Microarray and proteomic analysis of the cardioprotective effects of cold blood cardioplegia in the mature and aged male and female

Abstract

Recently we have shown that the cardioprotection afforded by cardioplegia is modulated by age and gender and is significantly decreased in the aged female. In this report we use microarray and proteomic analyses to identify transcriptomic and proteomic alterations affecting cardioprotection using cold blood cardioplegia in the mature and aged male and female heart. Mature and aged male and female New Zealand White rabbits were used for in situ blood perfused cardiopulmonary bypass. Control hearts received 30 min sham ischemia and 120 min sham reperfusion. Global ischemia (GI) hearts received 30 min of GI achieved by cross-clamping of the aorta. Cardioplegia (CP) hearts received cold blood cardioplegia prior to GI. Following 30 min of GI the hearts were reperfused for 120 min and then used for RNA and protein isolation. Microarray and proteomic analyses were performed. Functional enrichment analysis showed that mitochondrial dysfunction, oxidative phosphorylation and calcium signaling pathways were significantly enriched in all experimental groups. Glycolysis/gluconeogenesis and the pentose phosphate pathway were significantly changed in the aged male only (P < 0.05), while glyoxylate/dicarboxylate metabolism was significant in the aged female only (P < 0.05). Our data show that specific pathways associated with the mitochondrion modulate cardioprotection with CP in the aged and specifically in the aged female. The alteration of these pathways significantly contributes to decreased myocardial functional recovery and myonecrosis following ischemia and may be modulated to allow for enhanced cardioprotection in the aged and specifically in the aged female.

the present paradigm to alleviate surgically induced ischemia/reperfusion injury requires the use of cardioplegia (CP) solutions for the rapid electromechanical arrest of the myocardium. Most CP solutions use high potassium to depolarize the heart however, the effects of these solutions are not optimal. In a series of studies we have shown that the cardioprotective efficacy of CP varies in the mature and aged male and female heart (28, 29). Analysis of age and sex and age by sex effects on the cardioprotection afforded by CP indicated that postischemic functional recovery was significantly decreased and infarct size was significantly increased in the aged compared with the mature heart and was significantly increased in the aged female compared with the aged male heart (28, 29).

These data agree with studies showing women have a significantly greater operative risk compared with men (1) and that women have worse outcomes after cardiac surgery (19, 42). There is now sufficient evidence to show that after adjusting for all comorbidities, including body surface area, female sex is an independent predictor of increased mortality following coronary artery bypass surgery (3, 17, 18).

Taken together, these data demonstrate that currently optimized cardioprotective protocols effective in the mature and aged male heart may not be as efficacious in the aged female heart and that further study as to the mechanisms regulating cardioprotection in the aged female are required.

In previous studies we and others have shown that there are significant alterations in RNA and protein and that these changes are significantly greater in the aged compared with the mature heart and that specific changes were associated with the aged female (9, 26).

In this report we use microarray and high-throughput proteomic analysis to identify functionally enriched pathways associated with cardioprotection in the mature and aged male and female heart with specific emphasis on identification of the pathways limiting cardioprotection in the aged female.

METHODS

Animal model.

Mature (sexually mature, 15–20 wk; 3–4 kg) and aged (not senescent, >32 mo; 5–6 kg) male and female New Zealand White rabbits were used for in situ cardiopulmonary bypass. Rabbits were obtained from Millbrook Farm (Amherst, MA). All experiments were approved by the Beth Israel Deaconess Medical Center Animal Care and Use Committee and conformed to the US National Institutes of Health (NIH) guidelines regulating the care and use of laboratory animals (NIH publication no. 5377-3, 1996). All research was performed in accordance to the American Physiological Society Guiding Principles in the Care and Use of Animals.

Justification of animal model.

The rabbit heart provides a reproducible experimental model with documented relevance to human disease, not represented in rodent models (33). The rabbit heart closely parallels the human heart, with similar myosin heavy chain phenotype, metabolic rate, and lacking myocardial xanthine oxidase (7, 14). The use of 38 mo of age rabbits provides a model that we have previously shown to have significantly increased infarct size, significantly decreased postischemic functional recovery, significantly decreased mitochondrial function, and high energy phosphate preservation and resynthesis following ischemia/reperfusion compared with rabbits 12–15 wk of age (27–29).

In situ blood perfused heart model.

Rabbits (n = 72) were sedated with acepromazine (0.75 mg/kg im), and a 21-gage butterfly catheter was placed in the ear and looped and secured in place with tape and vet wrap. Ketamine (35 mg/kg) and xylazine (2.5 mg/kg) IV were administered via the catheter, which was also used intraoperatively to administer heparin (3 mg/kg iv), lidocaine (1% solution, 5 ml) and lactated Ringer's solution (10 ml·kg⁻¹·h⁻¹).

The surgical site was shaved and prepped with Betadine solution and 70% isopropyl alcohol, each applied in triplicate and patted dry with sterile gauze pads, and the entire animal was draped with sterile towels (except for the surgical sites). The larynx was anesthetized with a 1% lidocaine solution and the rabbit was intubated with an uncuffed endotracheal tube (pediatric size 2-0 or 3-0 ID), and the rabbit was placed on mechanical ventilation (oxygen 40%; tidal volume = 18 ml/kg; ventilation rate 16–18 breaths/min). Proper endotracheal tube placement was verified by auscultation and observation of condensation of the end of the tube. General anesthesia was induced with 3.0% isoflurane and maintained at 1.5% for the duration of the surgical procedure. The animal was secured on the operating room table with soft restraints.

Cardiopulmonary bypass.

A medial sternotomy was performed, the heart was exposed, and the pericardial sac opened. The cardiopulmonary bypass circuit was achieved using a Prolene (6-0) cardiovascular suture to form a single purse-string layer in the ascending aorta. An incision was made within the purse string to allow insertion of an aortic cannula (3.3 mm). The cannula was secured with a purse-string tourniquet. A 14-Fr venous cannula was inserted in the right atrium via the auricular appendage. The cardiopulmonary circuit consisted of a roller pump (American Optical, Southbridge, MA) and a neonate membrane oxygenator (Sorin Group USA, St. Louis, MO). The circuit was primed with whole blood obtained from a donor rabbit. During cardiopulmonary bypass when oxygenation was no longer occurring via ventilator, propofol (0.5–0.7 mg·kg⁻¹·min⁻¹ iv) was continuously infused via the marginal ear vein.

Blood donor rabbits.

Donor rabbits (n = 8) were sedated and anesthetized as described above. A medial sternotomy was performed and the heart was exposed; a large-bore needle was inserted into the heart and blood was withdrawn into a 60 ml syringe. The heart was then removed, and the animal died by exsanguination. The blood was collected in blood bags containing citric acid (Baxter, Deerfield, IL) and used immediately to prime the cardiopulmonary bypass circuit. The blood volume in the priming solution was calculated for the final hematocrit of the priming volume to be 18–20%. Mannitol (15%) and sodium bicarbonate (20 meq/l) were added for pH adjustment. Blood gases and hematocrit were monitored every 10–15 min with a Corning 238 pH/blood gas analyzer and a Corning 270 Co-oximeter (Chiron Diagnostics, Emeryville, CA).

Treatment groups.

Three treatment groups were investigated (Fig. 1). Control treatment animals were sedated and anesthetized and received a sternotomy; the aorta, right atrium\auricular appendage, and the apex were sham-manipulated in the same manner as for global ischemia (GI) and cardioplegia hearts. Control hearts received no ischemia or cardioplegia. Control hearts (n = 6 each experimental group) were removed from the animal following 180 min (30 min equilibrium, 30 min sham ischemia, and 120 min sham reperfusion). GI hearts (n = 6 each experimental group) received 30 min of GI. GI was achieved by cross-clamping of the aorta. CP hearts (n = 6 each experimental group) received cold blood cardioplegia [Deaconess Surgical Associates (DSA) solution {K⁺, 60 mM; MgSO₄, 8 mM; dextrose, 2.5 mM; THAM [Tris(hydroxymethyl) aminomethane], 10 mM in normal (0.9%) saline and 50 μM diazoxide (DZX)} mixed (1:4; vol:vol) with cold blood]. DZX was dissolved in dimethyl sulfoxide (DMSO, Fisher Scientific, Fair Lawn, NJ) before being added to DSA cardioplegia. The final concentration of DMSO was <0.1%. DMSO was added to control and GI hearts at the same concentration. CP was administered antegrade through the aortic root (20 ml/kg) to induce cardiac arrest. Following 30 min of CP the cross-clamp was released, cardiopulmonary bypass was ceased, and the hearts were reperfused for 120 min.

Fig. 1.

Experimental protocol. Rabbits were sedated and subjected to cardiopulmonary bypass consisting of 30 min equilibrium, 30 min global ischemia (GI), and 120 min reperfusion. Control hearts were sham-manipulated and received no ischemia or cardioplegia (CP). GI hearts received 30 min of GI achieved by cross-clamping of the aorta. CP hearts received cold blood CP administered antegrade through the aortic root to induce cardiac arrest prior to ischemia. Following 30 min of GI the aortic cross-clamp was released, cardiopulmonary bypass was ceased, and the hearts were reperfused for 120 min. The heart was excised and placed in an ice-cold bath of Krebs-Ringer solution and was then subjected to a brief Langendorff retrograde perfusion to remove all blood, a possible source of contamination. The hearts were then either used immediately for measurement of infarct size or quick frozen in liquid nitrogen and used for RNA and protein isolation. Experimental groups are found at left, and time in minutes is found at bottom.

Functional measurements.

Myocardial function was assessed in n = 6 hearts by sonomicrometry using two orthogonal pairs of dimension crystals arranged perpendicularly on the left ventricle. Percent regional segment shortening was recorded continuously in both crystal sets throughout the study protocol. Percent regional segment shortening was calculated as [(end diastolic regional length − end systolic regional length) / end diastolic regional length] × 100% (40).

Measurement of infarct size.

Infarct Size [% left ventricular (LV) mass] was determined in n = 6 hearts for each experimental group using 1% TTC as previously described (29).

Langendorff perfusion.

To ensure that the heart tissue did not contain blood, a possible source of contamination in subsequent molecular isolation and analysis, following 120 min of reperfusion the heart was excised from the animal and placed in an ice-cold bath of 4°C Krebs-Ringer solution containing NaCl 100 mM, KCl 4.7 mM, KH₂PO₄ 1.1 mM, MgSO₄ 1.2 mM, NaHCO₃ 25 mM, CaCl₂ 1.7 mM, glucose 11.5 mM, pyruvic acid 4.9 mM, and fumaric acid 5.4 mM and equilibrated with 95% O₂ and 5% CO₂ (pH 7.4 at 37°C), where spontaneous beating ceased within a few seconds. The aorta was cannulated with a polyethylene cannula, and the heart was subjected to Langendorff retrograde perfusion for 10 min through the aortic root with Krebs-Ringer solution to remove all blood. Langendorff perfusion was performed at a constant pressure of 75 cmH₂O in a water-jacketed chamber with myocardial temperature maintained at 37°C (29). The blood-free heart was removed from perfusion, and the atria, right ventricle, and large vessels removed. The LV was then quick frozen in liquid nitrogen and then stored in liquid nitrogen prior to the isolation of myocardial RNA and proteins from matched samples.

RNA and protein isolation.

Total tissue RNA was isolated, and quality and purity were assessed by spectrophotometric analysis and agarose gel electrophoresis as described previously (27). Total tissue ventricular myocardial protein was isolated and quality and purity was assessed by SDS-polyacrylamide gel electrophoresis (PAGE) as described previously (27).

Microarray and proteomic analysis.

All microarray and proteomic analyses were performed on matched samples (Fig. 2).

Fig. 2.

Experimental flow chart. Flow chart of transcriptomic and proteomic analysis. Transcriptomic and proteomic analyses were performed on matched samples.

Microarray analysis.

Microarray slide preparation/fabrication, print array, blocking, probe labeling and hybridization, scanning quality control, scanning and feature extraction, raw data analysis, and computational biology were performed at the Harvard FAS Center for Systems Biology.

Microarray analysis was performed as previously described using our nonredundant rabbit heart cDNA compendium consisting of 3,592 nonredundant cDNAs (mean insert size of 1.67 kb) (25). Putative identity was assigned only to cDNAs exhibiting matches to Homo sapiens or Oryctolagus cuniculus (rabbit) genes with E = 1e⁻²⁵ (minimum P = 10⁻¹⁰) and nucleotide sequence identity >95%. This approach allows for ultimate usage in humans, as 100% of our rabbit microarrays directly correspond to human, without the need for reisolation and sequence of the specific gene. This approach will allow for the rapid verification of specific RNAs in humans by RT-PCR as sequence identity, and thus oligomer sequence, is readily available. All cDNA sequences have been submitted to GenBank and are available under accession numbers EC618425-EC626095 and dbEST40150263-40157933 and at http://www.mccullylab.org-a.googlepages.com/.

Microarray data analysis.

The hybridized array images were analyzed with R and Bioconductor packages to extract normalized expression values for cDNA probes. The quality control analysis was performed with the “arrayQualityMetrics” package in R to identify the outliers. The high-quality arrays were further preprocessed and normalized using the Limma (Linear Models for Microarray Data) package from the Bioconductor project. The processed signal resulting from the scanner Genepix Feature Extraction software was read into Limma using the “read.maimages” function. The background corrected probe data was normalized using print-tip-loess and quantile normalization within and between arrays, respectively. The control and experimental groups for designed microarray experiments were compared by fitting a linear model for each gene and applying empirical Bayes smoothing to identify differentially expressed genes. The genes with P value <0.01 were considered significantly differentially expressed.

All microarray data discussed in this publication have been deposited in the National Center for Biotechnology Information's Gene Expression Omnibus (GEO) (8) and are accessible through GEO Series accession numbers GSE14576 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE14576) and GSE30261 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE30261).

Quantitative real-time reverse transcription-polymerase chain reaction.

Quantitative real-time reverse transcription-polymerase chain reaction (qRT-PCR) analysis was performed to validate changes in gene expression across experimental and treatment conditions using an Eppendorf Realplex Mastercycler and software package (Eppendorf North America, Westbury, NY). The iScript One-Step RT-PCR Kit with SYBR Green solution (Bio-Rad, Hercules, CA) was used for the qRT-PCR reactions according to manufacturer's instructions. In brief, 100 ng of total RNA and 600 nM of both forward and reverse primer were added to each reaction. Control, GI, and CP reactions were run for each primer set in triplicates. Control reactions without reverse transcriptase were also performed for each reaction.

Oligonucleotide primers (Table 1) were designed based on the laboratory's rabbit nonredundant cDNA library sequence data using Primer3 (http://frodo.wi.mit.edu/) and were synthesized by Oligos ETC., (Wilsonville, OR). Reaction kinetics were optimized for each primer set. Reverse transcription of RNA template occurred at 50°C for 10 min with inactivation at 95°C for 5 min. Amplification and detection occurred over 41 cycles (denaturing 95°C; 10 s, annealing 58–63°C; 10 s, extension 72°C; 20 s, with plate read at 78°C). Melting curves were performed for each reaction at the conclusion of the cycling parameters from 60–95°C. Fold changes in gene expression were calculated by the delta delta CT method (22).

Table 1.

Oligonucleotide primers used for real-time RT-PCR

Gene	GenBank Accession Number	Oligonucleotide Primers	Product Size, bp
18s	EC623896	Forward: 5′-atGGccGttcttaGttGGtG-3′	217
		Reverse: 5′-cGctGaGccaGtcaGtGtaG-3′
Succinate dehydrogenase complex, subunit A, flavoprotein	EC621662	Forward: 5′-caGacaGGaacccGtGattt-3′	200
		Reverse: 5′-aGctttGtGatGcatGctGt-3′
Voltage-dependent ion channel 2	EC619532	Forward: 5′-cccaGtttGtttccccctat-3′	242
		Reverse: 5′-GattGatccGatcctccaaa-3′
Cysteine- and glycine-rich protein 3	EC620325	Forward: 5′-GctcatGacaGcacGacaGt-3′	160
		Reverse: 5′-GGGactGttGGaactGGaGa-3′
ADP/ATP translocase 1	AB009386	Forward: 5′-GGcGctactttGctGGtaac-3′	157
		Reverse: 5′-atGatacaGtcGcccaGacc-3′
Isocitrate dehydrogenase 2	EC619860	Forward: 5′-tGGctcaGGtcctcaaGtct-3′	155
		Reverse: 5′-ctcaGcctcaatcGtcttcc-3′

Quantitative real-time RT-PCR genes analyzed, oligonucleotide primers, and product size.

Validation and identification of qRT-PCR products.

qRT-PCR products were verified by agarose gel and electrophoresis and the products were purified using the QIAquick PCR purification kit (QIAGEN, Valencia, CA) according to manufacturer's instructions. The purified products were sequenced using the BigDye Terminator v1.1 and v3.1 Cycle Sequencing Kits (Applied Biosystems, Foster City, CA) using the 3130 xl Genetic Analyzer (Applied Biosystems) (27).

Proteomic analysis.

High-throughput profiling of protein samples was performed at the Genomics and Proteomics Center Core facility of the Beth Israel Deaconess Medical Center using the 8-plex iTRAQ (AB Sciex, Foster City, CA) labeling protocol and standard MudPIT methodology coupled with the 4800 MALDI TOF/TOF Plus instrument to perform the mass spectrometry as previously described (27).

Unsupervised and quality control analysis of proteomics data.

The raw data were analyzed by the ProteinPilot v3.0 software (AB Sciex, Carlsbad, CA) using the Paragon algorithm (38), which enables rapid matching of MS/MS spectra and iTRAQ relative quantitation. Searches were performed against the latest available FASTA format SwissProt protein database [UniProtKB/Swiss-Prot release 2010_09 (10/07/2010) that was downloaded from the following website: http://www.uniprot.org/downloads]. Each protein received a confidence score, with the overall confidence score ranging from 0 to 100% based on the total and best ion score. The proteins with confidence score >90% and with at least one peptide of 95% identification confidence were used for further quality control and differential expression analysis. Each protein also achieved quantitative scores for each of the eight iTRAQ tags to calculate the relative expression levels. To compare results from one set of eight samples to the others, we compared all samples labeled with the iTRAQ 114, 115, 116, 117, 118, 119, and 121 to the iTRAQ 113-labeled reference sample (control sample). By using matched control samples we eliminated any constitutive protein expression and analyzed strictly the proteins expressed with ischemia and cardioplegia. Using the weighted average algorithm in the ProteinPilot software, we calculated the relative expression levels of proteins from the peptides relative abundance.

The quality control analysis was performed on the basis of relative expression values of different proteins to identify any outliers. The quality control analysis was performed using pair-wise correlation plots, box-plots, and principal component analysis (PCA). PCA projects multivariate data objects onto a lower dimensional space while retaining as much of the original variance as possible. This is necessary because in an analysis of proteomic data, due to a dimensionality problem, the number of proteins most often exceeds the number of samples by a considerable number. Each principal component is associated with an eigenvalue, which corresponds to the amount of variability explained by the corresponding principal component. An eigenvalue is a measure of strength of each component.

Supervised analysis.

To identify the differentially expressed proteins, the relative protein expression values were compared between groups (GI vs. control, CP vs. control, CP vs. GI). By using matched control samples we eliminated any constitutive protein expression and analyzed strictly the proteins expressed with ischemia and cardioplegia. Proteins were considered overexpressed in GI relative to control if the iTRAQ ratio of GI to control was >2.0 and if the corresponding maximum control to control ratio was less than the GI-to-control ratio. Similarly proteins were considered underexpressed in GI relative to control if the iTRAQ ratio of GI to control was <0.5 and if the corresponding minimum control-to-control ratio was higher than the GI-to-control ratio. Using the same method, we identified differentially expressed proteins for CP vs. control and CP vs. GI comparisons.

Gene ontology analysis.

Gene ontology (GO) analysis was performed using “GeneGO” (http://www.genego.com/). To carry out this analysis we used the MetaCore tool in GeneGO based on a proprietary manually curated database of human protein-protein, protein-DNA, protein-compound interactions, metabolic and signaling pathways, and the effects of bioactive molecules in gene expression. Each of the effected categories was provided with a P value as an indicator of the significance. The lower the P value the more significantly a category was affected. The P value was calculated using the formula for hypergeometric distribution. In this analysis categories with P values <0.05 were considered significantly affected.

Functional and pathway enrichment analysis.

GO categories in differentially expressed transcripts and proteins were identified using the Biological Processes and Molecular Functions Enrichment Analysis Database for Annotation, Visualization and Integrated Discovery (DAVID) (http://david.abcc.ncifcrf.gov/).

Canonical pathway enrichment analysis was performed using Ingenuity Pathways Analysis (IPA 4.0) (http://www.ingenuity.com/). All statistics and data analysis were performed by the Beth Israel Deaconess Medical Center Genetics Core of the Genomics and Proteomics Center.

Western blot analysis.

Western blot analysis was performed as previously described (39). In brief, protein samples (25 μg) were fractionated on 10% Novex Tris-Glycine gels (Invitrogen, Carlsbad, CA) and then electro-blotted to nitrocellulose membranes (Invitrogen). Protein equivalency, transfer efficiency, and membrane blocking were performed as previously described (39). Immunoblotting was performed with mouse monoclonal isocitrate dehydrogenase 2 (IDH2) antibody (1:2,000 dilution; Abcam, Cambridge, MA). Blots were detected using ECL-Plus (Amersham Pharmacia Biotech, Piscataway, NJ) with species-appropriate secondary antibodies. Densitometry analysis was performed using the Image J analysis software (http://rsbweb.nih.gov/ij/) (39).

Statistical analysis.

The number of experiments required for each group (n) was determined by Power Analysis, with an α = 0.05 (two-sided), probability of type 1 error, and β = 0.05, probability of type 2 error. Ignoring repeated measurements, six animals per group provide >85% power to detect a difference equal to twice the within group standard deviation (α = 0.05, two-sided), and 82% to detect a difference equal to 2.5 times the within group standard deviation (α = 0.0083, two-sided, the Bonferroni adjusted significance criterion for post hoc comparisons between four groups). For presence/absence of a finding, six animals per group would have 95% power to detect a factor occurring in at least 40% of the animals, while 10 animals per group would have 95% power to detect a factor occurring in at least 26% of the animals.

Statistical analysis for myocardial function, infarct size, qRT-PCR, and Western blot analysis was performed with SAS (version 6.12) software package (SAS Institute, Cary, NC). The mean ± SE of the mean for all data was calculated for all variables. Statistical differences between groups were evaluated by one-way ANOVA. Dunnett's test was used for comparisons between control and other groups to adjust for the multiplicity of tests.

RESULTS

In situ blood perfused heart.

Our results demonstrate that there was no significant difference in segmental shortening (SS) within or between any of the experimental groups during equilibrium. In control hearts, there was no significant difference in SS following sham ischemia and 120 min sham reperfusion compared with equilibrium within or between any of the treatment groups (Fig. 3). In GI hearts, SS was significantly decreased (P < 0.05 vs. equilibrium) following 30 min. ischemia and 120 min reperfusion in all experimental groups (Fig. 3). In CP hearts, SS following 30 min ischemia and 120 min. reperfusion was not significantly different than equilibrium in mature male, mature female, or aged male hearts. However, in aged female CP hearts SS was significantly decreased (P < 0.05) compared with equilibrium and was significantly decreased (P < 0.05) compared with all other experimental groups.

Fig. 3.

Segmental shortening. Segmental shortening (SS) was calculated as [(end diastolic regional length − end systolic regional length) / end diastolic regional length] × 100%. There was no significant difference in the control groups between equilibrium and reperfusion. SS was significantly decreased in GI with reperfusion (P < 0.05 vs. control) in all experimental groups. CP significantly enhanced SS in the mature male, mature female, and aged male with reperfusion (NS vs. control; P < 0.05 vs. GI). However, in the aged female SS was significantly decreased with reperfusion compared with control (P < 0.05 vs. control) and significantly decreased with reperfusion compared with the aged male (P < 0.05 vs. aged male). NS, not significant. *P < 0.05 vs. equilibrium, **P < 0.05 vs. aged male.

There was no significant difference in infarct size within or between any of the experimental groups in the control treatment receiving sham ischemia and sham reperfusion (Fig. 4). In GI, infarct size following 30 min ischemia and 120 min reperfusion was significantly increased (P < 0.05) in all experimental treatments compared with control (Fig. 4). Infarct sizes in the aged male and the aged female in GI were significantly increased (P < 0.05) compared with mature male and mature female.

Fig. 4.

Myocardial infarct size [% left ventricular (LV) mass]. Infarct size was determined using 1% TTC and expressed as a percentage of LV mass for each heart. There was no significant difference in the control groups. Infarct size was significantly increased (P < 0.05 vs. control) in GI and was significantly increased in aged compared with mature hearts (P < 0.05). CP significantly decreased infarct size (P < 0.05 vs. GI). There was no significant difference in infarct size between control and mature male or female with CP. However, infarct size in the aged hearts was significantly increased (P < 0.05 vs. control and CP mature). Infarct size was significantly increased in aged female with CP compared with aged male with CP (P < 0.05). *P < 0.05 vs. Equilibrium.

In CP, infarct sizes were significantly decreased (P < 0.05) in all experimental groups compared with GI (Fig. 4). CP significantly decreased infarct sizes in mature male and mature female experimental groups such that there was no significant difference compared with control. In contrast, CP hearts infarct sizes in aged male and aged female were significantly increased (P < 0.05) compared with control and compared with mature male and mature female CP groups (P < 0.05). Infarct size in the aged female in CP was significantly increased (P < 0.05) compared with the aged male. One animal was excluded from analysis of infarct size in the aged female CP group because it was an outlier (19.5% LV mass).

Age increased susceptibility to infarct size in ischemia; however, there was no difference between male and female. The effects of cardioprotection decrease susceptibility to ischemia in the mature and aged male and female; once again cardioprotection is greater in the mature than the aged. However, in the aged female the cardioprotection is significantly decreased compared with the aged male.

Microarray analysis.

Microarray analysis of 3,592 nonredundant cDNAs for mature and aged male and female hearts showed that there were no transcripts upregulated (>1.2-fold and P < 0.01) in GI compared with control, in agreement with our previous findings (28). No transcripts were commonly differentially expressed in all groups.

In the mature male, in CP compared with GI, there were 42 upregulated cDNAs and 54 downregulated cDNAs (>1.2-fold and P < 0.01). Functional annotation clustering revealed that fatty acid β-oxidation, fatty acid metabolism, and fatty acid oxidation were upregulated (enrichment score >1.4) in CP (28).

In the mature female comparison between CP and GI revealed 72 transcripts were upregulated and 13 downregulated (>1.2-fold and P < 0.01 for each); however, no functional annotation clusters were identified (Fig. 5A).

Fig. 5.

Differentially expressed genes in mature (A) and aged (B) females. The differentially expressed genes were identified by supervised analysis on the basis of P value <0.01 in each group; GI and CP groups were compared with control to account for constitutively expressed RNAs and CP vs. GI to show RNAs up- and downregulated in CP compared with GI. The log fold change (lFC) in gene expression is shown with pseudocolor scale (−3 to 3) with red denoting upregulation and green denoting downregulation. The columns represent lFC from comparisons and the rows represent the genes. Dendograms are shown on the left. Color scale is shown on the bottom.

In the aged male, our results showed that in CP vs. GI there were 19 transcripts upregulated and two downregulated (>1.2-fold and P < 0.01 for each). No functional annotation clusters were identified.

In the aged female, 30 transcripts were upregulated and four transcripts were downregulated (>1.2-fold and P < 0.01 for each) in CP vs. GI (Fig. 5B). Functional annotation clustering analysis (enrichment score >1.4) revealed upregulation of two annotation clusters: ATP metabolic process (P = 3.7 × 10–3) and purine ribonucleoside triphosphate metabolic process (P = 4.6 × 10–3). All results were confirmed by qRT-PCR analysis (Table 2).

Table 2.

qRT-PCR and microarray results, fold changes compared with control

Group	RT-PCR	Microarray	RT-PCR	Microarray	RT-PCR	Microarray	RT-PCR	Microarray	RT-PCR	Microarray
	SDH		VDAC2		CSRP3		ADP/ATP Translocase 1		IDH2
Mature male GI	1.16 ± 0.31	0.164	2.28 ± 0.07	0.300	1.16 ± 0.18	0.222	0.83 ± 0.12	−0.620	0.42 ± 0.06	−0.645
Mature Female GI	13.10 ± 0.01	−0.228	0.71 ± 0.47	−0.014	7.16 ± 0.02	0.258	1.38 ± 0.06	−0.163	0.65 ± 0.07	−0.378
Aged male GI	0.89 ± 0.22	−0.467	1.44 ± 0.72	0.336	1.13 ± 0.16	0.528	1.18 ± 0.51	0.523	0.80 ± 0.15	−0.225
Aged female GI	0.72 ± 0.19	−0.338	1.89 ± 0.57	0.242	1.02 ± 0.26	0.436	0.85 ± 0.27	−0.599	1.63 ± 0.15	0.347
Mature male CP	13.90 ± 0.12	1.800	1.24 ± 1.07	0.0	6.71 ± 0.09	1.700	2.76 ± 0.10	0.0	1.71 ± 0.06	0.0
Mature female CP	7.72 ± 0.03	−0.363	3.12 ± 0.19	0.376	5.78 ± 0.20	0.589	1.69 ± 0.18	1.737	0.50 ± 0.01	−0.731
Aged male CP	0.60 ± 0.09	−0.588	1.93 ± 0.37	0.407	0.73 ± 0.17	−0.360	1.06 ± 0.04	0.268	0.38 ± 0.17	−0.322
Aged female CP	0.42 ± 0.16	−0.339	7.66 ± 0.68	0.505	0.56 ± 0.10	−0.517	3.35 ± 0.05	−0.353	0.63 ± 0.16	−0.168

Validation of microarray results via qRT-PCR analysis. Results displayed as fold changes as compared to control. Differences between microarray and qRT-PCR are shown in boldface. GI, global ischemia; CP, cardioplegia.

Quality control analysis.

The total number of proteins identified in each experimental group with valid iTRAQ labels is shown in Table 3. The data were found to be normalized with no significant outliers. Box plot analysis showed alignment of average iTRAQ intensity values for all experimental groups (Fig. 6, A, D, G, and J). Pair-wise correlation plots indicated highly significant differences in expressed proteins in the mature male, mature female and aged male (Fig. 6, C, F, and I); whereas in the aged female (Fig. 6L) the r values for GI and CP were found to be insignificant, showing no difference between treatments.

Table 3.

Number of iTRAQ validated proteins

Experimental Group	Proteins, n
Mature male	239
Mature female	204
Aged male	225
Aged female	222

iTRAQ validated proteins: the number of proteins identified and analyzed in each group with valid iTRAQ labeling and at least 1 high confidence peptide (>95% confidence) are listed. Numbers listed in this table are values for total proteins per group.

Fig. 6.

Quality control analysis. A–C: mature male. D–F, mature female. G–I: aged male. J–L: aged female. Box plot analysis (A, D, G, J) showed alignment of average iTRAQ intensity values for all experimental groups in all experimental treatments. Principal component analysis (PCA) (B, E, H, K) indicated that in the mature male (B), mature female (E), and aged male (H) the proteins expressed in control (CTR), GI, and CP were separate and diverse for individual treatments, suggesting that the proteins had different expression levels. In contrast, in the aged female (K) the proteins expressed in GI and CP were clustered together, suggesting there was no difference between treatments. PC1, Principal Component 1, found on the x-axis, accounts for as much variability in the data as possible. PC2, Principal Component 2, found on the y-axis, accounts for as much variability as possible that is not related to PC1. CTR circled in green, GI circled in red, CP circled in orange. Pair-wise correlation plots (C, F, I, L) indicated highly significant differences in expressed proteins in the mature male (C), mature female (F), and aged male (I); whereas in the aged female (L) the r values for GI and CP were found to be insignificant, showing no difference between treatments.

Principal component analysis indicated that in the mature male, mature female, and aged male (Figs. 6, B, E, and H) the proteins expressed in control, GI, and CP were diverse and clustered for individual treatments suggesting that the proteins had different expression levels in each treatment. However, in the aged female the proteins expressed in GI and in CP were clustered together, suggesting no difference in the proteins expressed (Fig. 6K).

Overall proteomic analysis.

A total of 890 proteins were found to have valid iTRAQ labeling. From these 890 proteins, 362 proteins were found to be unique within the experimental groups and treatments. Results show that 132 proteins were commonly expressed in all experimental groups and treatments (Fig. 7A).

Fig. 7.

Venn diagrams of all experimental groups: mature male, mature female, aged male, and aged female. A: all proteins. A total of 362 high confidence proteins were identified in all groups and treatments; 132 proteins were commonly identified in all experimental groups. B: GI. A total of 64 proteins were identified in GI. No proteins were commonly detected by mass spectrometry in all experimental groups. C: CP. A total of 84 proteins were identified in CP. Four proteins were commonly detected in all experimental groups.

Hierarchical cluster analysis (Fig. 8, A, B, and C) showed distinct protein expression patterns between GI and CP in the mature male, mature female, and the aged male. However, the aged female hierarchical cluster analysis (Fig. 8D) showed no distinct protein expression patterns between GI and CP in agreement with PCA analysis (Fig. 6K). Proteomic results were confirmed by Western blot analysis (Fig. 9).

Fig. 8.

Hierarchical cluster analysis of differentially expressed proteins. A: mature male. B: mature female. C: aged male. D: aged female. The differentially expressed proteins were identified by supervised analysis on the basis of P value <0.01 in each group: CTR, GI, CP. The proteins are identified with short descriptions obtained from the SWISS-PROT database. The lFC in protein expression is shown with pseudocolor scale (−3 to 3) with red denoting upregulation and green denoting downregulation. The columns represent lFC comparisons and the rows represent the proteins. Dendograms are found on the left, experimental groups are found on the bottom, and protein names are found on the right.

Fig. 9.

Western blot results for isocitrate dehydrogenase 2 (IDH2). A: validation of proteomic results via Western blot (WB) analysis. Results displayed as fold changes compared with control. B: image of immunoblot, performed using mouse monoclonal IDH2 antibody (1:2,000 dilution, Abcam). Blots were detected using ECL-Plus (Amersham Pharmacia Biotech) with species-appropriate secondary antibodies. Densitometry analysis was performed using the Image J analysis software. Lanes: 1, control; 2, mature male; 3, mature female; 4, aged male; 5, aged female.

Proteomic changes with GI.

GI up- and downregulated 64 unique iTRAQ validated proteins (Fig. 7B). There were no proteins common to all experimental groups. The number of unique proteins altered in GI was significantly increased (P < 0.05) in both mature and aged females compared with males. Analysis of up- and downregulated proteins in GI indicated that there were nine unique proteins altered in the mature male and three unique proteins altered in the aged male. In contrast GI up- and downregulated 17 unique proteins in the mature female and 21 unique proteins in the aged female (Fig. 7B and Supplemental Table S1).¹

Proteomic changes with CP.

In CP there was a total of 84 unique up- or downregulated iTRAQ-validated proteins in all experimental groups (Fig. 7C and Supplemental Table S2). Four proteins were commonly detected by mass spectrometry in CP in all experimental groups. These commonly detected proteins were myosin-6 (upregulated), albumin (downregulated), glypican-6 (downregulated), and zinc finger protein 518A (downregulated).

Proteomic changes in CP vs. GI.

Comparison between CP and GI indicated that there were 35, 38, 28, and 44 proteins detected in both CP and GI for mature male, mature female, aged male, and aged female, respectively. The number of proteins commonly detected in both GI and CP, GI only and CP only are shown in Table 4.

Table 4.

Number of proteins expressed in each group and treatment

Group	GI & CP	GI Only	CP Only
Mature male	35	1	3
Mature female	38	5	5
Aged male	28	1	8
Aged female	44	2	8

The number of unique proteins detected by mass spectrometry in each group and treatment.

In the mature male and female and aged male and female there were three, five, eight, and eight proteins, respectively, that were uniquely detected by mass spectrometry in CP that were not detected in any other group or treatment; however, these proteins were not identified with any functional annotation clusters (Table 5).

Table 5.

Unique protein expression in each experimental group in CP vs. GI

Protein Name	Z Score
Mature Male
NADH dehydrogenase flavoprotein 1, mitochondrial	2.476
Plectin	0.431
Protocadherin beta-14	0.430
Mature Female
Phosphoglycerate mutase 1	6.863
Phospholipase B1, membrane-associated	0.465
Elongation factor 1-alpha 2	0.446
Calnexin	0.374
Fatty acid-binding protein, adipocyte	0.359
Aged Male
Hydroxyacyl-coenzyme A dehydrogenase, mitochondrial	24.140
Troponin I, cardiac muscle	4.939
Laminin subunit gamma-1	3.234
Glyceraldehyde-3-phosphate dehydrogenase	2.238
Fructose-bisphosphate aldolase A	2.128
Decorin	2.079
Ubiquitin-40S ribosomal protein S27a	0.383
Cytochrome c oxidase subunit 6B1	0.333
Aged Female
Alpha-actinin-2	2.680
Cytochrome b-c1 complex subunit Rieske, mitochondria	2.316
Sarcoplasmic/endoplasmic reticulum calcium ATPase 2	2.261
2-oxoglutarate dehydrogenase, mitochondrial	2.231
Calcium-binding mitochondrial carrier protein Aralar1	2.122
Leukotriene C4 synthase	0.440
Neuroblast differentiation-associated protein AHNAK	0.318
Hemoglobin subunit beta	0.036

Unique protein expression in each experimental group in CP vs. GI. Z score indicates the fold change compared with control.

Functional annotation clustering of proteins expressed in GI and CP.

Functional annotation clustering (P < 0.05, enrichment score >2.0) indicated that there were two commonly upregulated pathways in GI and CP in all experimental groups; the generation of precursor metabolites and energy and mitochondrial membrane.

Functional annotation clustering in GI.

In the mature male oxidative phosphorylation and cardiac muscle tissue morphogenesis were significantly upregulated in GI, while in the mature female nucleosome assembly and fatty acid metabolism were significantly upregulated (P < 0.05, enrichment score >2.0 for each). In the aged female oxidative phosphorylation, tricarboxylic acid cycle (TCA cycle), ATP catabolism, and cytoskeletal protein binding were significantly upregulated in GI (P < 0.05, enrichment score >2.0).

Cardiac muscle tissue morphogenesis and calcium binding were significantly downregulated in GI in the mature female (P < 0.05, enrichment score >2.0). No significant clusters were downregulated in GI in the mature male, aged male, or aged female.

Functional annotation clustering in CP.

In the mature male there was a significant upregulation (P < 0.05, enrichment score >2.0) of cellular respiration, oxidative phosphorylation, and the electron transport chain. In the aged male there was a significant upregulation (P < 0.05, enrichment score >2.0) of glycolysis, gluconeogenesis, glucose metabolic/catabolic processes, muscle protein, and muscle contraction. In the mature female there was a significant upregulation (P < 0.05, enrichment score >2.0) of ATP binding while in the aged female there was a significant upregulation (P < 0.05, enrichment score >2.0) of the TCA cycle, cellular respiration, muscle tissue morphogenesis, aerobic respiration, oxidative phosphorylation, ATP catabolism, and ATPase activity (Supplemental Table S3).

Muscle protein and cardiac muscle tissue morphogenesis were significantly downregulated (P < 0.05, enrichment score >2.0) in the mature male, while in the aged male protein degradation, induction of apoptosis, translation, and cell-cell signaling were significantly downregulated (P < 0.05, enrichment score >2.0) (Supplemental Table S3). There were no significantly downregulated pathways in either the mature female or the aged female.

Functional annotation clustering analysis of the proteins expressed in CP but not in GI.

There were no significant clusters in CP that were uniquely upregulated in the mature male or female. In the aged male, CP uniquely upregulated (P < 0.05, enrichment score >2.0) protein degradation, antiapoptosis, glycolysis, gluconeogenesis, and glucose catabolic and metabolic pathways. In the aged female the structural constituent of muscle, cytoskeleton organization, and protein complex assembly and biogenesis were uniquely upregulated with CP (P < 0.05, enrichment score >2.0).

DISCUSSION

The main finding of this study is the identification of the differing pathways, expressed after ischemia and after ischemia with cardioplegia, of cardioprotection in the mature and aged male and female. Of particular importance is the identification of the pathways for cardioprotection in the aged female.

The mechanisms modulating cardioprotection have remained elusive. Confounding identification of these mechanisms are the differences in absolute cardioprotection observed between the mature and aged heart and the aged female compared with the aged male. In this report we have used the in situ blood perfused heart model to allow for clinical relevance. Our results using this model are in agreement with our previous findings using the isolated perfused heart model and demonstrate that the cardioprotection afforded by CP is significantly decreased in the aged heart and is significantly decreased in the aged female compared with the aged male (27–29). These results are consistent with clinical data showing that following cardiac surgery there is a significant depression in each parameter of cardiac function and a significant increase in myocardial infarction in females compared with males (1, 4, 19, 42).

In this report we have used both microarray and large-scale proteomic analysis. Microarray analysis showed that there were no significant functional annotation clusters identified for the mature or aged male and that differing mitochondrial associated functional annotation clusters were identified in the mature male and aged female. We did not detect any changes in inflammation or inflammatory response. This is in contrast to Podgoreanu et al. (34), who have previously shown that in a rat model of cardiopulmonary bypass there was upregulation in inflammatory response transcripts, response to external stimuli, response to biotic stimuli and chemokine activity transcripts and downregulation in calcium ion binding and nucleocytoplasmic transport transcripts. Ruel et al. (36) also showed that in human atrial tissue samples there was an upregulation of inflammation/transcription activators, apoptotic genes, and stress genes and a downregulation of immunoglobulin genes associated with cardiopulmonary bypass and cardioplegia. However, Wang et al. (43) reported overall low expression levels in all transcripts but changes in specific angiogenesis-related genes that were preidentified for analysis. Our results may be accounted for in part by differences in animal models but most likely reflect differences in CP formulae. In our studies we have used depolarizing cardioplegia with the addition of DZX, a mitochondrial ATP-sensitive potassium channel opener. The use of DZX has been shown to enhance cardioprotection and decrease infarct size through modulation of mitochondrion volume and mitochondrial function (27–29). Thus changes in mitochondrial associated transcripts would be expected to be observed. The decreased infarct size in our studies and the time for reperfusion would also limit the activation of inflammation.

Examination of the mechanisms of cardioprotection by large-scale proteomic analysis showed distinct pathways in the mature and aged male and female. PCA indicates that in the mature male and female and the aged male the proteins expressed in control, GI and CP were separate and diverse. However, in the aged female the proteins expressed in GI and in CP were clustered together, suggesting no difference in the proteins expressed. These differences were also evident in hierarchical clustering data and with the functional data shown herein and provide a possible mechanism for the differences in cardioprotection observed between the mature and aged heart and in particular in the aged female heart.

Our results show that in the mature male, of the 29 differentially expressed proteins in CP there were 11 common proteins that were also expressed in GI. Similarly, in the mature female of the 25 differentially expressed proteins in CP, 9 proteins were commonly identified in GI. The aged male had too few proteins expressed in GI for comparison. However, in the aged female, 25 of the 30 differentially expressed proteins in GI were commonly identified in the 39 differentially expressed proteins of CP.

The pathways identified by proteomic analysis provide key insight into the mechanisms of cardioprotection. In the mature male and mature female mitochondrial pathways were identified as playing a key role in cardioprotection; whereas in the aged male protein degradation, antiapoptosis and mitochondrial pathways were identified as being associated with cardioprotection. In contrast, in the aged female, muscle and protein complex pathways were identified as important to cardioprotection. The relevance of these differing pathways is supported by our previous findings noting that the role of the mitochondrion in providing cardioprotection is modulated by age and sex, with the aged female being more severely compromised than the aged male (27–29). We have previously shown that in the aged female heart, mitochondrial function is reduced compared with mature and aged male and that myocardial function and myonecrosis is significantly increased. In the aged male the reduction of mitochondrial function would appear to be compensated by the downregulation of the pathways of protein degradation, induction of apoptosis, translation, and cell morphogenesis. In the aged female the reduction of mitochondrial function and the increase in mitochondrial calcium accumulation is uncompensated by these pathways, and therefore, muscle, cytoskeleton, and protein mechanisms must be upregulated to modulate myonecrosis and allow for survival of the myocardium following insult.

The importance of the mitochondrion and the mitochondrial proteome have been previously demonstrated; our data agree with the proteomic data of others who have shown changes in energy metabolism pathways following ischemia and reperfusion (6). In particular our data would agree with that of White et al. (44) who have shown that ischemia-reperfusion in the New Zealand White rabbit results in changes in the proteome and includes changes in contractile fiber and energy production, similar to our findings.

Our data also agree with that of Schwertz et al. (37), who show that after ischemia rabbit hearts had a discernible difference in the myocardial proteins expressed compared with control. We also show that in all experimental groups there was a marked difference in myocardial protein expression with GI and a significant increase in mitochondrial protein expression and mitochondrial annotation clusters with CP. The changes in mitochondrial-related mechanisms are supported by Wong et al. (45) who demonstrated, in isolated perfused hearts, that ischemic preconditioning alters the mitochondrial proteome.

The importance of the mitochondrion in cardioprotection is further shown by functional enrichment analysis using GO terms to identify the cellular components in which the differentially expressed proteins were overrepresented (Figs. 10 and 11 and Supplemental Tables S4 and S5).

Fig. 10.

Functional enrichment pathways for mature male (A), mature female (B), aged male (C), and aged female (D) GI vs. CTL. Functional enrichment analysis was performed by identifying the overrepresented gene ontology (GO) categories in differentially expressed proteins: this was done using the biological processes and molecular functions enrichment analysis available from the Database for Annotation, Visualization and Integrated Discovery (DAVID). Functional pathways are labeled on the x-axis, and the −log(P value) is on the y-axis. On the right is the ratio of proteins in pathway over total proteins, and the yellow line shows the ratio of each pathway. The red line is the threshold at P = 0.05. Any pathway that passes the red line is significantly enriched.

Fig. 11.

Functional enrichment pathways for mature male (A), mature female (B), aged male (C), and aged female (D) CP vs. CTL. Functional enrichment analysis was performed by identifying the overrepresented GO categories in differentially expressed proteins: this was done using the biological processes and molecular functions enrichment analysis available from the DAVID. Functional pathways are labeled on the x-axis, and the −log(P value) is on the y-axis. On the right is the ratio of proteins in pathway over total proteins, and the yellow line shows the ratio of each pathway. The red line is the threshold at P = 0.05. Any pathway that passes the red line is significantly enriched.

Functional enrichment analysis showed three significantly enriched pathways that were common (P < 0.05) with CP in all experimental groups. These pathways were mitochondrial dysfunction, oxidative phosphorylation, and calcium signaling (Fig. 11).

These data agree with our previous studies on the mechanisms of cardioprotection where we have used magnesium supplementation to limit calcium overload and to enhance the preservation and resynthesis of high-energy phosphates. We have also shown that despite the use of additives mitochondrial dysfunction and oxidative phosphorylation are modulated in cardioprotection and is affected by both age and sex (27–29). The pathways of calcium signaling have also been shown to play a key role in cardioprotection. Our studies have demonstrated that increased mitochondrial calcium accumulation affects DNA fragmentation and is associated with decreased postischemic functional recovery and increased infarct size (10, 30, 35, 41). These effects were exacerbated in the aged female despite intervention with CP (30).

It is interesting to note the differing enriched pathways between the aged male and the aged female. Glycolysis/gluconeogenesis and the pentose phosphate pathway were significantly enriched pathways in the aged male but not significant in any other experimental groups. Purine and pyruvate metabolism were not significant in the aged male but were significantly enriched in the mature and aged female (Figs. 10 and 11). Of significance, functional enrichment analysis showed that glyoxylate and dicarboxylate metabolism were significant pathways in the aged female only. In addition the TCA cycle, purine and pyruvate metabolism were significant in the aged female but not the aged male. Glycolysis and gluconeogenesis did not reach threshold in the aged female.

The differences in enriched and functional pathways in CP between the mature and aged and males and females are supported in previous reports by others (16, 11, 27). Our data show that fatty acid metabolism and elongation in the mitochondria is a significant pathway in the mature female, aged male, and aged female, allowing for an energy lucrative pathway for cardiac mitochondria. This would agree with Jafri et al. (16) and Fogle et al. (11) and our own studies (27) showing sex modulation in energy-producing pathways.

The mechanisms of energy production have long been known to be altered by ischemia and by reperfusion. Kohn and Garfinkel (20) have shown that the rate of fatty acid transport and oxidation is sufficient to support the homeostatic needs of the myocardium but is insufficient to meet these needs during ischemia. It has also been shown that increased levels of fatty acid oxidation during ischemia and reperfusion are detrimental to the myocardium (20, 24). Lopaschuk (23) has suggested that an increase in glucose metabolism during ischemia would decrease the amount of myocellular injury. Our data indicate that in the aged male glycolysis and gluconeogenesis are significant pathways along with fatty acid oxidation in CP. The utilization of these pathways would provide a mechanism for enhanced cardioprotection as outlined above. In contrast, in the aged female where glycolysis and gluconeogenesis are not significant pathways and energy production is shifted to glyoxylate and dicarboxylate metabolism our data predict that the response to ischemia would be limited.

The importance of glyoxylate and dicarboxylate metabolism in the aged female would suggest impairment of mitochondrial function that would limit or impair the ability of the aged female to respond to ischemia. Previous studies have suggested that a shift to glyoxylate and dicarboxylate metabolism is a marker for decreased tolerance to insult (5). These data would agree with our findings showing that at similar ischemic insult levels, the aged female fairs more poorly compared with the aged male and is associated with changes in mitochondrial function (28).

In addition, our data show that glutathione metabolism is not a significant pathway in the aged female, suggesting that the modulation or control of reactive oxygen species may be impaired. hypoxia-inducible factor (HIF)1α signaling and nitric oxide signaling in the cardiovascular system were also not significant in the aged female. HIF1α is abundant at low oxygen concentrations and has been shown to mediate the effects of hypoxia (2). The reduction of HIF1α signaling and nitric oxide signaling would compromise the mechanisms allowing for functional recovery during reperfusion and exacerbate myocellular injury; this would agree with our previous studies (29, 35).

Despite the use of matched samples we found no overlap between transcriptomic and proteomic data using either indirect or direct comparison. This is not uncommon as previous studies have shown that the correlation between transcription and translation is not linear (31) and that little or no direct overlap is seen between transcriptomic and proteomic data in mammals (15, 25, 32).

Foss et al. (12) showed a significantly greater overlap than expected by chance (P < 0.001) when comparing the proteomic and transcriptomic data of yeast. This showed that differentially expressed proteins were more likely to correspond to differentially expressed transcripts. However, this considerable overlap is not seen when examining data from mammals. Ghazalpour et al. (13) examined >5,000 peptides and 22,000 transcripts of recombinant mice and found only a modest correlation between the data sets. A statistically significant correlation was found in only 21% of the genes and 15% of the proteins. Murgiano et al. (31) found no direct overlap between transcriptomic and proteomic data in pigs.

The data on transcriptomic and proteomic overlap in mammals are limited and what are present are not faultless; however, there are many possible reasons as to why the data sets do not correlate; posttranscriptional parameters, posttranslational parameters, noise, and experimental errors (25). However, in addition to various biological factors, it should be considered that the poor correlation could be due to the inadequacy of available statistical tools to compensate for biases in the data collection methodologies as well as integrating the data from either proteomics or genomics studies (31).

In conclusion, our results indicate that age and sex modulate the cardioprotective effects afforded by CP and that significant changes in functionally enriched pathways are evident between the mature and aged male and female (Fig. 12). Our data also show that specific pathways associated with the mitochondrion and mitochondrial function modulates cardioprotection and that these pathways are significantly altered in the aged and in particular in the aged female compared with the aged male. The alteration of these pathways significantly contributes to decreased myocardial functional recovery and myonecrosis following ischemia and reperfusion. These data provide age- and sex-specific mechanisms that may be modulated to allow for enhanced cardioprotection.

Fig. 12.

Age and sex modulation of the cardioprotective effects afforded by CP. The mitochondrion plays an important role in the modulation of the cardioprotective effects afforded by CP. Common pathways for mature and aged male and female are shown in the representative mitochondria. Relative contribution in each age group is shown by line number and width. [Ca²⁺]_i, intracellular calcium concentration.

Study Limitations

In our study we have constructed a cDNA library made from rabbit heart LV tissue. We have isolated and 5′ sequenced 8,647 rabbit heart cDNAs and have identified and stored 3,592 nonredundant cDNAs with a mean insert size of 1.67 kb (27). It has been noted that the Broad Institute's OryCun 2.0 reference genome contains 22,971 total genes; however, we expressly used LV cDNAs with homology to humans to produce data that are immediately transferable to humans and a clinical approach.

The Genomics Core facility at Beth Israel Deaconess Medical Center does not have the most recent mass spectrometer available; with constantly progressing technology and limited budgets it is impossible always to have the most up-to-date instrumentation. The AB SCIEX 5800 MALDI-TOF/TOF is a recently released improvement (June, 2009) on the 4800 that has reported ∼10×-fold increase in sensitivity. Every effort was made with the instrumentation that we had available (Agilent 1200 HPLC, Dionex Ultimate Plus nanoLC and the AB Sciex 4800 Plus mass spectrometer) to maximize the dynamic range of the four 8-plex iTRAQ MudPIT experiments in this report. The 4800 Plus MALDI-TOF/TOF (a complete redesign of the 4700 instrument) used in the proteomics experiments in this report was released as a commercial product on 04/23/2007, and the Beth Israel Deaconess Medical Center purchased this instrumentation in March, 2008 for the Genomics Core. The proteomic experiments in this report were run in November, 2010. We acknowledge that there are newer mass spectrometers available, and we will take this into consideration for future proteomic experiments.

GRANTS

This study was supported by National Heart, Lung, and Blood Institute Grants HL-029077 and HL-103542.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: K.M.B., R.J.B., C.D., S.T.D., and J.D.M. performed experiments; K.M.B., M.K.B., S.T.D., T.A.L., and J.D.M. analyzed data; K.M.B., R.J.B., S.T.D., T.A.L., and J.D.M. interpreted results of experiments; K.M.B. and J.D.M. prepared figures; K.M.B. and J.D.M. drafted manuscript; K.M.B., S.L., and J.D.M. edited and revised manuscript; K.M.B., S.L., and J.D.M. approved final version of manuscript; J.D.M. conception and design of research.