Proteomic analysis of hepatic iron overload in mice suggests dysregulation of urea cycle, impairment of fatty acid oxidation, and changes in the methylation cycle

Jiri Petrak, Denisa Myslivcova, Petr Man, Radek Cmejla, Jana Cmejlova, Daniel Vyoral, Milan Elleder, Christopher D. Vulpe

Abstract

Liver iron overload can be found in hereditary hemochromatosis, chronic liver diseases such as alcoholic liver disease, and chronic viral hepatitis or secondary to repeated blood transfusions. The excess iron promotes liver damage, including fibrosis, cirrhosis, and hepatocellular carcinoma. Despite significant research effort, we remain largely ignorant of the cellular consequences of liver iron overload and the cellular processes that result in the observed pathological changes. In addition, the variability in outcome and the compensatory response that likely modulates the effect of increased iron levels are not understood. To provide insight into these critical questions, we undertook a study to determine the consequences of iron overload on protein levels in liver using a proteomic approach. Using two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) combined with matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS), we studied hepatic iron overload induced by carbonyl iron-rich diet in mice and identified 30 liver proteins whose quantity changes in condition of excess liver iron. Among the identified proteins were enzymes involved in several important metabolic pathways, namely the urea cycle, fatty acid oxidation, and the methylation cycle. This pattern of changes likely reflects compensatory and pathological changes associated with liver iron overload and provides a window into these processes.

  • hereditary hemochromatosis
  • liver disease
  • 2D-PAGE

iron overload disorders are a significant health concern. Liver iron overload can be found in chronic liver diseases such as alcoholic liver disease and chronic viral hepatitis or secondary to repeated blood transfusions. Inherited disorders can also result in iron overload. Hemochromatosis (HC), which results predominantly from mutations in the HFE gene (11), is the most common autosomally inherited disease of humans, and in Caucasians ∼1 in 200 individuals are homozygous for the major HFE mutation that can lead to iron loading (25). Phenotypically, HC is characterized by an inappropriately high rate of intestinal iron absorption (14). Iron is subsequently deposited in many tissues, but most notably in the liver, where it can lead to significant organ damage. Iron loading similar to HFE-associated disease can also result from mutations in ferroportin, transferrin receptor 2, hepcidin, and hemojuvelin (27). Other prominent disorders associated with elevated iron absorption and liver iron deposition include the iron loading anemias such as β-thalassemia (28) and African iron overload (15).

Although the overwhelming majority of individuals with inherited HC share a single identical mutation (C282Y) in the HFE protein, there is considerable individual variation in age of onset, severity of clinical features such as diabetes, cardiac disease, cirrhosis, the rate of iron accumulation in the liver, and the amount of iron loss via phlebotomy. At the extremes, individuals homozygous for the C282Y mutation have presented at an early age with fulminant liver disease, whereas others with the identical mutation have survived to their eighties with no overt signs of disease. The penetrance of the HFE genotype has been estimated to be as high as 50%, on the basis of data from a population-based study conducted in Busselton, Australia (25). A recent study from Beutler and colleagues (3) demonstrated that over 50% of C282Y homozygotes had increased total body iron. However, clinical penetrance of disease was much lower than expected, at ∼1%. Furthermore, data from studies of hfe and β2-microglobulin knockout mice have shown that the degree of hepatic iron loading varies considerably depending on the genetic background on which the knockout is placed (13, 40). The molecular mechanisms that underlie the observed phenotypes in iron overload and the processes mediating the variability in effects and clinical consequences remain unclear.

Animal models have provided some insight into the molecular mechanism of the pathology in iron overload. Iron overload in rodents induced by carbonyl iron-supplemented diet is known to result in a predominantly parenchymal iron deposition pattern similar to the iron deposition observed in hereditary HC. Current models suggest that cellular injury may be related to iron-generated reactive oxygen species causing peroxidation of lipids in membrane organelles, such as mitochondria (2) and lysosomes (23), thus limiting their functions. Oxidative damage to the membranes, proteins, and DNA is thought to contribute to hepatocyte necrosis and apoptosis, which in turn ultimately lead to fibrosis, cirrhosis, and liver cancer (31). Indeed, patients with iron overload are at significantly higher risk of developing hepatocellular carcinoma (10, 24). Despite the research effort, we remain largely ignorant of the cellular consequences of liver iron overload and the cellular processes that result in the observed pathological changes. In addition, the variability in outcome and the compensatory responses that likely modulate the effect of increased iron are not understood.

To provide insight into these critical questions, we undertook a study to determine the consequences of iron overload on protein levels in liver cells using a proteomic approach. Using two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) combined with matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS), we studied hepatic iron overload induced by carbonyl iron-rich diet in mice and identified 30 liver proteins whose quantity changes in condition of excess liver iron. Among the identified proteins were enzymes involved in several important metabolic pathways, namely urea cycle, fatty acid oxidation, and methylation cycle. This pattern of changes likely reflects compensatory and pathological changes associated with liver iron overload and provides a window into these processes.

MATERIALS AND METHODS

All chemicals and reagents used were from Sigma-Aldrich, unless stated otherwise.

Experimental animals and diets.

C57BL/6J mice were maintained in light- and temperature-controlled environment with free access to water. Sixteen 3-mo-old male mice were divided into two groups of eight animals and fed ad libitum with either defined standard rodent diet (C1000, Altromin, 180 ppm Fe) or the same diet supplemented with 2% carbonyl iron. Animals were euthanized after 8 wk and livers were removed. Four samples from each group were used for histology and histochemistry, the second half for the proteomic analysis and iron content measurements. For the proteomic analysis liver were snap-frozen in liquid nitrogen and homogenized by pestle in mortar. Liquid nitrogen was continuously added to prevent thawing of the sample during homogenization. Resulting individual liver powder samples were aliquoted and frozen at −80°C. The experimental procedures were approved by the Ethical Committee of the Institute of Hematology and Blood Transfusion in accordance with the current National Institutes of Health guidelines.

Histology, histochemistry, and electron microscopy.

Pieces of liver were fixed overnight in cold 10% paraformaldehyde in phosphate-buffered saline (PBS) and processed for both histology and electron microscopy. The latter samples were fixed in 1% OsO4, dehydrated, and embedded into Epon-Araldite mixture. Paraffin sections were stained with routine stains, including hematoxylin-eosin, van Gieson (detection of collagen), periodic acid-Schiff staining (glycogen staining), and Perl's reaction (detection of Fe3+) and observed by Nikon E800 microscope. In parallel, the liver samples were frozen in liquid nitrogen and cut in a cryostat. The following enzyme activities were evaluated on the cryostat sections: acid phosphatase, alkaline phosphatase, succinate dehydrogenases, NADH dehydrogenases, α-glycerophosphate dehydrogenases, and cytochrome oxidase (15). Lipids were stained using Oil red O, and lipopigments were detected by autofluorescence (filter block BV-2A: excitation 400–440 nm, barrier 470 nm).

Acid phosphatase activity was detected by azo-coupling reactions with naphthol AS-BI phosphate and hexazonium pararosaniline prepared from pararosaniline. Alkaline phosphatase activity was detected with naphthol AS phosphate and fast blue B (Lachema). Activities of dehydrogenases were tested using nitrotetrazolium blue (19). Cytochrome oxidase activity was tested with cytochrome c and diaminobenzidine (38).

Determination of liver iron content by atomic absorption spectrometry.

The liver iron content was determined by atomic absorption spectrometry with acetylene-air flame atomization. The analysis was performed with the Varian atomic absorption spectrometer (Mulgrave) with a deuterium background correction. Measurements were performed with the analytical line 248.3 nm in the spectral interval of 0.2 nm. Iron concentration was determined by the standard addition method. Sample digestion was accomplished in the MDS 2000 microwave sample preparation system (CEM) in Teflon cartridges by a mixture of nitric acid (5 ml) and hydrogen peroxide (2 ml) (both from Merck, ultrapure grade) for 20 min at pressure 120 psi. The resulting product was analyzed directly in the Teflon cartridges.

Sample preparation for 2D electrophoresis.

Liver samples were thawed in the lysis buffer (7 M urea, 2 M thiourea, 4% CHAPS, 60 mM DTT), and 1% ampholytes (IPG buffer pH 4–7, Amersham) containing protease inhibitor cocktail (EDTA Free, Roche Diagnostics) for 20 min at room temperature. Samples were centrifuged at 14,000 g for 20 min at room temperature. Supernatants were collected and protein concentration was determined by the Bradford method (Bio-Rad). Protein concentration in all samples was equalized to 5.8 mg/ml by dilution with the lysis buffer.

Two-dimensional electrophoresis.

Isoelectric focusing was performed with Bio-Rad Protean IEF cell, using 18-cm IPG strips (pH 4–7, Amersham). Samples were applied by rehydration loading. Each strip (4 gels for the control animals, 4 gels for the iron-overloaded animals; each strip originated from one animal) was rehydrated overnight in 345 μl of sample, representing 2.0 mg protein. Isoelectric focusing was performed for 60 kVh, with maximum voltage not exceeding 5 kV, current limited to 50 μA per strip and temperature set to 18°C. Focused strips were stored at −80°C. For the SDS electrophoresis strips were thawed, equilibrated, and reduced in equilibration buffer A (6 M urea, 50 mM Tris pH 8.8, 30% glycerol, 2% SDS, and 450 mg DTT per 50 ml of the buffer) for 15 min and then alkylated in equilibration buffer B (6 M urea, 50 mM Tris pH 8.8, 30% glycerol, 2% SDS, and 1.125 mg iodacetamide per 50 ml). Equilibrated strips were placed on the top of 10% PAGE and secured in place by molten agarose. Electrophoresis was performed in a Tris-glycine-SDS system using a Bio-Rad Protean II apparatus with external cooling (20°C). Gels were run at constant voltage, starting at 50 V for 30 min then 200 V for 1 h and 300 V for 3.5 h. Following electrophoresis, gels were washed two times for 15 min in deionized water to remove redundant SDS. Washed gels were stained in colloidal Coomassie blue (Simply Blue SafeStain, Invitrogen) overnight, then destained in deionized water.

Gel image analysis.

Coomassie blue-stained gels were scanned with a GS 800 calibrated densitometer (Bio-Rad). Image analysis was performed with Phoretix 2D software (Nonlinear Dynamics) in semimanual mode. Normalization of the gel images was based on total spot density. Integrated spot density values (spot volumes) were calculated (after background subtraction). Average values of spot volume (averages from the all 4 gels in the group) for each spot were compared between the two studied groups. Protein spots were considered changed (differentially expressed) only if they met both of the following criteria: average normalized spot volume difference between the two groups >1.5-fold, statistical significance of the change determined by the t-test, P < 0.05.

MALDI-MS and protein identification.

Differentially expressed proteins were excised from gels, cut into small pieces, and washed several times with 0.1 M 4-ethylmorpholine acetate (pH 8.1) in 50% acetonitrile (MeCN). After complete destaining the gel was washed with deionized water, shrunk by dehydration in MeCN, and reswollen again in water. The supernatant was removed and the gel was partly dried in a SpeedVac concentrator. The gel pieces were then reconstituted in a cleavage buffer containing 0.1 M 4-ethylmorpholine acetate, 10% MeCN, and sequencing-grade trypsin (50 ng/μl; Promega). After overnight digestion, the resulting peptides were extracted with 40% MeCN-0.5% trifluoroacetic acid. A saturated solution of α-cyano-4-hydroxycinnamic acid in aqueous 50% MeCN-0.2% trifluoroacetic acid was used as a MALDI matrix. A 0.5-μl of sample and 0.5 μl of matrix solution were placed on the sample target and allowed to dry at the room temperature. Positive ion MALDI mass spectra were measured on a Bruker BIFLEX II reflectron time-of-flight mass spectrometer (Bruker-Franzen) equipped with a SCOUT 26 sample inlet, a gridless delayed extraction ion source, and a nitrogen laser (337 nm, Laser Science). Ion acceleration voltage was 19 kV and the reflectron voltage was set to 20 kV. The spectrometer was calibrated externally using the monoisotopic [M+H]+ ions of peptide standards angiotensin II and insulin (Sigma). Software used and the data filtering method were BrukerDaltonics-FlexAnalysis 2.0, smoothing Savitzky Golay, 0.2 m/z, 1 cycle. Peak detection algorithm SNAP, S/N threshold 8. Proteins were identified by searching the peptide mass maps in the NCBInr database (release date 2005/06/01) using the search program MASCOT (http://www.matrixscience.com) with search criteria: Taxonomy-Mammalia, enzyme-trypsin, missed cleavage-1, complete modification-Cys (iodoacetamide), partial modification-methionine oxidation, mass tolerance-0.35 Da.

Western blotting.

Liver samples were thawed in the lysis buffer (1.5% Triton X-100, 140 mM NaCl, 10 mM HEPES pH 7.4) for 10 min on ice. Supernatant was collected after centrifugation at 15,000 g for 15 min at 4°C. Sixty micrograms of liver were loaded on a precast 4–20% acrylamide gel (Invitrogen) and electrophoresed in Tris-glycine-SDS buffer system in X-Cell Sure Lock apparatus (Invitrogen). Proteins were transferred to polyvinylidene difluoride membrane (Hybond-P, GE-Amersham) in a semidry blotting apparatus (Semi-phor, Hoeffer) at 0.8 mA/cm2 for 1 h. Membranes were blocked overnight in 5% non-fat milk in PBS supplemented with 0.1% Tween 20 (T-PBS). Membranes were probed with primary antibodies diluted 1:200 with T-PBS for 60 min at room temperature [polyclonal goat antibodies: anti-HMGCS1 (sc-32422), anti-dihydrodiol dehydrogenase (sc-20425), anti-arginase 1 (sc-18351) from Santa Cruz Biotechnology]. Membranes were then washed four times in 200 ml of T-PBS and incubated with secondary antibody [rabbit anti-goat IgG-HRP (sc-2768), Santa Cruz Biotechnology] diluted 1:5,000 with T-PBS for 60 min at room temperature. After intensive washing with T-PBS membranes were incubated with chemiluminescent substrate [LumiGlo (KPL)] according manufacturer instructions. Membranes were then exposed to Kodak Biomax XAR films.

RESULTS

Mice model of iron overload.

We induced liver iron overload in male mice for our subsequent proteomic studies by dietary iron overload. Sixteen male mice (3 mo old) were divided into two equal groups and fed ad libitum with either defined standard rodent diet (C1000, Altromin, 180 ppm Fe) or the same diet supplemented with 2% carbonyl iron for 8 wk. There were no significant differences in total body weights between control and iron-overload group at the end of the experiment. Liver iron concentration was almost ninefold increased in animals fed carbonyl iron-supplemented diet for 8 wk (P < 0.005). Animals in this group also displayed mild hepatomegaly (∼120 percent of control liver weight) with higher liver-to-body weight ratio compared with control group (P < 0.05) (Fig. 1).

Fig. 1.

Measurement of liver iron content by atomic mass spectrometry and liver-to-body weight ratio in animals fed control (Ctrl) and iron-rich (2% carbonyl iron) diets. Statistical significance was determined by t-test; *P < 0.05, **P < 0.005, n = 4.

Histology, histochemistry, and electron microscopy.

Histological examination of the iron overloaded mice and controls confirmed parenchymal iron loading with minimal histological changes. The Perl's reaction for iron was entirely negative in controls. In iron-overloaded liver samples there was very slight diffuse staining of hepatocytes (consistent with cytosolic ferritin) and discrete granular staining (hemosiderin) both in hepatocytes (in the peribiliary region) and in sinusoidal cells. The highest intensity of the staining was in peripheral zone, decreasing toward the centrozonal region, which was entirely negative. The iron-loaded liver showed signs of growth stimulation: anisokoria and increased number of hepatocytes with double nuclei and occasional mitoses (Fig. 2). The cytological and ultrastructure examination showed an increase in size and amount of hepatocytic organelles (namely mitochondria) toward the central part of the anatomical lobule. No signs of hepatocyte necrosis or apoptosis were found. Borderline microvesicular steatosis was observed, but no signs of fibrosis or Ito cell activation were present in the iron-loaded livers. The amount of glycogen was unchanged. No defects of hepatocyte biliary poles, mitochondria, or lysozomes were present, as judged on histochemical level by unchanged activities of alkaline phosphatase (marker of integrity of biliary pole of hepatocytes), acid phosphatase (marker of integrity of lysosomal compartment), and respiratory chain dehydrogenases and cytochrome oxidase.

Fig. 2.

Growth stimulation in liver of iron-overloaded animals. Arrow shows mitosis.

In summary, using an iron-supplemented diet we induced significant hepatic iron overload in the mice in a pattern similar to that seen in HC. Despite the almost ninefold increase in liver iron content we did not observe dramatic histological changes in the iron-overloaded liver. We would suggest that our subsequent proteomic analysis would be relevant to understanding of moderate iron overload in mildly injured liver.

Differential proteomic analysis (2D-PAGE/MALDI-MS).

We identified changes in protein levels coincident with the excess iron in the liver by differential proteomic analysis of control vs. iron-loaded liver. The comparison was done using 2D-PAGE. We analyzed liver homogenates of control and iron overload animal in quadruplicate (4 gels, each gel originating from 1 animal) for each group (4 gels for the control animals, 4 gels for the iron-overloaded animals). We reproducibly detected 1,020 (±25) spots on colloidal Coomassie blue (CCB)-stained gels (Fig. 3). Lower sensitivity of CCB compared with silver nitrate or fluorescent staining is counterbalanced by higher recovery of peptides from gel after digestion with trypsin. We further increased the applicability of CCB by “mild” electric field conditions during isoelectric focusing (slow voltage gradient in early phases of focusing, maximum voltage not exceeding 5,000 V). That enabled us to load as much as 2 mg of protein per strip without compromising the high-resolution separation.

Fig. 3.

Two-dimensional electrophoretic (2D-PAGE) separation of liver lysates from control and iron-overloaded animals. Isoelectric focusing performed on IPG strips, pH 4–7, was followed by denaturing 10% polyacrylamide SDS electrophoresis in a Tris-glycine buffer system. Gels were stained with colloidal Coomassie blue. Integrated spot density values (spot volumes) were calculated (after normalization and background subtraction). Average values of spot volume (averages from the all 4 gels in the group) for each spot were calculated and compared between the 2 studied groups. Protein spots were considered changed (differentially expressed) only if they met both of the following criteria: average normalized spot volume difference >1.5-fold, statistical significance of the change determined by the t-test, P < 0.05. Arrows with numbers indicate the spots with significant increase of the spot volume relative to the second group of gels.

Gel analysis with Phoretix 2D software (Nonlinear Dynamics) revealed 32 protein spots to be different between the two conditions (selection criteria: average normalized spot volume difference >1.5-fold, statistical significance of the change determined by the t-test, P < 0.05). Fourteen proteins were increased in iron overload whereas 18 were decreased. Relative differences ranged from 1.5 to as much as almost sevenfold. We identified 30 of the 32 protein spots by peptide mass fingerprint with MALDI-time-of-flight mass spectrometry (Table 1, Fig. 4). Only two very small and faint spots (nos. 16 and 30) failed to provide enough protein material for reliable identification. One spot (no. 25) contained two proteins and was excluded from further data interpretation. All of the identified molecules are relatively hydrophilic, soluble proteins. Highly hydrophobic membrane proteins are not separated well by the conventional 2D electrophoresis and therefore are not present on conventional 2D gels (35). We performed a validation of the observed expression changes for three identified proteins (arginase 1, 3-hydroxy-3-methylglutaryl-CoA synthase 1 and dihydrodiol dehydrogenase) by Western blotting (Fig. 5).

Fig. 4.

Matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS) spectrum of tryptic digest of arginase-1 (spot no. 24). Numbers in brackets indicate the first and last amino acids of the identified peptides. *, Unassigned peaks; T, trypsin autolysis products. Sequence of arginase 1 with the identified parts of the sequence (underlined) is shown in the top right corner.

Fig. 5.

Validation of differential expression by Western blotting. Relative expression levels of hydroxymethylglutaryl-CoA synthase 1 (HMGCS1; 58 kDa), arginase 1 (ARG1; 35 kDa), and dimeric dihydrodiol dehydrogenase (DDD; 47 kDa) in control and iron-loaded liver (8 wk, 2% carbonyl iron diet) was assessed by Western blotting. Decreased expression of HMGCS and DDD and increased expression of arginase 1 in iron-loaded liver confirms our 2D electrophoresis results.

View this table:
Table 1.

Liver proteins differentially expressed in response to nutritional iron overload

DISCUSSION

Iron storage and toxic stress.

We observed that the levels of the iron-storage protein ferritin was substantially increased in the iron-loaded livers (L-ferritin 6.7-fold, H-ferritin 3.7-fold) that provides further confirmation of liver iron overload in these mice. The L-ferritin-to-H-ferritin ratio in control and iron loaded liver was 4.5 and 8.6, respectively. Despite the increased ferritin levels we noted increased levels of two isoforms of GRP78/Bip protein in iron overload, which can be indicative of cellular stress in the liver. This chaperone molecule, belonging to the HSP70 family, participates in quality control of secretory proteins in the lumen of the endoplasmic reticulum (9). GRP78/Bip is inducible by several forms of toxic stress, and its upregulation in response to iron excess has been demonstrated previously in human cells (43). Upregulation of GRP78/Bip may indicate increased demand for refolding or retention of proteins in the endoplasmic reticulum of iron-overloaded cells.

Urea cycle.

We identified increased levels of three enzymes of the urea cycle (carbamoyl-phosphate synthase, ornithine carbamoyltransferase, and arginase) in iron overload (Fig. 6). Significantly, CAAT/enhancer-binding protein α (C/EBPα), a transcriptional regulator of the genes encoding these enzymes, has been previously shown to be increased by iron excess in mouse liver (8). C/EBPα also regulates expression of hepcidin, a key regulator of intestinal iron absorption, whose levels also increase in iron overload (8). Urea cycle (ornithine cycle) is an essential five-enzyme system that converts toxic ammonia, produced mainly by amino acid metabolism, into urea. Urea cycle produces arginine, a substrate for nitric oxide synthase and for biosynthesis of polyamines. A connection of urea cycle with iron metabolism has not been previously established, and the role of urea cycle in iron overload remains to be elucidated. However, it is noteworthy that the fifth enzyme of the urea cycle, arginase, competes with nitric oxide synthase for arginine, the sole substrate for nitric oxide production. Indeed, alterations of urea cycle affect nitric oxide production (36). Nitric oxide is known to play an important role in iron homeostasis by modulating the activity of iron regulatory proteins and nontransferrin iron transport (17, 34). Nitric oxide also alters intracellular iron metabolism through iron release from ferritin (33), and via the formation and export of dinitrosyl- and diglutathionyl-dinytrosyl-iron complexes (41). Conversely, expression of inducible nitric oxide synthase is negatively regulated by iron (42). Nitric oxide may therefore represent a key link between the urea cycle and iron homeostasis in the liver.

Fig. 6.

Changes in expression of enzymes of urea cycle (A) and fatty acid oxidation (B) as determined by spot volume. Spot volumes were calculated by Phoretix software after background subtraction and total spot density normalization. Statistical significance of the differences was determined by the t-test; *P < 0.05, **P < 0.005, n = 4.

Peroxisomal fatty acid oxidation and sterol metabolism.

Surprisingly, we noted decreased levels of three enzymes responsible for fatty acid oxidation in iron-overload livers (Fig. 6). Peroxisomal 2-hydroxyphytanoyl-CoA lyase catalyzes the carbon-carbon bond cleavage during α-oxidation of 3-methyl-branched fatty acids, peroxisomal enoyl CoA hydratase is a component of multifunctional protein-2 and catalyzes the second step in the peroxisomal β-oxidation pathway of fatty acid metabolism. Acetyl-CoA thioesterase 1 catalyzes the hydrolysis of acyl-CoA thioesters to the corresponding free fatty acid and CoASH.

Decreased levels of these enzymes could indicate impaired fatty acid catabolism in liver of iron-overloaded animals. Reduced fatty acid oxidation can contribute to the development of liver steatosis (32). Hepatic steatosis is often present in patients with hereditary iron overload (29, 37). It is noteworthy that microvesicular steatosis, although very mild, was present in the liver of iron overloaded mice in our experiment.

Alterations of fatty acid metabolism in response to iron overload has not been reported to our knowledge. However, marked changes in overall lipid metabolism, namely perturbations in plasma lipid transport and hepatic sterol metabolism (decreased activity of 3-hydroxy 3-methylglutaryl CoA reductase) have been demonstrated previously in iron overloaded rats (4). Our study confirmed these results, revealing a decline of both forms of 3-hydroxy 3-methylglutaryl CoA synthase (HMGCSA1, HMGCSA2), enzymes involved in de novo synthesis of cholesterol, which immediately precedes the 3-hydroxy 3-methylglutaryl CoA reductase in the pathway. Surprisingly, another enzyme of cholesterol metabolism, farnesyl diphosphate synthase, was increased in our study, indicating a generalized imbalance of sterol synthesis induced by iron overload. Together, our results suggest that lipid metabolism may be altered in iron overload.

Methylation cycle and methionine metabolism.

Our study revealed specific changes in the levels of two enzymes involved in methionine metabolism and methylation cycle. Adenosylmethonine (AdoMet)-dependent methylation has been shown to be central to many biological processes including gene regulation via DNA, protein methylation, and biosynthesis of phospholipids. S-adenosylhomocysteine (AdoHcy), that is formed after donation of the activated methyl group of AdoMet to a methyl acceptor is a strong competitor of all AdoMet-dependent methyltransferases. AdoHcy is removed by hydrolysis via S-adenosylhomocysteine hydrolase (which was decreased in iron overload). Glycine N′-methyltransferase (increased in iron overload) catalyzes the methylation of glycine by using S-adenosylmethionine (AdoMet) to form AdoHcy and sarcosine, a molecule that has no known physiological role in mammals. AdoHcy is a potent inhibitor of most AdoMet-dependent methyltransferases, and it has been suggested that glycine N′-methyltransferase functions to optimize the AdoMet-to-AdoHcy ratio, and consequently the cellular transmethylation potential (18). Interestingly, sarcosine dehydrogenase, the enzyme that recycles sarcosine back to glycine, was also decreased in iron overload in our study. We noted that this enzyme was present in two spots of different molecular weight and isoelectric point (nos. 1 and 4), which suggest a posttranslational modification.

We hypothesize that the observed decrease of S-adenosylhomocysteine hydrolase combined with increased level of glycine N′-methyltransferase in iron overload could lead to accumulation of AdoHcy, and consequently to the inhibition of numerous AdoMet-dependent methyltransferases. Decreased activity of these proteins can lead to altered phospholipid synthesis and methylation of proteins, small molecules, DNA, and RNA with a multitude of biological effects (6). It is noteworthy that mice lacking one of the key enzymes of methionine metabolism and methylation cycle, methionine adenosyl transferase, are severely affected with evidence of increased oxidative stress, hepatosteatosis, and increased incidence of liver tumors (20, 21). Considering the critical role of the methylation cycle and AdoMet in the regulation of liver function, we hypothesize that iron-induced alterations of methylation pathway could contribute to the liver damage observed in patients with iron overload diseases.

Enzymes of sugar metabolism.

Altered levels of three enzymes of sugar metabolism were detected in our study. Glucokinase was increased in iron overload, whereas pyruvate carboxylase and amylo-1,6-glucosidase (glycogen debranching enzyme) were decreased. Regulation of pyruvate carboxylase expression in response to iron levels has been demonstrated previously by Collins and coworkers (7). Since even small changes in expression of the hepatic glucokinase have a measurable impact on the blood glucose concentration (12, 16), we suggest that the observed changes could affect not only liver metabolism but also systemic glucose homeostasis. This could contribute to the high prevalence of diabetes in symptomatic hemochromatic patients, especially those in whom insulin resistance rather than impaired insulin secretion is observed (22).

Other proteins.

We observed changes in the levels of other proteins in diverse metabolic pathways. Glutathione S-transferase Mu 6, which is a member of a broad family of enzymes involved in elimination of many toxic compounds, was decreased in iron overload. Reduced activity of glutathione transferases have been demonstrated in mice fed iron-rich diet previously (30). δ-Aminolevulinate dehydratase (decreased in iron overload) is an enzyme of heme biosynthesis pathway and has been reported to decline in carbonyl iron-overloaded rats previously (1, 5). Iron excess also resulted in decreased levels of proteasome β3 subunit in our study. Recently, we observed decreased levels of another proteasome subunit (26S subunit 9) in iron-loaded human hepatoma cells (26). Lactamase β2 (decreased in iron overload) is a mammalian mitochondrial protein sharing sequence similarity to the β-lactamase/penicillin-binding protein family of serine proteases that are involved in bacterial cell wall metabolism. The physiological role of liver lactamase is unclear; however, disruption of the gene causes male-specific hepatic microvesicular steatosis (39). Dimeric dihydrodiol dehydrogenase (decreased) detoxifies aromatic hydrocarbons to corresponding catechols. Nothing is known about their possible connections with iron overload.

Previous to this study, a limited number of proteins have been identified whose levels are affected in iron overload. Employing proteomic approach, we identified 30 proteins with altered levels in the liver of mice with nutritionally developed iron overload. The proteins affected give us important clues into the significant metabolic changes in the iron-loaded liver. However, confirmation of the effects and the role of the particular metabolic pathways in molecular pathogenesis of iron overload remains to be determined in future studies. In our previous study on human hepatoma cells HepG2 we identified proteins changing expression in response to high doses of ferrous sulfate in cell culture (26). The lists of the differentially expressed proteins resulting from the present and the previous study do not share many proteins in common with exception of upregulated ferritin and decreased expression of a proteasome subunit. For instance, on the basis of the results from HepG2 cells, we hypothesized that liver cells respond to high iron levels by increased expression of several cytokeratins. However, our present data originating from physiological animal models did not confirm the hypothesis. Although both models provide valuable information about the cellular events triggered by high iron levels, the results are not directly comparable. Hepatoma cells grown in 1 mM ferrous sulfate can be considered an in vitro model of acute iron toxicity, nutritional iron overload in mice represents chronic and systemic iron overload processes.

GRANTS

This research was supported by grants from the Czech Science Foundation (GACR) 303/04/003 and 204/07/0830, Ministry of Health of the Czech Republic (MZCR/UHKT no. 023736, IGA MZCR NR8930-4) and the Ministry of Education, Youth and Sports of the Czech Republic (MSM Projects LC06044 and LC545 and MSM Project no. 0021620806).

Acknowledgments

The authors thank Mrtva Ryba.

Footnotes

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract