|Home | About | Journals | Submit | Contact Us | Français|
We have pioneered an in vitro pseudopod-generation model wherein suspended tumor cells are stimulated to form pseudopods into glass micropipettes in response to soluble collagen type IV (CIV). Pertussis toxin and removing intracellular calcium were found previously to be inhibitory to that process. We now extend those observations to dissect the roles of transmembrane calcium influx and circulating fatty acids on pseudopod extension. Removal of fatty acids from BSA in basal media resulted in abrogation of pseudopod formation, while reconstitution of free fatty acids restored cell pseudopod protrusion. We thus hypothesized that fatty acids may provide necessary pseudopod stimulatory signals. Addition of lysophosphatidic acid (LPA) to the fatty acid-free CIV solution or in an opposite pipette without CIV permitted approximately 50% pseudopod recovery in all pipette directions in a dose-dependent fashion. Thapsigargin (TG), an agent that releases internal calcium stores and causes opening of store-operated calcium channels, restored pseudopod protrusion up to 80% in CIV with fatty acid-free albumin. [Ca2+]i release was non-additive when cells were stimulated by TG and LPA, suggesting overlapping [Ca2+]i stores. The combination of TG and LPA in fatty acid-free albumin fully restored the pseudopod response to CIV. Addition of EGTA to chelate stimulatory media calcium blocked the pseudopod response to CIV in the presence of fatty acids. This indicates that pseudopod protrusion requires transmembrane calcium entry. Thus, extracellular lipids and calcium mobilization are required to complement CIV in pseudopod protrusion from suspended cells.
Active cell locomotion requires biochemical signaling, both intra- and extracellular in origin. A migrating tumor cell typically follows a gradient of soluble attractant molecules that induce crawling behavior, termed “chemotaxis” (Alberts et al., 1994). Such signaling molecules could be growth factors, cytokines or proteins found in the extracellular matrix (ECM), such as fibronectin, laminin and various collagens (Aznavoorian et al., 1990; Lester and McCarthy 1992; Schnaper and Kleinman 1993). Continuous cycling of protrusion and re-absorption of pseudopods following the direction of increasing chemoattractant concentration gradient characterizes the crawling behavior induced by these chemoattractants (Stossel, 1993; Bray, 1992). Thus, understanding the regulation of pseudopod protrusion may yield insight into cellular motility.
Chemotactic stimulation of A2058 human melanoma cells by type IV collagen (CIV) with fraction V BSA (FV BSA) has been studied using the micropipette system (Dong et al., 1994a,b; You et al., 1996; 1999). The combination of CIV and FV BSA significantly stimulated cell pseudopod protrusion into the micropipette, migration and [Ca2+]i release in vitro (Aznavoorian et al., 1990; Dong et al., 1994a; Savarese et al., 1992). We have previously observed that chelating extracellular Ca2+ with EGTA or [Ca2+]i with BAPTA-AM attenuated pseudopod extension significantly (Dong et al., 1994a). However, the effects of using FV BSA with its associated lipids were not addressed in these studies. We observed that removal of free fatty acids from the chemotactic solution by using purified fatty acid-free BSA (FAF BSA) significantly attenuated A2058 cell pseudopod protrusion in response to CIV. We therefore hypothesized that extracellular lipid homeostasis may affect cellular functioning; the role of cellular Ca2+ regulation in lipid-free conditions required further examination. We therefore used the A2058 human melanoma cell line as a model with which to study the influence of extracellular lipids and calcium regulation of tumor-cell pseudopod protrusion in response to CIV. Independent [Ca2+]i-mobilizing agents thapsigargin (TG) and L-α-lysophosphatidic acid-oleoyl (LPA) were used with CIV in the absence of extracellular lipids to effect cellular Ca2+ mobilization. Additional cellular Ca2+ release via these agents provided the means to assay the dependence of cell pseudopod protrusion on [Ca2+]i.
A2058 human melanoma cells were maintained in tissue culture as described previously (Aznavoorian et al., 1990; Stracke et al., 1991). Cells were detached when subconfluent by brief trypsinization and allowed to regenerate for 1 hr in culture media. Cells were resuspended in serum-free DMEM (Biofluids, Rockville, MD) containing 0.1% w/v Cohn FV BSA (Cohn et al., 1947) or purified FAF BSA with 0.02M HEPES (Sigma, St. Louis, MO) at a concentration of 3.5 × 105 cells/ml. Cells were allowed to regenerate for 1 hr prior to the assay. Cells were used between passages 14 and 20 for all experiments.
The procedures for single and dual micropipette assays have been described in detail elsewhere (Dong et al., 1994a; You et al., 1996, 1999). In brief, glass micropipettes with 6.6 ± 0.4 μm i.d. were back-filled with chemotactic solution containing 100 μg/ml CIV (Becton Dickinson, Bedford, MD) with 0.1% FV or FAF BSA in serum-free DMEM supplemented with 0.02 M HEPES. In dual pipette experiments, pipette A contained 100 μg/ml CIV/0.1% FV BSA and pipette B contained 100 μg/ml CIV/0.1% FAF BSA. The cell suspension and experiment chamber contained 0.1% FV BSA in DMEM with 0.02 M HEPES for all cases assayed. Micropipette tips were positioned against the surface of a cell suspended in experimental medium within a test chamber on the stage of an inverted microscope. A slight aspiration pressure was applied via the pipette, to form a tight seal between the cell and the micropipette tip. At this point, the cell came into direct contact with the focal stimulation. Real-time images were recorded and analyzed off-line to measure the extent of pseudopod protrusion as a function of elapsed time. All experiments were performed under ambient conditions.
In experiments involving TG (Biomol Research Laboratories, Plymouth Meeting, PA), TG was dispersed into anhydrous DMSO (Sigma) and diluted to final concentrations with the chemotactic solution as required. For dual micropipette experiments, pipette 1 contained 100 μg/ml CIV/0.1% FAF BSA and pipette 2 contained 1.0 μM TG/0.1% FAF BSA. LPA (Sigma) was diluted in serum-free DMEM containing 0.1% FAF BSA and 0.02 M HEPES to final concentrations with the chemotactic solution as required. For dual micropipette experiments, pipette 1 contained 100 μg/ml CIV/0.1% FAF BSA and pipette 2 contained LPA 10 μM/0.1% FAF BSA. The cell suspension in the experiment chamber contained 0.1% FV BSA in serum-free DMEM with 0.02 M HEPES for all cases assayed.
Fatty acid–loaded FAF BSA solutions were prepared in the following groups: oleic acid, palmitic acid, stearic acid and linoleic acid in a 3:3:3:1 ratio, all in serum-free DMEM with 0.02 M HEPES. Reconstitution of free fatty acids onto FAF BSA was previously described (Spector and Hoak, 1969). In brief, 0.5 mmole of fatty acid (oleic, palmitic, stearic and linoleic acids; Sigma) was dispersed into 100 ml n-hexane (>95% HPLC grade, Sigma) and dissolved completely. Celite (5 g, diatomaceous earth; Sigma) was suspended into the solution of n-hexane–containing fatty acid and was mixed using a magnetic stirrer. The solution was thoroughly dried under N2 and homogenized. The concentration of fatty acid on Celite was 0.1 mmole/g Celite. The mixture was stored desiccated at −20°C until assay.
On the day of experiments, fatty acid–loaded FAF BSA solution was prepared, assuming a 1.51 × 10-5 mole fatty acid/g BSA final concentration at a 60% efficiency of transfer (Spector and Hoak, 1969; Spector, 1986). Celite (0.252 g/g FAF BSA) was mixed with the appropriate amount of FAF BSA and serum-free DMEM with 0.02 M HEPES to make a 0.1% final BSA concentration. The mixture was stirred magnetically for 30 min, followed by centrifugation and filtration using a 0.45 μm syringe filter (Gelman, Ann Arbor, MI). The prepared BSA solution was used to resuspend CIV at the indicated concentration to produce the chemotactic solution in the micropipette.
A2058 cells were cultured onto 25mm round glass coverslips (pre-treated with 0.1% gelatin solution overnight) and maintained under standard culture conditions. The procedures for the digital calcium ratiometric assay are detailed elsewhere (Savarese et al., 1992; Grynkiewicz et al., 1985; Tsien and Harootunian, 1990). Axon Imaging Workbench 2.1 software (Axon Instruments, Foster City, CA) was used to control the excitation light (340 and 380 nm band pass filters) and to sample and record the emitted fluorescence (510 nm) images from fura-2-AM (Molecular Probes, Eugene, OR)–loaded cells in the field of view once every 4 sec. Background fluorescence was subtracted from each image. Ratio images of cells at rest were collected initially for the first minute, to establish a [Ca2+]i baseline. Then, the indicated solution was perfused into the chamber, and raw images were collected for 10 to 12 min. A calibration curve was constructed by acquiring 340/380 values (background subtracted) of 50 μM fura-2 pentapotassium salt solution using the Calcium Calibration Kit 1 (Molecular Probes) in off-line calibration of the ratio data.
Focal application of CIV/FV BSA via a micropipette resulted in a cell pseudopod protrusion of approximately 12 μm during the 1 hr time course of an assay (Fig. 1a). Elimination of extracellular lipids by replacing FV with FAF BSA resulted in an attenuated pseudopod protrusion to a mean of approximately 2 μm in length. This was almost indistinguishable from the basal negative control response in FV BSA only (without CIV; data not shown). Dual micropipette experiments were performed to test the local requirements for fatty acids (Fig. 1b). Whereas no pseudopod extended in response to lipid-free CIV solution in a single pipette experiment, a transcellular effect was observed when CIV was accompanied by FAF BSA in one pipette and FV BSA in the opposite, yielding equivalent but attenuated extension into each pipette. In all cases assayed, negative control experiments in which CIV was removed from the chemotactic solution were performed to ascertain that small aspiration pressure exerted by the pipette did not physically activate or initiate pseudopod protrusion.
A subconfluent monolayer of A2058 cells was serum-deprived (1 hr), then loaded with fura-2-AM and exposed to CIV in FV or FAF BSA containing medium (Fig. 2a). The majority of cells responded to CIV in FV BSA with a peak and plateau of [Ca2+]i in the range of about 600 nM, occurring in approximately 80 ± 5% of cells (10 to 14 independent experiments). Cells exposed to CIV/FAF BSA showed a slow rise in [Ca2+]i over time and did not reach the response of CIV/FV BSA–exposed cells in peak amplitudes. Peak internal release levels over baseline [Ca2+]i were, on average, approximately 359.5 ± 12.6 nM (CIV/FV BSA) and 208.1 ± 17.0 nM (CIV/FAF BSA) at p < 0.001 (Fig. 2b; unpaired Student's t-test, assuming unequal variance). Experiments in which CIV was removed from the chemotactic solution failed to elicit any [Ca2+]i release (negative control; data not shown).
Removal of extracellular lipids resulted in an attenuated pseudopod response and lowered cellular Ca2+ mobilization. To test whether FV BSA provided the lipids required for the cellular signaling that synergistically activated pseudopod protrusion to CIV, we added free fatty acids to the FAF BSA. As lipids were singularly reconstituted back to basal FAF BSA, pseudopod protrusion in response to CIV recovered to approximately 50% of CIV in FV BSA (Fig. 3a). While addition of individual fatty acids resulted in submaximal recovery, reconstitution of 4 types of fatty acid simultaneously resulted in approximately 90% pseudopod restoration (Fig. 3b).
Removing lipids from CIV chemotactic solution lowered cellular Ca2+ mobilization compared to CIV/FV BSA stimulation. To determine the sensitivity of cell pseudopod protrusion to cellular Ca2+ mobilization, the independent cellular Ca2+-mobilizing agents LPA and TG were used to boost the [Ca2+]i during a lipid-free CIV response.
LPA dose responses to pseudopod protrusion (5 to 15 μM) are shown in Figure 4a. LPA + CIV/FAF BSA stimulated pseudopod protrusion to approximately 50% of CIV/FV BSA. Removal of CIV from this condition failed to elicit any pseudopod response (data not shown). The dose response of TG (0.1 to 1.0 μM) is shown in Figure 4b. Pseudopod protrusion was restored to 8 to 10 μm during the 1 hr course of TG exposure. Saturating response was observed at TG 0.5 μM, where pseudopods reached approximately 80% of the CIV/FV BSA case. Compared with CIV/FV BSA cases, pseudopods exhibited slightly wider, translucent and space-filling morphologies when TG or LPA was applied. TG application without CIV resulted in basal activity (data not shown). The additive effects of LPA and TG on the CIV/FAF BSA background were tested (Fig. 4c). In each tested combination, reconstitution of full pseudopod extension was demonstrated even when subsaturating concentrations of agonist were included.
As shown previously, a single micropipette with CIV/FAF BSA failed to induce pseudopod protrusion. In these experiments, the cell suspension in the assay chamber contained FV BSA at all times. To ascertain that LPA was not contained in the FV BSA preparation, LPA was applied to a cell via a secondary pipette opposite the primary pipette containing CIV/FAF BSA. Transcellular application of LPA stimulated pseudopod protrusion into the primary pipette to approximately 30% to 45% of the CIV/FV BSA condition (Fig. 5a). Cells did not protrude pseudopods into the secondary pipette containing LPA only. If FV BSA carried LPA, transcellular effects would have been observed regardless of the presence of a secondary pipette since the cell suspension contained FV BSA at all times. Control experiments in which a second pipette contained only FV BSA failed to elicit any transcellular activation, ensuring that physical manipulation did not activate the cell response (data not shown).
To test whether cellular Ca2+ mobilization from the agonist applied via a secondary pipette was responsible for pseudopod transactivation, TG was applied in a manner similar to LPA via a secondary pipette opposite the primary pipette containing CIV/FAF BSA. A similar transcellular effect was also observed with this method of TG application (Fig. 5b). Taken together, these results indicated that LPA and TG at 10 and 1.0 μM, respectively, can induce whole-cell activation and drive transcellular pseudopod protrusion, likely via cellular Ca2+ mobilization.
The calcium mobilization activity of LPA (5 and 10 μM) and TG (1.0 μM) on an adherent A2058 cell monolayer in FAF BSA is shown in Figure 6. The peak initial release from internal stores of [Ca2+]i in response to LPA (5 and 10 μM) in FAF BSA was 351.8 ± 16.0 nM and 358.3 ± 19.5 nM, respectively (Fig. 6a,b; resting basal [Ca2+]i subtracted), followed by a lesser influx component. The initial peak [Ca2+]i level approached those of CIV/FV BSA. Percentages of cells responding to LPA/FAF BSA were similar at both dosages of LPA, 90 ± 5%. Further addition of CIV did not significantly increase the peak amplitude, indicating overlapping internal sources (Fig. 6c). Initial peak [Ca2+]i release over baseline for TG (1.0 μM) in FAF BSA was 508.0 ± 14.6 nM (Fig. 6d) with a similar response profile to LPA. Almost 100% of cells responded in this case. When CIV was added to this TG condition, peak [Ca2+]i over basal increased by approximately 100 nM (617.8 ± 17.2 nM over basal, Fig. 6e), indicating that CIV-stimulated [Ca2+]i release occurred partially from non-TG–inducible sources. However, further addition of 10 μM LPA did not induce additional Ca2+ release (597.1 ± 19.5 nM over baseline, Fig. 6f); therefore, it is likely that LPA-stimulated [Ca2+]i release occurred from CIV-sensitive and TG-inducible sources.
Extracellular calcium was removed by chelation with EGTA (3.2 mM), to lower extracellular free Ca2+ to approximately 65 nM at pH 7.4 (Bers et al., 1994). This level was chosen since it was below a typical resting basal [Ca2+]i (100 to 250 nM) encountered in normal A2058 cells. EGTA was included either in the cell suspension chamber (global) or within the micropipette (local) as a solution for the chemoattractant. Pseudopod protrusion in response to CIV/FV BSA was attenuated by both global and local applications of EGTA (Fig. 7a). Local application of EGTA in the micropipette solution was more effective (approx. 80% reduction) at abrogating pseudopod protrusion compared with global application (approx. 50% reduction).
Application of CIV/FV BSA chemotactic solution containing 3.2 mM EGTA onto an adherent A2058 cell monolayer produced a fast initial peak response (208.2 ±17.7 nM) over baseline, followed by an immediate and a rapid decline to basal [Ca2+]i (Figs. 7b,c). The later sustained [Ca2+]i plateau was completely eliminated when EGTA was applied with the chemotactic solution.
In this report, we demonstrate a requirement for extracellular lipids and extracellular calcium on A2058 cell pseudopod protrusion in response to CIV chemotactic stimulation. Alteration in extracellular lipid homeostasis resulted in a significant attenuation of initial [Ca2+]i release and a moderate decrease of the later influx. The causal link between the [Ca2+]i reduction and cell pseudopod abrogation appeared to be a synergistic activation required between CIV and lipids to result in cell pseudopod protrusion. We previously tested another tumor cell line, C8161 metastatic human melanoma, using our micropipette assay system (You et al., 1995). C8161 cells exhibited a similar pseudopod response to CIV, and extracellular lipids were also required to mediate the cell pseudopod protrusion (data not shown). Our observation indicates that extracellular lipids are required for ECM-activated tumor cell pseudopod protrusion in vitro. Our micropipette assay system requires application of a gentle aspiration pressure initially to hold a freely suspended cell on the tip of a micropipette. The exerted pressure remains on the order of 0.6 to 0.8 mm H2O initially; however, it is gradually reduced to 0 pressure by 15 to 20 min into the assay. Our laboratory has shown previously that cells require at least 2 orders of magnitude larger pressure application plus severe disruption of cortical actin structure (via application of cytochalasin B) for physical suction to have any measurable effect (Cao, 1997).
Our study demonstrates that the cellular activation required for pseudopod protrusion in response to CIV is a whole-cell event. The dual pipette technique was capable of transactivating a cell, suggesting that signals required for pseudopod protrusion likely consisted of distinctly intra- and extracellular elements. Reconstituting the intracellular component through application of [Ca2+]i-mobilizing agents via the secondary pipette resulted in pseudopod protrusion in the absence of lipids. Also, [Ca2+]i release was associated with the initial cell activation in response to CIV in an adherent monolayer of A2058 cells. It is possible that the fidelity of the signal-transduction cascade resulting in [Ca2+]i release and influx may be affected when lipids are removed from the stimulatory medium. Transcellular application of [Ca2+]i-mobilizing agents acted on pseudopod protrusion to cause intracellular signaling independent of extracellular chemotactic agonist coupling. This further isolates the role of lipids as key mediators that are essential for transducing ECM–agonist-initiated signaling into the intracellular machinery.
The requirement of an initial calcium increase and cell activation was shown by application of TG and LPA together with CIV in the absence of lipids. TG and LPA have been documented as potent [Ca2+]i-mobilizing agents (Kline and Kline, 1992; Short et al., 1993). We have shown previously that cell pseudopod protrusion is completely abolished by chelating [Ca2+]i using BAPTA (Dong et al., 1994a), further emphasizing that free [Ca2+]i is a necessary, but not sufficient, condition for pseudopod protrusion in response to CIV. In our results, TG application stimulated [Ca2+]i up to approximately 1.7-fold above the levels encountered in CIV/FV BSA. However, the resulting pseudopod protrusion was incomplete, implicating the need for additional intracellular signals. Combinatorial effects of LPA and TG with CIV/FAF BSA were observed to reconstitute pseudopod protrusion completely, while [Ca2+]i activation was not additive. These observations suggested that LPA provided the additional intracellular signaling element besides [Ca2+]i mobilization. LPA is a potent bioactive phospholipid capable of intercellular signaling, recruits its cognate 7-pass transmembrane receptor and is released from activated platelets (Van Corven et al., 1993; Eichholtz et al., 1993; Moolenaar, 1995). The pertinent functionality of LPA, such as activation of focal adhesion kinase, G protein–linked signal-transduction cascade and small GTPase Ras and Rho, has been suggested to play a significant role in cell proliferation and motility (Van Corven et al., 1993; Seufferlein and Rozengurt, 1994; Howe and Marshall, 1993; Chrzanowska-Wodnicka and Burridge, 1994; Moolenaar, 1995). It is therefore entirely possible that additional signaling afforded by LPA in our pseudopod protrusion assay counteracted the lack of extracellular lipid mediation by providing the missing signaling events required to reconstitute a full pseudopod response to CIV.
LPA is a phospholipid with an acyl side chain (-oleoyl); hence, the binding of LPA to BSA was an issue. More specifically, one of our major concerns was the possible presence of LPA in the FV BSA preparation. Surprisingly, definitive information as to this point in the literature is limited. LPA binds to BSA at a high affinity and is present in freshly prepared serum at 2 to 20 μM concentration. However, due to its small size and hydrophilicity from a free hydroxyl and phosphate moiety, it often escapes detection or is precipitated in the presence of the Ca2+ ion (Tigyi and Miledi, 1992; Thumser et al., 1994; Moolenaar, 1995). Our evidence suggests that FV BSA does not contain LPA, as shown in our dual pipette transactivation assay. Application of LPA in FAF BSA using a second pipette induced pseudopod protrusion into the primary pipette with CIV/FAF BSA. Considering that cell suspension in the assay chamber always contained FV BSA, if LPA was sequestered in the FV BSA preparation, then we would have expected a pseudopod response with CIV/FAF BSA applied via a single pipette. Such was not the case; therefore, LPA is unlikely to be present in FV BSA at physiologically relevant levels.
FV BSA contains methanol-extractable fatty acids, some of which can be bioactive (Rosseneu-Motreff et al., 1970; Moolenaar, 1995). BSA functions in vivo as a lipid reservoir (Spector, 1986; Spector et al., 1969; Saifer and Goldman, 1961); thus, the altered lipid content of FAF BSA might eliminate the necessary synergistic lipid agonists required for CIV-stimulated pseudopod protrusion. A question remains as to how such changes in the extracellular lipid environment may affect cellular homeostasis, namely [Ca2+]i regulation and cytoskeletal remodeling. There was the possibility that FAF BSA may somehow extract and sequester bioactive lipids from tumor cells. Bogdanov et al. (1993) used a method of phosphocholine extraction from the cell membrane with FAF BSA incubation. However, their system involved the use of a 20-fold higher FAF BSA concentration plus a 10 min incubation to extract artificially introduced (therefore loosely associated) phospholipids from the membrane. Our assay system used a significantly lower BSA concentration with instantaneous cell–pipette contact. The relative area of cell–pipette contact was also small, at approximately 3.8% of the whole cell-surface area. Based on these issues, we tended to discount this as a viable possibility. Our current study indicates that both cellular [Ca2+]i mobilization and pseudopod protrusion are attenuated in a lipid-free environment. Specifically, the initial [Ca2+]i release is abrogated and the later sustained Ca2+ influx is reduced substantially in the absence of lipids, suggesting that Ca2+ mobilization must either reach a defined threshold or be activated through more than 1 mechanism. This was also suggested by an attenuation of pseudopod protrusion observed while extracellular Ca2+ was chelated by EGTA. This confirmed the requirement for local calcium availability and further indicated that robust [Ca2+]i mobilization in response to CIV is a requirement in a coordinated cell response mediated by extracellular lipids. Indeed, CIV stimulation in the absence of lipids produces lower extracellular Ca2+ influx levels as well as the initial internal release. Taken together, these results indicate a novel role of extracellular lipids as mediators of the cell response to ECM protein.
In summary, we have dissected the signals required for A2058 cell pseudopod protrusion in response to CIV, demonstrating that it requires additional signaling events: mobilization of [Ca2+]i and a lipid-mediated signal. Attenuation of pseudopod protrusion and [Ca2+]i release in response to CIV stimulation in lipid-free conditions was reversible when LPA and TG were applied. Dual pipette experiments have shown that a suspended cell can be transactivated via application of LPA and TG from the opposite side to protrude pseudopods into the pipette containing CIV in FAF BSA medium. This suggests that global [Ca2+]i activation and cell pseudopod protrusion may be distinct events. These factors then require separate signaling pathways, which converge at a point mediated by extracellular lipids. Our results suggest that lack of lipids influences the initial [Ca2+]i release as well as the later influx by affecting the local dependence of extracellular Ca2+ by an unidentified mechanism. Effects of extracellular lipids on A2058 cell pseudopod protrusion have provided us with a useful tool to investigate cell activation. Our current study has revealed the significance of extracellular lipid homeostasis on cell pseudopod protrusion and [Ca2+]i regulation.
Grant sponsor: National Institutes of Health; Grant number: CA76434; Grant sponsor: National Science Foundation; Grant number: BES9502069.