| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Michael E. DeBakey Department of Surgery, Liver Transplant Center Laboratory, Baylor College of Medicine, Houston, Texas; and
Department of Biochemistry and Center for the Study of Host Resistance, McGill University, Montreal, Canada
Correspondence: John A. Goss, M.D., Department of Surgery, Liver Transplant Center Laboratory, Baylor College of Medicine, 6550 Fannin, Ste. 1621, Houston, TX 77030. E-mail: jgoss{at}bcm.tmc.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(TNF-) levels were measured, and liver sections were stained for polymorphonuclear leukocyte (PMN) accumulation. After 45 min of ischemia and 30/60 min of reperfusion of Nramp1+/+ and Nramp1-/- mice livers, we found significant increases in plasma pGOT activity and TNF- levels in Nramp1+/+ mice at 30 and 60 min of reperfusion, respectively, compared with sham controls and all Nramp1-/- mice. A significant accumulation of PMNs was also found in livers of Nramp1+/+ mice at 60 min of reperfusion compared with all other groups. We have shown that disruption of the Nramp1 gene attenuates I/R injury to the mouse liver during the early phase of warm I/R injury. An increased understanding of the role played by Nramp1 is particularly important in the liver, as this organ is subjected to a wide variety of injuries during hemorrhagic shock, partial resections, and transplantation.
Key Words: macrophages • solute • carrier • P-selectin • ICAM-1
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(TNF-) and interleukin-1 (IL-1); ref.6
]. This creates a KC-dependent, extracellular oxidative stress, which serves to activate hepatocytes and recruited PMNs. KC-generated ROI directly injure hepatocytes and lay the foundation for the more severe late phase (6–24 h) of hepatic I/R injury, attributed to PMN activation, transmigation, and production of ROI, which cause injury to activated hepatocytes in the close proximity [7
]. The mammalian Nramp (also known as Slc11a, solute carrier family 11) gene family is divided into two classes, Nramp1 and Nramp2 (also known as Nramp2/DMT1/DCTI). Nramp1 is exclusively expressed in phagocytic cells [e.g., monocytes/macrophages and granulocytes (PMNs); refs.8 , 9 ]. In contrast, Nramp2 is ubiquitously expressed in most tisssues and and cell types [10 , 11 ]. Most of the current literature addressing the biochemical function of Nramp1 and Nramp2 supports the notion that the polypeptides of both are able to transport divalent metals (e.g., Fe2+, Mn2+, and Zn2+) into/out of late endosomal and lysosomal compartments. Nramp1 was mapped to chromosome 1 in mice, and a Nramp-related sequence (also known as Nramp-rs) has been mapped to chromosome 17 [12 ]. The Nramp1 complimentary DNA (cDNA) comprises of 15 exons spanning 11.5 kb genomic DNA [13 ]. The gene encodes a 90–100 kDa integral membrane protein, containing 12 hydrophobic, putative transmembrane domains, a heavily glycosylated extracellular loop, several phosphorylated sites, a Src homology 3-binding region, and a consensus transport signature [8 , 14 , 15 ]. Immunocytochemical studies with protein-specific antibodies have shown that Nramp1 protein is expressed in late endosomal and lysosomal membranes in macrophages, and upon phagocytosis, Nramp1 is localized to the phagosomal membrane [16 , 17 ].
Nramp1 has been identified as identical to the Lsh/Ity/Bcg gene, the gene responsible for the natural resistance of inbred mice strains to infection by Salmonella typhimurium, Mycobacteria, and Leishmania donovani [18
19
20
]. Lsh/Ity/Bcg has been shown to influence the early phase of bacterial replication in macrophages. Expectedly, most studies done on Nramp1 dealt with mice innate immunity to bacterial infection, which addressed resistance/lack of resistance to bacterial proliferation in spleen and liver macrophages. These studies used monocyte/macrophage cell lines (RAW264.7 and J774a.1) [17
, 21
, 22
] and isolated mouse peritoneal/splenic macrophages from whole animals. Several studies have shown that Nramp1 is constitutively expressed in macrophage cell lines of the myeloid lineage (isolated peritoneal, splenic, and liver resident macrophages) and can be induced by treatment of macrophage cells with interferon-
(IFN-) or IFN- plus lipopolysaccharide [9
, 13
].
Expression of Nramp1 in isolated KC from mice livers and in whole livers was also demonstrated [23
, 24
]. Functional studies using isolated macrophages from Bcgs (susceptible) mice, Bcgr (resistant) mice, and RAW297 cells have demonstrated that the Nramp1 protein plays a pivotal role in early macrophage activation [17
, 21
, 22
] and modulates macrophage ROI and proinflammatory cytokine production [25
26
27
28
29
]. Therefore, as Nramp1 modulates KC activation, ROI production, and proinflammatory cytokine release—the cell mainly responsible for the early phase of warm I/R injury to the liver—we hypothesized that in the absence of Nramp1 in KC, the mouse liver will be protected from the early phase of warm I/R injury. Several lines of evidence support the rationale for undertaking this study: KC make up 30% of the cells in the liver sinusoids and are a major reservoir of resident macrophages [30
]; extracellular oxidative stress mainly as a result of activated KC was shown to be responsible for the early phase of warm I/R injury to livers in rats [6
], and late-phase injury was shown to be mainly a result of PMNs [7
]; isolated KC from post-ischemic rat livers (after 60 min of reperfusion) were shown to be primed (unlike hepatocytes and endothelial cells) and generated significantly greater ROI than isolated KC from sham-control rats [31
]; Nramp1 modulates macrophage activation [17
, 21
, 22
] and ROI production [25
26
27
28
29
]; isolated, experimentally activated peritoneal macrophages from Bcgr mice and RAW297 cells transfected with the Nramp1G169 gene were shown to generate significantly more ROI than isolated, experimentally activated peritoneal macrophages from Bcgs mice and Nramp1D169-transfected RAW297 cells, respectively [32
]; Nramp1 was shown to be exclusively expressed in macrophages (including KC) and PMNs [8
, 9
]; Nramp1 mediates iron metabolism in macrophages [33
34
35
]; intracellular iron levels in isolated KC were shown to regulate KC activation and TNF- production via nuclear factor-
B (NF-B) activation [36
]; and Nramp1 mediates iron metabolism in vivo and in vitro [33
, 35
, 37
, 38
]. Although results of earlier studies have supported a role (ROI-mediated killing of bacteria/iron deprivation) for Nramp1 in innate immunity to intracellular pathogens [32
, 39
, 40
] and linkage to inflammatory diseases such as rheumatoid arthritis [41
], this study is the first to implicate Nramp1 as a mediator of KC-dependent, early-phase I/R injury of the liver.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Warm hepatic I/R model
Animals were anaesthetized with Nembutal (50–60 mg/kg body weight, intraperitoneally; Sigma-Aldrich, St. Louis, MO), and under aseptic conditions, a laparotomy was performed to access the liver for mobilization (after dividing ligamenture attachments). To prevent portal vein pooling, 200 IU/kg heparin was administered intravenously. Partial warm-hepatic ischemia was produced for 45 min by placement of vascular microclips across the hilum of the median and left lateral lobes. Sham animals were subjected to anaesthesia, laparotomy, and mobilization of the liver only. After 45 min of ischemia, vascular microclips were removed to produce reperfusion. Blood and liver tissues were taken from sham controls at 30 and 60 min of reperfusion.
Plasma glutamate oxaloacetate transaminase (pGOT) activity
Blood samples were collected by heart puncture, and pGOT activity and plasma TNF- levels were determined with a commercially available kit (#DG158K-U, Sigma Diagnostics, St. Louis, MO).
Hepatic PMN accumulation
As an additional indicator of liver injury, accumulation of PMNs in mice livers following I/R was determined in formalin-fixed paraffin sections of the liver obtained at each reperfusion-sampling time point. A commercially available kit (91-C, Sigma-Aldrich) was used to stain for sinusoidal-sequestrated PMNs using the well-established Napthol AS-D chloroacetate esterase procedure, according to the manufacturer’s directions. At least four random sections from each group were analyzed by viewing (blindly) 50 random high-power fields (HPF, 400x) on each section. Results are expressed as number of PMNs/50 HPF.
Histological analysis of liver injury
hematoxylin and eosin (H&E) sections from formalin-fixed liver tissues obtained from sham-control mice and mice subjected to warm I/R were randomly selected and analyzed blindly (by pathologist Milton J. Finegold, Texas Children’s Hospital, Houston) for the degree of necrosis, vacuolization, glycogen depletion, zonal variations, and sinusoidal congestion as a measure of hepatic injury.
Western blot analysis of liver Nramp1 protein
Liver Nramp1-enriched membrane extracts were prepared essentially as described by Govoni et al. [43
]. A 20% homogenate (
100 mg tissue) in TM [10 mM Tris-HCl, 1 mM MgCl2, and cØmpleteTM protease inhibitor (Roche Diagnostics Corp., Indianapolis, IN)] buffer, pH 7.0, was made from snap-frozen liver samples harvested from sham-control mice and at 30 and 60 min of reperfusion. Homogenates were centrifuged at 400 g for 10 min at 4°C to eliminate unbroken cells and nuclei. The resulting supernatants were centrifuged at 100,000 g for 30 min at 4°C to pellet crude membrane fractions. Pellets were subsequently resuspended in 10 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, and cØmpleteTM protease-inhibitor buffer, pH 7.0, and were used for Western blot analysis. Proteins (27.4 µg) were separated on sodium dodecyl sulfate (SDS)-polyacrylamide gels (10%) and were transferred to nitrocellulose membranes (Schleicher and Schuel, Dassel). Equal loading and transfer to the membranes were confirmed by staining the blots with Ponceau S (Sigma-Aldrich). Nramp1 protein bands were detected on membranes by incubating with a Nramp1 polyclonal antibody (pAb) [35
] (a generous gift from Dr. Bruce S. Zwilling, Department of Microbiology, The Ohio State University, Columbus), raised in rabbits against a glutathione-S-transferase-Nramp1 fusion protein (containing Nramp1 amino acids 305–346 of the fourth extracytoplasmic "loop") and diluted 1:100. Nramp1 protein bands were visualized with horseradish peroxidase (HRP)-conjugated anti-rabbit secondary antibody (diluted 1:1000) from a commercially available enhanced chemiluminescence (ECL) kit (Amersham Life Sciences, Piscataway, NJ). Membranes were immunoblotted for ß-actin with an affinity-purified goat pAb diluted 1:500 (sc-1616, Santa Cruz Biotechnology, Inc., Santa Cruz, CA) and a HRP-conjugated anti-goat secondary antibody (sc-2350, Santa Cruz Biotechnology; diluted 1:1000). Actin bands were visualized with a commercially available ECL kit (Amersham Life Sciences). Nramp1 and actin bands were digitized using Scion Image Beta 3b software (Scion Corp., Frederick, MD), and densitometric values for Nramp1 bands were normalized using densitometric values obtained for actin bands.
Northern blot analysis of liver Nramp1 mRNA
Total RNA was extracted from liver tissues using an UltraSpec total RNA isolation kit (#BL-10050, Biotecx Laboratories Inc., Houston , TX). cDNA was transcribed with 4 µg total RNA, random hexamers, and a SuperScriptTM II preamplification system (#18089-011, Gibco-BRL, Life Technologies, Grand Island, NY) per the manufacturer’s protocol. To ensure that equal amounts of RNA were used for all samples, RNA was quantified spectrophotometrically, and its integrity was evaluated on 1% TreviGelTM 500 powder (Trevigen, Gaithersburg, MD) gel. A Nramp1 probe was made using specific primers for Nramp1 [EMBL/GenBank access no. L13732, sense 5'-TCTCCGCTGGACGCAT-3' (positions 569–590), antisense 5'-TGACCAAGGCTGAGTGCAGT-3' (positions 854–874), 303-bp product (Sigma-Genosys, Woodlands, TX)] and for polymerase chain reaction (PCR). Complementary DNA of Nramp1 was amplified by PCR under the following conditions: Nramp1, 30 cycles, 94°C for 60 s, 61°C for 60 s, and 72°C for 60 s. All PCR products were electrophoresed on gels made of 1% TrevigelTM 500 powder. The PCR product was visualized by post-staining for 30 min with GelStar nucleic acid gel stain (FMC Bioproducts, Rockland, MA). The 303-bp band was cut from gels and further purified and concentrated using SUPRECTM 01 and 02 columns (Pan Vera Corp., Madison WI), respectively. Purified PCR product was radiolabeled with [32P]dCTP (ICN, Irvine, CA) using a Nick translation kit (Amersham Pharmacia Biotech, Piscataway, NJ). To normalize data from our Nramp1 Nothern blot, a mouse ribosomal 18S probe (18S-Mouse DECAtemplateTM) was obtained from Ambion Inc. (Austin, TX) and radiolabeled using [32P]dCTP and a Nick translation kit (Amersham Pharmacia Biotech). Total RNA was extracted from liver tissues and stored at -80°C. RNA was quantified by UV-spectrophotometry, and equal amounts (10 µg) were resolved under denaturing conditions on 1% TreviGelTM 500 powder gel made in 1x TAE buffer. RNA bands were transferred to BrightStar-Plus positively charged nylon membrane (Ambion Inc.) by semi-dry blotting. Prehybridization and hybridization of the Nramp1 probe to membranes were done in ExpressHyb hybridization solution (Clontech, Palo Alto, CA) at 68°C. Membranes were washed twice in 50 ml 2x saline sodium phosphate EDTA buffer (SSPE) plus 0.1% SDS at room temperature for 30 min, followed by two additional 50 ml washes in 0.1x SSPE plus 0.1% SDS at 50°C. Prehybridization and hybridization of the 18S probe to membranes were done in ExpressHyb hybridization solution (Clontech) at 50°C. Membranes were washed twice in 50 ml 2x SSPE plus 0.1% SDS at room temperature for 30 min, followed by two additional 50 ml washes in 0.1x SSPE plus 0.1% SDS at 68°C. The washed blots were then exposed to Kodax X-OMAT film at -80°C. Nramp1 and 18S blots were digitized using Scion Image Beta 3b software (Scion Corp.), and density values for Nramp1 bands were normalized using density values obtained for 18S bands.
Immunohistochemical (IHC) analysis of liver Nramp1 protein
The Nramp1 "loop" antibody mentioned above (diluted 1:10) and a goat anti-rabbit biotinylated immunoglobulin G (IgG) secondary antibody diluted 1:400 (Vector Laboratories, Inc., Burlingame, CA) were used to stain paraffin sections for Nramp1. All primary and secondary antibody solutions used to detect Nramp1 contained 0.1% saponin. The commercially available Vectastain ABC kit (P-6100, Vector Laboratories) was used for immunodetection and visualization with Fast Red Substrate (alkaline phosphatase). Sections were counterstained with Mayer’s hematoxylin. The specificities of labeling were assessed by use of isotype-matched antibodies.
Liver glutathione (GSH) levels
Total GSH levels in livers were determined by the method of Tietze [44
], as previously described by Jaeschke et al. [7
]. Liver samples were snap-frozen in liquid nitrogen and stored at -80°C until used.
Extraction and Western blot analysis of liver heme oxygenase-1 (HO-1) and 2 (HO-2)
Liver extracts containing the HO-1 and HO-2 proteins were obtained by standard differential centrifugation. Briefly, a 20% homogenate was made from liver tissues (0.4 g) in a 50 mM Tris-HCl buffer, pH 8.0, containing 1 mM phenylmethylsulfonyl fluoride. The homogenates were centrifuged at 10,000 g for 15 min, and the supernatants were used for Western blot. Proteins (27.4 µg) were separated on SDS-polyacrylamide gels (10%) and transferred to nitrocellulose membranes (Schleicher and Schuel). Equal loading and transfer to the membranes were confirmed by staining the blots with Ponceau S (Sigma-Aldrich). HO-1 and HO-2 protein bands were detected on membranes by incubating with commercially available pAb for HO-1 (rabbit anti-rat pAb diluted 1:300, Affinity Bioreagents, Inc., Golden, CO) and HO-2 (rabbit anti-human pAb diluted 1:200, Santa Cruz Biotechnology). Protein bands were visualized with HRP-conjugated anti-rabbit secondary antibody (diluted 1:1000) from a commercially available ECL kit (Amersham Life Sciences).
Plasma TNF- levels
Plasma TNF- levels were determined in blood samples taken from sham-control mice and at 30 and 60 min of reperfusion using a commercially available kit (558870, BD Pharmingen, San Diego, CA). HO-1, HO-2, and actin bands were digitized using Scion Image Beta 3b software (Scion Corp.), and densitometric values for HO-1 and HO-2 bands were normalized using densitometric values obtained for actin bands.
Liver NF-B activation
Activation of the redox-sensitive NF-B in livers of Nramp1+/+ and Nramp1-/- mice was measured using a commercially available kit (Trans-AMTM NF-B p50 transcription factor assay kit, Active Motif, Carlsbad, CA). The assay was done according to the manufacturer’s procedure and as described by Renard et al. [45
]. Nuclear proteins were extracted according to the procedure of Osarogiagbon et al. [46
]. Roughly 100 mg snap-frozen liver tissue was homogenized in 0.4 ml cold TM buffer (using cØmpleteTM protease inhibitor, Roche Diagnostics). Homogenates were centrifuged at 2000 rpm for 30 s, and the supernatant was mixed with 200 µl lysis buffer, incubated at 4°C for 5 min, and centrifuged at 5000 rpm for 10 min. The nuclear pellets were reconstituted with lysis buffer and centrifuged at 14,000 rpm for 20 s at 4°C. Nuclear protein extract (15 µg) of each sample was used to assay for NF-B activation. To ensure the specificity of the assay, the wild-type consensus oligonucleotide provided by the manufacturer served as a competitor to NF-B binding.
IHC analysis of F4/80, P-selectin, and intercellular adhesion molecule-1 (ICAM-1)
Liver tissues harvested from sham-control mice and at 30 and 60 min of reperfusion were fixed in formalin or embedded in 2 methylbutane-cooled OCT compound (Sakura Finetek, Torrance, CA) and snap-frozen in liquid nitrogen. Paraffin sections of 5 µm were used for immunodetection of the macrophage-specific cell surface marker F4/80 and P-selectin. Cryostat sections (4 µm) were fixed in cold acetone and stained for ICAM-1. Commercially available primary [monoclonal antibodies (mAb)/pAb] and secondary antibodies were used to detect F4/80 (rat anti-mouse mAb IgG2b diluted 1:10, Research Diagnostics Inc., Flanders, NJ) and mouse anti-rat biotinylated secondary antibody IgG2b diluted 1:400 (Research Diagnostics), P-selectin (rabbit anti-human pAb CD62P diluted 1:50, BD Pharmingen), goat anti-rabbit biotinylated secondary antibody IgG diluted 1:400 (Vector Laboratories), ICAM-1 (rat anti-mouse mAb YN1/1.7.4 diluted 1:100, American Type Culture Collection, Mansassas, VA), and mouse anti-rat biotinylated secondary antibody IgG2b diluted 1:400 (Research Diagnostics). The commercially available Vectastain ABC kit (P-6100, Vector Laboratories) was used for immunodetection and visualization with 3'-diaminobenzidine tetrahydrochloride (peroxidase) or Fast Red Substrate (alkaline phosphatase). Sections were counterstained with Mayer’s hematoxylin. The specificities of labeling were assessed by use of isotype-matched antibodies.
Protein concentration
All protein concentrations were done according to the method of Lowry et al. [47
].
Statistics
Statistically significant differences between groups of mice were determined by ANOVA followed by SNK or TUKEY’S nonparametric tests. P values of 
0.05 were taken as significant differences between groups.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
0.005; Fig. 1
)]. Similarly, pGOT activity in Nramp1+/+ mice at 60 min of reperfusion was significantly greater than pGOT in sham controls of Nramp1-/- and Nramp1+/+ mice (**, P0.046; Fig. 1
). A significant accumulation of PMNs (394±10 PMNs/50 HPF) was found in livers of Nramp1+/+ mice at 60 min of reperfusion compared with all other groups (*, P0.002; Fig. 2
). Also, PMN accumulation in livers of Nramp1+/+ mice at 30 min of reperfusion was significantly greater than in livers of sham-control Nramp1-/- and Nramp1+/+ mice (**, P<0.047; Fig. 2
). However, PMN accumulation in livers of Nramp1-/- mice at 30 and 60 min of reperfusion was significantly greater than PMN accumulation in livers of sham-control Nramp1-/- and Nramp1+/+ mice (***, P<0.047; Fig. 2
).
|
|
|
|
Liver GSH levels and Western blot analysis of HO-1 and HO-2
To test whether Nramp1 disruption would blunt the oxidant stress seen in the liver during the early phase of I/R injury, we measured liver GSH and HO-1 protein levels. The change in tissue GSH is considered to be a good marker of oxidant stress [6
], and liver HO-1 has been shown to be induced by oxidant stress as a result of I/R [49
]. In this study, liver GSH levels were significantly reduced at 30 and 60 min of reperfusion in Nramp1+/+ mice compared with Nramp1+/+ sham-control mice (*, P<0.001 and **, P=0.017; Fig. 5A
). In contrast, no significant change in GSH levels was detected in livers of Nramp1-/- mice compared with Nramp1-/- sham-control mice. GSH levels in Nramp1-/- mice livers were found to be significantly lower compared with GSH levels in Nramp1+/+ mice livers (#, P<0.001; Fig. 5A
). As expected, HO-1 protein levels increased in livers of Nramp1+/+ mice at 30 and 60 min of reperfusion compared with Nramp1+/+ sham-control groups *, P 
0.014 and *P 0.041; (Fig. 5B
and 5C) . In contrast, no significant change in HO-2 protein levels was found in livers of Nramp1+/+ mice compared with HO-2 protein levels in livers of Nramp1+/+ sham controls (Fig. 5B
and 5C)
. Similarly, no appreciable increase in HO-1 or HO-2 protein levels was found in livers of Nramp1-/- mice compared with livers of Nramp1-/- sham controls (Fig. 5B
and 5D)
.
|
and liver NF-B activation levels and liver NF-B activation. Plasma TNF- levels in Nramp1+/+ mice were significantly different (*, P<0.05) at 60 min of reperfusion compared with plasma TNF- levels in the Nramp1-/- group at 60 min of reperfusion (Fig. 6A
). These results are consistent with reports of earlier studies showing a significant increase in TNF- following I/R [50
51
52
]. Similarly, NF-B activation in the livers of Nramp1+/+ mice was significantly different (#, P<0.05) at 60 min of reperfusion compared with NF-B activation in livers of Nramp1-/- groups (Fig. 6B)
.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, IL-1ß, CXC chemokines, oxidative burst, nitric oxide synthase, L-arginine flux, major histocompatibility complex, and tumoricidal and antimicrobial activities [53
54
55
]. Many of the above factors have also been implicated in I/R injury of the liver [50
, 56
57
58
]. Therefore, the purpose of this study was to investigate whether the Nramp1 gene plays a role in I/R injury of the liver. In this study, we found significant increases of pGOT activity, hepatic PMN accumulation, and histological degenerative evidence in livers of Nramp1+/+ mice compared with livers of Nramp1-/- mice after 45 min of ischemia and 30 or 60 min of reperfusion. Our data demonstrate that disruption of the Nramp1 gene in KC attenuates I/R injury to the mouse liver during the early phase of I/R injury. The major findings of this study are that disruption of Nramp1 blunts hepatic oxidative stress and delayed the proinflammatory response elicited by I/R of the liver during the early phase of I/R injury. The blunted oxidative stress is reflected in the lack of significant change to total liver GSH in Nramp1-/- mice compared with the significant decreases in total GSH found in livers of Nramp1+/+ mice at 30 and 60 min of reperfusion. Earlier studies have demonstrated that KC are mainly responsible for the oxidative stress in the liver during the early phase of warm I/R of the liver [6 ], and activated PMNs are mostly responsible for the oxidative stress seen in the latter phase of I/R injury [7 ]. In addition, recent evidence in livers reperfused after cold storage has shown that hepatocytes may also be a source oxidative changes capable of impairing liver function during reperfusion [59 ]. In this study, although lower levels of total GSH were found in livers of all Nramp1-/- mice compared with livers of Nramp1+/+ mice, the lack of significant change in total liver GSH of Nramp1-/- mice may be taken as the absence/attenuation of oxidative stress in the liver [60 ]. The effect of Nramp1 disruption on the antioxidant status of Nramp1-/- mice has not been determined and warrants further studies.
Additional support for a blunted/attenuated oxidative stress in livers of Nramp1-/- mice is the significant increase in HO-1 protein observed in livers of Nramp1+/+ mice at 60 min of reperfusion compared with the lack of change in HO-1 protein levels found in livers of Nramp1-/- mice. It has been reported that oxidative stress as a result of I/R of liver induces HO-1 via an activated protein-1 (AP-1) pathway [61
] and that HO-1 induction may serve a protective function. If the induction of HO-1 is taken as an indicator of oxidative stress to the liver, our data clearly show, although indirectly, attenuation of hepatic oxidative stress in Nramp1-/- mice. Further studies of I/R-induced oxidative stress in Nramp1+/+ and Nramp1-/- mice are currently in progress. The delayed/attenuated proinflammatory response to I/R observed in Nramp1-/- mice is also reflected in the significant increases in plasma TNF- and liver NF-B activation of Nramp1+/+ mice at 60 min of reperfusion compared with Nramp1-/- mice. NF-B activation in the liver following I/R has been reported [62
63
64
]. In addition, its activation has been reported as a requirement for an I/R-dependent, TNF- increase in the liver [62
]. The blunted, inflammatory response observed may be the result of an attenuated vascular oxidative stress attributed to KC during the early phase of I/R injury [60
], as Nramp1 plays a major role in early activation of KC [17
, 21
, 22
].
Earlier and more recent studies using the macrophage surface antigen F4/80 used it as a marker to identify/estimate mature resident tissue macrophages [65 66 67 68 69 ]. However, some recent studies have used the increased expression of F4/80 as an indication of macrophage activation [70 71 72 73 ]. In this study, we used F4/80 staining in the liver to identify KC [67 ]. If we consider F4/80 staining to be an indication of KC activation, our data from livers of Nramp1-/- mice at 30 min of reperfusion support earlier studies, which reported decreased expression of F4/80 with activation [74 75 76 77 ]. At the same time, we would not be able to explain the significantly lower expression of F4/80 in sham-control Nramp1-/- mice livers and livers reperfused for 30 and 60 min. Given the uncertainty surrounding the regulation of F4/80 at present, we are inclined to believe that F4/80 is a marker of certain subgroups of KC, which normally constitute 50 percent of KC [66 ]. The discrepancy of its expression is probably a result of its low expression on macrophages and the specificity of antibodies used for detection [66 ].
Existing reports suggest that P-selectin mediates I/R injury of the liver by assisting in the recruitment of PMNs [78 79 80 ]. However, other reports minimize its role in liver I/R injury and its role in recruiting PMNs in the inflamed liver vasculature [81 82 83 84 85 86 ]. Our study further supports a questionable, mandatory role for P-selectin-dependent sequestration of neutrophils in the liver sinusoids after I/R [86 ]. We did not detect any staining for P-selectin on sinusoidal endothelial cells or hepatocytes of any group and found no significant difference between groups in staining intensity of P-selectin in post-sinusoidal venules. Futhermore, hepatic PMN accumulation mediated by P-selectin expressed on endothelial cells of post-sinusoidal venules would not have contributed significantly to liver injury, as there is no experimental evidence supporting extravasation of these neutrophils to the liver parenchyma [84 , 86 ].
In addition, our data showing significant increases of hepatic PMN accumulation in Nramp1+/+ mice at 30 and 60 min of reperfusion are contrary to the limited staining found for ICAM-1 in liver sinusoids at 30 and 60 min of reperfusion. Numerous studies have reported that ICAM-1 mediates margination and transmigration of PMNs in the liver following an inflammatory stimuli [87 88 89 90 91 92 ]. However, other studies [81 , 93 94 95 96 ] have questioned the absolute role of ICAM-1 in PMN sequestration in hepatic vasculature. We cannot explain the limited staining found for ICAM-1 in livers of Nramp1+/+ mice at 30 and 60 min of reperfusion with concomitant and significant increases in PMN accumulation. However, we are not the first to report a paradoxical decrease in ICAM-1 levels following I/R of the liver [94 ]. Nonetheless, although the general mechanism of selectin-dependent rolling of PMNs followed by firmer adhesion to endothelial cells by integrins and ICAM-1 is applicable to the vasculature of some organs (heart, lung, intestine, and cremaster muscle), this is not the case for the liver [83 , 84 ]. In this study, the significant accumulation of PMNs in livers of Nramp1+/+ mice at 60 min of reperfusion is probably a result of physical trapping, resulting from the swelling of sinusoidal endothelial and KC. Enhanced hepatic PMN accumulation would have also occurred from greater hepatocyte injury and the existence of a chemokine gradients in Nramp1+/+ mice compared with Nramp1-/-. Alternatively, other yet-unidentified adhesion molecules may play a major role in the sequestration of PMNs in liver sinusoids. Finally, we found similar staining intensity for ICAM-1 in all Nramp1-/- groups, which suggests an absence of/or a significantly attenuated proinflammatory response.
Exactly how the presence/absence of Nramp1-mediated iron metabolism in KC controls the activation of KC and subsequent proinflammatory response as a result of liver I/R is still unclear and beyond the scope of this study. However, existing evidence suggests that KC intracellular iron concentration modulates redox-sensitive transcription factors such as NF-B and AP-1 activation, which in turn mediates the gene expression of proinflammatory cytokines such as TNF- and IL-1 [36
]. We propose that Nramp1 mediates the overall iron metabolism in KC [33
, 35
, 37
, 38
], which in turn modulates the activation of redox-sensitive transcription factors [36
], mRNA stability [97
], and gene expression of proinflammatory cytokines [29
] (Fig. 10
). Further in vitro studies to investigate KC activation and iron-handling capacity in the presence and absence of Nramp1 using isolated KC from post-ischemic and nonischemic livers are warranted. Nonetheless, it is reasonable to suggest that Nramp1 plays a major role in I/R injury of the liver. Given the acidosis of the liver tissue (as a result of ischemia-induced anaerobic glycolysis, proton released during adenosine 5'-triphosphate (ATP) hydrolysis, and inactivated V-ATPases; ref. [99
]), the increase in cytosolic iron known to occur with ischemia [100
, 101
], and Nramp1’s pH-dependent function of transporting divalent metals [106
], we propose that upon reperfusion in the presence of a fully functional Nramp1 protein, livers of Nramp1+/+ mice are more prone to I/R injury than livers of Nramp1-/- mice. The presence of a fully functional Nramp1 may contribute to the neutralization of KC cytosolic pH [102
] and the re-establishment and maintenance of iron homeostasis to steady-state levels [103
, 104
, 107
]. Also, in macrophages, Nramp1 is known to be involved in the acidification of subcellular vesicles [101
] (affecting V-ATPase recruitment) and direct contribution of iron to ferritin [104
]. Proteases released by KC become active (following inactivation of plasma antiproteases by ROI from KC and pH neutralization) and cause direct injury to hepatocytes [108
]. Nramp1 may also mediate the re-establishment of KC intracellular iron homeostasis by transporting iron out of KC cytoplasm (efflux, as proposed by Atkinson and Barton, ref. [103
]) or by transporting excess iron into subcellular vesicles (as proposed by Zwilling and co-workers, ref. [107
]). Given the perturbation in pH, both of these mechanisms may be taking place in KC under our experimental condition, as Nramp1 has been reported to shuttle divalent metals in either direction in a pH-dependent manner [106
]. Either way, in Nramp1+/+ mice livers, Nramp1-mediated re-establishment of iron homeostasis in KC will create intracellular iron concentrations permissive to the activation of the redox-sensitive transcription factor NF-B, mRNA stability, and the gene expression of the proinflammatory cytokines TNF- and IL-1, which are key mediators of liver I/R injury. Excess cytosolic iron levels in KC have been reported to inhibit KC cytokine release [105
]. Although other compensating mechanisms may exist in KC, which are able to mediate the neutralization of cytosolic pH and iron homeostasis, in the absence of a functional Nramp1 protein, KC in livers of Nramp1-/- mice may not be able to respond as quickly to the acidic environment and excess cytosolic iron. This would explain the delayed/blunted proinflammatory response observed in Nramp1-/- mice following liver I/R in this study. To further investigate our hypothesis, hypoxia/reoxygenation studies using isolated KC from Nramp1+/+ and Nramp1-/- mice livers, iron chelators, and pH modulators (e.g., bafilomycin/concanamycin) are in progress.
|
|
ACKNOWLEDGEMENTS |
|---|
Received February 20, 2002; revised June 13, 2002; accepted June 14, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and infection with BCG Immunology 82,457-464
B activation and cytokine gene expression by rat hepatic macrophages Am. J. Physiol. 272,G1355-G1364-/- mice J. Interferon Cytokine Res. 21,53-62[Medline]
B Nucleic Acids Res. 29,E21
production after total hepatic ischemia Transplantation 64,821-827[Medline]
