| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,
,1
* Inserm 602; Service de Biochimie et Biologie Moléculaire; Hôpital Universitaire Paul Brousse; Université Paris-Sud/XI, Villejuif Cedex; Assistance Publique-Hôpitaux de Paris, France;
Division of Biostatistics, University of California, Berkeley, USA,
Inserm U444; Université Paris-VI; Hôpital Tenon, Paris;
Centre Hépato-Biliaire; Hôpital Paul Brousse, Villejuif; and
|| Service d’Anatomie Pathologie; Hôpital Ambroise Paré, Boulogne-Billancourt, France
2 Correspondence: Biochimie et Biologie moléculaire, Hôpital Universitaire Paul Brousse 94800, Villejuif, France. E-mail: antoinette.lemoine{at}pbr.ap-hop-paris.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
= 2.5 106 (Bonferroni correction). We found that 60 genes were differentially expressed (36 over- and 24 underexpressed in preconditioning group). After IP, the most significantly overexpressed gene was IL-1Ra. This was confirmed by immunoblotting. Differentially expressed were genes involved in apoptosis (NOD2, ephrin-A1, and calpain) and in the carbohydrate metabolism. A significant increase in the amount of the anti-apoptotic protein Bcl-2 in preconditioned livers but no change in the cleavage of procaspase-3, -8, and -9 was observed. We also observed an increase in the amount in the inducible nitric oxide synthase. Therefore, the benefits of IP may be associated with the overproduction of IL-1Ra, Bcl-2, and NO countering the proinflammatory and proapoptotic effects generated during ischemia-reperfusion.—Barrier, A., Olaya, N., Chiappini, F., Roser, F., Scatton, O., Artus, C., Franc, B., Dudoit, S., Flahault, A., Debuire, B., Azoulay, D., Lemoine, A. Ischemic preconditioning modulates the expression of several genes, leading to the overproduction of IL-1Ra, iNOS, and Bcl-2 in a human model of liver ischemia-reperfusion.
Key Words: preconditioning • liver transplantation • gene expression • apoptosis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
. Cells from liver and other organs also undergo apoptosis after ischemia-reperfusion (3
4
5
6)
. In particular, ischemic injury of parenchymal cells results in a loss of mitochondrial respiratory chain activity and consequently ATP depletion (7
, 8)
. Sinusoidal endothelial cells are more susceptible than hepatocytes to hypothermic ischemia (6
, 9)
.
Ischemic preconditioning, which was first described for the heart by Murry et al. (10)
, is an endogenous mechanism consisting of brief periods of vascular occlusion that protects against subsequent ischemia reperfusion. Several experimental studies have reported that ischemic preconditioning has a beneficial effect on cold ischemic injury of different organs including heart, intestine, lung, and kidney (11
12
13
14
15
16
17
18
19)
. Despite intensive investigations, the mechanisms underlying this protective effect remain unclear. However, it has been suggested that the benefits of ischemic preconditioning could be mediated by the synthesis of vasoactive mediators, such as nitric oxide (19
20
21
22)
. In human liver surgery, ischemic preconditioning reduces the severity of postoperative hepatic injury as indicated by lower transaminase concentrations and less severe endothelial cell injury (23)
. TUNEL assays and electron microscopy have shown that this protective effect is linked to decreased apoptosis and the preservation of ATP content in liver tissue (24)
. Ischemic preconditioning may also attenuate liver graft injury by decreasing apoptosis of hepatocytes and the production of TNF- (25)
.
We studied the genetic changes caused by ischemic preconditioning during hepatectomy in humans. For this, we carried out microarray analysis using mRNA from liver biopsy samples randomly assigned to one of two groups: ischemic preconditioning before standardized partial hepatectomy and no ischemic preconditioning. The patients were living donors who donated liver tissue to a family member.
|
PATIENTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Surgical procedures
Procurement of liver graft
The right half of the liver was harvested in all cases. In brief, the graft included the right liver parenchyma with the right branch of the hepatic artery, the right branch of the portal vein and the right bile duct(s), and the right hepatic vein. There was no vascular clamping during the procedure and the middle hepatic vein was sometimes included in the graft.
Orthotopic liver transplant in the recipient
The transplant was performed as described previously. In brief, the native liver was totally removed with caval preservation (26)
. A temporary portacaval shunt was performed if deemed necessary by the transplant surgeon (27)
. The partial liver graft was then implanted. Cold ischemia time (CIT) was considered as the time between devascularization in the donor and portal reperfusion in the recipient. A liver biopsy was performed before closure of the abdomen to evaluate the degree of ischemia-reperfusion injury (postreperfusion biopsy). Ischemia-reperfusion injury was usually classified as moderate to severe (vs. absent) when at least 10% of hepatocytes were necrotic. It is mainly observed in the center of the lobule or disseminated throughout. No ischemia reperfusion injury was observed in any of the 14 biopsies. The degree of microvacuolar steatosis within hepatocytes was also evaluated by two independent pathologists. None of the biopsies exhibited steatosis.
RNA extraction and microarray analysis
A fragment of the liver biopsy was stored at –80°C until RNA was isolated. RNA was extracted using an RNeasy kit (Qiagen, Courtaboeuf, France) according to the protocol of the manufacturer. Integrity of the RNA was confirmed by running it on an Agilent bioanalyzer (Palo Alto, CA, USA). We obtained a sufficient quantity of good-quality RNA for microarray studies from all subjects, seven in each group. Total RNA (20 µg) was processed as described in the CyScribe Post Labeling Kit protocol (Amersham Biosciences, Buckinghamshire, UK). This protocol consists of two steps. The first step involves the incorporation of amino allyl-dUTP (AA-dUTP) during cDNA synthesis using an optimized nucleotide mix. The second step involves chemically labeling the amino allyl-modified cDNA with CyDye NHS-esters. As we used total eukaryotic RNA as a template for cDNA synthesis, we omitted the random nonamers to avoid copying rRNA. Priming with 3 µL anchored oligo (dT) directed the start of cDNA synthesis from the 5' end of the poly A tail. We carried out dual color microarray hybridization in which amino allyl-modified cDNA was coupled with Cy3 and Cy5 separately and the two probes were combined in the hybridization solution. This enabled simultaneous detection of hybridization signals and comparison of gene expression levels. We used Agilent® Human 1A OligoMicroarray (v2) (Agilent Technologies, Palo Alto, USA), designed to truly represent genes in the human genome. These arrays consist of 20,173 60-mer oligonucleotide probes, spanning conserved exons across the transcripts of the targeted full-length genes, covering a total of 18,716 well-known human genes. Microarrays were processed as described in the Agilent® 60-mer oligo microarray processing protocol. Arrays were scanned with a GenePix® Personal 4100A (Axon Instruments, Union City, USA), using the GenePix® Pro v4.1.1.40 software.
For each slide, mRNA from a patient in the preconditioning group was randomly labeled with Cy3 or Cy5 and mRNA from a patient in the control group was labeled with the other dye. This was repeated seven times. Thus, we excluded the consequences of ischemia-reperfusion and assessed only those of the preconditioning step.
Analysis of microarray data
Fluorescence intensities were first converted to numerical values using the ScanAlyse software (28)
. These raw values were then subjected to a log2 transformation and normalized using a robust, locally weighted regression procedure (29
, 30)
. In contrast to global normalization methods, local normalization methods take into account potential space- and intensity-dependent dye biases. The "marray" tool of the R Bioconductor package was used for the normalization (maNormMain, location normalization function "loess").
Differences in the expression of each sequence between the two groups were evaluated by a paired t test. The Bonferroni procedure for multiple testing was applied. Expression levels were considered to differ if P< 2.5 x 106. The "multest" tool of the R Bioconductor package (Ge, Y., Dudoit, S.) was used for statistical analysis.
Western blot analysis
Whole protein extracts were prepared as follows: 20 µg of frozen samples were homogenized in RIPA buffer at 4°C. This buffer contained 10 mM Tris pH 7.4, 150 mM NaCl, 0.02% sodium azide, 1% Triton X-100, 0.5% DOC, 1% SDS, 1M EDTA, 10 mg/mL aprotinin, 10 mg/mL leupeptin, 1M phenylmethylsulfonyl fluoride, 50 M NaF, and 1 mM DTT. Mixtures were centrifuged and supernatants collected.
Protein concentration was measured by the BCA assay (Pierce Biotechnology, Rockford, IL, USA). Aliquots of protein (50 µg) were boiled and subjected to SDS-PAGE on a 4–15% gradient precast gel (Bio-Rad Laboratories, Hercules, CA, USA) or on a 15% linear gel. For each antibody, the eight samples were run on the same gel to allow the comparison. Resolved proteins were transferred on to a PVDF membrane. Nonspecific interactions were blocked by incubating the membranes with 5% nonfat milk/TBS-Tween 0.1% for 1 h at room temperature.
The membranes were probed with the following antibodies: mouse monoclonal anti-caspase 3, clone 31A1067; mouse monoclonal anti-caspase 8 clone 12F5 (Alexis Biochemical Corp., Lausen, Switzerland); polyclonal anti-caspase 9 (Cayman chemical, Ann Arbor MI, USA); mouse anti-NOS2, iNOS clone 54 (BD Transduction Laboratories, San Diego, CA, USA); mouse anti-human Bcl-2 oncoprotein, clone 124, (DAKO, Glostrup, Denmark); polyclonal anti-NOS3, eNOS c-20 (Santa Cruz Biotechnology, Santa Cruz, CA, USA); goat anti human IL-1Ra (R&D systems Inc. Minneapolis, MN, USA); monoclonal anti-heat shock protein 70 (HSP70) and mouse monoclonal anti-ß-actin, clone AC-74 (Sigma-Aldrich, Saint Louis, MO, USA). After overnight incubation with each primary antibody, the membranes were incubated with horseradish peroxidase-conjugated IgG (DAKO, Glostrup, Denmark) for 1 h at room temperature. The proteins were then detected using an ECL Western blotting detection system (Pierce Biotechnology).
Images were captured with a FujiFilm Intelligent Dark Box II coupled to the IRLAS-1000 pro V2.5 software (FujiFoto Film Co. Ltd.). The density of the bands was compared using Image Gauge V4.0 software (FujiFoto Film Co. Ltd). For comparison, bands were delineated into squares of identical surface area and the intensity of signal was measured. The global background signal was subtracted. Medians and standard deviations were calculated and for Bcl-2, NOS2, and IL-1Ra results, Wilcoxon-Mann-Whitney nonparametric tests were performed.
Immunohistochemistry
Paraffin embedded tissue sections (4 µm) were dewaxed, rehydrated, and treated with 0.3% hydrogen peroxide in methanol. The sections were then incubated with blocking solution and then incubated overnight at 4°C with goat antihuman IL-1Ra antibody (dilution 1:25; R&D systems, Minneapolis, MN, USA). Sections were incubated with a biotin-labeled secondary antibody (DAKO, Glostrup, Denmark) and peroxidase conjugated avidine was added. Diaminobenzidin was used as the chromogen and hematoxylin as the nuclear counter stain. For negative controls, the primary antibody was either omitted or replaced by a suitable concentration of normal IgG of the same species.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Age, sex ratio, hepatic glycogen content, and cold ischemia time of the liver graft were similar in the two groups. None of the donors presented liver steatosis and all had high hepatic glycogen content.
|
Genes differentially expressed in liver tissues that had and had not been subjected to ischemic preconditioning before partial hepatectomy
Microarray data were analyzed using paired t test with a type I error rate fixed at = 2.5 x 106 (Bonferroni correction). Table 2
shows the sequences identified as being differentially expressed in the two groups. Sixty sequences were found to be differentially expressed. Of these, 36 were overexpressed and 24 were underexpressed in the Group Precond.
|
The gene encoding IL-1Ra was the most significantly overexpressed in preconditioned livers. We also observed an increase (in Group Precond) in the expression of genes involved in apoptosis, such as one coding for the NOD2 protein, a member of the NOD (nucleotide binding oligomerization domain, CARD15) family of apoptosis regulators. The gene encoding a serine threonine protein kinase that inhibits apoptosis (serum glucocorticoid regulated kinase) was also overexpressed in Group Precond as was the gene encoding a member of the TNF-R family (TNFRSF11A) that has never been shown to be involved in apoptosis. Conversely, the genes encoding Ephrin-A1, a related tyrosine kinase receptor induced by TNF- in endothelial cells, and the genes encoding apoptosis regulators such as HSP70 and for STAT1, which induce apoptosis, were underexpressed in Group Precond. The proapoptotic calpain gene was also underexpressed in this group.
Ischemic preconditioning also induced genes that encode proteins related to carbohydrate metabolism, particularly to glycolysis (enolase 3) and glucose oxidation (pyruvate dehydrogenase kinase 4). The products of the under- and overexpressed genes in this group are implicated in the metabolism of glucose, insulin, and lipids such as apolipoprotein AV, IGFBP1, ATP binding cassette subfamily A member 1, which plays a role in the transport of cholesterol and phospholipids, and cytochrome P-450 46 (cholesterol 24-hydroxylase).
Finally, we found that several genes encoding proteins involved in cell growth and/or differentiation were differentially expressed in the two groups. These genes notably included genes encoding proteins required for the GO-G1 switch and the cyclin E2 gene, which codes for a protein that interacts with CDK2 and CDK3 during G1. Several genes involved in the negative regulation of the cell cycle were underexpressed in Group Precond. For example, the transducer of ERBB2 1, encoding an anti-proliferative protein that interacts with ERRB2, thus negatively regulating cell proliferation, was found underexpressed. The mitogen-inducible gene 6, encoding a putative suppressor, which is expressed in a cell cycle stage-dependent manner, was also down-regulated in Group Precond. This gene mediates ERBB2 signaling and activates human MAPK10.
We have also observed a change in the expression of genes implicated in the heme synthase, or numerous genes coding for several zinc-finger containing proteins and ion/transporters genes.
Influence of ischemic preconditioning on protein expression
We used Western blot analysis to confirm the overexpression of IL-1Ra and to analyze the production of major proteins involved in apoptosis, such as Bcl-2, caspase-3, caspase-8 and caspase-9. Sufficient protein for Western blot analysis was available for four of the seven pairs of randomized liver tissues. An increase of the icIL-1Ra type II form of IL-1Ra was seen in GroupPrecond (three times greater than Group Control, Fig. 1
A) Wilcoxon-Mann-Whitney nonparametric test was performed (/2=0.025). The icIL-1Ra type I form was increased for both groups of liver biopsies.
|
A rise in expression of icIL-1Ra was also clearly observed in preconditioned livers compared with nonpreconditioned livers as assessed usingimmunohistochemistry (Fig. 2
). Preconditioned liver shows Il-1Ra protein overexpressed in the cytoplasm of hepatocytes belonging to zone 1 of the acinus, near the portal tract. There was twice as much Bcl-2, an anti-apoptotic protein, in the ischemic preconditioning group than in the control group. Finally, cleavage of procaspase-8 and procaspase-9 was observed in most of the samples. However, no differences were observed between the two groups of liver biopsies. (Fig. 3
A, B). We saw no cleavage of procaspase-3.
|
|
Effect of ischemic preconditioning on nitric oxide production
As nitric oxide synthesis has been proposed to mediate the beneficial effects of ischemic preconditioning, we assessed the expression of the gene encoding the inducible nitric oxide synthase (iNOS, NOS2) that produces NO. We observed a threefold increase in the levels of iNOS in the preconditioning group than in the control group. Wilcoxon-Mann-Whitney nonparametric tests was performed (/2=0.025). (Fig. 1B
). No difference in the amount of the constitutive form, NOS3, was observed.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, and to be specifically protected against by ischemic preconditioning (24)
. The same volume of liver was taken in each case and cold ischemia time was similar in all cases. All donors had large hepatic glycogen stores; this is important as the size of glycogen stores may affect the susceptibility of parenchymal cells to ischemia injury (32
, 33)
. However, ischemia lasted long enough to trigger detectable protein damage after reperfusion.
The most differentially expressed gene between the two groups was the Il-1 receptor antagonist, IL-1Ra. Previous studies have suggested that ischemic preconditioning modulates IL-1 production in steatotic livers undergoing ischemia-reperfusion. IL-1 is known to mediate acute inflammation (34)
. However, the effects of ischemic preconditioning on IL-1Ra have not been evaluated. IL-1Ra is known to inhibit the effects of IL-1 and IL-1ß by competing for type I and type II IL-1 receptors, resulting in reduced inflammation (35)
. Various experimental models, such as renal and cardiac ischemic injury, have demonstrated a beneficial anti-inflammatory effect of IL-1Ra administered directly or by transfection (36
, 37)
. It is also reported that IL-1Ra attenuates IL-1ß-induced apoptosis in various cell cultures (38)
. Therefore, we suggest that hepatic overexpression of IL1-Ra after ischemic preconditioning directly inhibits the effects of high levels of IL-1 during ischemia-reperfusion. This results in attenuation of the inflammatory response, thus reducing liver damage, liver necrosis, and apoptosis.
In experimental models of warm and cold ischemia, NO has been shown in normal livers to be involved in the protective effects of preconditioning against hepatic ischemia-reperfusion injury (9
, 14
, 39
, 40)
. Its effect on apoptosis is complex. Apoptosis is either induced or blocked, depending the cell type, stimulation conditions, and cellular NO concentration (41)
. We found that iNOS (NOS2) production was significantly higher in the preconditioning group than in the control group. Microarray analysis identified genes very similar to those identified in a previous study using cDNA arrays of NO-induced genes in mouse hepatocytes (42)
. These include mainly signal transduction-related genes, including members of the STAT family, which are important for signaling via the INF- and IL-6 receptors (43)
. They also include genes implicated in heme synthesis and numerous genes encoding zinc finger-containing proteins. We also observed that ischemic preconditioning altered the expression of several ion transporter genes, as observed in mouse hepatocytes transformed with an adenovirus expressing human iNOS (42)
. This suggests that preconditioning may influence, via NO synthesis, the ion channel and the subsequent intracellular signal transduction in hepatocytes (44
, 45)
. The relationship between NO generation and IL-1 production in liver has been previously demonstrated in different inflammatory processes, including ischemia-reperfusion (46
, 47)
. Thus, ischemic preconditioning, through NO production and/or IL1-Ra overexpression, may reduce the release of IL-1ß, found to be increased in fatty livers (46)
. Our study may explain the results of the recent randomized clinical study by Clavien et al. (24)
, showing that tolerance to ischemia-reperfusion injury was improved after preconditioning in patients with liver steatosis.
The protective role of ischemic preconditioning against ischemia-reperfusion injury has been linked to a decrease in apoptosis in animal models and in humans (23
24
25
, 39
, 48
, 49)
. However, we showed, by Western blotting, that procaspases-8 and -9 were similarly cleaved in both groups whereas no cleavage of the procaspase-3 was seen, despite being previously reported (48)
. Unfortunately, a more accurate assessment of this process, with more biopsies and quantification of caspase-3 activity, was not possible due to the small amount of liver tissue remaining after cDNA microarray assays. The pattern of gene modulation caused by ischemic preconditioning suggests there is activation of an anti-apoptotic program in hepatocytes. However, changes in the gene expression were not significantly associated with changes in the cleavage of caspases-3, 8 and 9. The NOD2 gene, a member of the NOD family, was more strongly expressed in the preconditioning group than in the control group. NOD proteins are members of the growing family of cytosolic factors related to the apoptosis regulator Apaf-1. NOD2 enhances the apoptosis induced by caspase-9, but the significance of this proapoptotic activity is unclear (50)
. Increase in NOD2 mRNA levels may regulate the host response to pathogens, a process that may be faulty in certain inflammatory diseases. A relationship between NOD2/CARD15 and the overproduction of IL-1ß mRNA has been demonstrated in intestinal inflammation (51)
. Genes encoding Ephrin-A1, a related tyrosine kinase receptor induced by TNF- in endothelial cells, and apoptosis inducers such as HSP70 and STAT1 were found to be underexpressed in the preconditioning group. However, at the protein level, HSP70 protein was overexpressed in preconditioned livers (data not shown). The calpain 10 gene was underexpressed in this group. Calpain activation has been reported to induce apoptosis in several models (52)
. During transplantation, calpain activity has been shown to increase with the duration of cold ischemia (6)
and graft nonviability (46
, 52)
.
The anti-apoptotic mediator, Bcl-2, was overexpressed in the preconditioning group. Indeed, a protective role of Bcl-2 has been reported in several models of ischemic preconditioning, with contradictory results (53
54
55
56
57)
. Bcl-2 is the most ubiquitous anti-apoptotic molecule and protects the cell against a large variety of apoptotic stimuli, such as hypoxic stress. It acts mainly by preventing mitochondrial permeability and the release of cytochrome c. It also inhibits various caspases (58
59
60)
. Although not constitutively expressed in the liver (61)
, it has been shown that Bcl-2 can be quickly up-regulated after inflammatory stress (62
, 63)
, as in our model. The increase of Bcl-2 has also been implicated in the balance between intracellular ATP depletion after the mitochondrial permeability transition (MPT) and ATP generation by glycolysis (64)
. This balance determines whether the cell undergoes necrosis or apoptosis after the ischemia-reperfusion-induced onset of the MPT (8
, 64)
. Recent studies have shown that liver cells also undergo apoptosis after ischemia-reperfusion (6
, 65)
, whereas ischemia-reperfusion has usually been considered to be only associated with necrotic cell death (2)
. When ATP loss is massive, as may be during ischemia-reperfusion, necrotic cell death rapidly occurs. Ours microarray assay results are consistent with glycolytic ATP generation after preconditioning. As described previously by Clavien and Selzner (24
, 53)
, this counters the loss of ATP during ischemia and hypoxia and prevents necrosis of hepatocytes. Bcl-2, per se, could also maintain relatively high intercellular ATP levels by preventing a decline in glycolysis in response to different stresses (54)
. Thus, in our model of ischemic preconditioning, Bcl-2 overexpression seems to be a key component in the protection of cells against the shift from apoptosis to necrosis in response to ischemia-reperfusion. However, our study also suggests that ischemic preconditioning may protect liver cells from autophagy due to deficiencies of extracellular and intracellular nutrients. Indeed, autophagy has been shown to play an important role in the cellular response to stress, including inflammation (66
, 67)
. Recently, it has been shown that Bcl-2 controls nonapoptotic programmed cell death that depends on autophagy genes (68)
.
In conclusion, we show that liver injury caused by ischemia-reperfusion could be reduced in patients who underwent ischemic preconditioning before standardized hepatectomy. The ischemic preconditiong induces the overexpression of IL1-Ra, Bcl-2 and increases the production of NO. These factors counter the proinflammatory and the proapoptotic effects of release of cytokines such as IL-1ß, and probably TNF-, known to be produced during ischemia-reperfusion, including liver steatosis. Ischemic preconditioning may extend liver cell life span by inducing autophagy, a type of programmed-cell death that is separate from apoptosis.
|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication November 24, 2004. Accepted for publication May 25, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
