~embrane-bound Protein Kinase C ~ o d u l a t e s Receptor Affinity and Chemotactic Responsiveness of Lewis Lung Carcinoma Sublines to an Elastin-derived Peptide*

The M27 and H59 variants of Lewis lung carcinoma differ in their responsiveness to the chemotactic elastin peptide Val-Gly-Val-Ala-Pro-Gly (VGVAPG). M27 cells, selected for metastasis to lung, are highly responsive to a positive gradient of VGVAPG. H59 cells, selected for metastasis to liver, do not migrate in response to VGVAPG. Although both cell types bind ra- diolabeled VGVAPG, Scatchard analysis of 12‘I-Tyr-VGVAPG binding reveals that M27 cells bind the chemoattractant with a Kd of 2.7 nM, whereas nonresponsive H59 cells bind the peptide with a & of 67 nM. These findings indicate that the failure of H59 cells to migrate in response to VGVAPG may be due to the reduced affinity of their VGVAPG receptors. Both receptor affinity and chemot~ct~c responsiveness to VGVAPG can be modulated in each of these two tumor cell lines by the levels of active membrane-associated protein kinase C. Treatment of nonresponsive H59 cells with 12-0-tetradecanoylphorbol 13-acetate increases the level of membrane-bound protein kinase C activity with a concomitant increase in VGVAPG binding affinity and induction of chemotactic responsive- ness to VGVAPG. Treatment of M27 cells with the protein kinase C inhibitor, staurosporine, performed in 100 pl of assay buffer containing 20 mM Tris (pH 7.5), 5 mM MgCl,, 1.0 mM CaC12, 5 p~ [Y-~'P]ATP, 10 pg/lOO p1 histone 111-s, and 10 pl of enzyme preparation. Phosphatidylserine (40 pg/ml) and 1,a-diolein (10 pg/ml) were suspended in assay buffer by sonication prior to addition to reaction tubes. Reactions were initiated by addition of [y3'P]ATP, allowed to proceed for 5 min at 30 "C, and terminated by addition of 5 ml of iced 25% (w/v) trichloroacetic acid. Precipitated protein was collected on Millipore filters (HA type) using vacuum filtration and washed three times with 5 ml of 25% trichloroacetic acid, and 32P on filters was measured in a Beckman model LS 1800 liquid scintillation counter.

The M27 and H59 variants of Lewis lung carcinoma differ in their responsiveness to the chemotactic elastin peptide Val-Gly-Val-Ala-Pro-Gly (VGVAPG). M27 cells, selected for metastasis to lung, are highly responsive to a positive gradient of VGVAPG. H59 cells, selected for metastasis to liver, do not migrate in response to VGVAPG. Although both cell types bind radiolabeled VGVAPG, Scatchard analysis of 12'I-Tyr-VGVAPG binding reveals that M27 cells bind the chemoattractant with a Kd of 2.7 nM, whereas nonresponsive H59 cells bind the peptide with a & of 67 nM.
These findings indicate that the failure of H59 cells to migrate in response to VGVAPG may be due to the reduced affinity of their VGVAPG receptors. Both receptor affinity and chemot~ct~c responsiveness to VGVAPG can be modulated in each of these two tumor cell lines by the levels of active membrane-associated protein kinase C. Treatment of nonresponsive H59 cells with 12-0-tetradecanoylphorbol 13-acetate increases the level of membrane-bound protein kinase C activity with a concomitant increase in VGVAPG binding affinity and induction of chemotactic responsiveness to VGVAPG. Treatment of M27 cells with the protein kinase C inhibitor, staurosporine, reduces VGVAPG binding affinity and abrogates the chemotactic response. We conclude that chemotactic responsiveness of M27 and H59 tumor cells is dependent upon high VGVAPG receptor affinity, which is strongly correlated to high levels of membrane-bound protein kinase C activity.
Increased motility and enhanced responsiveness to extracellular chemoattractants are associated with malignant transformation (1)(2)(3)(4). Whereas certain tumor cells synthesize and secrete autocrine motility factors (5,6), directional migration of tumor cells is induced by a variety of exogenous chemoattractants including the N-formylated oligopeptide Met-Leu-Phe (MLF)' (7,8), the C5a component of comple-* This work was supported by Grant CA37393 from the National Cancer Institute, United States Public Health Service. 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.
Chemotaxis in response to extracellular chemoattractants is generally mediated by the action of cell surface receptors (1). Specific fMLF receptors have been detected on Walker carcinoma cells and are postulated to mediate the fMLFinduced chemotactic response (17, 18). Dlrectiona~ migration in response to the extracellular matrix adhesion molecules laminin and fibronectin is mediated by cell surface receptors employed for cellular attachment to these glycoproteins (19,20). Although tumor cell migration in response to chemoattractants may enhance tumor invasion and metastasis, little is currently known regarding the biochemical mechanisms of receptor binding and signal transduction involved in tumor cell chemotaxis.
The most intensively studied system of chemoattractant action has been the response of neutrophils to MLF (reviewed in Ref. 21). Neutrophil stimulation with fMLF results in several responses normally associated with inflammation such as migration, aggregation, degranulation, and generation of oxygen radicals. These cellular responses are mediated by cell surface receptors that transduce a series of biochemical signals resulting in neutrophil activation (22)(23)(24)(25). The fMLF receptor exists in at least two affinity states, and treatment of neutrophils with guanine nucleotides results in conversion from the high to low affinity state (26,27). The fMLF receptor is linked to a M, 41,000 regulatory guanine nucleotide-binding protein, which is coupled to phospholipase C, a primary enzyme in the phosphatidylinositol pathway of transmembrane signal transduction (28-30). fMLF binding to its neutrophil receptor consequently results in phospholipase C-mediated hydrolysis of plasma membrane phosphatidylinositol 4,5-bisphosphate to generate inositol 1,4,5-triphosphate and diacylglycerol, Inositol triphosphate mobilizes calcium from intracellular stores, whereas diacylglycerol activates calcium/phosphatidylserinedependent protein kinase C. Current evidence suggests that cell surface receptor binding of fMLF induces motility via the concerted action of changes in calcium ion fluxes and phosp h o~l a t i o n of cellular substrates by protein kinase C (21,251. TPA is a phorbol diester tumor promoter that accelerates the rate of tumor progression following a mutagenic stimulus (31). The major cellular receptor for TPA is protein kinase C (32, 33), and TPA can substitute for endogenous diacylglycerol in activating the phosphotransfer~e activity of protein kinase C (34). TPA treatment of neutrophils causes activation of protein kinase C along with aggegation, generation of superoxide anion, and release of exocytotic granule components, but it does not induce increased cellular motility (35,36). fMLF-induced neutrophil migration is actually inhibited by long term treatment with TPA because the high levels of protein kinase C produce a negative feedback on receptormediated phospholipase C activation (37, 38). Protein kinase C activation is not sufficient for neutrophil migration, and the neutrophil chemotactic response to fMLF is not directly coupled to cellular activation. The mechanisms by which chemotaxis is controlled in nonleukocytic cells have not been as well defined. In this report, we have studied the role of protein kinase C in modulating receptor affinity and chemotactic responsiveness of murine Lewis lung carinoma cells that differ markedly in their chemotactic responses to an elastin-derived chemotactic peptide. VGVAPG, a repeated sequence in the elastin molecule, is a chemoattractant for macrophages, elastin-producing fibroblasts, and certain tumor cells (39-42). The M27 subline of Lewis lung carinoma, selected in vivo for lung metastasis (43), exhibits a particularly strong chemotactic response to VGVAPG (41, 42). M27 cells bind VGVAPG with a Kd of 2.7 nM, and a potential cell surface receptor for the chemotactic peptide has been identified as a M , 59,000 protein by crosslinking of radiolabeled peptide to M27 cells (42). We have begun to explore the biochemical mechanism of action of M27 cell chemotaxis to VGVAPG by comparing the responsiveness of the M27 cell line to a second subline of Lewis lung carcinoma (H59) that does not migrate in response to VGVAPG. Our results demonstrate that chemotactic responsiveness to VGVAPG depends on the affinity of the receptor to its ligand. Furthermore, we find that receptor affinity and, consequently, chemotactic responsiveness can be modulated by agents which alter the level of activated membrane-bound protein kinase C in these Lewis lung carcinoma cells.

EXPERIMENTAL PROCEDURES
Materials-Sodium "' I for peptide iodination and [y3'P]ATP (11 Ci/mmol) for kinase assays were from Amersham Corp. and Du Pont-New England Nuclear, respectively. TPA, histone 111-S, phosphatidylserine, and 1,2-diolein were purchased from Sigma. Staurosporine was from Boehringer Mannheim, and 4-a-TPA was from LC Services (Woburn, MA). The synthetic peptides VGVAPG and YVGVAPG were prepared using an Applied Biosystems 430A peptide synthesizer (Dr. J. Sasse, Shriner's Medical Center, Tampa, FL). Purity of the synthetic peptides was analyzed by reversed phase chromatography on a C8 column (Rainin Instrument Co. Inc., Woburn, MA) and was found to exceed 95% in each case.
Cell Culture-The M27 and H59 sublines of murine Lewis lung carcinoma were derived from metastatic colonies in lung and liver, respectively, following subcutaneous injection of Lewis lung carcinoma cells in C57BL/6 mice (43). These variants were obtained from Dr. P. Brodt (McGill University, Montreal, Canada) and cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum, glutamine/penicillin/streptomycin, and glucose as described previously (43). Cells were periodically re-isolated from lung or liver metastases following subcutaneous injection of 5 X 10' cells in C57BL/6 mice. Only cultured cells passaged fewer than 10 times were employed for these studies.
Migration Assays-Cell migration was evaluated by a modification of the method of Boyden (44) using 48-well chemotaxis chambers (Neuro Probe, Inc., Cabin John, MD) as described previously (42).
Measurement of Peptide Binding-Radiolabeling of the synthetic chemotactic peptide and characterization of peptide binding to M27 cells have been described in detail previously (42). Specific activity of '"I-Tyr-VGVAPG (110 Ci/mmol) was calculated from the amount of radioactivity in the recovered peptide and its absorbance at 280 nm upon reversed phase high pressure liquid chromatography purification. The peptides VGVAPG, YVGVAPG, and '261-Tyr-VGVAPG exhibit nearly identical chemotactic potencies for M27 tumor cells the method of Scatchard (45). (42). Radiolabeled peptide binding by tumor cells was analyzed by Measurement of Protein Kinase C Actiuity-Cells were removed from culture flasks with trypsin/EDTA. Trypsin was inactivated with medium containing 10% fetal bovine serum, and cells were collected by centrifugation at 1,000 X g. Cells were resuspended in 5 ml of a solution containing 20 mM Tris (pH 7.51, 0.25 M sucrose, 2 mM EDTA, 2 mM EGTA, 2 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 0.01% soybean trypsin inhibitor, and 0.01% leupeptin. The cell suspension was transferred to a Dounce homogenizer, and cells were disrupted by 10 strokes with a B pestle. The homogenizer was washed twice with 5 ml of extraction buffer, resulting in a total homogenate volume of 15 ml. The homogenate was separated into cytosolic and particulate fractions by centrifugation at 100,000 X g for 1 h. The supernatant (cytosolic) fraction was centrifuged in Centricon concentrators (Amicon Corp.) containing filters with a M, 10,000 cutoff. The particulate (membrane) fraction was extracted with 0.5% Triton X-100. Protein concentrations were estimated using the BCA protein assay reagent (Pierce Chemical Co.) titrated against a standard curve determined using bovine serum albumin. Both the cytosolic and soluble particulate fractions were then assayed in 10-p1 aliquots (20-50 pg) for protein kinase C activity.
Protein kinase C activity was determined using a previously described procedure (46) by measuring the increase in incorporation of 32P from [y-32P]ATP into histone with the addition of phosphatidylserine and diolein to the assay buffer. Assays were performed in 100 pl of assay buffer containing 20 mM Tris (pH 7.5), 5 mM MgCl,, 1.0 mM CaC12, 5 p~ [Y-~'P]ATP, 10 pg/lOO p1 histone 111-s, and 10 pl of enzyme preparation. Phosphatidylserine (40 pg/ml) and 1,a-diolein (10 pg/ml) were suspended in assay buffer by sonication prior to addition to reaction tubes. Reactions were initiated by addition of [y3'P]ATP, allowed to proceed for 5 min at 30 "C, and terminated by addition of 5 ml of iced 25% (w/v) trichloroacetic acid. Precipitated protein was collected on Millipore filters (HA type) using vacuum filtration and washed three times with 5 ml of 25% trichloroacetic acid, and 32P on filters was measured in a Beckman model LS 1800 liquid scintillation counter.

RESULTS
The M27 and H59 variants of Lewis lung carcinoma have been selected in vivo for preferential metastasis to lung and liver, respectively, upon subcutaneous injection in C57BL/6 mice (43). These two sublines have been shown to differ in patterns of cellular adhesion to extracellular matrix proteins and in patterns of lectin binding (47,48). M27 tumor cells are particularly responsive to the chemotactic elastin-derived peptide VGVAPG when tested in Boyden chamber assays (42). We have tested the H59 subline for chemotaxis to VGVAPG under identical assay conditions employed for M27 cells and found H59 cells to be nonresponsive to VGVAPG over a wide range of peptide concentrations (Fig. 1). The failure of VGVAPG to induce H59 cell migration is not due to an inability of H59 cells to adhere to a fibronectin-coated substrate, as these cells readily attach and spread on fibronectin substrates (47). H59 cells are capable of chemotaxis under these assay conditions, as they exhibit a moderate chemotactic response to trypsin-activated C5a when assayed on fibronectin-coated filters (data not shown).
VGVAPG has been shown to bind specifically to the surface of M27 cells (42). To investigate whether H59 cells were nonresponsive to VGVAPG because they lacked specific receptors, we measured the binding of a radiolabeled VGVAPG analog, "'I-Tyr-VGVAPG. Our previous results demonstrated that this ligand has identical chemotactic activity as VGVAPG and occupies the same binding site on M27 cells (42). Fig. 2A demonstrates that H59 cells bind the radiolabeled peptide and that this binding is saturable. Approximately 90% inhibition of '"I-Tyr-VGVAPG binding to H59 cells could be attained by addition of 10 pg/ml unlabeled VGVAPG to the binding reaction (data not shown). Total specific binding of '"I-Tyr-VGVAPG to H59 cells was seven times greater than binding to M27 cells, and approximately three times more peptide was required to saturate the H59 binding sites. Scatchard analysis of '"I-Tyr-VGVAPG binding to H59 and M27 cells is shown in Fig. 2B. For each cell type, a single class of binding sites is indicated by the data. The dissociation constants and the number of binding sites calculated from the Scatchard plots revealed differences in both receptor affinity and receptor number for the two cell types. M27 cells bound the peptide with higher affinity (2.7 nM) than did H59 cells (67 nM), whereas H59 cells exhibited more available binding sites (350,000) than did M27 cells (52,000). The potential involvement of protein kinase C in the chemotactic response of M27 tumor cells to VGVAPG was initially investigated by treating the cells with the protein kinase C-activating agent TPA. As is shown in Table I, exposure of M27 cells to TPA for 1 h prior to the chemotaxis assay resulted in neither potentiation nor amelioration of the chemotactic response. However, when M27 cells were treated with staurosporine, an inhibitor of protein kinase C activity (49), no chemotactic response to the elastin-derived peptide was observed over a broad range of effector concentrations (Fig. 3). The similarity in control levels (cells traversing the filter in the absence of peptide) of staurosporine-treated cells uersus untreated cells indicates that staurosporine did not reduce background cellular motility but effectively inhibited all specific response to the chemotactic peptide. These results suggest that protein kinase C activity is required for chemotactic responsiveness of M27 cells to VGVAPG.
When H59 cells were treated with TPA, an unexpected result was found. These normally nonresponsive cells now exhibited a chemotactic response to VGVAPG that was similar in magnitude and dose response to that observed for non-TPA treated M27 cells ( Effect of phorbol esters on chemotaxis to VGVAPG Tumor cells were exposed to 100 ng/ml TPA or 100 ng/ml 4-a-TPA for 1 h at 37 "C prior to start of migration assays. The indicated concentrations of VGVAPG were placed in the lower wells of the chemotaxis chamber. Data are expressed as the number of cells that traversed the filter during a 3-h incubation (mean f S. D., n = 8-12).  H59 cells to TPA (18 h) resulted in a reversion to a nonresponsive phenotype (data not shown). Checkerboard analysis (51) of VGVAPG-induced migration of TPA-treated H59 cells is shown in Table 11. Increased migration in the presence of a positive VGVAPG gradient between the upper and lower compartments of chemotaxis chambers confirmed a true directional chemotactic response. No increase in migration was noted when equal concentrations of VGVAPG were placed in both compartments, indicating that VGVAPG did not influence random cell motility. Whereas TPA treatment of H59 cells increased migration in the absence of VGVAPG, a positive gradient of VGVAPG induced the chemotactic response.

Cell line
In view of our finding that M27 cells possess higher affinity VGVAPG binding sites than those found on untreated H59 cells, we inquired whether exposure of H59 cells to TPA might alter the binding affinity of H59 VGVAPG receptors. A series of radioligand binding assays were then performed on H59 and M27 cells that had been treated with protein kinase Cmodulating agents. Specific '251-Tyr-VGVAPG binding to treated and untreated cells was measured and analyzed by the method of Scatchard (45). A summary of binding affinities and receptor numbers calculated from Scatchard plots are presented in Table 111. Exposure of H59 cells to TPA induced a 15-fold increase in the VGVAPG receptor affinity, which correlated with the acquisition of VGVAPG chemotactic responsiveness. In addition, lZ5I-Tyr-VGVAPG binding analysis revealed that the number of binding sites on TPA -TABLE I1 VGVAPG induces directional migration of TPA-treated H59 cells H59 cells were exposed to 100 ng/ml TPA for 1 h at 37 "C prior to migration assays. Gradients of peptide concentration were formed by adding VGVAPG at the final concentration indicated to the lower compartment or with the cells to the upper compartment.

TABLE I11
Effect ofprotein kinase C modulators on VGVAPG binding Tumor cells were exposed either to 100 ng/ml TPA for 1 h at 37 "C or to 10 nM staurosporine for 3 h at 37 "C prior to '"1-Tyr-VGVAPG binding assay. Specific binding of '"1-Tyr-VGVAPG was calculated as the amount of radioactivity displaced by the presence of 5 pg of unlabeled VGVAPG; specific binding was plotted as a function of the ratio of bound to free peptide. Dissociation constants (Kd) were calculated from -l/slope of the plots, and the number of receptor sites/cell (B,) was calculated from the z intercept values. Results are representative of a minimum of three independent binding assays for each condition. The effect of treatment on chemotaxis has been transposed from Table I  To ensure that TPA and staurosporine were indeed altering protein kinase C activity in these cells, we measured calcium/ phosphatidylserine-dependent kinase activity. Protein kinase C activities were determined for both cytosolic (including chelator-extractable activity associated with the plasma membrane) and membrane (particulate; solubilized with Triton X-100) fractions. Protein kinase C activity in untreated M27 cells and H59 cells treated with TPA are shown in Table IV. A comparison of kinase activities in untreated M27 and H59 cells shows that M27 cells had 3-fold more protein kinase C activity in the membrane fraction than did H59 cells, and the membrane/cytosol ratio was seven times higher for M27 cells than for H59 cells. Exposure of M27 cells to TPA did not significantly alter protein kinase C activity in either the membrane-bound or cytosolic fractions, as compared to activities for untreated M27 cells. In contrast, TPA treatment of H59 cells increased membrane-associated protein kinase C activity, with a concurrent decrease in cytosolic protein kinase C activity. The ratio of membrane to cytosolic protein kinase C activities observed for TPA-treated H59 cells resembled the ratio observed for M27 cells. Protein kinase C activity in both the soluble and particulate fractions of M27 cells was inactivated by treatment with staurosporine. These results demonstrate the association of VGVAPG-induced chemotaxis of both tumor sublines with localization of protein kinase C activity in the particulate cell fraction.
Our final experiments were designed to determine whether protein kinase C activity is stimulated in responsive or non-

TABLE IV Protein kinase C actiuity in tumor cells
Tumor cells were treated with 100 ng/ml TPA for 1 h at 37 "C or with 10 nM staurosporine for 3 h at 37 "C prior to cell lysis and fractionation. Soluble fractions (S) and detergent-solubilized particulate fractions (P) were assayed for protein kinase activity by incorporation of 32P into histone substrate. Calcium-independent 32P incorporation, measured in the presence of 2 mM EGTA in the assay buffer, has been subtracted. FIG . 4. Effect of VGVAPG on protein kinase C activity. M27 H59 cells ( B ) , or TPA-t.reated H59 cells (C) were exposed to 5 X lo-' M VGVAPG for the indicated times prior to disruption and separation into soluble and particulate fractions. Protein kinase C activity in soluble fractions (0) or Triton-solubilized particulate fractions (0) was measured as the rate of incorporation of [-y-"P] ATP into histone 111-S substrate, and the radioactivity incorporated in the absence of calcium and phospholipid has been subtracted. Data are presented as the mean results of two independent protein kinase assays in triplicate (S. D. < 5%).

cells ( A ) ,
responsive cells after treatment with the chemotactic peptide. Fig. 4 shows the effect on protein kinase C activity of VGVAPG exposure for various time periods prior to cell fractionation. Within 1 min of VGVAPG addition, protein kinase C activity in the soluble fraction of M27 cells rose by more than 100% (Fig. 4A). No increased kinase activity was detected in the M27 particulate fraction. For nonresponsive H59 cells, exposure to VGVAPG caused no increase in protein kinase C activity in either cell fraction (Fig. 4B). In TPAtreated H59 cells, however, VGVAPG once again caused a significant increase in the activity of protein kinase C in cytosolic fractions (Fig. 4C). A high level of membrane-bound kinase C activity thus correlates with high affinity binding of VGVAPG to its cell surface receptor and corresponds with the chemotactic response. The rise in cytosolic protein kinase C subsequent to ligand binding was observed only in responsive cells and suggests an additional role for protein kinase C in signal transduction upon VGVAPG occupancy of high affinity receptors.

DISCUSSION
The M27 and H59 variants of Lewis lung carcinoma exhibit differential chemotactic responsiveness to the elastin peptide VGVAPG. The inability of H59 cells to migrate toward a positive gradient of VGVAPG is not due to an absence of specific VGVAPG receptors on H59 cells, as both M27 and H59 cells effectively bind VGVAPG. Radioligand binding assays reveal, however, that M27 cells bind the chemotactic peptide VGVAPC with an affinity of 2.7 nM, as compared to 67 nM for VGVAPG binding to H59 cells. The ED5o of biologically active peptides frequently approximates the dissociation constant of the ligand receptor (52). This relationship has been observed for VGVAPG binding to M27 cells (Kd = 2.7 nM) and VGVAPG-induced chemotaxis (ED50 = 3.0 nM), yet VGVAPG concentrations which approximate and exceed the binding affinity of H59 cell VGVAPG receptors fail to induce H59 cell chemotaxis. The presence of specific VGVAPG receptors on the tumor cell surface is, therefore, not sufficient for chemotactic responsiveness.
VGVAPG receptor affinities and chemotactic responsiveness of M27 and H59 cells can be coordinately modulated by treatment of cells with agents that affect protein kinase C activity. Chemotaxis of M27 cells to VGVAPG can be abrogated by treatment with the protein kinase C inhibitor staurosporine. Exposure of M27 cells to staurosporine also de-creases the affinity of the VGVAPG receptor from 2.7 nM to 53 nM, indicating that endogenous levels of active protein kinase C may be necessary for the maintenance of high affinity binding sites and concomitant chemotactic responsiveness. Nonresponsive H59 cells can be induced to migrate toward VGVAPG after pretreatment with the protein kinase C-activating agent TPA. In this case, the induction of chemotactic responsiveness is associated with conversion of cell surface VGVAPG receptors to a higher affinity state. For both M27 and H59 cells, the ability to respond to the chemotactic stimulus of VGVAPG is associated with high affinity receptor binding and with maintenance or induction of protein kinase C activity.
When the levels and distributions of active protein kinase C were examined in each cell type, the above relationships were clarified. Responsive cells (M27 and TPA-treated H59 cells) express high affinity VGVAPG receptors and maintain a high ratio of membrane-associated to cytosolic protein kinase C activity. Nonresponsive cells (H59 and staurosporinetreated M27 cells) express lower affinity VGVAPG receptors and depressed levels of membrane-bound protein kinase C activity. We conclude that chemotactic responsiveness of the M27 and H59 variants of Lewis lung carcinoma requires a VGVAPG receptor in a high affinity state and that the affinity state of the receptor is correlated with the activation and distribution of protein kinase C.
Exposure of responsive M27 cells and TPA-treated H59 cells to the chemotactic ligand induces a rapid rise in cytosolic protein kinase C activity that is not seen in the nonresponsive H59 cells. Although the source of VGVAPG-induced cytosolic protein kinase C activity has not been determined, protein kinase C activity in the particulate cell fraction does not decrease after exposure of responsive cells to VGVAPG. The rapidity of the rise in cytosolic activity suggests the generation of a protein kinase C activator in response to VGVAPG binding. The second messenger diacylglycerol is known to activate protein kinase C in fMLF-stimulated neutrophils (21) and is, consequently, a potential candidate for a cytosolic protein kinase C activator in VGVAPG-responsive Lewis lung carcinoma cells.
There is some evidence that protein kinase C activity in particulate fractions is associated with the state of cellular activation. The appearance of protein kinase C activity in the particulate fraction, for example, is required for chemoattractant-stimulated respiratory burst in neutrophils (53). In addition, platelet activating factor induces protein kinase activity in the particulate fraction of platelets (54). The high levels of protein kinase C tightly bound to the particulate fraction of M27 tumor cells not exposed to TPA indicates a phorbol ester-independent mechanism for maintenance of membranebound protein kinase C activity. For H59 cells, which express less membrane-bound protein kinase C activity than M27 cells, insertion of protein kinase C in the membrane by treatment with TPA correlates with both high affinity binding and chemotactic responsiveness to VGVAPG. Some chemoattractants require "priming" with a co-chemotaxin t o elicit a maximal chemotactic response (25,55). In this regard, TPA may be said to "prime" H59 cells for chemotaxis to VGVAPG by increasing membrane-bound protein kinase C activity and increasing affinity of VGVAPG receptors. TPA also causes an increase in the background levels of H59 cell motility in the absence of chemotactic peptide, and this increased level of cell locomotion may be required for optimal response to VGVAPG. Unstimulated cell motility in M27 cells remains constant in the presence or absence of TPA or staurosporine, suggesting that for these cells protein kinase C activity is necessary for directional but not for random cell motility.
The relationships observed between protein kinase C activity and tumor cell chemotaxis to VGVAPG are not identical in all regards to those described previously for neutrophil chemotaxis to fMLF. Among the differences observed are: 1) TPA treatment of neutrophils increases cell surface expression of fMLF receptors (56, 57), whereas TPA decreases the VGVAPG receptor number on H59 cells and increases the affinity for VGVAPG binding. Down-regulation of transferrin and insulin receptors by TPA has been described previously (58,59). 2) Affinity of the neutrophil fMLF receptor is modulated by guanine nucleotides but not by TPA (26, 27, 571, whereas VGVAPG receptor affinity in both M27 and H59 cells is modulated by TPA and other agents that alter protein kinase C levels. 3) Induction of elevated protein kinase C by TPA treatment of neutrophils decreases chemotactic responsiveness to fMLF via negative feedback mechanisms (37, 38), whereas TPA has no effect on M27 cell chemotaxis to VGVAPG and enhances the chemotactic response of H59 cells to that peptide.
The mechanism by which activated membrane-bound protein kinase C can alter the affinity state of tumor cell VGVAPG receptor is not yet known. A variety of cellular substrates for protein kinase C has been identified previously and includes receptor proteins, regulatory proteins, and cytoskeletal components (32, 60, 61). Tumor cells also express other protein kinases, which might be activated by protein kinase C themselves. The differential levels of protein kinase C in untreated M27 and H59 cells suggest that the effector molecule should be phosphorylated in M27 cells but not in H59 cells.
High levels of membrane-bound protein kinase C have been observed for other tumor cell types, particularly after treatment with tumor-promoting phorbol esters. TPA increases membrane-bound protein kinase C in B16 melanoma sublines, and these increases correlate with increased lung metastasis after intravenous injection (62). A similar enhancement of metastatic potential was observed after TPA treatment of Lewis lung carcinoma cells (63). Phorbol ester-induced membrane-bound protein kinase C activity is also associated with increased drug resistance (64). In this case, effects of TPA could be reversed by calcium channel blockers and calmodulin antagonists. It appears that increased levels of membranebound protein kinase C are correlated with traits of highly aggressive tumor cells. The M27 and H59 Lewis lung carcinoma sublines differ from each other in endogenous levels of protein kinase C activity, in receptor affinity and chemotactic responsiveness to VGVAPG, and in their patterns of metastatic localization. These cells represent a valuable system for investigation of the mechanisms of tumor cell chemotaxis and the relationship of these mechanisms to the process of metastasis.