Acute Hypoxia Inhibits K_2P3/9 (TASK1/3) Independently of Cytosolic ATP in Carotid Body Type 1 Cells

ABSTRACT

Hypoxia stimulates ventilation through a chemosensory reflex that is initiated by inhibition of the O₂‐sensitive K⁺ channels such as K_2P3/9 in carotid body Type 1 glomus (CB) cells. Because K_2P3/9 activity is sensitive to cytosolic ATP concentration ([ATP]_cyt, K_1/2 ≈ 2.5 mM), it has been proposed that hypoxia suppresses mitochondrial ATP production and thereby reduces [ATP]_cyt to inhibit K_2P3/9 activity. To test this hypothesis, we developed a K_ATP channel‐based bioassay using a patch‐cramming approach to measure [ATP]_cyt in living cells. Co‐expression of Kir6.2(ΔC36‐R50G) and SUR1(G1485D) in HEK293 cells formed a functional K_ATP channel (designated R50G*) with markedly reduced ATP sensitivity. In inside‐out patches, normalized R50G* (${\mathrm{R}50\mathrm{G}}_{\mathrm{NL}}^{*}$) activity showed ~100‐fold lower ATP sensitivity (K_1/2 ≈ 2.0 mM) than that of the wild‐type K_ATP (K_1/2 ≈ 20 μM). Inside‐out patches from HEK293 cells containing R50G* were crammed into CB cells under normoxia (21% O₂), exposing the patch to the CB cell cytoplasm. Based on ${\mathrm{R}50\mathrm{G}}_{\mathrm{NL}}^{*}$ activity in crammed CB cells, [ATP]_cyt was estimated to be ~4.9 mM in normoxia. In CB cells, hypoxia (1% O₂, 15 s) did not affect ${\mathrm{R}50\mathrm{G}}_{\mathrm{NL}}^{*}$ activity, indicating that [ATP]_cyt remained stable during hypoxia. FCCP (1 μM), a mitochondrial uncoupler, increased ${\mathrm{R}50\mathrm{G}}_{\mathrm{NL}}^{*}$ activity, which corresponded to [ATP]_cyt of ~3.7 mM. As hypoxia (~1% O₂) inhibited K_2P3/9 activity by 65% within 5 s, our result demonstrates that inhibition of K_2P3/9 by hypoxia is independent of [ATP]_cyt and that ATP depletion is unlikely to be the initiating signal for hypoxia‐induced excitation of CB cells.

Practioner Points

• Acute hypoxia first inhibits O₂‐ and ATP‐sensitive K_2P3/9 (TASK1/3) in carotid body Type 1 cells, which initiates the arterial chemoreflex pathway leading to augmented ventilation.

• A possible mechanism by which acute hypoxia inhibits K_2P3/9 is a decrease in cytosolic [ATP]. Test of this mechanism using a K_ATP channel‐based bioassay together with patch cramming shows that cytosolic [ATP] is reduced by mitochondrial uncoupling, but unaffected by acute hypoxia.

• Thus, acute O₂ sensing by carotid body Type 1 cells occurs by a mechanism independent of cytosolic [ATP], and probably involves other mitochondrial signal(s) produced by acute hypoxia.

Abbreviations

[ATP]

ATP concentration

[Ca²⁺]_cyt

cytosolic calcium concentration

Cav

voltage‐dependent Ca²⁺ channel

CB cell

carotid body glomus cell

CGS

cystathionine‐γ‐synthase

COX

cytochrome c oxidase

FCCP

(Carbonyl cyanide 4‐(trifluoromethoxy)phenylhydrazone)

H₂S

hydrogen sulfide

HIF

Hypoxia‐Inducible Factor

K_2P3/9

heteromeric two‐pore domain acid‐sensing K⁺ channels (also known as TASK‐1/3)

Kir

inwardly rectifying K⁺ channel

NADH

nicotinamide adenine dinucleotide

NADPH

nicotinamide adenine dinucleotide phosphate

NOX4

nicotinamide adenine dinucleotide phosphate oxidase‐4

R50G*

Kir6.2(ΔC36‐R50G)‐SUR1(G1485D)

ROS

reactive oxygen species

SUR

sulfonylurea receptor

1. Introduction

Hypoxia inhibits the resting K⁺ conductance and depolarizes carotid body chemoreceptor Type 1 (CB) cells. This depolarization leads to the opening of voltage‐dependent Ca²⁺ channels (Cav) and elevation of cytosolic Ca²⁺ concentration ([Ca²⁺]_cyt), which augments the secretory activity of CB cells. The released transmitters activate receptors on afferent carotid sinus nerve terminals, resulting in increased ventilation and autonomic nervous system activity via neurons in the medullary respiratory and cardiovascular centers [1]. In CB cells, a major target of hypoxia is K_2P3/9 (also known as TASK1/3), the primary background O₂‐sensitive K⁺ channel that is highly expressed in these cells [2, 3]. The mechanism by which hypoxia inhibits K_2P3/9 is still not well understood.

One proposed mechanism is that hypoxia increases the generation of reactive oxygen species (ROS) and reduced form of nicotinamide adenine dinucleotide (NADH) from mitochondrial complex I, which inhibits K⁺ channels such as K_2P3/9 [4, 5]. A recent study showed that HIF2α is crucial for hypoxia sensing in CB cells, which involves the generation of ROS and NADH associated with HIF2α‐dependent expression of atypical mitochondrial electron transport chain subunits [6]. Mice lacking HIF2α showed no ventilatory response to hypoxia. Transcriptome analysis of O₂‐sensing CB cells also revealed a strong expression of HIF2α [6, 7]. In an earlier study, ROS species such as hydrogen peroxide and superoxide radicals, when directly applied at relatively high concentrations to the cytoplasmic side of inside‐out patches from CB cells, did not inhibit K_2P3/9 [8]. In contrast, in HEK cells expressing K_2P9 (TASK3), low levels of ROS and NADH were found to inhibit the TASK3 current [9], suggesting a possible biphasic effect of these molecules on TASKs. The effects of physiological levels of these signals on O₂‐sensitive K⁺ channels in CB cells will need to be determined to help identify their role in hypoxia sensing.

Another proposed mechanism involves a direct interaction between NADPH oxidase‐4 (NOX4) and K_2P3 (TASK1), leading to modulation of channel activity independently of ROS generation [10]. However, inhibition of K_2P3/9 by hypoxia has not been observed in heterologous expression systems such as HEK293 cells [8, 11], suggesting that cell‐specific redox signaling mechanisms required for hypoxia sensitivity may be absent in these systems. Consequently, the role of NOX4 in regulating K_2P3/9 under physiological hypoxic conditions in CB cells remains unclear and warrants additional investigation.

A distinct mechanism for hypoxia‐induced depolarization that does not involve K_2P3/9 has also been proposed. In this model, hypoxia increases hydrogen sulfide (H₂S) production, which activates the lactate receptor OlfR78, leading to increased cAMP levels and activation of cAMP‐gated channels via Gα_olf signaling [12, 13]. However, this pathway remains speculative, as the expression of cAMP‐gated channels in CB cells has not been established. Moreover, studies using pharmacological inhibitors of H₂S production and mice with a deletion of the cystathionine‐γ‐synthase gene (a key H₂S‐producing enzyme in CB cells) reported no effect on hypoxia‐induced inhibition of K_2P3/9 or on the elevation of [Ca²⁺]_cyt [2, 14]. These findings argue against a major role for H₂S in hypoxic inhibition of K_2P3/9. As H₂S itself is a potent inhibitor of mitochondrial cytochrome c oxidase [14], this raises the possibility that its effect on K_2P3/9 may be mediated indirectly through mitochondrial inhibition. This would be similar to the action of cyanide on mitochondrial cytochrome oxidase, which also suppresses K_2P3/9 activity [15].

Another proposed mechanism for hypoxia‐induced inhibition of K_2P3/9 is that hypoxia reduces mitochondrial ATP production and [ATP]_cyt, thereby decreasing K_2P3/9 activity [16]. This mechanism is quite plausible because K_2P3/9 activity has been reported to be modulated by [ATP]_cyt in excised patches [17]. Acute hypoxia has been shown to inhibit K_2P3/9 (< 5 s) rapidly. Therefore, if a decrease in [ATP]_cyt is the primary mechanism for hypoxia‐induced inhibition of K_2P3/9, hypoxia is expected to reduce [ATP]_cyt within a few seconds as well. A recent study using cultured CB cells expressing a genetically encoded fluorescent protein found that hypoxia increased the secretory activity of these cells without altering the cytosolic [ATP]/[ADP] ratio [9], suggesting that hypoxia sensing did not involve a change in [ATP]_cyt. Thus, the role of ATP in the hypoxia‐induced inhibition of K_2P3/9 remains controversial.

In the present study, we tested the hypothesis that hypoxia reduces [ATP]_cyt in freshly isolated CB cells, thereby accounting for the inhibition of K_2P3/9. To address this mechanism, we studied the temporal relationship between hypoxia and K_2P3/9 inhibition in more detail. We employed a mutant K_ATP channel‐based bioassay that involves patch‐cramming to estimate [ATP]_cyt in CB cells in normoxia and acute hypoxia. This strategy enabled a direct evaluation of whether acute hypoxia is associated with detectable changes in [ATP]_cyt under conditions known to suppress K_2P3/9 activity. Our study indicated that hypoxic inhibition of K_2P3/9 is not accompanied by a change in [ATP]_cyt, suggesting that ATP depletion is not the mechanism linking hypoxia to K_2P3/9 inhibition and initiating chemosensory signaling in CB cells.

2. Methods

2.1. Chemicals

All chemicals used in this study were purchased from Sigma‐Aldrich (St. Louis, MO, USA) unless stated otherwise.

2.2. Animals and Ethical Approval

All experiments were carried out according to the guidelines set by the Animal Care and Use Committee of Rosalind Franklin University under Institutional Animal Care and Use Committee protocol #18‐02, and conform to the principles and regulations as described by Grundy [18]. The animal care and use program is accredited by the American Association for Accreditation of Laboratory Animal Care (AAALAC) and approved by the Office of Laboratory Animal Welfare (A3279‐01). There are no ethical concerns. Pregnant Sprague–Dawley rats (16 days) were purchased from Envigo (Indianapolis, IN, USA) and acclimatized in‐house with the pups for ~3 weeks. Standard laboratory diet (Teklad Global 19% Protein Extruded Rodent Diet, Harlan #2019S) and drinking water were provided.

2.3. Cell Isolation

Rats (~3 weeks old) of either sex were deeply anesthetized by inhalation of isoflurane until cessation of breathing, and decapitated with a sharp razor blade. Six to eight CBs from the region of the carotid artery bifurcation were removed from 3 to 4 rats and placed in ice‐cold, low Ca²⁺/low Mg²⁺ phosphate‐buffered saline (low Ca²⁺/Mg²⁺ PBS: 137 mM NaCl, 2.8 mM KCl, 2 mM KH₂PO₄, 0.07 mM CaCl₂, 0.05 mM MgCl₂, pH 7.4). Each CB was cut into several pieces and placed in a solution containing trypsin (0.4 mg ml⁻¹) and collagenase (0.4 mg ml⁻¹) in low Ca²⁺/Mg²⁺ PBS and incubated at 37°C for ~25 min. Following trituration using a fire‐polished glass pipette tip, the dispersed cells were separated by centrifugation at low speed (1200 g for 5 min) and suspended in CB growth medium [Ham's F‐12, 10% fetal bovine serum, 23 mM glucose, 4 mM Glutamax‐I (L‐alanyl glutamine), 10 kU penicillin, 10 kU streptomycin, and 300 μg mL⁻¹ insulin]. Cells were plated on poly‐L‐lysine‐pretreated glass coverslips and incubated at 37°C for ~3 h in a humidified atmosphere of 95% air/5% CO₂. Cells were used the same day, within 6 h after plating.

2.4. Generation of K_ATP Channel Mutants

Mutant constructs of rat Kir6.2 (AB043638) and SUR1 (L40624) were generated by an external service provider (Cosmo Genetech, Seoul, Republic of Korea) using a PCR‐based site‐directed mutagenesis approach. The templates used for mutagenesis were Kir6.2 and SUR1 cDNAs cloned into pcDNA3.1(+) expression vectors. For Kir6.2, in addition to the wild‐type construct, three mutant constructs were generated. Each mutant lacked the terminal 36 amino acid residues at the C‐terminus (designated ΔC36) and contained a single amino acid substitution: Kir6.2ΔC36‐C166M, Kir6.2ΔC36‐I167M, and Kir6.2ΔC36‐R50G. The ΔC36 truncation was introduced by PCR using primers containing a premature stop codon positioned 36 residues upstream of the native termination site, and point mutations were introduced individually using mutagenic primers. The SUR1 subunit was used either as wild‐type or as a mutant (G1485D), which abolishes ADP sensitivity. The presence of the intended mutations and the absence of undesired mutations were confirmed by Sanger sequencing before use.

2.5. Transfection

HEK293 cells were co‐transfected with Kir6.2 and SUR1 constructs using Lipofectamine 3000 Transfection Reagent (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's protocol. Cells were maintained in DMEM supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin at 37°C in a humidified 5% CO₂ incubator, and electrophysiological recordings were performed 24–48 h after transfection.

2.6. Electrophysiology and Patch‐Cramming Protocol

Cell‐attached and inside‐out patch recordings were performed using a patch‐clamp amplifier (Axopatch 200B, Molecular Devices, San Jose, CA, USA) and analyzed with the pClamp program (version 10). The cell‐attached patch was first formed with a sylgard‐coated, tip‐polished, borosilicate glass pipette. The single‐channel current was initially filtered at 1 kHz using an eight‐pole Bessel filter (−3 dB) and transferred to a computer using the Digidata1440 interface. For studies of K_ATP channels in HEK293 cells and studies on the effect of cytosolic ATP on TASK in CB cells, the pipette and bath solutions contained (mM): 140 KCl, 1 MgCl₂, 5 EGTA, 11 glucose, and 10 HEPES (pH 7.4). Pipette potential was set at 60 mV to record inward single‐channel current. For studies on the effect of hypoxia on TASK channels in CB cells, the pipette solution contained (mM): 140 KCl, 1 MgCl₂, 5 EGTA, 11 glucose, and 10 HEPES (pH 7.4; high‐K⁺ solution) and the bath solution contained (mM): 117 NaCl, 5 KCl, 23 NaHCO₃, 1 MgCl₂, 1 CaCl₂, 10 HEPES, and 11 glucose (physiological solution). For cramming studies, inside‐out patches from HEK293 cells expressing R50G* channels were obtained in high‐K⁺ pipette and bath solutions, with the pipette potential set to +30 or +40 mV for recording inward single‐channel currents. Immediately upon the CB cells being crammed, the bath perfusion solution was switched to the physiological solution, and the pipette potential was set to 0 mV to record the R50G* inward current.

For hypoxia studies, the bath solution was continuously bubbled with a gas mixture (pH 7.4) containing 5% CO₂/21% O₂/74% N₂ (normoxia). The temperature of the bath perfusion solution at the recording chamber was maintained at 36°C–37°C for all experiments. Channel openings were analyzed to obtain channel activity (NP_o, where N is the number of channels in the patch, and P_o is the probability of a channel being open). Channel activity was determined from ~10 s of the current recording. Because CB cells express both ~16‐pS (TASK‐1) and ~34‐pS (TASK3 and TASK1/3) channels that we collectively refer to as K_2P3/9 or TASK1/3 [19], analysis for TASKs was performed by setting the open levels to multiples of the 16‐pS channel to detect all K_2P3/9 isoforms. Under the ionic conditions we used, only TASK channels are visibly active at rest in cell‐attached and inside‐out patch configurations (with ATP in the bath solution), allowing a clear comparison of relative channel activity before and after a stimulus. A1899 (Tocris Biosciences, Bristol, UK) was used to inhibit TASKs when necessary. In some CB cell patches, a non‐selective cation channel (NSC) with a long open time duration (> 100 ms) is also present, but NSC appears only after activation by increased [Ca²⁺]_cyt produced by stimuli such as hypoxia and mitochondrial inhibitors, and is insensitive to ATP [15, 20]. Channel analysis was performed using patches containing only TASKs, without NSC.

2.7. Measurement of [Ca²⁺]_cyt

Isolated cells plated on glass coverslips were incubated with 2 μM fura‐2 AM (acetoxymethyl ester; Thermo Fisher Scientific, Waltham, MA, USA) for 30 min at 37°C in culture medium. A coverslip with attached cells was placed in a recording chamber positioned on the stage of an inverted microscope (IX71; Olympus America Inc., Center Valley, PA, USA). Fura‐2 was alternately excited at 340 and 380 nm, and the emitted fluorescence over a 2 s period per data point was filtered at 510 nm and recorded using a charge‐coupled device (CCD)‐based imaging system running SimplePCI software (Hamamatsu Corp., Hamamatsu City, Shizuoka, Japan). The cells in the recording chamber were continuously perfused with a solution containing (mM): 117 NaCl, 5 KCl, 23 NaHCO₃, 1 MgCl₂, 1 CaCl₂, and 11 glucose (pH 7.4). Ratiometric data were calibrated by applying experimentally determined constants to the equation: [Ca²⁺] = K_d × β × (R − R _min)/(R _max − R) [21]. Values for R _max (12.0), R _min (0.1), and β (11.6) were determined in vitro, and a K_d value of 300 nM for fura‐2 was assumed [22]. Each experiment was performed using coverslips of cells prepared on different days, and no more than 2 coverslips were used per day.

2.8. Controlled Generation and Calibration of Hypoxic Conditions

In experiments conducted using a recording chamber open to the atmosphere, defined levels of hypoxia were generated by bubbling physiological bicarbonate‐buffered solutions with specific gas mixtures. Solutions were equilibrated with 5% CO₂ balanced with N₂ and supplemented with glucose oxidase (2 U/mL) and catalase (100 U/mL) to achieve anoxia. Severe hypoxia was generated using a gas mixture (Linde, Waukegan, IL, USA) of 5% CO₂/1% O₂/94% N₂, whereas normoxic conditions were maintained with 5% CO₂/21% O₂/74% N₂. The composition of the solutions was identical to that used in [Ca²⁺]_cyt measurement experiments. Reproducible oxygen levels within the recording chamber were confirmed using an oxygen meter (ISO2; World Precision Instruments, Sarasota, FL, USA). The temperature of the perfusion solutions was maintained at 36°C–37°C throughout hypoxia experiments. Across three independent trials, mean oxygen levels in the recording chamber were 0.4% (3 mmHg) for anoxia, 1.6% (12 mmHg) for severe hypoxia, and 21% (160 mmHg) for normoxia. The oxygen meter was calibrated to 0% using solutions bubbled with pure nitrogen for 60 min and to 21% using solutions gassed with air for 30 min at 37°C.

2.9. Data Analysis

The Hill equation (Y = [ATP] ⁿ /(K_0.5 + [ATP] ⁿ ); Y is the relative channel activity, n is the Hill coefficient, and K_1/2 is the concentration of ATP that produces half maximal response) was applied using the Origin 6.0 software (OriginLab Co., Northampton, MA, USA). To determine statistical significance, Student's t‐test (for comparison of two dot plots) and one‐way analysis of variance (for comparison of three or more sets of data) were performed using PRISM software (GraphPad, Boston, MA, USA). Post hoc testing following analysis of variance was based on the paired (or unpaired when comparing data sets from different coverslips) t‐test with Bonferroni correction. p values less than 0.05 were considered significant.

3. Results

3.1. Sensitivity of K_2P3/9 to [ATP]_cyt

The proposed mechanism in which hypoxia inhibits K_2P3/9 via reduction in [ATP]_cyt is based on the central premise that the K⁺ channel is sensitive to changes in [ATP]_cyt within the physiological range. Although the general effect of ATP on the background K⁺ channel in CB cells has been reported earlier in one study at ~24°C [16], this has not been confirmed and studied in detail by others. Here, we used a physiological temperature (37°C) and ionic conditions to study the effect of ATP on K_2P3/9 expressed in CB cells.

In cell‐attached patches, K_2P3/9 was highly active at rest in every CB cell tested and was the primary channel type that was present, as described in earlier studies ([2, 14, 16, 23, 24]; Figure 1Aa). K_2P3/9 activity declined rapidly upon formation of inside‐out patches (Figure 1Ab). When ATP was added back to the perfusion solution, the activity of K⁺ channels with kinetic properties (single‐channel conductance of ~32 pS and mean open time durations of 0.6–0.8 ms) similar to those observed in cell‐attached patches recovered fully (Figure 1Ac). No other types of ion channels were activated by the addition of ATP to the bath solution in CB cells. These results indicate that the changes in [ATP] were responsible for the decline and recovery of K_2P3/9 activity in inside‐out patches (Figure 1A). Cramming of the inside‐out patch into a CB cell also recovered K_2P3/9 activity, most likely due to the exposure of the inside‐out patch to high [ATP] that is present in the cell (Figure 1Bc). These findings are summarized in Figure 1C.

FIGURE 1.

K_2P3/9 sensitivity to [ATP]_cyt. (A) A cell‐attached (C/A) patch recording from a CB cell shows robust K_2P3/9 channel activity. Transition to the inside‐out (I/O) patch configuration, which removes cytosolic ATP, resulted in a marked reduction in K_2P3/9 activity. Subsequent addition of 4 mM ATP to the bath solution fully restored K_2P3/9 activity. Expanded single‐channel recordings are shown for (a) C/A mode, (b) I/O mode, and (c) I/O mode with 4 mM ATP. (B) Physical apposition (cramming) of the I/O patch onto the original CB cell membrane restored K_2P3/9 activity. Expanded single‐channel recordings (filtered at 1 KHz) are shown for (a) C/A mode, (b) I/O mode, and (c) cramming after I/O. (C) Summary graph showing the effects of C/A, I/O formation, ATP application, and cramming (cram) on K_2P3/9 activity (n = 5). (D) In I/O patches, K_2P3/9 (TASK) activity was plotted as a function of ATP concentration, revealing an approximately linear relationship between ATP levels (0–6 mM) and channel activity. (E) Addition of 1 μM A1899 to I/O patches caused inhibition of the channels activated by 4 mM ATP in I/O patches (n = 4). (F) Summary graph on the effect of ATP and A1899 on K_2P3/9 activity. Data are expressed as NPo (mean ± SD). *p < 0.05 compared with the corresponding control. ^† p < 0.05 compared with the ATP‐treated group. Arrows indicate transitions between recording conditions.

To assess in more detail the sensitivity of K_2P3/9 to ATP, inside‐out patches were exposed to ATP concentrations ranging from 0 to 6 mM. The relationship between K_2P3/9 activity and [ATP]_bath was nearly linear up to 6 mM. This experiment showed that a 56% reduction of [ATP]_bath would be necessary to inhibit K_2P3/9 by 50% (Figure 1D). In inside‐out patches with ATP in the bath solution, application of A1899 (1 mM), an inhibitor of K_2P3/9 [23, 25], decreased the channel activity by 82.2% ± 13.4% (Figure 1E,F). This information was used to assess whether hypoxia‐induced inhibition of K_2P3/9 was due to a reduction in [ATP]_cyt in CB cells (see sections below).

3.2. Acute Hypoxia Causes a Rapid Decrease in K_2P3/9 and a Rapid Increase in [Ca²⁺]_cyt

Hypoxia has been reported to decrease K_2P3/9 activity [26]. To determine the time course of inhibition in detail, cell‐attached patches from CB cells were perfused at a rapid flow rate with normoxic (21% O₂) and hypoxic (1% O₂) solutions while recording K_2P3/9 (Figure 2A). Acute hypoxia produced a very rapid inhibition of K_2P3/9 such that the activity was reduced by 65% ± 7% (n = 5) within 5 s. K_2P3/9 remained inhibited by hypoxia without any sign of desensitization for at least 5 min of recording. The effect of hypoxia was fully reversed upon exposure of the patch to normoxia, at similar rapid rates. A similar rapid and reversible effect on K_2P3/9 was also observed in response to anoxia (0% O₂), with a more potent inhibition (84% ± 11%; n = 5). The reduction in single‐channel amplitude observed in response to hypoxia and anoxia is expected, reflecting K_2P3/9 inhibition that depolarizes the cell and reduces the transmembrane potential gradient, as previously described [27].

FIGURE 2.

Effect of hypoxia (1% O₂) and anoxia (0% O₂) on K_2P3/9 activity in CB cells. (A) Cells in the cell‐attached state are perfused with normoxic solution at 0 mV pipette potential and 37°C. Cell‐attached patch shows a high basal level of K_2P3/9 activity. Switching to hypoxia (top) or anoxia (bottom) solution produced a rapid and reversible decrease in K_2P3/9 activity. Expanded current tracings (filtered at 1 kHz) show the apparent effects of hypoxia and anoxia on the frequency of K_2P3/9 openings. The reduction in single‐channel conductance is also present due to cell depolarization caused by K_2P3/9 inhibition. (B) Effect of high KCl (20 mM), severe hypoxia (1% O₂), and anoxia (~0% O₂) on [Ca²⁺]_cyt in CB cells. Cells in the recording chamber are perfused with normoxic solution at 37°C. Cells are briefly perfused with KCl (20 mM), hypoxia, and anoxia solutions. Individual cell responses are shown on top, and averaged responses are displayed on the bottom. Representative tracings from five experiments show the rapid onset and offset of Ca²⁺ responses.

Similar to the rapid effect of acute hypoxia (1% O₂) and anoxia (0% O₂) on K_2P3/9, hypoxia and anoxia produced rapid and reversible increases in [Ca²⁺]_cyt in CB cells (Figure 2B). The Ca²⁺ signal shown on the top panel shows changes in [Ca²⁺]_cyt in individual cells in response to 20 mM KCl, hypoxia, and anoxia. The averaged Ca²⁺ response is shown on the bottom panel. The effect of hypoxia on [Ca²⁺]_cyt has previously been shown to be due to cell depolarization that opens Cav [28]. These findings give us confidence that our freshly isolated CB cell preparations are highly responsive to acute hypoxia.

3.3. Differential ATP Sensitivity of K_ATP Channel Mutants

To determine whether acute hypoxia reduces [ATP]_cyt to inhibit K_2P3/9, we developed a bioassay to estimate [ATP]_cyt in CB cells in real‐time. As a starting point, Kir6.2 and SUR1 were co‐expressed in HEK293 cells to form a functional K_ATP channel [29]. Wild‐type Kir6.2/SUR1 showed high ATP sensitivity, with a K_1/2 of ~20 μM [30]. As shown in Figure 3A, representative current traces illustrate the wild‐type KATP channel's response to varying [ATP] in the bath solution perfusing an inside‐out patch.

FIGURE 3.

Effect of ATP on wild‐type and four K_ATP mutants expressed in HEK293 cells. (A) An inside‐out patch shows K_ATP activity of the wild‐type K_ATP channel (Kir6.2/SUR1) recorded at different [ATP]. (B) Schematic diagram of Kir6.2 variants. Full‐length wild‐type (WT) Kir6.2 consists of 390 amino acids and contains two transmembrane segments (M1 and M2) and a pore‐forming H5 region. C‐terminally truncated Kir6.2 lacking the final 36 amino acids (Kir6.2ΔC36) carrying the C166M mutation, I167M mutation, R50G mutation, and SUR1‐G1485D are shown. The SUR1mutant (G1485D) lacks sensitivity to ADP. (C) Relative K_ATP activity from wild‐type and four mutant K_ATP channels is plotted as a function of [ATP]. The activity observed at 0.001 mM ATP was taken as the maximum (1.0). Each point is the mean of 4 determinations. Standard deviation (not shown) was < 18% of the mean for each value. (D) ATP‐dependent channel activity of the R50G* mutant. An inside‐out patch shows K_ATP activity of the Kir6.2ΔC36‐R50G/SUR1‐G1485D (R50G*) mutant recorded at different [ATP] in the perfusion solution. The current was expanded and filtered at 0.6 kHz. (E). Relative K_ATP activities of the wild‐type, C166M, and R50G* mutants are plotted as a function of [ATP] (n = 3). The activity observed at 0.001 mM ATP was taken as the maximum (1.0). The data from the wild‐type K_ATP channel are shown as a solid square (K_1/2 = 20 μM). Each point is the mean of 3 determinations. Standard deviations (not shown) are < 20% of the mean value.

As [ATP]_cyt is believed to be in the millimolar range in most types of cells [31], it was necessary to construct a mutant K_ATP channel with K_1/2 that is also in the millimolar range. We used information from earlier studies reporting several mutations in Kir6.2 and SUR1 that form K_ATP channels with high K_1/2 values [32, 33, 34]. Four such mutants were constructed and co‐expressed in HEK293 cells. For Kir6.2, we deleted 36 amino acid residues from the C‐terminus, in addition to a single amino acid mutation (Figure 3B). SUR1 was used either as wild‐type or as a single‐point mutant (G1485D) that abolished sensitivity to ADP [35]. These mutants were co‐expressed in HEK293 cells, and their response to ATP was assessed in inside‐out patches perfused with solutions containing different ATP concentrations, with 1 mM free Mg²⁺. Figure 3C shows [ATP]‐response (channel activity) relationships for wild‐type and four mutant K_ATP channels. Wild‐type Kir6.2/SUR1 channels exhibited high ATP sensitivity, with a K_1/2 in the low micromolar range. In contrast, the Kir6.2ΔC‐C166M and Kir6.2ΔC‐I167M mutants showed rightward shifts in ATP sensitivity but retained K_1/2 values below 1 mM. The Kir6.2ΔC‐R50G/SUR1 mutant displayed a further rightward shift, indicating a substantial decrease in ATP sensitivity. Notably, co‐expression of Kir6.2ΔC‐R50G with SUR1‐G1485D mutation (hereon referred to as R50G*) produced the most shift, yielding a K_1/2 in the low millimolar range (~2 mM), within the physiological range of [ATP]_cyt.

R50G* mutant formed a channel with a K_1/2 of ~2.0 mM. A representative current response of the R50G* shows ~100‐fold lower ATP sensitivity than that of the wild‐type (Figure 3D,E). The K_1/2 values of Kir6.2ΔC‐C166M and Kir6.2ΔC‐I167M mutants remained below 1 mM (see Figure 3C). [ATP]–response analysis further revealed distinct differences in ATP sensitivity among wild‐type and mutant K_ATP channels (Figure 3D,E). Wild‐type channels showed high ATP sensitivity, with a K_1/2 of ~0.02 mM (n = 3). The Kir6.2ΔC‐C166M mutation produced a rightward shift in the ATP response curve, yielding a K_1/2 of 0.36 mM. By comparison, the R50G* displayed a substantially larger shift, with a K_1/2 of 2.10 mM, extending ATP sensitivity into the low‐millimolar range. Based on these properties, the R50G* mutant was selected for subsequent patch‐cramming experiments to estimate [ATP]_cyt.

3.4. Patch Cramming‐Based Estimation of [ATP]_cyt in Cells

A typical procedure for the patch‐cramming method is illustrated in Figure 4A. A cell‐attached patch is first formed on an HEK293 cell that expresses R50G*. In the cell‐attached mode, K_ATP channel activity was very low, probably due to the high [ATP]_cyt, which keeps the channels mostly closed. When an inside‐out patch was formed (Figure 4B), a rapid increase in channel activity occurred, as the [ATP] in the perfusion solution was reduced (~10 μM). Increasing the [ATP] in the perfusion solution to 1 and 3 mM reduced the channel activity in a concentration‐dependent manner.

FIGURE 4.

Patch‐cramming of an inside‐out patch into a HEK293 (test) cell. (A) A general scheme showing how patch‐cramming is performed. An inside‐out patch from a HEK293 cell containing R50G* K_ATP channels can be crammed into a test cell. (B) An example of this method is shown, in which the pipette is crammed into a nearby HEK293 cell. Patch cramming led to rapid inhibition of R50G* activity. Reversing cramming (thus re‐exposing the inside‐out patch to the bath solution) recovered channel activity.

An inside‐out patch obtained from a HEK293 cell expressing R50G* can be crammed into a different cell type (test cell) to estimate [ATP]_cyt in that cell. Cramming the patch into an intact HEK293 cell further decreased channel activity, indicating that [ATP]_cyt in the HEK293 cell was above 3 mM. From the relative [ATP]‐R50G* activity curve, one can estimate the [ATP]_cyt of the cell using the R50G* activity recorded in that cell. The normalized R50G* activity (${\mathrm{R}50\mathrm{G}}_{\mathrm{NL}}^{*}$) in crammed HEK293 cells was 0.15 ± 0.06 (n = 4), corresponding to a mean [ATP]_cyt of ~4.6 mM in HEK293 cells that were used in these experiments.

3.5. Quantitative Predictions and Experimental Determination of [ATP]_cyt During Hypoxia

A schematic overview of hypoxia‐induced inhibition of K_2P3/9 channels, elevation of [Ca²⁺]_cyt, and subsequent excitation of CB cells is shown in Figure 5A. Acute hypoxia (1% O₂) reduced K_2P3/9 activity by ~65%, consistent with previous reports [36]. Based on the ATP sensitivity of K_2P3/9 (Figure 1D), this degree of inhibition would require a reduction in [ATP]_cyt from ~5 to ~2 mM. According to the [ATP]–response relationship of the R50G* sensor, such a decrease in ATP is predicted to cause a greater than three‐fold increase in R50G* activity (Figure 5B). Similarly, anoxia (~0% O₂) reduced K_2P3/9 activity by ~80%, which would correspond to a decrease in [ATP]_cyt from ~5 to ~1 mM and is predicted to produce a greater than fivefold increase in R50G* activity (Figure 5B).

FIGURE 5.

Does hypoxia decrease K_2P3/9 by lowering [ATP]_cyt? (A) A schematic overview of hypoxia‐induced inhibition of K_2P3/9 channels, elevation of [Ca²⁺]_cyt, and release of transmitter in CB cells. (B) The graph shows the relative activity of R50G* at different ATP concentrations (1, 2, 5, and 10 mM). Based on this graph, the relative R50G* activity is expected to increase 3.5‐fold (0.13–0.46) if [ATP] decreases from 5 mM to 2 mM. R50G* activity would increase ~6 fold (0.13–0.78) if [ATP] decreases to 1 mM. (C) Effect of acute hypoxia (1% O₂) on R50G* activity. An inside‐out patch containing R50G* is crammed into a CB cell perfused with normoxic solution (21% O₂). In this cell, R50G* decreased markedly following cramming. The crammed cell is then perfused with the hypoxic solution (1% O₂) for ~15 s, followed by a return to normoxic solution. The inside‐out patch is then quickly removed from the cell, exposing it to the perfusion solution containing 1 mM ATP. (D) After the cramming of a CB cell and closure of R50G*, application of 1 μM FCCP activated R50G* after a delay. (E) After the CB cell was crammed and R50G* was closed, the CB cell was exposed to 1% O₂ for ~15 s, and then 1 μM FCCP was added under hypoxia. FCCP, but not hypoxia, caused reversible activation of R50G*.

To test these quantitative predictions, an inside‐out patch from a HEK293 cell expressing R50G* was crammed into an isolated CB cell under normoxic conditions (21% O₂). Before cramming, R50G* activity was moderately high in the inside‐out configuration with 1 mM ATP in the bath solution (Figure 5C). Upon cramming, channel activity decreased quickly, indicating that the patch was exposed to [ATP]_cyt greater than 1 mM. Under normoxic conditions, the ${\mathrm{R}50\mathrm{G}}_{\mathrm{NL}}^{*}$ activity in crammed CB cells was 0.12 ± 0.03 (n = 4), corresponding to a [ATP]_cyt of ~4.9 mM based on the calibration shown in Figure 5B. With the patch remaining in the crammed configuration, cells were exposed to hypoxia (1% O₂) for ~15 s. During hypoxia, R50G* activity did not increase. The ${\mathrm{R}50\mathrm{G}}_{\mathrm{NL}}^{*}$ activity was 0.12 ± 0.03 in normoxia and 0.13 ± 0.03 during hypoxia (n = 4, p > 0.05; Figure 5C).

To be certain that the crammed CB cells are functionally viable, and R50G* in the patch pipette in the crammed cells is capable of detecting the decrease in [ATP]_cyt, we tested the effect of carbonyl cyanide p‐(trifluoromethoxy) phenylhydrazone (FCCP, 1 μM), a mitochondrial uncoupler well known to rapidly reduce [ATP]_cyt in living cells [37, 38]. After the patch containing R50G* was crammed into a CB cell and the R50G* activity decreased, FCCP was applied to the CB cell. After a delay of 10–15 s, FCCP started to activate R50G* in the crammed patch, most likely due to the lowering of the [ATP]_cyt caused by mitochondrial uncoupling (Figure 5D). In these experiments, R50G* openings caused by FCCP in crammed CB cells showed single channel conductance (~70‐pS) and long‐lasting opening bursts similar to that observed before cramming. These results indicate that R50G* can respond to the FCCP‐induced reduction in [ATP]_cyt in crammed CB cells and that crammed CB cells have functional mitochondria. The relative activity of R50G* in the presence of FCCP was 0.22 ± 0.05 (n = 4), corresponding to a mean [ATP]_cyt of ~3.7 mM, estimated from the ${\mathrm{R}50\mathrm{G}}_{\mathrm{NL}}^{*}$‐[ATP] relationship. Thus, under the conditions of our experiment, FCCP reduced the [ATP]_cyt by ~24%. In a separate experiment, crammed CB cells were exposed to 1% O₂ for ~15 s, and then FCCP applied under hypoxic conditions. In two successful trials, acute hypoxia (1% O₂) did not cause activation of R50G* during the ~15 s period, but subsequent addition of 1 μM FCCP activated R50G* after a short delay (Figure 5E). Washout of FCCP in the continuous state of hypoxia led to closure of R50G*, suggesting that [ATP]_cyt has recovered. Together, these results suggest that crammed CB cells are viable and able to respond to inhibition of mitochondrial oxidative phosphorylation. These results further show that acute hypoxia (1% O₂) does not decrease [ATP]_cyt for the duration of the experiment. The question of whether a longer duration and more severe form of hypoxia than those used above decrease the [ATP]_cyt would be an interesting future topic.

4. Discussion

The present study addresses a long‐standing question in carotid body physiology: how does acute hypoxia inhibit K_2P3/9 (TASK1/3) channels in Type I (glomus) cells to initiate the chemosensory reflex and augment ventilation? Multiple intracellular signals, including CO, NOX4, ROS, NADH, AMP‐activated protein kinase, and H₂S, have been proposed to modulate K_2P3/9 activity, either directly or indirectly. Despite extensive investigation [8, 14, 39, 40], however, the identity of the proximal cytosolic signal linking hypoxia to channel inhibition has remained unclear. In this study, we focused on cytosolic ATP as a candidate signal, given that both hypoxia and ATP are known to modulate K_2P3/9 activity.

4.1. Re‐Evaluating the Hypoxia‐[ATP]‐K_2P3/9 Pathway

In this study, we have investigated the hypoxia‐[ATP]‐K_2P3/9 signaling pathway in more detail. Studies have shown that hypoxia inhibits K_2P3/9 and the resulting cell depolarization elicits a rise in [Ca²⁺] by opening Ca²⁺ channels, as also shown in Figure 2B [26, 41, 42]. These effects are very rapid, with maximal effects occurring within 5 s of hypoxic exposure of CB cells.

K_2P3/9 activity has been reported to be modulated by [ATP]_cyt [16]. K_2P3/9 activity decreased by 90% when the [ATP] in the solution perfusing the inside‐out patches was removed, and recovered as ATP was added to the bath solution, with a K_1/2 of 2.3 mM [16]. The kinetic and pharmacological properties of the channels that we refer to as K_2P3/9 have been investigated extensively and such studies have clearly demonstrated that K⁺ channels recorded in cell‐attached and inside‐out patches (with ATP in the bath solution) are the same channels belonging to the K_2P3/9 subfamily [K_2P3 (16‐pS), K_2P3/9 (32‐pS), and K_2P9 (32‐pS)]. In inside‐out patches, opening of a 70‐pS ATP‐sensitive K⁺ channel is occasionally observed (in the absence of ATP), and adding ATP quickly closes them as expected [43]. Therefore, available evidence clearly shows that added ATP only activated K_2P3/9 channels in inside‐out patches of CB cells.

In our own CB cell preparations, we found that ATP modulates K_2P3/9 with a K_1/2 of ~2.5 mM. Therefore, the hypoxia‐[ATP]‐K_2P3/9 signaling pathway seems quite reasonable as a mechanism underlying the inhibition of K_2P3/9 by hypoxia. In support of this mechanism, an earlier study found that hypoxia increased free [Mg²⁺], as measured using Mag‐Indo‐1 fluorescence. The rise in free [Mg²⁺] was attributed to a decrease in [ATP] in the cell [11]. It seems reasonable to assume that a reduction in [ATP] will have an opposite effect on [Mg²⁺], as Mg²⁺ is released to the cytoplasm from Mg²⁺ bound to ATP when [ATP] is reduced, for example, by hypoxia. However, it is possible that hypoxia augmented Mag‐Indo‐1 fluorescence without altering [ATP]. Since hypoxia elevates [Ca²⁺]_cyt and Mag‐Indo‐1 is highly sensitive to Ca²⁺, part of the signal may be attributable to the increase in [Ca²⁺]_cyt. Therefore, a different approach to measuring [ATP] is essential to confirm that hypoxia indeed reduces [ATP]. Consequently, we developed a K_ATP‐based bioassay to estimate the [ATP]_cyt in CB cells.

4.2. A Bioassay to Measure [ATP]_cyt

Cytosolic levels of cAMP and cGMP have been measured using an excised, inside‐out membrane patch containing cyclic nucleotide–gated ion channels as the biosensor [44]. The atrial G protein‐gated K⁺ (K_ACh) channel has also been used for studying [GTP] in living cells [45]. In these studies, a patch of membrane (an inside‐out patch) containing a single type of ion channel whose activity is sensitive to a cytosolic molecule is crammed into a cell of interest. Thus, the ion channel serves as a biosensor to estimate the level of the intracellular molecule in question. This technique is also useful for measuring real‐time changes in intracellular messenger levels in response to a stimulus with minimal invasion of cell function [46]. In our study, the K_ATP channel was used as a biosensor to detect the cytosolic [ATP].

Wild‐type K_ATP channels are usually highly sensitive to ATP, with a K1/2 in the micromolar range [47]. To assess [ATP] in living cells whose cytosol contains millimolar ATP, it was necessary to reduce the ATP sensitivity of the K_ATP channel > 50‐fold. Fortunately, earlier studies have identified specific mutations in the Kir6 portion of the K_ATP channel (composed of Kir6/SUR) that have reduced ATP sensitivity [34]. Based on this information, we constructed several mutant K_ATP channels, one of which (R50G*) showed ~100‐fold reduction of ATP sensitivity compared to the wild‐type K_ATP channel. [ATP]‐${\mathrm{R}50\mathrm{G}}_{\mathrm{NL}}^{*}$ relationship obtained from inside‐out patches from HEK293 cells showed a K_1/2 of ~2.0 mM for ATP. An inside‐out patch from a HEK293 cell expressing R50G* can then be crammed into an isolated cell of interest, exposing the inside‐out patch to the cytoplasm of that cell. From the ${\mathrm{R}50\mathrm{G}}_{\mathrm{NL}}^{*}$ activity in the crammed patch, [ATP] can then be interpolated from the [ATP]‐${\mathrm{R}50\mathrm{G}}_{\mathrm{NL}}^{*}$ relationship. By this method, the mean [ATP] in control HEK293 cells was estimated to be 4.6 mM. Using the same process, the mean [ATP]_cyt in CB cells was estimated to be 4.9 mM.

Successful cramming of an inside‐out patch containing R50G* channels into the CB cell was critical for our experiment. This procedure required practice for piercing the pipette tip into the cell through the plasma membrane (cramming). A quick manual movement of the patch pipette into the cell was critical for exposing the inside‐out patch to the cytosol. Sometimes the CB cell moved away as the pipette was forced into it. This was probably due to weak adhesion of the CB cell to the glass coverslip, which was perfused with solution at 37°C. In other instances, if the patch pipette moved below a certain speed, the pipette tip would form a dimple and fail to pass through the membrane. In our earlier study, an atrial inside‐out patch containing G‐protein‐gated K⁺ (K_ACh) channels was inactive in GTP‐free perfusion solution [45]. Upon cramming into another atrial cell, K_ACh channels became fully active without issues such as cell breakage or leakage, indicating the success of our cramming method. Most likely, the glass pipette wall of the inside‐out patch pipette and the target cell membrane form a tight seal after cramming, keeping the cell intact and non‐leaky.

The bioassay was considered successful when R50G* activity decreased rapidly after cramming and recovered fully after the pipette was withdrawn from the cell to the bath solution. This suggests that the integrity of the inside‐out patch membrane was not affected by either the cell plasma membrane or an intracellular organelle of the CB cell. Also, the channels that opened occasionally during the crammed period showed clearly identifiable single‐channel activity. The viability of the crammed CB cells was assessed using FCCP, a mitochondrial uncoupler known to rapidly reduce [ATP]_cyt [37, 38]. FCCP activated R50G* after a short delay (10–15 s), suggesting that the [ATP]_cyt remains relatively normal for a short time period before [ATP]_cyt declines to ~3.7 mM from ~5 mM. As FCCP can activate R50G* in crammed CB cells, it is safe to say that the crammed cells are viable and have a functional mitochondrial response to the mitochondrial uncoupler. The expected single‐channel conductance (~70 pS) of R50G* in crammed cells also indicates that the cells are able to maintain the resting membrane potential, which requires normal function of plasma membrane pumps and ion channels. These observations support the use of the bioassay to assess [ATP]_cyt not only in CB cells but also in other isolated cell types, and to record real‐time changes in [ATP]_cyt during short periods of external stimuli (such as acute hypoxia studied here).

4.3. Inhibition of K_2P3/9 by Acute Hypoxia Does Not Involve a Reduction of [ATP]_cyt

Our finding that R50G* activity in crammed cells is not affected by acute hypoxia (1% O₂) suggests that hypoxia that inhibits K_2P3/9 by 65% does not affect [ATP]_cyt. This is contrary to what we predicted based on an earlier report that hypoxia increased free [Mg²⁺], which may be associated with decreased [ATP] [16]. We believe that our K_ATP‐based bioassay provides a better assessment of changes in [ATP]_cyt because it does not rely on Mg²⁺‐ATP interactions in the cell. Our finding supports the recent study by Torres‐Lopez et al. [9], which found that hypoxia augments the secretory activity of CB cells without affecting intracellular [ATP/ADP]. Torres‐Lopez et al. used a genetically encoded fluorescent protein (PercevalHR) to monitor the ATP/ADP ratio in the CB cell. In this method, PercevalHR is expressed in CB cells using the adeno‐associated viral vector (AAV9). For infection, CB slices were incubated with viral particles for 48 h before the experiment. In our experiments, freshly isolated CB cells were used in the cramming experiments. Thus, one advantage of our cramming method is that the CB cells used in the experiments are freshly isolated, without the need to express a foreign gene that requires 48 h of incubation beforehand. We found that the sensitivity of K_2P3/9 in CB cells to hypoxia after 48 h in culture decreased by ~50% compared to that in freshly isolated cells (unpublished, Kim lab), making a strong case for using freshly isolated cells for hypoxia experiments. Another advantage of our method is that we can assess changes in [ATP]cyt, not the ATP/ADP ratio, although they are likely to be closely associated.

In the study by Torres‐Lopez et al. [9], the perfusion solution was bubbled with a gas mixture of 5% CO₂ and 95% N₂, producing an O₂ tension of 10–15 mmHg in the recording chamber. In our studies, hypoxia (1% O₂) was generated by bubbling the solution with a gas mixture of 5% CO₂/1% O₂/94% N₂, yielding an O₂ tension of ~12 mmHg measured with a commercial O₂ meter. These values are difficult to compare because of differences in experimental setup and the specific instruments that measure pO₂. Nevertheless, the level of hypoxia used by Torres‐Lopez et al. [9] seems comparable to that used in our experiments. Therefore, it is highly likely that the level of hypoxia used by Torres‐Lopez et al. [9] also strongly inhibited K_2P3/9 without affecting the ATP/ADP ratio in their system. The findings from both studies showing similar outcomes that acute hypoxia excites CB cells without affecting [ATP]_cyt or cellular ATP/ADP metabolism are encouraging. These findings help focus our efforts on other potential mechanisms by which hypoxia inhibits K_2P3/9 to initiate the chemoreceptor reflex.

4.4. What Hypoxia Signals Inhibit K_2P3/9 in CB Cells?

In inside‐out patches, we found that hypoxia does not affect K_2P3/9, indicating that K_2P3/9 and its associated molecules are not sensitive to hypoxia. As inhibitors of mitochondrial oxidative phosphorylation reduce K_2P3/9 activity, signals generated by mitochondria within the cell cytoplasm are likely to cause K_2P3/9 inhibition. Many hypotheses regarding signals involved in the hypoxia‐sensing mechanism, such as H₂S and AMPK, can be ruled out based on studies using genetically modified mice [14, 40, 48]. Gene deletion studies have shown that HIF2α is crucial for hypoxia sensing in CB cells, involving the generation of ROS and NADH, associated with HIF2α‐dependent expression of atypical mitochondrial electron transport chain subunits [6]. Transcriptome analysis has also shown high expression of HIF2α in O₂‐sensing CB cells [7], suggesting that ROS/NADH could serve a role in hypoxia sensing, as reported recently by Torres‐Lopez et al. [9].

The specific signals mediating the hypoxic excitation of CB cells may depend on the severity of hypoxia. For example, it seems plausible that inhibition of TASK by 1% O₂ and 0% O₂ involves distinct signaling pathways. Signals such as ROS‐NADH redox state may regulate TASK activity at mild to moderate levels of hypoxia, whereas additional mechanisms, such as a reduction in [ATP]_cyt, could further reduce TASK activity in response to severe hypoxia or anoxia. It could also be that hypoxia uses multiple cellular mechanisms to inhibit K_2P3/9, such that the contribution of each mechanism to the overall effect on the CB cell is small, but together, they become significant. Which signals are involved, and how much each contributes to the inhibition of K_2P3/9 in response to hypoxia, are key questions to be addressed in the future.

5. Conclusion

Using a K_ATP‐based bioassay in combination with patch‐cramming, we assessed [ATP]_cyt in CB cells. Acute hypoxia, which robustly inhibited K_2P3/9 channels, did not reduce [ATP]_cyt. Our findings demonstrate that cytosolic ATP depletion is unlikely to serve as the initiating signal for acute hypoxia‐induced excitation of CB cells.

Author Contributions

Donghee Kim: conceptualization, methodology, investigation, formal analysis, data curation, visualization, writing – original draft, funding acquisition, writing – review and editing. Dawon Kang: investigation, methodology, visualization, writing – review and editing, funding acquisition. Jiaju Wang: methodology, visualization, data curation, formal analysis, writing – review and editing.

Funding

This work was supported by National Research Foundation of Korea (RS‐2023‐00219399, RS‐2024‐00454609) and Rosalind Franklin University of Medicine and Science.

Disclosure

Dr. Jiaju Wang completed a part of this work while a PhD student at Rosalind Franklin University. Dr. Wang is currently an employee of Takeda Pharmaceuticals USA, and Takeda is not affiliated with or responsible for any part of this work.

Consent

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

All data supporting the conclusions of this study are included in the article. Additional data are available from the corresponding author upon reasonable request.