Controlled reoxygenation cardiopulmonary bypass is associated with reduced transcriptomic changes in cyanotic tetralogy of Fallot patients undergoing surgery

Abstract

In cyanotic patients undergoing repair of heart defects, high level of oxygen during cardiopulmonary bypass (CPB) leads to greater susceptibility to myocardial ischemia and reoxygenation injury. This study investigates the effects of controlled reoxygenation CPB on gene expression changes in cyanotic hearts of patients undergoing surgical correction of tetralogy of Fallot (TOF). We randomized 49 cyanotic TOF patients undergoing corrective cardiac surgery to receive either controlled reoxygenation or hyperoxic/standard CPB. Ventricular myocardium biopsies were obtained immediately after starting and before discontinuing CPB. Microarray analyses were performed on samples, and array results validated with real-time PCR. Gene expression profiles before and after hyperoxic/standard CPB revealed 35 differentially expressed genes with three upregulated and 32 downregulated. Upregulated genes included two E3 Ubiquitin ligases. The products of downregulated genes included intracellular signaling kinases, metabolic process proteins, and transport factors. In contrast, gene expression profiles before and after controlled reoxygenation CPB revealed only 11 differentially expressed genes with 10 upregulated including extracellular matrix proteins, transport factors, and one downregulated. The comparison of gene expression following hyperoxic/standard vs. controlled reoxygenation CPB revealed 59 differentially expressed genes, with six upregulated and 53 downregulated. Upregulated genes included PDE1A, MOSC1, and CRIP3. Downregulated genes functionally clustered into four major classes: extracellular matrix/cell adhesion, transcription, transport, and cellular metabolic process. This study provides direct evidence that hyperoxic CPB decreases the adaptation and remodeling capacity in cyanotic patients undergoing TOF repair. This simple CPB strategy of controlled reoxygenation reduced the number of genes whose expression was altered following hyperoxic/standard CPB.

reintroduction of high oxygen levels to cyanotic patients on cardiopulmonary bypass (CPB) leads to myocardial damage prior to ischemic cardioplegic arrest (31), suggesting that the injury seen following cardioplegic arrest may be in part due to CPB-induced reoxygenation injury, and this has also been demonstrated in other clinical studies (8, 9). One of the strategies proposed to avoid this injury is the use of “controlled reoxygenation” using normoxic CPB with a pump prime the Po₂ (oxygen tension in the blood) of which is matched to the patient's preoperative saturation. This has been shown to ameliorate reoxygenation injury and leads to almost complete functional recovery (17, 20). In a recent prospective randomized trial we have demonstrated that controlled reoxygenation on starting CPB in cyanotic patients undergoing cardiac surgery significantly reduces oxidative stress and myocardial cell injury, compared with hyperoxic CPB (5). Hypoxia changes the cardiac protein pattern, mainly through altered gene expression (6, 15). We have recently shown that chronic hypoxia in cyanotic children with tetralogy of Fallot (TOF) induced the expression of genes associated with apoptosis and reduced the expression of genes associated with normal myocyte contractility and function (15) and may be responsible for the susceptibility of cyanotic children to reoxygenation injury during and after surgery.

To our knowledge, no previous attempts have been made to determine the gene expression profiles associated with reoxygenation injury in cyanotic heart disease. Transcriptional profiling is a powerful tool for delineating complex patterns of gene expression in response to severe systemic stimuli and injury (37), and several recent studies have used microarray technology to examine the global myocardial stress response during cardiac surgery (27, 36, 39). The aim of this study was to investigate the effects of reoxygenation injury on the myocardial gene expression profiles of cyanotic patients with a diagnosis of TOF undergoing corrective heart surgery. Furthermore, we evaluated whether controlling reoxygenation during CPB could prevent the effect of hyperoxic CPB on the myocardial gene expression profile in these patients.

MATERIALS AND METHODS

Patients, clinical, and biochemical methods.

We randomized 49 cyanotic patients undergoing TOF or pulmonary atresia repair between January 2004 and November 2009 at the Bristol Royal Hospital for Children to receive either controlled reoxygenation (50–80 mmHg, n = 24) or hyperoxic/standard (150–180 mmHg, n = 25). All patients were in a stable condition without preoperative respiratory or inotropic support. “Controlled reoxygenation CPB” referred to a pump prime, the Po₂ of which is matched to the Po₂ of the patient. “Hyperoxic/standard CPB” referred to a pump prime prepared to the current “best practice” protocols, which has a Po₂ relatively hyperoxic for a cyanotic patient. Preoperative characteristics in the two groups are summarized in Table 1. The study was approved by the Hospital Research Ethics Committee, and parental informed consent was gained for all patients. Intraoperative anesthetic and operative techniques were standardized as previously reported (21). Cold blood (4–6°C) St Thomas' I-based cardioplegic solution (4:1 dilution blood/St Thomas' I crystalloid cardioplegia) was used for myocardial preservation, with the following composition (mM): 16 MgCl₂, 2 CaCl₂, 20 KCl, 147 NaCl, 1.0 procaine HCl. Additional cardioplegia was administered after each 20 min of aortic cross-clamping. Blood oxygen tension levels for assessing the efficacy of the controlled reoxygenation were measured in both groups immediately after intubation, at 10 and 30 min after starting CPB, at the end of the ischemic time, and 6 and 24 h postoperatively.

Table 1.

Baseline patients' characteristics and clinical outcomes

	Controlled Reoxygenation (n = 24)	Hyperoxic/Standard (n = 25)	P Value
Age, mo	10.5 ± 9.9	11.5 ± 6.5	0.1
Weight, kg	8.4 ± 2.3	8.2 ± 2.3	0.7
Preoperative saturations, %	80.3 ± 9.1	81.7 ± 5.9	0.5
Echo RV wall thickness, mm	6.5 ± 0.5	6.9 ± 0.7	0.2
Echo RV wall indexed, mm/m2	15.8 ± 3.6	17.3 ± 3.7	0.4
Echo VSD size, mm	11.5 ± 3.0	10.3 ± 1.3	0.2
CPB time, min	107.4 ± 30.3	109.6 ± 29.9	0.4
Cross-clamp time, min	61.7 ± 19.6	68.7 ± 15.3	0.1
Ventilation time, h	21.4 (21.7–50.2)	32.0 (30.1–73.2)	0.03
Dopamine support duration, h	24.0 (20.9–38.6)	37.0 (30.8–73.8)	0.05
Hospital stay, days	7.5 (6.7–15.8)	8.0 (6.9–17.8)	0.2
Lactate, mcg/l			>0.5
T1 = preoperatively	1.13 ± 0.3	1.03 ± 0.4
T2 = X-clamp off	1.13 ± 0.3	1.04 ± 0.3
T3 = 30 min X-clamp off	1.17 ± 0.4	1.05 ± 0.3
T4 = 2 h X-clamp off	1.19 ± 0.4	1.57 ± 0.8*
T5 = 6 h X-clamp off	1.28 ± 0.9	1.09 ± 0.4
T6 = 24 h X-clamp off	1.24 ± 0.4	1.28 ± 0.6

Data are means ± SD or median and interquartile range. RV, right ventricle; X-clamp, aortic cross clamp (cardioplegic ischemic time). *P < 0.05 vs. T1, T2, and T3 in the Hyperoxic/Standard group.

After the surgery, all patients were admitted to the pediatric intensive care unit (ICU) and were managed according to unit protocols (21, 32).

Serum lactate and 8-isoprostane levels (enzyme immunoassay; Cayman Chemicals, Ann Arbor, MI) were measured preoperatively, on removal of the aortic cross-clamp, and 30 min, 2, 6, and 24 h thereafter, as measurements of myocardial oxidative stress.

Continuous outcomes are summarized as an arithmetic mean and standard deviation if normally distributed or as median and interquartile range if skewed. Categorical data are presented as actual counts and percentages. Biochemical markers measured at multiple time points were analyzed by repeated-measures analysis of variance. A P value <0.05 was considered statistically significant.

Cardiac muscle biopsies.

Ventricular biopsy specimens (10 mg net weight) were collected from the apex of the right ventricle using a Trucut needle (21). The Pre-op biopsy was taken immediately after institution of CPB. The hyperoxic or controlled reoxygenation biopsies were collected at the end of ischemic time. Each specimen was immediately put in RNA Later solution (Qiagen, UK) and kept overnight at 4°C. The next day RNA Later solution was removed, and biopsies were kept at −80°C until RNA extraction.

RNA extraction.

Tissue was mechanically homogenized in lysis reagent (Qiagen, Crawley, UK) and Total RNA was purified with RNeasy Micro Kit (Qiagen) and eluted into 12 μl of RNase-free water. The concentration and purity of the total RNA samples were assessed by spectrophotometry (Nanodrop, Wilmington, DE). Only samples with a sufficiently high yield (≈1 μg of total RNA at a minimum concentration of 125 ng/μl) and purity (an A260/A280 ratio of close to 2) were further analyzed for integrity with a Bioanalyzer 2100 with RNA 6000 Nano Assay (Agilent Technologies, Stockport, UK). Samples that met the quality control criteria were used as templates for cRNA synthesis.

Gene microarrays.

For each experimental group (Pre-op, n = 7; Hyperoxic/standard, n = 5; and Controlled reoxygenation, n = 5), samples from individual patients were processed. Ventricular total RNAs (1 μg) from individual patients were used to generate biotinylated cRNAs. The quantity and size distribution of purified cRNA was assessed on a Bioanalyzer 2100 using RNA 6000 Nano Assay (Agilent Technologies) to ensure that the cRNA amplification was successful. Target fragmentation was achieved by incubation at 94°C for 35 min in fragmentation buffer (40 mM Tris-acetate, pH 8.1/100 mM KOAc/30 mM MgOAc). The size distribution of the fragmented labeled transcripts was assessed on the Agilent Technologies Bioanalyzer 2100 using the RNA 6000 Nano Assay. These cRNAs samples were used for hybridization to separate Affymetrix GeneChip arrays. For each experimental group, five samples from individual patients were processed. Hybridization of the labeled cRNA to the Affymetrix GeneChip Human Genome U133 Plus 2.0 Array was carried out for 16 h in the Affymetrix GeneChip Hybridization Oven 640. Then GeneChip arrays were stained and washed on the GeneChip Fluidics Station 450 (Affymetrix). The fluorescent signals were detected with an Affymetrix GeneChip Scanner 3000 and stored as high-resolution fluorescence intensity data file. These data were initially analyzed with Affymetrix GeneChip operating software GCOS 1.2, which generates an expression report file that lists the quality control parameters. All of these parameters were scrutinized to ensure that array data had reached the necessary quality standards (scaling factor of <3-fold; average background values at 20–100 and the ratio of 3′:5′ signal no more than 3 for housekeeping genes GAPDH and β-actin). All raw data have been submitted to the National Center for Biotechnology Information Gene Expression Omnibus database under the following accession numbers: GSE14956 and GSE38177.

Microarray data analysis.

Seven pre-op patients samples (4 pre-op hyperoxic and 3 pre-op controlled), five post-op hyperoxic samples and five post-op controlled samples were used. Separate microarrays were probed with independently generated target from each tissue from heart biopsies of either group at the beginning of surgery and at the end of ischemic time. The microarray data were normalized in accord with our standard protocol. Raw data (CEL files) were uploaded into ArrayStar software version 2.1 (DNASTAR) for normalization and statistical analysis. The robust multichip analysis algorithm (7) was used for background correction, quantile normalization, and median polish summarization. The statistical analysis was carried out using ArrayStar software followed by multiple testing corrections for false discovery rate. Transcripts were filtered on the basis of ≥1.6-fold difference.

Functional annotation analysis.

Differentially expressed genes were analyzed according to predefined pathways and functional categories annotated by KEGG (24) and GO (2) using the DAVID (Database for Annotation, Visualization, and Integrated Discovery) bioinformatics resource (11). All gene annotations were checked using online tools and databases (http://www.ncbi.nlm.nih.gov). Note that the databases are in a constant state of flux and that annotations are subject to updates, redefinition, and correction.

cDNA synthesis and real-time PCR.

For each experimental group (Hyperoxic/standard, n = 5 and controlled reoxygenation, n = 5), total RNA from individual patients was extracted. Complementary DNA was reverse transcribed from 1 μg of total RNA using Superscript III cDNA first-strand synthesis kit (Invitrogen, UK), diluted two fold and 1 μl used in real-time PCR reactions. Optimized primers of human SLC6A6 (QT00095655), NRAS1 (QT00076874), JUN1 (QT00242956), PIK3R1 (QT00023100), MAPK8 (QT01149512), NLK (QT00043148), COL1A2 (QT00072058), VCAN (QT00064064), PDE1A (QT00065877), MOSC1 (QT00084665), and CRIP3 (QT00089852) were purchased from Qiagen (UK). Amplification and detection of specific products were carried out with Roche Lightcycler 1.5 detection System. Each sample was performed in duplicate. 18S mRNA was used as endogenous control transcript in each sample. Relative expression ratios were calculated by the relative quantification real-time PCR method (34). Statistical analysis (unpaired t-test) was carried out using Instat 3 software, and a P value was calculated for comparison.

Western blotting.

For each experimental group (Hyperoxic/standard, n = 5 and controlled reoxygenation, n = 5), total proteins from individual patients were extracted. Western blotting was performed as previously described (15). Briefly, total protein extracts were prepared from ventricular biopsies, separated on SDS polyacrylamide gel, and transferred to Hybond nitrocellulose membrane (Amersham). Blocked membrane were incubated in primary polyclonal antibodies (rabbit anti-MOSC1, Lifespan Biosciences; rabbit anti-COL1A2, abcam; goat anti-TAUT, Santa Cruz Biotechnology), washed, and incubated in horseradish peroxidase-coupled anti-rabbit or anti-mouse secondary antibodies (Amersham). Membranes were exposed to Hyperfilm (Amersham), and protein bands were quantified with NIH Image J software. Statistical analysis (unpaired t-test) was carried out using Instat 3 software, and a P value was calculated for comparison.

RESULTS

Operative and postoperative data for both groups are summarized in Table 1. The two groups were comparable in terms of age, weight, CPB, and cross-clamp time. The pro-operative echocardiographic data show no differences between the two groups regarding right ventricular wall thickness. As expected, the blood oxygen levels were significantly increased in the hyperoxic/standard compared with the controlled reoxygenation CPB group (P = 0.01) demonstrating the efficacy of the CPB treatment strategy (Fig. 1A).

Fig. 1.

Oxygen pressure (A) and 8-isoprostane level (B) changes up to 24 h postoperatively in the hyperoxic/standard compared with the controlled reoxygenation cardiopulmonary bypass (CPB) groups. Time 1, preoperatively; time 2, X-clamp off; time 3, 30 min X-clamp off; time 4, 2 h X-clamp off; time 5, 6 h X-clamp off; time 6, 24 h X-clamp off. **P < 0.01 and *P < 0.05 vs. time 1 in the Hyperoxic/Standard group.

There were no deaths and no major morbidities in both groups. Patients receiving hyperoxic/standard CPB had a longer ventilation time (P = 0.03) and duration of dopamine support (P = 0.05) compared with patients in the controlled reoxygenation CPB group. Postoperative lactate levels were significantly raised at 6 h from the end of the cross clamp time compared with preoperative levels in the hyperoxic/standard CPB group (P < 0.05).

The 8-isoprostane levels were also significantly increased at the end of the cross clamp time compared with preoperative levels in the hyperoxic/standard group, remaining elevated up to 2 h from the end of the ischemic time (P < 0.05, Fig. 1B).

Gene expression.

We have used the Affymetrix GeneChip Human Genome U133 Plus 2.0 Array in this study as it represents the most comprehensive whole human genome expression array. It provides comprehensive coverage of the transcribed human genome on a single array with variants from >47,000 well-substantiated human genes (http://www.affymetrix.com).

Genes expressed in cyanotic heart tissue before and after hyperoxic/standard or controlled reoxygenation CPB.

The comparison of gene expression profiles in the cyanotic hearts before and after hyperoxic/standard CPB revealed 35 differentially expressed genes with three upregulated and 32 downregulated (Fig. 2). The products of upregulated genes included two E3 Ubiquitin ligases HECTD1 and DTX3 (Table 2). The products of downregulated genes included intracellular signaling kinases MAPK8 and NLK (Wnt signaling pathway), metabolic process proteins (KLF9, biliverdin reductase A, PPAPDC1B), and transport factors: SLC39A8, SLC25A30 (mitochondrion protein) and GULP (Table 2). In contrast, the comparison of gene expression profiles before and after controlled reoxygenation CPB revealed only 11 differentially expressed genes with 10 upregulated including extracellular matrix proteins (COL1A2 and IDE), transport factors, and one downregulated (Cortactin; Fig. 2, Table 3). Controlled reoxygenation reduced the number of genes whose expression was altered following hyperoxic/standard CPB.

Fig. 2.

A: summary of the regulated genes (≥1.6-fold) in post-op vs. pre-op controlled reoxygenation, post-op vs. pre-op hyperoxic procedure, and post-op hyperoxic vs. post-op controlled reoxygenation. B: regulated transcripts in the 3 comparisons. The histogram shows the proportion of upregulated vs. downregulated genes.

Table 2.

Genes exhibiting 1.6-fold or greater expression change in ventricular biopsy of post-op vs. pre-op hyperoxic patients

Gene	Symbol	NCBI Gene ID	Fold in Post-op	P Value
HECT domain containing 1	HECTD1	25831	1.791 up	0.0232
Zinc finger protein 106 homolog	ZFP106	64397	1.703 up	0.0172
Deltex 3 homolog (Drosophila)	DTX3	196403	1.674 up	0.00172
Solute carrier family 25, member 30	SLC25A30	253512	2.139 down	0.00218
Brain expressed X-linked 2	BEX2	84707	2.032 down	0.0198
Prion protein (p27–30)	PRNP	5621	1.993 down	0.0117
EF-hand calcium binding domain 2	EFCAB2	84288	1.976 down	0.0159
PAP associated domain containing 1	PAPD1	55149	1.890 down	0.0181
GULP, engulfment adaptor PTB domain containing 1	GULP1	51454	1.889 down	0.0178
Ubiquitin-like modifier activating enzyme 6	UBA6	55236	1.839 down	0.0172
TRK-fused gene	TFG	10342	1.770 down	0.0147
Biliverdin reductase A	BLVRA	644	1.759 down	0.00761
KIN, antigenic determinant of recA protein homolog	KIN	22944	1.732 down	0.00693
CDNA clone IMAGE:5286843			1.724 down	0.0203
Mesoderm specific transcript homolog	MEST	4232	1.695 down	0.00819
Solute carrier family 39 (zinc transporter), member 8	SLC39A8	64116	1.678 down	0.00967
Branched chain aminotransferase 1, cytosolic	BCAT1	586	1.672 down	0.0105
Heparan sulfate (glucosamine) 3-O-sulfotransferase 1	HS3ST1	9957	1.672 down	0.0176
Nemo-like kinase	NLK	51701	1.672 down	0.0249
Mitogen-activated protein kinase 8	MAPK8	5599	1.669 down	0.00734
RNA binding motif protein 7	RBM7	10179	1.664 down	0.0179
			1.662 down	0.00417
Ras homolog gene family, member U	RHOU	58480	1.660 down	0.0165
MIT, microtubule interacting and transport, domain containing 1	MITD1	129531	1.660 down	0.0201
CDNA FLJ40901 fis, clone UTERU2003704			1.637 down	0.00731
TBP-like 1	TBPL1	9519	1.635 down	0.0249
Calcyclin binding protein	CACYBP	27101	1.631 down	0.0234
Carbohydrate (chondroitin) synthase 1	CHSY1	22856	1.630 down	0.0229
PHD finger protein 14	PHF14	9678	1.626 down	0.0229
Kruppel-like factor 9	KLF9	687	1.618 down	0.0201
Ribonuclease H2, subunit B	RNASEH2B	79621	1.617 down	0.0222
ATG10 autophagy related 10 homolog	ATG10	83734	1.606 down	0.00197
Chromosome 21 open reading frame 51	C21orf51	54065	1.606 down	0.0154
Phosphatidic acid phosphatase type 2 domain containing 1B	PPAPDC1B	84513	1.603 down	0.0212
COP9 constitutive photomorphogenic homolog subunit 8	COPS8	10920	1.600 down	0.0109

Table 3.

Genes exhibiting 1.6-fold or greater expression change in ventricular biopsy of post-op vs. pre-op controlled reoxygenation patients

Gene	Symbol	NCBI Gene ID	Fold in Post-op	P Value
CDNA FLJ38472 fis, clone FEBRA2022148			2.246 up	0.00654
Ankyrin 3, node of Ranvier (ankyrin G)	ANK3	288	1.935 up	0.0205
Ubiquitously transcribed tetratricopeptide repeat, X chromosome	UTX	7403	1.856 up	0.0123
HECT domain containing 1	HECTD1	25831	1.815 up	0.0232
Golgi autoantigen, golgin subfamily a, 8B	GOLGA8B	440270	1.807 up	0.0228
CDNA clone IMAGE:5263531			1.783 up	0.021
Collagen, type I, alpha 2	COL1A2	1278	1.775 up	0.00884
RAN binding protein 2	RANBP2	5903	1.752 up	0.0206
Phosphoinositide-3-kinase, class 2, alpha polypeptide	PIK3C2A	5286	1.621 up	0.0126
Insulin-degrading enzyme	IDE	3416	1.617 up	0.0129
Cortactin	CTTN	2017	1.609 down	0.0157

Comparing gene expression in cyanotic heart tissue following hyperoxic/standard versus controlled reoxygenation CPB.

Our data revealed 59 differentially expressed genes in hyperoxic-standard vs. controlled reoxygenation CPB, with six upregulated and 53 downregulated (Fig. 2). The upregulated genes included PDE1A, MOSC1, and CRIP3 (Table 4). The downregulated genes were functionally clustered into four major classes: extracellular matrix/cell adhesion, transcription, transport, and cellular metabolic process. They included Versican, Collagen I alpha 2, JUN, KLF9, SLC6A6, and BCAT1 (Table 4).

Table 4.

Genes exhibiting 1.6-fold or greater expression change in ventricular biopsy of post-op hyperoxic vs. postop controlled reoxygenation patients

Gene	Symbol	NCBI Gene ID	Fold in Hyperoxic	P Value
Deiodinase, iodothyronine, type II	DIO2	1734	1.900 up	0.0235
Phosphodiesterase 1A, calmodulin-dependent	PDE1A	5136	1.813 up	0.00774
Actin, alpha 1, skeletal muscle	ACTA1	58	1.755 up	0.0168
MOCO sulphurase C-terminal domain containing 1	MOSC1	64757	1.740 up	0.0168
Cysteine-rich protein 3	CRIP3	401262	1.700 up	0.00171
GABA(A) receptor-associated protein like 1	GABARAPL1	23710	1.612 up	0.0123
CDNA FLJ38472 fis, clone FEBRA2022148			3.167 down	0.00935
CDNA FLJ40901 fis, clone UTERU2003704			2.658 down	0.00731
Branched chain aminotransferase 1, cytosolic	BCAT1	586	2.644 down	0.0105
CDNA FLJ38472 fis, clone FEBRA2022148			2.512 down	0.00654
Versican	VCAN	1462	2.365 down	0.0106
Mesoderm specific transcript homolog	MEST	4232	2.135 down	0.00819
GULP, engulfment adaptor PTB domain containing 1	GULP1	51454	2.031 down	0.0121
Inturned planar cell polarity effector homolog	INTU	27152	2.017 down	0.0157
Solute carrier family 25, member 30	SLC25A30	253512	1.982 down	0.00218
Neuroligin 4, X-linked	NLGN4X	57502	1.954 down	0.00423
Hypothetical protein LOC728555 /// hypothetical protein LOC730391	LOC728555 /// LOC730391	728555 /// 730391	1.927 down	0.0194
Ras homolog gene family, member U	RHOU	58480	1.895 down	0.0165
Neurotrimin	HNT	50863	1.876 down	0.0123
Vacuolar protein sorting 35 homolog	VPS35	55737	1.853 down	0.0217
TIA1 cytotoxic granule-associated RNA binding protein	TIA1	7072	1.791 down	0.0127
Glycoprotein M6B	GPM6B	2824	1.789 down	0.0137
Collagen, type I, alpha 2	COL1A2	1278	1.782 down	0.00884
Insulin-degrading enzyme	IDE	3416	1.755 down	0.00773
Calcitonin receptor-like	CALCRL	10203	1.747 down	0.00625
Asparagine-linked glycosylation 13 homolog	ALG13	55849	1.745 down	0.00513
Ankyrin 3, node of Ranvier (ankyrin G)	ANK3	288	1.735 down	0.0205
ATG10 autophagy related 10 homolog	ATG10	83734	1.722 down	0.00197
Fer-1-like 3, myoferlin	FER1L3	26509	1.719 down	0.00849
Ubiquitin-like modifier activating enzyme 6	UBA6	55236	1.708 down	0.0172
Solute carrier family 30 (zinc transporter), member 7	SLC30A7	148867	1.707 down	0.00933
Quinolinate phosphoribosyltransferase (nicotinate-nucleotide pyrophosphorylase (carboxylating))	QPRT	23475	1.706 down	0.0212
CDNA clone IMAGE:5270438			1.698 down	0.00574
Solute carrier family 35 (UDP-glucuronic acid/UDP-N-acetylgalactosamine dual transporter), member D1	SLC35D1	23169	1.697 down	0.021
Tetratricopeptide repeat domain 3	TTC3	7267	1.697 down	0.0118
Zinc finger, SWIM-type containing 7	ZSWIM7	125150	1.692 down	0.0242
Ectonucleoside triphosphate diphosphohydrolase 1	ENTPD1	953	1.691 down	0.0123
Zinc finger protein 302	ZNF302	55900	1.684 down	0.0233
Tau tubulin kinase 2	TTBK2	146057	1.682 down	0.0199
Kruppel-like factor 9	KLF9	687	1.679 down	0.0201
Family with sequence similarity 114, member A1	FAM114A1	92689	1.678 down	0.000927
KIAA1199	KIAA1199	57214	1.670 down	0.0113
Cdc2-related kinase, arginine/serine-rich	CRKRS	51755	1.667 down	0.00753
Zinc finger protein 397	ZNF397	84307	1.659 down	0.0207
Solute carrier family 6 (neurotransmitter transporter, taurine), member 6	SLC6A6	6533	1.658 down	0.015
Transcribed locus			1.655 down	0.0101
Pleckstrin homology domain interacting protein	PHIP	55023	1.654 down	0.0156
Interleukin 1 receptor-like 1	IL1RL1	9173	1.652 down	0.0112
			1.648 down	0.00417
Cytidine monophosphate-N-acetylneuraminic acid hydroxylase (CMP-N-acetylneuraminate monooxygenase) pseudogene	CMAH	8418	1.646 down	0.00472
CDNA FLJ36491 fis, clone THYMU2018197			1.645 down	0.0147
NIMA (never in mitosis gene a)-related kinase 3	NEK3	4752	1.638 down	0.00289
CDNA FLJ39389 fis, clone PLACE6003621			1.631 down	0.00536
Jun oncogene	JUN	3725	1.630 down	0.016
Neuroblastoma RAS viral (v-ras) oncogene homolog	NRAS	4893	1.621 down	0.0189
Aldehyde dehydrogenase 1 family, member A2	ALDH1A2	8854	1.619 down	0.00912
Zinc finger protein 347	ZNF347	84671	1.618 down	0.00557
Ankyrin repeat and SOCS box-containing 15	ASB15	142685	1.617 down	0.00804
Brain expressed X-linked 2	BEX2	84707	1.600 down	0.0198

Gene ontology biologic process annotations.

Functional annotation clustering of the genes, identified as downregulated following hyperoxic/standard compared with controlled reoxygenation CPB, unmasked four overrepresented major classes involved in myocardium normal function (Table 5). Indeed, cell adhesion/extracellular matrix, transport, transcription, and cellular metabolic process were downregulated in cyanotic myocardium. These ontological changes indicate a decrease in the adaptation and remodeling capacity of cyanotic hearts subjected to hyperoxic/standard compared with controlled reoxygenation CPB.

Table 5.

Gene Ontology enrichment

Pathway or GO ID	Term	Probes, n	P Value
KEGG	Erb signaling pathway	3	3.0E-2
GO:0031012	extracellular matrix	3	2.1E-1
GO:0007155	cell adhesion	4	4.3E-1
GO:0032268	regulation of cellular protein metabolic process	5	7.7E-2
GO:0006351	transcription	7	5.1E1
GO:0015031	protein transport	3	7.4E-1

Validation of microarray results with real-time PCR.

Microarray results were confirmed by real-time quantitative PCR on 10 selected genes that demonstrated high differential gene expression. The changes in expression levels of these genes showed a high degree of correlation with the microarray data (Fig. 3). Four out of the 10 examined genes showed a tendency to change similar to that of microarray experiment although not quite significant (Fig. 3).

Fig. 3.

Confirmation of the microarray results of changed genes in hyperoxic vs. controlled reoxygenation procedures. Changes in mRNA expression of COL1A2, VCAN, MAPK8, NLK, SLC6A6, NRAS1, PDE1, MOSC1, and CRIP3 were verified in 10 patients (5 using hyperoxic and 5 using controlled reoxygenation procedure) by quantitative real-time-PCR. Results are shown as means (±SE) fold-change compared with hyperoxia. *P < 0.05, **P < 0.01.

Western blotting of a selection of the identified genes' products.

In an attempt to predict protein expression levels of some of the identified genes, we went onto assessing the protein levels of these genes' products in the myocardium of children subjected to hyperoxic/standard and controlled reoxygenation CPB using Western blotting. Protein levels of MOSC1, TAUT, and COL1A2 showed no alteration in the myocardium of the two groups of patients (Fig. 4).

Fig. 4.

Protein levels of MOSC1, TAUT, and COL1A2 in ventricular biopsies of controlled reoxygenation vs. hyperoxic patients. Tissues were lysed to isolate protein content and Western blotting analysis performed probing for MOSC1 (A), TAUT (B), COL1A2 (C), and GAPDH. No significant changes were observed in controlled reoxygenation (C) compared with hyperoxic (H) samples. MOSC1, CRIP3, and COL1A2 bands were normalized to GAPDH levels. Data are means ± SE.

DISCUSSION

To our knowledge, this study is the first attempt to examine transcriptomic changes following the use of a novel strategy of controlled reoxygenation CPB in corrective heart surgery of cyanotic children. Our study confirmed that high levels of oxygen on starting CPB are associated with significant gene expression changes that are not seen when the reoxygenation is controlled. This was also associated with higher release of 8-isoprostane in the hyperoxic/standard CPB group, indicating an overall reduction in oxidative stress when oxygen levels are controlled on starting CPB in cyanotic patients. We have recently shown that controlling reoxygenation on starting CPB in cyanotic children undergoing cardiac surgery is associated with reduced reperfusion injury and oxidative stress compared with hyperoxic/standard CPB (5). There is substantial experimental evidence for the existence of a myocardial reoxygenation injury in the cyanotic, immature heart. Studies of hypoxemia/reoxygenation in immature piglets have provided evidence for oxygen-mediated myocardial injury as a result of hyperoxemia in the setting of previous cyanosis (33). The immature heart has a higher tolerance to hypoxia, but preceding hypoxia prior to ischemic cardioplegic arrest results in poor functional recovery and is associated with derangements in several metabolites (19, 23). Hypoxia also reduces the antioxidant reserve capacity, leading to a greater susceptibility to the oxidative stress of ischemia (e.g., cross-clamping of the aorta) and reoxygenation (12, 18, 38). Re-introduction of oxygen exacerbates this situation and eventually leads to a reduction in myocardial contractility (1, 12, 13, 16, 18, 38).

One major finding in our study is that controlled reoxygenation reduces the transcriptomic alterations observed following hyperoxic CPB. Most of the genes altered by hyperoxic CPB were downregulated genes involved in intracellular signaling, metabolic process, and transport, suggesting that hyperoxic CPB can have a deleterious effect on the myocardium as shown by the reduction in expression of genes involved in important cell processes. In contrast, most of the genes altered by controlled reoxygenation CPB were upregulated genes implicated in extracellular matrix remodeling, suggesting that this strategy could improve myocardium remodeling by increasing extracellular matrix proteins.

Most of the differentially expressed genes in hyperoxic/standard versus controlled reoxygenation CPB were downregulated. These genes functionally clustered into four major classes: extracellular matrix/cell adhesion, transcription, transport, and cellular metabolic process. These findings suggest a decrease in the adaptation and remodeling capacity of cyanotic hearts subjected to hyperoxic compared with controlled reoxygenation CPB. Additionally, a number of CDNA clones seem to play a role in reoxygenation injury. However, these clones are not well characterized.

It seems that the short time (70 min) between the starting of surgery and taking the second biopsy was enough to see the alterations at the mRNA levels observed by GeneChip and real-time PCR. However, this same time was too short to see any significant changes at the protein levels of the selected proteins: MOSC1, TAUT, and COL1A2. It is known that there's usually a delay between transcription and translation that can be hours to days. Because of the limitation of working on humans, it is not possible to examine the protein levels in human myocardium after hours or days from hyperoxic/standard or controlled reoxygenation CPB.

MOSC1 has a sulfite oxidase, a xanthine dehydrogenase and an aldehyde oxidase activities. Interestingly, the aldehyde oxidase (AO) produces hydrogen peroxide and, under certain conditions, can catalyze the formation of superoxide (14, 25, 26). When molecular oxygen acts as an electron acceptor in the AO-catalyzed oxidation of aldehydes or azaheterocycles, it undergoes a two-electron reduction to produce H₂O₂; however, a portion undergoes one-electron reduction to produce superoxide. Our study shows an augmentation of MOSC1 mRNA levels by hyperoxic/standard CPB, suggesting a possible role of this enzyme in the increase of oxidative stress observed when this CPB procedure is used. Very early studies have already demonstrated the ability of this enzyme to generate superoxide and suggested a role for it in reperfusion injury (29). More recent investigations have shown that MOSC1 is an important source of both superoxide and H₂O₂ in biological tissues (28).

Another interesting gene is the taurine transporter (SLC6A6, or TAUT). It is thought that taurine play an important role in ion movement, calcium handling, osmoregulation, and cytoprotection. Taurine depletion causes cardiomyocyte atrophy, mitochondrial and myofiber damage, and cardiac dysfunction, effects likely related to the actions of taurine (22). Indeed this investigation suggests that multiple actions of taurine, including osmoregulation, regulation of mitochondrial protein expression, and inhibition of apoptosis, collectively ensure proper maintenance of cardiac and skeletal muscular structure and function (22). Our data showed a reduction in mRNA levels by hyperoxic/standard CPB, suggesting a deleterious effect of this CPB protocol to the myocardium occurring probably by reduction of taurine cytoprotective action.

Clinical implications.

One of the strategies proposed to avoid reoxygenation injury is the use of controlled reoxygenation during CPB. Allen and associates (1) demonstrated an improvement in the antioxidant reserve capacity with lower levels of oxygen in cyanotic infants undergoing cardiac surgery, and Bulutcu and coworkers (4) obtained similar findings. Our results confirm these findings and provide the first direct evidence that an unintended oxygen-mediated myocardial genomic alteration injury occurs in cyanotic patients with TOF undergoing surgical repair. Reoxygenation injury is associated with the production of reactive oxygen species by the formation of oxygen-derived free radicals such as superoxide and peroxide. This leads to cell membrane degradation through lipid peroxidation (12, 18). It is now evident that oxidative stress is a major part of the cellular mechanism of the resulting myocardial damage (3, 10, 30, 35). 8-Isoprostane has been shown to be a reliable marker for the volume of myocardium exposed to oxidant stress during acute myocardial infarction (30, 35), as well as a quantitative marker of oxidant stress during coronary reperfusion (10). The increased 8-isoprostane levels in the hyperoxic/standard CPB group associated with the profound downregulation of key genetic pathways related to myocardial function suggests a direct effect of CPB oxygen levels on the degree of reperfusion injury and can possibly explain the increased need of postoperative inotropic support observed in these patients. The perfusion strategy of keeping Po₂ as close as possible to the patient's preoperative values resulted in a reduced oxidative stress and in a significant reduction in the genomic alterations observed with the hyperoxic/standard CPB group.

This novel and simple strategy of controlling reoxygenation on starting CPB does not interfere with the surgical procedure and, by limiting oxidative stress and reoxygenation injury, might lead to improvements in early clinical outcomes in the very high risk group of cyanotic infants and children undergoing cardiac surgery.

Limitations.

Our study cannot detect differences in clinical outcome between the two groups; the primary outcomes were related to differences in myocardial gene expression and biochemical markers of organ dysfunction. Recruitment to a larger trial to evaluate clinical outcomes as primary end points is ongoing at our institution.

This study investigated a single congenital pathology, and its findings cannot automatically be related to other cyanotic cardiac conditions.

GRANTS

This work was funded by the British Heart Foundation, the Garfield Weston Trust, the Bristol NIHR BRU in Cardiovascular Medicine, and the BUPA Foundation. M. T. Ghorbel was supported by an Intermediate Research Fellowship from the British Heart Foundation.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

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