BK_Ca channel regulates calcium oscillations induced by alpha-2-macroglobulin in human myometrial smooth muscle cells

Significance

The large-conductance, voltage-gated, calcium (Ca²⁺)-activated potassium channel (BK_Ca) plays an important role in regulating the membrane potential of uterine muscle cells. We demonstrate that BK_Ca interacts with the immunomodulator α-2-macroglobulin (α₂M) and its receptor low-density lipoprotein receptor-related protein 1 in human uterine muscle cells isolated from pregnant women. Furthermore, we report that activated α₂M regulates BK_Ca activity and that activated α₂M and BK_Ca together control Ca²⁺ oscillations in the cells, a process dependent on store-operated Ca²⁺ entry. This study reveals a previously unidentified modulator of the BK_Ca channel and may imply a link between inflammatory processes and excitation changes in the uterine muscle during pregnancy.

Abstract

The large-conductance, voltage-gated, calcium (Ca²⁺)-activated potassium channel (BK_Ca) plays an important role in regulating Ca²⁺ signaling and is implicated in the maintenance of uterine quiescence during pregnancy. We used immunopurification and mass spectrometry to identify proteins that interact with BK_Ca in myometrium samples from term pregnant (≥37 wk gestation) women. From this screen, we identified alpha-2-macroglobulin (α₂M). We then used immunoprecipitation followed by immunoblot and the proximity ligation assay to confirm the interaction between BK_Ca and both α₂M and its receptor, low-density lipoprotein receptor-related protein 1 (LRP1), in cultured primary human myometrial smooth muscle cells (hMSMCs). Single-channel electrophysiological recordings in the cell-attached configuration demonstrated that activated α₂M (α₂M*) increased the open probability of BK_Ca in an oscillatory pattern in hMSMCs. Furthermore, α₂M* caused intracellular levels of Ca²⁺ to oscillate in oxytocin-primed hMSMCs. The initiation of oscillations required an interaction between α₂M* and LRP1. By using Ca²⁺-free medium and inhibitors of various Ca²⁺ signaling pathways, we demonstrated that the oscillations required entry of extracellular Ca²⁺ through store-operated Ca²⁺ channels. Finally, we found that the specific BK_Ca blocker paxilline inhibited the oscillations, whereas the channel opener NS11021 increased the rate of these oscillations. These data demonstrate that α₂M* and LRP1 modulate the BK_Ca channel in human myometrium and that BK_Ca and its immunomodulatory interacting partners regulate Ca²⁺ dynamics in hMSMCs during pregnancy.

During pregnancy, the human myometrium remains relatively quiescent until near term, when it becomes more sensitive to contractile stimuli (1). At term, synchronized phasic contractions develop and increase in strength and frequency to facilitate labor. Phasic contractions require the myometrial smooth muscle cells (MSMCs) to alternate between states of contraction and relaxation. The primary trigger for initiating and maintaining spontaneous contractions is an increase in the intracellular level of the major charge carrier calcium (Ca²⁺) (2–4), which occurs by multiple mechanisms (5, 6), One, Ca²⁺ can enter through voltage-gated channels in response to membrane depolarization. Two, agonist stimulation can cause ligand binding to receptor-operated channels that allow extracellular Ca²⁺ to enter the cell to increase intracellular Ca²⁺ ([Ca²⁺]_i). Three, agonist [e.g., oxytocin (OXT)] can bind to receptors that induce Ca²⁺ release from intracellular stores, including the sarcoplasmic reticulum, by activating signal transduction pathways (2). Lastly, store-operated Ca²⁺ entry (SOCE) can occur in response to sarcoplasmic reticulum store depletion. The rise in [Ca²⁺]_i activates Ca²⁺-calmodulin, myosin light chain kinase, and the actomyosin machinery, thus leading to MSMC contraction (2, 7).

MSMC relaxation occurs by several mechanisms. First, myosin light chain kinase and myosin light chain phosphatase are inhibited by phosphorylation and dephosphorylation, respectively (8). Additionally, [Ca²⁺]_i is reduced by extrusion of Ca²⁺ from the cytosol by plasma membrane Ca²⁺-ATPases and sequestration of Ca²⁺ into internal stores by sarcoplasmic reticulum Ca²⁺-ATPases (6). Finally, K⁺ efflux through Ca²⁺-activated K⁺ channels repolarizes the MSMC membrane, thereby inducing closure of voltage-dependent Ca²⁺ channels and returning the cell to a resting state and [Ca²⁺]_i to basal levels (9).

The predominant K⁺ channel in the myometrium is the large-conductance, voltage-gated, Ca²⁺-activated K⁺ channel (BK_Ca), also known as MaxiK/hSlo/KCa1.1 (10–12). Opening of the BK_Ca channel provides a strong repolarizing current to maintain MSMCs at a polarized membrane potential, thus preventing voltage-gated Ca²⁺ influx and contraction. Conversely, pharmacological block of BK_Ca channels depolarizes human MSMCs (hMSMCs), causing activation of voltage-sensitive L-type Ca²⁺ channels and increased [Ca²⁺]_i (13). BK_Ca is activated by both depolarization of the plasma membrane and increases in [Ca²⁺]_i (13). In vascular smooth muscle cells, BK_Ca is also activated by Ca²⁺ sparks elicited by activation of ryanodine receptors in the sarcoplasmic reticulum (14), but this mechanism of BK_Ca activation does not occur in MSMCs (15). Although a role for this channel in maintaining the membrane potential in MSMCs, and potentially uterine quiescence, throughout pregnancy is well supported (9, 13, 16), BK_Ca channel protein expression is lower in myometrial biopsies isolated from women at term than in biopsies obtained from nonpregnant women (17). Additionally, laboring MSMCs have constitutively active BK_Ca channels in the absence of two key channel activators: high levels of Ca²⁺ and depolarizing stimuli (18). These findings indicate that we do not fully understand how BK_Ca is regulated throughout pregnancy. One strong possibility is that BK_Ca is regulated by association with other proteins. In fact, BK_Ca interacts with various plasma membrane and intracellular proteins and acts as a “coordinator” of cell signaling in other tissues (19–23). Such interactions add functional diversity to BK_Ca and may contribute to its cell- and tissue-specific regulation.

To identify proteins that interact with BK_Ca in human myometrium during pregnancy and assess their roles in uterine excitability, we affinity-purified BK_Ca channels from myometrium isolated from women at term and performed a proteomic analysis of the interacting proteins. We focused on a strong “hit” from this screen, the pan-proteinase inhibitor α-2-macroglobulin (α₂M), for three reasons: (i) α₂M plasma levels rise consistently and significantly increase during pregnancy (24); (ii) α₂M regulates cytokine production, which has been implicated in preterm and term labor (25); and (iii) α₂M is important for other aspects of pregnancy, including embryo implantation, in the mouse (26). However, α₂M’s role in human pregnancy and labor is not yet known. In its inactive form, α₂M can bind to certain cytokines or proteases, leading to proteolytic digestion of the “bait” region of α₂M and trapping of the cytokine/proteases. The proteolytic digestion conformationally changes α₂M to its active form (α₂M*), a state in which it can inhibit the bound endoprotease. This change also exposes the receptor recognition site/epitope for high-affinity binding to the α₂M* receptor, low-density lipoprotein receptor-related protein 1 (LRP1) [reviewed in Rehman et al. (27)]. Binding of α₂M* to LRP1 initiates endocytosis of the cytokine/protease–α₂M*–LRP1 complex, mediating the clearance of cytokines/proteases (28, 29), and also leads to an increase in [Ca²⁺]_i (30). Several cell types secrete α₂M, and, given its regulation by multiple factors, including cytokines and hormones (24), this protein may have cell-specific roles.

Here, we demonstrate that the BK_Ca channel associates with α₂M and its receptor LRP1 in hMSMCs. Electrophysiological measurements show that in the presence of OXT, α₂M* induces oscillatory increases in the open-state probability of the BK_Ca channel. Ca²⁺ imaging studies show that α₂M* binding to LRP1 induces Ca²⁺ oscillations and Ca²⁺ influx through store-operated Ca²⁺ channels (SOCs) in OXT-primed cells. Furthermore, BK_Ca channel activity can regulate the Ca²⁺ oscillations induced by α₂M*. Our findings that α₂M* regulates the BK_Ca channel and myometrial [Ca²⁺]_i dynamics provide evidence that immune-modulating signaling pathways interact with controllers of myometrial excitability.

Materials and Methods

Tissue Samples.

Human myometrial tissue samples from the lower uterine segment were obtained from nonlaboring women at term (≥37 wk gestation) during elective Cesarean section under spinal anesthesia. The recruited subjects had a history of repeat Cesarean sections with no spontaneous or induced labor. A total number of 105 myometrial biopsies were used for these studies. All subjects signed written consent forms approved by the Washington University in St. Louis Internal Review Board (approval no. 201108143). Tissues were obtained in 0.9% saline solution and processed within 1 h.

Isolation and Primary Culture of hMSMCs.

The hMSMCs were isolated and cultured as previously described (31). Briefly, the tissue was washed twice in cold Dulbecco’s PBS containing 50 μg/mL gentamicin and 5 μg/mL Fungizone (both from GIBCO-BRL). The tissue was then cut into 2- to 3-mm pieces, and explants were cultured in DMEM-Ham’s F-12 medium supplemented with 5% FBS, 25 μg/mL gentamicin, 2 ng/mL basic FGF, 3 ng/mL EGF (Lonza), 5 μg/mL Fungizone, and 5 μg/mL insulin (Sigma). Once explant colonies formed, they were expanded. For all experiments, hMSMCs were used at passage 1 or 2.

Protein Preparation from Tissues.

For total lysate (TL) preparations, human nonlaboring uterine tissue was homogenized in Triton-lysis buffer [1% Triton, 150 mM NaCl, 10 mM Tris (pH 8.0)] plus a complete protease inhibitor tablet (Roche) as previously described (32), and then spun at 800 × g for 15 min. The supernatant was cleared by spinning at 14,000 × g for 15 min. Isolation of membrane proteins followed the same procedure except that membrane preparation (MP) buffer [250 mM sucrose, 50 mM 3-morpholinopropane-1-sulfonic acid, 2 mM EDTA, 2 mM EGTA (pH 7.4)] was used, and the final supernatant was further centrifuged at 54,000 × g for 80 min. The pellet was then resuspended overnight with gentle agitation in 100–200 μL of Triton lysis buffer. All steps were performed at 4 °C.

Cross-Linking Abs to Beads.

Abs against BK_Ca (rabbit polyclonal; Santa Cruz Biotechnology), LRP1 (mouse monoclonal; Santa Cruz Biotechnology), α₂M (mouse monoclonal, R&D Systems), and rabbit and goat IgG (Sigma) were incubated with protein A/G beads (Santa Cruz Biotechnology) at room temperature for 1 h. The beads were pelleted at 1,000 × g for 2 min at 4 °C, washed three times with 0.01 M PBS (pH 7.5), and then incubated in 450 μM disuccinimidyl suberate (DSS; Pierce) in anhydrous dimethyl formamide for 1 h at room temperature. Beads were pelleted, and DSS was quenched by washing the beads in 50 mM Tris⋅HCl (pH 7.5) for 15 min at room temperature. After washing with 0.01 M PBS, the beads were suspended in 100 μL of 0.01 M PBS containing 0.025% sodium azide.

Affinity Purification and Sample Preparation for MS.

Myometrial tissue MP samples (200 μg) were precleared with protein A/G beads for 30 min at room temperature and then incubated with anti-BK_Ca– or rabbit IgG-coupled beads overnight at 4 °C in immunoprecipitate (IP) buffer [1% Triton X-100, 150 mM NaCl, 10 mM Tris (pH 7.5)] containing 100 mM PMSF. The lysate–Ab–bead complexes were washed three times with washing buffer [0.5% Triton X-100, 150 mM NaCl, 1 mM EDTA, 50 mM Tris⋅HCl (pH 7.5)], and the protein–Ab complexes were eluted from the beads with a 1 × 100-μL wash of 100 mM Gly (pH 2.5) and 2 × 100-μL washes of 100 mM Gly and 3 M urea (pH 2.5) at room temperature. Samples were neutralized with 1 M Tris at pH 8.0. The samples were dialyzed against 50 mM Tris and 8 M urea (pH 8.5), and buffer-exchanged two more times with 400 μL of 50 mM Tris at pH 8.0. Proteins were reduced in 10 mM DTT for 1 h at 37 °C, alkylated in 55 mM iodoacetamide (Thermo Scientific) for 1 h at room temperature in the dark, and digested overnight with 1 μg of trypsin gold (Promega) in 1 M urea. Each sample was spiked with a tryptic digest of BSA containing iodoacetic acid alkylated Cys residues (Michrom Bioresources) at a ratio of 1:75. Samples were acidified and desalted on Vydac C18 spin-columns (The Nest Group) and subjected to strong-cation exchange (SCX) fractionation on polysulfoethyl-A packed spin columns (The Nest Group) according to the manufacturer’s protocol. Briefly, desalted samples were dissolved into SCX buffer A [5 mM KHPO₄, 25% acetonitrile (ACN)] and loaded onto SCX spin-columns, and the tryptic digests were released from the SCX spin-columns by a three-step (20 mM, 40 mM, 60 mM) KCl elution gradient made from a mixture of SCX buffer A and SCX buffer B (5 mM KHPO₄, 25% ACN, 350 mM KCl). Salt-bumped eluted fractions were desalted on C18 microspin-columns (The Nest Group), dried down, and dissolved into MS loading buffer (1% acetic acid, 1% ACN).

MS.

Samples were subjected to liquid chromatography/tandem MS (LC-MS/MS) on an Agilent 6520 Accurate-Mass Quadropole Time-of-Flight mass spectrometer (Agilent Technologies) interfaced with an HPLC Chip Cube. The samples were loaded onto the large-capacity C18 Chip II (160-nL enrichment column, 9-mm analytical column) and subjected to LC-MS/MS analysis using a 60-min gradient from 1.5% to 35% buffer B (100% ACN, 0.8% acetic acid). The data-dependent settings (MS/MS) included a maximum of 10 ions per cycle at medium isolation width [∼4 atomic mass units (AMU)], and precursor masses were dynamically excluded for 30 s after five MS/MS in a 30-s time window. MS capillary voltage and temperature settings were set to 1,800 V and 330 °C, respectively.

Data Analysis-Spectrum Mill Analysis.

The raw.d files were searched against the UniProt human database using Spectrum Mill software, version B.04.00.127 (Agilent Technologies) with the following settings: precursor mass tolerance of 50 ppm, product mass tolerance of 300 ppm, and a maximum of two trypsin miscleavages. The search modifications included a static carbamidomethylation on Cys residues (C = 57.02146 AMU) and the following posttranslational modifications: oxidized methionine (M = 15.9949 AMU), phosphorylated Ser, Thr, and Tyr (STY = 79.9663 AMU), and ubiquitinated Lys (K = 114.0429 AMU).

Coimmunoprecipitation for Western Blotting.

MP and TL protein samples (200 μg) in 200 μL of IP buffer [10 mM Tris, 150 mM NaCl, 1% Triton X-100, 1 mM PMSF (pH 7.5)] were preadsorbed to 40 μL of Protein A/G beads for 30 min at 4 °C. The preadsorbed samples were then incubated with 60 μL of Ab cross-linked beads overnight at 4 °C with gentle agitation. The beads were pelleted at 1,200 × g for 2 min at 4 °C, washed for 3 × 5 min with Wash Buffer-A [0.5% Triton X-100,150 mM NaCl, 50 mM Tris⋅HCl, 1 mM EDTA (pH 7.5)], and eluted with 80 μL of 2× SDS buffer (for BK_Ca and α₂M Abs) or 100 mM Gly at pH 2.5 (for LRP1 Ab). Samples were resolved on 4–15% gradient gels and transferred to nitrocellulose membranes. Blots were blocked in 5% (wt/vol) nonfat dry milk in TBS-Tween 20 and probed with Abs against LRP1 (1:400) or α₂M (1:1,500) in blocking buffer overnight at 4 °C. Blots were washed in TBS-Tween at room temperature, probed with secondary anti-mouse or anti-goat Abs (1:10,000; Jackson Immunoresearch) for 1 h at room temperature, and developed with enhanced chemiluminescence (Denville Scientific Inc.).

In Situ Proximity Ligation Assay.

The hMSMCs were cultured in chambered slides (eight-well; LabTek), serum-deprived in 0.5% FBS for 24 h, washed with ice-cold PBS, and then fixed in 4% (wt/vol) paraformaldehyde (PFA) in PBS for 30 min at room temperature with gentle rocking. After 3 × 5-min washes in 0.01 M PBS, cells were permeabilized with 0.1% Triton X-100 for 15 min at room temperature, washed twice with PBS, and washed once in PBS containing 100 mM Gly to quench residual PFA. The slides were rinsed in milliQ water to remove residual salts, and in situ proximity ligation assay (PLA; Duolink PLA; Olink Biosciences) labeling was performed with Abs against the following proteins: BK_Ca, β1 (rabbit polyclonal; Abcam), α₂M, LRP1, and the small-conductance Ca²⁺-activated K⁺ channel (SK3, rabbit monoclonal; Alomone Labs) (all diluted 1:100 except for LRP1, which was diluted 1:250). The manufacturer’s protocol was followed exactly except that the cells were stained with nuclear dye TOPRO 3-iodide (1:1,000; Invitrogen) for 5 min at room temperature before the final wash in wash buffer B. Slides were dried at room temperature in the dark, mounted in Vectashield (Vector Laboratories), and stored in the dark at −20 °C until they were analyzed by confocal microscopy (FV500; Olympus). PLA signals were detected at 563 nm, and TOPRO 3-iodide was detected at 633 nm. Duolink ImageTool software (Olink Biosciences) was used for analysis. For accuracy of the quantitation, in certain images, clustered nuclei or nonspecific pixels outside the cells’ boundary were digitally removed before analysis. All data are presented as mean ± SEM.

Activation of α₂M.

For some experiments, α₂M (Sigma) was activated by incubation in 100 mM methylamine (Sigma) in 0.01 M PBS (pH 7.5) for 1 h at room temperature. Activated α₂M (α₂M*) was then dialyzed in at least three changes of PBS over 24 h at 4 °C as described previously (30). Inactive α₂M was processed identically except that methylamine was excluded. For other experiments, α₂M* was purchased from BioMac (no. 05-04). No differences in Ca²⁺ imaging data or patch-clamp recordings were observed between these two sources of α₂M*.

Electrophysiology.

Whole-cell patch-clamp recordings were performed on hMSMCs at room temperature in a bath solution containing 135 mM NaCl, 4.7 mM KCl, 5 mM Hepes, 10 mM glucose, 1 mM MgCl₂, and 2 mM Ca²⁺ (pH 7.4); the pipette solution contained 140 mM KCl, 0.5 mM MgCl₂, 10 mM Hepes, 1 mM EGTA, and 5 mM Mg-ATP (pH 7.2). Currents were elicited with 20-mV steps (100 ms) from −100 mV to +100 mV, from a holding potential of −80 mV, and were acquired at a sampling rate of 1 kHz using an Axopatch 200B amplifier and pCLAMP software (Molecular Devices). Currents were recorded, in the same cell, in the presence of 350 nM α₂M and after superfusion with 350 nM α₂M* for 3 min. Current densities (picoamperes/picofarads) were plotted as a function of membrane potential.

Single-channel recordings in the cell-attached configuration were performed at room temperature on hMSMCs in a bath solution containing 135 mM NaCl, 4.7 mM KCl, 5 mM Hepes, 10 mM glucose, 1 mM MgCl₂, and 2 mM Ca²⁺ (pH 7.4). Pipette solution contained 140 mM KCl, 20 mM KOH, 2 mM MgCl₂, and 10 mM Hepes (pH 7.2). Single-channel currents were recorded for 20 min before adding 100 nM OXT; then, after 10 min, either 350 nM α₂M or α₂M* was added to the recording chamber, and currents were acquired for at least another 30 min. A holding potential of +100 mV was maintained during the whole experiment. Recordings were acquired at 100 kHz and filtered at 5 kHz. Open probability (P_o), open dwell-time, and single-channel conductance were calculated for each condition using pCLAMP software (Molecular Devices).

Ca²⁺ Imaging.

The hMSMCs were loaded with 4 μM Fura-2-AM (Teflabs) and 0.1% Pluronic F-127 (Invitrogen) for 30 min at room temperature in the dark in the following buffer: 140 mM NaCl, 5.4 mM KCl, 0.8 mM MgCl₂, 1.8 mM CaCl₂, 10 mM NaOH, 10 mM Gly, 10 mM Hepes, 5.5 mM glucose, and 1.1 mM Na₂HPO₄ (pH 7.4). Cells were then washed with buffer devoid of fluorophore and incubated for an additional 40 min at room temperature to allow de-esterification of Fura-2-AM. The cells were then imaged on an inverted iMIC digital microscope (Till Photonics) using a 20×/0.75 objective (Olympus). The fluorescence excitation was provided by a Polychrome V monochromator (Till Photonics). A CCD camera (Cooke) was used to collect paired images at alternating excitation wavelengths (340/380 nm) through a 510-nm emission filter at 2.608-s intervals. After subtracting the matching background, the image intensities were divided by one another to yield ratio values for individual cells. Free cytosolic calcium concentration, [Ca²⁺]_i, was estimated in individual cells according to the formula

$${\left[{\text{Ca}}^{2+}\right]}_{\text{i}}=K_{\text{d}}\mathrm{*}\mathit{B*}\left(R-R_{\mathit{min}}\right)/\left(R_{\mathit{max}}-R\right),$$

where K_d is the indicator’s dissociation constant for Ca²⁺ (0.22 μM), R is ratio of fluorescence intensity at two different wavelengths (340/380 nm), R_max and R_min are the ratios of Ca²⁺-free and Ca²⁺-bound fura-2, respectively, and B is the ratio of the fluorescence intensity of the second excitation wavelength at zero and saturating Ca²⁺ concentrations (33). The calibration constants (R_min, R_max, and B) were determined on the same setup in calibration buffers (Invitrogen) containing Fura-2/K⁺ and either 10 mM EGTA or 39.6 μM free Ca²⁺.

For Ca²⁺ imaging studies, 100 nM OXT and 350 nM α₂M or α₂M* were added to the hMSMCs, and [Ca²⁺]_i was measured for at least 45 min. To test desensitization of the OXT receptor, 100 nM OXT was added a second time. Some measurements were taken in the presence of (i) nominal Ca²⁺ buffer (Ca²⁺-free), (ii) the L-type Ca²⁺ channel blocker nifedipine (20 μM; Sigma), or (iii) the broadly acting SOC blocker 2-aminoethoxydiphenylborane (2-APB, 40 μM) or N-[4-[3,5-Bis(trifluoromethyl)-1H-pyrazol-1-yl]phenyl]-4-methyl-1,2,3-thiadiazole-5-carboxamide (BTP2, 1 μM) (Tocris Biosciences). In some experiments, MSMCs were pretreated for 1 h with 40 μM 2-APB, 1 μM BTP2, or the inositol 1,4,5-triphosphate (IP₃) receptor inhibitor xestospongin C (10 μM; Tocris Biosciences). In another set of experiments, cells were pretreated with 500 nM of the LRP1 antagonist receptor-associated protein (RAP; Millipore) for 1 h before stimulation with OXT and α₂M*. Some of the Ca²⁺ recordings were performed in the presence of the BK_Ca selective blocker paxilline (500 nM; Tocris Biosciences) or opener NS11021 (500 nM; Glixx Laboratories), which were added to the cells after the oscillations were observed. The cells that did not respond to OXT and failed to produce rhythmic Ca²⁺ spikes after α₂M* stimulation were excluded from the data analysis in each experimental condition. All of the drugs tested were added to the media and remained in the recording chamber until the end of the experiments, with the exception of paxilline reversibility experiments, where the media in the chamber were replaced. The frequency of Ca²⁺ oscillations was measured as the time between consecutive peaks and was represented as the number of oscillations per hour. GraphPad Prism 6 software was used for data analysis. Replicates are expressed as number of cells and number of tissue samples.

Results

α₂M Is Expressed in Human Myometrium and Interacts with BK_Ca.

To identify proteins that interact with BK_Ca in myometrium from pregnant women, we generated TLs and MPs from term nonlabor human myometrial tissues and used a BK_Ca-specific Ab to affinity purify the BK_Ca channel and any interacting proteins. LC-MS/MS analysis of purified samples revealed several proteins known to interact with the BK_Ca channel, including the auxiliary β1 subunit, filamin-A, and caveolin-1 (18, 32, 34, 35). In addition, we identified human α₂M (UniProt ID no. P01023) as one of the top five multipeptide hits in the screen. Importantly, α₂M was not detected in negative control samples (tissue purified with IgG). Spectrum Mill software identified eight distinct peptides that covered 5.6% of the α₂M protein (84 of 1,474 amino acids) and 10 peptides that covered 8.4% of the BK_Ca protein (Table 1). The spectrum of the highest intensity α₂M peptide (MVSGFIPLKPTVK) is shown in Fig. S1.

Table 1.

BK_Ca channel and α₂M peptides identified by affinity purification and LC-MS/MS

Protein	Peptide	MS/MS score	Measured m/z	Expected m/z
BKCa	(R)AFFYCK(A)	14.07	418.1946	418.1952
	(K)AHLLNIPSWNWK(E)	12.61	493.6009	493.6002
	(K)EGDDAICLAELK(L)	4.28	667.3141	667.3093
	(R)EWETLHNFPK(V)	9.58	434.2136	434.2116
	(R)GGYSTPQTLANR(D)	5.12	632.8139	632.8101
	(R)IITQMLQYHNK(A)	14.91	463.5819	463.5804
	(R)NLVMPLR(A)	9.96	421.7473	421.7451
	(K)SSSVHSIPSTANR(Q)	7.60	448.2263	448.2241
	(K)VSILPGTPLSR(A)	16.74	570.3386	570.3344
	(K)YGGSYSAVSGR(K)	5.66	552.2580	552.2565
α2M	(K)AIGYLNTGYQR(Q)	7.84	628.3215	628.3179
	(K)ATVLNYLPK(C)	8.11	509.7963	509.7924
	(K)LPPNVVEESAR(A)	4.60	605.8199	605.8148
	(K)MVSGFIPLKPTVK(M)	12.20	472.9427	472.9377
	(R)QGIPFFGQVR(L)	10.45	574.8145	574.8148
	(R)SSGSLLNNAIK(G)	8.93	552.2998	552.2932
	(R)VGFYESDVMGR(G)	4.61	630.2902	630.2905
	(K)YGAATFTR(T)	9.36	443.7226	443.7207

Fig. S1.

MS/MS spectrum of a tryptic peptide from α₂M. Representative ionic spectrum of the highest scoring tryptic peptide, MVSGFIPLKPTVK, corresponding to α₂M (GenBank accession no. P01023, amino acids 1,385–1,397). The red arrowhead indicates the precursor mass m/z 472.7. The “b” and “y” ions are labeled in blue and red, respectively. The green lines represent masses of one or two amino acids from the identified peptide.

To confirm the interaction between BK_Ca and α₂M, we first verified that α₂M is expressed in human myometrial tissues. Reverse transcriptase quantitative PCR analysis revealed that α₂M, LRP1, and BK_Ca mRNAs were all expressed in myometrium samples isolated from term nonlabor and term laboring women. We observed no differences in expression between nonlabor and laboring states (Fig. S2A, list of primers used in Table S1). Additionally, immunoblot analysis revealed that α₂M, but not α₂M*, protein was present in nonlabor and laboring myometrium (Fig. S2B). We next examined the TL and MP samples by immunoblot and found that α₂M was present in both (Fig. 1A, input lanes). We then immunoprecipitated BK_Ca and confirmed by immunoblot that α₂M coimmunoprecipitated with the BK_Ca channel in the MP samples (Fig. 1A, BK_Ca lanes). However, very little α₂M coprecipitated with BK_Ca in the TL samples. Finally, we examined LRP1, the receptor through which α₂M* acts, and found that it was expressed in human myometrium (Fig. 1B, input lanes) and coimmunoprecipitated with the BK_Ca channel in both TL and MP samples (Fig. 1B, BK_Ca lanes). As expected, immunoprecipitations with anti-α₂M Ab (Fig. 1A, α₂M lanes) and anti-LRP1 Abs (Fig. 1B, LRP1 lanes) pulled down their respective targets in both TL and MP samples, whereas immunoprecipitations with IgG did not pull down either protein (Fig. 1 A and B, IgG lanes).

Fig. S2.

Expression of α₂M, LRP1, and BK_Ca in nonlabor and labor human myometrium. (A) Level of α₂M, LRP1, and BK_Ca (KCNMA1) mRNA was measured in term nonlabor and term labor myometrial tissue samples (n = 5 in each group) by qRT-PCR and normalized to topoisomerase 1 (TOP1). (B) Immunoblots of membrane proteins from term nonlabor and labor myometrial tissue probed with Abs specific to α₂M. Recombinant inactive α₂M and activated α₂M* serve as indicators of mobility on native gels (n = 3 in each group). MW, molecular weight.

Table S1.

Primers and probes sequences used in this study

Gene (RefSeq no.)	Primer	Primer sequence
KCNMA1 (NM_001014797)	Forward PCR primer	5′-CATTTGCCGTCAGTGTCCT-3′
	Reverse PCR primer	5′-CAGCAATCAGAGCCTCCAG-3′
	Taqman probe	5′-AGGGTCCGTATCAGGGTGAGGAT-3′
α2M (NM_000014)	Forward PCR primer	5′-GCTCATGAAGCCTGATGCT-3′
	Reverse PCR primer	5′-GTCTTCATCGTCCTGGTCATTC-3′
	Taqman probe	5′-TCTCGGCGTCCTCGGTTTACAAC-3′
LRP1 (NM_002332)	Forward PCR primer	5′-CCTGCAGAGATCAAATAACCTGT-3′
	Reverse PCR primer	5′-GTACCCAGGCAGTTATGCTC-3′
	Taqman probe	5′-AGGCCCCTGAGATTTGTCCACAG-3′
TOP1 (NM_003286)	Forward PCR primer	5′-TCATGCTTAACCCTAGTTCACG-3′
	Reverse PCR primer	5′-CGATACTGGTTCCGGATCTTG-3′
	Taqman probe	5′-TTCTGCCAGTCCTTCTCACCCTTG-3′

Fig. 1.

BK_Ca channel associates with α₂M and LRP1 in human myometrium and hMSMCs. (A and B) Representative immunoblots (IBs) of immunoprecipitates (IPs) of MPs and TLs from myometrial tissue from pregnant women. Immunoprecipitation was performed with Abs against the indicated proteins, and blots were probed with Abs specific to α₂M (A) and LRP1 (B). α₂M was detected at ∼180 kDa (A, SDS/PAGE), and LRP1 was detected at ∼500 kDa (arrow in B, native PAGE). The ∼150-kDa band in B is IgG. MW, molecular weight. (C) Representative PLA labeling of hMSMCs with the indicated single Abs and Ab combinations. (Scale bar, 10 μm.) (D) Average number of PLA signals in cells is as in C. Error bars indicate SEM (n = 200 each).

We further validated these molecular interactions by performing in situ PLAs (36) on hMSMCs. Punctate red fluorescent signals were detected when cells were exposed to Abs specific to BK_Ca plus α₂M or BK_Ca plus LRP1 (Fig. 1C), indicating that these pairs of proteins were located within 40 nM of each other. Signals were not detected in control experiments in which a single Ab was used. As expected, the Ab combinations of BK_Ca plus its auxiliary β1 subunit and α₂M plus LRP1 produced signals (Fig. 1C). Consistent with the idea that the interaction with LRP1 was specific for the BK_Ca channel, PLA using the Ab combination of SK3 channel (another prevalent potassium channel in hMSMCs) and LRP1 produced very few punctae (Fig. 1 C and D). Together, these data confirm that α₂M and LRP1 are expressed in hMSMCs and interact with the BK_Ca channel.

α₂M* Increases the P_o of BK_Ca in hMSMCs.

We next evaluated the functional significance of the interaction between BK_Ca and α₂M by performing whole-cell patch-clamp recordings in the presence of inactive or activated α₂M (α₂M or α₂M*, respectively). In this recording configuration, neither α₂M* nor α₂M significantly affected the BK_Ca current over a voltage range from −100 to +100 mV (Fig. 2A). Previous studies have shown that α₂M* increases [Ca²⁺]_i (30, 37, 38), a key regulator of the BK_Ca channel. Thus, we used the single-channel, cell-attached, patch-clamp configuration, in which cell constituents and signaling pathways are intact, to investigate whether α₂M* regulates BK_Ca activity. Under control conditions, hMSMCs showed low levels of BK_Ca channel activity (Fig. 2 B and C). To ensure that cells were able to release [Ca²⁺]_i, and thereby promote BK_Ca channel activation, cells were superfused with a high dose (100 nM) of the uterotonin OXT. At this dose, OXT increases [Ca²⁺]_i via G protein-coupled receptor activation and release of diacylglycerol and IP₃, and rapidly desensitizes the OXT receptors (39). Of the cells tested, 49% (36 of 74 cells from 16 myometrial samples) were sensitive to OXT. In these cells, OXT induced an increase in the P_o and dwell-time (from 1.29 to 4.47 ms) of the BK_Ca single-channel currents (Fig. 2 B and C). Treatment of hMSMCs with inactive α₂M did not alter P_o or dwell-time (Fig. 2B; 0 of 12 cells from six myometrial samples responded to α₂M). In contrast, 28% (10 of 36 cells from 16 myometrial samples) of cells responded to application of α₂M* by increasing both P_o and dwell-time of the BK_Ca channel openings. However, these increases were oscillatory (19.12 ± 4.85 oscillations per hour; Fig. 2 C–E). These results indicate that α₂M* activates BK_Ca channels, likely through intracellular pathways.

Fig. 2.

α₂M* causes oscillatory increases in BK_Ca channel activity in hMSMCs. (A) Plots of whole-cell current densities in hMSMCS evoked from −100 mV to +100 mV in 20-mV steps from a holding potential of −80 mV in the presence of 350 nM α₂M* (●) or α₂M (○) (n = 6). (Insets) Voltage-step protocol and representative recordings. (B–E) Representative cell-attached BK_Ca single-channel currents evoked by +100 mV, with 160 mM K⁺ in the pipette, in hMSMCs before and after addition of 100 nM OXT and 350 nM α₂M (B) or 100 nM OXT and 350 nM α₂M* (C). Arrows indicate phasic increases in BK_Ca currents. c, closed state of channels; o, open state of channels. (D and E) Single-channel analysis of P_o (D) and open dwell-time (E) of the BK_Ca channel in the presence of OXT and α₂M* (n = 10).

The Interaction of α₂M* with LRP1 Leads to Ca²⁺ Oscillations in hMSMCs.

Given the observations that α₂M* can cause an increase in [Ca²⁺]_i in osteoblasts (37), macrophages (38), and mouse cortical neurons (30), and the fact that [Ca²⁺]_i regulates BK_Ca activity, we wondered whether α₂M* causes an increase in [Ca²⁺]_i in hMSMCs. To examine this possibility, we used the Ca²⁺-sensitive dye Fura-2-AM to measure [Ca²⁺]_i in hMSMCs (Fig. S3 and Table S2). We first confirmed that the cells were able to respond to OXT, which is known to increase [Ca²⁺]_i. Addition of 100 nM OXT induced a short-duration (Fig. 3A; 4 ± 1.5 min) single-spike increase in [Ca²⁺]_i in ∼80% (901 of 1,127) of the cells analyzed. Although OXT was present throughout the entire experiment, it did not evoke any further response for up to 100 min. This was likely due to desensitization of the OXT receptor, because a second application of 100 nM OXT failed to produce any increase in [Ca²⁺]_i (Fig. 3A). When 350 nM α₂M* was applied after OXT, we observed Ca²⁺ oscillations at regular time intervals (Fig. 3B and Table 2). Whereas the baseline free [Ca²⁺]_i of the hMSMCs was 20.9 ± 13 nM, the α₂M*-induced [Ca²⁺]_i peaks were 298 ± 0.08 nM (Table 2). We did not observe α₂M*-induced Ca²⁺ oscillations in cells treated with α₂M* in the absence of OXT (Fig. 3C).

Fig. S3.

Live-cell Ca²⁺ imaging of hMSMCs. Representative images of Fura-2-AM intensity at various times in the presence of OXT and α₂M*. (A) Basal Ca²⁺ levels were observed (low Fura-2-AM excitation) before application of OXT. When OXT was added, high Fura-2-AM excitation was observed in most cells (B, arrows); this high excitation quickly declined (C). (D–F) Addition of α₂M* at 850 s induced a moderate oscillatory increase in Fura-2-AM signal in some of the OXT-sensitive cells (arrowheads). The bar in the top right corner of each panel shows Fura-2-AM intensities, which correspond to cytosolic Ca²⁺ concentration.

Table S2.

Summary of Ca²⁺ imaging experiments

Agents added (media conditions)† (n)	Cells responsive to OXT, %	Cells responsive to α2M*, %	Cells responsive to modulators %
OXT; OXT (6)	39/48 (81.25)‡	—	—
α2M*; OXT§ (14)	82/105 (78.09)	0/82 (0)	—
OXT; α2M* (43)	204/247 (82.59)	127/204 (62.25)	—
OXT; α2M* (RAP) (10)	80/102 (78.43)	0/80 (0)	—
OXT + α2M (10)	118/150 (78.67)	—	—
OXT; α2M* (Ca2+-free) (8)	55/70 (78.57)	0/55 (0)	—
OXT; α2M*; nifedipine (6)	58/73 (79.45)	36/58 (62.07)	0/36 (0)
OXT; nifedipine (3)	16/20 (80)	—	—
OXT; α2M* (2-APB) (3)	12/15 (80)	0/12 (0)	12/12 (100)
OXT; α2M* + 2-APB (5)	41/50 (82)	25/41 (61)	25/25 (100)
OXT; α2M* (BTP2) (3)	12/15 (80)	0/12 (0)	12/12 (100)
OXT; α2M* + BTP2 (3)	17/21 (80.95)	10/17 (58.82)	10/10 (100)
OXT; α2M* (xestospongin C) (6)	10/45 (22.22)¶	0/45 (0)	—
OXT; α2M*; paxilline (6)	53/67 (79.1)	32/53 (60.38)	32/32 (100)
OXT; paxilline (3)	23/29 (79.31)	—	—
OXT; α2M; paxilline (3)	27/34 (79.41)	—	—
OXT; α2M*; NS11021 (6)	56/70 (80)	35/56 (62.5)	35/35 (100)
OXT; NS11021 (3)	24/30 (80)	—	—
OXT; α2M; NS11021 (3)	25/32 (78.13)	—	—

n, number of patients.

^†Agents are listed in order added and separated by semicolons; media conditions were established before the first agent was added.

^‡Only the first OXT application elicited a response.

^§α₂M* was applied before OXT.

^¶Reduced OXT response.

Fig. 3.

α₂M* induces [Ca²⁺]_i oscillations in OXT-primed hMSMCs by binding LRP1. Representative live-cell Ca²⁺ imaging recordings of Fura-2-AM–loaded hMSMCs are shown. An increase in [Ca²⁺]_i is reported as a ratiometric measure of Fura-2-AM fluorescence at 340/380 nm [in arbitrary units (a.u.)], with background subtraction (Left, y axis) and estimated [Ca²⁺]_i concentration (Right, y axis) (details are provided in Materials and Methods). In all experiments, 100 nM OXT (▲) was added at the indicated times. (A) Only OXT was added. (B) α₂M* (●; 350 nM) was added after OXT. (C) α₂M* (350 nM) was applied before 100 nM OXT. (D) Cells were incubated with 500 nM RAP, a selective antagonist of LRP1, before OXT, and α₂M* was then added. (E) Inactive α₂M (○) was added after OXT.

Table 2.

Effect of BK_Ca modulators on α₂M*-induced Ca²⁺ oscillations

Agents added after OXT†	Percentage of α2M*-responsive cells‡	Fura-2 ratio (340 of 380)§	Estimated peak [Ca2+]i, nM§	Frequency, oscillations per hour§
α2M*	62.25 (127/204)	0.98 ± 0.07	298 ± 0.08	11 ± 0.78
α2M*; paxilline	60.38 (32/53)	0	0	0
α2M*; NS11021	62.5 (35/56)	0.93 ± 0.72	281 ± 0.09	20 ± 0.81¶

^†Agents are listed in order added and separated by semicolons.

^‡n = 43 patients.

^§Mean ± SEM.

^¶P < 0.05 compared with α₂M*.

To determine whether α₂M* acted through its receptor, LRP1, we pretreated hMSMCs with 500 nM RAP, a competitive antagonist of LRP1. In this case, the amplitude of the initial OXT-induced short-duration increase in [Ca²⁺]_i was reduced and the α₂M*-induced Ca²⁺ oscillations did not occur (Fig. 3D). Furthermore, inactive α₂M, which is unable to bind to LRP1, did not induce any change in [Ca²⁺]_i (Fig. 3E).

To identify the source of Ca²⁺, we measured [Ca²⁺]_i in the absence of extracellular Ca²⁺ (Ca²⁺-free buffer). Under these conditions, α₂M* was unable to produce oscillations, although the OXT-elicited peak of [Ca²⁺]_i still occurred (Fig. 4A). Next, we assessed whether the Ca²⁺ entered the cytoplasm through L-type Ca²⁺ channels, a major source of Ca²⁺ entry in MSMCs; through the sarcoplasmic reticulum; or by SOCE. Treating the cells with the L-type Ca²⁺ channel blocker nifedipine did not affect the oscillations (Fig. 4B). However, α₂M* was not able to elicit Ca²⁺ oscillations when hMSMCs were pretreated with either of two SOCE inhibitors, 40 μM 2-APB or 1 μM BTP2, although the OXT response still occurred (Fig. 4 C and E). Moreover, adding either 40 μM 2-APB or 1 μM BTP2 after α₂M*-induced Ca²⁺ oscillations were established abolished these oscillations, although the effect of BTP2 was slower than the effect of 2-APB (Fig. 4 D and F). Finally, blocking IP₃ receptors on the sarcoplasmic reticulum with xestopongin C inhibited both the OXT response and the Ca²⁺ oscillations (Fig. 4G). Together, our data indicate that in hMSMCs primed with OXT, α₂M*-induced Ca²⁺ oscillations require α₂M* binding to its receptor and influx of extracellular Ca²⁺ through SOCs.

Fig. 4.

Ca²⁺ oscillations evoked by α₂M* in hMSMCs require extracellular Ca²⁺ entry through SOCs. (A–G) Representative live-cell Ca²⁺ imaging recordings of Fura-2-AM–loaded hMSMCs. (A) Cells were preincubated in Ca²⁺-free extracellular medium before addition of OXT (▲) and α₂M* (●). (B) L-type Ca²⁺ channel blocker nifedipine (△) was added after OXT and α₂M*. Cells were preincubated with the SOC inhibitor 2-APB (C) or BTP2 (E) before addition of OXT and α₂M*. The 2-APB (D, gray triangle) or BTP2 (F, gray circle) was added after OXT and α₂M*. (G) Cells were pretreated with the IP₃ receptor antagonist xestospongin C (Xesto C) before addition of OXT and α₂M*.

BK_Ca Activity Modulates the α₂M*-Induced Ca²⁺ Oscillations.

In hMSMCs, Ca²⁺ influx depolarizes the plasma membrane and activates BK_Ca channels to restore the membrane potential. To determine whether BK_Ca activity affected the Ca²⁺ oscillations, we examined the effect of treating the cells with paxilline, a specific BK_Ca channel blocker, or NS11021, a specific BK_Ca channel opener. When 500 nM paxilline was added to the cells after establishment of α₂M*-induced Ca²⁺ oscillations, the oscillations were completely abolished (Fig. 5A and Table 2). This result was surprising because we assumed that paxilline would induce depolarization and increase the frequency of the Ca²⁺ oscillations. In fact, we observed no change in [Ca²⁺]_i after paxilline was added in the absence of α₂M* (Fig. 5C) or in the presence of inactive α₂M (Fig. S4A). In contrast, addition of 500 nM NS11021 increased the frequency of the oscillations by twofold (Fig. 5B and Table 2). NS11021 did not affect [Ca²⁺]_i in the absence of α₂M* (Fig. 5D) or in the presence of inactive α₂M (Fig. S4B). To determine whether the Ca²⁺ oscillations could be recovered, paxilline was washed out and replaced with buffer containing OXT and α₂M*. In ∼30% of cells, the Ca²⁺ oscillations returned, although they showed reduced amplitude (Fig. S5 and Table S3). Reapplication of OXT plus α₂M* (in the absence of paxilline) did not affect the already initiated α₂M*-induced Ca²⁺ oscillations (Fig. S5A).

Fig. 5.

BK_Ca channel modulates α₂M*-induced Ca²⁺ oscillations. Representative live-cell Ca²⁺ imaging recordings of Fura-2-AM–loaded hMSMCs are shown. The specific BK_Ca blocker paxilline (Pax) (A and C, □) or opener NS11021 (B and D, ♢) was added after OXT (▲) and α₂M* (●). (C and D) No α₂M* was added.

Fig. S4.

Neither a BK_Ca channel blocker nor an opener changes Ca²⁺ levels in the presence of inactive α₂M and OXT. Representative live-cell Ca²⁺ imaging recordings of Fura-2-AM–loaded hMSMCs. The specific BK_Ca blocker paxilline (A, Pax, □) or opener NS11021 (B, ♢) was added after OXT (▲) and inactive α₂M (○). a.u., arbitrary units.

Fig. S5.

BK_Ca channel blockade is partially reversible. Representative live-cell Ca²⁺ imaging recordings of Fura-2-AM–loaded hMSMCs. (A) Cells were treated with OXT (▲) and α₂M* (●), and OXT and α₂M* were then added again. (B and C) Paxilline (□) was added after OXT and α₂M*. After the oscillations were blocked, paxilline was replaced with OXT plus α₂M*. Examples are shown of a cell in which Ca²⁺ oscillations returned (although at reduced amplitude) (B) and one in which they did not (C).

Table S3.

Paxilline reversibility experiments

Agents added† (n)	Cells responsive to OXT, %	Cells responsive to α2M*, %	Cells responsive to paxilline, %	Cells showing oscillations after paxilline washout, %
OXT; α2M* (4)	24/30 (80)	15/24 (62.5)‡	—	—
OXT; α2M*; paxilline (3)	23/28 (82.14)	13/23 (56.52)	13/13 (100)	4/13 (30.77)§

n, number of patients.

^†Agents are listed in the order added and separated by semicolons.

^‡All cells showed Ca²⁺ oscillations after new media containing OXT and α₂M* were added.

^§Nine cells (69.23%) did not show oscillation recovery after paxilline removal.

Discussion

BK_Ca is the predominant K⁺ channel transcript in human myometrium (10), but its role in regulating uterine contraction is not well understood. In this study, we used an unbiased affinity purification and LC-MS/MS approach to identify proteins that interact with the BK_Ca channel in the human myometrium during pregnancy. This approach revealed an interaction between the BK_Ca channel and α₂M. Furthermore, we showed that the BK_Ca channel associates with the α₂M receptor, LRP1, in MSMCs from term pregnant women. Thus, we have identified a previously unknown BK_Ca channel complex in hMSMCs that may contribute to regulation of myometrial excitation.

Our observations comparing the effects of α₂M and α₂M* lead us to speculate that the BK_Ca channel regulates Ca²⁺ dynamics only in cells in an inflammatory state. For example, we found that pharmacological inhibition or activation of BK_Ca did not affect Ca²⁺ dynamics in cells treated with α₂M but abolished or increased, respectively, the frequency of Ca²⁺ oscillations in cells treated with α₂M*. Furthermore, other studies have shown that paxilline does not affect spontaneous contractions of the uterus (5, 6). On the basis of our data, we formulated the model illustrated in Fig. 6. First, OXT, a key uterotonin released in labor and used clinically for labor induction and augmentation (40), induces an initial rise of [Ca²⁺]_i through activation of IP₃ receptors (Fig. 6A). Second, LRP1 activation by α₂M* promotes extracellular Ca²⁺ entry via SOCE in conjunction with sarcoplasmic Ca²⁺ release and transient activation of BK_Ca channels (Fig. 6B). When BK_Ca channels are blocked, K⁺ efflux is prevented, thus lowering the driving force for Ca²⁺ influx into the cell (Fig. 6C). Conversely, pharmacological activation of BK_Ca channels facilitates Ca²⁺ influx, thereby increasing the frequency of α₂M*-elicited Ca²⁺ oscillations (Fig. 6D). Our observation that nifedipine did not abolish the α₂M*-induced Ca²⁺ oscillations leads us to speculate that BK_Ca activity-induced Ca²⁺ influx is due to SOCE rather than influx of Ca²⁺ through L-type Ca²⁺ channels.

Fig. 6.

Working model of Ca²⁺ dynamics regulation by the functional interaction between BK_Ca and α₂M*-LRP1 in hMSMCs. (A) OXT bound to its receptor induces production of IP₃, which binds its receptor and leads to release of Ca²⁺ from the sarcoplasmic reticulum. (Insets) Release is observed as a peak in [Ca²⁺]_i. (B) Release of Ca²⁺ from the sarcoplasmic reticulum activates SOCs. In the presence of α₂M* bound to LRP1, SOC activation induces a moderate transient Ca²⁺ influx and activation of BK_Ca channels. BK_Ca channel activation leads to K⁺ efflux, which facilitates further influx of Ca²⁺ through SOCs. This hyperpolarizes the membrane, increasing the driving force for Ca²⁺ entry and inducing the oscillatory rise in [Ca²⁺]_i, as shown in the Inset. (C) Blocking BK_Ca with paxilline prevents K⁺ efflux, so [Ca²⁺]_i is maintained at low levels. (D) In the presence of the BK_Ca opener NS11021, K⁺ efflux is increased, thereby increasing the Ca²⁺ influx rate to maintain the electrochemical gradient. DAG, diacylglycerol; Gα_q/11, α subunit of protein G_q/11; Gβγ, β and γ subunits of protein G; IP₃R, IP₃ receptor; OTR, OXT receptor; PIP₂, phosphatidylinositol biphosphate; PLC, phospholipase C.

Our findings run somewhat counter to traditional concepts about activity of the BK_Ca channel. Opening of the BK_Ca channel provides a strong repolarizing current to maintain MSMCs at a polarized membrane potential, thus preventing voltage-gated Ca²⁺ influx and contraction. Conversely, pharmacological block of BK_Ca channels depolarizes immortalized myometrial cells, causing activation of voltage-sensitive L-type Ca²⁺ channels and increased [Ca²⁺]_i (13). However, we found the opposite: In the presence of α₂M*, the channel opener NS11021 stimulated Ca²⁺ oscillations, creating a potentially procontractile state. At resting potential, BK_Ca displays a low open-state probability (32), and pharmacological block does not alter uterine activity (6). In light of our findings, we speculate that, depending on the inflammatory state, the BK_Ca channel can play two roles in the MSMCs. In a noninflammatory state, BK_Ca behaves in a fashion typically noted in excitable cells: The channel is activated at high depolarizing voltages, leading to K⁺ efflux, repolarization of the MSMC membrane, and deactivation of voltage-gated Ca²⁺ channels. Conversely, in an inflammatory state, such as we have mimicked with α₂M*, BK_Ca activity can regulate Ca²⁺ influx through SOCE to modulate phasic Ca²⁺ oscillations. Inhibition of BK_Ca inhibits SOCE, and, conversely, BK_Ca activation enhances SOCE. These effects appear to be voltage-independent (given the lack of effect of nifedipine) and are more typical of nonexcitable cells than of excitable cells. In the presence of α₂M*, the effects of paxilline on Ca²⁺ oscillations were immediate, suggesting involvement of plasma membrane BK_Ca channels. However, our data do not exclude the possibility that BK_Ca channels localized to other intracellular organelles contribute to Ca²⁺ dynamics in MSMCs (41, 42). In light of our findings, the role of the BK_Ca channel in Ca²⁺ signaling in hMSMCs should be revisited and considered in the context of other myometrial signaling events.

Regulation of the BK_Ca channel by α₂M* has been implied, but not directly shown, in other studies. For example, serum α₂M induced transient, discrete, hyperpolarizing spikes in isolated cells of a rat osteosarcoma cell line; these spikes were hypothesized to be due to opening of Ca^2+-dependent K⁺ channels by [Ca²⁺]_i (37). Our data showing that α₂M* activates the BK_Ca channel likely through increases in [Ca²⁺]_i rather than by direct activation of the channel are consistent with these findings. In contrast to our finding that α₂M* induces Ca²⁺ oscillations in MSMCs, others have reported that α₂M* caused stable [Ca²⁺]_i increases in macrophages (38), neurons (30), and trabecular meshwork cells (43). The mechanisms by which α₂M* causes Ca²⁺ influx appear to be cell type-specific. In neurons, the increases in [Ca²⁺]_i result from activation of the NMDA receptor (30), an unlikely candidate in MSMCs in which Ca²⁺ predominantly enters by voltage-gated Ca²⁺ channels (44, 45), with contributions from receptor-operated Ca²⁺ channels and SOCs (6, 46). In macrophages, Misra et al. (47) proposed that GRP78, a 78-kDa glucose-regulated protein, is an alternative α₂M receptor. In the absence of GRP78, α₂M* was unable to regulate IP₃ and cytosolic free Ca²⁺. GRP78 increases in human myometrium in response to inflammation (48), is associated with the BK_Ca channel in mouse cochlea (49), and is elevated in labor or preterm labor compared with nonlabor myometrium (48). We cannot rule out the possibility that GRP78 serves as a receptor for α₂M*. Nonetheless, our data showed that Ca²⁺ influx was inhibited in the presence of RAP, suggesting that LRP1 is the responsible receptor in hMSMCs. Although the identities of the SOCE proteins responsible for α₂M*-regulated Ca²⁺ entry are unknown, evidence from other studies in MSMCs indicate that TRPC (50), STIM, and Orai (51) may contribute to this current. The SOCE blockers 2-APB and BTP2 both inhibit TRPC channels (52), suggesting that entry through TRPC was, in part, a component of the SOCE entry needed for the Ca²⁺ oscillations we observed in the MSMCs. Whether these channels are a functional component of the α₂M*–LRP1–BK_Ca complex will be investigated in future studies.

We found that, in the absence of OXT, α₂M* failed to produce Ca²⁺ oscillations, indicating that the cells must first be primed by Ca²⁺ release from the sarcoplasmic reticulum. Further studies will establish whether this stimulus is specific to OXT. Because OXT was present throughout the entire Ca²⁺ imaging time, it could be argued that OXT exerted a sustained effect on the cells. However, we did not observe an increase in [Ca²⁺]_i upon a second exposure to OXT, and previous studies have shown that 100 nM OXT desensitizes OXT receptors (39). Moreover, inactive α₂M did not induce Ca²⁺ oscillations after OXT application. Alternatively, α₂M* might prevent OXT receptor down-regulation through its antiprotease activity. However, we only observed oscillations in the presence of α₂M*, which is in a locked conformation that cannot bind and clear proteases, thus excluding this possibility (53). We predict that in hMSMCs, a complex consisting of the OXT receptor, α₂M*, LRP1, and BK_Ca synergistically regulates Ca²⁺ dynamics, which, in turn, control excitation-contraction coupling. In support of this idea, previous studies have indicated that LRP1 (54) and the BK_Ca channel (32) can localize to specialized lipid-rich membrane domains that facilitate signaling events by positioning functionally interacting molecules in close proximity (55, 56). Given the physical association between BK_Ca, α₂M*, and LRP1, endocytosis mediated by α₂M*-LRP1 (57, 58) may result in internalization of BK_Ca to decrease membrane localization and may represent another mode of α₂M* action in MSMCs. This periodic removal of BK_Ca from the plasma membrane would provide an excitatory signal to elicit activation of SOCs and Ca²⁺ oscillations. Internalization of BK_Ca was recently demonstrated in arterial smooth muscle cells; in this case, angiotensin II signaling through PKC stimulated BK_Ca channel internalization (59). Likewise, in cardiomyocytes, ATP-sensitive K⁺ channels are internalized to the endosome, where they serve as a reserve to be trafficked to the surface after cardiac ischemia (60). Thus, in MSMCs, internalization may provide a mechanism for rapid control of BK_Ca channel surface localization to control excitability at the critical juncture between nonlaboring and laboring states.

Data from this and other studies (61) indicate that α₂M is present in the myometrium; however, we detected the inactive α₂M form only in native tissue. This could be due to rapid degradation of α₂M* by the endocytic machinery or destabilization of α₂M* during sample preparation. Furthermore, we envision that α₂M* secreted from other tissues, such as the decidua and placenta (24, 62), might contribute to the response in vivo. Our study indicates that, in addition to its roles in immune cell activation and migration (24), α₂M* may affect pregnancy outcomes via cross-talk with the BK_Ca channel. We performed our experiments on isolated hMSMCs, but, in vivo, gap junction-mediated connectivity and enhanced intercellular communication of the hMSMCs at the end of pregnancy (63) could magnify α₂M* action to promote uterine contraction. Both α₂M* and the BK_Ca channel are conserved across species, and α₂M* is the major endoprotease inhibitor in mammalian blood. However, whether α₂M* and BK_Ca functionally interact in the myometrium of other mammals was not investigated.

In human parturition, an inflammatory process drives cervical ripening and myometrial activation (64), but the process by which inflammation leads to changes in excitability of MSMCs is unclear. Proinflammatory cytokines, such as TNF-α, TGF-β, and IL-1β, have been implicated to be triggers of both term and preterm labor (65–67) and can bind to and activate α₂M (24). Thus, we speculate that the proinflammatory state at the time of labor causes α₂M activation and binding to LRP1, endocytosis of this complex, and an increase in [Ca²⁺]_i, thereby promoting uterine contractility. Moreover, IL-1β, a cytokine important in both term and preterm labor, enhances both basal entry and SOCE, but does not affect L- or T-type voltage-gated Ca²⁺ channels (25). Although the IL-1β–induced increase in [Ca²⁺]_i is likely due, in part, to the IL-1β receptor, additional signaling through the α₂M* pathway may provide a mechanism for signal enhancement. Inflammation elicited by intrauterine infections can also promote abnormal and premature contractions of the myometrium and preterm labor (68), but whether α₂M* is responsible for the uterine transition from quiescence to contractility is unclear. Further study to identify and dissect proteins and pathways that regulate ion channel activity to control uterine activity will continue to provide insights into the normal course of pregnancy.

SI Materials and Methods

RNA Isolation and Quantitative Reverse Transcriptase PCR.

Total RNA was isolated from myometrial tissues obtained by Cesarean section from women at term in established labor (spontaneous) or nonlaboring using an Aurum Total RNA Fatty and Fibrous Tissue Pack Kit (BioRad). The quality and integrity of RNA were validated by an Agilent 2100 Bioanalyzer RNA 6000 Pico Kit and Nanodrop (Agilent Technologies). RNA samples with 260 of 280 ratios ≥1.8 were used for further analysis. Relative gene expression was determined using TaqMan gene expression assays (Applied Biosystems) by quantitative reverse transcriptase PCR (qRT-PCR) following Minimum Information for Publication of Quantitative Real-Time PCR Experiments (MIQE) guidelines (69) with the ABI 7500 Fast Real-Time PCR System (Applied Biosystems). The qRT-PCR assay was performed using PrimeTime qPCR Assays (Integrated DNA Technologies, Inc.). Target gene expression of α₂M, LRP1, and KCNMA1 was normalized to the reference gene topoisomerase I (TOP1). Primers and probes used in this study are listed in Table S3.

Expression of α₂M and α₂M* in Term Labor and Term Nonlabor Myometrial Tissue Proteins.

MPs and TLs prepared from term labor and term nonlabor myometrial tissues were mixed with native sample buffer (BioRad) to a final protein concentration of 1 μg/μL. All samples were resolved on a 5% native gel and transferred to nitrocellulose. Immunoblotting was performed as described in Materials and Methods, probing with anti-α₂M Ab (1:1,500; R&D Systems) in blocking buffer overnight at 4 °C.