The FASEB Journal • Research Communication

Gene knockout of the KCNJ8-encoded Kir6.1 KATP

channel imparts fatal susceptibility to endotoxemiaGarvan C. Kane,*,† Chen-Fuh Lam,†,‡ Fearghas O’Cochlain,* Denice M. Hodgson,*Santiago Reyes,*,† Xiao-Ke Liu,*,† Takashi Miki,§ Susumu Seino,§ Zvonimir S. Katusic,†,‡

and Andre Terzic*,†,1

*Marriott Heart Disease Research Program Division of Cardiovascular Diseases, Department ofMedicine, †Department of Molecular Pharmacology and Experimental Therapeutics, ‡Department ofAnesthesiology, Mayo Clinic College of Medicine, Rochester, Minnesota, USA; and §Division ofCellular and Molecular Medicine, Kobe University Graduate School of Medicine, Kobe, Japan

ABSTRACT Sepsis, the systemic inflammatory re-sponse to infection, imposes a high demand for bodilyadaptation, with the cardiovascular response a keydeterminant of outcome. The homeostatic elementsthat secure cardiac tolerance in the setting of the sepsissyndrome are poorly understood. Here, in a model ofacute septic shock induced by endotoxin challenge withEscherichia coli lipopolysaccharide (LPS), knockout ofthe KCNJ8 gene encoding the vascular Kir6.1 KATPchannel pore predisposed to an early and profoundsurvival disadvantage. The exaggerated susceptibilityprovoked by disruption of this stress-responsive sensorof cellular metabolism was linked to progressive dete-rioration in cardiac activity, ischemic myocardial dam-age, and contractile dysfunction. Deletion of KCNJ8blunted the responsiveness of coronary vessels to cyto-kine- or metabolic-mediated vasodilation necessary tosupport myocardial perfusion in the wild-type (WT),creating a deficit in adaptive response in the Kir6.1knockout. Application of a KATP channel opener drugimproved survival in the endotoxic WT but had noeffect in the Kir6.1 knockout. Restoration of the dila-tory capacity of coronary vessels was required to rescuethe Kir6.1 knockout phenotype and reverse survivaldisadvantage in lethal endotoxemia. Thus, the Kir6.1-containing KATP channel, by coupling vasoreactivitywith metabolic demand, provides a vital feedback ele-ment for cardiovascular tolerance in endotoxicshock.—Kane, G. C., Lam, C-F., O’Cochlain, F., Hodg-son, D. M., Reyes, S., Liu, X-K., Miki, T., Seino, S.,Katusic, Z. S., Terzic, A. Gene knockout of the KCNJ8-encoded Kir6.1 KATP channel imparts fatal susceptibil-ity to endotoxemia. FASEB J. 20, 2271–2280 (2006)

Key Words: ATP-sensitive K� channel � Kir6.1 � metabolism� sepsis � septic shock � vasodilation

Severe sepsis, the leading cause of fatality in criticallyill patients, is characterized by an immunologicallydriven systemic response that underlies host protectionagainst microbial infection (1–3). Central to this syn-drome is a pronounced cytokine-induced cardiovascu-

lar state defined by a hypercontractile heart and alteredvascular tone (4). Despite diverse infectious pathogens,the outcome in severe sepsis is governed by the natureand degree of the host response, with the cardiovascu-lar status an ultimate determinant of survival (5–7).The Gram-negative bacterial cell wall endotoxin, LPS(lipopolysaccharide), evokes a surge in tumor necrosisfactor-alpha (TNF�), precipitating an inflammatorymediated cascade of pathophysiologic events that im-poses a high demand on the hyperdynamic septic heart(4, 6). Although there is an understanding of theproinflammatory cytokine signaling that triggers hemo-dynamic changes in endotoxic and, more generally,septic shock (1, 4), less is known of the molecularcomponents that ensure metabolic stability while main-taining cardiac workload.

Particularly for organs that are highly active metabol-ically, such as the heart, imposition of the demands ofstress necessitates augmentation of energetic supply. Asheart muscle displays saturating levels of oxygen extrac-tion under basal conditions, the cardiac capacity tomeet increased demand is limited to the responsivenessof the coronary circulation to vasodilate and increaseflow (8). Severe sepsis and septic shock are typified bypotent systemic vasodilation (9–11), yet the cellularcomponents that couple vasoreactivity with changes inmetabolic demand are unsolved.

ATP-sensitive K� (KATP) channels have emerged asprotein complexes with the capability to decode signalsof metabolic distress (12–14). By virtue of an intimateintegration with intracellular energetic networks and acapacity for high-fidelity processing of incoming meta-bolic signals, KATP channels regulate cellular excitabil-ity-dependent functions (12–16). These channels aresituated in high density within metabolically activetissues, where changes in the energetic state are con-veyed through the ATP-binding cassette sulfonylureareceptor (SUR) regulatory subunit to the K� channel

1 Correspondence: Division of Cardiovascular Diseases, De-partment of Medicine, Mayo Clinic College of Medicine,Rochester, MN 55905, USA. E-mail: terzic.andre@mayo.edu;

doi: 10.1096/fj.06-6349com

22710892-6638/06/0020-2271 © FASEB

pore, Kir6.x of the channel complex (12–17). It is thetissue-dependent channel pore that, requiring the reg-ulatory SUR chaperone to function, adjusts membranepotential to match demand and maintain cellular well-being (12–17). In nonvascular tissue, including themyocardium, where the pore-forming isoform is Kir6.2,defects in KATP channel subunits jeopardize cellularstress tolerance and predispose to injury (13–19). Inthe vasculature, the pore-forming subunit of the KATPchannel has been identified as the homologous butgenetically distinct Kir6.1 protein encoded by theKCNJ8 gene, and Kir6.1-containing channels have beenimplicated in the regulation of arterial smooth muscletone (20–23). The functional tissue specificity of Kir6.1is underscored by knockout of the KCNJ8 gene thattranslates into loss of plasmalemmal KATP channelactivity in the vasculature, in particular the coronarybed, but also maintenance of channel activity else-where, including the heart (23). Vascular Kir6.1 KATPchannels may thereby harbor the potential to translatesignals of demand, as occur under syndromes of sys-temic metabolic stress such as in severe sepsis, intooptimal cardiovascular tolerance.

To test this hypothesis, we probed the outcome ofgenetic disruption of Kir6.1 in a model of severe sepsiswith acute endotoxic shock. Knockout of the KCNJ8gene interrupted a vital vasodilatory adaptive mecha-nism in LPS-induced endotoxemia, producing deficitsin coronary flow, ischemic cardiac dysfunction, andpremature death. Demonstrating the requirement ofKir6.1 in securing coronary vasoreactivity identifies amolecular link between metabolic demand and ade-quate cardiac performance in endotoxic shock.

MATERIALS AND METHODS

Kir6.1-deficient mice

Generated by targeted disruption of the KCNJ8 gene encod-ing the pore-forming Kir6.1 subunit and backcrossed formore than five generations to a C57BL/6 background, micedeficient in vascular KATP channels (Kir6.1-KO) maintainnormal KATP channel current in nonvascular tissues, includ-ing the myocardium (23). Six- to 8-wk-old Kir6.1-KO micewere compared with age- and sex-matched WT littermatecontrols. All groups of mice were selected at the time ofweaning and/or genotyping by tail polymerase chain reaction(PCR) (3–4 wk of age) and were observed for an average of 2wk before experiments. Comparisons between groups wereperformed by ANOVA, Student’s t tests, nonparametric or logrank tests, using JMP software (SAS). Data are presented asmean � se; n refers to the sample size. P � 0.05 waspredetermined. Chemicals were obtained from Sigma (St.Louis, MO, USA). The Mayo Clinic Institutional Animal Careand Use Committee approved all protocols.

Endotoxic shock

The shock state of severe sepsis was modeled by administra-tion of E. coli LPS serotype 0111:B4 (4, 6). Dose-responsemortality curves were obtained over a range of LPS concen-trations (0.1–1000 �g/g) administered i.p. in 100–200 �l

sterile saline. Unless otherwise stated, for in vivo experimentsthe dose of LPS was 15 �g/g. To monitor survival, mice wereobserved without anesthesia hourly for 24 h, every 6 h foranother 48 h, then daily for a total of 7 days. The effect ofcalcium channel blockade on LPS-induced mortality in theKir6.1-KO was assessed by comparing survival in verapamil (5�g/g s.c. every 5 h) vs. saline-treated mice. The effect of theKATP channel opener pinacidil (10 �g/g) on LPS-inducedmortality (lipopolysaccharide 35 �g/g) in WT and Kir6.1-KOwas evaluated and compared with vehicle-treated controls.

Telemetry and electrocardiography

To continuously monitor cardiac activity and core tempera-ture in the conscious state, telemetry devices (Data SciencesInternational, St. Paul, MN, USA) were implanted in theperitoneum and leads were tunneled s.c. in a lead II config-uration under isoflurane anesthesia in WT and Kir6.1-KO(13). After recovery from surgery (24 h), signals were ac-quired at 2 kHz before and after LPS administration. Forcontinuous surface electrocardiographic (ECG) recordingsfollowing anesthesia with isoflurane (1%), WT and Kir6.1-KOmice had continuous ECG monitoring via limb lead elec-trodes. Tracings were recorded for 20 min before and up to6 h after LPS administration.

Histopathology

Light microscopy was performed on paraffin-embedded myo-cardial sections stained with hematoxylin-eosin from 4%formalin-fixed left ventricles (LV) taken from WT andKir6.1-KO mice 4–6 h after LPS or saline vehicle administra-tion. Transmitted electron microscopy (EM) was performedon ultramicrotome-cut, lead citrate-stained LV sections with aJEOL 1200 EXII electron microscope (24).

In vivo hemodynamics

Echocardiography with heart rate measurements (c256 and15L8, Acuson, Stockton, CA, USA) was performed in lightlysedated (1% isoflurane) WT and Kir6.1-KO before, 90, 180,and 360 min after LPS administration. Images were digitallyacquired and stored for off-line blinded analysis. Echocardio-graphic measurements of LV dimensions were recorded atend-diastole (EDD) and end-systole (ESD) from three con-secutive cardiac cycles using the leading edge method (24).LV fractional shortening (% fibrous sheath) was calculated as% fibrous sheath � [(EDD-ESD)/EDD] � 100. Ejection time(Et) was determined from the actual pulsed-wave Dopplertracings of LV outflow by measuring the interval from thebeginning of the acceleration to the end of the deceleration.The myocardial velocity of LV circumferential shortening (Vcfexpressed in circumferences per s) was calculated as Vcf �[(EDD-ESD)/EDD]Et. Blood pressure was measured underlight sedation (1% isoflurane) by tail-cuff (Columbus Instru-ments, Columbus, OH, USA; ref. 25) before and 90 min afterLPS challenge.

Blood analysis

Blood oxygen concentration, pH, lactate (iSTAT Corpora-tion, Abbott Park, IL, USA), and glucose (Glc) (Lifescan,Milpitas, CA, USA) were measured 6 h post-lipopolysaccha-ride or postsaline. Serum TNF� levels were quantified atbaseline, 90 and 180 min after LPS administration by ELISA(R&D Systems, Minneapolis, MN, USA).

2272 Vol. 20 November 2006 KANE ET AL.The FASEB Journal

Coronary flow

The aorta was cannulated in situ, heart excised, retrogradelyperfused at 90 mm Hg with Krebs-Henseleit buffer (bubbledwith 95% O2/5% CO2 at 37°C; pH 7.4), and paced at a rateof 600 beats/min (24). Coronary flow was continuouslymeasured with a T106 small animal blood flow meter (Tran-sonic Systems Inc., Ithaca, NY, USA). The effect of recombi-nant murine TNF� was assessed at least 20 min after flowstabilization. In a subset of hearts, TNF� was preceded by 10min perfusion of 10 �M glyburide in the WT or followed by1 �M verapamil in the Kir6.1-KO. Maximum vasodilation wasdefined as the peak increase in coronary flow observed in theWT in response to 10 �l of 1 mM adenosine after at least 10min of return to baseline flow on TNF� washout.

Aortic reactivity

Isolated aortic rings from WT and Kir6.1 knockout mice werestudied in parallel. Rings were mounted in a thermostaticorgan bath for isometric tension recordings and perfusedcontinuously with a Krebs-Henseleit buffer (bubbled with95% O2/5% CO2 at 37°C) as described (26). After anequilibration period of 30 min, the rings were progressivelystretched to their optimal passive tension as assessed by theresponse to 0.1 M KCl. Vasomotor function of aortic ringsfrom WT and Kir6.1 KO (n�6–8 each) were studied withconcentration-dependent response curves generated in re-sponse to cumulative addition of HCl to alter pH or to theKATP channel opener pinacidil (10�9 to 10�4 M).

RESULTS

Knockout of KCNJ8 predisposes toLPS-induced death

Infection by a pathogen provokes sepsis, with mortalitydetermined by the host response (1, 2). The shock stateof severe sepsis was simulated by i.p. administration ofthe endotoxin E. coli LPS in WT mice and in littermateslacking the pore-forming subunit of vascular KATPchannels (Kir6.1-KO) through deletion of the KCNJ8gene. In response to LPS, all mice displayed decreased

activity, piloerection, periocular discharge, and diar-rhea, typical early signs of endotoxemia. However,compared with the more tolerant WT (6 out of 22) orto vehicle-challenged Kir6.1-KO (0/10), in Kir6.1-KOmice an absence of vascular KATP channels predisposedto a prompt and massive mortality after 15 �g/g LPSadministration (22/22, P�0.0001; Fig. 1A). Kir6.1-KOmice began to die within the first 3 h, with 50% deathobserved at 7 h and mortality in the whole cohortconfirmed within 24 h (Fig. 1A). In WT, no fatality wasobserved in the initial 24 h, with the majority (73%)surviving long-term (Fig. 1A). In the absence of Kir6.1,the high susceptibility to endotoxin-mediated stress wasconfirmed over a range of LPS concentrations (Fig. 1B;P�0.009), with a readily demonstrable early deathfollowing LPS (Fig. 1C) in the Kir6.1-KO (n�39) com-pared with the WT (n�61; P�0.0055). The LD50 to LPSat 12 h (Fig. 1C) was 27-fold higher for Kir6.1-KO (6.7�g/g) than the WT (185.4 �g/g; P�0.0001). Thesefindings indicate a vital role for intact Kir6.1 in the hostresponse to LPS challenge.

Lack of Kir6.1 compromises cardiac adaptiveperformance in endotoxic shock

The development of cardiovascular impairment is acritical determinant of mortality in sepsis and septicshock (3–6). Death in the Kir6.1-KO following LPS waspreceded by a progressive decline in cardiac activitycaptured on telemetry (n�4; Fig. 2A, B). This was incontrast to the WT (n�3) in which cardiac activity waspreserved above preLPS levels (Fig. 2A, B). WT micehad normal electrocardiograms both before and afterLPS challenge (Fig. 2C). Electrocardiograms of theKir6.1-KO at baseline showed intermittent transientabnormalities due to lack of coronary KATP channelactivity (23) and demonstrated, under LPS challenge,severe and persistent ST segment change, an electro-physiological marker of the development of globalmyocardial ischemic injury (Fig. 2C). At autopsy, hearts

Figure 1. Survival disadvantage in endotoxicKir6.1-KO. The E. coli LPS endotoxin producedpremature and pronounced mortality in Kir6.1knockout (Kir6.1-KO) compared with time-matched WT mice (A, P�0.0001). Comparedwith WT, the Kir6.1-KO displayed a high sus-ceptibility to a dose range of LPS (B, P�0.01)with a readily demonstrable early death (C:P�0.006).

2273ENDOTOXIC KIR6.1 KNOCKOUT

from Kir6.1-KO mice displayed early features of myo-cyte coagulation necrosis with cytoplasmic hypereosino-philia within 4–6 h of LPS administration (n�8),pathological findings that were absent in endotoxic WT(n�8; Fig. 3A) or saline-treated Kir6.1-KO hearts (n�6;not illustrated). Distinct from the normal architectureof WT hearts, the ultrastructure of cardiomyocytes inLPS-challenged Kir6.1-KO demonstrated disarray (Fig.3B) with swollen, amorphous deposit-laden mitochon-dria (Fig. 3B, inset). Such pathological findings inKir6.1-KO hearts, indicative of ischemic cardiomyocyteinjury, translated into contractile dysfunction docu-mented by serial echocardiography. While both the WT(n�10) and Kir6.1-KO (n�10) were normal beforeLPS, only the WT augmented cardiac performanceearly in response to LPS (% fractional shorteningincreased from 39�2 to 49�4; P�0.05) and main-tained normal contractility through 3 and 6 h (Fig. 3C,E). In contrast, Kir6.1-KO mice displayed no earlypositive inotropy (% fractional shortening of 40�2 vs.42�4, P�0.64; Fig. 3C, D), and progressed to leftventricular failure documented by deterioration of

both left ventricular fractional shortening (Fig. 3D) andcircumferential shortening velocity (Fig. 3E). Cardiacdysfunction in Kir6.1-KO was mediated by a markedelevation in end-systolic left ventricular diameter(54�13% increase from baseline, P�0.02), whereas noleft ventricular dilation was noted in the WT (–3.3�5%,P�0.6; Fig. 3C). Furthermore, Kir6.1-KO was deprivedof tachycardia that accompanied the augmented car-diac performance in the endotoxic WT (Fig. 2A).Three hours after LPS, WT increased heart rates by16 � 6% (P�0.05), a response not observed inKir6.1-KO (–1�7%; P�0.92). In fact, in Kir6.1-KO at6 h, heart rates had declined by 33 � 6% (P�0.02) to360 � 19 beats per min while rates remained normal inWT at 476 � 22 beats per min (P�0.62). Thus, inresponse to the demands imposed by LPS, Kir6.1-KOmice failed to augment cardiac performance, and de-veloped marked ST segment elevation preceding andpersisting with the progression to extreme bradycardia,extensive cardiomyocyte injury, and severe contractiledysfunction.

Compared with the WT (n�4) that maintained a

Figure 2. Compromised cardiac activity after LPS in Kir6.1-KO. After LPS administration and preceding death, cardiac activityprogressively declined in the Kir6.1-KO, but not in WT, as captured by telemetry for individual cases (A) or collectively inrespective cohorts (B). WT manifested normal P, QRS, and T wave patterns on continuous lead II and III electrocardiogramsat baseline and following LPS administration (C, upper row). Kir6.1-KO (KO), which in most of the time monitored showunremarkable electrocardiograms at baseline (characteristic but short-lived and episodic ST changes are not shown),demonstrated under LPS challenge marked and persistent ST segment elevation (C, lower row), highlighted in tracingsmagnified in insets.

2274 Vol. 20 November 2006 KANE ET AL.The FASEB Journal

body temperature at 37.3 � 0.2°C and 35.3 � 1.0°Cbefore and 6 h post-lipopolysaccharide (P�0.2), thecardiac-impaired endotoxic Kir6.1-KO (n�3) devel-oped hypothermia, dropping from 36.5 � 0.5°C to29.3 � 1.0°C (P�0.02). These systemic perturbations inthe endotoxic Kir6.1-KO (n�5) were associated with agreater degree of metabolic acidosis characterized bylower serum pH (7.14�0.02 vs. 7.22�0.05; P�0.05)and higher serum lactate (4.0�1.6 mM vs. 1.8�0.2 mM;P�0.05) than seen in the WT (n�6). Noncardiovascu-lar parameters, including hypoglycemia, were compa-rable between the WT (serum Glc from 166�15 to52�14 mg/dl; P�0.001) and Kir6.1-KO (from 152�4to 44�6 mg/dl; P�0.001), as was blood oxygen con-centration (68.6�9 mm Hg in WT and 62.3�13 mmHg in Kir6.1-KO; P�0.67). Thus, vascular Kir6.1-con-taining KATP channels contribute to attainment of theenergetically demanding hyperdynamic cardiac state

and maintenance of metabolic stability in endotoxicshock.

Loss of cytokine-induced vasodilation in Kir6.1-KOcompromises coronary flow

The septic response is distinguished by a proinflamma-tory mediator cascade that activates cellular defensemechanisms required to combat infection (1). A criti-cal early common denominator of this response toinfectious organisms and their endotoxins is the cyto-kine TNF�, which acts as an early mediator of thevasodilatory septic shock syndrome (27). Before LPSadministration, both Kir6.1-KO and WT mice had un-detectable serum TNF� levels that rose to a similarextent after LPS injection, peaking at 90 min (Fig. 4A).This suggests that the early cardiac dysfunction ob-

Figure 3. Ischemic cardiac dysfunction precedes LPS-induced death in Kir6.1-KO. Myocyte necrosis on light (20�; arrows inpanel A) and electron (B) microscopy with mitochondrial swelling (B, inset) in Kir6.1-KO (KO), but not WT myocardial samplesat 4–6 h following LPS. B) Inset bar corresponds to 1 �m. Unlike WT, endotoxic-KO mice failed to initially augment cardiacperformance, and progressed to severe depression of ventricular function measured by left ventricular fractional shortening(LVFS, P�0.05; C, D) and circumferential shortening velocity (Vcf, P�0.05; C, E). C) Full and open arrows indicate diastolic andsystolic dimensions, respectively. D) * and **, significant difference from time 0 and corresponding WT, respectively. Grayshading depicts range of normal contractile function.

2275ENDOTOXIC KIR6.1 KNOCKOUT

served in the endotoxic Kir6.1-KO was independentfrom the LPS-induced serum TNF� dynamics, implyinga compromise in the host response to an otherwiseequivalent cytokine surge.

Execution of cardiac performance under stress ne-cessitates adjustment in coronary flow to meet demand(8). The TNF�-initiated hyperdynamic syndrome ofsevere sepsis characterized by increased cardiac outputwith vasodilation (4, 5, 27, 28). While having similarrates of coronary flow at baseline (P�0.98), WT andKir6.1-KO displayed a differential cytokine-induced va-sodilatory response. Recombinant TNF�, administeredat a dose (4 �g/l) comparable to the peak serum levelobserved during LPS challenge (Fig. 4A), inducedcoronary vasodilation only in WT hearts (Fig. 4B, C).This local TNF�-induced increase in flow was notobserved in Kir6.1-KO (at 4 �g/l or even higher dosesof 8 or 20 �g/l TNF�) or in WT hearts pretreated withthe KATP channel inhibitor, the sulfonylurea glyburide(Fig. 4B, C). This indicates that unimpeded KATP chan-nel activity is required to secure vascular smooth muscletone regulation under a central early mediator ofendotoxic shock. Furthermore, mirroring the lack of

response at the coronary level, Kir6.1-KO did notsystemically vasodilate (mean arterial pressure from95�8 mm Hg before to 103�15 mm Hg after LPS,P�0.6), unlike the WT mice that manifested pro-nounced systemic vasodilation indicated by a 29 � 4%fall in mean arterial pressure (85�6 mm Hg to 58�3mm Hg, P�0.0001; Fig. 4D). Thus, Kir6.1 is requiredfor adaptive vasodilatory response to endotoxic shock.

Metabolic vasoreactivity lost in Kir6.1-KO

The downstream consequence of the cytokine-inducedseptic shock syndrome, the metabolic state of acidosis isitself a contributor to the vasodilatory state. Unlike theresponse in WT vessels, which relaxed to the modestreductions in pH observed in endotoxemia, in theKir6.1-KO or glyburide-treated WT, acidosis did notinduce vasorelaxation (Fig. 5A). These differences per-sisted whether the endothelium was intact or not (notillustrated). Moreover, the potent endogenous vasodi-lator adenosine, which has been implicated in themaintenance of organ perfusion in sepsis (29, 30) and

Figure 4. Loss of cytokine-induced vasodilationin Kir6.1-KO. Comparable levels of the cytokineTNF� in sera of WT and Kir6.1-KO mice beforeand after induction of the endotoxic state (A).For each time point, 3–5 WT or Kir6.1-KO micewere sampled. WT hearts responded to TNF�with coronary vasodilation (B, C). No increasein coronary flow under TNF� was observed inKir6.1-KO hearts or WT hearts pretreated withthe KATP channel inhibitor, glyburide (Gly; B,C). LPS-challenged Kir6.1-KO mice demon-strated no systemic vascular response measuredby systolic blood pressure (D).

2276 Vol. 20 November 2006 KANE ET AL.The FASEB Journal

cardioprotection in the septic heart (31), inducedrobust coronary vasodilation in WT hearts (Fig. 5B). Yetthe coronary vascular bed of the Kir6.1-KO failed tovasodilate on adenosine administration (Fig. 5B), dis-playing vasoconstriction analogous to the response toTNF� (Fig. 4C). These findings indicate a generaldeficit in vasodilation in the absence of Kir6.1, which isotherwise critical in the WT, where vital cytokine- ormetabolic-mediated coronary vasoreactivity supportsmyocardial perfusion in response to demand.

Potassium channel opener-induced survival benefit inendotoxic shock

Systemic administration of a KATP channel openerresulted in arterial (including coronary) vasodilation,an action notably absent in Kir6.1-KO animals (Fig. 6A;ref. 23). In the WT, at a dose of LPS that approximates

the lethal dose50 (Fig. 1B), provision of pinacidil, theprototypic and clinically available potassium channelopener, abrogated LPS-induced mortality (Fig. 6B, C).Indeed, in the WT pinacidil reduced 24 h mortalityfrom 60% to zero (Fig. 6B) and provided enduringprotection against endotoxic shock, increasing long-term survival by 72% (Fig. 6C). Administration ofpinacidil had no protective benefit in Kir6.1-KO mice(Fig. 6B, C). Thus in the setting of life-threatening LPSchallenge, whereas deletion of Kir6.1 shifts the survivalcurve toward an exaggerated mortality, use of a potas-sium channel opener imparts an increased survival,indicating the vital role of adequate vascular KATPchannel function under endotoxemia.

Rescue of adaptive vascular response offsetsLPS-induced mortality in Kir6.1-KO

As the absence of KATP channel activity predisposes tocalcium-dependent smooth muscle constriction (21,32, 33), rescue of the vasodilatory response inKir6.1-KO was achieved by bypassing the defectivecapacity of vessels to secure flow (Fig. 7A). Applicationof the calcium channel antagonist, verapamil, toKir6.1-KO hearts challenged with TNF� reinstated cor-onary vasodilation (Fig. 7A) to a level comparable tothat of the flow response of the WT (Fig. 4B). Further-more, calcium channel inhibition negated the dispro-portionate LPS-induced mortality in the Kir6.1-KO(P�0.05, Fig. 7B), resulting in a significant increase inmedian survival time from 7 to 24 h. Restoration of thevascular response to endotoxin-mediated stress afterablation of Kir6.1 is thus necessary to secure adequateorgan perfusion, translating into prevention of prema-ture and exaggerated mortality in endotoxic shock.

DISCUSSION

In the adaptation to stress-imposed demand, the hostrelies on reactive processes to achieve a homeostaticstate at a superior level of performance (34–37). Themolecular determinants that secure an elevated state ofadaptation to a spectrum of environmental challenges,and in particular septic threats imposed by endotoxins,remain poorly defined. Here we identify Kir6.1 as anessential molecular component in the response tolethal endotoxemia. Targeted disruption of the KCNJ8gene that encodes the Kir6.1 pore and ablates vascular,but not nonvascular, KATP channel activity (23) com-promised coronary flow, precipitating cardiac damageand death after LPS challenge. Although with progres-sion of the sepsis syndrome to severe sepsis and septicshock, excessive systemic vasodilation is deleterious andrelated to poor outcome (2–5), the present findingspoint to the significance of adequate, Kir6.1-supportedcoronary perfusion in maintaining cardiac function.

KATP channels have been implicated as endogenousmechanisms of cardiac stress tolerance, harnessing thecapacity to translate signals of cellular distress, through

Figure 5. Loss of metabolic-induced vasodilation in Kir6.1-KO. Vasorelaxation induced by modest reductions in pH, afeature of the sepsis syndrome, in WT isolated aortic ringswere absent in the presence of the KATP channel blockerglyburide or in the Kir6.1-KO (A). Coronary vasoconstrictionto adenosine in Kir6.1-KO at odds to vasodilation in WT,demonstrated here in isolated perfused hearts (B).

2277ENDOTOXIC KIR6.1 KNOCKOUT

metabolic sensing capabilities, into cardioprotectivestress adaptation (38–40). While the stress-responsiverequirement of KATP channels has been suggested fortissues with Kir6.2-containing channels such as themyocardium (13, 41–43), the present findings impli-cate a comparable cardioprotective role for Kir6.1-containing KATP channels in the coronary vasculaturecontributing to a cardiac reserve necessary to respondto the metabolic stress imposed by endotoxic shock.The predisposition to precipitous mortality in endotox-emia in the KCNJ8 knockout correlated with a declinein electrical and mechanical cardiac activity, andmarked electrophysiologic and histopathologic evi-dence of myocardial necrosis not observed in the moretolerant WT. While multiple factors may influence theKir6.1-KO myocardial response, a major mechanismcontributing to myocardial dysfunction and cell deathin the Kir6.1-KO was mapped to a breakdown in thecoupling of metabolic demand and vasoreactivity. Theimportance of Kir6.1, required for KATP channel-medi-ated coronary dilatation (23), was underscored by theaberrant coronary response to the cytokine TNF� ormetabolic challenges of acidosis or adenosine in theabsence of the channel protein. Restoration of ade-quate coronary perfusion through L-type calcium chan-nel inhibition with improved survival in the Kir6.1-KO,and the benefits of potassium channel opener applica-tion in the septic WT further suggested a vital role foradequate Kir6.1-mediated KATP channel function in thecardiovascular response to endotoxic challenge.

The findings of hypothermia and acidosis, underendotoxic challenge, in the Kir6.1-KO appear linked toa progressive impairment in cardiac output, althoughdirect effects of Kir6.1 knockout on centers of homeo-static integration, such as in the hypothalamus, cannotbe excluded. In fact, distribution of the Kir6.1-contain-ing KATP channels throughout tissue beds may impli-

Figure 6. KATP channel opener mediated protection against endotoxic shock. Pinacidil induced vasorelaxation (% change frombaseline) in WT but not Kir6.1-KO (P�0.001; A). Pinacidil (pin) eliminated 24 h mortality in septic WT but had no effect ondeath rates in Kir6.1-KO (B). Pinacidil (pin) provided persistent survival benefit in WT (P�0.007; C) but not in Kir6.1-KO(P�0.9; C) LPS-challenged mice. Rescue experiments are based on 26 WT and 13 Kir6.1-KO mice.

Figure 7. Rescue of Kir6.1-KO phenotype by restoration ofthe coronary dilatory capacity. The calcium channel antago-nist, verapamil, restored coronary vasodilation in TNF�-challenged Kir6.1-KO (A) negating premature LPS-inducedmortality (B). Rescue experiments are based on a total of 14mice.

2278 Vol. 20 November 2006 KANE ET AL.The FASEB Journal

cate broader consequences contributing to the Kir6.1knockout phenotype under LPS challenge. Indeed, theabsence of systemic hypotension early after LPS admin-istration, at a time of preserved cardiac output, indi-cated a lack in the Kir6.1-KO of the drop in systemicvascular resistance-pathognomonic of the distributiveshock state. This finding parallels the vasoconstrictionand restoration of blood pressure seen with the sulfo-nylurea glyburide in established endotoxic hypotension(9, 44). While it is recognized that excessive systemicvasodilation late in the syndrome of severe sepsis isassociated with poor outcome, it is increasingly beingidentified that early in the hyperdynamic phase of thesyndrome, coronary vasodilation may be a requiredsupport of myocardial function (45–48). Further anal-ysis would determine the generalizability of the role ofKir6.1-containing KATP channels as necessary homeo-static molecules in endotoxemia to the wider syndromeof sepsis and other states of shock.

Through genetic ablation of channel subunits, vascu-lar KATP channels have recently been linked to themaintenance of arterial tone with knockout mice proneto vascular spasm and arrhythmic death (23, 33). Theimportance of intact Kir6.1-containing vascular chan-nels in securing optimal myocardial stress responsethereby establishes a framework for diagnostic and/ortherapeutic intervention in critical illness. In this re-gard, defects in KATP channel proteins and/or modu-lators of channel function (23, 33, 49–53) warrantinvestigation as determinants of hemodynamics indisease.

We thank Dr. M. C. Aubry for review of pathologicalspecimens, Mayo Clinic Translational Ultrasound ResearchCore for use of the echocardiographic machine, and JonNesbitt and Nadezda Alekseyeva for technical assistance. Thiswork was supported by grants from the National Institutes ofHealth, Marriott Heart Disease Research Program, Ted NashLong Life Foundation, Ralph Wilson Medical Research Foun-dation, Mayo Clinic and Japanese Ministry of Education,Science, Sports, Culture and Technology.

REFERENCES

1. Cohen, J. (2002) The immunopathogenesis of sepsis. Nature420, 885–891

2. Martin, G. S., Mannino D. M., Eaton, S., and Moss, M. (2003)The epidemiology of sepsis in the United States from 1979through 2000. N. Engl. J. Med. 348, 1546–1554

3. Levy, M., Fink, M., Marshall, J., Abraham, E., Angus, D., Cook,D., Cohen, J., Opal, S., Vincent, J., and Ramsay, G. (2003)SCCM/ESICM/ACCP/ATS/SIS 2001 International Sepsis Defi-nitions Conference. Intensive Care Med. 29, 530–538

4. Natanson, C., Eichenholz, P. W., Danner, R. L., Eichacker, P. Q.,Hoffman, W. D., Kuo, G., Banks, S. M., MacVittie, T. J., andParrillo, J. E. (1989) Endotoxin and tumor necrosis factorchallenges in dogs simulate the cardiovascular profile of humanseptic shock. J. Exp. Med. 169, 823–832

5. Rivers, E., Nguyen, B., Havstad, S., Ressler, J., Muzzin, A.,Knoblich, B., Peterson, E., and Tomlanovich, M. (2001) Earlygoal-directed therapy in the treatment of severe sepsis andseptic shock. N. Engl. J. Med. 345, 1368–1377

6. Suffredini, A. F., Fromm, R. E., Parker, M. M., Brenner, M.,Kovacs, J. A., Wesley, R. A., and Parrillo, J. E. (1989) The

cardiovascular response of normal humans to the administra-tion of endotoxin. N. Engl. J. Med. 321, 280–287

7. Ammann, P., Maggiorini, M., Bertel, O., Haenseler, E., Joller-Jemelka, H. I., Oechslin, E., Minder, E. I., Rickli, H., and Fehr,T. (2003) Troponin as a risk factor for mortality in critically illpatients without acute coronary syndromes. J. Am. Coll. Cardiol.41, 2004–2009

8. Berne, R. M. (1964) Regulation of coronary blood flow. Physiol.Rev. 44, 1–29

9. Landry, D. W., and Oliver, J. A. (2001) The pathogenesis ofvasodilatory shock. N. Engl. J. Med. 345, 588–595

10. MacMicking, J. D., Nathan, C., Hom, G., Chartrain, N., Fletcher,D. S., Trumbauer, M., Stevens, K., Xie, Q-W., Sokol, K., Hutchin-son, N., et al. (1995) Altered responses to bacterial infection andendotoxic shock in mice lacking inducible nitric oxide synthase.Cell 81, 641–650

11. Shesely E. G., Maeda, N., Kim, H. S., Desai, K. M., Krege, H.,Laubach, V. E., Sherman, P. A., Sessa, W. C., and Smithies, O.(1996) Elevated blood pressures in mice lacking endothelialnitric oxide synthase. Proc. Natl. Acad. Sci. U. S. A. 93, 13176–13181

12. Nichols, C. G. (2006) KATP channels as molecular sensors ofcellular metabolism. Nature 440, 470–476

13. Zingman, L. V., Hodgson, D. M., Bast, P. H., Kane, G. C.,Perez-Terzic, C., Gumina, R. J., Pucar, D., Bienengraeber, M.,Dzeja, P. P., Miki, T., et al. (2002) Kir6.2 is required foradaptation to stress. Proc. Natl. Acad. Sci. U. S. A. 99, 13278–13283

14. Miki, T., and Seino, S. (2005) Roles of KATP channels asmetabolic sensors in acute metabolic changes. J. Mol. Cell.Cardiol. 38, 917–925

15. Zingman, L. V., Alekseev, A. E., Bienengraeber, M., Hodgson,D., Karger, A. B., Dzeja, P. P., and Terzic, A. (2001) Signaling inchannel/enzyme multimers: ATPase transitions in SUR modulegate ATP-sensitive K� conductance. Neuron 31, 233–245

16. Selivanov, V. A., Alekseev, A. E., Hodgson, D. M., Dzeja, P. P.,and Terzic, A. (2004) Nucleotide-gated KATP channels inte-grated with creatine and adenylate kinases: amplification, tun-ing and sensing of energetic signals in the compartmentalizedcellular environment. Mol. Cell. Biochem. 256–257, 243–256

17. Inagaki, N., Gonoi, T., Clement, J. P., Namba, N., Inazawa, J.,Gonzalez, G., Aguilar-Bryan, L., Seino, S., and Bryan, J. (1995)Reconstitution of IKATP: An inward rectifier subunit plus thesulfonylurea receptor. Science 270, 1166–1170

18. Kane, G.C., Liu, X. K., Yamada, S., Olson, T. M., and Terzic, A.(2005) Cardiac KATP channels in health and disease. J. Mol. Cell.Cardiol. 38, 937–943

19. Suzuki, M., Li, R. A., Miki, T., Uemura, H., Sakamoto, N.,Ohmoto-Sekine, Y., Tamagawa, M., Ogura, T., Seino, S., Mar-ban, E., and Nakaya, H. (2001) Functional roles of cardiac andvascular ATP-sensitive potassium channels clarified by Kir6.2-knockout mice. Circ. Res. 88, 570–577

20. Quayle, J. M., Nelson, M. T., and Standen, N. B. (1997)ATP-sensitive and inwardly rectifying potassium channels insmooth muscle. Physiol. Rev. 77, 1165–1232

21. Yamada, M., Isomoto, S., Matsumoto, S., Kondo, C., Shindo, T.,Horio, Y., and Kurachi, Y. (1997) Sulphonylurea receptor 2Band Kir6.1 form a sulphonylurea-sensitive but ATP-insensitiveK� channel. J. Physiol. 499, 715–720

22. Repunte, V. P., Nakamura, H., Fujita, A., Horio, Y., Findlay, I.,Pott, L., and Kurachi, Y. (1999) Extracellular links in Kirsubunits control the unitary conductance of SUR/Kir6.0 ionchannels. EMBO J. 18, 3317–3324

23. Miki, T., Suzuki, M., Shibasaki, T., Uemura, H., Sato, T.,Yamaguchi, K., Koseki, H., Iwanaga, T., Nakaya, H., and Seino,S. (2002) Mouse model of Prinzmetal angina by disruption ofthe inward rectifier Kir6.1. Nat. Med. 8, 466–472

24. Hodgson, D. M., Zingman, L. V., Kane, G. C., Perez-Terzic, C.,Bienengraeber, M., Ozcan, C., Gumina, R. J., Pucar, D.,O’Coclain, F., Mann, D. L., et al. (2003) Cellular remodeling inheart failure disrupts KATP channel-dependent stress tolerance.EMBO J. 22, 1732–1742

25. O’Cochlain, D., Perez-Terzic, C., Reyes, S., Kane, G. C., Behfar,A., Hodgson, D., Strommen, J., Liu, X. K., van den Broek, W.,Wansink, D. G., et al. (2004) Transgenic overexpression ofhuman DMPK accumulates into hypertrophic cardiomyopathy,

2279ENDOTOXIC KIR6.1 KNOCKOUT

myotonic myopathy and hypotension traits of myotonic dystro-phy. Hum. Mol. Genet. 13, 2505–2518

26. D’Uscio, L. V., Baker, T. A., Mantilla, C. B., Smith, L., Weiler, D.,Sieck, G. C., and Katusic, Z. S. (2001) Mechanism of endothelialdysfunction in apolipoprotein E-deficient mice. Arterioscler.Thromb. Vasc. Biol. 21, 1017–1022

27. Tracey, K. J., Beutler, B., Lowry, S. F., Merryweather, J., Wolpe,S., Milsark, I. W., Hariri, R. J., Fahey, T. J., Zentella, A., Albert,J. D., et al. (1986) Shock and tissue injury induced by recombi-nant human cachectin. Science 234, 470–474

28. Cunnion, R. E., Schaer, G. L., Parker, M. M., Natanson, C., andParrillo, J. E. (1986) The coronary circulation in human septicshock. Circulation 73, 637–644

29. Ohta, A., and Sitkovsky, M. (2001) Role of G-protein-coupledadenosine receptors in downregulation of inflammation andprotection from tissue damage. Nature 414, 916–920

30. Motew, S. J., Mourelatos, M. G., Miller, R., Ferguson, J., and Law,W. R. (1997) Evidence that adenosine contributes to mainte-nance of hepatosplanchnic blood flow during peritoneal sepsisin rats. Shock 7, 439–446

31. Ismail, J., and McDonough, K. (1998) The role of coronary flowand adenosine in postischemic recovery of septic rat hearts.Am. J. Physiol. 275, H8–H14

32. Daut, J., Maier-Rudolph, W., von Beckerath, N., Mehrke, G.,Gunther, K., and Goedel-Meinen, L. (1990) Hypoxic dilation ofcoronary arteries is mediated by ATP-sensitive potassium chan-nels. Science 247, 1341–1344

33. Chutkow, W. A., Pu, J., Wheeler, M. T., Wada, T., Makielski,J. C., Burant, C. F., and McNally, E. M. (2002) Episodic coronaryartery vasospasm and hypertension develop in the absence ofSur2 KATP channels. J. Clin. Invest. 110, 203–238

34. McEwen, B. S. (1988) Stress, adaptation, and disease. Allostasisand allostatic load. Ann. N. Y. Acad. Sci. 840, 33–44

35. Mann, D. L. (2003) Stress-activated cytokines and the heart:from adaptation to maladaptation. Annu. Rev. Physiol. 65, 81–101

36. Bell, J. (2004) Predicting disease using genomics. Nature 429,453–456

37. Zingman, L. V., Hodgson, D. M., Alekseev, A. E., and Terzic, A.(2003) Stress without distress: homeostatic role for KATP chan-nels. Mol. Psychiatry 8, 253–254

38. Carrasco, A. J., Dzeja, P. P., Alekseev, A. E., Pucar, D., Zingman,L. V., Abraham, M. R., Hodgson, D., Bienengraeber, M., Puceat,M., Janssen, E., et al. (2001) Adenylate kinase phosphotransfercommunicates cellular energetic signals to ATP-sensitive potas-sium channels. Proc. Natl. Acad. Sci. U. S. A. 98, 7623–7628

39. Abraham, M. R., Selivanov, V. A., Hodgson, D. M., Pucar, D.,Zingman, L. V., Wieringa, B., Dzeja, P. P., Alekseev, A. E., andTerzic, A. (2002) Coupling of cell energetics with membranemetabolic sensing. Integrative signaling through creatine kinasephosphotransfer disrupted by M-CK gene knock-out. J. Biol.Chem. 277, 24427–24434

40. Alekseev, A. E., Hodgson, D. M., Karger, A. B., Park, S.,Zingman, L. V., and Terzic, A. (2005) ATP-sensitive K� channelchannel/enzyme multimer: metabolic gating in the heart. J.Mol. Cell. Cardiol. 38, 895–905

41. Suzuki, M., Sasaki, N., Miki, T., Sakamoto, N., Ohmoto-Sekine,Y., Tamagawa, M., Seino, S., Marban, E., and Nakaya, H. (2002)Role of sarcolemmal KATP channels in cardioprotection againstischemia/reperfusion injury in mice. J. Clin. Invest. 109, 509–516

42. Gumina, R. J., Pucar, D., Bast, P., Hodgson, D. M., Kurtz, C. E.,Dzeja, P. P., Miki, T., Seino, S., and Terzic, A. (2003) Knockoutof Kir6.2 negates ischemic preconditioning-induced protectionof myocardial energetics. Am. J. Physiol. 284, H2106–H2113

43. Kane, G. C., Behfar, A., Yamada, S. Y., Perez-Terzic, C.,O’Cochlain, F., Reyes, S., Dzeja, P. P., Miki, T., Seino, S., andTerzic, A. (2004) KATP channel knockout compromises meta-bolic benefit of exercise training resulting in cardiac deficits.Diabetes 53, S168–175

44. Landry, D. W., and Oliver, J. A. (1992) The ATP-sensitive K�

channel mediates hypotension in endotoxemia and hypoxiclactic acidosis in dog. J. Clin. Invest. 89, 2071–2074

45. Groeneveld, A. B., van Lambalgen, A. A., van den Bos, G. C.,Bronsveld, W., Nauta, J. J., and Thijs, L. G. (1991) Maldistribu-tion of heterogeneous coronary blood flow during canineendotoxin shock. Cardiovasc. Res. 25, 80–88

46. Grandel, U., Sibelius, U., Schrickel, J., Schmidt, D., Buerke, M.,Fink, L., Bournelis, E., Heep, M., Mayer, K., Bohle, R., et al.(2001) Biosynthesis of constitutive nitric oxide synthase-derivednitric oxide attenuates coronary vasoconstriction and myocar-dial depression in a model of septic heart failure induced byStaphylococcus aureus alpha-toxin. Crit. Care Med. 29, 1–7

47. Avontuur, J., Bruining, H., and Ince, C. (1995) Inhibition ofnitric oxide synthesis causes myocardial ischemia in endotox-emic rats. Circ. Res. 76, 418–425

48. Lopez, A., Lorente, J. A., Steingrub, J., Bakker, J., McLuckie, A.,Willatts, S., Brockway, M., Anzueto, A., Holzapfel, L., Breen, D.,et al. (2004) Multiple-center, randomized, placebo-controlled,double-blind study of the nitric oxide synthase inhibitor546C88: effect on survival in patients with septic shock. Crit. CareMed. 32, 21–30

49. Bienengraeber, M., Olson, T., Selivanov, V., Kathmann, E.,O’Cochlain, F., Gao, F., Karger, A., Ballew, J., Hodgson, D.,Zingman, L., et al. (2004) ABCC9 mutations identified inhuman dilated cardiomyopathy disrupt catalytic KATP channelgating. Nat. Genet. 36, 382–387

50. Ashcroft, F. M. (2005) ATP-sensitive potassium channelopa-thies. J. Clin. Invest. 115, 2047–2058

51. Jahangir, A., and Terzic, A. (2005) KATP channel therapeutics atthe bedside. J. Mol. Cell. Cardiol. 39, 99–112

52. Zager, R. A., Johnson, A. C., Lund, S., Hanson, S. Y., and Abrass,C. K. (2006) Levosimendan protects against experimental en-dotoxemic acute renal failure. Am. J. Physiol. 290, F1453–F1462

53. Kane, G.C., Behfar, A., Dyer, R. B., O’Cochlain, D. F., Liu, X.-K.,Hodgson, D. M., Reyes, S., Miki, T., Seino, S., and Terzic, A.(2006) KCNJ11 gene knockout of the Kir6.2 Katp channelcauses maladaptive remodeling and heart failure in hyperten-sion. Hum. Mol. Genet. 15, 2285–2297

Received for publication April 24, 2006.Accepted for publication June 16, 2006.

2280 Vol. 20 November 2006 KANE ET AL.The FASEB Journal