Research ArticlePhysiology

The LKB1–AMPK-α1 signaling pathway triggers hypoxic pulmonary vasoconstriction downstream of mitochondria

See allHide authors and affiliations

Science Signaling  02 Oct 2018:
Vol. 11, Issue 550, eaau0296
DOI: 10.1126/scisignal.aau0296

Hypoxia amps AMPK

Most arteries in the body vasodilate in response to hypoxia. In contrast, pulmonary arteries constrict in response to hypoxia to ensure that blood is diverted to oxygen-rich areas, a process called hypoxic pulmonary vasoconstriction. When this process is prolonged, it can lead to pulmonary hypertension and, ultimately, heart failure. Analysis of genetically manipulated mice by Moral-Sanz et al. showed that hypoxic pulmonary vasoconstriction under mild hypoxia required the α1 subunit of AMPK, which was activated downstream of LKB1 during hypoxia. During severe hypoxia, the α2 subunit of AMPK was also required. AMPK supports hypoxic pulmonary vasoconstriction, in part, by phosphorylating and inhibiting the K+ channel KV1.5 in pulmonary arterial myocytes. By contrast, in most systemic arteries, hypoxia does not inhibit KV, which may aid vasodilation rather than vasoconstriction. These results delineate a key role for LKB1-activated AMPK-α1 and AMPK-α2 in hypoxic pulmonary vasoconstriction.


Hypoxic pulmonary vasoconstriction (HPV), which aids ventilation-perfusion matching in the lungs, is triggered by mechanisms intrinsic to pulmonary arterial smooth muscles. The unique sensitivity of these muscles to hypoxia is conferred by mitochondrial cytochrome c oxidase subunit 4 isoform 2, the inhibition of which has been proposed to trigger HPV through increased generation of mitochondrial reactive oxygen species. Contrary to this model, we have shown that the LKB1–AMPK-α1 signaling pathway is critical to HPV. Spectral Doppler ultrasound revealed that deletion of the AMPK-α1 catalytic subunit blocked HPV in mice during mild (8% O2) and severe (5% O2) hypoxia, whereas AMPK-α2 deletion attenuated HPV only during severe hypoxia. By contrast, neither of these genetic manipulations affected serotonin-induced reductions in pulmonary vascular flow. HPV was also attenuated by reduced expression of LKB1, a kinase that activates AMPK during energy stress, but not after deletion of CaMKK2, a kinase that activates AMPK in response to increases in cytoplasmic Ca2+. Fluorescence imaging of acutely isolated pulmonary arterial myocytes revealed that AMPK-α1 or AMPK-α2 deletion did not affect mitochondrial membrane potential during normoxia or hypoxia. However, deletion of AMPK-α1, but not of AMPK-α2, blocked hypoxia from inhibiting KV1.5, the classical “oxygen-sensing” K+ channel in pulmonary arterial myocytes. We conclude that LKB1–AMPK-α1 signaling pathways downstream of mitochondria are critical for the induction of HPV, in a manner also supported by AMPK-α2 during severe hypoxia.


Hypoxic pulmonary vasoconstriction (HPV) (1) diverts blood flow from oxygen-deprived to oxygen-rich areas of the lung, thus matching perfusion with ventilation (2). However, when triggered by prolonged and widespread hypoxia, such as during ascent to altitude (3) or disease (such as cystic fibrosis), HPV can cause hypoxic pulmonary hypertension (HPH), leading to right heart failure (4).

HPV is initiated by airway hypoxia rather than by vascular hypoxemia (5) through constriction of precapillary resistance arteries by signaling pathways that are intrinsic to their smooth muscle and endothelial cells (68) and are independent of blood-borne mediators or the autonomic nervous system (9, 10). The initiation phase of acute HPV is driven by smooth muscle constriction, which is later augmented by an endothelium-derived vasoconstrictor (6, 11). The threshold for HPV is ≈80 mmHg O2 (12), beyond which the magnitude of HPV increases with the degree of hypoxia. The unique oxygen sensitivity of pulmonary arterial myocytes is conferred by mitochondrial cytochrome c oxidase subunit 4 isoform 2 (COX4I2), the deletion of which blocks HPV (13). HPV has been proposed to be triggered by increases in mitochondrial reactive oxygen species (ROS) upon inhibition of mitochondrial metabolism at the COX-COX4I2 complex (13). In this respect, however, it is evident that COX-COX4I2 deletion will not only compromise mitochondrial ROS production during hypoxia but also raise the threshold at which hypoxia evokes increases in the cytoplasmic AMP:ATP/ADP:ATP ratios (13). This is critical because we previously proposed that HPV is mediated not by ROS but by the adenosine 5′-monophosphate–activated protein kinase (AMPK) (14). Supporting this notion, small-molecule AMPK activators induce pulmonary artery constriction by mechanisms similar to those used by hypoxia (14), and each of these stimuli is mutually exclusive with respect to inhibition of KV1.5 (15), the archetypal oxygen-sensing K+ channel in pulmonary arterial myocytes (8, 16). That said, ROS have been proposed to activate AMPK directly (17, 18), although this view has been controversial (19, 20).

AMPK is a cellular energy sensor that occurs as heterotrimeric complexes comprising one catalytic α subunit (α1/α2) in combination with one β (β1/β2) and one γ (γ1/γ2/γ3) subunit; these subunit isoforms generate up to 12 different combinations, with different sensitivities to metabolic stress and the capacity to regulate different downstream targets (21). In response to metabolic stresses such as hypoxia and mitochondrial poisons, binding of AMP to the γ subunit increases AMPK activity by up to 10-fold through allosteric activation. However, it is the upstream kinase liver kinase B1 (LKB1) that is most critical to AMPK activation under these conditions. LKB1 is thought to be constitutively active (22, 23). Upon binding of AMP and/or adenosine 5′-diphosphate (ADP) to the AMPK-γ subunit, phosphorylation of Thr172 in the α subunit is enhanced, and dephosphorylation of Thr172 is inhibited, yielding up to a 100-fold increase in AMPK activity through a net increase in phosphorylation of Thr172 by LKB1. All three of these effects are blocked by high concentrations of adenosine 5′-triphosphate (ATP). AMPK activation may also be triggered independently of metabolic stresses through increases in intracellular Ca2+ and increases in Thr172 phosphorylation mediated by the calmodulin-dependent kinase, CaMKK2 (21). Once activated, AMPK phosphorylates targets that switch off nonessential ATP-consuming processes and switch on catabolic pathways generating ATP, thereby compensating for deficits in ATP supply caused by, for example, inhibition of mitochondrial oxidative phosphorylation.

We sought to determine whether AMPK signaling pathways within pulmonary arterial myocytes initiate HPV downstream of mitochondria, by studying transgenic mice deficient in LKB1, CaMKK2, or AMPK. Our findings show that activation of the LKB1–AMPK-α1 signaling pathway within pulmonary arterial smooth muscles is critical for the initiation of HPV.


We sought to determine the role of AMPK in the induction of HPV, the initiating phase of which is primarily driven by mechanisms of constriction intrinsic to pulmonary arterial myocytes (6). The genes encoding the AMPK-α1 (Prkaa1) or AMPK-α2 (Prkaa2) catalytic subunits were conditionally deleted (fig. S1) by crossing floxed mice with those expressing Cre recombinase from the transgelin promoter. AMPK-α1 and AMPK-α2 knockout mice did not exhibit overt phenotypes, although mild left, but not right, ventricular myopathy was evident after AMPK-α2 deletion alone (table S1). This finding is consistent with previous reports of a dominant role for AMPK-α2 in the regulation of cardiomyocyte metabolism and growth (24, 25).

However, AMPK-α1 or AMPK-α2 deletion attenuated HPV, which was monitored using spectral Doppler ultrasound analysis of systolic peak velocity and velocity time integral (VTI) within the main pulmonary artery. This method provides an index of dynamic changes in pulmonary vascular flow (26). Only AMPK-α1 deletion markedly attenuated HPV in mice exposed to both mild (8% O2; fig. S2, A and B) and severe hypoxia (5% O2; Fig. 1, A to D; fig. S2, A and B; and fig. S3, A and B). Moreover, the magnitude of HPV increased with the severity of hypoxia in floxed, control mice but not in AMPK-α1 knockouts (Fig. 1A and fig. S2, A and B). By contrast, AMPK-α2 deletion attenuated HPV evoked by severe (5% O2; Fig. 1, A to D) but not mild (8% O2) hypoxia (fig. S2, A and B, and fig. S3, A and B). Furthermore, the degree to which HPV was attenuated during severe hypoxia appeared to be slightly more marked for AMPK-α1 than for AMPK-α2 knockouts, although this difference was not significant (Fig. 1D; fig. S2, A and B; and fig. S3, A and B). Collectively, these data suggest that AMPK-α1 may be critical to the induction of HPV, whereas AMPK-α2 supports HPV in some other way during severe hypoxia. In contrast, neither AMPK-α1 nor AMPK-α2 deletion affected increases in pulmonary vascular reactivity induced by intravenous injection of serotonin (Fig. 1, E to H, and fig. S3, C and D), a pulmonary-selective vasoconstrictor (27) that promotes pulmonary hypertension by different mechanisms to hypoxia (28). Moreover, AMPK-α1 and AMPK-α2 deletion inhibited pulmonary vascular responses during hypoxia but did not affect either resting systemic arterial blood pressures or hypoxia-evoked decreases in systemic arterial blood pressures (29). Thus, deletion of AMPK-α1 and AMPK-α2 subunits within arterial myocytes selectively inhibits HPV. This view is supported by the findings of others, which have established that neither AMPK-α1 nor AMPK-α2 deletion affects normoxic or ischemic contracture of cardiac muscles (24, 30) and that there is substantial redundancy of function with respect to the regulation by AMPK-α1 and AMPK-α2 of cardiac metabolism and growth (24, 25, 31).

Fig. 1 The LKB1-AMPK but not the CaMKK2-AMPK signaling pathway is required for HPV.

(A to C) Top panels show spectral Doppler analyses of peak velocity versus time for the systolic waveform within the main pulmonary artery of (A) AMPK-α1 and AMPK-α2 floxed controls (AMPK-α1/α2 FLX), (B) AMPK-α1 knockout, and (C) AMPK-α2 knockout mice during normoxia, hypoxia (12, 8, or 5% O2), and recovery. Bottom panels show example records of peak velocity during normoxia, hypoxia, and recovery. (D) Bar chart shows the means ± SEM for the maximum change in peak velocity observed during 5% O2 for AMPK-α1/α2 FLX, C57Bl6, CaMKK2 KO, hypomorphic LKB1 floxed (LKB1 KD), AMPK-α1 KO, and AMPK-α2 KO mice; n = 3 to 7 mice per group, **P < 0.01, ***P < 0.001. (E to G) Top panels show spectral Doppler analyses of peak velocity versus time for the systolic waveform within the main pulmonary artery of (E) AMPK-α1/α2 FLX controls, (F) AMPK-α1 knockout, and (G) AMPK-α2 knockout mice during intravenous injection of 5-HT (serotonin). (H) Bar chart shows the means ± SEM for the maximum change in peak velocity observed for AMPK-α1/α2 FLX, AMPK-α1 KO, and AMPK-α2 KO mice; n = 3 to 5 mice per group.

We could not generate viable homozygous deletion of the gene encoding LKB1 (Stk11) using the trangelin-Cre promoter but did assess HPV in a strain of mice bearing homozygous floxed alleles of Stk11, which globally exhibit hypomorphic (≈90% reduced) expression of LKB1 (32). Spectral Doppler analysis of pulmonary arterial systolic waveforms revealed that HPV was markedly attenuated during exposure of these mice to severe hypoxia (5% O2; Fig. 1D; fig. S3, A and B; and fig. S4A). By contrast, global CaMKK2 deletion did not affect hypoxia-evoked reductions in peak velocity (Fig. 1D; fig. S3, A and B; and fig. S4B). Therefore, HPV is mediated by LKB1-AMPK signaling independently of the Ca2+/CaMKK2 pathway. Supporting this view, a previous study using the hypomorphic LKB1 model used here has shown that LKB1 deficiency compromises cardiac metabolism but not cardiac contractility, ischemic contracture, or recovery from ischemia (33).

In isolated pulmonary arterial myocytes, mitochondrial membrane potentials were not affected after deletion of either AMPK-α1 or AMPK-α2 (Fig. 2, A and B). These gene deletions also did not affect mitochondrial membrane hyperpolarization evoked by hypoxia or oligomycin, which was of similar magnitude to that observed by others (Fig. 2, C to F) (13). Consequently, hypoxia will increase AMP/ATP ratios in pulmonary arterial myocytes (14) that will in turn lead to rapid AMPK activation (1 to 2 min) (34). Consistent with this scenario and a requirement for AMPK, AMPK-α1 deletion in acutely isolated pulmonary arterial myocytes attenuated KV1.5 current inhibition in response to severe hypoxia [≈6% PO2 (partial pressure of oxygen)] (Fig. 3, A to G, and fig. S5, A to C), oligomycin, and the AMPK-α1–selective activator C13 (Fig. 3G) (35). These effects were not due to baseline changes in KV current amplitudes or kinetics (table S2). AMPK-α2 deletion did not affect KV1.5 current inhibition during hypoxia (Fig. 3, C, F, and G), supporting the view that AMPK-α1 and not AMPK-α2 drives HPV. The presence of KV1.5 currents was confirmed using the selective KV1.5 channel blocker diphenyl phosphine oxide-1 (DPO-1) (Fig. 3G and fig. S6, A to F). Therefore, with respect to KV1.5 current regulation, there is no redundancy in the system. Support for this view is provided by our findings showing that although AMPK-α1 deletion blocked KV1.5 inhibition during hypoxia, AMPK-α2 deletion did not affect the time course of KV1.5 inhibition (fig. S7, A to D), the onset of which was evident at 1 min and reached a maximum between 2 and 5 min of hypoxia. However, the true time course for KV1.5 inhibition during hypoxia will likely be faster because the time for complete superfusate exchange and equilibration within the experimental chamber was ≈3 to 5 min (fig. S7, A to D). Thus, during hypoxia, AMPK-α1 but not AMPK-α2 mediates reductions in KV1.5 channel activity in pulmonary arterial myocytes during hypoxia, downstream of mitochondrial inhibition.

Fig. 2 Deletion of AMPK-α1 or AMPK-α2 does not affect mitochondrial membrane potential.

(A) Images show a bright-field image of a pulmonary arterial myocyte, deconvolved three-dimensional reconstructions of a Z-stack of images of MitoTracker fluorescence, tetramethylrhodamine ethyl ester (TMRE) fluorescence, and a merged image from AMPK-α1/α2 FLX, AMPK-α1, and AMPK-α2 knockout mice. (B) Bar chart shows the means ± SEM for TMRE fluorescence (n = 23 to 34 cells from 4 mice per group). (C to E) Records of rhodamine 123 (Rhod123) fluorescence ratio (F/F0) plotted against time, recorded from pulmonary arterial myocytes of (C) AMPK-α1/α2 FLX, (D) AMPK-α1, and (E) AMPK-α2 knockout mice during normoxia and hypoxia. The nonquenching mode was confirmed by adding the mitochondrial uncoupler carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP; 10 μM). (F) Bar charts show the means ± SEM for the maximum change in Rhod123 fluorescence during application of hypoxia (~6% O2) and oligomycin (3 μM). n = 3 to 6 cells from 3 mice per group. AU, arbitrary units.

Fig. 3 Deletion of AMPK-α1 but not AMPK-α2 prevents the inhibition of KV1.5 currents by hypoxia and mitochondrial poisons.

(A to C) Top panels show the voltage steps. Bottom panels show raw records of KV currents activated at each voltage step during normoxia and hypoxia. (D to F) Plots (means ± SEM) for the current-voltage (I/Imax) relationship of steady-state KV currents recorded (average from the last 20 ms of voltage step) during normoxia and hypoxia (~6% O2) in acutely isolated pulmonary arterial myocytes from (D) AMPK-α1/-α2 floxed (control), (E) AMPK-α1 smooth muscle knockout (AMPK-α1 KO), or (F) AMPK-α2 smooth muscle knockout (AMPK-α2 KO) mice. (G) Bar charts show the means ± SEM for the peak inhibition in KV current at 0 mV (I/Imax) for pulmonary arterial myocytes from each genotype during hypoxia and after extracellular application of the mitochondrial poison oligomycin (3 μM), the AMPK-α1–selective agonist C13 (30 μM), and the selective KV1.5 blocker DPO-1 (1 μM); n = 4 to 9 cells from 3 mice per group; *P < 0.05, **P < 0.01.

In contrast to the apparent selectivity of KV1.5 regulation by endogenous AMPK-α1 in mouse pulmonary arterial myocytes, we have previously shown that exogenous application of recombinant heterotrimers comprising either AMPK-α1 or AMPK-α2 can phosphorylate and regulate KV1.5 in rat pulmonary arterial myocytes and human KV1.5 expressed in human embryonic kidney (HEK) 293 cells (15). We have now confirmed sequences surrounding Ser559 and Ser592 of the human KV1.5 alpha subunit (Fig. 4A) as reasonable fits to the AMPK recognition motif using SCANSITE4, as well as an algorithm available on GitHub ( (36, 37). These motifs lack the expected hydrophobic residue at position P-5, although Ser592 does have a large uncharged side chain (Asn) at this position. To identify the AMPK recognition sites that were phosphorylated, we treated the α subunit of the channel from transfected HEK293 cells with protein phosphatase to remove endogenous phosphate groups before incubating immunoprecipitates with bacterially expressed AMPK (α2β2γ1 complex) (38). The stoichiometry of phosphorylation using [γ-32P]ATP was estimated to be 3 mol/mol for wild-type (WT) KV1.5, indicative of multiple sites of phosphorylation (Fig. 4, B and C). Moreover, the stoichiometry of phosphorylation was reduced on mutating Ser559 (2.1 mol/mol) or Ser592 (1.2 mol/mol) to alanine, whereas mutation of both sites had a similar effect to mutating Ser592 alone, which was also reflected in the average fold change in phosphorylation relative to wild type with S559A, S592A, and S559A/S592A (Fig. 4, A to D). Thus, AMPK directly phosphorylates the α subunit of KV1.5 at Ser592 and, to a lesser extent, at Ser559. To assess the role of these sites in KV1.5 regulation by AMPK, we examined the effect of the AMPK-α1–selective agonist C13 on HEK293 cells transiently expressing WT KV1.5 α subunit and KV1.5 α subunits incorporating these single and double mutations. C13 inhibited steady-state currents carried by WT human KV1.5 α subunits (Fig. 4, E and F). This inhibition was virtually abolished by the S559A/S592A double mutation (Fig. 4, E to G), by the S559A mutation (Fig. 4G), and, to a lesser extent, by the S592A single mutation (Fig. 4G). In short, Ser559 phosphorylation confers similar KV1.5 current inhibition as the S559A/S592A double mutation, whereas S592A is less effective despite the greater effect of the S592A mutation on channel phosphorylation than the S559A mutation.

Fig. 4 AMPK inhibits recombinant KV1.5 by direct phosphorylation of two AMPK recognition motifs incorporating Ser559 and Ser592.

(A) Alignment of recognition motif for AMPK with sequences of KV1.5. Phosphorylation sites are marked in pink; β, basic residue; ϕ, hydrophobic residue. Schematic indicates position of residues within the KV1.5 α subunit. Coomassie blue stain (B) and autoradiogram (C) show the effect of S559A (n = 3 independent experiments), S592A (n = 3 independent experiments), and S559A/S592A (n = 2 independent experiments) mutations on KV1.5 phosphorylation in cell-free assays by AMPK in the presence of AMP. (D) Fold change in phosphorylation (mutant/wild type) of WT KV1.5 and KV1.5 mutants incorporating S559A, S592A, and S559A/S592A. **P < 0.01, ***P < 0.001. (E) Example records show the effects of C13 on whole-cell K+ currents measured by voltage ramp and voltage step protocols, during extracellular application of 30 μM C13 in HEK293 cells expressing WT KV1.5 and the KV1.5 S559A/S592A mutant. (F) Time course for the reductions in whole-cell K+ currents carried by WT and S559A/S592A mutant KV1.5 channels during extracellular application of 30 μM C13. (G) Bar chart shows the means ± SEM for reductions in steady-state K+ currents (average from last 20 ms of voltage step) carried by WT and mutant KV1.5 channels 5 min after extracellular application of 30 μM C13; n = 5 to 7 cells per group; *P < 0.05, **P < 0.01, ***P < 0.001. DMSO, dimethyl sulfoxide.


The present investigation showed that divergent LKB1-AMPK signaling pathways coordinated HPV downstream of mitochondria. HPV was suppressed by deletion of either AMPK-α1 or AMPK-α2, whereas reductions in KV1.5 channel activity in pulmonary arterial myocytes during hypoxia were blocked by deletion of AMPK-α1 alone. Furthermore, only AMPK-α1 deletion in smooth muscles blocked the induction and progressive augmentation of HPV by mild and severe hypoxia. By contrast, AMPK-α2 deletion attenuated HPV only during severe hypoxia, indicative of a role for AMPK-α2–containing heterotrimers in supporting HPV during severe hypoxic stress. Outcomes are therefore consistent with our previous studies using the nonselective AMPK antagonist compound C (39), although any action of this agent should be considered with caution because compound C inhibits at least 10 other kinases more potently than AMPK (40). Nevertheless, it is notable that in isolated rat pulmonary artery rings, two phases of constriction are evident during hypoxia, both of which are initiated concomitantly and are thus superimposed on each other (12). Phases 1 and 2 of HPV are attenuated in pulmonary arteries preincubated with compound C, although only the attenuation of phase 2 is statistically significant (39). However, it is open to debate whether the phase 1 constriction represents a component of the true physiological response of pulmonary arteries in situ in the lung, given that HPV in the rat lung is abolished by agents that inhibit phase 2 but not phase 1 of HPV in acutely isolated pulmonary arteries (12).

Although the mode of action by which AMPK-α2 complexes support HPV remains unclear, there is no redundancy of function with respect to either the induction by AMPK-α1 of myocyte contraction during hypoxia or the inhibition of KV1.5. This lack of redundancy is surprising because intracellular dialysis of AMPK heterotrimers containing either AMPK-α1 or AMPK-α2 inhibits KV1.5 in rat pulmonary arterial myocytes (15). This suggests that subcellular location rather than intrinsic target specificity must determine outcomes, with AMPK-α1 complexes in some way “tethered” proximal to the KV1.5 α subunit. Unfortunately, AMPK antibodies that are currently available are not sufficiently specific to directly test this idea. Although we cannot be certain that AMPK directly phosphorylates KV1.5 in hypoxic pulmonary arterial myocytes, given that KV1.5 has been reported to be regulated through tyrosine phosphorylation (41, 42), our data do show that AMPK phosphorylates and regulates KV1.5 through serine and not through tyrosine residues. In short, although multiple kinase pathways may affect KV1.5 activities in pulmonary arterial myocytes, the AMPK-dependent pathway is critical to HPV.

The induction of HPV requires LKB1 but is triggered independently of the Ca2+/CaMKK2 pathway, highlighting that AMPK activation occurs in response to energy stress rather than to changes in cytoplasmic Ca2+. This finding suggests that hypoxia-evoked increases in AMP:ATP/ADP:ATP ratios (14, 19, 20) are critical to the process of AMPK activation by hypoxia in pulmonary arterial myocytes. Although a requirement for LKB1 does not rule out a role for mitochondrial ROS in HPV (13, 17, 18), it provides further support for the view that mitochondrial ROS do not activate AMPK directly, although ROS may do so indirectly by inhibiting mitochondrial oxidative phosphorylation (19, 20). In this context, the possibility that ROS-dependent increases in cytoplasmic Ca2+ concentration might contribute to AMPK activation (43) seems more remote, given that CaMKK2 is not required for HPV.

Therefore, it would appear that the induction of HPV is primarily governed by the following factors. First, Cox4I2 is selectively expressed in pulmonary arterial smooth muscles, which, unlike COX4I1, does not confer COX inhibition by ATP (44, 45), such that the rate of oxygen consumption and thus ATP supply through mitochondrial oxidative phosphorylation will not increase during hypoxia as ATP levels fall (4548). Expression of AMPK subunits, particularly that of AMPK-α1, is higher in pulmonary than in systemic arterial smooth muscles (14). Moreover, our data also highlight the potential for the regulation of different targets by divergent AMPK signaling pathways, in a manner that may be facilitated by the selective expression and subcellular targeting of multiple heterotrimeric AMPK subunit combinations.

Thus, our observations implicate AMPK activation as the primary mediator of acute HPH consequent to widespread airway hypoxia, an idea consistent with the effects of small interfering RNA knockdown of AMPK in cultured pulmonary arterial myocytes and pharmacological interventions in vivo (49). Moreover, arterial myocyte proliferation during hypoxia and thus HPH has been suggested to result from the coordinate regulation by AMPK-α1 and AMPK-α2 of autophagy initiation and inhibition of apoptosis, respectively (49). Paradoxically, however, others have proposed that mechanistic target of rapamycin complex 2 (mTORC2)–dependent reductions in total AMPK may facilitate myocyte proliferation and thus hypoxic and idiopathic pulmonary hypertension through mTORC1 activation (50) in a manner that may be exacerbated by endothelial AMPK insufficiency (51). One possible explanation for these contrary findings is that the progression of pulmonary hypertension is governed by temporal fluctuations in AMPK activity and/or expression, a scenario that will require further studies for validation.

The present investigation shows that LKB1-AMPK signaling pathways mediate HPV subsequent to the inhibition of oxidative phosphorylation during hypoxia (13) through the phosphorylation and regulation of targets, such as KV1.5, outside of those canonical pathways by which AMPK coordinates cell-autonomous metabolic homeostasis. Therefore, excess AMPK activity or expression may predispose individuals to increases in pulmonary vascular perfusion pressure at high altitudes and may contribute to the initiation, as well as the progression, of HPH (52).



Experiments were performed in accordance with the regulations of the United Kingdom Animals (Scientific Procedures) Act of 1986. All studies and breeding were approved by the University of Edinburgh and performed under UK Home Office project licenses. Both male and female mice were used, and all were on a C57/Bl6 background. Numbers of mice (≥3 per measure) used are indicated for each experiment.

Gene deletion

Global, dual knockout of the genes encoding AMPK-α1 (Prkaa1) and AMPK-α2 (Prkaa2) is embryonic lethal. We therefore used conditional deletion of the genes for the AMPK-α1 or AMPK-α2 subunits, using mice in which the sequence encoding the catalytic site of either of the α subunits was flanked by loxP sequences (53). To direct AMPK deletion to cells of the smooth muscle lineage, these were crossed with mice in which Cre recombinase was under the control of the transgelin (smooth muscle protein 22α) promoter {the Jackson Laboratory [stock number 017491, Tg(Tagln-cre)1Her/J in C57BL/6:129SJL background]}. These transgelin-Cre mice do not exhibit Cre expression in endothelial cells and therefore provide for selective gene deletion in smooth muscle. It should be noted, however, that transient developmental expression of transgelin has been observed in atrial and ventricular myocytes. Therefore, genomic recombination will also occur in these cells, although they do not express transgelin in the adult (54). However, neither atrial nor ventricular myocytes contribute directly or indirectly to acute HPV. Therefore, although imperfect, the use of transgelin-Cre mice provided us with the necessary level of specificity to determine the role of AMPK in response to hypoxia of pulmonary arterial myocytes.

We detected the presence of WT or floxed alleles and Cre recombinase by polymerase chain reaction (PCR). Ten-microliter samples were run on 2% agarose gels with 12.5 μl of GoTag Green Master Mix M712B (Promega) in tris-borate EDTA buffer against a 100–base pair (bp) DNA ladder (GeneRuler, Fermentas) using a Model 200/2.0 Power Supply (Bio-Rad). Gels were imaged using a Genius Bio Imaging System and GeneSnap software (Syngene). We used four primers for Cre (transgene, 5′-GCGGTCTGGCAGTAAAAACTATC-3′ and 5′-GTGAAACAGCATTGCTGTCACTT-3′; internal positive control, 5′-CTAGGCCACAGAATTGAAAGATCT-3′ and 5′-GTAGGTGGAATTTCTAGCATCATCC-3′; expected size for WT = 324 bp, expected size for Cre = 100 bp). Two primers were used for each AMPK catalytic subunit: AMPK-α1, 5′-TATTGCTGCCATTAGGCTAC-3′ (forward) and 5′-GACCTGACAGAATAGGATATGCCCAACCTC-3′ (reverse) (WT = 588 bp, floxed = 682 bp); AMPK-α2, 5′-GCTTAGCACGTTACCCTGGATGG-3′ (forward) and 5′-GTTATCAGCCCAACTAATTACAC-3′ (reverse) (WT = 204 bp, floxed = 250 bp).

The presence of a WT or floxed Stk11 allele, which encodes for LKB1, was detected using two primers: 5′-CCAGCCTTCTGACTCTCAGG-3′ (p200) and 5′-GTAGGTATTCCAGGCCGTCA-3′ (p201). Deletion of LKB1 in smooth muscle proved to be embryonic lethal. Therefore, we studied homozygous LKB1 floxed (STK11flx/flx) mice, which are hypomorphic for LKB1 in all cell types and exhibit ≥90% lower LKB1 expression than LKB1+/+ littermates (32).

To genotype CaMKK2 WT and CaMKK2 knockouts, the presence of a WT allele was detected using two primers, 5′-CAGCACTCAGCTCCAATCAA-3′ (KKBeta1) and 5′-GCCACCTATTGCCTTGTTTG-3′ (KKBeta2), to generate an expected band of 347 bp. To detect CaMKKβ knockout, a KK2-1 primer was used with KK2-3, 5′-TAAGCACAAGCACTCATTCC-3′, to generate an expected band of 470 bp. The PCR protocol used for all genotype primers was 92°C for 5 min, 92°C for 45 s, 56°C for 45 s, 72°C for 60 s, and 72°C for 7 min for 35 cycles and then 4°C as the holding temperature.

RNA from control tissues (lung) and arterial smooth muscle was extracted using the High Pure RNA Tissue Kit (Roche) following the manufacturer’s guidelines, and the concentration was determined using the Nanodrop 1000 spectrophotometer (Thermo Fisher Scientific). Thereafter, complementary DNA (cDNA) synthesis was carried out using the Transcriptor High Fidelity cDNA Synthesis Kit (Roche) following the manufacturer’s instructions.

Smooth muscle cell isolation

Adult mice (2 to 6 months of age) were sacrificed by cervical dislocation. The heart and lungs were removed en bloc and placed in physiological salt solution (PSS) of the following composition: 135 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1 mM CaCl2, 10 mM glucose, and 10 mM HEPES (pH 7.4). Single smooth muscle cells were isolated from second- to third-order branches of the main intrapulmonary artery. For cell isolation, endothelium-denuded arteries were transferred into a nominally calcium-free PSS containing papain (1 mg/ml; Sigma-Aldrich), dithio-dl-threitol (0.8 mg/ml; Sigma-Aldrich), and bovine serum albumin (0.7 mg/ml; Sigma-Aldrich) and preincubated at 37°C for 6 min. The tissue was then washed three times in cold PSS without enzymes, and single smooth muscle cells were isolated by gentle trituration with a fire-polished Pasteur pipette. Cells were stored in suspension at 4°C until required.

PCR analysis

Two microliters of cDNA in ribonuclease-free water was made up to 20 μl with FastStart Universal SYBR Green Master (5-carboxy rhodamine-x; 10 μl, Roche), Ultra Pure Water (6.4 μl, Sigma-Aldrich), and forward and reverse primers for AMPK-α1 and AMPK-α2. The sample was then centrifuged and 20 μl was added to a MicroAmp Fast Optical 96-Well Reaction Plate (Greiner Bio-One), the reaction plate was sealed with an optical adhesive cover (Applied Biosystems), and the plate was centrifuged. The reaction was then run on a sequence-detection system (Applied Biosystems) using AmpliTaq Fast DNA Polymerase, with a 2-min initial step at 50°C followed by a 10-min step at 95°C and then 40× 15-s steps at 95°C. This was followed by a dissociation stage with 15 s at 95°C, 20 s at 60°C, and 15 s at 95°C. Negative controls included control cell aspirants for which no reverse transcriptase was added and aspiration of extracellular medium and PCR controls. None of the controls produced any detectable amplicon, ruling out genomic or other contamination.

Ultrasound measurements

Mice were rendered anesthetized using isoflurane, induced at level 5, and set to level 2 for procedures. Isoflurane was mixed with 21% O2 set at 1 liter/min for normoxia and 12, 8, or 5% O2 for the hypoxic challenge (0.05% CO2, balanced with N2). Using the Vevo 770 RMV 707B transducer (VisualSonics Inc.) with a center frequency of 30 MHz, a parasternal long axis (PLAX) view of the heart of anesthetized mice was obtained and then modified to bring the pulmonary artery into view for spectral Doppler wave measurements. Peak velocity (in millimeters per second) and VTI of pulmonary arterial flow were used as a noninvasive approximation of changes in systolic pulmonary vascular resistance and pressure as described by others (26), and this was confirmed here by infusion of a pulmonary-selective vasoconstrictor, serotonin, which was applied intravenously through the tail vein (27). Measurements of spectral Doppler velocities were taken under normoxia (21% O2; 0.05% CO2 balanced with N2) and during an acute hypoxic challenge (12, 8, or 5% O2; 0.05% CO2 balanced with N2) of about ~2 to 5 min duration followed by a return to normoxia.

Because of the base-to-apex contraction of the right ventricle (RV), it is difficult to accurately quantify RV function other than by fractional shortening, which was therefore assessed using a two-dimensional echocardiographic measurement (55, 56). An M-mode across the anterior and posterior left ventricle walls was used to determine fractional shortening of the left ventricle from a standard PLAX view. To determine RV fractional shortening, an M-mode was analyzed using ultrasound taken across the RV free wall and the septum from a modified PLAX view (57).

Mitochondrial imaging and mitochondrial membrane potential

Fluorescence microscopy was used together with 100 nM Mitotracker Green (excitation, 490 nm; emission, 516 nm) and 100 nM TMRE (excitation, 549 nm; emission, 575 nm) or rhodamine 123 (10 μg/ml) (excitation, 507 nm; emission, 529 nm); cells were incubated in PSS each at 37°C for 30 min. Whole-cell rhodamine 123 fluorescence was analyzed before (baseline) and under specified conditions for a maximum of 15 min. Nonquenching conditions were confirmed at the end of every experiment by adding the mitochondrial uncoupler FCCP (10 μM). Images of the emitted fluorescence were acquired at 37°C and 0.1 Hz using a Zeiss Axiovert 200 with a Zeiss Plan-Apochromat 63× 1.4 numerical aperture (NA) oil immersion objective for TMRE/Mitotracker Green or with Zeiss Fluar 40× 1.3 NA oil immersion objective for rhodamine 123. The microscope was coupled to a Hamamatsu Orca charge-coupled device camera and Sutter DG5 light source with appropriate filter sets. Analysis was completed using Volocity 5.5.1 image acquisition and analysis software.


K+ currents were recorded by whole-cell patch clamp and assessed using voltage ramps (−60 to +60 mV; HEK293 cells) and single (−60 to +50 mV; HEK293 cells) or multiple (−80 to +0 mV, 10-mV increments; pulmonary arterial myocytes) voltage steps. Recordings were conducted at 37°C. The standard perfusate (pH 7.4) was composed of 135 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1 mM CaCl2, 10 mM glucose, and 10 mM HEPES (pH 7.4). Patch pipettes had resistances of 4 to 6 megohms. Series resistance was monitored and compensated (60 to 80%) after achieving the whole-cell configuration. If a greater than 20% increase occurred during the recording, then the experiment was terminated. The pipette solution (pH 7.2) consisted of 140 mM KCl, 10 mM EGTA, 1 mM MgCl2, 10 mM HEPES, 4 mM Na2ATP, and 0.1 mM Na3GTP. Cells were normally superfused at a rate of 1 ml/min when drug treatments were tested. To test the effects of hypoxia, cells were superfused at 3 ml/min with bath solution steadily bubbled with 95% N2/5% CO2 [hypoxia = 6.2 ± 0.3% O2 in the experimental chamber, as measured with an optical oxygen meter (FireStingO2, PyroScience)]. All experiments on pulmonary arterial myocyte K+ currents were recorded in the presence of paxilline (1 μM) to block the large conductance voltage- and calcium-activated K+ channel. Signals were sampled at 10 kHz and low-pass–filtered at 2 kHz. Conductance values (G) were calculated from the equation G = I/(V − EK), where the Nernst equilibrium potential (EK) was calculated as −89 mV at 37°C. Normalized conductance/voltage profiles for KV currents were fitted to a single Boltzmann function with the form G = Gmax/(1 + exp[−(VV0.5)/k]), where Gmax is the maximal conductance, V0.5 is the test potential for half-maximal conductance (G0.5), and k represents the slope of the activation curve. Data are expressed as I/Imax, where Imax is the current magnitude recorded at the onset of a given experimental intervention. Voltage-clamp acquisition and analysis protocols were performed using an Axopatch 200A amplifier/Digidata 1200 interface controlled by Clampex 10.0 software (Molecular Devices). Offline analysis was performed using Clampfit 10.0 (Molecular Devices).

HEK293 cell transfection

HEK293 cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% (v/v) fetal bovine serum and 1% (v/v) penicillin/streptomycin. Cells were transfected with 4 μg of pcDNA3.1 encoding a WT or mutant human KV1.5 (KCNA5) using 10 μl of Lipofectamine 2000 (Thermo Fisher Scientific) and used 24 to 48 hours later.

Phosphorylation assays

AMPK recognition motifs were identified by medium stringency screen on SCANSITE4 ( and then confirmed and scored on GitHub (, as previously described by others (37). We performed phosphorylation assays using AMPK purified from bacteria and phosphorylated with CaMKK2 (10 U/ml) in the presence of 200 μM AMP for 30 min at 30°C, as described previously (38).

Statistical comparisons

Means ± SEM were compared using GraphPad Prism 6 and t test for paired observations and one-way analysis of variance (ANOVA) with Dunnett or Bonferroni post hoc analysis for multiple comparisons tests. P values < 0.05 were considered significant, assuming mice of a given genotype or cells thereof do not differ in their mean phenotype.


Fig. S1. End point reverse transcription PCR confirms AMPK deletion.

Fig. S2. AMPK-α1 deletion blocks HPV during moderate and severe hypoxia.

Fig. S3. Effect of deletion of LKB1, AMPK-α1, or AMPK-α2 on peak velocity and VTI.

Fig. S4. HPV requires LKB1 but not CaMKK2.

Fig. S5. Deletion of AMPK-α1 but not AMPK-α2 blocks KV1.5 current inhibition by hypoxia.

Fig. S6. Deletion of AMPK-α1 blocks KV1.5 current inhibition by hypoxia.

Fig. S7. Deletion of AMPK-α1 but not AMPK-α2 blocks KV1.5 current inhibition by hypoxia.

Table S1. Basal cardiopulmonary hemodynamics.

Table S2. Electrophysiological characteristics of KV currents.


Acknowledgments: We thank K. Sakamoto (Nestlé Institute of Health Sciences, Lausanne, Switzerland) for providing additional supplies of the AMPK activator (C13) at a critical time. Funding: A British Heart Foundation Programme Grant (29885 to A.M.E.) was the primary source of funding for this study. Initial studies received additional support from a Wellcome Trust Programme Grant (081195 to A.M.E.). Contributions from F.A.R. and D.G.H. were also supported by a Wellcome Trust Investigator Award (204766 to D.G.H.). Author contributions: J.M.-S., S.A.L., A.T., C.M., and A.M.E. carried out ultrasound experiments and analysis. J.M.-S. and A.M.E. carried out mitochondrial imaging. J.M.-S. and A.M.E. carried out electrophysiology and analysis. S.A.L., J.M.-S., and A.M.E. completed quantitative PCR. F.A.R. and D.G.H. carried out protein phosphorylation and analysis. S.M. bred, maintained and genotyped mice, and carried out blood pressure measurements. B.V. and M.F. developed and supplied AMPK-α 1/2 floxed mice. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.

Stay Connected to Science Signaling

Navigate This Article