Vago-Splenic Axis in Signal Transduction of Remote Ischemic Preconditioning in Pigs and Rats

Abstract

Rationale:

The signal transduction of remote ischemic conditioning is still largely unknown.

Objective:

Characterization of neurohumoral signal transfer and vago-splenic axis in remote ischemic preconditioning (RIPC).

Methods and Results:

Anesthetized pigs were subjected to 60 minutes of coronary occlusion and 180 minutes of reperfusion (placebo+ischemia/reperfusion [PLA+I/R]). RIPC was induced by 4×5/5 minutes of hindlimb I/R 90 minutes before coronary occlusion (RIPC+I/R). Arterial blood samples were taken after placebo or RIPC before I/R. In subgroups of pigs, bilateral cervical vagotomy, splenectomy, or splenic denervation were performed before PLA+I/R or RIPC+I/R, respectively. In pigs with RIPC+I/R, infarct size (percentage of area at risk) was less than in those with PLA+I/R (23±12% versus 45±8%); splenectomy or splenic denervation abrogated (splenectomy+RIPC+I/R: 38±15%; splenic denervation+RIPC+I/R: 43±5%), and vagotomy attenuated (vagotomy+RIPC+I/R: 36±11%) RIPC protection. RIPC increased phosphorylation of STAT3 (signal transducer and activator of transcription 3) in left ventricular biopsies taken at early reperfusion. Splenectomy or splenic denervation, but not vagotomy, abolished this increased phosphorylation. In rats with vagotomy, splenectomy, or splenic denervation, RIPC (3×5/5 minutes of hindlimb occlusion/reperfusion) or placebo was performed, respectively. Hearts were isolated, saline perfused, and subjected to 30/120-minute global I/R. With RIPC, infarct size (percentage of ventricular mass) was less (20±7%) than with placebo (37±6%), and vagotomy, splenectomy, or splenic denervation abrogated RIPC protection (38±12%, 36±9%, and 36±7%), respectively. Rat spleens were isolated, saline perfused, and splenic effluate (SEff) was sampled after infusion with carbachol (SEff_carbachol) or saline (SEff_saline). Pig plasma or SEff was infused into isolated perfused rat hearts subjected to global I/R. Infarct size was less with infusion of RIPC+I/R_plasma+ (24±6%) than with PLA+I/R_plasma (40±8%), vagotomy+PLA+I/R_plasma (39±11%), splenectomy+PLA+I/R_plasma (35±8%), vagotomy+RIPC+I/R_plasma (40±9%), splenectomy+RIPC+I/R_plasma (33±9%), or splenic denervation+RIPC+I/R_plasma (39±8%), respectively. With infusion of SEff_carbachol, infarct size was less than with infusion of SEff_saline (24 [19–27]% versus 35 [32–38]%).

Conclusions:

Activation of a vago-splenic axis is causally involved in RIPC cardioprotection.

Timely reperfusion is the only way to salvage ischemic myocardium from infarction and to limit infarct size (IS). However, reperfusion per se adds an additional irreversible component to ischemic damage and thus contributes to final IS.¹ Despite substantial improvements in prevention and therapy of acute myocardial infarction during the past decades, the incidence of myocardial infarction in an increasingly aging population is not declining, and many survivors of acute myocardial infarction develop heart failure.² Therefore, treatment of myocardial ischemia/reperfusion (I/R) injury remains a major unmet medical need, and adjunct cardioprotective strategies are required.³

Brief episodes of myocardial I/R, which do not cause irreversible injury per se, protect the myocardium from consequences of I/R injury by activation of molecular self-defense mechanisms.⁴ Such protection by ischemic conditioning can not only be induced locally in the heart but also in tissues remote from the heart (remote ischemic conditioning). Remote ischemic conditioning can be induced prior (pre)⁵ or during (per)⁶ ongoing myocardial ischemia or during early reperfusion (postconditioning).^7–9 Although cardioprotection by remote ischemic conditioning is operative in all species tested to date, including humans,¹⁰ the translation from animal experiments to clinical practice remains challenging and has been disappointing to date.¹¹ The best available evidence for the efficacy of cardioprotective maneuvers exists for remote ischemic conditioning.^8,10,12 In fact, there is one recent study with improved clinical outcome as the primary end point in patients with reperfused ST-segment-elevation myocardial infarction when undergoing remote ischemic perconditioning.^13,14

The signal transfer of remote ischemic conditioning from the stimulus site to the target organ and the signal transduction within the myocardium are still unclear in their details.⁹ Remote ischemic conditioning is a systemic response: it can be induced from different tissues and organs, and protection is not only operative in the heart but also in other organs.⁹ Both neuronal and humoral pathways are involved in the signal transfer from the periphery to the heart.^9,15,16 Vagal nerves are indispensable for the transfer of cardioprotection by remote ischemic conditioning in rats and rabbits.^17,18 In pigs, electrical vagal nerve stimulation attenuates myocardial I/R injury, indicating that vagal activation induces cardioprotection also in larger animals.^19,20 Also in patients with acute ST-segment-elevation myocardial infarction, low-level electrical vagal nerve stimulation at the right tragus reduces myocardial damage, as reflected by reduced release of CK-MB (creatine kinase muscle brain) and myoglobin during 72 hours.^21,22 Humoral communication in remote ischemic conditioning has been evidenced through transfer of the cardioprotective signal with blood plasma from one heart to another, even across species.^23–26 Neuronal and humoral pathways interact.^9,15,16 In rats, activation of vagal nerves induces the release of GLP-1 (glucagon-like peptide-1) from the gut, and a GLP-1 agonist reduces IS in isolated perfused rat hearts.¹⁸ However, there is no evidence for GLP-1 receptors on cardiomyocytes²⁷ or activation of cardiac vagal nerves by GLP-1.²⁸ Also, the GLP-1 agonist exenatide did not reduce myocardial IS in patients with ST-segment-elevation myocardial infarction when given before reperfusion.²⁹

Vagal nerve stimulation decreases the formation of proinflammatory cytokines by activation of a spleen-mediated neurohumoral mechanism.^30,31 Anti-inflammatory cytokines are released into the circulation in response to remote ischemic conditioning.⁹ However, the precise source of such humoral mediators is unclear,⁹ and the spleen as a source of cardioprotective substances has to date not been considered in the context of remote ischemic conditioning. Indeed, the spleen is a central relay organ of systemic immune responses, and the spleen is under control of the autonomic nervous system.³² Interestingly, a cardioprotective effect of splenic nerve stimulation through release of humoral factors was already demonstrated in the 1950s^33,34 but then not further pursued. Recently, the spleen has been implicated in neuroprotection against stroke by remote ischemic preconditioning (RIPC) in rats because prior splenectomy abrogated the protection, which was associated with an altered profile of circulating lymphocytes.³⁵

We have now used our established anesthetized pig model to characterize a vago-splenic axis for the cardioprotection by RIPC, without and with vagotomy, splenectomy, or splenic denervation. For subsequent transfer experiments to isolated, saline-perfused rat hearts, the vago-splenic axis was confirmed in rats. Plasma from pigs, which underwent vagotomy, splenectomy, or splenic denervation and subsequent RIPC, respectively, was then transferred to isolated perfused rat hearts with the aim to determine the role of humoral factor(s) in the vago-splenic axis. Finally, we established an isolated perfused rat spleen model to investigate the release of cardioprotective mediators. The activation of the STAT3 (signal transducer and activator of transcription 3)^25,26 by remote ischemic conditioning was used as readout to estimate the impact of the vago-splenic axis on RIPC intracellular signal transduction in the myocardium.

Methods

The authors declare that all supporting data of this exploratory study are available in the article and its Online Data Supplement.

Experiments in pigs and rats were performed between August 2014 and August 2018.

The experimental protocols were approved by the Bioethical Committee of the district of Düsseldorf (G1413/14, G1625/17, and G1655/18) and conform to the Position of the American Heart Association on Research Animal Use, adopted on November 11, 1984. Unless otherwise specified, materials were obtained from Sigma-Aldrich (Deisenhofen, Germany). Our study did not focus on sex differences, and myocardial I/R injury is more severe in males. Therefore, only male animals were used in the present study.

The experimental protocols of pigs in situ and of isolated perfused rat hearts and the methods for measurement of hemodynamics and regional blood flow, quantification of IS, and protein expression were standard,^36,37 as described previously.^25,26,38,39 For details see, Online Data Supplement.

Pigs in Situ

Experimental Preparation

In anesthetized Göttingen minipigs, a silicon-coated suture was placed around the left anterior descending coronary artery distal to its second diagonal branch for later coronary occlusion. In subgroups of pigs, either bilateral cervical vagotomy, splenectomy, or splenic denervation, respectively were performed before RIPC or a respective placebo maneuver (Figure 1). Vagotomy was completed within ≤2 minutes, splenectomy within <10 minutes, and splenic denervation within <30 minutes, respectively, and placebo and RIPC protocols were time matched for these interventions.

Figure 1. Experimental protocols.

I/R indicates ischemia/reperfusion; PLA, placebo; RIPC, remote ischemic preconditioning; SD, splenic denervation; SE, splenectomy; and V, vagotomy.

Vagotomy

The left and right cervical vagal nerves were exposed through a midline cervical incision. After their transection, an immediate rise in heart rate was taken to indicate successful vagal denervation.

Splenectomy

The abdominal cavity was opened through a left subcostal incision. Splenic ligaments were dissected and the short gastric arteries and veins ligated and transected to mobilize the spleen. The splenic artery and vein were then dissected, ligated, and the spleen removed. The abdominal cavity was closed with a continuous 2-layer 2.0 suture.

Splenic Denervation

The spleen was exposed and mobilized, as described above. Denervation was achieved by transection of visible nerves and topical application of phenol.⁴⁰ The splenic artery and vein were dissected proximal to their perihilar branching. Perivascular tissue was removed, and visible nerves were transected. The dissected vessels were coated circumferentially with an 88% phenol solution using the cotton end of a wooden applicator. The abdominal cavity was closed with a continuous 2-layer 2.0 suture.

Protocols in Pigs

RIPC and placebo protocols were performed in contemporary random order.

RIPC Protocol

A tourniquet was placed around the left hindlimb and tightened for 5 minutes; pale skin was taken to indicate leg ischemia. The tourniquet was quickly released after 5 minutes of ischemia and the limb reperfused for 5 minutes. Skin blush indicated reperfusion. The I/R cycle in the hindlimb was performed 4×. Arterial blood (200 mL) was sampled 60 minutes after completion of the last I/R cycle in the hindlimb. Blood samples were taken immediately before the coronary occlusion because any protective substances would need to be present then to exert their protective function. Blood was sampled into vials containing lithium heparin (B. Braun Melsungen, Germany) and immediately centrifuged at 4°C with 800g for 10 minutes. The separated plasma was again centrifuged at 4°C with 4500g for 10 minutes and then stored at −80°C for later use. The mean storage time of plasma samples was no longer than a maximum of 14 months, with a mean of 3±4 months; we did not observe changes in protective properties over time. Systemic hemodynamics and regional myocardial blood flow were measured at baseline. The suture around the left anterior descending coronary artery was then carefully tightened against a soft silicone plate. At 5 and 55 minutes of coronary occlusion, systemic hemodynamics and regional myocardial blood flow were measured again. After 60 minutes of coronary occlusion, reperfusion was induced by quick release and removal of the suture, as confirmed by the disappearance of the light blue color and the reappearance of red color on the surface of the reperfused myocardium. Systemic hemodynamics were again measured at 30, 60, and 120 minutes of reperfusion. Reperfusion was continued for 180 minutes. Ventricular fibrillation during the protocol was immediately terminated by electrical countershock.⁴¹

Placebo Protocol

The placebo protocol was identical to that of RIPC, except that the conditioning maneuver on the hindlimb was omitted.

RIPC and placebo were performed in pigs without and with vagotomy and splenectomy, respectively; a placebo protocol was omitted in the splenic denervation group. Surgical protocols (vagotomy, splenectomy, and splenic denervation), RIPC, and placebo were performed before induction of myocardial I/R, respectively (RIPC+I/R, n=10; vagotomy+RIPC+I/R, n=8; splenectomy+RIPC+I/R, n=6; splenic denervation+RIPC+I/R, n=5; PLA+I/R, n=8; vagotomy+PLA+I/R, n=8; splenectomy+PLA+I/R, n=7).

Rats in Situ

Experimental Preparation

Lewis rats (male; 200–380 g; 2.5–3.5 months; local animal facility) were anesthetized with an intraperitoneal injection of ketamine/xylazine (100 mg per 10 mg/kg). Spontaneously breathing animals received oxygen-enriched air, were placed on a thermistor-controlled heating pad, and covered with drapes to prevent hypothermia. The heating pad was adjusted to keep rectal temperature between 36.5°C and 38.0°C. The anesthetic depth was assessed from the pedal withdrawal reflex, respiration, and heart rate. In subgroups of rats, vagotomy, splenectomy, or splenic denervation, respectively, were performed before RIPC or a respective placebo maneuver (Figure 1). Vagotomy, splenectomy, and splenic denervation were completed within 15 to 20 minutes, respectively, and placebo and RIPC protocols were time matched for these interventions. Surgical trauma per se had no influence on IS in preliminary experiments (Online Figure I).

Vagotomy

The cervical vagal nerves were exposed through a midline cervical incision and transected. Transection of vagal nerves was omitted in the sham surgery. The skin incision was closed with a continuous 4.0 suture.

Splenectomy

A left paramedian laparotomy was performed. Terminal branches of the splenic artery and vein were ligated near the splenic hilus with a 3.0 silk suture. The spleen was then removed. Splenectomy was omitted in the sham surgery. The abdominal incision was closed with a continuous 2-layer 4.0 suture.

Splenic Denervation

The spleen was exposed, as described above. Using a stereomicroscope (LS 6000IC; Beckman Coulter, Krefeld, Germany), the origin of the splenic artery in the celiac trunk was identified and dissected. The splenic artery was coated with an 88% phenol solution using a small piece of soaked surgical gauze. The spleen was then pulled gently toward the midline incision, and both tips of the spleen were dissected to transect nerves entering the splenic tips.⁴²

Protocols in Rats

RIPC and placebo protocols were performed in contemporary random order. These protocols were performed in situ, whereas myocardial I/R was induced in isolated rat hearts, because we wanted to validate the use of our isolated rat heart and isolated spleen preparation for the transfer experiments and show that RIPC is indeed elicited in rats and signals through the spleen to the target organ heart.

RIPC Protocol

One third of the initial anesthetic drug dosage was again injected intraperitoneally to maintain anesthesia. A tourniquet was placed around the left hindlimb and tightened for 5 minutes; dark blue skin color was taken to indicate leg ischemia. The tourniquet was quickly released after 5 minutes of ischemia and the limb reperfused for 5 minutes. Skin blush indicated reperfusion. The I/R cycle in the hindlimb was performed 3×. Thirty minutes after the last I/R cycle, unfractionated heparin (300 IU/kg; heparin-Natrium-2500-ratiopharm; Ratiopharm GmbH, Ulm, Germany) was injected intraperitoneally to attenuate coagulation. A bilateral thoracotomy was performed and the heart rapidly excised, placed in ice-cold saline, and immediately mounted on a Langendorff apparatus.

Placebo Protocol

The placebo protocol was identical to that of RIPC, except that tightening of the tourniquet was omitted. Injection of heparin and excision of the heart corresponded to the respective timing in the RIPC group. RIPC or placebo were performed without and with vagotomy, splenectomy, or splenic denervation, respectively, and hearts were isolated, perfused at constant pressure, and subjected to global I/R (GI/R; sham+RIPC+GI/R, n=16; vagotomy+RIPC+GI/R, n=9; splenectomy+RIPC+GI/R, n=11; splenic denervation+RIPC+GI/R, n=7 or sham+PLA+GI/R, n=18; vagotomy+PLA+GI/R, n=11; splenectomy+PLA+GI/R, n=10; splenic denervation+PLA+GI/R, n=8).

Transfer Experiments in Rats

The heart and spleen of rats, which had not undergone prior RIPC or placebo protocols in situ, were isolated. Again, Lewis rats (male; 200–380 g; 2.5–3.5 months) were used. Rats were anesthetized with an intraperitoneal injection of sodium pentobarbital (800 mg/kg, Narcoderm; CP-Pharma, Burgdorf, Germany, supplemented with unfractionated heparin 300 IU/kg). Hearts were isolated, mounted on a Langendroff apparatus, and perfused at constant pressure. Immediately after excision of the heart, the abdominal cavity was opened and the spleen excised rapidly and transferred into ice-cold saline, supplemented with heparin (250 IU/mL).

Isolated Perfused Rat Spleens

Using a stereomicroscope, the origin of the splenic artery and the celiac trunk were identified and dissected. The celiac trunk was then cannulated, and the left gastric artery and the common hepatic artery were ligated. Adhering pancreatic tissue was dissected, and pancreatic branches of the splenic artery were ligated. The spleen was placed into a warmed humidified chamber and perfused with a modified Krebs-Henseleit buffer (in mmol/L: NaCl 118.0, KCL 4.7, MgSO₄ 16, KH₂PO4 1.2, glucose 5.6, NaHCO₃ 24.9, sodium pyruvate 2.0, CaCl₂ 1.0; gassed with 95% O₂ and 5% CO₂ in a prewarmed reservoir) at a constant pressure of 50 to 55 mm Hg. Visible washout of blood into the terminal splenic veins and homogeneous discoloration of the spleen from deep red to light red were taken to indicate sufficient perfusion. The temperature of the perfusate and the chamber was monitored and kept between 37.0°C and 37.5°C. After 5 minutes of perfusion, rat splenic effluate (SEff) was sampled into a prewarmed reservoir, which was continuously gassed with 95% O₂ and 5% CO₂, until a final volume of 100 mL was collected. SEff was then transferred into 15 mL polypropylene tubes (Fisher Scientific, Hampton, NH) and centrifuged at 4°C with 2400g for 15 minutes. The cell-free supernatant was transferred into a prewarmed (37.5°C) reservoir and again gassed with 95% O₂ and 5% CO₂ for 15 minutes. Subsequently, SEff was titrated to 2.0 mmol/L CaCl₂, filtered (0.45 μm pore size; Merck Millipore, Burlington, MA), and gassed 95% O₂ and 5% CO₂ for further 15 minutes in a prewarmed reservoir (37.5°C) before use. A SEff sample was taken and stored at −80°C for later use.

Protocols in Isolated Perfused Rat Spleens

Carbachol Infusion

Using a syringe pump, 2 mL saline supplemented with 500 pmol/L carbachol was infused into the isolated spleen at constant flow of 2 mL/minute. One hundred-milliliter effluate was sampled after carbachol infusion (SEff_carbachol). The final carbachol concentration in SEff_carbachol was ≤10 pmol/L. This estimated final concentration of carbachol in the SEff had been elaborated in preliminary experiments not to reduce IS in isolated perfused rat hearts. For that, saline supplemented with carbachol (1, 10, 20, 100, 1000, and 100 000 pmol/L) was infused into isolated perfused rat hearts.

Saline Infusion

Using a syringe pump, 2 mL saline was infused into the isolated spleen at constant flow of 2 mL/minute. Again, 100 mL effluate was sampled after saline infusion (SEff_saline).

Humoral Factors Released Into the SEff

The concentrations of humoral factors in SEff (after saline perfusion without and with carbachol, respectively) were determined using enzyme immunoassays. Concentrations of humoral factors were determined in the 5-fold concentrated SEff (60°C for 90 minutes; Concentrator plus/Vacufuge plus; Eppendorf AG, Hamburg, Germany). Temperature and duration of the concentration procedure had been elaborated in preliminary experiments to not impact on humoral factor concentration or to induce crystallization of the saline. Standards and samples were added to microplates, which were precoated with the specific antibody against the respective protein. For the detection of IL (interleukin)-1α, IL-1β, IL-6, IL-10, SDF-1α (stromal cell-derived factor-1α), TNF-α (tumor necrosis factor-α; all from R&D Systems, Abingdon, United Kingdom),⁴³ and GLP-1⁴⁴ (Elabsience, Houston, TX), the respective enzyme-linked antibody and a substrate solution were supplemented. After adding a substrate and a specific indicator solution, the enzyme-substrate reaction resulted in a yellow product. The color intensity was proportional to the concentration of the protein. The yellow color intensity was then measured at 450 nm using a spectrophotometer (Microplate Reader 680; Bio-Rad, Munich, Germany). Concentrations of the respective proteins were quantified by comparison with a standard curve.

Protocols in Isolated Perfused Rat Hearts Used for Transfer Experiments

For time control experiments (n=5), hearts were perfused for an equal duration to that of an experimental protocol, that is 180 minutes.

Pig plasma samples were filtered (5 μm pore size; Machery-Nagel, Düren, Germany) and added with a syringe pump to the perfusate (dilution 1:10 volume ratio) before passing the heat exchanger. Pig plasma samples were infused into the isolated rat heart, and GI/R was induced by 30 minutes of full stop of perfusion and subsequent 120 minutes of reperfusion (see the Online Data Supplement for details): RIPC+I/R_plasma+GI/R (n=13); vagotomy+RIPC+I/R_plasma+GI/R (n=7); splenectomy+RIPC+I/R_plasma+GI/R (n=7); splenic denervation+RIPC+I/R_plasma+GI/R (n=6); PLA+I/R_plasma+GI/R (n=8); vagotomy+PLA+I/R_plasma+GI/R (n=8); splenectomy+PLA+I/R_plasma+GI/R (n=7). SEff samples (SEff_carbachol+GI/R, n=22 and SEff_saline+GI/R, n=18) were also infused into the isolated rat heart (Figure 1). Pig plasma samples and SEff samples were infused for 8 minutes, followed by 2 minutes of washout to avoid adherence of proteins from stagnant plasma/SEff in the Langendorff apparatus. Saline infusion served as control, respectively (saline+G/IR, n=8). Timing of plasma and SEff infusion, total infused volume, and plasma dilution, respectively, had been elaborated in preliminary experiments. Perfusion pressure, temperature, pH, and oxygen saturation of the Krebs-Henseleit buffer were not affected by pig plasma or SEff infusion, respectively.

Statistics

Investigators assessing IS (pig and rat) and performing Western blot analyses were blinded with respect to the performed protocols. Investigators performing isolated rat heart experiments were blinded with respect to the prior in situ protocols. Investigators performing isolated rat heart transfer experiments were blinded with respect to the protocol from which the specific pig plasma or SEff sample originated.

Our study was exploratory in nature; therefore, a formal power analysis was not performed, and sample size was estimated from our experience with similar studies.^25,38 Data were tested for normality with the Kolmogorov-Smirnov test (SigmaStat 3.5; Erkrath, Germany). The assumption of normality was confirmed for most analyzed data sets. Data are presented as mean±SD or as median [interquartile range]. Transmural myocardial blood flow, systemic hemodynamics in pigs, coronary flow (CF), and left ventricular developed pressure (LVDP) in isolated rat hearts and STAT3_tyr705 phosphorylation were analyzed by 2-way ANOVA for repeated measures (protocol, time). One-way ANOVA was used to analyze area at risk in pigs, IS in pigs, and in isolated rat hearts and CF and LVDP at baseline in isolated rat hearts. Fisher least-significant-difference post hoc tests were used to compare individual mean values when the ANOVA indicated a significant difference. IS data in isolated perfused rat hearts receiving SEff infusion without or with carbachol were not normally distributed and compared using the Mann-Whitney U test. Differences were considered significant at the level of P <0.05, and exact P values are given for P values when ≥0.001 for IS and STAT3_tyr705 phosphorylation.

Results

Experiments in Pigs

Systemic Hemodynamics, Transmural Blood Flow, Area at Risk, and IS in Pigs

Eight pigs were retrospectively excluded from analysis (Online Table I). Plasma samples taken before ischemia from 6 excluded pigs were, nevertheless, used in isolated rat heart transfer experiments because in these pigs, the placebo or RIPC procedure had been performed before the exclusion criteria were met. Heart rate, left ventricular pressure (LVP), and maximal rate of rise of LVP were not different between groups without vagotomy (Online Table II). Vagotomy increased heart rate, LVP, and maximal rate of rise of LVP (Online Figure II; Online Table II). LVP decreased with the onset of ischemia and remained below baseline up to the end of the protocol in all groups (Online Table II). The withdrawal of 200 mL arterial blood 60 minutes after completion of the last I/R cycle in the hindlimb was associated with a transient reduction of LVP by ≤10 mm Hg, which recovered with saline infusion for 4 to 5 minutes. Transmural myocardial blood flow at baseline and during ischemia and area at risk were not different between groups (Figure 2, table). With PLA (PLA+I/R), IS was 45±3%. IS was significantly less with RIPC (RIPC+I/R, 24±4%). Vagotomy, splenectomy, and splenic denervation attenuated IS reduction by RIPC (RIPC+I/R): vagotomy+RIPC+I/R: 36±4%, splenectomy+RIPC+I/R: 38±6%, splenic denervation+RIPC+I/R: 43±2% (Figure 2). When performing a post hoc pairwise comparison between the vagotomy+PLA+I/R and vagotomy+RIPC+I/R (49±2% versus 36±4%; Figure 2), a statistically significant difference was maintained. IS in pigs with vagotomy and RIPC (vagotomy+RIPC+I/R) appeared to cluster in 2 different groups with IS <40% and >40% of area at risk. However, heart rate before and after vagotomy and during ischemia was not different between these 2 subgroups (Online Figure II). Vagotomy or splenectomy, respectively, did not impact on IS per se in the placebo protocols (Figure 2). The IS in placebo protocols were similar to our recently published IS in comparable protocols.^25,26,38

Figure 2. Infarct size in pigs.

Closed symbols: mean±SD. AAR indicates area at risk; I/R, ischemia/reperfusion; PLA, placebo; RIPC, remote ischemic preconditioning; SD, splenic denervation; SE, splenectomy; TMBFb/i, transmural myocardial blood flow at baseline (b) and during ischemia (i); and V, vagotomy. *P<0.001 vs PLA+I/R; **P<0.001 vs V+PLA+I/R; ***P<0.001 vs SE+PLA+I/R; ‡P=0.011 vs V+PLA+I/R; †P=0.016 vs RIPC+I/R; ††P=0.008 vs RIPC+I/R; †††P=0.001 vs RIPC+I/R (1-way ANOVA and Fisher least-significant-difference post hoc tests); #P<0.001 vs TMBFb (2-way ANOVA for repeated measures and Fisher least-significant-difference post hoc tests).

STAT3_tyr705 Phosphorylation

The expression of phosphorylated STAT3_tyr705, normalized to that of total STAT3 protein, was higher at 10 minutes of reperfusion than at baseline with RIPC+I/R but not with PLA+I/R (Figure 3A). In RIPC protocols, splenectomy (splenectomy+RIPC+I/R; Figure 3B) or splenic denervation (splenic denervation+RIPC+I/R; Figure 3C) but not vagotomy (vagotomy+RIPC+I/R; Figure 3A) abolished the RIPC-induced increase in STAT3_tyr705 phosphorylation at 10 minutes of reperfusion.

Figure 3. Phosphorylation of signal transducer and activator of transcription 3 (STAT3_tyr705) at baseline before coronary occlusion and at early reperfusion in pigs undergoing a placebo protocol before myocardial ischemia/reperfusion (PLA+I/R) or remote ischemic preconditioning (RIPC+I/R) with prior (A) bilateral cervical vagotomy (V+RIPC+I/R), (B) splenectomy (SE+RIPC+I/R), or (C) splenic denervation (SD+RIPC+I/R).

A–C, Top, middle, bottom: Membrane stained with Ponceau S, immunoreactivity signals for phosphorylated and total STAT3. The phosphorylation of STAT3_tyr705 in biopsies taken from the area at risk at baseline (empty symbols) and 10 min reperfusion (black symbols) was normalized to the respective total STAT3 (n=4 each). Data are mean±SD; 2-way ANOVA for repeated measures and Fisher least-significant-difference post hoc tests. *P value vs baseline.

IS in Isolated Perfused Rat Hearts After In Situ Experiments

Fourteen rats were excluded from analysis (Online Table I). Baseline values for CF and LVDP were not different between groups (Online Table III). The recovery of CF and LVDP after ischemia was not different between groups (Online Table III). With sham+PLA+GI/R, IS was 37±1%, and IS was significantly less with sham+RIPC+GI/R (21±2%; Figure 4). Surgical protocols did not impact on IS in placebo and RIPC protocols per se when compared with rats without prior surgery (Online Figure I). Therefore, sham surgery protocols (sham+vagotomy+GI/R and sham+splenectomy+GI/R) were pooled (Figure 4) and used as reference, respectively. Vagotomy, splenectomy, or splenic denervation abrogated the protective effect of RIPC (Figure 4), respectively. In placebo protocols, the respective surgical procedure had no impact on IS per se (sham+PLA+GI/R: 37±1%; vagotomy+PLA+GI/R: 33±3%; splenectomy+PLA+GI/R: 36±2%; splenic denervation+PLA+GI/R: 34 ±2%; Figure 4).

Figure 4. Infarct size in isolated perfused rat hearts after in situ protocols.

Closed symbols: mean±SD. Triangles represent rats with sham vagotomy and circles with sham splenectomy (sham+PLA+GI/R; sham+RIPC+GI/R). GI/R indicates global ischemia/ reperfusion; PLA, placebo; RIPC, remote ischemic preconditioning; SD, splenic denervation; SE, splenectomy; sham, sham surgery; and V, vagotomy. *P<0.001 vs PLA+GI/R; †P<0.001 vs sham+RIPC+GI/R (1-way ANOVA and Fisher least-significant-difference post hoc tests).

IS in Isolated Perfused Rat Hearts After Transfer Experiments

Baseline values for CF and LVDP were not different between groups (Online Table IV). Infusion of pig plasma slightly decreased CF and LVDP, with only partial recovery during washout (Online Table IV); infusion of SEff_carbachol increased CF slightly (Online Table IV). With infusion of plasma from pigs after RIPC, IS was 24±2% (RIPC+I/R_plasma+GI/R; Figure 5). Such protection was not observed with plasma from pigs after placebo (PLA+I/R_plasma+GI/R: 40±3%; Figure 5). Infusion of plasma from pigs after RIPC with vagotomy, splenectomy, or splenic denervation did not reduce IS, respectively (vagotomy+RIPC+I/R_plasma+GI/R: 40±9%; splenectomy+RIPC+I/R_plasma+GI/R: 33±9%; splenic denervation+RIPC+I/R_plasma+GI/R: 39±8%; Figure 5). IS after infusion of plasma from pigs with placebo and prior vagotomy or splenectomy, respectively, was similar to that of placebo (vagotomy+PLA+I/R_plasma+GI/R: 38±4%; splenectomy+PLA+I/R_plasma+GI/R: 33±4%; Figure 5). The reduction in IS was not different with plasma from the 2 above subgroups of pigs with vagotomy and RIPC, in which there appeared to be 2 subgroups (Online Figure III). With infusion of SEff_carbachol, IS was and less than that with infusion of SEff_saline (SEff_carbachol+GI/R: 24 [19–27]% versus SEff_saline+GI/R:_: 35 [32–38]%; Figure 6A). In time control experiments, only negligible infarction was detected (Figure 6B). Infusion of saline and saline supplemented with 1, 10, and 20 pmol/L carbachol before induction of GI/R, respectively, did not impact IS (Figure 6B), whereas IS was less with infusion of saline supplemented with 100, 1000, or 100 000 pmol/L carbachol, respectively (Figure 6B).

Figure 5. Infarct size in isolated perfused rat hearts with infusion of pig plasma.

Closed symbols are mean±SD. GI/R indicates global ischemia/ reperfusion; PLA, placebo; RIPC, remote ischemic preconditioning; SD, splenic denervation; SE, splenectomy; and V, vagotomy. *P<0.001 vs PLA+I/R_plasma+GI/R; **P<0.001 vs V+PLA+I/R_plasma+GI/R; ***P=0.013 vs SE+PLA+I/R_plasma+GI/R; †P<0.001 vs −RIPC+I/R_plasma+GI/R; ††P=0.030 vs −RIPC+I/R_plasma+GI/R; †††P<0.001 vs −RIPC+I/R_plasma+GI/R (1-way ANOVA and Fisher least-significant-difference post hoc tests).

Figure 6. Infarct size in isolated perfused rat hearts with infusion of splenic effluate (A) or carbachol (B).

A, Data are presented as minimum and maximum (crosses), interquartile range from 25% to 75% (box), mean (square), and median (line) in a box plot. *P<0.001 vs SEff_saline+GI/R; Mann-Whitney U test. B, Closed symbols: mean±SD. GI/R indicates global ischemia/reperfusion; SEff, splenic effluate; and TC, time control. *P<0.001 vs TC; †P<0.001 vs saline+GI/R; ‡P<0.001 vs 1 pmol/L carbachol+GI/R; §P<0.001 vs 10 pmol/L carbachol+GI/R; ∥P<0.001 vs 20 pmol/L carbachol+GI/R (1-way ANOVA and Fisher least-significant-difference post hoc tests).

Humoral Factors in the SEff

In the 5-fold concentrated SEff samples, humoral factor concentrations (IL-1α, IL-1β, IL-6, IL-10, SDF-1α, TNF-α, and GLP-1) were all lower than the respective enzyme immunoassay’s limit of detection, respectively (data not shown).

Discussion

We have identified the spleen with its innervation as an indispensable relay organ for RIPC’s cardioprotective signal transduction (Figure 7). Of course, the decisive role of the spleen in cardioprotection by RIPC does not exclude the transfer of cardioprotective substances from one isolated heart with brief I/R to another naive isolated heart without any involvement of the spleen.²³

Figure 7. Vago-splenic axis in remote ischemic preconditioning.

Peripheral sensory nerves in yellow, sympathetic nerves in purple, and vagal nerves in green.

In rats, bilateral cervical vagotomy, splenectomy, or splenic denervation each abrogated cardioprotection by RIPC. Pharmacological activation of muscarinic receptors in the isolated perfused rat spleen induced the release of cardioprotective factors into the effluate, supporting a vago-splenic axis. A vago-humoral interaction for RIPC signal transfer has been reported in previous studies in rats; however, these studies failed to identify the actual source of released cardioprotective factors.^17,18

In pigs, however, the situation is more complex than in rats. As described before, RIPC-induced cardioprotection is associated with myocardial STAT3 activation,^25,39 and there is humoral transfer of cardioprotection between species.^25,26 Different from rats, in pigs, the vagal signal transfer of RIPC appears to be not mandatory for cardioprotection because bilateral cervical vagotomy only attenuated IS reduction by RIPC. However, looking at the individual IS of pigs with vagotomy and RIPC, there appeared to be 2 subgroups, either with full protection or without. The immediate rise in heart rate after vagotomy was not different between these subgroups, excluding incomplete vagotomy. In vagotomized pigs with RIPC, myocardial STAT3 was activated as in pigs without vagotomy. Obviously, STAT3 activation per se is not sufficient for cardioprotection by RIPC but may well be required. Although vagotomy did apparently not fully abolish RIPC cardioprotection in the pig myocardium, release of humoral cardioprotective factors was, nevertheless, attenuated because there was no longer IS reduction with humoral transfer to isolated saline-perfused rat hearts. This apparent lack of humoral signal transfer, however, may largely reflect the dilution of the infused plasma (1:10) into the isolated saline-perfused rat heart with its artificially increased CF, which would not be the case in the pig in situ. Currently, all studies reporting a causal role of vagal nerves in signal transfer of RIPC cardioprotection were performed in rodents.^16–18 However, vagal tone is largely different between rodents and larger mammals, as reflected by the difference in heart rate by an order of magnitude,⁴⁵ and thus, a species-dependent difference is conceivable for the vagal contribution to RIPC cardioprotection. Electrical vagal nerve stimulation in pigs reduces IS.^19,20 Thus, vagal nerves are clearly causally involved but not mandatory for RIPC cardioprotection in pigs.

As in rats, splenectomy or splenic denervation in pigs fully abrogated RIPC cardioprotection and RIPC-induced myocardial STAT3 activation, and there was no humoral transfer of cardioprotection to isolated perfused rat hearts. Apparently, unlike vagal nerves, the spleen is an indispensable relay organ of the cardioprotective signal transduction of RIPC. Members of the cytokine and growth hormone family are known to activate STAT,^4,46 and several humoral factors have been associated with cardioprotection or STAT activation,⁴³ such as IL-10.^9,47 In patients undergoing elective bypass surgery associated with prior RIPC, IL-1α tended to increase.⁴³ IL-10 is associated with RIPC cardioprotection,^47,48 attenuates myocardial I/R injury,⁴⁹ and is released by the spleen.^50,51 To date, we have failed to detect any of the previously described factors in the SEff after pharmacological muscarinic receptor activation with carbachol. However, we used enzyme immunoassays with a limited detection range and a more sensitive, nonbiased approach, for example, mass spectrometry may be needed to identify the released and potentially also unknown cardioprotective factors. For our data derived from the isolated spleen preparation, we must caution that (1) any humoral factor would be largely diluted with the artificial saline perfusion of the isolated spleen and even further with infusion into an isolated saline-perfused heart; (2) we can, therefore, not extrapolate to the potential stoichiometry of any released factor in situ; and (3) we cannot distinguish between a primary release of humoral factor(s) from the spleen or a secretion of such factor(s) from cells during our in vitro preparation of the spleen effluate. We also acknowledge that, in addition to the transfer of cardioprotection with cell-free plasma from pigs or effluate from isolated rat spleens to the isolated perfused rat heart, circulating leukocytes may also contribute to cardioprotection by RIPC, as previously suggested for neuroprotection.³⁵

The role of the sympathetic nervous system for RIPC cardioprotection is not entirely clear.¹⁶ In our splenic denervation experiments, we were unable to distinguish between parasympathetic and sympathetic innervation, which interact in the control of inflammatory responses to various noxious stimuli.³² Vagal nerve stimulation releases acetylcholine in the spleen,⁵² and the carbachol-induced release of cardioprotective factors into the effluate in our isolated spleen model emphasizes the role of the vago-splenic axis.

On the RIPC stimulus, circulating protective factors are released also in the periphery. In mice, endothelium-derived NO was released in response to RIPC, circulated as nitrite, and mediated cardioprotection by NO transfer into the myocardium.⁵³ In pigs, NO is not involved in cardioprotection by local ischemic preconditioning.⁵⁴ However, NO facilitates acetylcholine release from parasympathetic nerve fibers through a presynaptic action.⁵⁵ Therefore, peripherally released NO may have induced release of acetylcholine from splenic parasympathetic neurons, despite bilateral cervical vagal transection and have mediated subsequent release of cardioprotective humoral factor(s) other than NO from the spleen in situ. The lack of transferable cardioprotection with plasma from vagotomized pigs to isolated saline-perfused rat hearts, however, may again have been related to the substantial dilution in our transfer experiments.

Conclusions and Perspectives

Robust translation of remote ischemic conditioning into clinical practice requires a better understanding of the underlying signal transduction. We here demonstrate that a vago-splenic axis is causally involved in RIPC complex signal transduction in pigs and rats. The spleen is a central relay organ in both species and is indispensable for RIPC cardioprotection. Specific cardioprotective factor(s) released after RIPC from the spleen and additional factor(s), which may be released independently of vagal control in the periphery, remain to be identified. Experiments with pharmacological NOS (NO synthase) inhibition could clarify the role of NO for direct cardioprotection and protection through activation of a vago-splenic axis in RIPC. Further studies are also warranted to identify/quantify these factor(s) and to investigate their interaction with the vago-splenic axis, as well as their long-term effects on post-myocardial infarction remodeling and impending heart failure. It is possible that the protective factors released from the spleen in turn activate intracardiac ganglia, which again release acetylcholine in the heart to induce protection.^15,16,56 The contribution of intracardiac ganglia, which possibly release acetylcholine, can be studied in the future with use of atropine.

To date, the spleen has been considered to be essentially involved in repair and remodeling after acute myocardial infarction; the effects of splenic activation were mostly deleterious in experimental models,^57–59 as well as in patients with acute coronary syndrome.⁶⁰ Our study is the first to report an immediate involvement of the vago-splenic axis in acute myocardial I/R injury with a beneficial impact on IS. In a way, splenic activation thus mirrors the ischemic conditioning phenomenon, in that early and brief ischemia and splenic activation induce protection, whereas subsequent prolonged ischemia and splenic activation induce injury. Interestingly, World War II veterans with splenectomy had increased mortality from acute myocardial infarction on follow-up,⁶¹ supporting a protective role of the spleen in myocardial I/R.

Novelty and Significance.

What Is Known?

• Remote ischemic conditioning protects the heart from myocardial ischemia/reperfusion injury and reduces infarct size.

• The signal transfer from the stimulus site in a peripheral organ to the target site in the ischemic/reperfused heart involves neuronal and humoral pathways.

What New Information Does This Article Contribute?

• The spleen is an indispensable relay organ in the signal transfer of remote ischemic conditioning.

• Remote ischemic conditioning activates efferent vagal nerves to stimulate the spleen, which then releases humoral cardioprotective substances.

Brief episodes of ischemia/reperfusion in a remote tissue or organ protect the heart from injury by ischemia/reperfusion and reduce infarct size. The transfer of the protective signal from the peripheral tissue or organ to the heart involves neuronal and humoral pathways. In the present study, remote ischemic conditioning of the hindlimb in pigs reduced infarct size from subsequent coronary occlusion and reperfusion. Protection was abrogated by bilateral cervical vagotomy, by splenectomy or by splenic denervation, implicating a signal transduction through vagal stimulation of the spleen. Vagomimetic stimulation of the isolated perfused rat spleen induced the release of cardioprotective substances. The present study identifies the spleen as a central relay organ of cardioprotection by remote ischemic conditioning, which is under vagal control and can release cardioprotective substances on vagal activation.

Nonstandard Abbreviations and Acronyms

CF

coronary flow

GI

global ischemia

GLP-1

glucagon-like peptide-1

I/R

ischemia/reperfusion

IL

interleukin

IS

infarct size

LVDP

left ventricular developed pressure

LVP

left ventricular pressure

NOS

NO synthase

PLA+I/R

placebo+ischemia/reperfusion

RIPC

remote ischemic preconditioning

SDF-1α

stromal cell-derived factor-1α

SEff

splenic effluate

STAT3

signal transducer and activator of transcription 3

TNF-α

tumor necrosis factor-α