KCa3.1-Dependent Hyperpolarization Enhances Intracellular Ca2+ Signaling Induced by fMLF in Differentiated U937 Cells

Formylated peptides are chemotactic agents generated by pathogens. The most relevant peptide is fMLF (formyl-Met-Leu-Phe) which participates in several immune functions, such as chemotaxis, phagocytosis, cytokine release and generation of reactive oxygen species. In macrophages fMLF-dependent responses are dependent on both, an increase in intracellular calcium concentration and on a hyperpolarization of the membrane potential. However, the molecular entity underlying this hyperpolarization remains unknown and it is not clear whether changes in membrane potential are linked to the increase in intracellular Ca2+. In this study, differentiated U937 cells, as a macrophage-like cell model, was used to characterize the fMLF response using electrophysiological and Ca2+ imaging techniques. We demonstrate by means of pharmacological and molecular biology tools that fMLF induces a Ca2+-dependent hyperpolarization via activation of the K+ channel KCa3.1 and thus, enhancing fMLF-induced intracellular Ca2+ increase through an amplification of the driving force for Ca2+ entry. Consequently, enhanced Ca2+ influx would in turn lengthen the hyperpolarization, operating as a positive feedback mechanism for fMLF-induced Ca2+ signaling.

fMLF-dependent effects in macrophages and neutrophils are mediated by an increase in intracellular calcium concentration ([Ca 2+ ] i ) [7][8][9] and by changes in the membrane potential mM): NaCl 140, KCl 5, MgCl 2 1, CaCl 2 2, HEPES 10, D-glucose 10 and pH 7.4 adjusted with NaOH. The low Na + bath solution used contained (in mM): NaCl 5, NMDGCl 100, KCl 5, MgCl 2 1, CaCl 2 2, HEPES 10, sorbitol 50, D-glucose 10 and pH 7.4 adjusted with NaOH. The 105 mM K + solution contained (in mM): NaCl 5, KCl 105, MgCl 2 1, CaCl 2 2, HEPES 10, sorbitol 50, D-glucose 10 and pH 7.4 adjusted with NaOH. The bath solution without Ca 2+ was prepared adding 5 mM EGTA in absence of CaCl 2 . The 65 mM K + solution was prepared adding 65 mM KCl and 80 mM NaCl. Bath solutions were changed by a gravity-fed perfusion system and the solution level in the chamber was kept constant by a peristaltic pump. Voltage-and current-clamp experiments were recorded with EPC-7 amplifier (List-Medical, Darmstadt, Germany). Data were digitized at 10 kHz and low-pass filtered at 1 kHz. pClamp 8.0 software (Molecular Device Corp., Sunnyvale, CA, USA) was used for data acquisition and analysis. Patch electrodes (4 MO resistance) were pulled from borosilicate glass (Warner Instruments) using a BB-CH puller (Mecanex SA, Geneva, Switzerland). All experiments were performed at room temperature.

Calcium measurements
Intracellular Ca 2+ concentration changes were measured using the Ca 2+ indicators Fluo-3 AM, Fluo-4 AM and Indo-1 AM (Invitrogen, Waltham, MA, USA), as indicated in the figures. U937 cells after 48 h of differentiation were incubated with 2 μM of the given indicators dissolved in RPMI 1640 without FBS for 30 min in the dark at 37°C. Cells were then kept for 30 min in the presence of RPMI 1640 with 10% FBS in the dark. U937 cells were washed 2 times with PBS and suspended or bathed with the same solutions used and described for the electrophysiological measurements. Cells were exposed to the different drugs, [K + ], EGTA or control conditions for approximately 5 min before starting the fluorescence measurements. fMLF was perfused throughout the experiments as shown in each figure. Fluorescence was measured using three different systems depending on the type of experiment and Ca 2+ indicator used. Sinergy 2 multi-mode microplate reader (BioTek, Winooski, VT, USA) was used to measured fluorescence changes in 50-100x10 3 U937 cells for each condition and acquisition was made using Gene5 Data Analysis software (BioTek). Disk Spinning Confocal Unit (Olympus Corp., Tokyo, Japan) was used to quantify fluorescence changes in the cytoplasm of U937 cells and Cell v2.8 software (Olympus Corp.) was used for acquisition. Two fluorometers were used to measured fluorescence changes in U937 cells incubated with Indo-1 AM and acquisition was performed with AxoScope 8.2 software (Molecular Devices). For Fluo-3 and Fluo-4 the cells were excited at 485 nm and emission was detected at 520 nm. Fluorescence changes were expressed as (Ft-F0)/F0, where Ft was the fluorescence at any time and F0 was the basal fluorescence during 2 min before stimulus. For Indo-1 cells were excited at 340 nm and the acquisition was made at 405 nm and 485 nm. Fluorescence changes were expressed as the ratio 405/485 and the results were expressed as R1/R0, where R1 was the ratio at any time and R0 was the basal ratio during 2 min before stimulus.

Reagents
All reagents were of analytical grade and were purchased from Sigma (St. Louis, MO, USA) and Merck (Darmstadt, Germany). fMLF and dibutyryl cAMP were purchased from Sigma (St. Louis, MO, USA) and dissolved in DMSO.

Statistics
Data are represented as mean ± SEM. Unpaired Student's t test was used to compare between two groups. For comparing three or more groups, one-way analysis of variance (ANOVA) followed by Bonferroni's post-hoc test was used. Statistical significance was set at p < 0.05.

fMLF increases [Ca 2+ ] i and hyperpolarizes U937 cells
To monitor changes in [Ca 2+ ] i induced by fMLF, differentiated U937 cells were loaded with Fluo-3. As depicted in Fig 1A, fMLF (100 nM) in the presence of 2 mM external Ca 2+ induced a rapid increase in intracellular Ca 2+ followed by a slower decay phase. Under 0 external Ca 2+ conditions (no Ca 2+ added + 5 mM EGTA) the slower decay phase disappears. As shown in Fig 1B left panel, normalized peak fMLF-induced increase in [Ca 2+ ] i in the absence of external Ca 2+ was not different to that observed with a solution containing 2 mM Ca 2+ . However, the normalized area under the curve (AUC) for Ca 2+ increase was 94% reduced in the absence of external Ca 2+ (Fig 1B right panel).
As previously reported, fMLF hyperpolarizes the transmembrane potential (V m ) in macrophages [10]. Thus, we asked whether fMLF induces changes in V m in differentiated U937 cells. Fig 1C depicts representative current clamp records in the presence of 2 mM external Ca 2+ (left panel) or 0 external Ca 2+ conditions (right panel). As shown, cells responded with a rapid (Ca 2+ : -0.131 ± 0.019 mV/ms; 0 Ca 2+ : -0.079 ± 0.005 mV/ms) hyperpolarization upon exposure to 100 nM fMLF reaching~-70 mV. Despite continuous exposure to fMLF V m repolarizes rapidly after approximately 8 min. However, in the absence of external Ca 2+ the duration of the hyperpolarization was 4-fold shorter (left panel) without a significant difference in the peak V m reached. Fig 1D shows an average of the data from experiments similar to those shown in Fig 1C. Of note, in non-differentiated U937 cells (i.e. not treated with dibutyryl cAMP), fMLF was unable to induce changes in intracellular Ca 2+ and V m (S1 Fig), suggesting that the observed effect is dependent upon expression of the FPR as it has been shown that non-differentiated U937 cells do not express them [24].
To determine whether the [Ca 2+ ] i increase and hyperpolarization induced by fMLF followed a similar time course, differentiated U937 cells were exposed to fMLF (100 nM) in the absence of external Ca 2+ for >5 min, then bathed with a solution containing 2 mM Ca 2+ . As depicted in Fig 1E and 1F, transient increase in [Ca 2+ ] i and hyperpolarization followed a similar temporal course, suggesting that both cellular responses might be linked by a common molecular mechanism.
fMLF-dependent hyperpolarization is secondary to a Ca 2+ -dependent K + conductance Because fMLF treatment changed V m to a value compatible with the reversal potential for K + (E K +), we studied whether a K + conductance underlies the fMLF-induced hyperpolarization. Fig 2A shows a representative voltage-clamp record of an experiment performed to determine the ionic nature of the hyperpolarizing current. To that end, differentiated U937 cells were stimulated with 100 nM fMLF and the current was recorded at E K + (-85 mV) and E Cl -(4 mV), alternately. Therefore, if fMLF activates a K + current, this current has to be observed at 4 mV. We observed an outwardly directed current of 33.20 ± 4.59 pA/pF at 4 mV. Furthermore, this current was completely abolished when cells were exposed to a bath solution containing 105 mM K + , which modified E K + to a value close to 4 mV. Next, to confirm whether a K + conductance underlies this hyperpolarization, we performed current clamp experiments, as shown by a representative current clamp record in Fig 2B. Cells where first exposed to 100 nM fMLF in a high-K + (65 mM) external solution and then to a normal-K + (5 mM) external solution. As shown, fMLF produced a small and brief hyperpolarization in high-K + conditions, whereas under normal-K + conditions a robust hyperpolarization was recorded. Furthermore, experiments performed with external tetraethylammonium (TEA 20 mM) and 140 mM internal Cs + Because the fMLF-induced hyperpolarization was found to be dependent on [Ca 2+ ] i as well as the similarity of the time course of Ca 2+ and V m changes, we hypothesized that the K + conductance activated upon fMLF exposure should be also Ca 2+ dependent. Fig 2C and   K + current. In the absence of external Ca 2+ , the K + current was briefly activated by fMLF (100 nM) and was reactivated upon reintroduction of 2 mM Ca 2+ (Fig 2C). In addition, after depletion of intracellular Ca 2+ reservoirs with thapsigargin (1 μM), activation of the K + current upon fMLF exposure was not detected in the absence of external Ca 2+ . However, upon external Ca 2+ reintroduction, the K + current was rapidly activated (Fig 2D).
K Ca 3.1 is the K + conductance activated by fMLF in U937 cells Immune cells express the intermediate and small-conductance Ca 2+ -dependent potassium channels (K Ca ) [19,20,[25][26][27]. To determine which K Ca is activated by fMLF in U937 cells, we used a pharmacological approach. The intermediate-conductance K Ca channel (K Ca 3.1) is blocked specifically and with high affinity (nanomolar range) by TRAM-34 and non-specifically by charybdotoxin and clotrimazole and is potentiated by DC-EBIO [28]. On the other hand, the smallconductance K Ca channels (K Ca 2.1, K Ca 2.2 and K Ca 2.3) are insensible to TRAM-34, charybdotoxin and clotrimazole, although enhanced by DC-EBIO [28]. Fig 3A show current traces recorded with a voltage-clamp ramp protocol (-100 to 100 mV) in U937 cells. As depicted (left panel), the current elicited by fMLF (E rev -78.5 mV; E K + -85 mV) was blocked (64%) by 100 nM TRAM-34 and potentiated (35%) by 10 μM DC-EBIO (right panel), suggesting that the molecular entity underlying fMLF-induced K + conductance in these cells is K Ca 3.1. Fig 3B summarizes the effect of K Ca blockers and DC-EBIO on the K + current elicited by fMLF in U937 cells. We next tested whether intracellular Ca 2+ could directly activate currents in U937 cells. To that end, conventional whole cell experiments were performed with 1 μM Ca 2+ in the pipette solution. Fig  3C depicts representative current traces elicited by a ramp protocol (-100 to 60 mV) in the presence and absence of 10 μM clotrimazole. As shown, clotrimazole inhibited the Ca 2+ -activated current, confirming the presence of a Ca 2+ -activated K + current in these cells. For unambiguous identification of the molecular identity of this Ca 2+ -activated K + current, we performed knock down experiments. Using the same experimental conditions as described above, lentiviral transduced cells with either a control sh (scrambled sequence) or a set of three K Ca 3.1-targeted sh were used. As depicted in Fig 3D and 3E, the Ca 2+ -activated current K + induced by 1 μM intracellular Ca 2+ was decreased in cells carrying the shK Ca

Discussion
In this study, we demonstrate that fMLF induces in differentiated human U937 cells a Ca 2+dependent hyperpolarization via activation of the K + channel K Ca 3.1. In addition, we found that this hyperpolarization enhances fMLF-induced intracellular Ca 2+ increase through an amplification of the driving force for Ca 2+ entry. Consequently, enhanced Ca 2+ influx would in turn lengthen hyperpolarization, operating as a positive feedback mechanism for fMLFinduced signaling.
Our data demonstrate that plasma membrane hyperpolarization induced by fMLF in U937 cells is due to the activation of a K + current as V m reaches -71 mV, a value close to the E K + (-85 mV). In addition, when [K + ] e is increased to 65 mM, which changes E K + to -20 mV, the magnitude of the hyperpolarization is significantly decreased. We also found that this K + conductance corresponds to a Ca 2+ -dependent K + channel because the activation of the current follows the same temporal course of the intracellular Ca 2+ increase. Also, absence of Ca 2+ influx and inhibition of Ca 2+ release from intracellular reservoirs hindered the activation of the nM: 40 ± 9%; clotrimazole, 10 μM: 9 ± 3%; and DC-EBIO, 10 μM: 135 ± 10% of the control current; n = 3-5 for each condition, one-way ANOVA, Bonferroni's post-hoc test, *** p < 0.001). C. Representative currents traces of the Ca 2+ -activated current (1 μM Ca 2+ in the pipette) elicited by a ramp protocol (-100 to 60 mV) under control conditions (pre) and after exposure to 10 μM clotrimazole (post). D. Representative currents traces of the Ca 2+ -activated current (1 μM Ca 2+ in the pipette) elicited by a ramp protocol (-100 to 60 mV) in sh-control (shControl) and sh-K Ca 3.1 U937 cells (shK Ca 3.1). E. Summarized data showing the effect of K Ca 3.1 knock down in Ca 2+ -induced K + currents (shControl: 11.7 ± 1.3 pA/pF; shK Ca 3.1: 5.8 ± 1.5 pA/pF, n = 3-4 for each condition, respectively, * p < 0.05). F. Summarized data of the effect of K Ca 3.1 knock down on fMLF-induced (100 nM) changes in membrane potential measured using the nystatin perforated patch-clamp technique (shControl: -62.3 ± 6 mV; shK Ca 3.1: -23.5 ± 3.8 pA/pF, n = 4 for each condition, ** p < 0.01).
doi:10.1371/journal.pone.0139243.g003 fMLF-induced K + conductance. Finally, based on the pharmacological profile (inhibition by TRAM-34 and potentiation by DC-EBIO) [28,29] as well as functional knock down we identified, for the first time, K Ca 3.1 as the molecular entity responsible for fMLF-induced K + current in macrophage-like differentiated U937 cells. fMLF-induced Ca 2+ increase occurs via release from intracellular reservoirs and Ca 2+ influx. Several studies report that upon fMLF binding to its receptor Ca 2+ is released from intracellular reservoirs trough IP 3 receptor activation [4,7,12]. This release promotes an initial, fast and transient increase in [Ca 2+ ] i followed by a sustained phase dependent on Ca 2+ influx. Although we did not explore the molecular entities underlying fMLF-induced Ca 2+ entry, it is interesting to note that clotrimazole is almost 3-fold more effective in reducing fMLF-induced Ca 2+ increase than TRAM-34, a value similar to that observed under 0 external Ca 2+ plus EGTA. The robust effect of clotrimazole on the Ca 2+ increase (S4 Fig) could be explained by the fact that clotrimazole besides blocking K + conductances also blocks TRPM2, known to be expressed in U937 cells [30], in the same dose range [28,31]. In addition, store operated channel entry (SOCE) might play a role in the influx of Ca 2+ , as K Ca 3.1 and SOCE are known to be tightly coupled in human macrophages [32].
Immune cell function depends on the magnitude and duration of the increase in the [Ca 2+ ] i [12,33]. The driving force for Ca 2+ entry is a critical factor to boost Ca 2+ signaling. In our experiments, fMLF shifts V m from -13 mV to -71 mV, implying roughly a 44% increase in the driving force for Ca 2+ entry. Because K Ca 3.1 is responsible for the hyperpolarization and, consequently, the increase in the driving force, the inhibition of this channel significantly decreases intracellular Ca 2+ increase. This decrease occurs only in the presence of external Ca 2+ , indicating that Ca 2+ entry is decreased by a reduction in the driving force. In fact, we observed that in the presence of external Ca 2+ K Ca 3.1 inhibition by 1 μM TRAM-34 decreased fMLF-induced Ca 2+ increase by 29% (Fig 4B). Interestingly, a similar result was observed in microglial cells, in which K Ca 3.1 inhibition by TRAM-34 (1 μM) was found to reduce the increase of Ca 2+ i triggered by extracellular UTP in approximately 30% [22].
As mentioned above K Ca 3.1 has a role in the Ca 2+ signaling induced by fMLF. Thus, the maximal effect of the K Ca 3.1 blocker on Ca 2+ signaling should be observed at a concentration sufficient to abolish fMLF-dependent hyperpolarization. However, we observed that to block the fMLF-induced hyperpolarization 0.1 μM TRAM-34 was enough, while the increase in the [Ca 2+ ] i triggered by fMLF was only significantly diminished with 1 μM TRAM-34. Despite the inconsistency regarding the concentration required to block fMLF-induced hyperpolarization and Ca 2+ increase, our results do support a relevant role of K Ca 3.1. It should be noted that fMLF-induced Ca 2+ signaling was diminished by 49% when hyperpolarization was prevented by increasing [K + ] e , which confirms that half of the increase in the [Ca 2+ ] i is mediated by a K +dependent hyperpolarization. Also, although TRAM-34 at 1 μM is still considered a specific blocker of K Ca 3.1, at this concentration TRAM-34 could block other voltage-gated K + channels, which have an IC 50 of approximately 5 μM [29]. Therefore, it can be concluded that the effect of TRAM-34 on Ca 2+ signaling is mediated by K Ca 3.1, although we can not discard the role of other K + channels.
It has been previously shown that K Ca 3.1 plays several roles in the immune response [34][35][36]. For example, K Ca 3.1 facilitates mast cell degranulation in mice [35], migration of dendritic cells [34] and promotes atherogenesis in mice due to an infiltration by macrophages and T lymphocytes in plaques [36]. All of the above mentioned processes share a common mechanism; an enhancement in Ca 2+ i increase by boosting the driving force for Ca 2+ entry. Therefore, our study suggests that K Ca 3.1 might enhance chemotaxis, phagocytosis, reactive oxygen species as well as cytokine production in monocytes and macrophages upon stimulation with chemotactic peptides.
In summary, our data indicate that K Ca 3.1 is the molecular entity responsible for the hyperpolarization observed upon fMLF exposure in differentiated U937 cells, thereby controlling the driving force for Ca 2+ entry and thus, modulating Ca 2+ i signaling in these cells through a positive feedback mechanism.
Supporting Information