Absence of gp130 in dopamine β-hydroxylase-expressing neurons leads to autonomic imbalance and increased reperfusion arrhythmias

Abstract

Inflammatory cytokines that act through glycoprotein (gp)130 are elevated in the heart after myocardial infarction and in heart failure. These cytokines are potent regulators of neurotransmitter and neuropeptide production in sympathetic neurons but are also important for the survival of cardiac myocytes after damage to the heart. To examine the effect of gp130 cytokines on cardiac nerves, we used gp130^DBH-Cre/lox mice, which have a selective deletion of the gp130 cytokine receptor in neurons expressing dopamine β-hydroxylase (DBH). Basal sympathetic parameters, including norepinephrine (NE) content, tyrosine hydroxylase expression, NE transporter expression, and sympathetic innervation density, appeared normal in gp130^DBH-Cre/lox compared with wild-type mice. Likewise, basal cardiovascular parameters measured under isoflurane anesthesia were similar in both genotypes, including mean arterial pressure, left ventricular peak systolic pressure, dP/dt_max, and dP/dt_min. However, pharmacological interventions revealed an autonomic imbalance in gp130^DBH-Cre/lox mice that was correlated with an increased incidence of premature ventricular complexes after reperfusion. Stimulation of NE release with tyramine and infusion of the β-agonist dobutamine revealed blunted adrenergic transmission that correlated with decreased β-receptor expression in gp130^DBH-Cre/lox hearts. Due to the developmental expression of the DBH-Cre transgene in parasympathetic ganglia, gp130 was eliminated. Cholinergic transmission was impaired in gp130^DBH-Cre/lox hearts due to decreased parasympathetic drive, but tyrosine hydroxylase immunohistochemistry in the brain stem revealed that catecholaminergic nuclei appeared grossly normal. Thus, the apparently normal basal parameters in gp130^DBH-Cre/lox mice mask an autonomic imbalance that includes alterations in sympathetic and parasympathetic transmission.

the autonomic nervous system regulates cardiac function by balancing the actions of sympathetic and parasympathetic (vagal) inputs to the heart, which are the final common pathways for all cardiac autonomic control. Sympathetic neurons from the stellate ganglia stimulate heart rate (HR), cardiac conduction, and force of contraction through the release of norepinephrine (NE) and activation of β₁-adrenergic receptors (β₁ARs). Conversely, parasympathetic neurons inhibit HR through the release of acetylcholine (ACh) and activation of M₂ muscarinic ACh receptors. Autonomic transmission to the heart is tightly regulated by the central nervous system (5, 26) and by local interactions in the heart (22, 39, 41). The normal balance between cardiac autonomic neurons is disrupted by myocardial infarction, which leads to increased sympathetic and decreased parasympathetic transmission in the heart.

Inflammatory cytokines, including interleukin-6 (IL-6), cardiotrophin-1 (CT-1), leukemia inhibitory factor (LIF), and ciliary neurotrophic factor (CNTF), are elevated in the heart after myocardial infarction (1, 7, 11) and heart failure (7, 19). These cytokines, which share the glycoprotein (gp)130 receptor subunit (14, 35), are potent regulators of neurotransmitter and neuropeptide production in sympathetic neurons (4, 23, 25, 28). Indirect evidence has suggested that these cytokines may contribute to the imbalance between NE release and reuptake that develops after acute myocardial infarction and in heart failure (19, 24, 25). However, gp130 cytokines are also important survival and hypertrophy factors for cardiac myocytes, and disruption of gp130 in the heart is fatal (3, 15, 43, 44).

To selectively examine the effect of gp130 cytokines on neurons rather than myocytes, we used gp130^DBH-Cre/lox mice, which have a selective deletion of the gp130 cytokine receptor in neurons expressing dopamine β-hydroxylase (DBH) (34). We anticipated that basal neuronal and cardiovascular parameters would be normal in these mice but that postischemic remodeling of sympathetic neurons in the heart would be altered. We found that the absence of neuronal gp130 caused altered basal autonomic control of the heart and increased reperfusion arrhythmias. Parasympathetic transmission was decreased in addition to changes in sympathetic transmission.

METHODS

Materials.

Tyramine hydrochloride, carbachol, propranolol, atropine sulfate, and hexamethonium chloride were obtained from Sigma (St. Louis, MO). Dobutamine was obtained from Hospira (Lake Forest, IL). Antibodies were obtained as follows: rabbit anti-tyrosine hydroxylase (TH) was from Chemicon (Temecula, CA); rabbit anti-protein gene product (PGP)9.5 was from Accurate Chemicals (Westbury, NY); goat anti-rabbit Alexa fluor 488 was from Molecular Probes (Eugene, OR); rabbit anti-β₁AR was from Affinity Bioreagents (Golden, CO); rabbit anti-actin was from Sigma; and goat anti-rabbit horseradish peroxidase was from Pierce (Rockford, IL).

Animals.

Wild-type (WT) C57BL/6J mice were obtained from Jackson Laboratories (Sacramento, CA). gp130^DBH-Cre/lox [gp130 knockout (KO)] mice and ROSA26R mice were generated as previously described (33, 34). All mice were kept on a 12:12-h light-dark cycle with ad libitum access to food and water. Male and female mice (12–18 wk old) were used for all experiments. Animals from the two genotypes were age and gender matched for each experiment. All procedures were approved by the Institutional Animal Care and Use Committee and complied with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Pub. No. 85-23, Revised 1996).

Immunohistochemistry of ventricles.

Unoperated control animals were euthanized with an isoflurane overdose, and hearts were removed, fixed for 1 h in 4% paraformaldehyde, rinsed in PBS, cryoprotected in 30% sucrose overnight, and frozen in mounting media for sectioning. Transverse sections (10 μm) were thaw mounted onto charged slides. To reduce fixative-induced fluorescence, sections were treated with a 10 mg/ml solution of sodium borohydride in PBS (3 × 10 min). Slides were then rinsed in PBS and blocked in 2% BSA and 0.3% Triton X-100 in PBS for 1 h. Sections were incubated with rabbit anti-TH (1:300) overnight, rinsed 3 × 10 min in PBS, and then incubated for 1.5 h with goat anti-rabbit Alexa fluor 488 (1:300). Sections were rinsed 3 × 10 min in PBS, incubated for 30 min in CuSO₄ in 50 mM ammonium acetate to reduce background, rinsed 3 × 10 min in PBS, coverslipped, and then visualized by fluorescence microscopy.

For innervation density analysis, photos were obtained using consistent camera settings from WT and KO sections that had been processed together. Identical areas of the heart were analyzed for each mouse: two regions of the right ventricle, one of the septum, and three regions of the left ventricle (LV) were quantified from each section by threshold discrimination using ImageJ software. Photos for analysis were obtained from 3 sections/heart, with each section separated from the others by at least 250 μm. Sections were quantified from five mice of each genotype.

Immunohistochemistry of the brain stem.

Unoperated control mice were overdosed with pentobarbital sodium (200 mg/kg) and perfused transcardially with heparinized saline (1,000 U/ml) followed by 10 ml of 3.8% acrolein in 2% paraformaldehyde and then followed by 50 ml of 2% paraformaldehyde (in 0.1 M phosphate buffer). Brains were extracted, and the brain stem was blocked and placed into 2% paraformaldehyde for 30 min. Brain stems were sectioned (40 μm) on a vibrating microtome and collected into 0.1 M phosphate buffer. Free floating coronal sections through the medulla were processed for the immunoperoxidase localization of TH. Tissue sections were sequentially incubated in 1% sodium borohydride in 0.1 M phosphate and 0.5% BSA in 0.1 M Tris-saline for 30 min before an incubation with polyclonal rabbit anti-TH antibody (1:1,000, Chemicon) in 0.1% BSA and 0.3% Triton X-100 in 0.1 M Tris-saline for 40 h at 4°C. The bound antibody was detected with a 30-min incubation in biotinylated goat anti-rabbit IgG (1:400, Vector Laboratories, Burlingame, CA) in 0.1% BSA in 0.1 M Tris-saline. Between incubations, sections were rinsed with either 0.1 M Tris-saline or 0.1 M phosphate buffer. All incubations were carried out on a shaker table. Sections were then mounted on gelatin-coated slides and coverslipped with DPX mounting medium (Sigma).

Slides were blinded to the observer, and images were captured with an Olympus DP71 digital camera mounted to an Olympus BX51 microscope. Images of the A1, C1, and subpostremal region of the nucleus tractus solitarius (NTS) were captured that best matched each area between mice. TH-immunoreactive cells were counted in a single section for each area from each mouse. Independent t-tests were performed in SigmaStat to determine if the number of TH-immunoreactive cells in the A1, C1, or dorsal vagal complex (subpostremal NTS and area postrema) differed between WT C57BL/6J and gp130^DBH-Cre/lox mice.

HPLC analysis of NE.

NE levels were measured in unoperated control animals by HPLC with electrochemical detection as previously described (24). Heart tissue was homogenized in perchloric acid (0.1 M) containing 1.0 μM of the internal standard dihydroxybenzylamine to correct for sample recovery. Catecholamines were purified by alumina extraction before analysis by HPLC. Detection limits were ∼0.05 pmol with recoveries from the alumina extraction of >60%.

Immunoblot analysis.

TH and PGP9.5 levels in unoperated control LVs were quantified via Western blot analysis as previously described (24). Tissue was homogenized in lysis buffer, size fractionated by SDS-PAGE, and transferred onto membranes for blotting. Samples were run in duplicate. Each blot was incubated with rabbit polyclonal anti-TH (1:1,000) followed by an incubation with goat anti-rabbit horseradish peroxidase (1:5,000). Immunoreactive bands were visualized by chemiluminescence using SuperSignal Dura (Pierce), and band intensity was recorded by a −40°C charge-coupled device camera. Blots were stripped and reblotted with polyclonal rabbit anti-PGP9.5 (1:1,000) followed by an incubation with goat anti-rabbit horseradish peroxidase (1:5,000). β₁AR levels were determined by incubating the blots with rabbit anti-β₁AR (1:2,000) normalized to rabbit anti-actin (1:2,000). Data were analyzed using LabWorks software (UVP, Upland, CA).

Real-time PCR.

Stellate ganglia were harvested from unoperated control mice and stored immediately in RNAlater. RNA was isolated from individual stellate ganglia using the Ambion RNAqueous micro kit. Total RNA was treated with DNase and quantified by optical densiometry at 260 nm, and 100 ng of total RNA were then reverse transcribed. Real-time PCR was performed using the ABI TaqMan Universal PCR Master Mix in the ABI 7500. Samples were assayed using prevalidated Taqman gene expression assays for mouse TH, NE transporter, and GAPDH (as a normalization control). For the PCR amplification, 2 μl of reverse transcription reactions were used in a 20-μl reaction. Each sample was assayed in duplicate. Standard curves were generated with known amounts of RNA from control tissue ranging from 0.8 to 100 ng. Values for TH and NE transporter were normalized to the internal control (GAPDH) from the same sample.

Hemodyamics.

Unoperated control mice were anesthetized with 4% isoflurane and maintained with 2–3% isoflurane. Mice were intubated and placed on a rodent ventilator. Body temperature was monitored and maintained at 37 ± 0.2°C. A microtipped pressure transducer (1.0-Fr, Millar) was inserted into the right carotid artery, and arterial pressure data were collected for 3–5 min using a PowerLab data-acquisition system. The pressure transducer was advanced into the LV for measurements of LV pressure. A small polyvinyl catheter was placed in the left jugular vein for drug administration. When the animal was stable, it was given a bolus dose of hexamethonium chloride (5 mg/kg in 20 μl) to abolish ganglionic transmission, which decreased all cardiovascular parameters in both genotypes and prevented reflex bradycardia after tyramine administration. After a new baseline had been established after ganglionic block, a bolus dose of tyramine hydrochloride was administered (200 μg/kg in 20 μl) to assess the cardiac response to the release of endogenous NE. Saline control injections confirmed that the 20-μl volume bolus dose did not trigger changes in arterial pressure or HR. In another set of mice, ganglionic block was established with hexamethonium, and increasing doses (4, 8, 16, 32, and 64 μg/kg, volumes: 5–20 μl) of the β-agonist dobutamine were then administered. The next dose was injected as soon as the increasing HR leveled off. LV peak systolic pressure (LVPSP), dP/dt_max, and dP/dt_min were analyzed using ChartPro software.

HR analysis.

Unoperated control mice were anesthetized with 4% isoflurane and maintained on a nose cone with 2% isoflurane. Body temperature was monitored and maintained at 37 ± 0.2°C, and ECG leads were placed to monitor HR (positive lead left leg, negative lead right arm, and ground right leg). The animal's head was covered with a black cloth to prevent light stimulation of sympathetic reflexes. When body temperature and HR were stable, animals were injected with either atropine sulfate (2 mg/kg), propranolol hydrochloride (4 mg/kg), or carbachol (50 μg/kg) intraperitoneally, and HR was monitored for 15 min. Before carbachol injections, mice were given hexamethonium chloride (5 mg/kg ip) to prevent reflex tachycardia. HRs were recorded and analyzed using ChartPro software.

Vagal nerve stimulation.

Unoperated control mice were anesthetized with 4% isoflurane and maintained on a nose cone with 2% isoflurane. Body temperature was monitored and maintained at 37 ± 0.2°C, and ECG leads were placed to monitor HR (positive lead left leg, negative lead right arm, and ground right leg). The right vagus nerve was isolated at the cervical level and placed across stainless steel electrodes. The uncut nerve was stimulated at 5 V and 5 Hz with a pulse width of 1 ms for 5 s (Grass S88X stimulator). HRs were recorded and analyzed using ChartPro software. This relatively low frequency of nerve stimulation was used because higher frequencies caused irregular heart rhythms and/or death in all of the gp130 KO mice tested. Although the stimulation of uncut vagal afferents may alter the activation of sympathetic efferents, the injection of atropine (2 mg/kg ip) abolished nerve-induced bradycardia in both gentoypes (data not shown).

β-Galactosidase staining.

Tissue from unoperated mice was fixed in 0.4% glutaraldehyde for 2–4 h. Staining was carried out overnight in a solution containing 0.1% sodium deoxycholate, 0.2% Nonidet P-40, 5 mM potassium ferrocyanide, 5 mM potassium ferricyanide, and 1 mg/ml X-Gal.

Statistics.

Student's t-tests were used for a single comparison between two groups (e.g., WT vs. gp130 KO). Two-way ANOVA with a Bonferroni posttest was used to compare across genotypes and treatment groups (e.g., WT and gp130 KO with or without drug treatment). Statistics were calculated using Prism 5.0 or SigmaStat software.

Myocardial ischemia-reperfusion.

Ischemia-reperfusion was carried out as previously described (8). Adult mice were placed in an induction chamber and anesthetized with 4% isoflurane. Mice were intubated, mechanically ventilated, and maintained with 1–2% isoflurane mixed with 100% oxygen. Core body temperature was maintained at ∼37°C, and a three-lead ECG (positive lead left leg, negative lead right arm, and ground right leg) was monitored throughout the surgery and for 15 min after reperfusion using a PowerLab data-acquisition system (ADInstruments).

The mouse was turned to a right lateral decubitus position, and a thoracotomy performed in the fourth intercostal space with the aid of a dissecting microscope. A ligature (8-0 nylon or the equivalent on a taper needle) was placed around a proximal segment of the left anterior descending coronary artery. The ends of the suture were passed through a tube (polyethylene-10) with a blunted end to prevent tissue damage. The ligature was tightened to induce regional myocardial ischemia, which was confirmed by ECG changes, regional cyanosis, and wall motion abnormalities. After 45 min, the coronary ligature was released, and reperfusion was confirmed by visible epicardial hyperemia.

ECG analysis.

ECG data collected during the 15 min after reperfusion were analyzed for arrhythmias. Ventricular arrhythmias were analyzed according to the Lambeth Convention guidelines for the analysis of arrhythmias in animal models of ischemia and reperfusion (38). Premature ventricular complexes (PVCs) were defined as a single premature QRS complex in relation to the P wave, whereas ventricular tachycardia (VT) was defined as a run of four or more PVCs (38). This definition does not take into account the R-R interval. However, all mice who exhibited a series of four or more PVCs had at least one series of PVCs with constant R-R intervals that met a more stringent definition of VT. We quantified the number of animals in each group that had at least one episode of VT after reperfusion, and we noted the duration of VT. One KO mouse developed VT during the occlusion and was treated with lidocaine to resolve the arrhythmia. That mouse was excluded from the reperfusion arrhythmia analysis due to the drug treatment.

RESULTS

The gp130 cytokine receptor was deleted in DBH-expressing neurons so that we could examine the effect of cytokines on sympathetic neurons. Basal sympathetic parameters appeared normal in neuronal gp130 KO mice. TH is the rate-limiting enzyme in NE synthesis, and TH mRNA and protein levels were normal in sympathetic neurons from gp130^DBH-Cre/lox (KO) mice compared with WT controls (Fig. 1). Likewise, mRNA encoding the NE transporter was normal in cardiac sympathetic neurons from gp130 KO mice, and LV NE content was identical compared with WT mice (Fig. 1).

Fig. 1.

Basal noradrenergic parameters were normal in neuronal glycoprotein (gp)130 knockout (KO) mice. WT, wild-type mice. A and B: tyrosine hydroxylase (TH; A) and norepinephrine (NE) transporter (NET; B) mRNA were quantified in cardiac sympathetic neurons and normalized to GAPDH mRNA in the same sample (means ± SE; n = 4). C: TH protein in the left ventricle (LV) normalized to the pan-neuronal marker protein gene product (PGP)9.5 (means ± SE; n = 6). D: NE content (in pmol/mg tissue) in the LV (means ± SE; n = 6).

Consistent with these molecular and chemical parameters, basal cardiovascular parameters were also normal in neuronal gp130 KO mice. Mean arterial pressure was identical in gp130 KO mice compared with age-matched WT mice (Fig. 2). LVPSP, dP/dt_max, and dP/dt_min were also similar in both genotypes (Fig. 2), suggesting that the lack of gp130 in noradrenergic neurons did not alter basal autonomic function.

Fig. 2.

Basal cardiovascular parameters were normal in neuronal gp130 KO mice. A: mean arterial pressure (MAP) was similar in both genotypes (means ± SE; n = 10). B: LV peak systolic pressure (LVP) was similar in both genotypes (means ± SE; n = 10–11). C and D: ventricular function, as measured by dP/dt_max (C) and dP/dt_min (D), was similar in both genotypes (means ± SE; n = 10–11).

Ischemia-reperfusion surgery, however, revealed unexpected differences between gp130 KO mice and WT controls. Over 90% of all mice that underwent ischemia-reperfusion surgery exhibited a change in the ECG characterized by widening of the QRS complex beginning ∼30 s after reperfusion (Fig. 3B). In addition to that common alteration in the ECG, 1 WT mouse of 45 total WT mice (2%) developed VT after reperfusion. In contrast, 13 of 52 (25%) gp130 KO mice developed VT after reperfusion, with several mice exhibiting VT that lasted for >1 min. The morphology and rate of PVCs differed between animals and across time within the same animal. Several examples are shown in Fig. 3, C–F. The cardiac innervation can contribute to ventricular arrhythmias, so this led us to examine autonomic control of the heart more carefully in gp130 KO mice.

Fig. 3.

Reperfusion ventricular tachycardia (VT) in gp130 KO mice. A: sinus rhythm in a mouse before left anterior descending coronary artery occlusion. B: ECG from the same mouse as in A after reperfusion. Note the widening of the QRS complex, which was present in >90% of mice from both genotypes after reperfusion. Scale bar = 0.25 s. C–F: examples of premature ventricular complexes after reperfusion from four different gp130 KO mice.

Since an altered pattern of sympathetic innervation results in rhythm instability in the heart (17), we examined the pattern and density of sympathetic innervation of the ventricles using immunohistochemistry for TH. The pattern and density of TH-immunoreactive fibers in gp130 KO hearts was identical to WT hearts (Fig. 4). Both right ventricles and LVs had sympathetic innervation that appeared normal.

Fig. 4.

Sympathetic innervation density was normal in neuronal gp130 KO hearts. A and B: TH-positive sympathetic nerve fibers in WT (A) and gp130 KO (B) LVs. Scale bar = 100 μm. C and D: sympathetic innervation density in the right ventricle (C) and LV (D) of WT and neuronal gp130 KO mice (means ± SE; n = 5).

We further examined sympathetic transmission in the heart using pharmacological antagonists and agonists. The basal HR was identical in both genotypes (Fig. 5A), as was the drop in HR after the administration of the β-antagonist propranolol (Fig. 5B). In contrast, stimulation of NE release with tyramine generated a smaller HR increase in gp130 KO mice (Fig. 5C). Tyramine also stimulated a smaller increase in LVPSP (Fig. 5D), whereas changes in dP/dt_max and dP/dt_min trended lower but were not significantly different (data not shown). NE content was normal in gp130 KO mice, so we examined β-receptor responsiveness with the agonist dobutamine. A low dose of dobutamine stimulated significantly smaller effects in gp130 KO hearts compared with WT hearts (Fig. 6A), but higher doses generated similar responses in both genotypes (Fig. 6, B and C). β₁AR levels were lower in gp130 KO ventricles than WT ventricles (Fig. 6D), consistent with the differences in responsiveness to tyramine and dobutamine.

Fig. 5.

gp130 KO hearts exhibited an impaired response to tyramine-induced NE release. A and B: heart rates [HRs; in beats/min (bpm)] were monitored in WT and neuronal gp130 KO mice. A: basal HRs were identical in both genotypes (means ± SE; n = 10–11). B: propranolol blockade of β-receptors decreased HRs to a similar extent in both genotypes (means ± SE; n = 10–11). C and D: tyramine infusion stimulated NE release from sympathetic terminals and increased HRs (C) and LVPs (D) in WT and neuronal gp130 KO mice. The increases in HR and LVP were significantly smaller in gp130 mice (means ± SE; n = 5). *P < 0.05.

Fig. 6.

gp130 KO hearts exhibited an impaired response to β-agonists and decreased β₁-adrenergic receptor (β₁AR) levels. A: a subsaturating dose (4 μg/kg) of the β-agonist dobutamine stimulated a smaller increase in dP/dt_max in gp130 KO hearts than in WT hearts (means ± SE; n = 5–6). *P < 0.05. B and C: full dose-response curves showed that higher doses of dobutamine stimulated identical increases in dP/dt_max (B) and HR (C) in both genotypes. D: β₁AR and actin levels were quantified by Western blot analysis in WT and gp130 KO LVs (means ± SE; n = 6). *P < 0.05.

DBH-Cre driver mice were generated to selectively delete genes in noradrenergic neurons. However, DBH is expressed during development in some postganglionic parasympathetic neurons that are derived from the neural crest (21), raising the possibility that cardiac parasympathetic neurons might be affected. We crossed DBH-Cre driver mice with ROSA reporter mice and discovered that recombination took place in cardiac parasympathetic neurons (Fig. 7A), leading to the deletion of gp130 in cardiac ganglion cells. In contrast, dorsal root ganglia sensory neurons do not undergo recombination, and substance P gene expression is identical in WT and gp130 KO mice (data not shown).

Fig. 7.

Recombination in cardiac parasympathetic ganglia; vagal transmission is altered in gp130 KO mice. A and B: DBH-Cre mice (A) or control mice (B) were crossed with ROSA reporter mice to identify cells in which DBH-Cre is expressed. β-Galactosidase staining in the postnatal day 4 heart revealed recombination in DBH-Cre cardiac parasympathetic ganglia but not in control ganglia. Blue sympathetic axons projecting to the ventricle are also visible in A. C: atropine blockade of muscarinic receptors produced a significantly smaller tachycardia in gp130 KO mice (means ± SE; n = 9). ***P < 0.001. D: hexamethonium blockade of ganglionic transmission lowered HRs in both genotypes. gp130 KO mice exhibited a smaller bradycardia, but the difference was not significant (means ± SE; n = 10–11; P = 0.053). E: vagal nerve stimulation triggered significantly greater bradycardia in gp130 mice than in WT mice (means ± SE; n = 5). **P < 0.01. F: the muscarinic agonist carbachol stimulated similar bradycardia in both genotypes (means ± SE; n = 5).

To determine if the absence of gp130 altered parasympathetic transmission in the heart, we inhibited cholinergic transmission with the muscarinic antagonist atropine. Atropine generated a significantly smaller rise in HR in gp130 KO mice, suggesting that parasympathetic transmission was blunted (Fig. 7C). To determine if the postganglionic nerves were impaired, we electrically stimulated the vagus nerve and measured the HR response. Vagal nerve stimulation provoked a significantly larger bradycardia in gp130 KO mice than in control mice (Fig. 7E). An injection of atropine (2 mg/kg) abolished nerve-stimulated bradycardia in both gentoypes (data not shown). The enhanced vagal bradycardia in gp130KO mice was not due to altered cardiac responsiveness, because the muscarinic agonist carbachol stimulated a similar bradycardia in both genotypes (Fig. 7F).

The combination of an enhanced response to vagal stimulation and impaired response to atropine blockade suggested that cardiac parasympathetic neurons were not receiving adequate stimulation from the brain stem. Cytokines acting through gp130 promote proliferation and neuronal differentiation during central nervous system development (14, 37), raising the possibility that the development of noradrenergic neurons in the brain stem was disrupted in gp130 KO mice. We stained brain stem sections for TH to identify catecholaminergic cells. We found that the catecholaminergic nuclei in the A1, C1, and NTS appeared normal in gp130 KO mice (Fig. 8), and the number of TH-immunoreactive cells in these three regions were indistinguishable from those in WT mice (bilateral NTS: WT 109.2 ± 17.2 and KO 134.2 ± 10.7; unilateral A1: WT 10.0 ± 1.4 and KO 10.8 ± 0.2; and unilateral C1: WT 14.8 ± 1.6 and KO 13.8 ± 1.5, means ± SE, n = 4). Thus, the altered parasympathetic outflow was not due to the absence of a catecholaminergic cell group in the gp130 KO brain stem.

Fig. 8.

Peroxidase staining of TH in the medulla of WT and gp130 KO mice resulted in comparable patterns and numbers of labeled cells. A: representative TH immunoreactivity in the medulla of a WT mouse. Catecholamine-labeled cells were seen in the nucleus tractus solitarius (NTS) and area postrema (AP) and additionally in ventrolateral regions such as the A1 and C1. The dashed box is shown at higher magnification in B. Scale bar = 500 μm. B: catecholamine-positive cells are apparent in the AP as well as throughout the NTS. C: the pattern of labeling seen in the AP and NTS of this gp130 KO mouse was very similar to the pattern and number of labeled cells seen in the WT mouse in B. Scale bars = 250 μm in B and C.

DISCUSSION

We began experiments with gp130^DBH-Cre/lox mice to test the role of inflammatory cytokines in postischemia neuronal plasticity and remodeling. We expected that basal autonomic parameters would be normal in these mice but that the postinfarct regulation of neurotransmitter and neuropeptide expression would be altered in the days after myocardial infarction. However, neuronal gp130 KO mice had a high rate of developing severe arrhythmias including polymorphic PVCs during reperfusion, suggesting that the lack of neuronal gp130 during development made them susceptible to reperfusion-induced VT. We discovered that the lack of neuronal gp130 resulted in an autonomic imbalance that was masked by compensatory adaptations.

The gp130 receptor mediates the actions of a family of cytokines that have significant effects on both cardiac myocytes and sympathetic neurons (3, 4, 15, 23, 25, 28, 43, 44). gp130 cytokines are critical survival and hypertrophy factors for cardiac myocytes (31, 40), and the lack of gp130 during development leads to severe ventricular hypoplasia. The lack of gp130 in the adult heart leads to myocyte apoptosis, heart failure, and death after pressure overload (15). IL-6, LIF, CT-1, and CNTF are all produced in the heart and have been implicated in cardiac remodeling in myocardial infarction and heart failure (1, 7, 11, 19). These cytokines are also potent regulators of neurotransmitter and neuropeptide production in sympathetic neurons (4, 23, 25, 28) and play a critical role in the regulation of neurotransmitters and neuropeptides after nerve injury (12). We hypothesized that cytokines in the heart stimulated changes in sympathetic transmitter and peptide expression after myocardial infarction and used the neuronal gp130 KO mice to test that hypothesis. Our data revealed that these cytokines are important during autonomic development in addition to their potential role after nerve injury. Future studies using an inducible DBH-Cre transgene will allow us to distinguish developmental effects from cytokine actions in the adult.

The deletion of gp130 in DBH-positive neurons disrupts the development of the cholinergic phenotype in neurons innervating sweat glands (34), but the expression of noradrenergic markers and neuropeptides is normal in sympathetic ganglia from neuronal gp130 KO mice (12). Altered regulation of neuropeptide and transmitter expression only becomes apparent in neuronal gp130 KO mice after nerve injury, when cytokine signaling and STAT3 phosphorylation is absent, and the peptide expression profiles differ significantly from those of WT mice (12). Thus, we did not expect to observe a cardiac sympathetic phenotype in control gp130 KO mice. Basal sympathetic and cardiovascular parameters appeared normal in gp130 KO mice, but pharmacological experiments identified an underlying autonomic imbalance.

The autonomic imbalance in gp130 KO mice was most clearly revealed by the administration of atropine to block parasympathetic transmission in the heart. The HR response to atropine in gp130 KO mice was half that of WT mice, consistent with decreased ACh release. The muscarinic agonist carbachol triggered similar bradycardia in both genotypes, confirming that parasympathetic transmission was altered rather than cardiac responsiveness. DBH is expressed during development in some neural crest-derived parasympathetic neurons (21), and DBH expression has been recently identified in cholinergic adult cardiac ganglion cells (16). We confirmed that DBH-Cre-induced recombination occured in cardiac parasympathetic neurons. Cytokines stimulate the expression of cholinergic genes in sympathetic neurons (30, 34, 42), and we suspected that the lack of gp130 in cardiac parasympathetic neurons resulted in decreased ACh production in those cells. However, vagal stimulation provoked enhanced bradycardia in gp130 KO hearts, suggesting that ACh synthesis was elevated in the cardiac ganglion and that the parasympathetic drive from the brain stem was impaired.

Cytokines stimulate the proliferation of precursor cells and the differentiation of some classes of neurons in the central nervous system (14, 37). Catecholaminergic nuclei in the brain stem are involved in autonomic control of the cardiovascular and other systems (9, 10, 29), and disruption of the proliferation or differentiation of those cells might account for the altered parasympathetic outflow. TH immunohistochemistry in the brain stem confirmed the presence of a normal number of catecholaminergic cells in the NTS, A1, and C1 but did not rule out functional deficits in those cells and/or their circuits.

In contrast to the deficits in parasympathetic transmission revealed by atropine, the β-antagonist propranolol revealed seemingly normal noradrenergic transmission. However, further analysis suggested subtle changes in sympathetic transmission that were masked by a series of compensatory changes. For example, forcing the release of NE from sympathetic nerve terminals with tyramine provoked a smaller cardiac response in gp130 KO mice compared with WT mice, as did low doses of the β-agonist dobutamine. The lack of responsiveness seemed to be due to decreased β₁AR levels in the hearts of gp130 KO mice. Decreased receptor expression is consistent with the small shift observed in dobutamine dose-response curves, which revealed that higher doses of agonist stimulated identical changes in both genotypes. However, low β₁AR levels and decreased responsiveness are unexpected in mice that develop severe ventricular arrhythmias at a higher than normal rate, since β₁AR antagonists have been protective in animal studies and in humans who have undergone cardiac ischemia and reperfusion (2, 6, 13, 27). These data raise two additional questions: Why are the basal HR and cardiac function normal if β₁AR levels are low in gp130 KO hearts while the NE content is normal? and Why does the β-receptor antagonist propranolol cause the same drop in HR in both genotypes? One possible explanation for all of these surprising results is that although NE content is normal in gp130 KO mice, NE release is elevated. This would contribute to arrhythmias, would compensate for the low β₁ receptors under basal conditions, and might even cause the decrease in β₁AR expression. Additional studies of NE turnover are ongoing to determine if NE release is altered in gp130 KO mice.

In summary, autonomic and cardiovascular function in mice lacking gp130 in DBH-positive neurons appear normal under basal conditions. However, perturbing the system with surgery or with pharmacological interventions reveals significant changes in autonomic transmission, including impaired cardiac parasympathetic transmission. Decreased parasympathetic activity in the heart correlates with an increased risk of arrhythmia and death in humans (18, 20, 32, 36). It is particularly interesting that neuronal gp130 KO mice, which exhibit decreased parasympathetic tone, have increased susceptibility to severe reperfusion arrhythmias.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants HL-068231 (to B. A. Habecker) and HL-056301 (to S. A. Aicher).