Kca3.1 Activation Via P2y2 Purinergic Receptors Promotes Human Ovarian Cancer Cell (Skov-3) Migration

Disorders in cell signaling mediated by ATP or histamine, activating specific membrane receptors, have been frequently associated with tumorigenesis. Among the elements of response to purinergic (and histaminergic) signaling, ion channel activation controls essential cellular processes in cancer, such as cell proliferation, motility, and death. Here, we studied the effects that ATP had on electrical properties of human ovarian adenocarcinoma cells named SKOV-3. ATP caused increase in intracellular Ca2+ concentration ([Ca2+]i) and, concurrently, it evoked a complex electrical response with a conspicuous outward component. This current was generated through P2Y2 receptor activation and opening of K+ channels, KCa3.1, as indicated by electrophysiological and pharmacological analysis, as well as by immunodetection and specific silencing of P2Y2 or KCa3.1 gene by esiRNA transfection. Low µM ATP concentration increased SKOV-3 cell migration, which was strongly inhibited by KCa3.1 channel blockers and by esiRNA-generated P2Y2 or KCa3.1 downregulation. Finally, in human ovarian tumors, the P2Y2 and KCa3.1 proteins are expressed and co-localized in neoplastic cells. Thus, stimulation of P2Y2 receptors expressed in SKOV-3 cells promotes motility through KCa3.1 activation. Since P2Y2 and KCa3.1 are co-expressed in primary tumors, our findings suggest that they may play a role in cancer progression.

Evidence supports a relationship between alterations in the purinergic or histaminergic signaling systems and the cancer process in several cell types 1,2 . Thus, stimulation of specific, ATP-sensitive membrane receptors, named P2 receptors, inhibits cell growth and/or promotes apoptosis in various cancer cells such as breast cancer 3 , cervical cancer 4 , glioma 5 , and prostate cancer 6 , among many others. However, purinergic stimulation might also have the opposite effect as it can promote cell proliferation, either in distinct cancer cell types or even in the same model when tested in different experimental conditions. These divergent effects have been thought to reflect ATP availability in the tumor environment together with a specific combination of purinergic membrane receptors expressed in a particular cell type 1 , and in addition, they would be strongly influenced by the expression of a distinctive set of effector proteins, such as G proteins, protein kinases, and membrane ion channels. Histaminergic signaling that is altered in cancer has also been proposed as an important paracrine and autocrine regulator of proliferation 2 , as well as a mediator of cancer progression acting on cell migration, angiogenesis, and modulation of the immune response.
Previous studies indicated that ion channel function might be one of the modifications suffered in cancer; their activation or inhibition, for example, affects various important functional processes in the context of cancer [7][8][9][10] . Altered expression of a diversity of K + channels in human breast cancer cells, in human astrocytomas and glioblastomas, and in human ovarian cells including SKOV-3 have been documented in distinct cell models 11,12 . Although ion channel activation through purinergic receptor stimulation is a well-known phenomenon, its role in cancer has not been thoroughly analyzed. Here, we undertake an analysis of the effects mediated by ATP (and histamine) on the electrical properties of human ovarian cancer cells named SKOV-3 13 , a well-studied cell model that expresses molecular markers of epithelial to mesenchymal transition, a phenomenon associated with tumor metastasis 14 .
SKOV-3 cells are endowed with P2 receptors of the two known subtypes: those forming receptor-channels named P2X 15 , as well as G protein-coupled receptors named P2Y. ATP application generates in SKOV-3 an increase in the intracellular Ca 2+ concentration ([Ca 2+ ] i ) via its release through P2 receptor stimulation 16 , and a similar [Ca 2+ ] i increase is evoked by histaminergic signaling activation; the effect of this [Ca 2+ ] i increase on membrane conductance, however, remains to be explored. On the other hand, the expression and function of K + channels correlate with the cancer progression in SKOV-3 cells, as some specific K + -channel subtypes, such as two-pore K + channels, are upregulated 17,18 .
Here, we carried out electrophysiological studies of SKOV-3 cells stimulated by ATP and other drugs, and found that specific stimulation of P2Y 2 receptors generated mainly an outward current response carried by K + and that this was mimicked by histamine. We also showed that the K Ca 3.1 channel activation was a prompt, electrical response to ATP or histamine and that it promoted SKOV-3 cell migration, while specific silencing of K Ca 3.1 or P2Y 2 gene downregulated protein expression and strongly reduced both the electrical response and cell motility. Finally, we provide evidence that both K Ca 3.1 channels and P2Y 2 receptors are expressed in SKOV-3 cells and in neoplastic cells in human ovarian tumor biopsies. Thus, we propose that K Ca 3.1 channels are important for the tumorigenic process, specifically by promoting cellular migration. This information suggests that K Ca 3.1 channels might be a useful target for the development of diagnostic and therapeutic strategies against ovarian cancer.

ATP triggers complex electrical membrane responses in SKOV-3 cells. Electrophysiological exper-
iments were made in SKOV-3 cells between passages 1 to 6, within a period of 48-72 h in culture. Single cells, or those with no more than 1 or 2 in contact, were chosen in order to avoid extensive cell coupling, although it has been shown previously that SKOV-3 cells have a low level of cell-to-cell coupling mediated through gap junction channels under similar culture conditions 19 . Thus, SKOV-3 cells presented a C m of 48.9 ± 1.5 pF and R m of 127.5 ± 17.8 MΩ (359 cells) as monitored in whole cell configuration using the standard internal and external solutions. SKOV-3 cells held at −40 mV a value close to their resting membrane potential (−38.8 ± 2.4 mV) were systematically tested for different agonists. ATP (1-100 µM) application elicited in 92.3 ± 3% of the cells a complex electrical response consistently composed of 3 main components (denoted components 1-3 in Fig. 1A), which was associated with an increase in membrane conductance: a fast, inward response of low amplitude (1) followed by a robust outward current (2) that developed together with a smaller, slower, and relatively less frequent inward current (3). Component 2 was the main membrane current elicited by ATP, and it had an average amplitude of 26.24 ± 2.5 pA/pF (n = 186 cells). Typically, the current response to ATP started to decrease after 2 to 3 applications with washing intervals of 3-4 min, and the response disappeared completely after 5 to 8 sequential applications (data not shown).
Other agonists were tested for their ability to activate membrane currents (Fig. 1B). Among them, UTP (n = 33) or histamine (n = 9) also generated current responses that mimicked mainly component 2 (12.91 ± 2.1 pA/pF and 7.3 ± 1.05 pA/pF; for UTP and histamine, respectively). It appears that current component 1 is specific for ATP. All other agonists, such as acetylcholine, adenosine or dopamine (Fig. 1B), failed to activate any response (4 cells in each case). Thus, these data strongly suggest that current responses were specific for purinergic and histaminergic receptors. We also explored the ATP sensitivity of the ovarian adenocarcinoma cell line named CaOV-3 (n = 10), where 10 µM ATP application generated similar current components 1 and 2, with an average amplitude of 0.47 ± 0.007 pA/pF and 17.4 ± 3.8 pA/pF, respectively. The trace shows a typical electrical response generated in SKOV-3 cells by ATP application; in cells held at −40 mV, three main components are identified in the response: 1) a fast inward current with low amplitude, 2) a robust outward current, and 3) a slow inward current. All response components were associated with an increase in membrane conductance. The bar graph shows the current density recorded for each component in a sample of 39 cells. (B) Traces show that component 2 was also activated by UTP or histamine application. All agonists were applied at 10 µM, and the bar graph summarizes the outward current density recorded in 60 cells tested for the various agonists. In general, in this and subsequent traces, responses were obtained in cells held at −40 mV, and upper bars indicate the drug application time in each case. Horizontal lines indicate zero current.
As illustrated in Fig. 2B, it was also demonstrated that the electrical response generated by ATP application was dose dependent; in these experiments, cells held at -40 mV were perfused with increasing concentrations of ATP within the range of 0.1 to 30 µM with wash intervals of 4-5 min. The peak response at each concentration was normalized against the maximal current that occurred around 3 µM. The dose-response curve gave an estimated EC 50 of 399 ± 11.4 nM for component 2 elicited by ATP.
Outward membrane response to ATP was mimicked by purinergic agonists. In order to distinguish among different P2 receptors, a battery of agonists was tested, all at 3 µM concentration (Fig. 3A). In Fig. 3B the bars illustrate the current response elicited by the different drugs, normalized against that obtained by ATP application in the same cell. This gave the following sequence of potency: ATP > UTPγS > UTP > ATPγS ≥ Ap4 A > 2-MeSATP ≫ Bz-ATP ≥ 5 Br-UDP ≥ ADP ≥ 2-MeSADP ≫> UDP; with no response to adenosine. All agonists tested generated mainly the outward current, although the more potent drugs, such as UTP, also elicited component 3 in a way similar to ATP, but not component 1. This strongly suggested that the receptor involved in generating the outward current was one of the P2Y subtypes sensitive to UTP with a pharmacological profile close to that displayed for the P2Y 2 subtype.
Three different P2Y-type receptors sensitive to UTP are known in different species 20,21 ; these are named the P2Y 2 , P2Y 4 and P2Y 6 receptors. Thus, an RT-PCR analysis was made to specifically amplify the transcripts for these receptor subtypes in SKOV-3 cells, and it showed that P2Y 2 and P2Y 4 were well expressed while P2Y 6 was not amplified (Fig. 3C). Immunocytochemistry against these receptors in SKOV-3 cells indicated that P2Y 2 and P2Y 4 proteins were also well expressed in 90% of cells tested (Fig. 3C). However, it is well known that the human P2Y 4 receptor is not activated by ATP 22 . All these data strongly suggested a main role of P2Y 2 in ATPor UTP-elicited outward current response of SKOV-3 cells. This was also supported by the inhibitory effect of AR-C118925 (10-30 µM), an antagonist with high specificity for P2Y 2 receptors 23 , that reduced 80-88% of the response elicited by ATP (3 µM) and more than 98% of the response to UTP (3 µM) (Fig. 3D,E); while NF340 or MRS2179, specific antagonists for P2Y 11 and P2Y 1 subtypes 21 , respectively, did not affect SKOV-3 responses generated by ATP.
Ionic basis for the outward current generated by ATP. Current-voltage (I/V) relationships were built by applying voltage steps in control conditions and at the peak of the outward current activated by ATP; the differences between these two relations were plotted as illustrated in Fig. 4A. The I/V curve showed that the outward current activated had a reversal potential (E rev ) of -95 ± 5 mV (n = 54), which corresponded closely to that for K + ions in the recording conditions. Similar voltage dependency was also observed in experiments where cells were held at different potentials and then tested by applying 3 µM ATP. As shown in Fig. 4B, component 2 of the response inverted polarity close to −100 mV. Using the voltage-stepping protocol and increasing the concentration of K + (substituted for Na + ) in the external solution, the current E rev shifted to less negative values; thus in 60 mM K + it was −20 ± 9 mV, and in 30 mM K + the E rev was −35 ± 5 mV. As shown in Fig. 4C the logarithmic relation of E rev versus extracellular K + concentration was a line with a slope of −59 mV, in agreement with the Nernst equation. Also, I/V curves were built during the development of component 3 after the complete wash out of component 2; in this condition inward currents had an E rev of −2 ± 1 mV (n = 10) and were not affected by T16Ainh-A01 (100 µM; n = 5), a specific Ca 2+ -dependent Cl − channel blocker.
Effect of K + channel blockers and expression of K Ca 3.1 channels. Consistent with a conducting pathway for K + ions during component 2, the ATP response was inhibited by 5 mM tetraethylammonium (TEA + ) or 5 mM 4-aminopyridine (4-AP), unselective K + channel blockers (18 cells for each blocker), by 67 ± 7% and 70 ± 4%, respectively (Fig. 5B). Similar results were obtained for outward currents activated by UTP. Together, these results indicated that purinergic activation opened a conductance selective for K + ions; thus, we tested various drugs that act selectively on K + channels that have been shown to either be expressed in SKOV-3 cells or be involved in tumor biology ( Fig. 4D-H). Specific blockers of two-pore K + channels known to be overexpressed in SKOV-3 cells 17,18 , such as 20 µM curcumin, 100 µM L-methionine, 100 µM TPenA (a TREK-1 and TREK-2 blocker), and 20 µM methanandamide (a TASK-3 blocker), showed only small inhibitory effects (ranging from 12 to 26%) on the K + current elicited by ATP (or UTP; n = 22) that were not significant when compared with the amplitude of control currents (Fig. 4D). However, application of 10 µM TRAM-34, a specific blocker for intermediate-conductance Ca 2+ -dependent K + channels 24 (K Ca 3.1), produced a strong inhibition of 83 ± 9% (n = 12) of the current response elicited by ATP or UTP (Fig. 4D,E), while 1 µM apamin, a potent blocker of the small-conductance Ca 2+ -dependent K + channel (K Ca 2) subtype, had a smaller inhibitory effect of 26 ± 10% (n = 12) that was not significant (Fig. 4D). Accordingly, the drug 1-EBIO (100-300 µM), a specific opener for K Ca 3.1 channels, produced the generation of outward currents in SKOV-3 cells in a manner that was dependent on drug concentration as well as on the free Ca 2+ concentration in the internal solution ( Fig. 4F-H) [25][26][27] . Also, it was noted that in cells where initial application of 300 µM 1-EBIO was not able to generate a response (6 out of The graph also shows results of using UTP as agonist and AR-C118925 as antagonist (n = 16; *p < 0.05).
18 cells), pre-stimulation with a low dose of 0.5 µM ATP (256.6 ± 61 pA) primed the cells to respond to a second application of 1-EBIO (396 ± 66.7 pA; Fig. 4H). These results indicated that purinergic stimulation of P2Y receptors elicited the activation of K + channels sensitive to TRAM-34 and 1-EBIO, which are specific drugs for K Ca 3.1 channels.
A Western blotting was performed to confirm K Ca 3.1 protein expression in SKOV-3 cells. The results are illustrated in Supplementary Figure S1A where it is shown that an antibody specific against this channel subtype gave a band with the expected weight of 55 KDa for the K Ca 3.1 channel protein. Immunocytochemical analysis of SKOV-3 cells also showed strong, specific label in most cells in the culture; a positive control using neurons from cortex 28 is shown in Supplementary Figure S1B    (D) A battery for distinct K + -channel inhibitors was tested on the response, including non-specific blockers of K + channels (TEA + and 4-AP), those specific for two-pore K + channels, and blockers apamin and TRAM-34 specific for K Ca 2 and K Ca 3.1, respectively; bars indicate the proportion of current inhibited by the different antagonists (*p < 0.05). (E) Traces show membrane current recorded during the outward peak response to ATP (upper traces) and that obtained in the same cell by co-applying ATP together with TRAM-34 (10 µM), a specific blocker of K Ca 3.1 channels (lower traces). The graph shows data (means ± SEM) obtained in 30 cells under the same protocol in which the basal currents were subtracted for each case, averaged and plotted. (F) Currents elicited by 1-EBIO (100 or 300 µM), a K + channel opener specific for K Ca 3.1 channels, applied at two different intracellular Ca 2+ concentrations: low and high (estimated concentration of 10 and 300 nM, respectively). The bar graph shows the averaged current density obtained in each condition monitored in 14 cells. (G) Traces illustrate a case in which initially, 1-EBIO was unable to generate any current (upper trace); however, after a response elicited by ATP, a second application of 1-EBIO (lower trace) at the same concentration generated a meaningful outward response. cells kept P2Y 2 receptor expression (Supplementary Figure S2). esiRNA-transfected SKOV-3 cells, including the control group, were electrophysiologically monitored for responses elicited by ATP or UTP, the more potent agonists; ADP, the weak agonist; and 1-EBIO, the K + -channel positive modulator (Fig. 5C). Both groups of esiRNA-transfected cells showed a strong decrease in electrical response elicited by drug application. A pool of 85 cells from 2-3 different transfections (24-72 h) (Fig. 5D,E) showed that current responses remained low at least for 72 h. For example, 24 h after P2Y 2 -esiRNA transfection, ATP response decreased by 90.65 ± 4.7%, UTP by 87.75 ± 6.5%, while ADP and 1-EBIO decreased by 83.15 ± 3.9% and 96 ± 3.1%, respectively. Similar results were obtained in K Ca 3.1-esiRNA transfected cells. It was consistently observed that knocking down P2Y 2 receptor also produced a strong decrease in 1-EBIO response (16.92 ± 7.4 pA; n = 19). Nevertheless, as that shown in Fig. 4H, after ATP (3 µM) application, a larger response was activated in P2Y 2 -esiRNA treated cells tested with a second 1-EBIO (300 µM) superfusion, although amplitudes were significantly decreased compared with control responses (96.65 ± 42.4 pA vs. 539.3 ± 75.1 pA; n = 15). This was not observed in K Ca 3.1-esiRNA transfected cells where a second 1-EBIO application, after ATP, was ineffective (not shown).
SKOV-3 silenced for P2Y 2 or K Ca 3.1 protein were fluorometrically monitored for [Ca 2+ ] i changes elicited by agonists (ATP, UTP, or ADP) (Fig. 6). The bar graph in Fig. 6A shows that when tested for ATP both groups of esiRNA-transfected cells, P2Y 2 or K Ca 3.1, showed a significant reduction of [Ca 2+ ] i increase compared with CNT cells. In the case of P2Y 2 -esiRNA, the reduction was of 69.3 ± 2.98% while for K Ca 3.1-esiRNA it was of 54.3 ± 4%. A similar result was obtained in cells tested with UTP, where [Ca 2+ ] i increase was reduced by 51 ± 4% and 98 ± 1.5%, respectively. Thus, downregulation of K Ca 3.1 alone in SKOV-3 cells unexpectedly seemed to affect [Ca 2+ ] i increase elicited by both agonists, suggesting an effect of membrane potential and/or Ca 2+ influx on the response. In addition, responses elicited by ADP were reduced compared with the control group, although changes were not statistically significant in such cases.
All these results indicated that downregulation of P2Y 2 or K Ca 3.1 eliminated the electrical response to ATP or UTP, and that strongly reduced the [Ca 2+ ] i increase regularly generated by the same agonists. K Ca 3.1 channel blockage and cell migration elicited by ATP. In several cell types, an important role attributed to K Ca 3.1 is its involvement in the migration phenomenon 29 for example in microglia 30 , glioblastoma 31 as well as in fibroblasts and melanoma cells 32 . First, we asked whether stimulation by ATP promoted cell migration in the concentration range that generated the current response. The results from the transwell migration assay are illustrated in Fig. 7A, where we quantified the migration of SKOV-3 cells for 16 h in the absence or presence of 0.3 µM, 0.6 µM or 3 µM ATP. We observed that migration of cells increased significantly in a dose-dependent manner in the presence of ATP at 0.3, 0.6 and 3 µM to 151.1 ± 19.8%, 160.7 ± 15.6%, and 181.6 ± 22.4%, respectively. Control experiments showed that ATP in the concentrations used did not affect cell proliferation or survival within the 16-h incubation period (using the MTS assay), in agreement with previous studies 16 . Migration of SKOV-3 cells was also stimulated by UTP and histamine, suggesting a close correlation of the effects exerted by receptors that generated a [Ca 2+ ] i increase, current generation and migration induction. Thus, we tested whether the K + current carried through K Ca 3.1 channels was also involved in the ATP-mediated migration increase. This is illustrated in Fig. 7C where it is shown that 100 µM 1-EBIO applied alone increased the phenomenon, while K + channel blockers applied together with ATP, such as TEA + , 4-AP or the specific blocker TRAM-34, inhibited cell migration compared to that of ATP alone. Moreover, the former blocker decreased migration with respect to basal level. Similar to the lack of ATP effect on proliferation and cell survival, TRAM-34 alone had no effect in control experiments with SKOV-3.
SKOV-3 cells that were esiRNA-treated were also tested for motility using a similar protocol; the results are illustrated in Fig. 7D. Groups of esiRNA-transfected cells were tested for migration with 3 µM ATP, 3 µM UTP or 300 µM 1-EBIO. The results showed that in P2Y 2 -or K Ca 3.1-esiRNA transfected cells, cell migration was strongly reduced by all drugs tested (compared with CNT group), including the group of P2Y 2 -esiRNA transfected cells tested with 1-EBIO. This suggested that P2Y 2 receptor was required for complete channel activation, a notion that was in agreement with our electrophysiological results. K Ca 3.1 and P2Y 2 receptor expression in biopsies of human ovarian tumors. It was also of importance to explore whether or not the main protein elements involved in the generation of the electrical and migration responses were expressed in tumor samples from human ovarian tumors. Figure 8 illustrates that K Ca 3.1 and P2Y 2 were expressed in biopsies of papillary serous carcinoma tumors (Patient 1, IC16-532-6) with a high degree of hypertrophy and hyperplasia. It is evident that human tissue expressed abundant K Ca 3.1 channel protein with preferential localization in neoplastic cells; nevertheless, some other structures were labeled as well, mainly stromal cells without apparent neoplastic phenotype. The P2Y 2 receptor signal was also localized primarily in areas where tumor cells presented K Ca 3.1 protein expression, and several regions showed clear co-localization. Similar results were obtained in biopsies from 5 more patients (Patients numbers IC16-4831-1, IC11-738, IC11-7619-3, IC16-1288, and IC16-1050-9; see Supplementary Figure S3).

Discussion
In this study we provide evidence that activation of specific receptors for ATP or histamine elicited an electrical response in SKOV-3 cells, a well-established cell model for human ovarian adenocarcinoma, with epithelial-like morphology. The electrical response was generated, most probably, as a consequence of the [Ca 2+ ] i increase generated by its release from internal stores, through the activation of specific receptors coupled to G proteins 16 . This idea is supported first, because a [Ca 2+ ] i increase was readily observed when the cells were stimulated by ATP or histamine at µM concentrations, as reported in previous studies, and also because the main electrical response corresponded to the generation of a Ca 2+ -dependent ionic current. The SKOV-3 response elicited by ATP and histamine had multiple phases; however, an outward current component was the most consistent and prominent. This response was mimicked by applying several purinergic agonists that, together, indicated participation of P2Y receptors sensitive to UTP. Nevertheless, comparatively small and fast inward responses elicited by ATP, but not by UTP (or ADP), seemed to indicate involvement of P2X channels in this particular response. The main UTP-sensitive P2Y receptors are the P2Y 2 , P2Y 4 and P2Y 6 subtypes, and expression analysis by PCR and immunocytochemistry showed that P2Y 2 and P2Y 4 were the main receptors expressed in SKOV-3 cells. However, given that the human P2Y 4 receptor is not activated by ATP 22 , the potency sequence for the distinct agonists  Biopsies from ovarian carcinoma tissue were collected, processed, and stained with hematoxylin-eosin and/ or labeled with K Ca 3.1-and P2Y 2 -specific antibodies that were revealed by second antibodies coupled to green and red fluorescent dyes, respectively. Nuclei in blue labeled with DAPI. Positive co-expression was detected in ovarian neoplastic cells (see merged image), whereas no signal was observed in control assays in which the primary antibody was omitted. Expression of K Ca 3.1 was also observed in stromal cells that did not show neoplastic morphology (*); however, P2Y 2 did not co-localize with K Ca 3.1 in these cells. Patient 1, IC16-532-6. supported a main role for P2Y 2 receptor involvement. For example, the P2Y 4 receptors expressed in humans are also highly sensitive to Ap4A, but in SKOV-3 cells this agonist was much less effective than ATP or UTP; moreover, pharmacological data excluded involvement of P2Y 6 since UDP was a weak agonist of the response 20 . Thus, nearly equipotent activation by ATP, UTP, and UTPγS, very low sensitivity to UDP, together with middle potency for Ap4A or ADP clearly indicated a main contribution of the P2Y 2 subtype in the electrical responses elicited by purinergic agonists in SKOV-3 cells. This result also agrees with the EC 50 of 400 nM for ATP found here for the outward response generated, since the range reported for the half-maximal effective concentration for P2Y 2 ranged in different cell systems from 100 to 500 nM 33,34 . The P2Y 2 antagonist, AC-R118925, strongly inhibited the response generated by ATP or UTP, while other antagonistic drugs with high specificity for P2Y1 or the P2Y11 subtypes had no effect on SKOV-3 response. It is also important to emphasize the lack of effect by transmitters such as adenosine, acetylcholine or dopamine, suggesting the specificity of the SKOV-3 responses to ATP and histamine. This result is compatible with studies demonstrating that SKOV-3 cells do not present either expression of muscarinic receptors or a Ca 2+ increase elicited by carbacholine, a muscarinic agonist 35 .
The outward-directed current response was carried mainly by K + ions, as demonstrated by monitoring its reversal potential, which was close to the potential estimated for K + ions in the recording conditions of −96.5 mV, as well as by changing the extracellular K + concentration, which shifted the reversal potential of the response as predicted by the Nernst equation. Moreover, outward current responses elicited by ATP were potently blocked by TEA + or 4-AP, two unspecific K + channel blockers. Treatment of cells with more specific drugs indicated that the main pathway involved did not correspond to activation of two-pore K + channels (TASK-3, TREK-1 and TREK-2), or to Ca 2+ -dependent K + channels sensitive to apamin, some of which are proposed to be related to the cancer process in SKOV-3 cells 17,18 . However, TRAM-34, a specific drug for Ca 2+ -dependent K + channels of intermediate-conductance corresponding to the K Ca 3.1 subtype, proved to be an effective blocker of the SKOV-3 response. Moreover, 1-EBIO, a specific positive modulator for this subtype of channels, showed an effect that was, as expected, dependent on the intracellular Ca 2+ concentration. Thus, the mechanism that generates the ATP response in SKOV-3 cells more likely involves the activation of P2Y 2 receptors, which produces a [Ca 2+ ] i increase and the subsequent opening of Ca 2+ -dependent K + channels sensitive to TRAM-34 and 1-EBIO; these K + channels are of the K Ca 3.1 subtype expressed in SKOV-3 cells, as confirmed by Western blot analysis and immunofluorescence. Also, apparently K Ca 3.1 channels were primed by P2Y 2 activation, facilitating their opening. This regulatory mechanism was not studied further here; however, it has been shown that K Ca 3.1 is post-translationally modulated by phosphorylation of a histidine residue in the C-terminal through nucleoside diphosphate kinase B 36 . A similar mechanism might be responsible for the effect observed during ATP-elicited response in SKOV-3 cells.
All functional and pharmacological evidence was also supported by results of specific downregulation of P2Y 2 or K Ca 3.1 protein expression using the esiRNA-transfection method. Silencing P2Y 2 or K Ca 3.1 gene produced a strong downregulation of the respective protein expression, which had a strong impact on generation of electrical response elicited by ATP or UTP that was eliminated. Moreover, a significant reduction in [Ca 2+ ] i increase elicited by these agonists was observed using fluorometric analysis. In average, reduction in [Ca 2+ ] i increase in cells transfected with P2Y 2 -esiRNA was not as robust as that observed for the electrical response, indicating that [Ca 2+ ] i increase was probably insufficient for K Ca 3.1 channel activation, and also probably to deficiencies in the P2Y 2 receptor mechanism that primed the channels. Another observation was that K Ca 3.1 esiRNA-transfected cells also showed a reduction in [Ca 2+ ] i response when ATP or UTP was applied. This might indicate a role of K + channels in [Ca 2+ ] i increase, probably through membrane potential regulation that would affect voltage-dependent mechanisms and/or reduce Ca 2+ influx, among several other possibilities. Similarly, it has been reported that downregulation of nucleoside diphosphate kinase B expression reduces K Ca 3.1 channel activity in CD4 T cells and decreases Ca 2+ influx 36 .
Furthermore, E rev obtained during the development of the inward current component seems to indicate the flux of Cl − ions. However, in preliminary results a specific blocker for Ca 2+ -dependent Cl − channels of the TMEM16-A subtype did not inhibit this component; hence, further studies are necessary to define its nature and molecular identity. Finally, a similar current response was observed in the ovarian carcinoma cell line CaOV-3.
A relationship between K Ca 3.1 channel activation and the process of cell migration has been documented in both normal and pathological conditions 29 . Therefore, we have tested whether or not opening K Ca 3.1 channels with ATP was an effective way to increase SKOV-3 migration at the low ATP concentrations that activated the electrical response. The results showed that 3 µM ATP application promoted SKOV-3 cell migration and that this was also blocked by TRAM-34 and mimicked by 1-EBIO. A strong inhibition of cell migration was also achieved by downregulation of either P2Y 2 or K Ca 3.1 protein, an effect that was observed on motility promoted by either ATP, UTP or 1-EBIO. This strongly suggests that an increase of extracellular ATP in the low µ-molar range would have an important consequence in the tumor microenvironment; an effect that would be influenced by a concomitant increase in histamine, which might act directly or indirectly on specific receptors (e.g., by promoting the release of ATP as shown in other cell systems) 37 .
Together, the results presented here support the idea that an increase in ATP concentration within the tumor microenvironment might alter the function of channels such as K Ca 3.1. This would promote cell motility, an effect possibly potentiated by several factors including an increase in the expression of the molecules involved, specifically P2Y 2 receptors and K Ca 3.1 channels in neoplastic cells. In fact, in distinct cancer cell types it has been shown that cell motility in general is highly dependent on [Ca 2+ ] i , acting through a complex machinery of molecules which are, in most cases, Ca 2+ -dependent 29 . Among the molecules involved, those that allow the flux of ions through the cell membrane are essential for the process. Both Ca 2+ -dependent K + channel and Clchannel activation are required for cell motility, allowing the needed movement of water across the membrane and the cell volume changes, which are membrane mechanisms that are also active during the process of metastasis 10,38 . There is evidence indicating that this might be the case in some cancer cell types specifically regarding K Ca 3.1 channels; for example, brandykinin activation of K Ca 3.1 channels in human glioma cells promotes their migration in both in vitro and in vivo models 31,39 . In addition, elevated levels of K Ca 3.1 expression in breast cancer cells 9 and cell renal carcinoma 40 correlate with tumor grade and metastatic status. Here, we also show that molecules involved in the SKOV-3 electrical response were robustly expressed in biopsies from human ovarian tumors. A strong expression of both P2Y 2 receptors and K Ca 3.1 channels was observed in human ovarian tumors, and they specifically co-localized in neoplastic cells. Thus, it is proposed that opening of K Ca 3.1 channels by ATP (and histamine) in human ovarian cancer cells might be one of the mechanisms deregulated during the cancer process, possibly explaining the effects that this transmitter has during tumorigenesis.
It is well known that K Ca 3.1 channels play a main role in the migration of microglial cells, as part of the normal immune response in the nervous system 30 , as well as in human dendritic cells 41 and in activated human T cells 42 . In this context, there are physiological conditions in which ovarian cells might require the capacity to migrate; for example, during ovulation the ovarian superficial epithelium suffers a rupture that allows oocyte release, which occurs through a mechanism similar to an acute inflammatory reaction 43 . After the gamete is released the ovarian surface wound is repaired by the epithelium. The latter process requires cell migration, as indicated by the profile of genes expressed during ovulation, that shows a strong correlation with important epithelial functions, such as the inflammation reaction, angiogenesis, extracellular matrix remodeling and cell-to-cell contact 44 . Both ATP and histamine signaling have been proposed to be involved in ovulation 15 ; in this manner, transmitter-activated membrane ionic currents in the ovarian superficial epithelium might participate during this specific physiological condition, a hypothesis that requires further studies. Nevertheless, an important role for K Ca 3.1 in general epithelial secretory activity is expected, as suggested by its broad expression in most of the epithelial cells analyzed 45 .
In summary, in this study we show that purinergic and histaminergic stimulation generated electrical responses in SKOV-3 cells through the opening of membrane ion channels. This has two important implications: first, two main transmitters that are commonly found in increased concentrations in the tumoral microenvironment exerted direct actions on membrane conductivity pathways; and second, activation of one of these pathways, the K Ca 3.1 channel subtype, was directly involved in SKOV-3 cell motility, an important phenomenon in cancer. Since the P2Y 2 receptor and the K Ca 3.1 channel are co-expressed in neoplastic cells from human ovary, we propose that they may be useful tumor markers as well as targets for therapy to halt ovarian cancer progression.

Methods
Cell cultures. The human ovarian cancer cell line SKOV-3 was purchased from American Type Culture Collection (ATCC; Manassas, VA, USA). Cells were cultured at 37 °C in a humidified atmosphere containing 95% air: 5% CO 2 in RPMI1640 growth medium with L-glutamine (Mediatech, Manassas, VA, USA), supplemented with 10% fetal bovine serum (FBS; Gibco, Waltham, MA, USA) and 1% antibiotic-antimicotic mix (streptomycin, penicillin and amphotericin B; Life Technologies, USA). SKOV-3 cells were plated onto 12-mm-diameter coverslips in 12-well culture dishes and after 48 h in culture were used for electrical recording or other experimental protocols such as migration assays or specific protein detection by Western blot or immunolabeling. In some experiments, human ovarian adenocarcinoma CaOV3 cells obtained from ATCC were cultured in a similar manner for electrophysiological recording. As control cells, cortical neurons were cultured from brain cerebral cortex of E18 Sprague-Dawley rat embryos 46 . Briefly, neurons were dissociated enzymatically and mechanically, seeded on poly-D-lysine-coated coverslips and maintained in neurobasal medium supplemented with 10% FBS. Cultures were processed for immunocytochemistry after 7 days in vitro.
Electrophysiology. Whole-cell recordings were performed at room temperature (23-26 °C) using the Axon 200B patch-clamp amplifier (Axon Instruments; Sunnyvale, CA, USA). Currents were regularly recorded at a holding membrane potential of −40 mV, digitized and stored using the A/D converter Digidata 1400 and pClamp 10 software (Axon Instruments) for subsequent analysis. The extracellular bath solution adjusted to pH 7.4 contained the following (in mM): 140 NaCl, 3 KCl, 1 CaCl 2 , 1 MgCl 2 , and 10 HEPES. Patch-clamp pipettes (3)(4)(5) were filled with internal solution adjusted to pH 7.4 containing (in mM): 130 KCl, 5 NaCl, 2 EGTA, 1 MgCl 2 , 10 HEPES, 2 Mg-ATP, and 0.2 Na-GTP. The estimated Ca 2+ concentration in this solution was 10 nM. In some experiments, the estimated free Ca 2+ concentration was increased to 300 nM by changing the EGTA concentration to 2 mM and adding 1.5 mM CaCl 2 . External solutions with various K + concentrations were prepared by equimolar substitution of NaCl by KCl to 10, 30 or 60 mM or the KCl concentration was decreased to 1 mM without compensation. Agonists and other drugs were added to the external solution from stock solutions to reach the desired dilution and applied through superfusion. In most cases, peak currents generated at -40 mV by drug superfusion were used in the analysis. Current-voltage (I/V) relationships were built by changing the membrane potential from −120 to + 60 mV in 20-mV steps (150 ms) while the cells were held at −40 mV, and the peak membrane current values at the beginning (20 ms) of each step were plotted as in Fig. 4. In some other cases (Fig. 2B) the membrane potential was held at a desired value while drugs were superfused. Reverse Transcription Polymerase Chain Reaction. Total RNA from SKOV-3 cells was purified using the guanidine isothiocyanate method. First strand cDNA was synthesized using 2 μg of DNase-treated RNA as template, 1 µg of oligo(dT), 1.5 µg of random hexamers, and reverse transcriptase. The cDNA was used as template in a polymerase chain reaction to amplify cDNA fragments for β-actin, P2Y 2 , P2Y 4 , and P2Y 6 transcripts. All the PCR programs started at 96 °C for 2 min and finished at 72 °C for 5 min. The amplification cycles consisted in 40 s at 96 °C, 40 s at the specific annealing temperature for each primer set, and 40 s at 72 °C.
Western blot. Cultured SKOV-3 cells were scraped in Laemmli buffer and boiled for 5 min. For electrophoresis, samples were fractionated in a 10% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane (BioRad; Hercules, CA, USA). Membranes were blocked for 1 h at room temperature in 150 mM NaCl, 20 mM Tris, pH 7.4, and 0.1% Tween 20 (TBS-T) containing 5% nonfat dry milk and then incubated overnight at 4 °C with the appropriate mouse monoclonal antibody (1:1000) directed against the K Ca 3.1 channel protein (ALM-051, Alomone; Jerusalen, Israel). After washing with TBS-T, membranes were incubated for 1 h at 37 °C with HRP-conjugated goat anti-rabbit antibody (Zymed; Grand Island, NY, USA) in TBS-T. The immunoreactive proteins were detected by chemiluminescence, and images were analyzed with ImageJ Software.
Migration assay. Transwell assays were performed using 12-well plates containing 8-mm polyethylene hanging cell culture inserts (Millipore; Billerica, MA, USA). The cells were seeded at the apical side of the chamber. The basolateral side was filled with medium supplemented with drugs according to the particular experimental conditions. After 16 h of incubation with drugs, cells attached on the lower face of the inserts were fixed with 4% paraformaldehyde in PBS for 10 min and stained with 10 µg/ml propidium iodide. The samples were visualized under a microscope, pictures were taken and images were analyzed using ImageJ software (NIH) for quantification.
Cell proliferation assay. To analyze cell proliferation and viability, mitochondrial activity of the whole culture was assessed by using the 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2 (4-sulfophenyl)-2H-tetrazolium salt (MTS) assay (Promega; Wisconsin, USA). For this, cells were cultured in 96-well plates in RPMI-1640-10% FBS, and after 24 h they were transferred to serum-free RPMI-1640 for 8 h. After this, an appropriate stimulus was applied, and the cultures were incubated for an additional 16 h. Finally, the MTS assay was performed as described by the manufacturer. Results were quantified as fold increase of absorbance in experimental conditions relative to unstimulated cells. esiRNA transfection. SKOV-3 cells were plated at 4 × 10 4 cells per well in 24-well plates and allowed to attach overnight. Endonuclease prepared small interfering RNAs (esiRNAs) were commercially synthesized (Sigma-Aldrich, St. Louis MO, USA) using either 420 bp length of P2Y 2 receptor (EHU156731; P2Y 2 -esiRNA) or 472 bp length of K Ca 3.1 channel gene (EHU035251; K Ca 3.1-esiRNA), targeting human sequences NM_002564 and NM_002250, respectively. Then, for P2Y 2 receptor or K Ca 3.1 channel knockdown, cells were transfected with 150 ng per well of P2Y 2 -esiRNA or K Ca 3.1-esiRNA, respectively, by using Lipofectamine 3000 (Invitrogen, Grand Island NY, USA) to deliver esiRNA into the cells following the method indicated by the manufacturer. As a non-targeting control (referred to as CNT), was transfected in the same condition the esiRNA of enhanced green fluorescent protein (EHUEGFP, Sigma-Aldrich), all responses of transfected SKOV-3 cells with either P2Y 2 -esiRNA or K Ca 3.1-esiRNA were compared versus those of the CNT group. Twenty-four to 72 h after transfection, esiRNA-treated SKOV-3 cells maintained in culture were recorded electrically, or used for immunocytochemistry or fluorometric assays, as well as for migration quantification using the methods described above.
Immunohistochemistry in human ovarian cancer biopsies. To explore whether human ovarian tumors express P2Y 2 receptor and K Ca 3.1 channel proteins, biopsies of patients with ovarian carcinoma were analyzed by immunohistochemistry. The samples were obtained from the Instituto Nacional de Cancerología (INCAN) México, where clinical histories of all patients are archived, in accordance with ethical procedures approved by the Bioethics Committee. The samples used in the present study correspond to 6 patients diagnosed with ovarian carcinoma. The characteristics of the carcinomas analyzed are as follows: Patient 1 (IC16-532-6), 44 years of age diagnosed with low-grade papillary serous carcinoma; Patient 2 (IC16-4831-1), 61 years of age diagnosed with high-grade papillary serous carcinoma; Patient 3 (IC11-7381), 56 years of age with endometrioid G3-type carcinoma; Patient 4 (IC11-7619-3), 70 years of age diagnosed with high-grade serous carcinoma; Patient 5 (IC16-1288), 37 years of age with endometrioid G2-type carcinoma; and Patient 6 (IC16-1050-9), 57 years of age diagnosed with high-grade papillary serous carcinoma. Paraffin-embedded human ovarian biopsies were cut at 10-µm intervals and the slices attached to gelatinized slides. Paraffin was eliminated by incubating with xylene, and slices were rehydrated by passing them through a series of five concentrations of ethanol (from 100% to 50%) and washed with PBS. Antigens were exposed by incubating in 10 mM sodium citrate, pH 6, for 15 min and then equilibrated in PBS. For immunohistochemistry, samples were incubated overnight with the appropriate antibodies (1:80 anti-K Ca 3.1, and 1:100 anti-P2Y 2 , both from Alomone, Jerusalem, Israel), and non-specific sites were blocked with 3% BSA in PBS. The next day, samples were washed with PBS and incubated for 1 h at room temperature with anti-mouse Alexa Fluor 488 (1:100; Jackson ImmunoResearch, West Grove, PA, USA) or anti-rabbit Cy5 (1:100; Life Technologies, Carlsbad, CA, USA). After three washes with PBS, samples were stained with 1:1000 DAPI (Molecular Probes). Finally, the samples were mounted on VectaShield and analyzed by confocal microscopy.
Tissue samples were also stained with the hematoxylin and eosin technique, and sections were visualized and analyzed under a microscope; representative images were acquired with a Leica ICC50 HD (Leica Microsystems, Wetzlar, Germany). Statistical Analysis. AII data are expressed as mean ± S.E.M. The means of two groups were compared using a Student's t-test, or when appropriate, by analysis of variance followed by post-hoc comparisons of individual means using the Bonferroni correction. Statistical analysis was performed using GraphPad Prism software. Differences were considered to be significant at P < 0.05.