Regulation of Interleukin-1 beta secretion from Macrophages via modulation of Potassium ion (K⁺) channel activity

Abstract

A causal relationship exists between macrophage cholesterol levels and inflammation, e.g., Interleukin-1β (IL-1β) secretion. A decrease in intracellular K⁺ is essential for inflammasome activation/IL-1β secretion and, herein, we examined the hypothesis that cellular cholesterol affects K⁺-channel activity and K⁺-efflux using mouse peritoneal macrophages (MPMs) and human/THP1 macrophages. An increase in cellular cholesterol led to a significant increase in K⁺ currents (>350% in both MPM and THP1). Enhancing cholesterol efflux returned K⁺ currents back to basal levels with corresponding increase in intracellular K⁺ (11.2 – 14.5%) and reduced IL-1β secretion (32 – 62%). These data demonstrate a novel mechanism by which cellular cholesterol modulates inflammation/inflammasome via regulation of K⁺-channel activity and intracellular K⁺ levels. Attenuation of IL-1β secretion by Nateglinide/Repaglinide further suggests involvement of Kir6 channels.

Introduction

Several lines of evidence suggest a close relationship between cellular cholesterol content/homeostasis and inflammatory status of macrophages. Hyper responsiveness of macrophages from hypercholesterolemic patients towards chemotactic stimuli and increased adhesion to vessel walls provided the first evidence for the role of cholesterol metabolism in directing an inflammatory response (1). Since then, the relationship between macrophage cholesterol levels and their inflammatory status is increasingly being strengthened and the role of cellular cholesterol transporters (e.g., ABCA1, ABCG1) and extracellular cholesterol acceptors (e.g., HDL) that regulate removal of cellular cholesterol is examined (2). Consistently, bacterial lipopolysaccharide (LPS)-induced sepsis is exacerbated in ABCA1−/−LDLR−/− mice compared to LDLR−/− mice indicating increased LPS-signaling in cholesterol-rich ABCA1 deficient cells (3). Deficiency of ABCA1 also leads to increased inflammatory gene expression and increased signaling via toll-like receptor-4 (TLR4) (4) as well as enhanced pro-inflammatory response of macrophages (5). Supplementation of diabetogenic diet with cholesterol led to increased macrophage infiltration into adipose tissue (6), increase in systemic inflammation (7) and exacerbation of hepatic steatosis and inflammation (8). These findings directly demonstrate a role for cellular cholesterol in inflammation-linked disease processes. Cholesterol transporters ABCA1 and ABCG1 also protect macrophages from apoptosis following efferocytosis providing yet another link between macrophage cholesterol homeostasis to macrophage function (2). However, the mechanisms underlying cellular cholesterol mediated changes in macrophage phenotype and inflammatory status are not completely defined.

Extreme hydrophobicity of cholesterol limits its localization to cellular membranes with plasma membrane containing the largest amount of un-esterified or free cholesterol (FC) where it regulates membrane fluidity and consequently the functions of membrane associated proteins. Increase in cholesterol-enriched lipid rafts in the plasma membrane due to deficiency of membrane cholesterol transporters, ABCA1 or ABCG1, leads to increased activity of TLR4 and LPS-mediated activation of inflammatory signaling (4, 5). Similarly, accumulation of excess FC in intracellular membranes such as endoplasmic reticulum (ER) leads to increased ER stress and apoptosis (9). In addition to the effects of intracellular cholesterol on cellular membrane structure and function, uptake of extracellular cholesterol crystals by macrophages activates the NLRP3 inflammasome resulting in increased secretion of interleukin-1β (IL-1β) and this process is thought to involve phagolysosomal damage (10). However, it has recently been shown that human macrophages avidly phagocytose cholesterol crystals and store the ingested cholesterol as cholesteryl esters (CEs) (11). It needs to be emphasized that while CEs represent the intracellular storage form of cholesterol, CE within intracellular lipid droplets are in constant flux with FC (associated with cellular membranes) in the continuous CE cycle (12). Enhanced hydrolysis of stored CE and subsequent efflux of released FC by over expression of cholesteryl ester hydrolase (CEH) not only decreases foam cell formation and diet-induced atherosclerosis (13) but CEH-mediated reduction in macrophage cholesterol content also leads to decreased systemic inflammation; greater than 20-fold reduction in circulating IL-1β levels are observed in macrophage-specific CEH transgenic (CEHTg) mice (14). Consistently, CE loading of macrophages in vitro led to a dramatic increase in IL-1β secretion which was attenuated in CEHTg macrophages (14). Targeted reduction in macrophage cholesterol content by CEH over expression, therefore, attenuates pro-inflammatory pathways.

Secretion of IL-1β, regarded as the master cytokine of inflammation (15), is tightly regulated requiring two signals. Signal 1 or priming is required to induce NF-κB dependent transcription of IL-1β and this is mediated via activation of TLRs, (e.g., TLR4) by LPS during classical inflammation or by endogenous lipids (oxLDL or free fatty acids, FFA) during sterile inflammation. Signal 2 is subsequently required to assemble the inflammasome complex consisting of a nucleating pattern recognition receptor (PRR), the adaptor protein ASC and the Caspase-1 enzyme that cleaves pro-IL-1β. NLRP3 is the prototypical PRR in inflammasome and is a member of the NLR family. NLRs are cytoplasmic and unlike the TLRs, NLR proteins respond to a variety of DAMPs (danger-associated molecular patterns, both microbial and metabolic) that do not share any obvious structural similarities. The intracellular mechanisms involved in the recognition of these disparate DAMPs are not completely defined. Although perturbation of lysosomal or mitochondrial membranes resulting in the release of cathepsins or reactive oxygen species (ROS) respectively, are thought to be involved it is not established as yet whether one or both of these mechanisms is directly involved in inflammasome activation or that both activate a common downstream pathway. Masters and O’Neill suggested that an ionic flux resulting from plasma membrane perturbation is sensed by NLRP3 such that it can oligomerize and cooperatively activate Caspase-1 (16). Reduction in cellular K⁺ is a common activator of Caspases and inhibition of K⁺ efflux by increasing extracellular K⁺ effectively reduces Caspase-1 activation (17) and secretion of IL-1β (18). However, the mechanisms underlying regulation of cellular K⁺ efflux remain elusive.

In the present studies, we examined the hypothesis that modulation of cellular cholesterol levels regulates final secretion of IL-1β by modulating K⁺ efflux and thereby affecting cellular K⁺ levels. Using in vivo generated macrophage foam cells (macrophages isolated from western diet fed mice) or in vitro generated foam cells (loading of macrophages with acetylated LDL, AcLDL, in vitro), we show an increase in K⁺ currents with increased cholesterol loading that led to a corresponding decrease in cellular K⁺ levels. Consistent with the pivotal role of K⁺ efflux in inflammasome activation, this cellular cholesterol level-dependent decrease in cellular K⁺ led to an increase in IL-1β secretion. The observed K⁺ channel inhibitor-dependent attenuation of IL-1β secretion further establishes the important role of these channels in regulating this pro-inflammatory process.

Materials and methods

Animals:

Macrophage-specific CEHTg mice in LDLR−/− background were developed and characterized in our laboratory (13) and transgenic (LDLR−/−CEHTg) or non-transgenic controls (LDLR−/−) littermates were used. To facilitate increase in macrophage cholesterol content (or in vivo generation of macrophage foam cells), at 10 weeks of age, mice were fed a high fat high cholesterol containing western-type diet (WD, TD 88137). Due to CEH-mediated increase in mobilization of intracellular CE, macrophages from WD-diet fed LDLR−/−CEHTg represent a model system with reduced cellular cholesterol. All procedures were approved by VCU IACUC.

Isolation and culture of macrophages:

Peritoneal exudates were collected 4–5 days after intraperitoneal injection of 4% Thioglycollate broth and plated in suitable tissue culture dishes. The growth medium (DMEM supplemented with 10% FBS and penicillin/streptomycin) was replaced after 4h to remove the non-adherent cells and adherent mouse peritoneal macrophages (MPMs) were used for all the studies.

Human THP1 macrophages:

THP1 monocytes obtained from ATCC were grown and differentiated using PMA as described (19). In some experiments, 3-days after differentiation, THP1 macrophages were incubated with AcLDL (50 µg/ml) in serum-free medium for 48h to facilitate cholesterol loading. Cholesterol depletion/efflux was initiated by replacing the medium with complete growth medium containing 10% FBS as extracellular cholesterol acceptor. Cells were used after 48 h. Cellular cholesterol levels were determined as described earlier (19).

Measurement of K⁺ currents:

Standard whole-cell patch clamp methods were used and all recordings were made at room temperature with micropipettes with a tip diameter of 2–3 μm and resistance of 2–4 MΩ. Junction potentials were corrected and a 3M KCl-agar bridge served as the ground electrode. Freshly isolated MPMs or THP1 macrophages were plated on glass coverslips and adherent cells were directly used for whole cell patch clamp studies. Typical seal resistances were 5–10 GΩ. After obtaining the whole-cell configuration, MPMs were dialyzed for at least 5 min before data were collected. Successive 500 ms pulses were made from a holding potential of −60 mV to test potentials from −100 to +60 mV in +10 mV increments and currents were recorded with an Axoclamp 200B and Digidata 1322A under pClamp 9 (MDS Analytical Technology) and digitized (5 KHz) after low-pass filtering (Bessel, 2 KHz). Current-voltage (I-V) relationships were plotted from quasi steady-state currents. Pipet solution contained 140 mM K⁺ and the bath solution contained 5 mM K⁺ representing the physiological intracellular and extracellular K⁺ concentrations, respectively. Furthermore, to maintain physiological relevance, no channel selective inhibitors were used to isolate ion fluxes through any particular channel as described by Di et al to identify TWIK2 or Kcnk6 as the K⁺ channel involved in macrophage K+ efflux (20). The other components of the pipette solution were: 1 mM CaCl₂, 2 mM MgCl₂, 5 mM K₂EGTAH₂, 10 mM glucose, and 10 mM HEPES (pH 7.4) and that of the bath solution were: 140 mM NaCl, 5 mM KCl, 2 mM CaCl₂, 2 mM MgCl₂, 10 mM glucose, and 10 mM HEPES (pH 7.4). Currents were normalized for cell size by dividing current amplitude (pA) by membrane capacitance (pF) to obtain current density. Chord conductance (Siemens/Farads) was calculated as follows: Current density (pA/pF) at +60mV/change in voltage in Volts between reversal potential and +60 mV.

Measurement of intracellular K⁺ levels:

Relative changes in intracellular K⁺ levels were determined using K⁺ reporter dye as described earlier (21, 22). MPMs were isolated and plated in 12-well tissue culture plates (1.5×10⁶ cells/well). After removing non-adherent cells, adherent MPMs were incubated overnight at 37ºC. Where indicated, MPMs were exposed to LPS (50 ng/ml overnight) followed by 20 µM Nigericin for 20 min or ALUM, 200 µg/ml for 6 h. After washing with PBS, cells were loaded with PBFI-AM (5µM) using PowerLoad™ concentrate from Invitrogen in the presence of Probenecid to inhibit organic-anion transporters (that can extrude the dye and thus contribute to poor loading) according to the manufacturer’s instructions. Following incubation at room temperature for 45 minutes in the dark, cells were washed with FACS buffer containing Probenicid, scraped into the same wash buffer and collected by centrifugation. After re-suspending in wash buffer, 1 µl of cell viability dye, 7AAD was added and samples analyzed by FACS. Mean fluorescent intensity of 7AAD-PBFI+ live cells was determined as the measure of intracellular K⁺ levels.

Secretion of IL-1β:

MPMs or differentiated THP1 macrophages were exposed to LPS (50 ng/ml overnight) followed by 20 µM Nigericin for 20 min or ALUM, 200 µg/ml for 6 h or ATP (5 mM) for 30 minutes. Medium was replaced with fresh growth medium and incubation continued for additional 3h. IL-1β secreted into the culture medium was measured by ELISA (mouse or human IL-1β ELISA Ready-SET-Go!™ kit, respectively). In some experiments, LPS exposed cells were pretreated with the following K⁺ channel inhibitors for 1h prior to induction of IL-1β secretion by Nigericin; KCl (50 mM), Nateglinide (10 µM – 1 mM) (23), Repaglinide (5 µM – 0.5 mM) (23), Ouabain (1 mM) (24), Bumetanide (100 µM) (24); IBTX (100 nM) (25), 4-AP (2 mM) (24), Terfenadine (1 µM) (26), ML-133 (20 mM) (27), SCH 23390 (10 µM) (28) or Quinine (1 µM) (20). Inflammasome inhibitor Glyburide (0.5 mM) was used as a positive control. Cell viability following all indicated treatments was >90% as assessed by MTT viability assay.

Statistical analyses:

All data were analyzed using GraphPad Prism software. Statistical significance of difference between groups was determined by ANOVA and Tukey’s multiple comparison tests were performed to evaluate the significant difference between groups. Significance of differences between two groups was determined using non-parametric t-test and p<0.05 was considered statistically significant.

Results

Cellular cholesterol levels modulate K⁺ channel activity in macrophages:

We earlier reported a significant reduction in inflammation and circulating IL-1β levels in macrophage-specific CEHTg mice; these mice have significantly lower macrophage cholesterol levels especially following a dietary challenge with high fat high cholesterol containing WD (14). While increase in cellular cholesterol content due to deficiency of ABCA1 is shown to increase lipid rafts and enhance TLR4 signaling (5), CD36-dependent uptake of oxidized LDL and likely formation of intracellular cholesterol crystals is thought to activate inflammasome by lysosomal destabilization (22). Given the central role of K⁺ efflux in inflammasome activation, we examined the effects of WD-dependent modulation of intracellular cholesterol levels on total K⁺ channel activities to determine the subsequent effects on regulating K⁺ efflux.

Single cell patch clamp studies were conducted using MPMs isolated from chow or WD fed LDLR−/− or LDLR−/−CEHTg mice with varying cellular cholesterol levels to monitor total K⁺ currents under physiological concentrations of intra- and extracellular K⁺, 140 and 5 mM, respectively. Furthermore, to maintain physiological relevance, movement of other ions was not blocked during these measurements as described earlier by Di et al (20). Consistent with WD-mediated hypercholesterolemia, MPM from WD-fed LDLR−/−mice showed significant increase (>10 fold) in total cellular cholesterol compared to MPM from chow-fed mice (Figure 1A). Stimulation of CE mobilization and cholesterol efflux by transgenic expression of CEH, led to a significantly lower total cellular cholesterol in MPM from WD-fed LDLR−/−CEHTg mice. MPMs isolated from chow fed LDLR−/− or LDLR−/−CEHTg mice displayed small K⁺ current densities (Figure 1B and C). WD feeding, that increases macrophage cholesterol content, significantly increased K⁺ currents in MPMs from LDLR−/− mice. In contrast, no significant difference in K⁺ currents was noted in MPMs from LDLR−/−CEHTg mice (Figure 1B) consistent with significantly lower accumulation of CE in these MPMs (1.69 ± 0.16 vs 3.08 ± 0.59). The statistical significance of the observed differences between the current density/voltage (pA/pF vs mV) plots obtained for the indicted groups was either evaluated by determining the area under the curve (AUC) for the positive currents (Figure 1C) or by calculating the chord conductance (Figure 1D). These data demonstrate that K⁺ currents are significantly modulated by a WD-induced increase in cellular cholesterol in MPMs from LDLR−/− mice.

Figure 1. Increased cellular cholesterol and K⁺ currents in MPMs isolated from WD-fed mice:

MPMs were harvested from chow or WD fed LDLR−/− or LDLR−/−CEHTg mice and used to determine K⁺ currents under physiological K⁺ concentration gradient as described under “Methods”. Panel A: Total cellular cholesterol levels in MPMs isolated from indicated genotypes and diets (Mean ± SD, n=4); Panel B: Current traces upon stepping for 500 ms from −100 mV to +60 mV in 10 mV increments from a holding potential of −60 mV under physiological K⁺ gradient (140 mM in the pipette and 5 mM in the bath). Panel C: Current-voltage (I-V) relationships (Mean ± SD, n = 6–8, MPM from 3 mice were analyzed with a minimum of 2 cells used per mouse) from −100 mV to +60mV. Panel D: AUC for the positive currents for individual I-V curve included in Panel A was calculated and data are presented as Mean ± SD for the indicated genotype and diet groups. Panel D: Chord conductance was calculated as described under “Methods” and data are presented as Mean ± SD. One-way ANOVA and Tukey’s multiple comparison tests were performed and dissimilar letters indicate P<0.05.

To further establish whether these observed changes in K⁺ currents correspond to changes in intracellular K⁺ levels during activation of inflammasome, changes in cellular K⁺ levels were subsequently measured using PBFI, a fluorescent K⁺ indicator, in MPMs isolated from WD-fed mice of both genotypes following activation with LPS and Nigericin; Figure 2A shows a left shift in PBFI fluorescence indicative of decrease in cellular K⁺ levels following activation. Although LPS + Nigericin treatment induced K⁺ efflux and reduced intracellular K⁺ levels in both genotypes, consistent with reduced K⁺ currents, a significantly smaller decrease in cellular K⁺ was noted in MPMs from LDLR−/−CEHTg mice (Figure 2B). Correspondingly, there was a significant reduction in IL-1β secretion from MPMs obtained from LDLR−/−CEHTg mice. (Figure 2C). It needs to be emphasized that significant reduction in circulating IL-1β levels were only observed in LDLR−/−CEHTg mice following WD-feeding (14) and therefore, in these studies MPM from only WD-fed mice were used to evaluate the role of CEH-mediated increase in CE mobilization and cholesterol efflux in modulating cellular K⁺ levels.

Figure 2. Higher K⁺ efflux and consequently lower intracellular K⁺ and higher IL-1β secretion in MPMs from WD-fed LDLR−/− mice:

MPMs were harvested from WD fed LDLR−/− or LDLR−/−CEHTg mice and used to determine intracellular K⁺ levels by FACS as described under “Methods”. Panel A: Representative histograms showing 7AAD negative live cells (Left). Compared to untreated cells (No/T, middle), a left shift indicative of reduced intracellular K⁺ indicator PBFI is seen following stimulation with LPS and Nigericin (Right). Panel B: Mean Fluorescent Intensity of PBFI in 7AAD negative live cells was determined as a measure of intracellular K⁺ concentration and the data (Mean ± SD, n=6, MPM from 6 mice were used and each measurement performed in triplicate) are expressed as % No treatment control for the indicated genotypes. Panel C: LPS+Nigericin induced IL-1β secretion was determined and the data (Mean ± SD, n=6, MPM from 6 mice were used and each measurement performed in triplicate) are expressed as % LDLR−/−. One-way ANOVA and Tukey’s multiple comparison tests were performed and dissimilar letters indicate P<0.05.

To determine whether similar effects of cholesterol loading on K⁺ currents are also relevant to human macrophages and not restricted to MPMs, human THP1 macrophages were incubated with AcLDL to facilitate cholesterol loading and foam cell formation followed by incubation with 10% serum to induce cholesterol depletion/efflux. As shown in Figure 3A, incubation with AcLDL led to a significant increase in cellular cholesterol levels. Induction of efflux reduced cellular cholesterol back to control levels. K⁺ currents were monitored in untreated controls (No AcLDL), cells loaded with cholesterol by incubation with AcLDL as well as cells in which depletion/efflux of cholesterol was facilitated. AcLDL-dependent cholesterol loading significantly increased K⁺ currents and this increase was reversed by depletion/efflux of cholesterol (Figure 3B and C). Consistently, the AUC for the positive currents as well as chord conductance were significantly higher in AcLDL-loaded THP1 macrophages.

Figure 3. Cholesterol loading and/or efflux mediated changes in cellular cholesterol, K⁺ currents, intracellular K⁺ and IL-1β secretion in human THP1 macrophages:

THP1 macrophages were differentiated with 100 nM PMA (Control), then loaded with AcLDL (50 µg/ml) and either used immediately after loading (+AcLDL) or 48 h after initiation of cholesterol efflux with medium containing 10% FBS as extracellular cholesterol acceptor (+AcLDL+Efflux). Panel A: Total cholesterol content of the cells after indicated treatments (Mean ± SD, n=3). Panel B: Current traces upon stepping for 500 ms from −100 mV to +60 mV in 10 mV increments from a holding potential of −60 mV under physiological K⁺ gradient (140 mM in the pipette and 5 mM in the bath). Panel C: Current-voltage (I-V) relationships (Mean ± SD, n=5 cells for each condition) from −100 mV to +60mV for indicated groups. Panel D: AUC for the positive currents for individual I-V curve included in Panel A was calculated and data are presented as Mean ± SD for the indicated groups. Panel E: Chord conductance was calculated as described under “Methods” and data are presented as Mean ± SD. One-way ANOVA and Tukey’s multiple comparison tests were performed and dissimilar letters indicate P<0.05.

In accordance with this cholesterol loading and depletion-mediated changes in K⁺ currents, intracellular K⁺ levels were significantly lower in cholesterol-loaded cells (+ AcLDL) and reverted back to control levels following cholesterol depletion/efflux (Figure 4A). This causal relationship between cellular K⁺ levels and IL-1β secretion was also confirmed by the observed 2-fold increase in IL-1β with cholesterol loading (+AcLDL) and significant reduction following serum-mediated cholesterol depletion/efflux (Figure 4B). Consistently, increased proteolytic cleavage of Pro-caspase to caspase was noted upon cholesterol loading with AcLDL (Figure S1).

Figure 4. Higher K⁺ efflux and consequently lower intracellular K⁺ and higher IL-1β secretion in AcLDL loaded THP1 macrophages:

THP1 macrophages without any cholesterol loading (No AcLDL), with AcLDL loading (+AcLDL) or with stimulated efflux following AcLDL loading (+AcLDL+Efflux) were used to measure intracellular K⁺ levels and IL-1β secretion following treatment with LPS and Nigericin; untreated cells were used as No treatment control. Panel A: Mean Fluorescent Intensity of PBFI in 7AAD negative live cells was determined as a measure of intracellular K⁺ concentration and the data (Mean ± SD, n=6) are expressed as % No treatment control for the indicated groups. Panel B: LPS+Nigericin induced IL-1β secretion was determined and the data (Mean ± SD, n=6) are expressed as % No treatment Control. One-way ANOVA and Tukey’s multiple comparison tests were performed and dissimilar letters indicate P<0.05.

Taken together, these data from MPMs and THP1 suggest that an increase in cellular cholesterol (either by WD-feeding or by AcLDL loading) leads to increased K⁺ currents resulting in decreased cellular K⁺ levels, and, in turn, increased IL-1β secretion. Nigericin, the most commonly used Signal 2 for inflammasome activation and an ionophore, was used in these studies. Whether cellular cholesterol content directly affects the action of nigericin cannot be ruled out.

Inhibition of K⁺ efflux reduces secretion of IL-1β by MPMs induced by different activating signals:

Cholesterol crystal-mediated disruption of lysosomes in oxLDL loaded macrophages is thought to underlie cholesterol loading-mediated inflammasome activation. Based on increase in K⁺ channel activity by cholesterol loading of macrophages, we next examined whether activation of inflammasome by crystal-mediated lysosomal damage could also be modulated by inhibiting K⁺ efflux. Secretion of IL-1β following activation with LPS and ALUM crystals was monitored and activation of K⁺ efflux by Nigericin (an ionophore and an established as well as widely used activator of inflammasome) (29) was used as a positive control. K⁺ efflux was inhibited by increasing extracellular K⁺ concentration by the addition of KCl. Glyburide, an established inflammasome inhibitor, was used as a positive control for inhibition of inflammasome activation/IL-1β secretion. It should be noted that while glyburide is a K_ATP channel inhibitor, its ability to inhibit inflammasome is independent of the cyclohexylurea domain required to interact with ATP-sensitive K_ATP channels (30). As shown in Figure 5A, in the presence of KCl, significant reduction in IL-1β secretion was seen when inflammasome was activated by ALUM crystals similar to the effects seen with Nigericin (Figure 5B) indicating that activation of inflammasome via lysosomal damage by ALUM crystals is equally responsive to modulation of K⁺ efflux. Furthermore, inhibition of IL-1β secretion by increasing extracellular K⁺, and thereby reducing K⁺ efflux, was comparable to that seen with bona fide inflammasome inhibitor glyburide.

Figure 5. Secretion of IL-1β is equally susceptible to changes in intracellular K⁺ efflux whether ALUM or Nigericin is used as Signal 2.

MPMs were harvested from wild type C57BL/6 mice and treated with LPS (50 ng/ml, overnight). K⁺ efflux was blocked by pre-treatment for 1h by KCl (50 mM); pre-treatment with inflammasome inhibitor Glyburide (0.5 mM) was included as positive control. IL-1β secretion was then induced by either ALUM, 200 µg/ml for 6 h (Panel A) or 20 µM Nigericin for 20 min (Panel B) and levels of IL-1β in conditioned medium was determined. Data (Mean ± SD, n=6, MPM from 6 mice were used and each measurement performed in triplicate) are presented as %LPS + ALUM or %LPS + Nigericin. Panel C: Mean Fluorescent Intensity of PBFI in 7AAD negative live cells was determined as a measure of intracellular K⁺ concentration and the data (Mean ± SD, n=6 MPM from 6 mice were used and each measurement performed in triplicate) are expressed as % No treatment control for the indicated groups. One-way ANOVA and Tukey’s multiple comparison tests were performed and dissimilar letters indicate P<0.05.

To further explore whether activation with ALUM leads to a decrease in intracellular K⁺ levels, cellular K⁺ levels were measured in LPS treated cells following stimulation with ALUM; Nigericin was used as a positive control. A significant and quantitatively similar decrease in cellular K⁺ levels was noted with both ALUM and Nigericin (Figure 5C) indicating that despite differences in the intracellular site(s) of action, both these activators have similar effects on cellular K⁺ levels. Taken together with data shown in Figures 1–3, these data suggest that cholesterol accumulation dependent effects on K⁺ channel activity and resulting changes in cellular K⁺ levels represent an additional mechanism by which cellular cholesterol levels, even if acting by mechanism(s) analogous to ALUM crystals, induce an inflammatory phenotype in macrophages. It should be pointed out that the main objective of this series of studies was to establish the central role of K⁺ efflux, irrespective of the signal 2 used, in inflammasome activation and therefore only MPM from C57BL/6 mice were used. Unlike the data shown in Figure 1–3, MPMs from LDLR−/− or LDLR−/−CEHTg mice were not used for these studies. Although a likely limitation, collectively these studies (Figures 1–5) provide a strong support for the concept that processes modulating cellular K⁺ levels (e.g., cellular cholesterol-dependent modulation of K⁺ channel activity) regulate inflammasome activation/IL-1β secretion.

Potential K⁺ channels involved in the efflux of K⁺ from macrophages:

Currently the identity of the K+ channel involved in inflammasome activation is not established and in an attempt to identify the K⁺ channel(s) involved, we next examined the effects of a series of well characterized K⁺ channel inhibitors (Figure 6A) on IL-1β secretion from MPMs and THP1 macrophages. Specific inhibitors of Kir6 channels namely Nateglinide and Repaglinide potently inhibited IL-1β secretion in both MPMs (Figure 6B) as well as THP1 macrophages (Figure 6C) while other inhibitors had negligible effects; KCl and Glyburide were used as positive controls. Based on the dose-response curves for Nateglinide and Repaglinide for MPMs and THP1 macrophages (Figure 7), 50% inhibition of IL-1β secretion was observed with ~500 µM Nateglinide and >50 µM Repaglinide. This is consistent with the reported IC₅₀ values for Nateglinide (800 µM) as well as Repaglinide (21 µM) (23). These data point to a likely involvement of Kir6 channels in mediating K⁺ efflux that underlies inflammasome activation but does not rule out the possible involvement of other channels not specifically examined here.

Figure 6. Identification of putative K+ channel involved in IL-1β secretion from MPMs and THP1 macrophages by using pharmacological inhibitors.

Panel A: Target channels for the inhibitors used. MPMs from C57BL/6 mice (Panel B) and THP1 macrophages (Panel C) were treated with LPS (50 ng/ml) overnight and then pre-treated with indicated K⁺ channel inhibitors for 1h. Nigericin (20 µM for 20 min) induced IL-1β secretion was determined. Data (Mean ± SD, n=6) are presented as %LPS + Nigericin. One-way ANOVA and Tukey’s multiple comparison tests were performed and dissimilar letters indicate P<0.05.

Figure 7. Dose-dependent inhibition of IL-1β secretion from MPMs or THP1 macrophages by Nateglinide and Repaglinide.

MPMs from C57BL/6 mice (Panel A and B) and THP1 macrophages (Panel C and D) were treated with LPS (50 ng/ml) overnight and then pre-treated with indicated concentrations of Nateglinide or Repaglinide for 1h. Nigericin (20 µM for 20 min) induced IL-1β secretion was determined. Data (Mean ± SD, n=3) are presented as %LPS + Nigericin. One-way ANOVA and Tukey’s multiple comparison tests were performed and dissimilar letters indicate P<0.05.

While these studies were being conducted, Di et al reported the involvement of TWIK2 K⁺ channel in ATP-mediated activation of inflammasome (20). To determine whether, in addition to Kir6, if inhibition of TWIK2 channel also leads to decrease in IL-1β secretion, MPMs or THP1 macrophages were primed with LPS and activated with ATP or Nigericin. Treatment with TWIK2 inhibitor quinine did not significantly reduce ATP or Nigericin stimulated IL-1β secretion in either cell types (Figure 8). However, blocking K+ efflux by increasing extracellular K+ concentration (+KCl) significantly reduced IL-1β secretion similar to the effect seen with bona fide inflammasome inhibitor glyburide. To rule out the possibility that absence of TWIK2 expression was the cause for the lack of effect of quinine on IL-1β secretion, expression of TWIK2 in MPMs and THP1 macrophages was confirmed by Real time RT-PCR (Ct values in the range of 26–28 for TWIK2 for β-actin Ct value of 15–16; No Ct value for no template control) as well as by immunocytochemistry. As shown in Figure S2, both MPM and THP1 macrophages showed immunoreactivity for TWIK2 protein confirming its expression in these cell types. Collectively, these data suggest that while TWIK2 is expressed in MPMs and THP1 macrophages, its inhibition by quinine had negligible effect on ATP or Nigericin stimulated IL-1β secretion from LPS-primed cells.

Figure 8. Pharmacological inhibition of TWIK2 channel by quinine does not affect Nigericin or ATP stimulated IL-1β secretion from LPS-primed cells.

MPMs (Panel A) or THP1 macrophages (Panel B) were treated with LPS (50 ng/ml) overnight and then pre-treated with indicated K+ channel inhibitors for 1h. Nigericin (N, 20 µM for 20 min) or ATP (5 mM for 30 min) induced IL-1β secretion was determined. Data (Mean ± SD, n=6) are presented as %LPS + Nigericin or ATP. *P<0.05.

Discussion

Although positive correlation between cellular cholesterol content and inflammatory status, especially of macrophages, is well established, the underlying mechanisms are not completely defined. The two likely mechanisms studied thus far are increased inflammatory signaling via TLR4 due to deficiency of cholesterol transporters ABCA1 or ABCG1 and subsequent increase in lipid raft cholesterol (5) and increased inflammasome activation via intracellular cholesterol crystal-dependent lysosomal destabilization (31). Data presented herein establishes cholesterol-dependent modulation of K⁺ efflux as an additional mechanism by which intracellular cholesterol levels modulate secretion of pro-inflammatory cytokine IL-1β. Single cell patch clamp studies show that K⁺ currents were modulated by changes in cellular cholesterol content resulting in corresponding changes in cellular K⁺ levels as well as secretion of IL-1β. Furthermore, these data also indicate that direct inhibition of K⁺ efflux can attenuate IL-1β secretion irrespective of the second signal (e.g., crystal-mediated lysosomal destabilization or direct stimulation of K⁺ efflux by K⁺ ionophore such as Nigericin) of inflammasome activation underscoring the importance of processes that affect K⁺ currents and the resulting K⁺ efflux (e.g., changes in cellular cholesterol levels) in regulating inflammation.

Secretion of IL-1β is under very tight regulation and requires priming (signal 1 for pro IL-1β expression) and activation (signal 2 for Caspase-1 dependent cleavage of pro IL-1β for generation of mature IL-1β that is secreted). Therefore, the biological end point and measure of inflammasome activation is the presence of secreted IL-1β. Diverse DAMPs have been identified as bona fide Signal 2. Increase in intracellular cholesterol levels and subsequent lysosomal disruption by the likely formation of cholesterol crystals has been described as one of the potential mechanisms by which intracellular cholesterol levels act as the signal 2 for inflammasome activation (31). However, K⁺ efflux is thought to be a common denominator for Caspase-1 activation by diverse DAMPs and decreasing intracellular K⁺ is sufficient to activate NLRP3 inflammasome in vitro (32). Consistently, exposure to ALUM or Nigericin led to a comparable decrease in cellular K⁺ levels and IL-1β secretion was equally attenuated when K⁺ efflux was inhibited (Figure 5). These data provide strong support for the hypothesis that cellular perturbations that can affect intracellular K⁺ levels would modulate Caspase-1 activation and subsequent IL-1β secretion. By modulating the activity of K⁺ channels and thereby altering intracellular K⁺ levels (Figures 1–3), cellular cholesterol, therefore, serves as the signal 2 of inflammasome activation.

The extreme hydrophobicity of cholesterol limits the pathways by which cellular cholesterol can potentially exert diverse effects and changes in membrane properties leading to subsequent changes in activity/function of membrane associated proteins is often the most likely mechanism. For example, cellular cholesterol homeostasis is regulated by the cholesterol content of the ER membrane via controlled processing of SREBPs (33). Similarly, increased accumulation of cholesterol due to deficiency of cholesterol transporters, namely, ABCA1 or ABCG1 is shown to increase lipid raft cholesterol resulting in increased surface expression of TLR4 and increased LPS-induced downstream signaling (5). It should also be noted that although excess cholesterol is stored as CE within cytoplasmic lipid droplets, a continuous intracellular CE cycle maintains a constant flux of FC generated by CE hydrolysis between cellular membranes and lipid droplets (12) that can modulate the function of membrane proteins/channels. The observed differences in total K⁺ currents in response to cholesterol loading, either by WD feeding (Figure 1) or direct AcLDL loading (Figure 3) suggest that K⁺ channel activity is modulated by cellular cholesterol content. K⁺ channels are associated with the plasma membrane and are often regulated by membrane cholesterol content (34) and lipid composition (35) and changes in cellular cholesterol levels affect the activity of K⁺ channels (36). Furthermore, Martens et al described differential targeting of K⁺ channel subtypes to raft or non-raft regions within a single membrane as a unique compartmentalization-dependent regulation of K⁺ channel activity (37). Differential targeting of the same K⁺ channel to raft or non-raft regions in different cell types is also described (38). Two different mechanisms are currently thought to be responsible for the sensitivity of ion channels to cholesterol, namely, specific cholesterol-channel protein interactions or regulation of channels by the physical properties of the membrane bilayer (39). The identity of the K⁺ channel(s) involved in K⁺ efflux central to inflammasome activation is currently not established precluding the identification of the mechanism by which cholesterol modulates the K⁺ currents. Inhibition profiling shown in Figure 6, provides evidence for the involvement of Kir6 channels in regulating IL-1β secretion. However, it should be noted that in the present study we measured K⁺ currents under physiological conditions and did not attempt to characterize Kir6 channels by electrophysiological methods. Therefore, the data presented do not rule out the possible involvement of other K⁺ channels associated with MPMs or THP1 macrophages. Future studies with targeted genetic ablation of Kir6 channels are likely to determine the relative role of these channels in K⁺ efflux and resulting inflammasome activation.

The purinergic P2X7 receptor, ligated by DAMPs such as ATP (40), is also a non-selective cation channel with a pore forming motif resembling K⁺ channels (41). Thus, it has been proposed to be the putative K⁺ channel mediating inflammasome activation (42). However, this would require P2X7 to act both as a DAMP (e.g., ATP) sensor as well as K⁺ channel. Di et al recently reported the identification of TWIK2 instead of P2X7 as the likely ATP-induced K⁺ efflux channel in macrophages (20). However, in the present study pharmacological inhibition of TWIK2 with quinine did not reduce ATP or Nigericin-induced IL-1β secretion from LPS-primed MPMs or THP1 macrophages although expression of TWIK2 in these cells was confirmed both at mRNA and protein levels. Targeted future studies are required to establish the molecular identity of the K⁺ channel(s) involved in inflammasome activation and are likely to provide more detailed information on the mechanisms involved in cellular cholesterol-dependent regulation of K⁺ currents and K⁺ efflux.

It is noteworthy that Kir6 inhibitors tested here are used as short acting insulin secretagogues that have also been shown to reduce systemic inflammation (22, 43) and Assloni et al showed that controlling post-prandial hyperglycemia by mitiglinide improves cluster of oxidative and inflammatory markers in diabetic patients (44). However, it remains unclear whether these effects are direct or reduction in inflammation is an indirect consequence of reduced glucose levels (45).

The causal relationship between high cholesterol levels and K⁺ currents and their role in several pathophysiological processes is increasingly being recognized. Cholesterol enrichment-dependent increase in amplitude of outward K⁺ currents in inner ear hair cells is thought to link serum hypercholesterolemia to hearing loss (46). Similarly, high fat diet mediated increase in K⁺ currents in atrial myocytes links obesity-related arrhythmias to lipid-induced changes in K⁺ channel modulation (47). Deng et al have also reported upregulation of K_Ach cardiac currents by hypercholesterolemia (48). The data presented herein adds to this growing list of pathophysiological processes linked to hypercholesterolemia by modulation of K⁺ channel activity. In this regard, it is noteworthy that targeted inhibition of K⁺ channels is being evaluated as a therapy for inflammation linked autoimmune diseases (49).

In summary, modulation of macrophage cholesterol content either by WD-feeding or direct loading with modified LDL (AcLDL) or by stimulating cholesterol efflux (by transgenic over-expression of CEH) regulates K⁺ currents resulting in changes in cellular K⁺ levels. While disparate intracellular danger signals with little or no molecular or chemical similarity act as signal 2 for inflammasome activation, Caspase-1 dependent cleavage of pro-IL-1β is directly regulated by cellular K⁺ levels emphasizing the importance of regulating this critical step. The data presented herein establishes a link between cellular cholesterol levels, K⁺ currents, intracellular K⁺ levels and IL-1β secretion identifying yet another important mechanism by which cellular cholesterol levels are linked to the cellular inflammatory status. Continuing characterization of these underlying mechanisms and identification of specific K⁺ channels is likely to identify novel pathways to decrease or resolve chronic inflammation associated with hypercholesterolemia and metabolic inflammation-associated diseases such as atherosclerosis, diabetes and hepatic steatosis.

Abbreviations:

ABCA1

ATP-binding cassette transporter A1

ABCG1

ATP-binding cassette transporter G1

AcLDL

Acetylated low-density lipoprotein

CE

Cholesteryl esters

CEH

Cholesteryl ester hydrolase

CEHTg

CEH transgenic

DAMP

Danger-associated molecular patterns

ER

Endoplasmic reticulum

FC

Un-esterified or free cholesterol

HDL

High density lipoprotein

IL-1β

Interleukin 1β

LPS

lipopolysaccharide

MPM

Mouse peritoneal macrophages

NF-κB

Nuclear factor kappa B

NLRP3

Nucleotide-binding oligomerization domain, Leucine rich Repeat and Pyrin domain containing

PRR

Pattern recognition receptor

ROS

Reactive oxygen species

TLR4

Toll like receptor 4

WD

High fat high cholesterol containing western diet (TD88137)