mast cells; GBSS, Gey’s balanced salt solution; HPF, high-power field of view (1000×); NMDG, N-methyl-d-glucamine; HMC, human mast cell.