In vitro-targeted gene identification in patients with hepatitis C using a genome-wide microarray technology

HepatologyVolume 49, Issue 2 p. 378-386 Viral HepatitisFree Access In vitro–targeted gene identification in patients with hepatitis C using a genome-wide microarray technology† Susanne Hagist, Susanne Hagist Department of Internal Medicine, Salem Medical Center, University of Heidelberg, Heidelberg, Germany These authors contributed equally to this study.Search for more papers by this authorHolger Sültmann, Holger Sültmann Department of Molecular Genome Analysis, German Cancer Research Center, Heidelberg, Germany These authors contributed equally to this study.Search for more papers by this authorGunda Millonig, Gunda Millonig Department of Internal Medicine, Salem Medical Center, University of Heidelberg, Heidelberg, GermanySearch for more papers by this authorUlrike Hebling, Ulrike Hebling Department of Internal Medicine, Salem Medical Center, University of Heidelberg, Heidelberg, GermanySearch for more papers by this authorDörthe Kieslich, Dörthe Kieslich Department of Internal Medicine, Salem Medical Center, University of Heidelberg, Heidelberg, GermanySearch for more papers by this authorRupert Kuner, Rupert Kuner Department of Molecular Genome Analysis, German Cancer Research Center, Heidelberg, GermanySearch for more papers by this authorSabrina Balaguer, Sabrina Balaguer Department of Molecular Genome Analysis, German Cancer Research Center, Heidelberg, GermanySearch for more papers by this authorHelmut-Karl Seitz, Helmut-Karl Seitz Department of Internal Medicine, Salem Medical Center, University of Heidelberg, Heidelberg, GermanySearch for more papers by this authorAnnemarie Poustka, Annemarie Poustka Department of Molecular Genome Analysis, German Cancer Research Center, Heidelberg, GermanySearch for more papers by this authorSebastian Mueller, Corresponding Author Sebastian Mueller sebastian.mueller@urz.uni-heidelberg.de Department of Internal Medicine, Salem Medical Center, University of Heidelberg, Heidelberg, Germany fax: (49)-0-6221-484-494.Department of Internal Medicine, Salem Medical Center, University of Heidelberg, Zeppelinstraße 11–33, 69121 Heidelberg, Germany===Search for more papers by this author Susanne Hagist, Susanne Hagist Department of Internal Medicine, Salem Medical Center, University of Heidelberg, Heidelberg, Germany These authors contributed equally to this study.Search for more papers by this authorHolger Sültmann, Holger Sültmann Department of Molecular Genome Analysis, German Cancer Research Center, Heidelberg, Germany These authors contributed equally to this study.Search for more papers by this authorGunda Millonig, Gunda Millonig Department of Internal Medicine, Salem Medical Center, University of Heidelberg, Heidelberg, GermanySearch for more papers by this authorUlrike Hebling, Ulrike Hebling Department of Internal Medicine, Salem Medical Center, University of Heidelberg, Heidelberg, GermanySearch for more papers by this authorDörthe Kieslich, Dörthe Kieslich Department of Internal Medicine, Salem Medical Center, University of Heidelberg, Heidelberg, GermanySearch for more papers by this authorRupert Kuner, Rupert Kuner Department of Molecular Genome Analysis, German Cancer Research Center, Heidelberg, GermanySearch for more papers by this authorSabrina Balaguer, Sabrina Balaguer Department of Molecular Genome Analysis, German Cancer Research Center, Heidelberg, GermanySearch for more papers by this authorHelmut-Karl Seitz, Helmut-Karl Seitz Department of Internal Medicine, Salem Medical Center, University of Heidelberg, Heidelberg, GermanySearch for more papers by this authorAnnemarie Poustka, Annemarie Poustka Department of Molecular Genome Analysis, German Cancer Research Center, Heidelberg, GermanySearch for more papers by this authorSebastian Mueller, Corresponding Author Sebastian Mueller sebastian.mueller@urz.uni-heidelberg.de Department of Internal Medicine, Salem Medical Center, University of Heidelberg, Heidelberg, Germany fax: (49)-0-6221-484-494.Department of Internal Medicine, Salem Medical Center, University of Heidelberg, Zeppelinstraße 11–33, 69121 Heidelberg, Germany===Search for more papers by this author First published: 28 January 2009 https://doi.org/10.1002/hep.22677Citations: 15 † Potential conflict of interest: Nothing to report. AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinked InRedditWechat Abstract Iron in association with reactive oxygen species (ROS) is highly toxic, aggravating oxidative stress reactions. Increased iron not only plays an important role in the progression of hereditary hemochromatosis (HH) but also in common liver diseases such as chronic hepatitis C. The underlying mechanisms of hepatitis C virus (HCV)-mediated iron accumulation, however, are poorly understood. We introduce an in vitro–targeted approach to identify ROS/iron-regulated genes in patients with HCV using a genome-wide DNA microarray. The sensitivity of the 32,231 complementary DNA clone-carrying microarray was approximately 20% as estimated by detecting target genes of the genome-wide transcription factor hypoxia inducible factor 1α. Upon in vitro challenge to iron and oxidative stress, 265 iron-related and 1326 ROS-related genes could be identified in HepG2 cells; 233 significantly regulated genes were found in patients with mild (HCV) or severe (HH) iron deposition. Notably, 17 of the in vitro–selected genes corresponded to the genes identified in patients with HCV or HH. Among them, natriuretic peptide precursor B (NPPB) was the only iron-regulated gene identified in vitro that was differentially regulated between HCV and HH. Reverse-transcription polymerase chain reaction confirmed most of the microarray-identified genes in an even larger group of patients (n = 12). In patients with HCV, these included genes that are associated with RNA processing (MED9/NFAT, NSUN2), proliferation, differentiation, hypoxia, or iron metabolism (ISG20, MIG6, HIG2, CA9, NDRG1), whereas none of the nine known iron-related genes showed significant differences between HCV and HH. Conclusion: Although high-density microarray technology is less suitable for routine liver diagnosis, its use in combination with prior in vitro selection is a powerful approach to identify candidate genes relevant for liver disease. (HEPATOLOGY 2009;49:378–386.) Approximately 170 million people worldwide are infected with the hepatitis C virus (HCV).1 Up to one-third of these patients progress to cirrhosis within 20 years, and they are at high risk to develop hepatocellular carcinoma (HCC). Currently, the best available therapy is the combination of pegylated interferon alpha and ribavirin,2 which is effective in only 50% of patients, while many either have contraindications to interferon-based therapies or fail to respond to treatment. A better understanding of the host–virus-initiated disease mechanisms is required to develop therapeutic approaches for these patients. Oxidative stress with the release of reactive oxygen species (ROS) is considered a key factor that triggers and contributes to the progression of HCV, and increased markers of oxidative stress such as oxidized glutathione and the lipid peroxidation marker malondialdehyde are commonly observed in the livers of these patients.3-6 Potential sources of ROS formation include inflammatory cells and HCV proteins such as the HCV core protein7 or NS5A.7, 8 Iron is an important cellular toxin and drastically aggravates the toxicity of ROS via the so-called Fenton chemistry. Iron-mediated toxicity can be best observed in patients with hereditary iron overload such as hereditary hemochromatosis (HH) that have a >200-fold increased risk to develop HCC.9 Pathological iron deposition, however, is not only restricted to patients with HH but is also commonly observed (≈60%) in patients with HCV.5, 6, 10-13 In fact, a recent study from several transplantation centers identified iron as an important independent risk factor for the development of HCC in patients with HCV.14 Despite major progress in understanding human iron regulation and hereditary iron overload,15 the mechanisms of hepatic iron deposition in HCV are unresolved.6, 16 Thus, in transgenic mice that express the HCV polyprotein and develop HCC in the presence of mild iron overload,17 the central iron sensor hepcidin18 is decreased despite iron loading and increased inflammatory cytokines (tumor necrosis factor α, interleukin-1, and interleukin-6).18, 19 However, in HCV-infected patients, no altered expression of hepcidin messenger RNA (mRNA) (and thus impaired iron export) but rather an increased hepatocyte iron uptake was observed20 probably due to increased mRNA levels of transferrin receptor 1 (TfR1) and TfR2.21 These diverse findings suggest that, in contrast to posttranscriptional mechanisms,22 additional (post)translational mechanisms could contribute to the mild iron overload found in patients with HCV as recently observed under conditions of sustained nontoxic levels of the central ROS H2O2.23 Complementary DNA microarrays are well established tools to obtain RNA expression profiles24 and they have been applied to study liver diseases such as cancer and drug-mediated hepatotoxicity but also HCV infection (e.g., see Barton and Stivers25 and Girard et al.26). In contrast, large genome-wide microarrays (hereafter referred to simply as microarrays) have not yet been applied in patients with liver disease. They are derived from a complete human gene library and comprise large numbers of unknown transcripts. We present an in vitro–targeted approach to explore such a genome-wide microarray platform in searching for iron and ROS-regulated genes in HCV patients with mild iron overload (as compared with HH). Target genes of the genome-wide transcription factor hypoxia inducible factor 1α (HIF1α) were used to estimate the sensitivity of the microarray platform. Interestingly, most of the identified genes were not among known ROS- and iron-regulated genes but were relevant for RNA processing and proliferation. Thus, genome-wide microarrays allow, in combination with prior in vitro selection, a more efficient identification of candidate genes that are relevant for liver disease. Abbreviations HCC, hepatocellular carcinoma; HCV, hepatitis C virus; HH, hereditary hemochromatosis; HIF1α, hypoxia inducible factor 1α; mRNA, messenger RNA; ROS, reactive oxygen species; RT-PCR, reverse-transcription polymerase chain reaction; TfR1, transferrin receptor 1. Materials and Methods Overall Study Design. The overall study design is shown in Fig. 1. We first characterized the microarray that included allocation of genes, evaluation of hybridization quality, and various filter criteria (Fig. 1, left upper panel). We next estimated sensitivity and reproducibility of the microarray by its ability to identify target genes of the genome-wide acting transcription factor HIF1α in two independent experiments (Fig. 1, right upper panel). Iron- and ROS-regulated genes were then identified in HepG2 cells by exposing them to iron-overload and iron-deficient conditions or ROS (H2O2). The results were compared with a list of known ROS-regulated (n = 197) and iron-regulated genes (n = 41) (Supplementary Table 4). Samples of patients with HCV (n = 12), HH (n = 8), and controls (n = 6) were then processed and hybridized. Hybridizations were ranked according to quality, and four patients of each group with the best hybridization quality and homogenous clinical presentation were chosen for final analysis. Selected genes (n = 20) were confirmed via reverse-transcription polymerase chain reaction (RT-PCR) in an extended group of patients with HCV (n = 12). Cellular experiments/patients including number of hybridizations are given in Table 1. Figure 1Open in figure viewerPowerPoint Overall study design of in vitro–targeted gene selection using a genome-wide microarray. Upon microarray characterization (left panel), sensitivity was estimated by analyzing target genes of the genome-wide transcription factor HIF1α under hypoxia-mimetic conditions (right panel). Iron- and ROS-regulated genes were then identified in HepG2 cells and compared with known iron- and ROS-related genes. Patients with HCV and mild iron overload were analyzed and compared with patients with HH (severe iron overload) and in vitro–identified genes. Finally, 20 selected genes were analyzed via RT-PCR in a larger patient group. Table 1. Overview of Patient/Cellular Microarray Experiments Aim Cells/Patients Experimental Approach Number of Experiments/Patients Samples per Experiment Hybridizations per Sample Samples/Patients Used for Analysis Studies on Cells HIF1α target genes in cells HepG2 Desferroxamin 2 2 2 2 Hemin 2 2 2 2 Iron-regulated genes HepG2 Iron overload 2 2 2 2 Iron depletion 2 2 2 2 ROS-regulated genes HepG2 5 μM H2O2 3 2 2 3 10 μM H2O2 3 2 2 3 control cells HepG2 2 2 2 2 Studies on patients HCV-regulated genes Patient HCV 12 — 2 4 HH-regulated genes Patient HH 8 — 2 4 Control patients Patient Control 6 — 2 4 Human Liver Specimens. An average total mRNA of 2344 ng (minimum 340–13,500 ng) could be obtained from 105 patients that presented to the liver center. Twenty-three of 105 (22%) were below 1 μg RNA, which is required for microarray analysis. Thus, mRNA amplification was necessary to obtain sufficient mRNA for diagnostic microarray analysis. The study was reviewed and fully approved by the institutional review board of the University of Heidelberg. Approximately 5 mm of the liver biopsy cylinder using a 1.4-mm Menghini needle was used for mRNA preparation, leaving 10 mm for routine histological diagnosis. The liver biopsy specimens were immediately snap-frozen in liquid nitrogen and later used for RNA preparation. Patient Selection. After quality control, hybridizations of four out of eight HH patients underwent subsequent analysis (Table 1). As shown in Table 2, these four HH patients had a typical clinical setting of HH confirmed by laboratory tests, increased hepatic iron index, and at least one mutation of the hemochromatosis gene HFE. Furthermore, hybridizations of four out of 12 HCV patients underwent subsequent analysis. These patients were all of HCV genotype I, had a viral load >800,000 U/mL, and mildly increased liver function tests. In addition, they had histological mild iron deposits, slightly increased serum ferritin levels without concomitant signs of inflammation (normal CRP, see Table 2). Four patients with normal liver histology and normal liver tests were used as controls. Reasons for liver puncture of these controls were exclusion of Wilson's disease in four cases with positive family history. Table 2. Baseline Characteristics of Controls and Patients with HH and HCV Used for Microarray Analysis and RT-PCR Parameter Control HCV HFE Mean SD Mean SD Mean SD Age 34.38 ± 9.32 41.04 ± 14.79 57.7 ± 6.5 RNA ng/μL 178.80 ± 210.57 88.28 ± 43.36 57.8 ± 19.3 mRNA purity 2.03 ± 0.01 2.03 ± 0.05 2.0 ± 0.0 Abs. mRNA 4,470.00 ± 5,264.33 2,208.88 ± 1,084.12 1,445.0 ± 481.6 GOT (U/L) 9.33 ± 2.82 22.73 ± 7.93 12.60 ± 3.57 GPT (U/L) 14.67 ± 8.18 61.45 ± 84.03 18.00 ± 7.58 LDH (U/L) 122.50 ± 8.50 134.25 ± 22.73 137.25 ± 40.64 CK (U/L) 31.50 ± 15.50 31.75 ± 16.70 42.33 ± 20.53 GGT (U/L) 10.33 ± 7.59 27.40 ± 20.55 15.25 ± 5.97 ALP (U/L) 98.67 ± 15.11 111.78 ± 60.92 101.25 ± 24.13 CHE (KU/L) 5.35 ± 0.45 5.69 ± 1.41 5.26 ± 1.35 Quick (%) 102.50 ± 4.15 103.42 ± 12.71 107.00 ± 6.04 aPTT (s) 25.35 ± 1.56 25.49 ± 1.43 25.10 ± 3.40 INR-TPZ 1.00 ± 0.00 0.99 ± 0.06 0.98 ± 0.04 CRP (mg/L) 2.45 ± 0.78 2.00 ± 0.00 3.58 ± 2.73 Na+ (mmol/L) 138.75 ± 1.30 141.00 ± 2.97 137.50 ± 0.87 K+ (mmol/L) 3.90 ± 0.40 4.14 ± 0.42 4.01 ± 0.39 Triglyc. (mg/dL) 176.67 ± 93.88 118.33 ± 60.48 179.67 ± 12.28 Chol. (mg/dL) 192.67 ± 15.54 184.67 ± 36.55 178.67 ± 23.70 Leukos (/nL) 6.98 ± 1.95 6.32 ± 1.04 7.36 ± 2.33 Hb (g/dL) 14.48 ± 1.84 13.87 ± 1.82 13.55 ± 1.41 Hke (L/L) 0.40 ± 0.05 0.40 ± 0.04 0.38 ± 0.04 Crea. (mg/dL) 0.76 ± 0.07 0.77 ± 0.14 0.82 ± 0.16 Harnst. (mg/dL) 28.00 ± 10.46 24.70 ± 7.96 35.25 ± 9.34 Ferritin (μg/L) 138.67 ± 83.48 320.20 ± 345.21 1,080.50 ± 238.33 Transferrin (g/L) 2.57 ± 0.19 2.95 ± 0.64 2.20 ± 0.32 TRSG (%) 28.53 ± 6.82 34.42 ± 58.80 ± 29.07 Bill.ges. (mg/dL) 0.50 ± 0.28 0.57 ± 0.27 0.47 ± 0.38 a1AT (g/L) 1.32 ± 0.00 1.64 ± 0.41 1.06 ± 0.00 Coerulopl. (g/L) 0.22 ± 0.03 0.31 ± 0.10 0.28 ± 0.07 HCV-RNA 0.00 ± 0.00 834,008.20 ± 1,041,483.11 ± Fe/kg 6,619.50 ± 4,229.41 FeIndex 2.09 ± 1.42 Cell Experiments to Identify Iron- and ROS-Regulated Genes. An overview of in vitro experiments is shown in Table 1. For the in vitro–targeted identification of iron-regulated genes, HepG2 cells were exposed for 24 hours to conditions of iron overload (100 μM hemin) and iron depletion (100 μM desferroxamine). To identify ROS-regulated genes, cultured cells were exposed to 5 μM (nontoxic) or 10 μM (subtoxic) H2O2. H2O2 was generated enzymatically using the glucose oxidase in combination with catalase as described.23 To select HIF1α target genes, HepG2 cells were exposed with the hypoxia-mimetic desferroxamine (100 μM) for 24 hours (see Results). Western Blotting for HIF1α. Western blots were performed as described23 using 1:250 HIF-1α mouse (Biosciences, Heidelberg, Germany) or 1:500 β-actin (Sigma) antibodies. Microarray Analysis. We used human global microarrays containing 32,231 complementary DNA clones (Human UniGeneSet RZPD3.0; www.imagenes-bio.de). Altogether, the microarray enclosed approximately 20.000 nonredundant genes (13,528 genes [63.5%] with assigned gene name and approximately 6,000 unknown transcripts). Array processing, RNA isolation and amplification, hybridization, and data analysis as well as all quality controls were performed as described.27 Obvious hybridization errors were excluded prior to data analysis. At least two successful hybridizations were required per each experiment. To reduce systematic errors, we explored a universal gene probe for all hybridizations (Stratagene) including cell and patient samples. Data Analysis and Filter Criteria. The data were normalized and expressed as fold up-regulation or down-regulation. To identify iron-regulated genes, the ratio between iron-overload and iron-deficient conditions were calculated (Table 1). For ROS-regulated genes and HIF1α-target genes, controls were compared with cells treated with H2O2 or the hypoxia-mimetic desferroxamine. Samples from patients (HH, HCV) were compared with controls or each other (Supplementary Tables 6–8). Because output of significant gene lists from significance analysis of microarrays can vary considerably,28 we established more restrictive filter criteria/conditions for our study: (1) minimum of four cell samples/patient samples, (2) >2-fold regulation, and (3) standard deviation categories: <20% (strict), <60% (weak, only for patient samples). These filter criteria yielded a much smaller fraction (>5%) of the significance analysis of microarrays identified genes (not shown) and could be confirmed via RT-PCR in an even larger patient group in more than 80% (see results section). Real-Time Quantitative PCR. RNA concentration from the samples described above was adjusted after photometric measurement. A total of 500 ng of RNA were reverse-transcribed using M-MLV reverse transcriptase and 100 pmol of oligo-dT primers (Promega, Madison, WI). Relative mRNA transcript levels from triplicate reactions were quantified using the ABIPrism 7900HT (Applied Biosystems, Foster City, CA) applying hydrolysis technology probes based on the UPL library from Roche (Supplementary Table 1). The analyses were performed by using the SDS software of Applied Biosystems. The expression level of the gene β2-microglobulin (B2M) was used for normalization using the algorithm by Pfaffl.29 PUBMATRIX Cluster Analysis of Identified Genes. PUBMATRIX (http://pubmatrix.grc.nia.nih.gov/) was used to study whether identified genes and their potential functions had been described (e.g., HCV identified genes with HCV) for analysis. Results are shown in Supplementary Table 9 for HCV-, HH-, ROS-, and iron-regulated genes and specific keywords. Results HIF1α-Controlled Target Genes to Estimate Microarray Sensitivity. Because complete RT-PCR validation of microarrays is not possible, we used target gene expression of the genome-wide transcription factor HIF1α to estimate the sensitivity and reproducibility of the microarray. HIF1α was used because (1) it is a key transcription factor that controls more than 150 genes30 and (2) hypoxia-mimetic experimental conditions (e.g., by desferroxamine) lead to a profound and reproducible induction of HIF1α.31 HepG2 cells were treated for 24 hours with 100 μM desferroxamine and up-regulation of HIF1α was assessed via western blotting (Fig. 2). Fifty-five known HIF1α-controlled genes could be allocated on the microarray (Supplementary Table 2). Notably, 12 (21.8%) of these genes were identified in the desferroxamine-treated cells. Seven (58%) of these 12 genes could be reproduced in an independent second experimental approach. These findings suggest that the microarray in combination with our restrictive selection criteria recognizes approximately 20% of all genes, representing a feasible sensitivity for a genome-wide microarray. Figure 2Open in figure viewerPowerPoint Sustained induction of the genome-wide transcription factor HIF1α in the presence of the hypoxia-mimetic desferroxamine. Cultured cells (HepG2) were exposed to 100 μM desferroxamine over 24 hours, and HIF1α expression was determined via western blotting. Identification of Iron-Regulated Genes in HepG2 Cells. We next exposed cultured hepatoma cells (HepG2) for 24 hours to iron-depleted (100 μM desferroxamine) or iron-repleted (100 μM hemin) conditions and extracted the differentially regulated genes. Both iron-depleted and iron-repleted conditions are routinely used to study iron-regulated mechanisms, because cell proliferation and serum media can considerably modify iron concentrations under in vitro conditions. Table 3 gives an overview of the in vitro–identified genes. In the two independent experiments, 196 and 103 genes were identified with an overlap of 27 (13%-26%) genes (Supplementary Table 3). Altogether, the two independent experiments identified 265 iron-regulated genes. Forty-one genes could be allocated on our microarray that are known to be involved in iron metabolism such as TfR1 or ferritin (Supplementary Table 4). However, none of these 41 iron-related genes were among the 265 genes identified in the iron-based in vitro approach. The divalent metal transporter NRAMP1 was the only identified iron-related protein when weaker selection criteria were applied (1.5-fold regulation and standard deviation <0.8). Table 3. Overview of In Vitro–Identified Genes Known ROS- and Iron-Related Genes In Vitro–Identified Genes In Vitro–Identified EST's In Vitro–Identified and Known Genes Patient-Identified Genes HH HCV HCV vs HH 42 63 144 Patient In Vitro–Identified Genes H202 197 1326 394 10 3 3 10 Iron 41 265 70 0 1 1 1 Identification of ROS-Regulated Genes in HepG2 Cells. We next applied an in vitro search strategy for ROS-related genes by exposing cultured cells continuously to an H2O2 generating system over 24 hours at either nontoxic (<5 μM [n = 3]) or subtoxic (10 μM [n = 3]) concentrations of H2O223 (Table 1). H2O2 was chosen since it is a central ROS that easily converts into other more reactive ROS within cells. Seven hundred thirty genes were down-regulated, and 596 were up-regulated. Altogether, 1,326 genes were regulated, among them 124 under toxic conditions (Supplementary Table 5). The rather high number of ROS-identified genes is due to several reasons: First, we used more conditions (toxic versus nontoxic). Second, more independent experiments were performed (six versus two). Finally, redox- sensitive transcription factors are more abundant as compared with, for example, iron-sensitive factors and respond in a less specific manner to a variety of stress conditions. One gene (EML1) was identified in three, 42 genes in two, and the rest were identified only in one of six independent experiments. Notably, the in vitro approach recruited 10 genes from a list of 197 known-redox regulated genes (Table 3). Gene Expression in Patients with Chronic Hepatitis C and HH. We next identified differentially expressed genes in patients with mild iron (HCV) or severe iron overload (HH) compared with the control group (n = 4 each). Special care was taken to select homogenous patient groups without comorbidities (see Materials and Methods). The number of identified genes and their relation to the in vitro–selected genes are shown in Table 3. Sixty-three significantly regulated genes were found in HCV versus control, 42 genes in HH versus control, and 144 in HH versus HCV (Supplementary Tables 6–8). About 50% of these genes are known, while the rest are unknown transcripts. We then compared these patient-identified genes with the iron- and ROS-identified genes from HepG2 cells. Seventeen genes were both differentially regulated in patients and cells and most of them (n = 10) were ROS-related (Table 4). Only two iron-regulated genes from the in vitro approach were significantly regulated in patient samples (natriuretic peptide precursor B in HCV, and vascular endothelial growth factor in HH). Among the 41 known iron-regulated genes (Supplementary Table 4), the following genes were also significantly regulated in patients: TFR2 in HH and TFR1, ERB4, FTH1, and transferrin in HCV. Hemojuvelin and homogentisate oxidase were the only iron-related genes that were differentially regulated between patients with HH and HCV. Table 4. Genes that Are both Significantly Regulated in an ROS- and Iron-Targeted In Vitro Approach and in Patients with Mild to Severe Iron Accumulation (HH or HCV) Gene Abbreviation Complete Gene Name Iron H2O2 HH versus Control HCV versus Control HCV versus HH Chromosome RZPD clone ID Genbank ID ANAPC10 Anaphase promoting complex subunit 10 3.0 0.5 4:q31 IMAGp998N21595 N29883 (gb). N42530 (gb) BCL7B B cell CLL/lymphoma 7B 0.4 0.3 3.9 7:q11.23 IMAGp998M23417 H69096 (gb). H69097 (gb) CFTR Cystic fibrosis transmembrane conductance regulator. ATP-binding cassette (subfamily C, member 7) 2.1 2.4 7:q31.2 IMAGp9981216073 AI890417 (gb) CNR1 Cannabinoid receptor 1 (brain) 0.4 0.5 6:q14-q15 IMAGp998M08139 R13505 (gb), R20626 (gb) DKK3 Dickkopf homolog 3 (Xenopus laevis) 0.5 2.7 11:p15.2 IMAGp998C22269 H07079 (gb), H07080 (gb) FLJ14494 RNA pseudouridylate synthase domain containing 4 0.4 11:q24.2 IMAGp998F024693 AI309208 (gb) LOC283745 Hypothetical protein LOC283745 2.5 2.5 15:q14 IMAGp998L183557 AA868190 (gb) NPPB Natriuretic peptide precursor B 2.1 0.5 2.5 1:p36.2 IMAGp998P034427 AI186050 (gb) NSUN2 NOL1/NOP2/Sun domain family, member 2 0.5 0.2 5:p15.31 IMAGp998I221937 AA447911 (gb), AA448850 (gb) PDE4D Phosphodiesterase 4D, cAMP-specific (phosphodiesterase E3 dunce homolog, Drosophila) 2.1 2.3 5:q12 IMAGp998N155835 AI660657 (gb) PTPRF Protein tyrosine phosphatase, receptor type, F 2.0 2.4 1:p34 IMAGp998C16113 T96626 (gb), T96737 (gb) SLC17A7 Solute carrier family 17 (sodium-dependent inorganic phosphate cotransporter), member 7 0.4 0.4 19:q13 IMAGp998K18147 R15334 (gb). R42231 (gb) TPM1 Tropomyosin 1 (alpha) 0.4 15:q22.1 IMAGp998H054248 AI056202 (gb) TTLL3 Tubulin tyrosine ligase-like family, member 3 0.5 2.8 3:p25.3 IMAGp998M20727 W33081 (gb). W33082 (gb) VEGF Vascular endothelial growth factor 2.3 2.0 6:p12 IMAGp998H24733 W44690 (gb). W44691 (gb) EST's 0.3 IMAGp998L195286 AI473450 (gb) Complementary DNA FLJ36315 fls. clone THYMU2005240, highly similar to 40S RIBOSOMAL PROTEIN S6 0.4 2.2 5 IMAGp998L113814 AA905507 (gb) Validation by RT-PCR. To validate our microarray data, we performed RT-PCR for 20 identified genes. These genes (including β2-microglobulin) were selected for the following reasons: (1) they were significantly regulated in patients with HH or HCV (n = 6), (2) they were identified in the in vitro approaches by H2O2 or iron (n = 4), or (3) they were known iron-regulated genes of potential interest (n = 9) (Table 5). Eight additional HCV patients with genotype 1B were included to increase the statistical robustness for the analysis of HCV (final, n = 12). Confirmed results are shown in Table 5. EIF2C4 and KCNJ11 were confirmed via RT-PCR in HH patients, ISG20 and NSUN2 in HCV patients, and KCNJ11 and MED9 in HCV versus HH. Only ZNF165 (HCV) was not statistically relevant. Of the nine known iron-related genes, ACO1, FTH1, transferrin, and TfR1, but not ferroportin, were significantly downregulated in patients with HCV, confirming the microarray-obtained results above (Supplementary Table 4) for TfR1, transferrin, and FTH1. TfR1, ferritin, and NRAMP2 were down-regulated in patients with HH at a lower level of significance (P < 0.1). Comparing HCV and HH, none of the known iron-related genes reached a considerable level of significance. Clearly discriminative between HH and HCV were mRNA-processing MED9/NFAT5, the interfer