sGCα₁β₁ attenuates cardiac dysfunction and mortality in murine inflammatory shock models

Abstract

Altered cGMP signaling has been implicated in myocardial depression, morbidity, and mortality associated with sepsis. Previous studies, using inhibitors of soluble guanylate cyclase (sGC), suggested that cGMP generated by sGC contributed to the cardiac dysfunction and mortality associated with sepsis. We used sGCα₁-deficient (sGCα₁^−/−) mice to unequivocally determine the role of sGCα₁β₁ in the development of cardiac dysfunction and death associated with two models of inflammatory shock: endotoxin- and TNF-induced shock. At baseline, echocardiographic assessment and invasive hemodynamic measurements of left ventricular (LV) dimensions and function did not differ between wild-type (WT) mice and sGCα₁^−/− mice on the C57BL/6 background (sGCα₁^−/−B6 mice). At 14 h after endotoxin challenge, cardiac dysfunction was more pronounced in sGCα₁^−/−B6 than WT mice, as assessed using echocardiographic and hemodynamic indexes of LV function. Similarly, Ca²⁺ handling and cell shortening were impaired to a greater extent in cardiomyocytes isolated from sGCα₁^−/−B6 than WT mice after endotoxin challenge. Importantly, morbidity and mortality associated with inflammatory shock induced by endotoxin or TNF were increased in sGCα₁^−/−B6 compared with WT mice. Together, these findings suggest that cGMP generated by sGCα₁β₁ protects against cardiac dysfunction and mortality in murine inflammatory shock models.

in the united states, sepsis and septic shock develop in 750,000 people annually, of whom >210,000 die (1), thereby making sepsis a leading cause of death in intensive care units. Invading microorganisms (including gram-positive and -negative bacteria, viruses, fungi, and parasites) induce the release of a wide variety of proinflammatory mediators, causing a systemic inflammatory response syndrome. Systemic inflammatory response syndrome can develop independently of an infection, for example, in cases of pancreatitis or trauma, as well as after cardiopulmonary bypass. It is recognized that refractory hypotension and myocardial depression contribute significantly to the morbidity and mortality of sepsis and septic shock (33, 43, 57). Similarly, inflammatory shock induced by TNF is characterized by cardiovascular collapse (8, 34).

Abundant evidence suggests that nitric oxide (NO) has critical roles in the regulation of myocardial function and structure (5, 24, 48) and blood pressure (23, 50). NO plays a pivotal role in cytokine-induced myocardial dysfunction (17, 30, 52), either attenuating or exacerbating the adverse hemodynamic sequelae associated with systemic inflammation. NO is synthesized from l-arginine by a family of three enzymes referred to as NO synthases (NOSs). NOS1 and NOS3 are constitutively expressed in a variety of cells, including neuronal cells, endothelial cells, and cardiomyocytes. NOS2, or inducible NOS, was first identified in macrophages and has been detected in a wide variety of cells (including cardiomyocytes) exposed to endotoxin and cytokines, such as TNF, and produces high levels of NO in inflammatory settings (40). High levels of NO produced by NOS2 have been implicated in the development of cardiac dysfunction associated with inflammatory shock (26, 53). However, NOS2-derived NO has also been found to have protective effects in TNF-induced (9) and endotoxin-induced (59) shock models, curbing oxidative stress and apoptosis. Similarly, increased NO levels, associated with overexpression of NOS3 in cardiomyocytes, attenuated myocardial dysfunction in mice challenged with endotoxin (25). However, NOS3-derived NO was also shown to have adverse cardiovascular effects in endotoxin-induced (13) and anaphylactic (7, 10) shock models. Importantly, the serious adverse events in clinical trials studying the effect of NOS inhibitors on survival in patients with septic shock appeared to be primarily of cardiac origin (12, 20, 36, 46).

The multiple targets of NO contribute to the widely varying and often conflicting effects of this molecule. One of the primary targets of NO is soluble guanylate cyclase (sGC), a heme-containing heterodimeric enzyme, consisting of one α-subunit and one β-subunit, that generates the secondary messenger molecule cGMP (4). The regulatory effect of cGMP on myocardial contractile function (22, 27, 39, 49, 56) and ventricular myocyte contractility (21, 51) is well established. Although the sGCα₁β₁ heterodimer is considered to be the principal cardiovascular isoform (38), recent studies have suggested that low levels of cGMP generated by sGCα₂β, are sufficient to mediate many of NO's cardiovascular effects (6, 18, 37, 42, 54).

In shock, detrimental (11, 14, 28, 32, 52, 58) and protective (2, 29, 31) effects on cardiac function have been attributed to NO-dependent cGMP signaling. These dual effects of cGMP appear to depend, at least in part, on the timing of sGC inhibition: cGMP might be detrimental in one phase of inflammatory shock but protective in another (15, 16). Alternatively, it cannot be excluded that the source of cGMP (sGCα₁β₁ or sGCα₁β₂) determines its impact on cardiac function. Furthermore, interpretation of these studies is complicated by the incomplete specificity and systemic effects of the pharmacological agents used to inhibit sGC activity, e.g., 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one, LY-83583, and methylene blue (MB).

In this study, we used sGCα₁-deficient (sGCα₁^−/−) mice (6) to unequivocally determine the role of sGCα₁β₁ in the development of cardiac dysfunction associated with two inflammatory shock models: endotoxin- and TNF-induced shock. We found that sGCα₁^−/− mice are sensitized to the cardiac dysfunction and mortality associated with inflammatory shock, which suggests that sGCα₁β₁-derived cGMP is cardioprotective in murine inflammatory shock models.

MATERIALS AND METHODS

Mouse models.

sGCα₁^−/− mice were generated on the 129S6 background (sGCα₁^−/−S6) as previously described (6) and were back-crossed eight generations onto a C57BL/6 background (sGCα₁^−/−B6). sGCα₁^−/− mice carry a targeted deletion of the sixth exon of the gene encoding sGCα₁, resulting in the expression of a mutant, catalytically inactive protein.

For the endotoxin-induced inflammatory shock model, male wild-type (WT) C57BL/6 mice and sGCα₁^−/−B6 mice were challenged with Escherichia coli 0111:B4 endotoxin (25 mg/kg ip). For TNF-induced shock, male WT and sGCα₁^−/− mice on a C57BL/6 background were challenged with recombinant murine TNF (0.35 mg/kg iv) produced in E. coli and purified to homogeneity (9). All animal procedures were conducted in accordance with guidelines published in the Guide for the Care and Use of Laboratory Animals (National Research Council, National Academy Press, Washington, DC, 1996) and were approved by the Massachusetts General Hospital Subcommittee on Research Animal Care and by the Animal Ethics Committee of Ghent University.

sGC enzyme activity.

sGC enzyme activity was measured as described elsewhere (6). Lung tissues were harvested from 12- to 14-wk-old mice and subsequently homogenized. Supernatants (containing 50 μg of protein) were incubated for 10 min at 37°C in a reaction mixture with or without 10 μmol/l of l2-(N,N-diethylamino)-diazenolate-2-oxide. cGMP in the reaction mixture was measured using a commercial radioimmunoassay (Biomedical Technologies, Stoughton, MA). sGC enzyme activity is expressed as picomoles of cGMP produced per minute per milligram of protein in lung extract supernatant.

Hemodynamic measurements.

sGCα₁^−/−B6 and WT mice, between 10 and 14 wk of age, were anesthetized by intraperitoneal injection with ketamine (100 mg/kg), fentanyl (50 μg/kg), and pancuronium (2 mg/kg), intubated, and mechanically ventilated (fraction of inspired O₂ = 1, 10 μl/g, 120 breaths/min). A saline-filled catheter was inserted into the left carotid artery for infusion of saline (2 ml/h). For measurement of cardiac function, the chest was opened, and a pressure-volume conductance catheter (model SPR-839, Millar Instruments, Houston, TX) was introduced through the apex into the left ventricle (LV), as described previously (25). Heart rate (HR), LV end-systolic and end-diastolic pressures (LVESP and LVEDP, respectively), and LV volumes were measured. The maximum and minimum first derivative of developed LV pressure (dP/dt_max and dP/dt_min, respectively), stroke work (SW), maximal developed power (P_max), time constant of isovolumic relaxation, cardiac output (CO), stroke volume (SV), and arterial elastance (E_a, defined as the ratio of LVESP to SV) were calculated. dP/dt_max divided by instantaneous pressure (IP) was calculated and is considered a relatively load-independent index of contractility.

In separate groups of sGCα₁^−/−B6 and WT mice (n = 8 and 9, respectively), sodium nitroprusside (SNP, 2.5, 5, and 10 μg/kg), BAY 41-2272 (100 μg/kg), or vehicle (PBS for SNP and 4% DMSO and 8% cremophor in PBS for BAY 41-2272) was administered intravenously (50-μl bolus in the jugular vein), and LVESP was measured.

Noninvasive blood pressure measurements.

Mean arterial blood pressure (MAP) was measured noninvasively in conscious WT and sGCα₁^−/−B6 mice (n = 5 each) with use of a tail-cuff system (Kent) warmed to 37°C. Mice were studied at 12–14 wk of age. Mice were habituated to the blood pressure measurement device for 7–10 days and underwent 2 cycles of 10 measurements per day for 10 days for blood pressure determination.

Echocardiographic measurements.

Transthoracic echocardiograms were obtained using a 13-MHz probe (Vivid 7, GE Medical Systems) in awake WT and sGCα₁^−/−B6 mice 14 h after a challenge with saline (n = 5 and 5, respectively) or 25 mg/kg endotoxin (n = 13 and 18, respectively), as described previously (48). Measurements were made by an observer who was blinded to the experimental group. LV ejection fraction was measured on two-dimensional images using the area-length method. HR, LV end-systolic internal diameter, and LV fractional shortening were measured using an M-mode echocardiogram obtained at the midpapillary level, as described previously (48).

Measurement of gene expression.

Total RNA was extracted from LV tissue using TRIzol reagent (Invitrogen), and cDNA was synthesized using Maloney's murine leukemia virus RT (Promega). NOS2, IL-6, and 18S rRNA transcript levels were measured by real-time PCR using a Mastercycler ep realplex 2 (Eppendorf). Primers were designed for 18S (5′-CGGCTACCACATCCAAGGAA-3′ and 5′-GCTGGAATTACCGCGGCT-3′). For NOS2 and IL-6, TaqMan primer sets (Applied Biosystems) were used. Changes in the relative gene expression normalized to levels of 18S rRNA were determined using the relative cycle threshold method.

Immunoblot analysis.

LV tissue samples were homogenized in 2 ml of RIPA buffer (Boston BioProducts) supplemented with 1% protease inhibitor cocktail (Sigma) and microcentrifuged for 20 min at 20,000 g. Supernatant proteins (15 μg) were fractionated on 10% SDS-polyacrylamide gels and transferred to polyvinylidene difluoride membranes. Membranes were blocked for 1 h in Odyssey blocking buffer (LI-COR) and incubated overnight with primary rabbit antibodies against sGCα₁, sGCα₂, or sGCβ₁ (diluted 1:1,000; Abcam). Immunoblotting with a rabbit GAPDH antibody (Cell Signaling) was performed to confirm equal loading. Bound antibody was detected with anti-rabbit infrared dye (catalog no. 680, LI-COR) using an Odyssey infrared imaging system (LI-COR).

Cardiomyocyte isolation.

Cardiomyocytes were isolated as described previously (35). Briefly, mice were treated with heparin (200 U ip) and, 5 min later, were euthanized with pentobarbital sodium (10 mg ip). The chest was opened, and the heart was removed quickly and placed in cold (4°C) isolation buffer containing 137 mmol/l NaCl, 5.4 mmol/l KCl, 0.5 mmol/l MgCl₂, 10 mmol/l HEPES, 10 mmol/l glucose, 5 mmol/l taurine, and 10 mmol/l 2,3-butanedione monoxime, with pH titrated to 7.4 with NaOH. The aorta was mounted on a cannula for retrograde perfusion at 3.5 ml/min. After cannulation, the heart was perfused with 3 ml of cold isolation buffer to clear blood from the coronary system. The arrested heart was subsequently mounted on a perfusion apparatus and perfused at 37°C with isolation buffer for 2 min and then with isolation buffer containing a mixture of digesting enzymes [0.5 mg/ml collagenase B and 0.5 mg/ml collagenase D (Roche Diagnostics) and 0.1 mg/ml type XIV protease (Sigma-Aldrich)]. The enzyme-containing solution was perfused for 7 min or until coronary pressure (measured continuously with a manometer) was <20 mmHg. At the end of perfusion, the heart was removed from the cannula, and the LV was isolated, chopped into small pieces in isolation buffer at room temperature, and then gently agitated for another 5 min. The cell suspension was filtered through nylon gauze (200-μm mesh) and sedimented in a 50-ml tube for 15 min, and the supernatant was replaced three times with a higher-Ca²⁺ (200, 500, and 1,000 mmol/l) solution. Cells were transferred in a physiological solution (see below) and kept at room temperature until use.

Measurement of contractility and Ca²⁺ handling in isolated cardiomyocytes.

LV cardiomyocytes were dissociated from WT and sGCα₁^−/−B6 mice 14 h after challenge with saline or endotoxin (n = 10 each). In studies of cardiac myocyte contractility and Ca²⁺ handling, only cells that had a rod shape and clear striations, were quiescent in the absence of stimulation, showed visible twitches at 6 Hz, and did not have spontaneous Ca²⁺ release events and waves were included. Furthermore, all cells with a resting sarcomere length <1.65 μm were excluded. Measurements of sarcomere shortening (cell shortening, as a percentage of diastolic length) and intracellular Ca²⁺ (ΔCa_i, measured as the difference between the peak systolic-to-diastolic fura 2 ratios) were performed simultaneously and as described previously (25). Briefly, isolated cardiac cells were loaded with the membrane-permeable Ca²⁺ indicator fura 2-AM (Molecular Probes; 1 μmol/l, 25 min) and superfused with a physiological solution containing (in mmol/l) 137 NaCl, 5.4 KCl, 1.2 Ca²⁺, 0.5 MgCl₂, 10 HEPES, and 5.5 and 0.5 probenecid (to improve fura 2-AM retention), with pH titrated to 7.4 with NaOH. Cells were externally paced at 6 Hz (MyoPacer, Ionoptix) at 37°C, and sarcomere shortening and fura 2 ratiometric data were recorded and analyzed using IonWizard software (IonOptix). The integrated acquisition system (IonOptix) includes a Nikon TS100 inverted microscope, a MyoCam-S digital charge-coupled device video camera (with a 500-Hz acquisition rate), a HyperSwitch dual-excitation light source (with a paired 340-380-nm excitation rate of 250 Hz), and a photomultiplier tube. For statistical analysis, data were compared using an unpaired Student's t-test.

Survival analysis.

Survival after a challenge with endotoxin was studied in age-matched C57BL/6 WT and sGCα₁^−/−B6 mice. To avoid dehydration, saline (1 ml) was given intraperitoneally to all mice at time 0 and at 24 h after endotoxin challenge. Survival after a challenge with TNF was studied in age-matched WT and sGCα₁^−/− mice on the C57BL/6 background.

Statistical evaluation.

Unless stated otherwise, data were compared using multiple comparisons by two-way ANOVA. After ANOVA, Bonferroni's post hoc test was used to compare separate groups. Values are means ± SE. P < 0.05 was considered significant.

RESULTS

sGC enzyme activity, sGC expression levels, blood pressure, and cardiac function in sGCα₁^−/−B6 mice.

Similar to observations in sGCα₁^−/−S6 mice (6), NO-induced increases in sGC enzyme activity were severely attenuated in lung extracts from sGCα₁^−/−B6 mice compared with WT mice (Fig. 1). NO donor compounds reduced blood pressure in sGCα₁^−/−B6 mice, albeit to a lesser extent than in WT mice (Fig. 2A). BAY 41-2272 reduced blood pressure in WT, but not sGCα₁^−/−B6, mice (Fig. 2B). Expression of the sGCα₂ gene in aortic and lung tissue did not differ in WT and sGCα₁^−/−B6 mice (Fig. 3). In contrast to male sGCα₁^−/−S6 mice, male sGCα₁^−/−B6 mice do not develop hypertension, measured invasively in anesthetized mice or noninvasively in awake mice (Table 1). At baseline, cardiac function parameters, measured invasively (Table 1) or noninvasively (Fig. 4), were similar between sGCα₁^−/−B6 and WT mice.

Fig. 1.

Soluble guanylate cyclase (sGC) enzyme activity. Unstimulated (baseline) and 2-(N,N-diethylamino)-diazenolate-2-oxide (DEA-NO)-stimulated sGC enzyme activity in lung extracts of 10- to 12-wk-old wild-type (WT) and sGC-deficient (sGCα₁^−/−B6) mice. Values are means ± SE (n = 6 for all groups). *P < 0.001 vs. WT DEA-NO. #P < 0.001 vs. baseline.

Fig. 2.

Acute effects on left ventricular (LV) end-systolic pressure (LVESP) of compounds acting on the nitric oxide (NO)-cGMP pathway. A: effect of sodium nitroprusside (SNP, 2.5, 5, or 10 μg/kg iv) on LVESP in WT (n = 9) and sGCα₁^−/−B6 (n = 8) mice. Values are means ± SE. *P < 0.05. #P < 0.01 vs. WT. B: sGC stimulator BAY 41-2272 decreased LVESP in WT, but not sGCα₁^−/−B6, mice. Values are means ± SE. *P < 0.001 vs. WT BAY 41-2272. #P < 0.001 vs. WT vehicle.

Fig. 3.

Levels of sGCα₂ expression, as measured by quantitative RT-PCR, in extracts from aorta (A) and lung (B) of WT and sGCα₁^−/− mice on the 129S6 (S6) and the C57BL/6 (B6) backgrounds. Values are means ± SE of number of mice indicated below each bar.

Table 1.

LV function in WT and sGCα₁^−/−B6 mice 14 h after challenge with saline or endotoxin

	WT		sGCα1−/−B6
	Control	LPS	Control	LPS
n (invasive)	15	11	11	10
Body wt, g	24±0	24±0	23±1	24±0
HR, beats/min	647±12	563±13*	654±17	605±14†
LVESP, mmHg	111±2	103±5	112±3	117±6‡
LVEDP, mmHg	6±1	3±0*	6±1	3±1†
dP/dtmax, mmHg/s	14,003±610	11,226±919†	14,012±919	10,679±870†
dP/dtmin, mmHg/s	−13,690±513	−11,578±977	−13,642±730	−12,204±644
dP/dtmax/IP	160±6	152±9	163±12	133±6†
SW, mmHg·ml	1.9±0.1	1.8±0.3	2.0±0.1	1.1±0.2*‡
Pmax, mW	17±1	17±3	17±1	10±1*§
τ, ms	5.3±0.2	5.5±0.2	5.5±0.3	5.6±0.1
CO, ml/min	12±1	12±2	13±1	8±1*‡
SV, μl	19±1	21±3	20±1	13±1*§
Ea, mmHg/μl	6±0	6±1	6±0	10±1*§
n (tail cuff)	5		5
MAP, mmHg	107±1		108±1

Values are means ± SE. n, Number of 10- to 12-wk-old wild-type (WT) and soluble guanylate cyclase-α₁-deficient (sGCα₁^−/−B6) mice; HR, heart rate; LVESP, left ventricular (LV) end-systolic pressure; LVEDP, LV end-diastolic pressure; dP/dt_max and dP/dt_min, maximum and minimum rates of developed LV pressure; IP, instantaneous pressure; SW, stroke work; P_max, maximal power; τ, time constant of isovolumic relaxation; CO, cardiac output; SV, stroke volume; E_a, arterial elastance; MAP, mean arterial pressure.

^*P < 0.01;

^†P < 0.05 vs. control mice of the same genotype.

^‡P < 0.05;

^§P < 0.01 vs. septic WT mice.

Fig. 4.

Echocardiographic analysis of LV function. LV fractional shortening (FS, A), ejection fraction (EF, B), LV end-systolic internal diameter (LVID_s, C), and heart rate (HR, D) were measured in WT and sGCα₁^−/−B6 mice 14 h after challenge with saline or 25 mg/kg endotoxin (LPS). Values are means ± SE. *P < 0.001; #P < 0.01 vs. saline. $P < 0.01; &P < 0.05 vs. WT LPS.

Impact of sGCα₁ deficiency on myocardial dysfunction associated with endotoxin-induced inflammatory shock.

At 14 h after endotoxin challenge, sGCα₁^−/−B6 and WT mice developed hypothermia (27 ± 1°C and 29 ± 1°C) and bradycardia, measured noninvasively (Fig. 4D) and invasively (Table 1). The 14-h time point was chosen, because all mice were alive, and myocardial dysfunction was apparent. The decrease in LV fractional shortening, LV ejection fraction, and the increase in LV end-systolic internal diameter after endotoxin challenge, as measured by echocardiography, were more marked in sGCα₁^−/−B6 than in WT mice (Fig. 4, A–C). LVESP was higher in sGCα₁^−/−B6 than in WT mice after endotoxin challenge (Table 1). Although dP/dt_max, a measure of LV contractility that is relatively sensitive to afterload, was similarly depressed in WT and sGCα₁^−/−B6 mice after endotoxin challenge, dP/dt_max/IP, a load-independent parameter of LV contractility, was depressed to a greater extent in sGCα₁^−/−B6 than in WT mice. Similarly, SW and P_max, both indexes of LV contractility, were decreased in sGCα₁^−/−B6, but not WT, mice after endotoxin challenge. The relaxation time constant, a measure of diastolic function, tended to be prolonged in WT and sGCα₁^−/−B6 mice after endotoxin challenge. Importantly, CO was decreased in sGCα₁^−/−B6, but not WT, mice after endotoxin challenge. This decrease in CO was accompanied by a lower SV. Finally, E_a, a measure of afterload, was increased in sGCα₁^−/−B6, but not WT, mice after endotoxin challenge.

Effects of endotoxin on LV expression of genes encoding NOS2 and IL-6.

NOS2 and IL-6 mRNA levels in cardiac tissue extracts did not differ between WT and sGCα₁^−/−B6 mice at baseline (Fig. 5). NOS2 gene expression levels were not increased in the LV of WT or sGCα₁^−/−B6 mice 14 h after endotoxin challenge. IL-6 levels were similarly increased in WT and sGCα₁^−/−B6 mice 14 h after endotoxin injection (Fig. 5B).

Fig. 5.

Expression analysis of inflammatory markers. mRNAs encoding type 2 nitric oxide synthase (NOS2, A) and IL-6 (B) in the LV of WT and sGCα₁^−/−B6 mice were measured by quantitative RT-PCR 14 h after challenge with saline (n = 13 and 11, respectively) or 25 mg/kg endotoxin (LPS, n = 10 and 11, respectively). Values are means ± SE. *P < 0.01 vs. saline.

LV expression of sGC.

LV protein levels of sGCα₁, sGCα₂, and sGCβ₁ did not differ between saline- and endotoxin-treated WT mice or between saline- and endotoxin-treated sGCα₁^−/−B6 mice (Fig. 6). Expression levels of sGCβ₁ were lower in the LV of saline- and endotoxin-treated sGCα₁^−/−B6 than in WT mice (Fig. 6C).

Fig. 6.

Quantification of immunoblot analyses of sGCα₁ (A), sGCα₂ (B), sGCβ₁ (C), and GAPDH (D) in LV protein extracts of WT and sGCα₁^−/−B6 mice 14 h after challenge with saline or 25 mg/kg endotoxin (LPS). Blots represent results from 3 mice in each group. Values are means ± SE (n = 5 per group). *P < 0.001 and #P < 0.05 vs. WT in the same treatment group.

Effects of endotoxin on contractility and Ca²⁺ handling in cardiomyocytes isolated from WT and sGCα₁^−/−B6 mice.

At rest, sarcomere length was 1.77 ± 0.01 and 1.78 ± 0.01 μm for cardiomyocytes isolated from saline-treated WT and sGCα₁^−/−B6 mice, respectively, and 1.77 ± 0.01 and 1.79 ± 0.01 μm for cardiomyocytes isolated from LPS-treated WT and sGCα₁^−/−B6 mice, respectively. Cardiomyocyte contractility, as assessed by percent cell shortening (Fig. 7A), and Ca²⁺ handling, as measured by peak Ca²⁺ transient amplitude (ΔCa_i, Fig. 7B), were similar in cardiomyocytes isolated from saline-challenged WT and sGCα₁^−/−B6 mice paced at 6 Hz. Diastolic sarcomere length was similar in cardiomyocytes isolated from WT and sGCα₁^−/−B6 mice at baseline and after endotoxin challenge. However, the decrease in cell shortening was more marked in cardiomyocytes isolated from sGCα₁^−/− than from WT mice (Fig. 7A) 14 h after endotoxin challenge. Similarly, endotoxin challenge decreased ΔCa_i to a greater extent in sGCα₁^−/−B6 than in WT cardiomyocytes.

Fig. 7.

Percent cell shortening (CS, A) and amplitude of Ca²⁺ transient (ΔCa_i, B) in cardiomyocytes isolated from WT and sGCα₁^−/−B6 mice 14 h after challenge with saline or 25 mg/kg endotoxin (LPS) and paced at 6 Hz. Values are means ± SE of number of cells and mice (cells/mice) indicated below each bar. *P < 0.05; #P < 0.01 vs. saline-challenged mice of the respective genotype. $P < 0.05 vs. endotoxin-challenged WT mice.

Effects of endotoxin on survival in WT and sGCα₁^−/−B6 mice.

Decreased LV function was accompanied by increased mortality in endotoxin-treated sGCα₁^−/−B6 mice, as demonstrated by Kaplan-Meier survival curves (Fig. 8A).

Fig. 8.

Survival analysis. Kaplan-Meier curve shows survival after challenge of WT and sGCα₁^−/−B6 mice with endotoxin (LPS, A) or TNF (B). #P < 0.05; *P < 0.01 vs. WT (by log-rank test).

Effects of TNF on survival in WT and sGCα₁^−/−B6 mice.

Rectal body temperatures were measured after a challenge with TNF, a second model of inflammatory shock. At 9 h after a challenge with TNF, hypothermia was more pronounced in sGCα₁^−/−B6 than in WT C57BL/6 mice (27 ± 3°C vs. 31 ± 4°C, P < 0.01). Similar to our observations in endotoxin-treated mice, TNF-induced inflammatory shock was associated with a higher mortality in sGCα₁^−/−B6 than in WT mice (Fig. 8B).

DISCUSSION

In this study, mice deficient in sGCα₁ were used to investigate the function of cGMP generated by sGCα₁β₁ in murine models of endotoxin- and TNF-induced shock. The data suggest that cGMP generated by sGCα₁β₁ attenuates myocardial dysfunction and lethality associated with inflammatory shock in mice: mortality was higher and cardiac dysfunction was more pronounced in sGCα₁^−/−B6 than in WT mice. Furthermore, the adverse effects of endotoxin challenge on contractility and Ca²⁺ handling in cardiomyocytes were more marked in cardiomyocytes isolated from sGCα₁^−/−B6 than from WT mice. The use of genetically modified mice allowed us to avoid some of the limitations associated with studies focusing on the contribution of sGC to the pathogenesis of endotoxic shock with use of pharmacological inhibitors such as MB, 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one, and LY-83583, all of which modulate the activity of other enzymes in addition to sGC and none of which are sGC isoform specific. To assess the impact of sGCα₂β₁ on cardiac function in sepsis, a study of sGCα₂-deficient mice is necessary.

We previously reported the generation and cardiovascular phenotype of sGCα₁^−/− mice on a mixed Swiss/129S6 or a 129S6 background (sGCα₁^−/−S6 mice). sGCα₁^−/−S6 mice develop gender-specific and testosterone-dependent hypertension (6). However, neither WT 129S6 nor sGCα₁^−/−S6 mice developed cardiac dysfunction when challenged by endotoxin, Staphylococcus aureus, or cecal ligation and puncture (data not shown), suggesting that 129S6 mice are particularly resistant to adverse effects of inflammatory shock on cardiac function. Because C57BL/6 strains are known to be sensitive to endotoxin, we back-crossed the sGCα₁^−/−S6 mice onto a C57BL/6 background (sGCα₁^−/−B6). Interestingly, although sGCα₁^−/−S6 and sGCα₁^−/− mice on a mixed Swiss/129S6 background develop hypertension (6), we found that sGCα₁^−/−B6 mice are normotensive, which is consistent with another report describing only modestly elevated blood pressures in independently generated sGCα₁^−/− mice on the C57BL/6 background (37). Together, these data strongly suggest the existence of genetic modifiers of the hypertensive effects of sGCα₁ deficiency. Similarly, although increased cardiac contractility and E_a, as well as impaired ventricular relaxation, were observed in sGCα₁^−/−S6 mice (6), cardiac dysfunction was not evident in sGCα₁^−/−B6 mice at baseline. Importantly, although male sGCα₁^−/−S6 male mice, but not male sGCα₁^−/−B6 mice, develop hypertension, the blood pressure response to a low dose of an NO donor compound was similarly attenuated in sGCα₁^−/−S6 and sGCα₁^−/−B6 mice, suggesting that a difference in the hemodynamic response to NO does not underlie the different blood pressure phenotype. Furthermore, sGC enzyme activity, at baseline and activated by NO, was similarly attenuated in lung extracts from sGCα₁^−/−S6 and sGCα₁^−/−B6 mice compared with WT mice. Expression levels of sGCα₂ did not differ between WT and sGCα₁^−/− mice on the 129S6 or the C57BL/6 background, indicating that a compensatory increase in expression levels of sGCα₂ in sGCα₁^−/−B6 mice is not responsible for the absence of hypertension observed in sGCα₁^−/−B6 mice. Similar to observations in sGCα₁^−/−S6 mice (6), the effect of BAY 41-2272 on blood pressure was abolished in sGCα₁^−/−B6 mice. Together, these data demonstrate that sGCα₂ is not able to mediate BAY 41-2272-induced decreases in blood pressure in sGCα₁^−/− mice, implying that BAY 41-2272 is an isoform-specific sGCα₁β₁ stimulator in vivo in mice.

Myocardial depression is a well-recognized manifestation of organ dysfunction in sepsis (47). Therefore, we studied cardiac function in WT and sGCα₁^−/−B6 mice after endotoxin challenge. Both echocardiographic parameters (fractional shortening and ejection fraction) measured in awake mice and invasive hemodynamic parameters (dP/dt_max/IP, SW, and P_max) measured in anesthetized mice demonstrated that sGCα₁ deficiency exacerbates contractile dysfunction associated with endotoxin-induced shock. SV was decreased in sGCα₁^−/−B6, but not WT, mice 14 h after a challenge with endotoxin, as reflected by a substantial decrease in CO. Together, these findings indicate that sGCα₁β₁-derived cGMP has a protective effect on myocardial function in a setting of endotoxic shock. Interestingly, E_a, an index of arterial vascular load, increased in sGCα₁^−/−B6, but not WT, mice after a challenge with endotoxin. This finding is consistent with a previous report of an endotoxin-induced increase in E_a in NOS3-deficient, but not WT, mice (25). Together, these findings suggest that, during inflammatory shock in mice, systemic vasoconstriction occurs in an NO-cGMP-dependent manner. More research is warranted to determine whether a higher afterload in sGCα₁^−/−B6 than in WT mice contributes to or is a consequence of increased heart failure associated with inflammatory shock in sGCα₁^−/−B6 mice.

To test the hypothesis that the worsened cardiac dysfunction associated with inflammatory shock in sGCα₁^−/−B6 mice is mediated by increased inflammation in sGCα₁^−/−B6 mice compared with WT mice, we measured gene expression levels of two markers of inflammation: IL-6 and NOS2. LV expression levels of NOS2 and IL-6 did not differ between saline-treated WT and sGCα₁^−/−B6 mice. IL-6 expression levels were similarly induced in the LV of endotoxin-treated WT and sGCα₁^−/−B6 mice, suggesting that a difference in inflammatory response in the LV to endotoxin challenge is not responsible for the observed difference in cardiac phenotype. In contrast to IL-6 mRNA levels, cardiac NOS2 gene expression did not differ before or 14 h after endotoxin challenge in WT or sGCα₁^−/−B6 mice. These findings are consistent with a prior report that endotoxin challenge increases cardiac NOS2 gene expression only transiently, returning to baseline after 12 h (19). A challenge with endotoxin did not alter LV protein levels of sGCα₁, sGCα₂, or sGCβ₁ in WT or sGCα₁^−/−B6 mice. sGCβ₁ protein levels were lower in saline-and endotoxin-treated sGCα₁^−/−B6 than in WT mice. These findings are in accordance with previously published findings suggesting that loss of sGCα₁ results in decreased stability of sGCβ₁ (6, 37).

sGC is expressed in multiple cell types, including cardiomyocytes (55), vascular smooth muscle cells, endothelial cells (3), and neuronal cells (38). Any of the cell types expressing sGC might be involved in modulating myocardial function during inflammatory shock. Studies of the cellular mechanisms mediating sepsis-induced myocardial depression in the intact animal are complicated by the multiple interacting cell types involved, as well as by the potentially confounding influences of neurohumoral factors. We studied Ca²⁺ handling and contractility of isolated cardiomyocytes in vitro. Cardiomyocyte contractility and Ca²⁺ handling were similar in cardiomyocytes isolated from saline-challenged WT and sGCα₁^−/−B6 mice, recapitulating the in vivo hemodynamic data showing no difference in cardiac function between saline-challenged WT and sGCα₁^−/−B6 mice. Contractility was impaired to a greater extent in cardiomyocytes isolated from endotoxin-treated sGCα₁^−/−B6 than from endotoxin-treated WT mice. Similarly, although endotoxin decreased ΔCa_i in cardiomyocytes isolated from both genotypes, the reduction was greater in cardiomyocytes isolated from sGCα₁^−/−B6 than from WT mice. Together, these data show that cardiomyocytes reproduced the greater contractile deficit induced by endotoxin in sGCα₁^−/−B6 than in WT mice. We recently reported that endotoxin induces marked impairment of Ca²⁺ handling in cardiomyocytes and that this impairment is attenuated by increasing myocardial NOS3 levels (25), suggesting that NO protects cardiomyocytes from endotoxin-induced dysfunction. In the present study, the observation that endotoxin-induced cardiomyocyte dysfunction is exacerbated in sGCα₁^−/− mice compared with WT mice suggests that the ability of NO to prevent contractile dysfunction of murine cardiomyocytes may be, at least partially, dependent on sGCα₁β₁. Taken together, our results indicate that the more marked endotoxin-induced myocardial dysfunction in sGCα₁^−/−B6 than in WT mice may be attributable to greater impairment of cardiomyocyte function.

Importantly, we observed that more pronounced cardiac dysfunction in endotoxin-induced inflammatory shock in sGCα₁^−/−B6 than in WT mice was associated with decreased survival. To unequivocally link decreased survival in sGCα₁^−/−B6 mice with greater cardiac dysfunction, a study of mice with cardiac specific sGCα₁ deficiency is necessary. However, in patients with sepsis, cardiac dysfunction was identified to be an important predictor of mortality (41, 44, 45). Furthermore, increasing myocardial NO levels not only prevented myocardial dysfunction in murine models of inflammatory shock, but it also increased the survival rate (25). Our findings suggest that the ability of cardiac NO to protect against mortality in endotoxic shock may be dependent on sGCα₁β₁.

TNF is a pleiotropic cytokine that has a strong antitumor activity in vitro and in vivo. Systemic administration of TNF induces a systemic inflammatory response syndrome, characterized by bowel necrosis, liver damage, and severe hypotension leading to death. Previous studies suggested that TNF exerts its lethal effects via sGC activation: inhibition of sGC with MB or LY-83583 protected against TNF-induced lethality (11). However, similar to our finding in the endotoxin-induced shock model, mortality was more marked in sGCα₁^−/−B6 than in WT mice treated with TNF. Taken together, these findings demonstrate that cGMP generated by sGCα₁β₁ protects, at least partially, from myocardial dysfunction and death in endotoxin- and TNF-induced shock models.

In conclusion, inflammatory shock impaired Ca²⁺ handling and cell shortening to a greater extent in cardiomyocytes isolated from sGCα₁^−/−B6 than from WT mice. Ventricular dysfunction associated with inflammatory shock was more marked in sGCα₁^−/−B6 than in WT mice. Furthermore, mortality was higher in sGCα₁^−/−B6 than in WT mice in endotoxin-induced and TNF-induced inflammatory shock. These results identify a significant role for cGMP synthesized by sGCα₁β₁ in the modulation of ventricular function in a setting of inflammatory shock and identify activation of sGCα₁β₁ as a potential therapeutic strategy in septic shock.

GRANTS

This study was supported by a postdoctoral fellowship award of the American Heart Association Northeast Affiliate and a postdoctoral fellowship from the Massachusetts Biomedical Research Corporation (E. S. Buys), grants from the Fonds Wetenschappelijk Onderzoek-Vlaanderen (FWO-Vlaanderen) and the Geconcerteerde Onderzoeksacties (P. Brouckaert), and National Heart, Lung, and Blood Institute Grants HL-42397, HL-70896, and HL-71987. A. Cauwels is a postdoctoral fellow of the FWO Vlaanderen.