Abstract
Objective: Adaptation to ischemia by brief episodes of ischemia and reperfusion (preconditioning) of the heart protects the heart against sustained ischemia, where the transcription factor nuclear factor kappa-B (NFκB) appears crucial for the protection. Preconditioning of the heart may even be evoked by brief episodes of ischemia and reperfusion in other organs. The present study investigates a possible role for NFκB and inducible nitric oxide synthase (iNOS) in adaption to ischemia by remote, delayed protection. Methods: Mice (wild-types, or with targeted deletions of the NFκB p105 or the iNOS gene) were subjected to cycles of occlusion and reperfusion of both hind limbs, and 24h later their hearts were isolated and Langendorff-perfused with induced global ischemia and reperfusion. Infarct size was measured. Skeletal muscles from ischemized limbs as well as hearts were also collected for polymerase chain reaction (PCR) and electromobility shift assay (EMSA). Results: Hind limb preconditioning protected left ventricular function and reduced infarct size during reperfusion in wild-type mice. Nuclear translocation of NFκB was detected in both heart and preconditioned skeletal muscle 1–2h after the preconditioning episodes (EMSA); while cardiac mRNA for iNOS gradually increased in a 24-h time course after hind limb preconditioning (real-time PCR). When hind limbs of mice with targeted deletions for the p105 subunit of NFκB or the iNOS gene were preconditioned, no beneficial effect was observed in the heart. Conclusions: Delayed cardioprotection induced by hind limb preconditioning involves signaling through NFκB and iNOS.
1 Introduction
Ischemic preconditioning of the heart profoundly protects cell viability when the adaption takes place less than 2h before the sustained ischemia (classic preconditioning), or when the sustained ischemia is 24–72h after the preconditioning episode (delayed preconditioning). A preconditioning-like myocardial protection can be evoked by systemic stimuli such as lipopolysaccharide, hypoxia, hyperoxia, and nitric oxide [1–3]. It has also become evident that short episodes of ischemia and reperfusion of other organs such as mesenterium, kidney, or hind limbs will protect the heart (remote preconditioning) [1,4]. The perspective of preconditioning research is to find the underlying mechanism of action, and to pharmacologically exploit this in patients for increased myocardial protection.
Mechanisms underlying delayed preconditioning of the heart are the most well studied. Short episodes of ischemia and reperfusion lead to release of trigger substances such as nitric oxide, low doses of reactive oxygen intermediates, adenosine, bradykinin, and/or prostacyclin [3,5]. The triggers may cause activation of kinase cascades where protein kinase C, and probably downstream tyrosine kinase and members of the MAP kinase cascade are important for signaling [6,7]. Protein kinases may activate transcription factors, of which particularly nuclear factor kappa-B (NFκB) has been investigated in both classic and delayed models in the heart [8,9]. NFκB is a redox sensitive transcription factor regulating a battery of inflammatory genes such as inducible nitric oxide synthase (iNOS) and inducible cyclooxygenase. The NFκB family consists of the members p50, p52, p65 (RelA), c-Rel, and RelB which form various homo- and heterodimers, where the most common active form is the p50 or p52/RelA heterodimer. p105 is a precursor to p50 [10]. NFκB dimers reside in the cytoplasm in an inactive form bound to inhibitory proteins known as IκB, which are phosphorylated in response to diverse stimuli. The phosphorylated IκBs are then ubiquitinated and proteolytically degraded. This process activates NFκB, which translocates to the nucleus and binds to promoter or enhancer regions of specific genes, initiating transcription.
NFκB is translocated to the nucleus in the preconditioned heart, and pharmacological blocking of its translocation inhibits preconditioning [8,9]. The beneficial effect of NFκB in this context may be through upregulation of a beneficial gene, where inducible cyclooxygenase [11], manganese superoxide dismutase, as well as iNOS [12] are implicated as mediators of infarct size reduction afforded by delayed ischemic preconditioning of the heart. Another possibility is that NFκB activation during the brief ischemic episodes induces an inhibition of NFκB activation during sustained ischemia through increased of IκB [13]. It is also possible that the apoptosis limiting effect of NFκB activation is important for the protection afforded [14].
The mechanisms by which remote preconditioning protects the heart are less well understood. The target of the present study was to investigate the role of NFκB and iNOS in a mouse model of remote, delayed preconditioning induced by hind limb occlusion 24h prior to heart isolation and Langendorff-perfusion with induced global ischemia and reperfusion. Mice with targeted deletions of the NFκB p105 subunit, or the iNOS gene were employed for this purpose and compared with wild-types.
2 Materials and methods
2.1 Animals
The study was performed in accordance with the European Convention for the Protection of Vertebrate Animals used for Experimental and other Scientific Purposes (Strasbourg 1985), and was approved by the Animal Ethics Committee of Northern Stockholm (Stockholms Norra Djurforsöksetiska Nämnd). Female mice with homozygously targeted deletion of the NFκB p105 gene, as well female homozygous iNOS knockout mice were purchased from The Jackson Laboratory (Maine, USA, B6, 129P-Nfkb1〈tm1Bal〉 and B6, 129P-Nos2〈tm1lau〉) with wild-type age-matched females as controls (B6, 129PF1/J-A≪A-J/A〈W〉FX or F1). The mice were approximately 4 months old when employed for experiments, and were without any phenotypical or developmental deviations.
2.2 Hind limb preconditioning and isolated heart perfusion
The mice were anaesthetized intraperitoneally with pentobarbital (60mg/kg), and a bilateral external tourniquet was applied around the upper hind limb joint for 5min followed by 5min reperfusion for a total of 6 cycles. Twenty-four hours after preconditioning or sham treatment the mice were re-anaesthetized, and hearts were isolated and Langendorff-perfused. Retrograde perfusion with gassed (5% CO2, 95% O2) Krebs–Henseleit buffer was performed at a constant pressure of 55mmHg. Left ventricular pressures, coronary flow, and arrhythmias were registered. The data were collected by a continuous data collection system.
2.3 Protocol
Hearts were stabilized for 25min, and the balloon volume adjusted to have a LVEDP of 5mmHg. Forty minutes of global ischemia was induced by clamping the inflow tubing followed by 60min of reperfusion. Wild-type mice with hind limb preconditioning, wild-type sham treated, NFκB p105 knockouts with hind limb preconditioning, NFκB sham treated, and iNOS knockouts either preconditioned or sham treated (n=7 of each) were investigated for infarct size and hemodynamics. Additional tissue was sampled for supplementary analysis.
2.4 Measurement of infarct size
At the end of reperfusion, hearts were perfused with 1% triphenyltetrazolium chloride solution (TTC, Sigma Chemical Co., St Louis, MO), fixed in 4% formaldehyde and cut into 1mm transverse slices. Digitized images were obtained by a magnified video system (LEICA Qwin, Leica Imaging Systems Ltd, Cambridge, England), and the percentage of infarcted area calculated semi-automatically (Adobe PhotoShop 5.0, Adobe, San Jose, CA).
2.5 Electromobility shift assay
Samples of skeletal muscle and hearts were collected before start of preconditioning, and 1 and 2h after later. Hearts were collected after 25min perfusion, 40min ischemia, and 20min reperfusion (triplicate samples at all time points and groups). Nuclear proteins were extracted and incubated with a 32P labeled probe containing the NFκB binding site (Promega) as previously described in detail [13]. DNA-protein complexes were electrophoresed on a 4% polyacrylamide gel. For supershift analysis, a rabbit polyclonal anti-p50 antibody or a rabbit polyclonal anti-p65 antibody (both Santa Cruz Biotechnology, Santa Cruz, CA) were employed. For competition analysis, unlabelled probe in 25- or 50-fold excess was added prior to radiolabeled probe (results not shown).
2.6 Real-time PCR
Hearts from preconditioned wild-types were collected before limb ischemia, and 3, 6, 12, 18, and 24h after preconditioning (two animals at each time point) for real-time polymerase chain reaction (PCR) analysis of iNOS. Total RNA was extracted with an RNA isolation kit (Ultraspec™ RNA, Biotecx, Houston, TX) and reversely transcribed. cDNA was amplified by real-time PCR with 1× TaqMan Buffer (PE Biosystem, Foster City, CA). For the amplification of β-actin the forward primer 5′-AGA GGG AAA TCG TGC GTG AC-3′, reverse primer 5′-CAA TAG TGA TGA CCT GGC CGT-3′ and probe 5′-CAC TGC CGC ATC CTC TTC CTC CC-3′ were used. For the amplification of iNOS, the forward primer 5′-CAG CTG GGC TGT ACA AAC CTT-3′, reverse primer 5′-CAT TGG AAG TGA AGC GTT TCG-3′, and probe 5′-CGG GCA GCC TGT GAG ACC TTT GA-3′ (PE Biosystems) were used. Each sample was analyzed in duplicates using ABI Prism 7700 Sequence Detector (PE Biosystems) and correlated to a standard curve. The reactions were performed in MicroAmp Optical 96-Well Reaction Plates (PE Biosystems).
2.7 Statistical analysis
Data are presented as mean±SEM. Repeated measure ANOVA was employed for evaluation of hemodynamics, with a Scheffe's post hoc test. Infarct size was compared by unpaired t test. P≪0.05 was considered significant.
3 Results
3.1 NFκB translocation after limb preconditioning
When nuclear protein extracts from heart and hind limb skeletal muscle were analyzed with EMSA, no activation of NFκB was apparent prior to preconditioning (Fig. 1A). However, 1 and 2h after end of limb preconditioning a nuclear factor binding specifically to a NFκB consensus sequence appeared in both heart and skeletal muscle. The composition of the protein–DNA complex was analyzed by adding antibodies to the p65 or p50 subunits of the NFκB family. The p50, but not the p65, antibody caused a retardation of the probe (Fig. 1A).
3.2 Cardiac iNOS mRNA induction
mRNA was extracted from hearts serially after preconditioning the hind limbs, and the synthesized cDNA was amplified with real-time PCR using β-actin as a reference gene. The ratio between iNOS and β-actin increased gradually after preconditioning the hind limbs, reaching the highest level 24h after the preconditioning episode, corresponding to time of Langendorff-perfusion (Fig. 1B).
3.3 Heart function
3.3.1 Left ventricular pressures
Left ventricular developed pressure (LVDP) decreased during reperfusion after 40min global ischemia. Preconditioning the hind limbs 24h prior to heart isolation attenuated this decrease (P≪0.05) (Fig. 2). In mice with targeted gene deletions of the p105 subunit of NFκB, the depression of LVDP during reperfusion was attenuated compared with wild-types (P≪0.05). However, no beneficial effect of preconditioning could be found (Fig. 2). In analogy, mice with targeted deletions of the iNOS gene had less depression of LVDP during reperfusion than wild-types (P≪0.05), but no beneficial effect of preconditioning was apparent (Fig. 2).
LVEDP increased during reperfusion of wild-types, compared with pre-ischemic values and thereafter gradually decreased. The increase of LVEDP was attenuated by preconditioning (P≪0.001) (Fig. 2). Mice lacking the p105 gene had less severe increase of LVEDP than wild-types (P≪0.001), but no additional attenuation was afforded by preconditioning. Deletion of the iNOS gene did not influence LVEDP compared with wild-type. No beneficial effect on LVEDP was afforded by preconditioning (Fig. 2).
3.3.2 Coronary flow, heart rate, first derivatives of pressure and occurrence of arrhythmias
No significant differences could be detected between groups in these parameters.
3.3.3 Infarct size
After 1h reperfusion after global ischemia, approximately 35% of myocardial tissue was evaluated to be infarcted by TTC staining in wild-type mice. Preconditioning reduced this to about 24% (P≪0.05) (Fig. 3). Infarct size in mice deleted for p105 or iNOS were not significantly different from wild types. Preconditioning did not reduce the amount of infarcted tissue in these animals (Fig. 3).
3.3.4 Translocation of NFκB during global ischemia and reperfusion
Possible NFκB activation during Langendorff-perfusion and its modification by preconditioning was evaluated by EMSA of nuclear proteins from hearts of wild-type and p105 knockout animals. During Langendorff-perfusion before ischemia a faint band was visible after hybridizing the NFκB probe with nuclear extracts from hearts of some wild-type animals. At the end of 40min global ischemia in wild-type controls, however, NFκB was activated, and some activation was also apparent after 20min of reperfusion (Fig. 4). Limb preconditioning abolished the activation of NFκB during ischemia (Fig. 4). In cardiac nuclear protein extracts of animals with targeted deletion of the p105 gene, no bands appeared on the gels (Fig. 4).
4 Discussion
The main findings of the present study were that brief episodes of ischemia and reperfusion to the hind limbs protected the heart 24h later against induced global ischemia. Preconditioning of the hind limbs caused activation of NFκB in both the ischemic-reperfused hind limbs and in the heart, where increased mRNA for iNOS was detected. When preconditioning was performed in mice with targeted deletions of the NFκB p105 or the iNOS gene, no adaptation to ischemia was found, but the animals had a slight improvement of left ventricular performance during reperfusion compared with wild-types. In conjunction with induced global ischemia and reperfusion, NFκB was activated in wild-type controls, but not in preconditioned hearts. NFκB activation could not be detected in the p105 knockout animals. Delayed cardiac protection with hind limb preconditioning involves signaling via NFκB, and iNOS may be one end-effector of the protection afforded.
From the present data, a dual role of NFκB in the heart is indicated; detrimental during ischemia-reperfusion injury without adaptation, but a key factor in adaptation to ischemia by hind limb preconditioning. This dual role is in accordance with the current literature. NFκB is activated by myocardial ischemia and reperfusion [15]. Inhibition of its gene products such as leukocyte adhesion molecules, cytokines, and chemokines during reperfusion protects the heart against reperfusion injury [16–18]. More direct evidence for a detrimental role of NFκB is supplied by Morishita et al. [19], who transfected rats intracoronary with a double stranded oligonucleotide containing the NFκB cis-element before coronary artery ligation. The decoy inhibited NFκB-activation during reperfusion, and concomitantly reduced infarct size [19]. When the decoy was used for transfection and heterotopic transplantation 3 days prior to Langendorff-perfusion with Krebs-Henseleit buffer containing rat leukocytes, improved cardiac function during reperfusion was found together with reduced neutrophil adherence and tissue interleukin 8 production [20].
Signaling to NFκB such as generation of oxygen free radicals and activation of protein kinase C, p38 MAP kinase, tyrosine kinase, and MAPKAP kinase 2 are crucial for the preconditioning response in various experimental models [6,7]. A role for NFκB in ischemic preconditioning of the heart is recently suggested in both classic [8] and delayed [9] models in rats and rabbits. NFκB is activated during the preconditioning episodes, and pharmacological inhibition of NFκB abolishes the cardioprotection [8,9]. In a model of classic preconditioning in the rat, the p50 subunit of NFκB is indicated to be the important unit through pharmacological studies [8]. In the present study, animals with targeted gene deletion of the p105 precursor to p50 were employed. Adaptation to ischemia was lost in these animals, indicating that the p50 subunit of NFκB is crucial for the response. In the present study, the beneficial effects on left ventricular function may have been secondary to reduction of necrosis.
We can as yet only speculate on how ischemic adaptation of the hind limb is transmitted to protection of the heart, as very few papers have attempted to address the underlying mechanisms of remote preconditioning. It is recently shown that cross-perfusion with donor blood from a preconditioned animal could evoke a protection in the host [21], indicating that blood-borne transmission of protection is likely. One factor indicated as a trigger of remote preconditioning is adenosine, as an adenosine receptor blocker abolishes the beneficial effects. No study has previously investigated the involvement of NFκB or genes it regulates in remote preconditioning.
Nitric oxide has been suggested as both a trigger and a mediator of the preconditioning response [3]. iNOS is induced in cardiac tissue as a response to preconditioning of the heart [22], as its mRNA was in the present study after hind limb preconditioning. Mice with deletion of the iNOS gene could not be preconditioned by hind limb adaptation. This is in accordance with others' findings using pharmacological inhibition of iNOS or genetic deletion of the gene [5,12,22].
One previous paper using a model of remote, delayed preconditioning targeting the brain did not protect hearts of iNOS KO mice [23], nor did it preserve in vitro vessel reactivity in arterial rings from iNOS KO [23]. Details of why iNOS contributes to remote preconditioning in the present paper is, however, not clear. The vasorelaxing effect of nitric oxide appears not to be important, as there were no differences in coronary flow between groups. Thus, other protective effects of NO must have been important.
Myocardial protection by NFκB activation may be caused by induction of a NFκB-regulated mediator as discussed above. However, myocardial protection by NFκB activation could also be caused by a downregulation of the inflammatory response during reperfusion, as NFκB activation increases its own inhibitor IκBα [13]. In the present study, hearts of preconditioned animals had less NFκB activation during the sustained ischemia. Accordingly, in HUVECs preconditioned by hydrogen peroxide, reduced upregulation of cytokines and leukocyte adhesion molecules after subsequent stimulation with TNF-α was found [24].
What is the relation between the present study and the daily cardiac surgery scenario? A considerable limitation of this work is that the step from murine physiology to humans is large. Studies on animals with genetic deletion or overexpression are a powerful scientific tool to understand gene function, and mice are the species where genetic engineering is well established. Unfortunately, murine physiology is in some aspects different from other mammals, and the findings must be extrapolated with caution. Furthermore, we chose an isolated heart model for induced ischemia and reperfusion. The advantage of the isolated heart is excellent evaluation of heart function without having to interpreter effects of preload and afterload; the disadvantage is a denervated, non-working heart lacking blood–cell interactions.
Studying the cellular and molecular mechanisms of increased cell defense towards injury may provide us with knowledge how to exploit this pharmacologically. The vision is to obtain ‘the pill the day before surgery’. The signaling of remote preconditioning may give us information how preconditioning of one organ may influence other organs, and in theory we may obtain a ‘whole body preconditioning’. A general increase of the endogenous cell defense may reduce the deleterious effects of cardiopulmonary bypass and surgery, and thus contribute to reduced morbidity and mortality.
Upper panel: A representative electromobility shift assay with a radiolabeled NFκB probe of nuclear protein extracts obtained from heart (H) or skeletal muscle (M) of mice subjected to bilateral hind limb preconditioning (6 cycles of 5min occlusion through an external tourniquet followed by 5min reperfusion). Samples taken before preconditioning (0), at the end of the procedure (1) and 2h after start of the procedure (2) of wild-type mice are shown. Sample M2 was supershifted when an antibody to the p50 subunit of NFκB was added, while a p65 antibody did not retard the labeled DNA-complex. The last lane shows sample H2 with the p50 antibody. Lower Panel: mRNA was extracted from hearts of wild-type mice before limb preconditioning as described above, and serially after start of preconditioning. The mRNA was amplified with real-time PCR with primers for iNOS and β-actin as a reference gene. The ratio between mRNA for iNOS and β-actin is shown as individual values from two experiments.
Left ventricular developed (LVDP) and end-diastolic (LVEDP) pressures in Langendorff-perfused mice hearts isolated 24h after sham treatment (C) or limb preconditioning (LPC) as described in legend to Fig. 1, and subjected to 40min of global ischemia followed by reperfusion. Mice were either wild-type (B6, 129), deleted for the p105 subunit of the NFκB gene, or deleted for the iNOS gene (n=7 in each group, mean±SEM). *P≪0.05 when comparing preconditioned animals with controls, #P≪0.05 when comparing mice with targeted deletions to wild-types (C versus C). S, stabilization; BI, before ischemia.
Infarct size after 60min reperfusion of the mouse hearts described in legend to Fig. 2. The bar graph shows the infarcted tissue in % of total left ventricular tissue after triphenyl tetrazolium chloride-staining of 1mm sections of the whole hearts, and adding together areas of all sections for further calculations (n=7 in each group, mean±SEM). *P≪0.05 when comparing preconditioned animals with controls. There were no significant differences between wild-type and gene deleted animals.
A representative EMSA of cardiac nuclear protein extracts from wild-type (B6,129) or NFκB p105 knockout (p105−/−) animals with a radiolabeled NFκB probe. Hearts were harvested after 25min Langendorff-perfusion (p), at the end of 40min global ischemia (i), or after 20min of reperfusion (r) 24h after sham treatment (C) or hind limb preconditioning (LPC). When densitometry was performed on three-independent experiments, the values for B6,129 mice were LPCp, 3±0.5×104; LPCi, 1±0.5×104; LPCr, 4±0.8×104; Cp, 2±0.8×104; Ci, 6±2.5×104; Cr, 3±1.5×104 (mean±SEM).
This work has been supported by grants from the Swedish Medical Research Council (11235 and 12665), The Swedish Heart-Lung Foundation, the Foundations Fredrik o Ingrid Thuring, Tore Nilsson, Åke Wiberg, Ragnhild and Einar Lundströms Memory, and the Karolinska Institutet. Guohu Li has been supported by a grant from the Wenner-Gren Foundations.
