DNA chips facilitate genomic-wide exploration of gene expression. The authors hypothesized that ischemic (IPC) and anesthetic preconditioning (APC) would differentially modulate gene expression in hearts.
Affymetrix rat U34A gene chips were used to explore the transcriptional response to IPC and APC, sustained ischemia (110 min) without reperfusion, and time-matched perfusion in isolated rat hearts. IPC was induced by three cycles of 5 min of ischemia, and APC was induced by 1.5 minimum alveolar concentration isoflurane (110 min). For each heart, a separate chip was used for hybridization. Data were analyzed for significant > or = 2.0-fold changes in gene expression. Microarray results were confirmed by quantitative real-time reverse-transcription polymerase chain reaction.
Of the 8,799 genes represented on U34A, 217 transcripts in the APC group, 234 in the IPC group, and 29 in the ischemia group displayed significant > or = 2.0-fold up-regulation in messenger RNA levels, and 185 transcripts in the APC group, 55 in the IPC group, and 49 in the ischemia group displayed significant > or = 2.0-fold down-regulation. Many of these transcripts were unknown genes. A high number of commonly regulated genes were found in IPC and APC (39 up-regulated, 17 down-regulated). Genes commonly regulated included those associated with cell defense (heat shock protein 10, aldose reductase, Bcl-xS). Conversely, a pool of protective and antiprotective genes was differentially regulated in APC versus IPC (heat shock protein 27/70, programmed cell death 8), suggesting trigger-dependent transcriptome variability.
The novel microarray technology provides evidence for distinct cardioprotective phenotypes in IPC and APC. The observed transcriptional changes raise the possibility of a second window of protection by volatile anesthetics. The authors' molecular portraits are the first global genomic comparison between IPC and APC.
SIMILAR to brief ischemic episodes, volatile anesthetics promote and establish a protective cellular state in cardiac tissue by complex multiple fast-acting phosphorylation signaling steps. 1This phenomenon, called classic or early preconditioning , is transient and lost over the ensuing 1–2 h. However, the preconditioning stimulus further initiates and triggers a second long-lasting genomic response leading to increased de novo protein synthesis. 2,3Although less protective than early preconditioning, this delayed cardiac protection, also called late preconditioning or the second window of protection , may be effective for up to 72 h and thus be more promising in terms of therapeutic implications.
Although many similarities exist between ischemic (IPC) and anesthetic preconditioning (APC), fundamental differences with respect to key signaling steps are increasingly emerging. These regard activation and translocation of protein kinase C isoforms to subcellular targets 4as well as the role of mitogen-activated protein kinases in triggering and mediating IPC- and APC-induced cardioprotection. 5Moreover, differences in the memory phase of early preconditioning were previously reported for both types of preconditioning, suggesting differential signal transduction pathways. 1Finally, the results of three recent studies addressing the question of whether APC would be capable of eliciting late preconditioning similar to IPC were contradictory. No delayed cardiac protection was observed in a dog study 24 h after sustained exposure to isoflurane, 6whereas a delayed cardioprotective role of isoflurane was found in a rabbit model 7and against cytokine-induced injury in endothelial and vascular smooth muscle cells. 8
Gene expression profiling is an emerging type of research and enables investigators to comprehensively assess the molecular response of thousands of genes across the entire genome with the potential to dissect complex physiologic processes such as the phenomenon of preconditioning. 9–11To date, no data are available with respect to trigger-dependent changes in gene expression patterns in APC versus IPC. Therefore, the current study served (1) to determine the molecular portraits of APC and IPC at the transcriptional level and (2) to discover transcriptome variability between volatile anesthetic- and ischemia-induced cardiac preconditioning. Specifically, it was hypothesized that IPC and APC would activate distinct protective and injury-related genetic programs. For this purpose, the powerful genetic technology of microarrays was applied.
The data collected in this investigation not only unravel a high number of in-parallel up-/down-regulated known and unknown genes in APC and IPC, but they also provide evidence for significant trigger-dependent transcriptome variability. Furthermore, they may serve as a new platform for designing future hypothesis-driven research projects in search of cardioprotective mechanisms.
Materials and Methods
This study conformed to the guidelines of the Animal Care and Use Committee of the University of Zurich (Zurich, Switzerland).
Perfusion Protocols of Isolated Rat Hearts
Hearts from heparinized (500 U intraperitoneal) and decapitated male Wistar rats (weight, 250 g) were quickly removed and perfused in a noncirculating Langendorff apparatus with Krebs-Henseleit buffer containing the following ingredients: 155 mm Na+, 5.6 mm K+, 138 mm Cl−, 2.1 mm Ca2+, 1.2 mm PO43−, 25 mm HCO3−, 0.56 mm Mg2+, 11 mm glucose. The buffer was saturated with 95% O2–5% CO2(pH = 7.4, 37°C). Hearts were perfused at a constant pressure of 80 mmHg. A water-filled balloon-tipped catheter was inserted into the left ventricle through the left atrium. Left-ventricular end-diastolic pressure was adjusted to 3–7 mmHg during the initial equilibration, and the volume of the balloon was not changed thereafter. The distal end of the catheter was connected to a performance analyzer (Plugsys Modular System; Hugo Sachs, March-Hugstetten, Germany) by way of a pressure transducer. Perfusion pressure, epicardial electrocardiogram, and coronary flow (Transit Time Flowmeter type 700; Hugo Sachs) were simultaneously recorded on the same performance analyzer. All recorded data were digitized and processed on a personal computer using the software IsoHeart (Hugo-Sachs). Hearts were allowed to equilibrate for 10 min and to beat spontaneously in all experiments. The following experimental groups were defined: APC stimulus, IPC stimulus, ischemia, and control. For each experimental group, five hearts were prepared, and functional parameters were recorded. To obtain a strong stimulus on gene expression, APC was induced by administration of 110 min of isoflurane at 2.1% (vol/vol; 1.5 minimum alveolar concentration [MAC] in rats) (fig. 1), as previously described. 6–8The buffer solution was equilibrated with isoflurane using an Isotec 3 vaporizer (Datex-Ohmeda, Tewksbury, MA) with an air bubbler. The delivered vapor concentration of isoflurane was continuously controlled by the infrared gas analyzer Capnomac Ultima (Datex-Ohmeda). The applied concentration of isoflurane was also measured in the buffer solution using a gas chromatograph (Perkin-Elmer, Norwalk, CT): 2.1% isoflurane (vol/vol; 1.5 MAC in rats at 37°C), 0.52 ± 0.04 mm. IPC was induced by three episodes of ischemia interspersed by 5 min of reperfusion and followed by 85 min of reperfusion (fig. 1). In the ischemia group, hearts were subjected to 110 min of no-flow ischemia after 10 min of equilibration. The control group consisted of five hearts subjected to time-matched perfusion for 120 min.
Isolation of RNA and cDNA Synthesis
After completion of the protocols, left ventricular tissue was quickly frozen in liquid nitrogen and stored at −80°C. Frozen heart tissue (100 mg) was powdered in liquid nitrogen and homogenized in TRIzol LS reagent (Invitrogen, Basel, Switzerland) and chloroform–isoamyl alcohol (Fluka, Buchs, Switzerland). After centrifugation, the aqueous phase was mixed with isopropanol and precipitated overnight at −20°C. The pellet was washed with isopropanol, dried at 37°C, and eluted in diethyl pyrocarbonate–treated water. Single-strand complementary DNA (cDNA) synthesis from total RNA was performed using Superscript II (Invitrogen) in the presence of the T7-(T) 24 RNA polymerase promoter primer (Microsynth GmbH, Balgach, Switzerland). Double-strand cDNA was synthesized with a Superscript kit (Invitrogen). Biotin-labeled antisense cRNA was synthesized in vitro using a high-yield RNA transcript labeling kit (BioArray; Enzo, Farmingdale, NY).
Oligonucleotide Array Hybridization and Scanning
This project was conducted in close collaboration with the Functional Genomics Center of the University of Zurich. Gene expression profiling was performed with Affymetrix Rat Genome U34A array (Affymetrix, Santa Clara, CA). Gene chip U34A contains a total of 8,799 probe sets representing approximately 7,000 known rat genes and 1,000 expressed sequence tags (ESTs; unknown genes). Five independent gene chips for each group were used so that a total of 20 gene chips were analyzed. Importantly, no messenger RNA (mRNA) samples of different hearts were pooled. The biotin-labeled cRNA was fragmented in fragmentation buffer (200 mm Tris-acetate, 50 mm KOAc, and 150 mm MgOAc, at pH 8.1) and hybridized to the microarray in hybridization solution containing 15 μg fragmented cRNA in MES buffer (0.1 m MES, 1.0 m NaCl, and 0.01% Triton X-100, at pH 6.7) and herring sperm DNA. All arrays were placed in a hybridization oven at 60 rpm and 45°C for 16 h. Afterward, arrays were repeatedly washed at 22°C with SSPE-T (0.9 m NaCl, 60 mm NaH2PO4, 6 mm EDTA, and 0.005% Triton X-100, at pH 7.6), and subsequently with 0.1 MES at 45°C for 30 min. The gene chips were then stained with a streptavidin–phycoerythrin conjugate (Molecular Probes, Leiden, The Netherlands) and washed. To enhance the signals, the arrays were further stained with antistreptavidin antibody followed by staining with a streptavidin–phycoerythrin conjugate and again washed. All gene chips were scanned at a resolution of 3 μm, using a specifically designed confocal scanner (model 900154; Affymetrix).
Data Analysis
Briefly, for each gene, 16–20 probes with “perfect match” 25-bp oligonucleotides and 16 probes with paired “mismatched” oligonucleotides (designed with a single mismatch in the center position) were used to measure the transcript level of the gene (low, moderate, and high expression). Comparison of the hybridization signal from the perfect match and mismatched probes allows elimination of nonspecific cross-hybridization signals after background subtraction. Image data were analyzed using Affymetrix Microarray Analysis Suite detection algorithm (version 5.0), which assesses probe pair saturation, calculates a detection P value, and assigns a low, moderate, or high expression level to each individual probe set. Average differences between perfect match and mismatch signals were used to calculate the relative mRNA levels of transcripts. A global scaling method was applied to normalize signals in each gene chip minimizing the overall variability in hybridization intensities. The purpose of this normalization is to minimize the systemic variations in the measured gene expression levels and to allow the comparisons across different gene chips. Scaled data were further normalized by dividing all raw measurements for each gene by the mean raw signal intensity for the same gene as obtained in the five control hearts (normalized ratio intensity value, fold change) (GeneSpring, version 5.1; Silicon Genetics, Redwood City, CA). Significantly differentially regulated genes were defined as those genes that had a normalized ratio intensity of 2.0 or greater, i.e. , 2.0-fold up-/down-regulation in expression level of replicate experiments compared with the control group. The stringent criteria of a twofold cutoff and a P value less than 0.0125 were selected to reduce false-positive findings.
Verification of Candidate Gene Expression Using Quantitative Real-time RT-PCR
To confirm the gene expression patterns of microarray hybridization, reverse-transcription (RT) polymerase chain reaction (PCR) was performed for nine selected genes, of which the expression level was significantly altered (≥ 2.0-fold) by APC, IPC, or ischemia, respectively. The forward and reverse sequences of the used primers are listed in table 1. For each gene-specific amplification, 20 μl cDNA was diluted in water (1:10) before use as a template for the QuantiTect SybrGreen RT-PCR kit (Qiagen, Hilden, Germany). RT-PCR quantification and determination of expression levels were performed on an ABI Prism 7700 Sequence Detector Real-Time PCR machine (Perkin-Elmer, Foster City, CA). Amplification reactions were conducted with an initial step at 90°C for 3 min followed by 20–35 cycles. All PCR reactions were performed in triplicate, and α tubulin was used as a reference control. In addition, the predicted size of PCR products was confirmed by agarose gel electrophoresis.
Statistical Analysis
Data are expressed as mean ± SD. Functional parameters at identical time points and signal intensities of individual transcripts from microarray analyses were compared between groups by unpaired t tests. A post hoc Bonferroni test for multiple comparisons was performed to determine statistical significance. Accordingly, P < 0.0125 was considered to be statistically significant. SigmaStat (version 2.0; SPSS Science, Chicago, IL), Affymetrix Microarray Analysis Suite (version 5), and GeneSpring (version 5.1) were used for the statistical analysis.
Results
Functional Parameters in APC- and IPC-triggering Protocols
The characteristic hemodynamic profiles of the triggering protocols and the respective control groups are depicted in figure 2. Left ventricular developed pressure was significantly decreased and coronary flow was increased in APC hearts, whereas IPC hearts exhibited short transient increases in left ventricular end-diastolic pressure and coronary flow (fig. 2). Prolonged normothermic ischemia was associated with a marked increase in left ventricular end-diastolic pressure.
APC and IPC Elicit Profound Changes in Gene Expression Pattern in Isolated Perfused Rat Hearts
Of a total of 8,799 probe sets displayed on the Affymetrix U34A microarray chip, 402 were significantly (≥ 2.0-fold) altered in APC as compared with the control group, of which 116 (29%) represented ESTs (unknown genes). Two hundred eighty-nine were significantly (≥ 2.0-fold) regulated in IPC (containing 60 ESTs [21%]), and 78 were significantly altered in the ischemia group (containing 18 ESTs [23%]). APC up-regulated 217 genes (containing 95 ESTs [44%]) and down-regulated 185 genes (containing 21 ESTs [11%]) by 2.0-fold or greater. IPC up-regulated 234 genes (containing 49 ESTs [20%]) and down-regulated 55 genes (containing 11 ESTs [20%]) by 2.0-fold or greater. Prolonged ischemia up-regulated 29 genes (containing 9 ESTs [31%]) and down-regulated 49 genes (containing 9 ESTs [18%]) by 2.0-fold or greater. A graphical representation of the gene regulation between the various groups is shown as scatter plots in figure 3. Interestingly, highly expressed genes were less regulated by IPC, APC, or even prolonged normothermic ischemia than genes with a moderate or low expression level. Collectively, APC and IPC triggered profound modulations of gene expression in cardiac tissue after only a short time.
APC and IPC Display Significant Overlapping in Up-/Down-regulated Genes
Anesthetic preconditioning and IPC commonly up-regulated 39 genes (containing 10 ESTs [26%]) and jointly down-regulated 17 genes (containing 2 ESTs [12%]) by 2.0-fold or greater (fig. 4). Many fewer up-/down-regulated genes were observed between APC versus ischemia and IPC versus ischemia, respectively. APC and ischemia commonly up-regulated two genes (no ESTs) and concomitantly down-regulated six genes (containing 1 EST [16%]) by 2.0-fold or greater, whereas IPC and ischemia commonly up-regulated four genes (no ESTs) and jointly down-regulated five genes (containing 2 ESTs [40%]) by 2.0-fold or greater (fig. 4). Notably, only two genes were jointly up-regulated by 2.0-fold or greater in all three treatments (APC, IPC, and ischemia), but none of the transcripts was commonly down-regulated.
APC and IPC Exhibit Transcriptome Variability Consistent with Two Distinct Cardioprotective Phenotypes
Representative genes up-/down-regulated by 2.0-fold or greater are listed in tables 2–7. Genes were sorted according to the six functional clusters: (1) cell defense/death, (2) cell structure/motility, (3) growth/remodeling/inflammatory response, (4) metabolism, (5) redox state/respiratory chain, and (6) signaling/communication. To confirm the results of microarray hybridization experiments, RT-PCR was used as an independent method of measuring gene expression levels. RT-PCR results of nine selected genes revealed an almost perfect concordance with microarray data and hence clearly confirmed the accuracy of the gene chip data.
Anesthetic preconditioning and IPC showed significant changes in all of the clusters. APC and IPC commonly up-regulated the cytoprotective heat shock protein (Hsp) 10. In contrast, APC down-regulated Hsp-70 precursor and Hsp-90, whereas IPC up-regulated Hsp-70. Both APC and IPC down-regulated Bcl-xS, a proapoptotic protein, whereas prolonged ischemia did not, and commonly up-regulated proteasome, a product involved in cell protection. Conversely, APC exclusively increased Hsp-40 and defender against cell death 1 protein, whereas IPC decreased proapoptotic GATA-binding protein 4. Interestingly, APC but not IPC increased gene expression of proapoptotic protein programmed cell death 8. Similarities and disparities were also observed in the redox state/respiratory chain functional cluster. Antioxidant catalase and cytochrome p450 were exclusively up-regulated in APC, but glutathione peroxidase 1 and rhodanese were up-regulated in IPC. Antioxidant aldose reductase, glutathione S-transferase, and ubiquinone nicotinamide adenine dinucleotide dehydrogenase were up-regulated in both types of preconditioning. Prominent changes in metabolic enzymes and growth-related factors were present in APC and IPC. Profound differences between APC and IPC were also observed in signaling components. Whereas both APC and IPC increased receptor tyrosine phosphatase and nonreceptor tyrosine kinase and decreased RhoA-serine/threonine kinase α and inducible nitric oxide synthase, only APC increased expression of p38 mitogen-activated protein kinase α and decreased phosphodiesterase 4B and endothelial nitric oxide synthase expression levels.
APC and IPC Promote Expression of Gene Products Involved in Delayed Cardioprotection
To address the question of whether gene products previously reported to participate in delayed cardioprotection would be differentially regulated in APC and IPC, microarray data were also examined for genes with changes of less than 2.0-fold. The use of change of 2.0-fold or greater may result in neglecting genes with less than 2.0-fold but statistically significant changes. Figure 5depicts microarray data for nuclear factor κB, cyclooxygenase 2, manganese superoxide dismutase, inducible nitric oxide synthase, aldose reductase, Bcl-xS, Hsp-27, and Hsp-70 in all groups. Whereas cyclooxygenase 2 and manganese superoxide dismutase were not regulated by APC and IPC at the investigated time point, both types of preconditioning up-regulated aldose-reductase and down-regulated Bcl-xS. In contrast, Hsp-27 and Hsp-70 were solely up-regulated in IPC. Nuclear factor κB was significantly up-regulated in APC but also showed a tendency toward up-regulation in IPC. These data suggest that APC might be capable to promote the development of delayed cardioprotection as observed in IPC. This notion is further supported by the fact that APC, similarly to IPC, up-regulated a high number of ribosomal proteins, albeit to a lesser degree (7 in APC vs. 38 in IPC), increasing the efficiency of the cellular machinery for protein synthesis. Interestingly, 6 ribosomal proteins were similarly up-regulated in APC and IPC.
Discussion
The principal new findings of this comprehensive study are as follows. First, administration of isoflurane at clinically relevant concentrations profoundly alters the genetic program in the myocardium, comparable with brief ischemic episodes. This change in gene expression pattern directly reflects the pharmacologic power of volatile anesthetics and further implies sustained biologic effects by this class of drugs. Second, APC and IPC exhibit a high number of commonly regulated genes, including those associated with cell defense, structure and motility, growth and remodeling, metabolism, redox state, and signaling. In contrast, they share few similarities with gene regulation under prolonged ischemia. Surprisingly, the degree of concordance with sustained ischemia was not higher in IPC (nine transcripts), although containing a 15-min exposure to ischemia, than in APC (eight transcripts). Moreover, APC and IPC similarly modulated transcription of those gene products that were previously reported to be involved in delayed cardioprotection, providing some molecular evidence for a possible, albeit less pronounced, development of delayed protection in APC. Third, despite these similarities, fundamental differences with respect to protective and antiprotective genes were identified in APC and IPC and imply distinct cardioprotective mechanisms. Among the most prominent differentially regulated genes were Hsp-27 and Hsp-70, antioxidant proteins, and signaling components. Finally, numerous regulated but unknown genes were discovered in APC. These genes may be valuable candidates to expand and foster a more integrative understanding of APC in the future and may serve to generate new hypotheses in search of cardioprotective mechanisms in APC.
Gene Chip Analysis
Recently, attention has been directed toward cDNA microarray technology, which offers a genetic approach to explore molecular mechanisms on a genomic-wide scale. 9,10Microarray technology makes it possible to study simultaneously thousands of genes in the same sample. In the current study, Affymetrix gene chip U34A, containing more than 8,000 known and unknown genes, was used to characterize the expression profiles after administration of APC and IPC triggers. Using rigorous and complex statistical analysis, 9the expression levels of the individual genes were determined by means of their specific hybridization characteristics. Notably, in this study, five chips for each group were used to overcome inevitable variability in expression levels across experiments and to minimize false-positive results. No data from different experiments were pooled. Moreover, microarray data from selected genes could be clearly confirmed by the independent method of RT-PCR highlighting the accuracy of data obtained by gene chip analysis.
Activation of Distinct Protective Gene Programs by APC and IPC
Activation of a genetic survival program occurred in both preconditioning protocols. Heat shock proteins act as protein chaperones, protecting vital protein structure and function. 12They are strategically located in the cytoplasm, peroxisomes, mitochondria, and endoplasmic reticulum. Whereas Hsp-27 scavenges cytochrome c, inhibits autoactivation of caspase 3, and blocks Fas/Fas ligand proapoptotic pathway, 13Hsp-70 in collaboration with Hsp-40 prevents mitochondrial cytochrome c release and inhibits caspase 9 activation via Apaf-1, 14thereby protecting against cell death. Another chaperone, Hsp-10, is predominantly located in mitochondria and acts as a functional partner of Hsp-60, opposing the deleterious effects of cell death promoting Bax. 15In our study, APC and IPC similarly up-regulated Hsp-10. Unlike APC, IPC exclusively up-regulated Hsp-27 and Hsp-70, which was previously reported to also prevent caspase-independent cell death by c-Jun N-terminal kinase activation. 16Unexpectedly, APC down-regulated Hsp-70 precursor and Hsp-90, which, however, may reflect the prolonged application of the preconditioning stimulus in APC.
Bcl-2 family members regulate the delicate balance between cell death and survival. They are divided by function and homologous sequences into proapoptotic and antiapoptotic members. Proapoptotic Bcl-2 family proteins such as Bcl-xSare normally sequestered at the endoplasmic reticulum and nuclear membrane, and on noxious stimulation, they translocate to mitochondria, where they facilitate pore formation and the release of apoptogenic substances. 17Our data now provide evidence that APC and IPC markedly down-regulate antiprotective Bcl-xS. Intriguingly, the proteasome, a major machine for proapoptotic protein degradation, was also up-regulated in both types of preconditioning. 18Also, AP-1, a component previously linked to cell death, was repressed in APC and IPC. 19In contrast to IPC, APC differentially regulated several other cell defense–related factors, including programmed cell death 8 and defender against apoptotic death 1 (Dad1). 20The significance of these factors is poorly defined in IPC and not yet determined in APC and clearly requires further elucidation. Collectively, changes in these critical proteins will clearly affect the final outcome after prolonged ischemic injury.
The production of reactive oxygen species is both essential and detrimental to life. Previous studies have shown that oxidative signaling is intricately involved in the establishment of the preconditioned state. 21,22However, delayed preconditioning induces a marked cellular defense against oxidative damage to proteins. The most striking common changes in the expression of antioxidant proteins were the increases in the chemoprotective enzymes glutathione S-transferase, which was down-regulated in prolonged ischemia, and aldose reductase. Aldose reductase is a member of the aldo-keto reductase superfamily, metabolizes toxic aldehydes generated by lipid oxidation, and is an obligatory mediator of delayed protection in IPC. 23Despite these similarities between APC and IPC, the results of this study now indicate that APC- and IPC-induced cardioprotective phenotypes may differentially cope with oxidative stress. APC up-regulates catalase and cytochrome p450, whereas IPC up-regulates glutathione peroxidase 1 and rhodanese.
Characteristic patterns of growth-related and inflammatory response–related factors were modulated in IPC and APC. The importance of cytokines in the second window of IPC was previously established. 24APC induced tumor necrosis factor member 13 and decreased—among other growth-related factors—endothelial selectin, platelet–endothelial cell adhesion molecule, and vascular endothelial growth factor. Alterations in these factors may be implicated in recently reported APC-induced decreased adhesion of leukocytes after ischemia–reperfusion. 25In contrast, IPC up-regulated insulin-like growth factor protein 3, which may inhibit the effects of insulin-like growth factor I. 26Both types of preconditioning up-regulated the antiinflammatory serine protease inhibitor.
High rates of fatty acid oxidation significantly contribute to myocardial ischemic injury. In accordance with this concept, APC and IPC rather repressed key enzymes of cellular β oxidation and increased expression of enzymes involved in glycolysis. In contrast, APC and IPC increased the expression level of carnitine palmitoyltransferase I, a key regulator of the long-chain fatty acid oxidation. Interestingly, inhibition of carnitine palmitoyltransferase I was previously shown to augment palmitate-induced myocyte apoptosis. 27
Molecular Evidence for Delayed Cardioprotection by APC
In the current study, several factors previously reported to be involved in delayed cardiac protection were evaluated with respect to their transcriptional response on APC and IPC triggering. Apart from Bcl-2 family members and aldose reductase, these comprise nitric oxide, 28inducible nitric oxide synthase, 29cyclooxygenase 2, 30protein kinase C, 31Src family of protein tyrosine kinases, 32and nuclear factor κB. 33Although the interval between the application of the trigger stimuli and the measurement of the transcriptional response was limited in the model used and may not have allowed the detection of a fully developed delayed cardioprotective phenotype, the data of this study are consistent with the development of some sort of delayed protection in APC. Das et al. 34previously evaluated the adaptation of cellular defenses after one or four cycles of 5 min of ischemia interspersed by 5 min of reperfusion in a buffer-perfused rat heart model. Using the highly sensitive Northern hybridization technique for a limited number of transcripts, they found up-regulation of Hsp-27, Hsp-70, Hsp-89, catalase, manganese superoxide dismutase, and glutathione reductase after four preconditioning cycles but not after a single cycle. In our study, a significant increase in nuclear factor κB expression was present in APC and there was a clear, albeit nonsignificant, tendency to increased levels in IPC. The oxidant-sensitive transcription factor nuclear factor κB is involved in the early activation of multiple genes encoding defense proteins and was repeatedly shown to be critically involved in delayed cardioprotection. 35Nuclear factor κB directly regulates the expression of inducible nitric oxide synthase 36and cyclooxygenase 2, 37two important mediators of delayed protection. The fact that nuclear factor κB exhibits a biphasic expression activity on oxidative stress, with an initial transcriptional inhibition followed by a marked activation, 37may explain the observed down-regulation of inducible nitric oxide synthase at the early investigated time point in our study. From a biologic point of view, decreased inducible nitric oxide synthase activity, as observed at this early stage after triggering, may rather reflect a protective mechanism at the transcriptional level associated with early preconditioning. Previous studies have clearly linked increased inducible nitric oxide synthase activity to enhanced cell death. 38In accordance with this concept is the notion that actinomycin D, an inhibitor known to be effective at the transcriptional level, was previously reported to diminish protection elicited by early IPC. 39Many transcripts for ribosomal proteins were up-regulated in both types of preconditioning, albeit in IPC significantly more than in APC. To date, it is unclear as to why so many ribosomal proteins were up-regulated after preconditioning. It may be speculated that this genomic response simply signifies more ribosomal biogenesis in response to the “damaging” aspects of preconditioning. Alternatively, global increase in ribosomal protein expression may mirror enhanced protein synthesis heralding the second window of protection. Collectively, APC and IPC enhance the transcription of gene products associated with delayed protection. Failure to detect regulation of some critical mediators of delayed protection in our study may be a result of the relatively short time interval between the application of the trigger stimuli and the determination of expression levels.
Transcriptional Changes in Signaling Components
We were recently able to visualize translocation of protein kinase C isoform to subcellular targets in IPC and APC. 4Moreover, our laboratory provided the first evidence that mitogen-activated protein kinases differentially act in triggering and mediating IPC and APC. 5We now extend these findings by showing that differences between IPC and APC also occur at the transcriptional level. Whereas APC solely increased the expression of the injury-related p38 mitogen-activated protein kinase α, APC and IPC repressed p38 mitogen-activated protein kinase–activated protein kinase 2. Although previous reports showed increased gene expression of p38 mitogen-activated protein kinase–activated protein kinase 2 after IPC, 40our observation may reflect the biphasic regulation of this transcript. Interestingly, protein kinase C-δ was exclusively up-regulated in IPC. As a common feature in APC and IPC, receptor-bound tyrosine phosphatase and nonreceptor tyrosine kinase were up-regulated, whereas RhoA-binding serine/threonine kinase α was repressed. Again, these findings extend at a transcriptional level previous observations of complex interactions between protein tyrosine kinases and their counterplayers in APC. 41
A recent study in isolated rat hearts reported transient increases of B-type natriuretic peptide (BNP) and reported infarct-size limiting effects of this hormone via opening of ATP-sensitive potassium channels. 42In our study, no significant transcriptional regulation of atrial natriuretic peptide was detected after triggering. Also, BNP expression was unaffected by IPC and APC. These findings suggest that BNP expression is not affected by administration of preconditioning triggers, strengthening its diagnostic role as a sensitive marker for cardiac dysfunction after a prolonged ischemic hit, as recently demonstrated in a clinical preconditioning study. 43
Gene Chip Analysis in Cardiac Preconditioning
To date, only one study has been conducted comparing gene expression profiles between pharmacologic and ischemia-induced preconditioning. Rokosh et al. 11compared ischemic with nitric oxide–induced preconditioning in mice using oligonucleotide microarrays containing more than 6,000 known and 6,000 unknown genes. In this study, 12% of the IPC-regulated genes were concordant with pharmacologic preconditioning by nitric oxide, which is similar to our observation, with 19% of IPC-regulated genes being commonly regulated in APC. Moreover, a number of genes such as serine protease inhibitor or nicotinamide adenine dinucleotide–ubiquinone oxidoreductase B15, previously reported to be regulated by IPC, were confirmed in the current study. 44However, identification of all unknown genes in IPC and APC and functional studies of candidate genes using specific blockers or selective knockout will be necessary to precisely determine their roles.
Study Limitations
The following remarks should be added: (1) A left ventricular balloon was used to monitor left ventricular function during the preconditioning process. However, this may jeopardize coronary perfusion and subsequently alter gene expression. Differential alterations in gene expression pattern were previously reported in subendocardium versus subepicardium, implying the existence of a spatial gradient. 45(2) Because of the limited long-term biologic stability of the isolated heart preparation, the cardiac tissue was analyzed after a total of 120 min. This time was sufficient to allow up-/down-regulation of genes. However, the interval might have been too short to develop the full range of genetic responses. Also, no time course of gene expression profiling was established. (3) Isoflurane was administered over a more extended period than in most previously reported preconditioning protocols. However, this approach reflects the clinical situation in which patients are treated for hours with volatile anesthetics and may enhance the preconditioning stimulus to induce significant transcriptional changes. 7,8(4) Finally, buffer perfusion is an artificial condition, and isolated hearts may undergo a short ischemic period.
In conclusion, the current study provides the first comprehensive gene expression profile associated with APC as compared with IPC. Among many similarities, the data collected suggest distinct differences in gene expression between APC and IPC. Understanding the function of unknown genes in the complex biologic process of pharmacologic and ischemic cardiac preconditioning will open new avenues for optimizing perioperative cardioprotection in humans.