Hepatic gene expression in patients with obesity ...

2 downloads 0 Views 204KB Size Report
Measure: Differential hepatic gene expressions in patients with NASH as compared with ...... Benyon R C, Arthur M J. Extracellular matrix degrada- tion and the ...
Liver International 2005: 25: 760–771 Printed in Denmark. All rights reserved

Copyright r Blackwell Munksgaard 2005

Clinical Studies

DOI: 10.1111/j.1478-3231.2005.01117.x

Hepatic gene expression in patients with obesity-related non-alcoholic steatohepatitis

Younossi ZM, Gorreta F, Ong JP, Schlauch K, Del Giacco L, Elariny H, VanMeter A, Younoszai A, Goodman Z, Baranova A, Christensen A, Grant G, Chandhoke V. Hepatic gene expression in patients with obesity-related non-alcoholic steatohepatitis. Liver International 2005: 25: 760–771. r Blackwell Munksgaard 2005 Abstract: Background: Non-alcoholic fatty liver disease (NAFLD) is among the most common causes of chronic liver disease. NAFLD includes a spectrum of clinicopathologic syndromes that includes non-alcoholic steatohepatitis (NASH) that has potential for progression. The pathogenesis of NASH is poorly characterized. Aim: This study was designed to identify differences in hepatic gene expression in patients with NASH and to relate such differences to their clinical characteristics. Design: Consecutive patients undergoing bariatric surgery were prospectively recruited. Extensive clinical data and two liver biopsy specimens were obtained at the time of enrollment. A single hepatopathologist reviewed and classified the liver biopsies. Patients with excessive alcohol use and other causes of liver disease were excluded. A group of 29 NASH patients, 12 with steatosis alone, seven obese controls and six non-obese controls were selected for further investigation. Customized cDNA microarrays containing 5220 relevant genes were designed specifically for this study. Microarray experiments were run in triplicate for each sample and a selected group of genes were confirmed using real-time PCR. Outcome Measure: Differential hepatic gene expressions in patients with NASH as compared with controls. Results: Thirty-four genes with significant differential expression were identified in patients with NASH when compared with non-obese controls. Moreover, 19 of these genes showed no significant expression differences in obese vs. non-obese controls, suggesting a stronger association of these genes to NASH. Conclusions: Several differentially expressed genes in patients with NASH are related to lipid metabolism and extracellular matrix remodeling. Additionally, genes related to liver regeneration, apoptosis, and the detoxification process were differentially expressed. These findings may help clarify the molecular pathogenesis of NASH and identify potential targets for therapeutic intervention.

Non-alcoholic fatty liver disease (NAFLD) is an important cause of chronic liver disease worldwide (1–8). NAFLD includes a spectrum of clinicopathologic syndromes with varying degrees of risk for progression to cirrhosis (1, 9–11). For example, non-alcoholic steatohepatitis (NASH) is considered potentially progressive, whereas patients whose liver biopsies show steatosis alone seem to follow a more benign course with little or no progression (1, 9–15). Furthermore, burned-out NASH may be an important cause of cryptogenic cirrhosis (16–18). Although initially described in 1980, the current rise in the prevalence of NAFLD and NASH may be related to parallel 760



Zobair M. Younossi1,2, Francesco Gorreta2, Janus P. Ong1,2, Karen Schlauch1,2, Luca Del Giacco2, Hazem Elariny1,2, Amy VanMeter2, Abraham Younoszai1,2, Zachary Goodman3, Anna Baranova2, Alan Christensen2, Geraldine Grant2 and Vikas Chandhoke2 1 Center for Liver Diseases, Inova Fairfax Hospital, Falls Church, VA, USA, 2Center for the Study of Genomics in Liver Diseases, Molecular and Microbiology Department, George Mason University, Manassas VA, USA, 3Armed Forces Institute of Pathology, Washington, DC, USA

Key words: NASH – gene expression – micro arrays Zobair M. Younossi, MD, MPH, Center for Liver Diseases, Inova Fairfax Hospital, 3300 Gallows Road, Falls Church, VA 22042, USA. Tel: 11 703 698 3182 or 11 703 208 6650 Fax: 11 703 698 3482 or 11 703 208 6655 e-mail: [email protected] Received 28 September 2004, accepted 11 January 2005

increases in the prevalence of obesity and metabolic syndrome (19–30). Although the pathogenesis of NAFLD remains largely unknown, insulin resistance (IR) seems to contribute to its progression (31, 32). Several endogenous and exogenous factors may influence this differential progression in NAFLD (32–34). For example, exogenous factors enhancing oxidative stress may compound hepatic injury in some patients, whereas genetic factors may be more important in others (32, 34–44). Additionally, an interaction among these factors may influence the disease course and tip the balance in favor of progression (4).

Gene expression in non-alcoholic steatohepatitis The goal of this study was to elucidate steps in the complex pathogenesis of NASH through a genomic approach using the microarray technology (45–47). This study attempts to relate the patterns of hepatic gene expression in patients with NASH to those seen in non-NASH controls. Our hope is that the identification of some of the differentially expressed genes will help clarify the pathogenesis of NASH and potential targets for new treatments for NASH. Design, setting and patient population

Patient selection

This study was based on our ongoing epidemiology of NAFLD (EPI-NAFLD) database, which was created by enrolling consecutive patients undergoing bariatric surgery at Inova Fairfax Hospital from October 2001 to May 2002. Extensive clinical and laboratory data were collected on each patient after obtaining informed consent. Patients with evidence of excessive alcohol use (410 g/day) or other causes of liver disease were excluded. IR was defined as a homeostatic model assessment index 42.2 (48), and diabetes mellitus was defined as a clinically established diagnosis treated with diet control or anti-diabetic medications or both. Two liver biopsy specimens were obtained at the time of surgery. One specimen was sent to the study pathologist (Z.G.) for routine pathologic assessment and the other was immediately snap-frozen with liquid nitrogen and used later for RNA extraction. Control liver specimens were obtained from potential liver transplant organ donors or from patients undergoing hepatic resection for liver mass (e.g., hepatic hemangioma), neither of whom had clinical or histologic evidence of chronic liver disease. This study was reviewed and fully approved by the Institutional Review Board of Inova Fairfax Hospital.

tion, Kupffer cell hypertrophy, apoptotic bodies, focal parenchymal necrosis, glycogen nuclei, hepatocellular ballooning, and Mallory bodies. These features were graded as follows: 0 5 none; 1 5 mild or few; 2 5 moderate; 3 5 marked or many. Fibrosis was assessed with the Masson trichrome stain. Portal fibrosis and intralobular pericellular fibrosis were graded as 0 5 none; 1 5 mild; 2 5 moderate; or 3 5 marked. When present, bridging fibrosis was noted as few or many bridges, and cirrhosis was identified when parenchymal nodules surrounded by fibrous tissue were noted. Cirrhosis was further categorized as incomplete or established, depending on the degree of loss of acinar architecture (41). Using this pathologic protocol, NASH was defined when, in addition to steatosis, at least one unequivocal Mallory body was identified on the hematoxylin and eosin stain and/or some degree of zone 3 pericellular fibrosis or bridging fibrosis identified on the trichrome stain (41). Microarray technique

RNA preparation Total RNA was extracted from 20 to 30 mg of liver tissue by using the RNeasys Mini Kit (Qiagen, Valencia, CA) and treated by DNAfreet (Ambion, Austin, TX). Abundance and ratio between 28S and 18S rRNA were monitored both by agarose gel electrophoresis and Agilent 2100 Bioanalyzer (Agilent, Palo Alto, CA). Total RNA pooled from 10 different human cell lines (Universal Human Reference, Stratagene, La Jolla, CA, Cat # 740000) was used throughout the experiment as a reference RNA sample. Liver RNA and reference RNAs (1.5 mg) were amplified with the MessageAmp aRNA Kit (Ambion).

Pathologic assessments

Reverse transcription and sample labeling aRNA (4 mg) was reverse transcribed and labeled according to The Institute for Genomic Research (TIGR) protocol (http://www.tigr.org/tdb/microarray/protocolsTIGR.shtml).

Each liver biopsy was fixed in formalin, routinely processed, sectioned, and stained with hematoxylin and eosin and Masson’s trichrome. All biopsies were evaluated by a single hepatopathologist (Z. G.). The degree of steatosis was assessed in hematoxylin- and eosin-stained sections and graded as an estimate of the percentage of tissue occupied by fat vacuoles as follows: 0 5 none; 1 5 up to 5%; 2 5 6–33%; 3 5 34–66%; 4 5 4 66%. Other histologic features evaluated in hematoxylin and eosin sections included portal inflammation, lymphoplasmacytic lobular inflammation, polymorphonuclear lobular inflamma-

Array fabrication A custom microarray was produced from 5220 human cDNA clones (Research Genetics, Carlsbad, CA). These selected clones represented many genes involved in inflammation pathways and genes related to liver diseases as well as 329 ESTs and  1000 other cDNAs of unknown function. A complete list of genes with their accession numbers is available at http://www.gmu.edu/ centers/genomics/research/keys. cDNA clone inserts were amplified directly from clones in culture using GF200F (5 0 -CTGCAAGGCGATTAAGTTGGGTAAC) and GF200R (5 0 -GTGAGCGGA761

Younossi et al. TAACAATTTCACACAGGAAACAGC) universal primers and electrophoresed on agarose gels to monitor the yield and the specificity of the amplification reactions. PCR products were purified, dried and then re-suspended in 30 ml of 3  saline sodium citrate (SSC). The cDNAs were printed on poly-L-lysine-coated slides in a single replicate by using Gene Machines OGR-03 OmniGrid Microarrayer with SMP3 pins (Telechem International, Sunnyvale, CA) along with controls consisting of no-template PCR amplifications (n 5 67) and empty wells (n 5 384). Additionally, as cDNA libraries might contain clones with mismatched descriptions (49), 10 clones from the gene lists were randomly selected and sequenced. The identity of all clones was confirmed. Prehybridization and hybridization Spotted cDNAs were first rehydrated in a humidity chamber until glistening (1  SSC, 1 min and 30 s), denatured (95 1C for 4 s), and then UV crosslinked to the slide surface. Slides were incubated at 45 1C for 45 min in prehybridization buffer containing 5  SSC, 0.1% SDS, 1% bovine serum albumin (Sigma, Research Triangle Park, NC). After washing in MilliQ water, the slides were dipped in isopropanol and air-dried. The labeled cDNA probes were re-suspended in 23 ml of hybridization buffer (25% formamide, 1  SSC, 0.1% SDS). Each sample probe was combined with a reference probe in a total volume of 46 ml, denatured at 95 1C for 3 min, and applied to a prehybridized microarray slide. The microarray slide was incubated at 45 1C overnight in a sealed hybridization chamber (Corning Costar, Acton, MA). Slides were washed twice in 1  SSC, 0.2% SDS (10 min, 45 1C), twice in 0.1  SSC, 0.1% SDS (10 min, 45 1C), twice in 0.1  SSC (10 min, 45 1C), rinsed in MilliQ water and dried. Image analysis Hybridized slides were scanned with the confocal laser scanner ScanArrayt Express HT (Packard BioScience, Boston, MA) using settings of 75% of photomultiplier tube, 75% of laser power, and a 10 mm pixel resolution. Images were acquired by ScanArrayt Express 2.0 software (PE Life Sciences, Boston, MA) and processed with QuantArray 3.0 software (PE Life Sciences) to quantify intensity levels and local background for both Cy3 and Cy5 channels of each spot. Quality control and processing of microarray data

Microarray data were generated for NASH patients and the non-NASH controls. The amount of available RNA allowed us to run triplicate arrays for almost all samples, yielding 121 total arrays. 762

The quality of all microarrays was assessed with a series of control methods. A heatmap image of each microarray was generated to identify notable spatial variation, uneven hybridization, arraywide low or high overall abundance levels, high or uneven background intensity values and various contaminations. Arrays with any identifiable problems were excluded from the study. Additionally, arrays with unreliable spot replication, or uncharacteristically high negative control values were also excluded. This quality control process yielded 77 arrays of good quality for 29 NASH patients, seven obese non-NASH controls and six non-obese controls. Spot replicates on each microarray were also inspected for reproducibility. The average Pearson’s correlation coefficient of these replicate pairs was 0.74 across all arrays used in this study. Variability across triplicate ratios was also computed as a measure of quality control. After normalization, the average standard deviation of 72 155 ratio triplicates was 0.25. To eliminate as much systematic variation as possible from the expression data, several data processing and normalization steps were also applied. Local background, as supplied by QuantArrayt image analysis software, was subtracted first from each channel for each individual gene. Genes with background values greater than intensity values in either channel were discarded from each array (average percentage of genes discarded was less than 0.3%). Each microarray was filtered to determine which intensity levels were significantly greater than background values, and thereby deemed detected (50). A perchip background threshold for each channel defined as the mean negative control intensity of each channel on the array provided a useful threshold for these customized microarrays (50). Intensity values lower than the threshold in either channel were set equal to the threshold in that channel to avoid the computation of artificially large expression ratios (50). Dye bias was measured on a per-chip basis, by performing a linear regression on the blank well and negative control spot values. Every chip exhibited a clear, significant linear trend in Cy5 vs. Cy3 measures. Expression ratios (Cy5/Cy3) were computed, and normalized using a linear median normalization (division by the per-chip median ratio). Average ratios were then computed across the set of replicated arrays of each sample. Statistical analyses

A reference design allowed comparisons of individual samples to a common reference, as well

Gene expression in non-alcoholic steatohepatitis as cross-comparison of several different patient groups. Gene expression activity across groups was assessed by comparing average expression ratios of each group. Genes exhibiting more than twofold difference between group averages were identified, and Mann–Whitney rank-sum tests were performed on the median expression ratios to determine statistical significance. A multiple testing correction (49) controlled the false discovery rate, which is the expected proportion of false positives in the rejected hypotheses. Genes with significant differences between groups were identified and examined with respect to their role in specific molecular pathways. A standard hierarchical clustering procedure, written in the R statistical program language (http://cran. r-project.org/), was also applied to the microarray data to identify and investigate possible patterns in gene expression profiles across and within disease groups (51, 52).

real-time RT-PCR. For normalization purposes, 18S ribosomal RNA level was tested in parallel with the genes of interest (Table 1) (53). Reactions were performed in a 96-well format in the BioRad iCycler iQ Real-Time Detection System (BioRad Laboratories, Hercules, CA). For each gene, three independent PCR experiments from the same reverse transcription sample were performed. The presence of a single specific PCR product was verified by melting curve analysis and confirmed on agarose gel. Results

Patient characteristics

Clinical and demographic data for the NASH cohort (N 5 29), steatosis alone (N 5 12), obese controls (n 5 7) and non-obese controls (n 5 6) are summarized in Table 2. Differential gene expression

Real-time RT-PCR verification

The differential expression of a group of genes identified by microarray analysis was validated by

Group comparisons A number of comparisons were performed to detect genes potentially involved in NASH. The

Table 1. Selected genes differentially expressed in microarrays confirmed by real-time PCR Gene name and symbol

Primer sequences

3-hydroxy-3-methylglutaryl-Coenzyme A synthase 2 (HMGCS2)

F: TCTCTGGCTCGCCTGATGTTC R: AGTGCTTTATCCAGGTCCTTGTTG F: ACCCAGAAAACTTGGGCATT R: CAGAGTTTTCTGAATATAATTC F: AAAGCTGATGTCCTGACCAC R: CAATATGGGATCCCTGATGA F: CATCTCCCATCTTCCTCTCGC R: GTTCAGCATTCACACTTTCCAG F: AGGAATTCCCAGTAAGTGCG R: GCCTCACTAAACCATCCAA

Acyl-CoA synthetase long-chain family member 4 (ACSL4/FACL4) Catalase (CAT) Activating transcription factor 3 (ATF3) 18S (49)

Table 2. Clinical and demographic data of NASH cohort and controls

Age Hip/waist ratio Body mass index ALT AST ALT/AST Fasting serum glucose (mg/dl) Fasting serum cholesterol (mg/dl) Fasting serum triglyceride (mg/dl) Fasting serum insulin (mU/ml) HOMA Percent female Percent Caucasian Percent with type II diabetes Percent hyperlipidemia

NASH (n 5 29)

Steatosis alone (n 5 14)

Obese controls (n 5 7)

Non-obese controls (n 5 6)

39.9  9.1 1.1  0.1 50.3  9.2 37.3  29.3 m/l 31.9  26.1 1.2  0.4 113.4  33.6 199.7  35.6 179.25  69.3 12.3  7.6 3.7  2.7 69 69 21 21

38.8  8.1 1.1  0.1 47.6  6.8 18.9  9.0 18.0  5.3 1.0  0.3 104.3  60.1 186.1  36.0 120.1  67.0 5.3  4.4 1.5  1.9 86 64 21 29

40.5  14.7 1.1  0.2 42.3  4.1 26.0  16.3 m/l 23.9  7.6 1.0  0.4 92.4  33.4 191.3  42.9 125.2  79.1 7.3  2.2 1.8  0.7 100 71 0 43

45.2  17.8 N/A 24.4  4.2 20.6  10.0 m/l 27.5  12.8 m/l 0.9  0.5 109.3  24.6 N/A N/A N/A N/A 50 100 17 N/A

N/A, data not available; NASH, non-alcoholic steatohepatitis.

763

Younossi et al. first analysis compared NASH expression profiles to non-obese control’s profiles. Because obesity is a potential risk factor to NASH, and thus may act as a confounding factor, NASH patients were also compared with the obese controls without NAFLD. Furthermore, differential expression between the obese controls and non-obese controls was also evaluated to identify genes most likely to be related to obesity. In this analysis, genes that were differentially expressed in NASH patients and obese controls were of greatest interest. Genes that were differentially expressed in NASH patients vs. the non-obese controls and in the obese controls vs. the non-obese controls were considered to be characteristic of liverspecific transcriptome deregulation related to obesity (see Discussion). Differential gene expression values that reached statistical significance are summarized in Tables 3–5. Compared with non-obese controls, patients with NASH exhibited 34 differentially expressed genes (4twofold, adjusted P-value o0.05). The expression of the majority of these genes (19/34) was not significantly different in the obese and non-obese controls, suggesting NASH-specific gene expression. Some of these identified genes encode key enzymes of lipid metabolism (acylcoenzyme A (CoA) synthetase long-chain family member 4 and mitochondrial 3-hydroxy-3methylglutaryl-CoA synthase), and extracellular matrix remodeling (chondroitin sulfate proteoglycan core protein 2 (CPSG2) and syndecan 4 (SDC4)). Furthermore, genes involved in liver regeneration, apoptosis and the detoxification process were also differentially expressed, suggesting the possible involvement of these processes in the pathogenesis of NASH. In addition to the experiments described above, microarray profiling of NAFLD patients with steatosis alone was also included. These NAFLD patients had histologic evidence for steatosis alone but did not fulfill the strict NASH criteria applied in this study. In comparison with NASH, the group of NAFLD patients with steatosis alone was characterized by downregulation of insulin-like growth factor binding protein 1 (IGFbp-1), encoding for IGFbp-1 and upregulation of acyl-CoA synthetase long-chain family member 4 (ACSL4), encoding for long-chain fatty-acid-CoA ligase 4. Clustering the gene expression profiles A standard hierarchical clustering procedure was applied to the gene expression profiles of the NASH and control groups across the differentially expressed genes that are listed in Tables 3–5. Using the Euclidean distance metric as a dissim764

ilarity measure, and Ward’s Minimum Variance agglomeration method, the NASH and non-obese controls are clearly separated into two distinct clusters (51). The same procedure was applied to the NASH and obese control groups. Using this clustering procedure, NASH patients were clearly separated from the non-obese controls, but not from the obese controls (data not shown). Verification of gene regulation via real-time RT-PCR Changes in gene expression profiles observed in the microarrays were validated by real-time RT-PCR analysis by using the same total RNAs analyzed in the cDNA microarray experiments. 18S ribosomal RNA level was used to normalize the experiments (71). Differences in the dynamics of the expression values obtained with the two approaches are related to differences in sensitivity between the two techniques (usually spot signal saturation limits the dynamic range of detection in microarrays). Four genes were selected for real-time RTPCR confirmation: activating transcription factor 3 (ATF3), catalase (CAT), 3-hydroxy-3-methylglutaryl-CoA synthase (HMGCS2) and ACSL4. This verification confirmed that all the RT-PCR results were in agreement with the microarray data (Fig. 1). Discussion

The aim of this study was to analyze differential gene expression in patients with NASH in whom we had also collected extensive clinical and histological data. This integrative approach, combining gene expression data and clinical data, has enabled us to generate genotype–phenotype associations for patients with biopsy-proven NASH. As obesity is a risk factor for NASH, the gene expression analyses included obese patients with NASH as well as both the obese and non-obese controls. This step-wise approach to gene expression analysis may lead to a better understanding of the processes involved in the pathogenesis of NASH. We have identified several potentially important gene expression patterns in patients with biopsy-proven NASH. This analysis highlighted deregulation of the ketogenesis pathway in NASH patients. In fact, these results suggest upregulation of mitochondrial HMGCS2 in patients with biopsy-proven NASH. This enzyme is the most important rate-limiting enzyme for ketogenesis (54). Upregulation of this gene supports the theory that an increase in ketone body formation, lipolysis, and fatty acid (FA) oxidation are involved in the pathogenesis of NASH

Gene expression in non-alcoholic steatohepatitis Table 3. Differential gene expression comparing NASH and non-obese controls

Gene name Metabolism Acyl-CoA synthetase long-chain family member 4 3-hydroxy-3-methylglutarylCoenzyme A synthase 2 (mitochondrial) Delta aminolevulinate synthase 1 S100 calcium-binding protein A8 (calgranulin A, MRP8) Serine dehydratase Glutamic-oxaloacetic transaminase 1, soluble (aspartate aminotransferase 1) Insulin signaling Insulin-like growth factor-binding protein 2 (36 kDa) Serum/glucocorticoid regulated kinase Transcription factors jun B proto-oncogene v-jun avian sarcoma virus 17 oncogene homolog

Gene name

NASH vs. non-obese controls

P-values

Biological function

ACSL4

2.555

0.020

HMGCS2

2.061

0.009

Central enzyme controlling the unesterified arachidonic acid (AA) level in cells. Rate-limiting enzyme of the HMG-CoA pathway of fatty acid metabolism (ketogenesis)

ALAS1

0.495

0.013

S100A8

0.344

0.004

SDS GOT1

0.369 0.384

0.004 0.006

IGFBP2

0.385

0.0265

SGK

0.4

0.042

JUNB

0.455

0.009

JUN

0.202

0.004

0.412

0.020

0.442

0.042

0.482

0.042

Potent regulator in liver cell growth. Upregulated in liver regeneration Transmembrane heparan sulfate proteoglycan localized to focal adhesions. Also regulates inositol phospholipid binding and signaling

0.452

0.004

Counterregulator of inflammatory processes

0.444

0.0064

A vasodilatory peptide protecting liver against organ damage via oxidative stress

PLSCR1

0.459

0.042

SH3BGRL2

2.093

0.009

PRSS3 AFP

2.105 2.647

0.030 0.009

Participates in transbilayer movement of phosphatidylserine and other phospholipids. Also possesses nuclear functions and promotes Src kinase activation through the EGF receptor Unknown. Protein belongs to thioredoxin-like protein superfamily Digestive degradation of trypsin inhibitors Marker of liver carcinogenesis, major serum protein synthesized during fetal life

v-ets avian erythroblastosis virus ETS2 E26 oncogene homolog 2 Extracellular matrix and associated signaling Fibrinogen-like 1 FGL1 Syndecan 4 (amphiglycan, ryudocan)

SDC4

Inflammation and suppression of inflammation Interleukin 1 receptor antagonist IL-1RN Cellular defenses against the oxidative stress Adrenomedullin ADM Other Phospholipid scramblase 1

SH3 domain-binding glutamic acidrich protein like 2 Serine protease 3 (mesotrypsin) a-fetoprotein

Heme biosynthetic pathway. ALAS1 is regulated by AP-1 complex Positive regulator of NADPH oxidase activation and extracellular transporter of arachidonic acid. S100A8 is chemotactic for myeloid cells L-Serine metabolism Amino acid metabolism; urea and tricarboxylic acid cycles

Modulates the availability of unbound IGF1 for interaction with IGF1R SGK inactivates ubiquitin ligase Nedd4-2. SGK upregulates by IGF1 TF activating synthesis of acute phase poteins in the liver. TNF-a increases JunD TF activated by JNK kinase controlling expression of many genes involved in the immune response. In ob/ ob mice JNK kinase not able to be activated by partial hepatectomy TF cooperating with the AP-1 transcription factor

Differential expression values, P-values (adjusted for multiple tests using the Benjamini–Hochberg method (49)) and biological function of the differentially expressed genes between NASH patients and non-obese controls. These genes are not differentially expressed between non-obese controls and controls, and are thus potentially more NASH specific. NASH, non-alcoholic steatohepatitis.

(55). Additionally, insulin may affect the expression of HMGCS2, suggesting the potential role of insulin resistance in regulating this pathway. A subgroup analysis provided further support for this hypothesis. This subgroup analysis showed that NASH patients with IR had higher HMGCS2 expression than NASH patients without IR (NASH samples with IR vs. non-obese

controls yielded a gene expression ratio of 2.23; NASH samples without IR vs. non-obese controls yielded a ratio of 1.99, adjusted P-value o0.025). These results are in agreement with previous data linking IR to NAFLD (31, 32, 55). Another upregulated gene that plays an important role in lipid metabolism is ACSL4/ FACL4, which is involved with lipid metabolism 765

Younossi et al. Table 4. Differentially expressed genes comparing both NASH vs. non-obese controls as well as obese controls vs. non-obese controls

Gene name

Gene name

NASH vs. non-obese controls

Obese controls vs. non-obese controls

Metabolism Nicotinamide N-methyltransferase

NNMT

0.321

0.446

ODC1

0.46

0.464

IGFBP1

0.133

0.254

Modulates the availability of unbound IGF1 for interaction with IGF1R. Most strongly downregulated gene in liver of Cyp1a2  /  mice

ATF3

0.126

0.129

Transcriptional repressor induced by many stress signals. ATF3 represses gluconeogenic enzymes in liver. ATF3 incerased in liver regeneration

0.492

0.349

Activates both extracellular signal-regulated kinase- and protein kinase B-mediated signaling pathways. Serum level is upregulated in liver cirrhosis, fibrosis and alcoholic liver disease

0.454

0.492

0.346

0.346

0.457 0.458

0.452 0.46

TF playing role in cell survival vs. apoptosis and the anti-inflammatory response Promotes the instability of the TNFa and GM-CSF mRNAs. TTF downregulation has strong proinflammatory effect Counterregulator of inflammatory processes Profibrogenic cytokine with anti-inflammatory and immunosuppressive effects. TGFb1(1/  ) mice form vascular lipid lesions which were accompanied by local invasion of macrophages

2.02

2.883

Ornithine decarboxylase 1 Insulin signaling Insulin-like growth factor-binding protein 1

Transcription factors Activating transcription factor 3

Extracellular matrix and associated signaling Chitinase 3-like 1 (cartilage CHI3L1 glycoprotein-39)

Inflammation and suppression of inflammation Nuclear factor, interleukin NFIL-3 3 regulated Zinc-finger protein 36 ZFP36 (tristetraproline) Interleukin 15 receptor, a Transforming growth factor, b1

IL-15RA TGFb1

Cellular defenses against the oxidative stress Catalase CAT

Glutathione S-transferase A4

GSTA4

2.18

2.178

Lactotransferrin

LTF

2.06

2.05

Transferrin

TF

2.07

2.115

ABR

0.382

0.338

PNRC1

0.39

0.456

Other Active BCR-related gene Proline-rich nuclear receptor coactivator 1

Biological function

N-methylation of nicotinamide and structurally related pyridines. NNMT increased in fulminant hepatic failure Polyamine biosynthesis

Antioxidant gene shown to be upregulated in the liver in response to CYP2E1-dependent oxidative stress Cellular defenses against the oxidative stress. Increased in response to CYP2E1-dependent production of mitochondrial ROS, to TNFa, IL-6 and EGF Iron-binding protein with antimicrobial activity. Its expression is upregulated in response to inflammatory stimuli. Hepatic iron promotes oxidative stress Major iron-transporting protein in vertebrates mainly synthesized in the liver. Hepatic iron promotes oxidative stress Multifunctional regulators of the Rho GTPbinding protein family Modulates transcriptional activation of multiple nuclear receptors, including ER, PR, GR, RAR, etc.

Lists all genes (and their differential expression) that were found to be significantly differentially expressed between both NASH and non-obese controls and obese controls and non-obese controls. These genes may be related to obesity but not to NASH itself. NASH, non-alcoholic steatohepatitis.

regulation. Acyl-CoA synthetase activity increases the uptake of FAs by catalyzing their activation to acyl-CoA esters (56). Acyl-CoA is utilized for triglyceride formation in the liver and its high levels inhibit Kreb’s cycle, stimulating ketogenesis. Furthermore, ACSL4/FACL4 con766

trols the level of free arachidonic acid (AA) regulating eicosanoid production (57). Although ACSL4/FACL4 is normally expressed in the liver at a very low level, its over-expression has been reported in human hepatocellular carcinoma (HCC) (58), suggesting its potential role in the

Gene expression in non-alcoholic steatohepatitis Table 5. Differential gene expression comparing NASH and obese controls Gene name Metabolism Ornithine aminotransferase (gyrate atrophy) Acyl-coenzyme A (CoA) dehydrogenase, short/branched chain Transcription factors POU domain, class 2, associating factor 1 Extracellular matrix and associated signaling Matrix metalloproteinase 2 (gelatinase A, 72 kDa gelatinase, 72 kDa type IV collagenase)

Chondroitin sulfate proteoglycan 2 (versican) Angiogenesis and hematopoiesis c-fos-induced growth factor (vascular endothelial growth factor D) Regulator of differentiation (in S. pombe) 1 ADP-ribosylation factor-like 6 interacting protein Other Hypothetical protein dJ462O23.2 Chromosome 12 open-reading frame 14

Gene name

NASH vs. obese controls

Adjusted P-values

Biological function

OAT

0.397

0.002

ACADSB

0.492

0.0002

Mitochondrial matrix enzyme controlling L-ornithine level and producing L-glutamate Mitochondrial enzyme oxidizing straight chain or branched chain acyl-CoAs in the metabolism of fatty acids or branched chain amino acids

POU2AF1

2.672

0.0006

B-cell-specific transcription factor

MMP-2

2.3

0.026

CSPG2

2.037

0.025

Gelatinase secreted by stellate cells and liver myofibroblasts on exposure to the proinflammatory cytokines TGFb1 and IL-6. Pericellular generation of active MMP-2 drives local degradation of normal liver matrix. MMP-2 activated in human fibrotic liver. ECM component shown to play a role in the formation of connective tissue fibers in liver fibrosis (1). Stimulated by TP53

FIGF

2.209

0.0006

Potent angiogenic factor, also in the liver

ROD1

2.048

0.0004

ARL6IP

0.493

0.002

RNA-binding protein that blocks differentiation in hematopoietic cells RAS-like protein involved in hematopoietic maturation possibly by protein transport or membrane trafficking

DJ462O23.2 C12orf14

2.083 2.105

0.002 0.030

Unknown Homolog of mouse TERA gene expressed in teratocarcinomas

Differential expression values, P-values (adjusted for multiple tests using the Benjamini–Hochberg FDR method (49)), and biological function of the differentially expressed genes between NASH patients and obese controls. NASH, non-alcoholic steatohepatitis.

Fig. 1. Real-time RT-PCR confirmation. Real-time RT-PCR (solid bars) was performed on a random selection of 23 samples of the three groups compared in this study. Microarray values (shaded bars) refer only to the selected samples and not to the complete population.

767

Younossi et al. inhibition of apoptosis and the promotion of carcinogenesis. It is noteworthy that differences in expression of ACSL4 were found to be statistically significant in NASH patients compared with those patients with steatosis alone, suggesting its potential role with the progression to NASH. Moreover, an increase in ACSL4/FACL4 may also be responsible for upregulation of the afetoprotein (AFP) gene observed in our analysis. In fact, the ACSL4/FACL4 enzyme exerts a pivotal role in controlling the unesterified AA level in the cells, rapidly converting the intracellular AA into AA-CoA esters, which are responsible for regulation of the transactivating hepatocyte nuclear factor 4 (HNF4) activity. One of the HNF4 responsive genes is the transcription factor encoding gene HNF1, which is directly involved in the AFP gene expression regulation (58). As it is widely known, AFP is an important serum marker for HCC, and is also found, to some extent, in patients with hepatic necrosis (59). Both these lines of evidence suggest that ACSL4/FACL4 gene expression may play a role in NASH-related liver injury and NASHrelated HCC (60). Upregulation of the AFP gene may also be affected by downregulation of the c-jun (v-jun) and jun B AP-1 complex subunits seen in our analysis. These two genes, c-jun and jun B, are also involved in regulating collagenase gene activity, which regulates hepatocyte proliferation by controlling the extracellular matrix. Therefore, the downregulation of AP-1 complex may decrease collagenase expression, further impairing the liver’s normal proliferative ability in patients with NASH (61–65). In addition to the aforementioned genes, our analysis highlights several other differentially expressed genes involved in the matrix assembly, structure, and cell adhesion. For instance, the upregulation of the gene CPSG2 is especially interesting. This gene encodes versican, a member of the group of aggregating proteoglycans. Abnormal accumulation of versican may influence early repair processes, as reported in pulmonary (63) and hepatic (64) fibrosis. Additionally, the activity of the SDC4 gene, encoding transmembrane heparan sulfate ryodokan, was also decreased in NASH patients. This proteoglycan localizes at focal adhesions, potentially pointing to the disorganization of tissue architecture in patients with NASH. In addition to CPSG2 activation and SDC4 downregulation, note that matrix metalloproteinase 2/gelatinase A (MMP2) was also significantly upregulated in NASH patients. Gelatinase A is secreted by hepatic 768

stellate cells (HSC) on exposure to pro-inflammatory cytokines TGFb1 and IL-6. Late phases of liver injury upon HSC activation are characterized by pericellular generation of MMP-2 and local degradation of the normal liver matrix, while degradation of the fibrillar collagens that accumulate in liver fibrosis remains inhibited (65). According to our observations, changes in the extracellular matrix gene expression were especially prominent in the subgroup of NASH patients whose liver biopsies showed significant pericellular fibrosis compared with NASH patients without pericellular fibrosis (Fig. 2). One important aspect of our results relates to the differential gene expression of NASH patients as compared with the obese controls. Although some of these genes may be associated with obesity, their relevance to the pathogenesis of NASH cannot be underestimated (1–20). In this context, our results showed downregulation of IGFbp-1 in patients with NASH. This protein is rapidly and abundantly induced in the regenerating liver. Downregulation of IGFbp-1 may suggest that pathways involved in hepatocyte proliferation as well as prevention of hepatocyte injury and apoptosis may be impaired in these patients (66). However, in patients with steatosis alone, downregulation of IGFbp-1 was even more pronounced. This finding suggests that downregulation of IGFbp-1 negatively affects the regenerative capacity of the liver in patients with NAFLD, but is not a hallmark of NASH.

Fig. 2. Comparisons of NASH fibrosis and non-fibrosis patients for genes CPSG2, MMP-2, and SDC4. Microarray gene expression value of NASH patients without fibrosis (shaded) and NASH patients with fibrosis (solid black). NASH, non-alcoholic steatohepatitis; CPSG2, chondroitin sulfate proteoglycan core protein 2; MMP-2, matrix metalloproteinase 2/gelatinase A; SDC, syndecan 4.

Gene expression in non-alcoholic steatohepatitis ATF3 is another interesting gene that was differentially expressed in NASH patients. This gene is involved in the regulation of phosphoenolpyruvate carboxykinase (PEPCK) (67). In addition to its potential role in controlling p53 with its potential impact on carcinogenesis, ATF3 downregulation may result in important alterations in PEPCK as well as gluconeogenesis (68). Additionally, these gene expression data also show an increase in activity of CAT and glutathione S-transferase A4 (GSTA4) genes. Both genes play a role in the detoxification of lipid peroxidation products and are upregulated in response to the reactive oxygen species (ROS) produced by a number of potential mechanisms including FA b-oxidation (69, 70). In our analysis, CAT and GSTA4 were upregulated in both NASH and obese controls when compared with the non-obese controls. This finding suggests that the altered expression of these two genes in NASH may be a reflection of increase in oxidative stress accompanying obesity and NASH. Similarly, a group of anti-inflammatory genes is found to be under-expressed in the obese subjects. Downregulation of these genes may contribute to the development of inflammatory milieu contributing to the pathogenesis of NASH in obese individuals. In addition to differential gene expression analyses, we performed a number of clustering analyses. Clustering the expression profiles of all significantly differentially expressed genes across NASH patients and non-obese controls clearly separates the diseased and controls into two distinct groups, indicating that expression patterns of NASH patients are different from those of non-obese controls. However, similar clustering did not lead to clear separation of NASH patients from obese controls, suggesting overlapping gene expression profiles involved in the pathogenesis of NASH and its main risk factor, obesity. Our data revealed some differences in the hepatic gene expression profiles reported in a previous study (62). Although these reported differences can be explained by the differences in gene expression methodology and the steps used for its quality control, the difference in the patients’ selection criteria seems to be the most important one. Future studies of hepatic gene expression with larger number of patients from the entire spectrum of NAFLD (from steatosis alone to NASH to NASH-related cirrhosis) should provide additional information and clarify these differences. In conclusion, this in-depth, stepwise approach to the gene expression in liver specimens from

patients with biopsy-proven NASH suggests the involvement of several important genes related to lipid metabolism, b-oxidation, IR, and extracellular matrix signaling. Our findings point to impaired hepatocyte regeneration and apoptosis in NASH, and suggest a theoretical link between NASH and HCC. Further studies are required to assess possible roles of these genes in the pathogenesis of NASH. If verified, the identification of these genes may promote our understanding of the progressive form of NAFLD (or NASH) and provide potential targets for future therapy. Acknowledgments We wish to acknowledge Cooperative Human Tissue Network (CHTN), Mentor, Ohio for providing us with non-obese liver tissue for the use of controls in this study. The study was partially funded though Liver Outcomes Research Fund, Center for Liver Diseases, Inova Health System, Falls Church, VA.

References 1. Matteoni C A, Younossi Z M, Gramlich T, Boparai N, Liu Y C, McCullough A J. Nonalcoholic fatty liver disease: a spectrum of clinical and pathological severity. Gastroenterology 1999; 116: 1413–9. 2. Falck-Ytter Y, Younossi Z M, Marchesini G, McCullough A J. Clinical features and natural history of nonalcoholic steatosis syndromes. Semin Liver Dis 2001; 21: 17–26. 3. Ludwig J, Viggiano T R, McGill D B, Oh B J. Nonalcoholic steatohepatitis: mayo clinic experiences with a hitherto unnamed disease. Mayo Clin Proc 1980; 55: 434–8. 4. Younossi Z M, Diehl A M, Ong J P. Nonalcoholic fatty liver disease: an agenda for clinical research. Hepatology 2002; 35: 746–52. 5. Younossi Z M. Nonalcoholic fatty liver disease. Curr Gastroenterol Rep 1999; 1: 57–62. 6. Younossi Z M, Gramlich T, Liu Y C, Matteoni C, Petrelli M, Goldblum J, Rybicki L, McCullough A J. Nonalcoholic fatty liver disease: assessment of variability in pathologic interpretations. Mod Pathol 1998; 11: 560. 7. Reid A E. Nonalcoholic steatohepatitis. Gastroenterology 2001; 121: 710–23. 8. Burt A D, Mutton A, Day C P. Diagnosis and interpretation of steatosis and steatohepatitis. Sem Diagn Pathol 1998; 15: 246–58. 9. Bacon B R, Farahvash M J, Janney C G, Neuschwander-Tetri B A. Nonalcoholic steatohepatitis: an expanded clinical entity. Gastroenterology 1994; 107: 1103–9. 10. Lee R G. Nonalcoholic steatohepatitis: a study of 49 patients. Hum Pathol 1989; 20: 594–8. 11. Teli M R, James O F, Burt A D, Bennett M K, Day C P. The natural history of nonalcoholic fatty liver: a follow-up study. Hepatology 1995; 22: 1714–9. 12. Day C P, James O F W. Steatohepatitis: a tale of two ‘hits’? Gastroenterology 1998; 114: 842–5. 13. James O F, Day C P. Non-alcoholic steatohepatitis (NASH): a disease of emerging identity and importance. J Hepatol 1998; 29: 495–501. 14. Powell E E, Cooksley W G, Hanson R, Searle J, Halliday J W, Powell L W. The natural history of nonalcoholic steatohepatitis: a follow-up study of forty-two patients for up to 21 years. Hepatology 1990; 11: 74–80.

769

Younossi et al. 15. Angulo P, Keach J C, Batts K P, Lindor K D. Independent predictors of liver fibrosis in patients with nonalcoholic steatohepatitis. Hepatology 1999; 30: 1356–62. 16. Caldwell S H, Oelsner D H, Iezzoni J C, Hespenheide E E, Battle E H, Driscoll C J. Cryptogenic cirrhosis: clinical characterization and risk factors for underlying disease. Hepatology 1999; 29: 664–9. 17. Ong J, Younossi Z M, Reddy V, Price L L, Gramlich T, Mayes J, Boparai N. Cryptogenic cirrhosis and posttransplantation nonalcoholic fatty liver disease. Liver Transpl 2001; 7: 797–801. 18. Bugianesi E, Leone N, Vanni E, Marchesini G, Brunello F, Carucci P, Musso A, De Paolis P, Capussotti L, Salizzoni M, Rizzetto M. Expanding the natural history of nonalcoholic steatohepatitis: from cryptogenic cirrhosis to hepatocellular carcinoma. Gastroenterology 2002; 123: 134–40. 19. Hay J E, Czaja A J, Rakela J, Ludwig J. The nature of unexplained chronic aminotransferase elevations of a mild to moderate degree in asymptomatic patients. Hepatology 1989; 9: 193–7. 20. Adler M, Schaffner F. Fatty liver hepatitis and cirrhosis in obese patients. Am J Med 1979; 67: 811–6. 21. Nomura H, Kashiwagi S, Hayashi J, Kajiyama W, Tani S, Goto M. Prevalence of fatty liver in a general population of Okinawa, Japan. Jpn J Med 1988; 27: 143–9. 22. Hulcrantz R, Glaumann H, Lindberg G, Nilsen L H. Liver investigation in 149 asymptomatic patients with moderately elevated activities of serum aminotransferases. Scand J Gastroenterol 1986; 21: 109–13. 23. Araujo L M B, DeOliveira A, Nunes D S. Liver and biliary ultrasonography in diabetic and non-diabetic obese women. Diabetes Metabolism 1998; 24: 458–62. 24. El-Hassan A Y, Ibrahim E M, Al-Mulhim F A, Nabhan A A, Chammas M Y. Fatty infiltration of the liver: analysis of prevalence, radiological and clinical features and influence on patient management. Br J Radiol 1992; 65: 774–8. 25. Nonomura A, Mizukami Y, Unoura M, Kobayashi K, Takeda Y, Takeda R. Clinicopathologic study of alcohollike liver disease in non-alcoholics; non-alcoholic steatohepatitis and fibrosis. Gastroenterol Jpn 1992; 27: 521–8. 26. Daniel S, Ben-Menachem T, Vasudevan G, Ma C K, Blumenkehl M. Prospective evaluation of unexplained chronic liver transaminase abnormalities in asymptomatic and symptomatic patients. Am J Gastroenterol 1999; 94: 3010–4. 27. Ratziu V, Giral P, Charlotte F, Bruckert E, Thibault V, Theodorou I, Khalil L, Turpin G, Opolon P, Poynard T. Liver fibrosis in overweight patients. Gastroenterology 2000; 118: 1117–23. 28. Lonardo A, Bellino M, Tartoni P, Tondelli E. The bright liver syndrome: prevalence and determinants of a ‘bright’ liver echo pattern. Ital J Gastroenterol Hepatol 1997; 29: 351–6. 29. Brolin R E, Bradley L J, Taliwal R V. Unsuspected cirrhosis discovered during obesity operations. Arch Surg 1998; 133: 84–8. 30. Wanless I R, Lentz J S. Fatty liver hepatitis (steatohepatitis) and obesity: an autopsy study with analysis of risk factors. Hepatology 1990; 12: 1106–10. 31. Younossi Z, Gramlich T, Matteoni C, Boparai N, McCullough A. Non-alcoholic fatty liver disease in patients with type II. Diabetes Clin Gastro Hepatol 2004; 2(3): 262–5. 32. Chitturi S, Farrell G C. Etiopathogenesis of nonalcoholic steatohepatitis. Semin Liver Dis 2001; 21: 27–41.

770

33. Weltman M D, Farrell G C, Hall P, Ingelman-Sundberg M, Liddle C. Hepatic cytochrome P450 2E1 is increased in patients with nonalcoholic steatohepatitis. Hepatology 1998; 27: 128–33. 34. Tilg H, Diehl A M. Cytokines in alcoholic and nonalcoholic steatohepatitis. N Engl J Med 2000; 343: 1467–76. 35. Raucy J L, Lasker J M, Kraner J C, Salazar D E, Lieber C S, Corcoran G B. Induction of cytochrome P450IIE1 in the obese overfed rat. Mol Pharmacol 1991; 39: 275–80. 36. Zhang C, Walker L M, Hinson J A, Mayeux P R. Oxidant stress in rat liver after lipopolysaccharide administration: effect of inducible nitric-oxide synthase inhibition. J Pharmacol Exp Ther 2000; 293: 968–72. 37. Leclercq I A, Farrell G C, Field J, Bell D R, Gonzalez F J, Robertson G R. CYP2E1 and CYP4A as microsomal catalysts of lipid peroxides in murine nonalcoholic steatohepatitis. J Clin Invest 2000; 105: 1067–75. 38. Weltman M D, Farrell G C, Liddle C. Increased hepatocyte CYP2E1 expression in a rat nutritional model of hepatic steatosis with inflammation. Gastroenterology 1996; 111: 1645–53. 39. Fong D G, Nehra V, Lindor K D, Buchman A L. Metabolic and nutritional considerations in nonalcoholic fatty liver. Hepatology 2000; 32: 3–10. 40. Rashid A, Wu T C, Huang C C, Chen C H, Lin H Z, Yang S Q, Lee F Y, Diehl A M. Mitochondrial proteins that regulate apoptosis and necrosis are induced in mouse fatty liver. Hepatology 1999; 29: 1131–8. 41. Gramlich T, Kleiner D, McCullough A, Matteoni C, Boparai N, Younossi Z M. Pathologic features associated with fibrosis in non-alcoholic fatty liver disease. Human Pathol 2004; 35(2): 196–9. 42. George D K, Goldwurm S, MacDonald G A, Cowley L L, Walker N I, Ward P J, Jazwinska E C, Powell L W. Increased hepatic iron concentration in nonalcoholic steatohepatitis is associated with increased fibrosis. Gastroenterology 1998; 114: 311–8. 43. Younossi Z M, Gramlich T, Bacon B R, Matteoni C A, Boparai N, O’Neill R, McCullough A J. Hepatic iron and nonalcoholic fatty liver disease. Hepatology 1999; 30: 847–50. 44. Ong J P, Younossi Z M, Speer C, Olano A, Gramlich T, Boparai N. Chronic hepatitis C and superimposed nonalcoholic fatty liver disease. Liver 2001; 21: 266–71. 45. Guttmacher A E, Collins F S. Genomic medicine – a primer. N Engl J Med 2002; 347: 1512–20. 46. Honda M, Kaneko S, Kawai H, Shirota Y, Kobayashi K. Differential gene expression between chronic hepatitis B and C hepatic lesion. Gastroenterology 2001; 120: 955–66. 47. Collins F S, Guttmacher A E. Genetics moves into the medical mainstream. JAMA 2001; 286: 2322–4. 48. Dessein P H, Joffe B I, Stanwix A E, Botha A S, Moomal Z. The acute phase response does not fully predict the presence of insulin resistance and dyslipidemia in inflammatory arthritis. J Rheumatol 2002; 29: 462–6. 49. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B 1995; 57: 289–300. 50. Epstein C B, Hale IV W, Butow R. Numerical methods for handling uncertainty in microarray data: an example analyzing perturbed mitochondrial function in yeast. Methods Cell Biol 2001; 65: 439–52. 51. Ward J H. Hierarchical grouping to optimize an objective function. J Am Stat Assoc 1963; 58: 236–44. 52. Eisen M B, Spellman P T, Brown P O, Botstein D. Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA 1998; 95: 14863–8.

Gene expression in non-alcoholic steatohepatitis 53. Grace M B, McLeland C B, Gagliardi S J, Smith J M, Jackson W E I, Blakely W F. Quantitative RT-PCR assay using simultaneous measurement of four fluorosphores for radiation-responsive gene targets. Clin Chem 2003; 49: 1467–75. 54. Fausto G. HEGARDT. Mitochondrial 3-hydroxy-3-methylglutaryl-CoA synthase: a control enzyme in ketogenesis. Biochem J 1999; 338: 569–82 (printed in Great Britain) 569. 55. Sanyal A J, Campbell-Sargent C, Mirshahi F, et al. Nonalcoholic steatohepatitis: association of insulin resistance and mitochondrial abnormalities. Gastroenterology 2001; 120(5): 1183–92. 56. Gargiulo C E, Stuhlsatz-Krouper S M, Schaffter J E. Localization of adipocyte long-chain fatty acyl-CoA synthetase at the plasma membrane. J Lipid Res 1999; 40(5): 881–92. 57. Hertz R, Magenheim J, Berman I, Bar-Tana J. Fatty acyl-CoA thioesters are ligands of hepatic nuclear factor-4a. Nature 1998; 392: 2. 58. Magee T R, Cai Y, El-Houseini M E, Locker J. J Biol Chem 1998; 273(45): 30024–32. 59. Lazarevich N L. Molecular mechanisms of alpha-fetoprotein gene expression. Biochemistry (Mosc) 2000; 65(1): 117–33. 60. Bugianesi E, Leone N, Vanni E, et al. Expanding the natural history of nonalcoholic steatohepatitis: from cryptogenic cirrhosis to hepatocellular carcinoma. Gastroenterology 2002; 123(1): 134–40. 61. Dong J M, Nordloh P W, Chiu J F. The mechanism of the bidirectional regulation of the rat alpha-fetoprotein gene by glucocorticoid hormone. Mol Cell Endocrinol 1989; 66(1): 109–14. 62. Sreekumar R, Rosado B, Rasmussen D, Charlton M. Hepatic gene expression in histologically progressive nonalcoholic steatohepatitis. Hepatology 2003; 38: 244–51.

63. Bensadoun E S, Burke A K, Hogg J C, Roberts C R. Proteoglycan deposition in pulmonary fibrosis. Am J Respir Crit Care Med 1996 Dec; 154(6 Part 1): 1819–28. 64. Szende B, Lapis K, Kovalszky I, Timar F. Role of the modified (glycosaminoglycan producing) perisinusoidal fibroblasts in the CCl4-induced fibrosis of the rat liver. In Vivo 1992; 6(4): 355–61. 65. Benyon R C, Arthur M J. Extracellular matrix degradation and the role of hepatic stellate cells. Semin Liver Dis 2001; 21(3): 373–84. 66. Leu J I, Crissey M A, Taub R. Massive hepatic apoptosis associated with TGF-beta1 activation after Fas ligand treatment of IGF binding protein-1-deficient mice. J Clin Invest 2003; 111: 129–39. 67. Hai T, Wolfgang C D, Marsee D K, Allen A E, Sivaprasad U. ATF3 and stress responses. Gene Exp 1999; 7: 321–35. 68. Yang S Q, Lin H Z, Mandal A K, Huang J, Diehl A M. Disrupted signaling and inhibited regeneration in obese mice with fatty livers: implications for nonalcoholic fatty liver disease pathophysiology. Hepatology 2001; 34: 694– 706. 69. Irshad M, Chaudhuri P S. Oxidant–antioxidant system: role and significance in human body. Indian J Exp Biol 2002; 40(11): 1233–9. 70. Mari M, Cederbaum A I. Induction of catalase, alpha, and microsomal glutathione S-transferase in CYP2E1 overexpressing HepG2 cells and protection against short-term oxidative stress. Hepatology 2001; 33(3): 652–61. 71. Hurteau G J, Spivack S D. mRNA-specific reverse transcription-polymerase chain reaction from human tissue extracts. Anal Biochem 2002; 307: 304–15.

771