Introduction

Myocardial infarction is a major cause for morbidity and mortality and often results in subsequent heart failure. Albeit many advances in prompt revascularization have been accomplished, only very few procedures for tissue protection to prevent further progressive heart dysfunction turned out to be successful. A promising procedure may represent ischemic precondition—and even remote preconditioning—which can prepare the heart muscle for an ischemic period of metabolic deficiency and cellular stress [13]. Previous studies demonstrated that ischemic preconditioning substantially relies on the activation of the transcription factor hypoxia-inducible factor (HIF) and that genetic deletion of HIF abrogates protective effects of preconditioning [6].

The heterodimeric transcription factor HIF is constitutively expressed, whereas the protein stability of the α subunits of HIF-1 or HIF-2 is oxygen-dependently regulated. In hypoxia, HIFα levels accumulate and stimulate the hypoxia-inducible gene expression. This involves genes regulating glycolysis, glucose metabolism, angiogenesis, cell survival, etc. [31, 37]. An enzyme family of oxygen- and 2-oxoglutarate-dependent dioxygenases regulates the stability of HIFα [18]. The prolyl-4-hydroxylase domain (PHD) enzymes 1–3 hydroxylate HIFα in a strictly oxygen-dependent manner, which marks the protein for ubiquitination and proteasomal degradation [19]. Both HIFα isoforms and in particular PHD2 and 3 are expressed in cardiomyocytes [29, 39]. Genetic inactivation of PHD2 in cardiomyocytes including conditional knockout or siRNA transfection led to HIF-1α accumulation and increased ischemia tolerance of infarcted hearts [15, 17, 25]. Besides a decreased myocardial infarct size, an improved fractional shortening 3 weeks after the ischemic event could be observed in these mice [15].

In the meantime, new small-molecule PHD inhibitors (PHIs) were developed and successfully applied, e.g., in acute organ injuries, like ischemic or toxic kidney injury [3, 14, 36]. For the heart, a decreased myocardial infarct size was also reported after treating rats with the non-specific PHI dimethyloxalylglycine (DMOG) given 24 h before coronary occlusion [40] or ischemia reperfusion [10]. Since DMOG is a non-selective oxoglutarate analogue, it inhibits, besides the PHDs, also other 2-oxoglutarate-dependent dioxygenases including the collagen-modifying hydroxylases [8]. In fact, N-oxalylglycine and its dimethyl ester analogue DMOG were described to inhibit prolyl-4-hydroxylase activity before the PHD enzymes were even discovered [1]. Therefore, the data obtained with DMOG have to be interpreted with caution. Another study applied the newly developed PHI GSK360A for 28 days starting first at 48 h after left anterior descending artery (LAD) ligation. In this study, myocardial remodeling was improved and vascularity was better maintained [2]. In contrast, genetic chronic activation of HIF-1α or HIF-2α in the mouse heart resulted in heart failure mimicking ischemic cardiomyopathy [16, 23], which might limit long-term application of PHI. Just recently, Ong et al. reported that also short-term pre-conditional HIFα stabilization via application of GSK360A protects the heart from acute myocardial infarction [26]. In the clinical setting, application of a PHI in response to an ischemic event is more realistic compared to a pre-conditional application. Thus, we wondered how far short-term application of a specific HIFα-stabilizing PHI leads to significant HIF activation in the heart and whether HIFα stabilization may have effects on myocardial infarction outcome measures directly after infarction and in the post-infarction period. We applied 2-(1-chloro-4-hydroxyisoquinoline-3-carboxamido) acetate (ICA), which has been chemically characterized before and has been successfully used in other acute ischemic models including kidney ischemia [30]. Characterization of ICA revealed a manifold higher affinity towards PHD2 compared to N-oxalylglycine and a higher specificity towards the PHDs over the asparaginyl hydroxylase factor-inhibiting HIF (FIH) [33, 34]. The tissue-protective effect of ICA was tested pre-conditionally or immediately post-conditionally after permanent ligation of the LAD in mice.

Materials and methods

ICA synthesis

ICA was obtained in a six-step synthesis as described earlier [30].

Animal experimentation and echocardiography

Animal experimentations were performed with male, 8–10-week-old C57BL/6 mice (Jackson Laboratories). All protocols regarding animal experimentation were conducted according to the German animal protection law and approved by the responsible governmental authority (Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit in Oldenburg; animal experimentation number 33.9-42502-04-10/0069) and conform to the Directive 2010/63/EU of the European Parliament. LAD ligations were performed by an investigator, who was blinded regarding the PHI treatment of the mice as described previously [15]. The measurement of infarct size was performed by an additional investigator, who was likewise blinded. Mice were anesthetized using 1.5 % isoflurane (Forene, Abbott), applied by a blunt intubation cannula (intubation cannula, stainless steel with Y-adapter, 1.2-mm outer diameter, 30-mm length; Hugo Sachs Elektronik, Harvard Apparatus GmbH) connected to a mechanical ventilator (MiniVent; Hugo Sachs Elektronik, Harvard Apparatus GmbH). Anesthesia was considered adequate when the animal stayed still quietly, did not respond to external stimuli, and did not show the palpebral reflex. Pain management was controlled by intraperitoneal (i.p.) injection of 0.06 mg/kg body weight (BW) buprenorphine 1 h and daily treatment with 1.33 mg/ml metamizole in the drinking water starting 2 days before the surgery.

Mice received two doses of ICA (40 mg/kg BW) or vehicle i.p., either 6 and 1 h before myocardial infarction or 1 and 5 h after myocardial infarction. Two-dimensional images and M-mode tracings were recorded from the parasternal long axis view at midpapillary level (Vevo 2100TM, MS400, Visual Sonics). Heart rate, end-diastolic area, end-systolic area as well as anterior and posterior wall thicknesses were determined. Fractional area shortening (FAS) of the left ventricle was defined as the (end-diastolic area − end-systolic area)/end-diastolic area. FAS was used as marker for cardiac function. During echocardiography, mice were anesthetized with 1 % isoflurane.

Measurement of infarct size

Six hours after LAD ligation, the mice were given heparin (250 IU), anesthetized with 2 % isoflurane, and sacrificed by cervical dislocation. Subsequently, the hearts were excised. Total infarct size was determined by using Evans blue and 2,3,5-triphenyltetrazolium chloride (TTC) [15]. Briefly, the ascending aorta was cannulated with a 20-gauge tubing adapter, and 1 % Evans blue was perfused into the aorta and coronary arteries to delineate the total area at risk (AAR). The Evans blue dye was uniformly distributed to those areas of the myocardium, which were well perfused; hence, the area of the myocardium that was not stained with Evans blue was defined as the total AAR. The left ventricle was separated from the rest of the heart and sectioned into three transverse slices. Sections were incubated in 2 % TTC for 20 min at 37 °C to identify viable tissue, which appears in red compared to the white/pale area of necrosis (AON). Infarct quantification was performed on digital photographs (SMZ 1500, Nikon, Tokyo, Japan) using ImageJ (NIH, Bethesda, MD). AON and AAR were determined as the average percent area per slice and were then related to individual slice weight. Total infarct size was calculated as AON/total AAR.

Immunohistochemistry

Paraffin sections (2 to 4 μm thick) were dewaxed in xylene and rehydrated in a series of ethanol washes. The following primary antibodies were used for immunodetection: polyclonal rabbit anti-HIF-1α (1:10,000, Cayman Chemical), polyclonal rabbit anti-HIF-2α (1:10,000; PM9 [38], a kind gift of P.H. Maxwell, University of Cambridge), polyclonal rabbit anti-PHD2 (1:10,000; Novus Biologicals), polyclonal rabbit anti-PHD3 (1:10,000, Novus Biologicals), polyclonal rabbit anti-heme oxygenase-1 (HO-1; 1:15,000, Enzo Life Sciences). A biotinylated secondary anti-rabbit antibody (1:1000, Dako) and a catalyzed signal amplification system (Dako) based on the streptavidin-biotin-peroxidase reaction were used for staining according to the manufacturer’s instructions.

ODD-Luc engineered heart muscle

ODD-Luc engineered heart muscles (EHMs) were generated as described earlier [12]. In brief, EHMs were assembled from purified cardiomyocytes derived from a murine transgenic ODD-Luc embryonic stem cell line and mouse embryonic fibroblasts at a ratio of 70:30. Mouse embryonic fibroblasts were isolated from 13–16-day-old embryos of pregnant NMRI mice after CO2 euthanasia. ICA and a solvent control were added to the ODD-Luc EHMs after a 7-day differentiation process under mechanical loading. Drug treatment was performed for 1 or 6 h. Subsequently, EHMs were removed from the stretchers, transferred in luciferin (Caliper) solution (1 mg/ml in PBS), and incubated for 1 min. The EHMs were subsequently placed on a 14.5-cm cell culture dish, covered with a glass slide, and imaged in a light-tight chamber with a luminescence image analyzer (LAS-3000).

Red blood cell count, hemoglobin, and hematocrit

Red blood cell count, hemoglobin concentration, and hematocrit were determined using the CELL-DYN Sapphire hematology analyzer system (Abbot Laboratories).

Statistical analysis

Data are presented as mean ± SEM. We determined statistical differences by two-tailed Student’s t test. A p value less than 0.05 was considered statistically significant.

Results

ICA increases HIFα abundance and HIF target gene expression in cardiac tissue

We have previously established ODD-Luc hypoxia reporter cardiomyocytes, which express the oxygen-dependent degradation domain of HIF-1α fused to the firefly luciferase under the control of the constitutively active CMV promoter [12]. These can be applied for generating EHMs, which are capable of monitoring sensitively and quantitatively the HIF response via bioluminescence in vitro. To analyze the efficacy for the induction of a HIF response after PHI treatment, ODD-Luc EHMs were generated and incubated with ICA for 1 and 6 h (Fig. 1a, b). Bioluminescence increased after PHI treatment, demonstrating that in cardiac tissue, ICA can robustly stimulate the HIF system. This was additionally verified by in vivo experiments with resting mice (Fig. 2a). Treatment of mice with ICA (40 mg/kg BW), which was injected i.p. once, in vivo resulted in a transient HIF-1α and HIF-2α accumulation in the heart. Most interestingly, the HIFα induction was rapid but transient; 1 h after injection, a clear HIF-1α and HIF-2α response was visible, which was even more pronounced at 6 h and was no longer visible at 24 h after treatment. HIF-1α was seen predominantly in cardiomyocytes but also in some stromal cells, whereas HIF-2α stabilization occurred predominantly in stromal cells which appeared morphologically to be endothelial cells in most cases (Fig. 2b). PHD3 and HO-1 are HIF target genes, which are upregulated as a consequence of HIFα stabilization and subsequent transcription activation of the HIFα/β complex. We could indeed see increased HO-1, PHD2, and PHD3 protein signals in the heart after ICA treatment, which persisted up to 24 h (Fig. 3). Whereas HO-1 was expressed predominantly in stromal cells, PHD2 and PHD3 were expressed ubiquitously. Compared to a strong induction of PHD3, PHD2 induction after ICA treatment was mild. Taken together, these data demonstrate that ICA is inducing a functional HIF program in the heart with elevated levels of HIFα up to 6 h after a single injection of the PHI.

Fig. 1
figure 1

ODD-Luc EHMs respond to ICA treatment with enhanced Luc signals. a ODD-Luc EHMs were incubated with 100 μM, 1 mM ICA or a vehicle control for 1 or 6 h as indicated. Color-coded image of the ODD-Luc EHMs (blue is low, red is high Luc signal). b Luc signal intensity of vehicle- or ICA-treated ODD-Luc EHMs was quantified. n = 4 EHMs per treatment group; shown are means ± SEM, *p < 0.05

Full size image
Fig. 2
figure 2

ICA treatment results in transient HIF-1α and HIF-2α stabilization. a HIF-1α and HIF-2α, immunohistochemistries of heart sections were performed under control conditions or 1, 6, and 24 h after ICA treatment (40-fold magnification). b Microscopy pictures of HIF-1α and HIF-2α immunohistochemistry of heart sections, which were obtained 6 h after ICA treatment (100-fold magnification). HIF-1α and HIF-2α immunohistochemistries were repeated with samples from three independent animals. Representative figures are shown

Full size image
Fig. 3
figure 3

ICA treatment results in induction of HIF target genes. Heme oxygenase-1 (HO-1), PHD2, and PHD3 immunohistochemistries of heart sections were performed under control conditions or 1, 6, and 24 h after ICA treatment (40-fold magnification). HO-1, PHD2, and PHD3 immunohistochemistries were repeated with samples from three independent animals. Representative figures are shown

Full size image

Short-term treatment with ICA affects neither cardiac function nor hematocrit in resting mice

Because PHIs induce erythropoietin synthesis, long-term application is linked to erythrocytosis, which may represent a new therapeutic option to treat renal anemia [9, 4]. In case of cardiac tissue protection, it is, however, important that a HIF response in the heart can be achieved without affecting the erythrocyte concentration and hematocrit to prevent an increased afterload. In line with this, a short-term ICA treatment of two dosages separated by a 5-h time interval failed to increase red blood cell count, hemoglobin concentration, and hematocrit (Fig. 4a–d). Additionally FAS, which is a marker for cardiac function as well as anterior and posterior wall thicknesses, which are markers for cardiac hypertrophy, were analyzed for up to 3 weeks (Fig. 4e–g). Neither cardiac function nor cardiac wall thickness was significantly altered by the PHI treatment.

Fig. 4
figure 4

Cardiac function, hematocrit, and erythrocyte and hemoglobin concentrations are not affected by short-term ICA treatment. a Schematic sketch showing the experimental time frame starting with the initial echocardiography. b Red blood cell concentration, c hemoglobin concentration, and d hematocrit of mice 21 days after ICA treatment and of vehicle-treated control animals. The numbers in the bars indicate the number of mice analyzed. Echocardiographic measurements were performed before ICA treatments and serially 7, 14, and 21 days after treatments. e Fractional area shortening (FAS) and f anterior and posterior wall thicknesses were analyzed. n = 9–10 per treatment group; shown are means ± SEM

Full size image

Pre- and post-conditional ICA treatment decreases the myocardial infarct size after LAD ligation

Tissue protection from ischemia was demonstrated in genetically modified mice with constitutively high HIF-1α levels [15, 21]. These animals had a pre-conditional high HIF response, which protects the tissue from the subsequent insult. In line with these mouse models, a decrease in the total infarct size was seen 6 h after LAD ligation when resting mice were treated with ICA 6 and 1 h before LAD ligation (Fig. 5a–c). Pre-conditional application of a tissue-protective intervention is restricted to rare conditions of a planned ischemic procedure like transplantation etc. but is not useful for acute cardioprotection after myocardial infarction. Therefore, we next tested if post-conditional application of ICA 1 and 5 h after LAD ligation would affect the total infarct size likewise (Fig. 5d, e). Compared to the vehicle control, post-conditional application of ICA indeed significantly reduced the total infarct size, demonstrating that after the ischemic insult, there is a therapeutic time window. Cardiac function was severely impaired after LAD ligation, which is exemplified by a decreased FAS 7 and 14 days after surgery (Fig. 6). Tissue protection after treatment with ICA was translated into a partially better preserved heart function 14 days after LAD ligation.

Fig. 5
figure 5

Protective effect of ICA in myocardial infarction. a Schematic sketch showing the experimental time frames of pre- and post-conditional ICA application. b Hearts of pre-conditionally treated mice were harvested 6 h after LAD ligation. The total infarct size was calculated based on Evans blue perfusion and TTC staining. c Representative mid-myocardial cross sections of stained left ventricles analyzed in b are shown. d Hearts of post-conditionally treated mice were harvested 6 h after LAD ligation. The total infarct size was calculated based on Evans blue perfusion and TTC staining. e Representative mid-myocardial cross sections of stained left ventricles analyzed in d are shown. For visualization purposes, the stained areas in c and e were color marked (perfused area: blue outline; vital AAR: red outline; AON: grey outline). The numbers in the bars indicate the number of mice analyzed. n = 14–17 per treatment group; shown are means ± SEM, *p < 0.05

Full size image
Fig. 6
figure 6

Fractional area shortening is partially rescued after ICA treatment. Mice were treated with ICA (n = 8) or with vehicle control (vehicle, n = 6) 6 h before undergoing LAD ligation. Before LAD ligation as well as 7 and 14 days after surgery, fractional area shortening (FAS) was determined by echocardiography. Shown are means ± SEM, *p < 0.05

Full size image

Discussion

Activation of the HIF signaling pathway has a therapeutic potential for the treatment of ischemic diseases [27]. There are several reports demonstrating that treatment with known inhibitors of the PHD enzymes ameliorate a subsequent acute damage in models mimicking renal ischemia or stroke [24, 28, 30, 32]. Genetic approaches have additionally supported the idea that HIF-1α stabilization increases the tolerance towards hypoxia [22]. Regarding cardiac ischemia, previous studies suggest that chronic treatment with PHI after myocardial infarction improves cardiac function by supporting remodeling and increasing the vascularity in the peri-infarct region [2]. This effect was described in rats treated with the PHI GSK360A, which was given for 28 days starting first at 48 h after ligation of the LAD. Short-term protective effects of GSK360A on myocardial infarct size were seen after pre-conditional application [26]. Genetic HIF-1α or HIF-2α overexpression in the mouse heart demonstrates that chronic activation of HIF results in heart failure mimicking ischemic cardiomyopathy [16, 23]. This effect might limit the possibility for a long-term treatment with PHIs and thus narrow the therapeutic time window. Additionally, cardio-selective PHD inhibitors are currently not available. Therefore, heart function after long-term treatment has to be analyzed in the context of hematological effects like increased hematocrit, which alters the mechanical load of the heart. In this regard, our data demonstrate most importantly that short-term pre- as well as post-conditional treatment of mice with the PHI ICA decreased the myocardial infarct size. The tissue-protective effect was achieved after treatment with the inhibitor done just twice over a period of 6 h. ICA treatment was associated with a significant stabilization of HIF-1α and HIF-2α in the heart, which peaked transiently at 6 h after injection. Therefore, short-time treatment with a PHD inhibitor might be sufficient to stabilize HIF-1α without affecting systemic hematocrit based on different pharmacokinetic profiles of this small-molecule drug in the heart compared to the kidney. In this regard, it should be noted that the HIFα kinetic in the kidney after ICA treatment, which was reported previously, differed significantly from the kinetic seen in the heart. Whereas the response in the heart was transient, HIFα accumulation in the kidney can be observed later and up to 72 h after ICA injection [30]. In this study, ICA induced HIF-1α in all nephron segments with predominant high levels in the distal nephron. HIF-2α was detectable additionally in interstitial and glomerular cells. This might at least in part explain that in renal ischemia, pre-conditional application but not post-conditional application, in sharp contrast to myocardial ischemia, resulted in tissue protection [35].

Although the net effect regarding tissue protection was similar in the pre- versus post-conditional ICA treatment conditions, the underlying mechanisms still might be different [11]. Cardioprotection initiated through ischemic pre- and post-conditioning has been associated with a myriad of cellular and subcellular adaptive responses including endogenous stimulation of adenosine receptors, iNOS and HO-1, activation of survival kinases, opening of ion channels, attenuation of mitochondrial activity, etc. [5]. HIF itself as well as several of its target genes has been associated with ischemic pre- and post-conditioning [7]. Since ICA resulted in a clear stabilization of HIF-1α, HIF-2α, and its target genes like HO-1 and PHD3, it is highly likely that the tissue-protective effect is mediated via HIF. The protective effect, however, is not necessarily limited to the stabilization of HIFα in the cardiomyocytes as suggested by the described cardiomyocyte-specific genetically modified mouse models. Interestingly, besides HIF-1 and HIF-2α stabilization in cardiomyocytes after ICA treatment, HIF-2α stabilization occurred additionally in endothelial cells, which may contribute to the observed protective effects. Cardioprotection by HIFα stabilization in endothelial cells was previously demonstrated in a PHD2 hypomorph mouse model, which, in contrast to the described cardiomyocyte-specific deletion of PHD2, has additionally significantly lower PHD2 protein levels in the cardiac endothelial cells [20]. Tie2 signaling in the endothelial cells contributed significantly to the ischemic cardioprotective effect in this mouse model.

After acute myocardial infarction, the heart undergoes a detailed remodeling process, which also involves the ischemic border zones. Heart function after ischemia is affected not only by the acute loss of tissue but also by a beneficial remodeling process. Cardiac fibroblasts are highly involved in adjusting heart structure and remodeling. Detailed follow-up studies involving this cell type need to be performed to clarify if inhibiting PHD activities also would beneficially support the remodeling process after infarction.

Taken together, our data demonstrate that post-myocardial infarction has a therapeutic time window, which might be used for acute tissue protection via post-conditional application of PHI.