Gene Expression Profiling Reveals the Mechanism and Pathophysiology of Mouse Liver Regeneration

Comprehensive analysis of the changes in gene expression during liver regeneration was carried out by using an in-house microarray composed of 2,304 distinct mouse liver cDNA clones. Mice were subjected to partial two-thirds hepatectomy, and changes in mRNA levels were monitored up to 48 h. Of the 2,304 genes analyzed, 496 genes showed expression levels measurable at all time points after the partial hepatectomy. 317 genes were up- or down-regulated 2-fold or more at least at one time point during liver regeneration and were classified into eight clusters based on their expression patterns. With a more stringent cut-off value of ±2 S.D., 68 genes were listed and were classified into five clusters. In these two analyses with different clustering criteria, functionally categorized genes showed similar cluster distributions. Genes involved in protein synthesis and posttranslational processing were significantly enriched in the cluster characterized by rapid gene activation and subsequent persistence. This suggests the importance of modulating the efficiency of protein supply and/or altering the composition of protein population from the early phase of hepatocyte proliferation. Genes for two major liver functions, i.e. plasma protein secretion and intermediate metabolism were enriched in distinct clusters exhibiting the features of gradual gene activation and sustained repression, respectively. Therefore, these genes are differentially regulated during the regeneration, possibly leading to changes in the flow of amino acids and energy from enzyme proteins to plasma proteins in their synthesis. Thus, clustering analysis of expression patterns of functionally classified genes gave insights into mechanism and pathophysiology of liver regeneration. Comprehensive analysis of the changes in gene expression during liver regeneration was carried out by using an in-house microarray composed of 2,304 distinct mouse liver cDNA clones. Mice were subjected to partial two-thirds hepatectomy, and changes in mRNA levels were monitored up to 48 h. Of the 2,304 genes analyzed, 496 genes showed expression levels measurable at all time points after the partial hepatectomy. 317 genes were up- or down-regulated 2-fold or more at least at one time point during liver regeneration and were classified into eight clusters based on their expression patterns. With a more stringent cut-off value of ±2 S.D., 68 genes were listed and were classified into five clusters. In these two analyses with different clustering criteria, functionally categorized genes showed similar cluster distributions. Genes involved in protein synthesis and posttranslational processing were significantly enriched in the cluster characterized by rapid gene activation and subsequent persistence. This suggests the importance of modulating the efficiency of protein supply and/or altering the composition of protein population from the early phase of hepatocyte proliferation. Genes for two major liver functions, i.e. plasma protein secretion and intermediate metabolism were enriched in distinct clusters exhibiting the features of gradual gene activation and sustained repression, respectively. Therefore, these genes are differentially regulated during the regeneration, possibly leading to changes in the flow of amino acids and energy from enzyme proteins to plasma proteins in their synthesis. Thus, clustering analysis of expression patterns of functionally classified genes gave insights into mechanism and pathophysiology of liver regeneration. Liver is unique in the ability to regenerate rapidly even in adulthood. A number of studies have been done to reveal the genes responsible for liver regeneration. Cytokines such as interleukin-6 and tumor necrosis factor α, hormones and growth factors including insulin, norepinephrine, hepatocyte growth factor, and epidermal growth factor, and a number of transcription factors have been shown to be involved in liver regeneration (1Michalopoulos G.K. DeFrances M.C. Science. 1997; 276: 60-66Crossref PubMed Scopus (2917) Google Scholar, 2Yamada Y. Kirillova I. Peschon J.J. Fausto N. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 1441-1446Crossref PubMed Scopus (842) Google Scholar, 3Fausto N. J. Hepatol. 2000; 32: 19-31Abstract Full Text PDF PubMed Google Scholar). While administration of these proteins and/or overexpression of their genes induced the rapid regeneration, disruption of the genes in mice resulted in severe impairment of the regeneration (1Michalopoulos G.K. DeFrances M.C. Science. 1997; 276: 60-66Crossref PubMed Scopus (2917) Google Scholar, 2Yamada Y. Kirillova I. Peschon J.J. Fausto N. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 1441-1446Crossref PubMed Scopus (842) Google Scholar, 3Fausto N. J. Hepatol. 2000; 32: 19-31Abstract Full Text PDF PubMed Google Scholar). Despite the clarification of the importance of these genes for liver regeneration, the complex mechanism for the regeneration by the interplay of many factors remains to be investigated. The expression patterns of many genes associated with liver regeneration were discussed previously (4Taub R. FASEB J. 1996; 10: 413-427Crossref PubMed Scopus (322) Google Scholar). However, a simultaneous and comprehensive analysis of the expression profiles of these genes has been difficult because of technical limitation. Recently, DNA microarray technology has been developed and shown to be a powerful tool for analyzing the expression profiles of many genes at one time (5DeRisi J.L. Iyer V.R. Brown P.O. Science. 1997; 278: 680-686Crossref PubMed Scopus (3706) Google Scholar). By using this technique, Su et al. (6Su A.I. Guidotti L.G. Pezacki J.P. Chisari F.V. Schultz P.G. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 11181-11186Crossref PubMed Scopus (170) Google Scholar) revealed gene expression profiles during the priming phase of liver regeneration up to 4 h after partial hepatectomy (PHx), 1The abbreviations used are: PHx, partial hepatectomy; Hsp, heat-shock protein; HNF, hepatocyte nuclear factor; HFH, HNF-3/fork head homolog. and demonstrated the changes in expression of 185 genes. After the priming phase, hepatocytes proliferation starts at the periportal area of the lobular architecture and then proceeds to the perivenular areas (1Michalopoulos G.K. DeFrances M.C. Science. 1997; 276: 60-66Crossref PubMed Scopus (2917) Google Scholar, 7Assy N. Minuk G.Y. J. Hepatol. 1997; 26: 945-952Abstract Full Text PDF PubMed Scopus (105) Google Scholar). Since the peaks of DNA synthesis in hepatocytes after PHx are around 24–40 h (1Michalopoulos G.K. DeFrances M.C. Science. 1997; 276: 60-66Crossref PubMed Scopus (2917) Google Scholar, 3Fausto N. J. Hepatol. 2000; 32: 19-31Abstract Full Text PDF PubMed Google Scholar), here we analyzed the changes in gene expression up to 48 h by using an in-house microarray harboring mouse liver cDNAs. 317 genes were shown to be up- or down-regulated during the course following PHx, and their clustering analysis revealed striking features of expression profiles of functionally classified genes, giving insights into roles and regulation of genes in liver regeneration. Mice and mRNA Preparation—Male C57BL6 mice aged 5–7 weeks were obtained from CLEA Japan, Inc. (Tokyo, Japan) and were subjected to partial two-thirds PHx under ether anesthesia as described (8Higgins G.M. Andersen R.M. Arch. Pathol. 1931; 12: 186-202Google Scholar). Two mice were sacrificed at each time point (2, 6, 12, 24, and 48 h after PHx), and livers were resected. Liver sections removed by original hepatectomy operation of respective mice were used as a control. Total RNA was prepared from livers using TRIZOL reagent (Invitrogen), and subjected to isolation of poly(A)+ RNA using the Oligotex-dT30 mRNA purification kit (TaKaRa Shuzo Co., Kyoto, Japan) according to the manufacturer's instructions. Preparation of the cDNA Microarray—A cDNA microarray chip, consisting of 2,304 cDNA (1,504 known genes and 800 unknown genes) was made as described previously (5DeRisi J.L. Iyer V.R. Brown P.O. Science. 1997; 278: 680-686Crossref PubMed Scopus (3706) Google Scholar, 9Yoshikawa T. Nagasugi Y. Azuma T. Kato M. Sugano S. Hashimoto K. Masuho Y. Muramatsu M. Seki N. Biochem. Biophys. Res. Commun. 2000; 275: 532-537Crossref PubMed Scopus (62) Google Scholar, 10Schena M. Shalon D. Davis R.W. Brown P.O. Science. 1995; 270: 467-470Crossref PubMed Scopus (7668) Google Scholar). 2,304 unique clones were selected from about 9,000 sequenced clones in an oligo-capped cDNA library (11Suzuki Y. Yoshitomo-Nakagawa K. Maruyama K. Suyama A. Sugano S. Gene (Amst.). 1997; 200: 149-156Crossref PubMed Scopus (233) Google Scholar) of the mouse liver. PCR-amplified cDNA products were mixed with nitrocellulose in dimethyl sulfoxide and then spotted onto carbodiimide-coated glass slides using robotics SPBIO-2000 (Hitachi Software Engineering Co., Yokohama, Japan). Microarray Analysis—Fluorescent cDNA probes (Cy3- or Cy5-labeled) were prepared from 2 μg of poly(A)+ RNA. Hybridization and fluorescence detection were performed essentially as described previously (9Yoshikawa T. Nagasugi Y. Azuma T. Kato M. Sugano S. Hashimoto K. Masuho Y. Muramatsu M. Seki N. Biochem. Biophys. Res. Commun. 2000; 275: 532-537Crossref PubMed Scopus (62) Google Scholar). Images were analyzed with Quant Array (GSI Lumonics, Nepean, Canada) and DNASIS Array (Hitachi Software Engineering) according to the manufacturer's instructions. The intensities of four areas between diagonal spots were used as the background for each spot. The mean and S.D. of background levels were calculated, and the genes whose intensities were less than mean plus 2 S.D. of background levels at any of the time points were excluded from further analysis. The Cy5/Cy3 ratios of all spots on the microarray were normalized by dividing them by the median. Averaged data of the two animals were subjected to statistical analysis. Northern Blot Analysis—20 μg of total RNA were electrophoresed through a 1% agarose-formaldehyde gel, transferred to a nitrocellulose membrane (Amersham Biosciences) overnight and cross-linked with irradiation. Probes were generated using Megaprime DNA labeling system (Amersham Biosciences) and [α-32P]dCTP. Probes and blots were hybridized in Rapid-Hyb buffer (Amersham Biosciences). Experiments were performed twice and quantified using BAS 2000 (Fuji Photo Film Co., Tokyo, Japan). Statistical Analysis and Annotation of Gene Function—The normalized Cy5/Cy3 ratios in the microarray analysis were log2-transrformed and used to classify the patterns of serial changes of the gene expression. Genes whose expression levels varied at least 2-fold or by 2 S.D. at any of the time points were subjected to hierarchical clustering analysis, using the algorithm of Euclid and Ward in GeneMaths software (Applied Maths BVBA, Sint-Martens-Latem, Belgium). The molecular functions of the genes were assigned referring to GENE ONTOLOGY™ (www.geneontology.org/) and GeneCards™ (bioinfo.weizmann.ac.il/cards/). To examine statistical significance for frequencies of genes of each functional group in each cluster, values of the other groups in the relevant cluster, and values of the other clusters in the relevant group, were each combined, and the resultant combined values were compared with the relevant value with Fisher's exact test by computing for the 2 × 2 table with the statistical program StatView (SAS Institute Inc., Cary, NC). Application of the cDNA Microarray Analysis for Detection of Changes in Gene Expression during Liver Regeneration—To examine the sequential changes in gene expression during liver regeneration comprehensively, we performed cDNA microarray analysis. We selected the PHx to cause the regeneration, because the initiation time of the regeneration is very clear in this model and because effects on gene expression by tissue injury or inflammation observed in other models using chemical compounds such as carbon tetrachloride or acetyl aminofluorene are minimal in the remnant intact liver in the PHx model. Poly(A)+ RNAs derived from the liver at 2, 6, 12, 24, or 48 h after PHx and the control liver in duplicate were subjected to Cy3 and Cy5 labeling, respectively, coupled with cDNA synthesis. Both cDNAs were mixed in an equal amount, and hybridized with a microarray. We used an in-house microarray (9Yoshikawa T. Nagasugi Y. Azuma T. Kato M. Sugano S. Hashimoto K. Masuho Y. Muramatsu M. Seki N. Biochem. Biophys. Res. Commun. 2000; 275: 532-537Crossref PubMed Scopus (62) Google Scholar) harboring 2,304 mouse liver cDNA clones, facilitating efficient detection of changes in gene expression in the target organ. Out of the 2,304 genes analyzed, we regarded expression of 496 genes as meaningful, because their expression levels were higher than background levels in all microarray experiments. 317 genes were up- or down-regulated 2-fold or more at least at one time point during the regeneration. To determine the validity of results obtained by the microarray analysis, 10 randomly selected genes were subjected to Northern blot analysis. While mRNAs for two genes (glutathione peroxidase and itih-4) were under the detectable level in Northern analysis (data not shown), mRNA levels of the other eight genes (S-adenosylmethionine synthetase, claudin-1, squalene synthase, virus-like retro-element, leptin receptor, insulin-like growth factor-binding protein 1, haptoglobin, and α-1 acid glycoprotein 1B) were almost comparable between microarray and Northern analysis (Fig. 1), generally verifying the competency of the microarray analysis for detection of changes in gene expression after PHx. However, some discrepancy was present between the results of microarray and Northern experiments. Following analyses on microarray data were done with gene numbers large enough for statistical treatment. Serial Changes in Expression Levels of the Genes in Relation to Their Functions—The 317 genes with the changes in their expression levels after PHx were subjected to clustering analysis based on expression patterns and were classified into eight clusters (Table SI (see Supplemental Material) and Fig. 2A). Time course of changes in mRNA levels in each cluster can be roughly characterized as follows (Fig. 2B): cluster A, rapid increase and persistence; cluster B, transient up-regulation and gradual decrease; cluster C, gradual increase; cluster D, transient down-regulation and rebound; cluster E, rapid increase and rapid return; cluster F, transient down-regulation and gradual increase; cluster G, general decrease; cluster H, transient down-regulation and persistence. We also categorized the 317 genes based on their functions, referring to GENE ONTOLOGY™ and GeneCards™, and classified them into 11 groups (Table SI in Supplemental Material and Fig. 3): cell growth and maintenance, cytoskeleton and cell membrane, metabolism, secreted proteins, signal transduction, transcription and processing, translation and processing, mitochondrion, others, and expressed sequence tags (EST). Frequencies of these functionally classified genes in each cluster are shown in Fig. 3, with statistical significance evaluated by Fisher's exact test for high and low frequencies. Hitherto, we have used the 2-/0.5-fold changes in expression levels as cut-off values for seemingly reliable changes, which were exhibited by 17, 29, 23, 23, and 15% (amounting to 317 genes in total) of intensity-measurable 496 genes at the time points of 2, 6, 12, 24, and 48 h, respectively. The 2-/0.5-fold changes corresponded with 1.41/-1.12, 1.51/-0.75, 1.24/-0.97, 1.12/-1.34, and 1.21/-1.32 S.D. at each time point. We also tested a more stringent cut-off value of ±2 S.D., to verify the cluster-function relationship (Fig. 4). The numbers of genes that exceeded this cut-off value occupied 3.8, 4.8, 5.6, 3.2, and 5.0% (amounting to 68 genes in total) of the 496 genes at each time point. The 68 genes were classified into five clusters (a–e) by hierarchical clustering analysis. Their features are as follows: cluster a, rapid increase and rapid return; cluster b, rapid increase and persistence; cluster c, transient up-regulation and gradual decrease; cluster d, gradual increase; cluster e, general decrease. Again, frequencies of functionally categorized genes in these clusters are represented with their statistical significance (Fig. 4, right panels). Together with the results of Fig. 3, high frequencies were reproducibly observed in following gene distributions: genes for the category “translation and processing” in the cluster “rapid increase and persistence” (clusters A and b in Figs. 3 and 4, respectively); “secreted proteins” in “gradual increase” (clusters C and F and d); “metabolism” in “general decrease” (clusters G and e). We will discuss these characteristic gene distributions below, mainly based on the gene cluster classification in Table SI in Supplemental Material and Figs. 2 and 3. Activation of Genes in the Category “Translation and Processing” in the Early Stage of Liver Regeneration—Among genes in the cluster A with properties of rapid activation and following persistence, significantly frequent were members of the category “translation and processing” (Fig. 3), which are a ribosomal protein (number 19 in Table SI in Supplemental Material), elongation factors (numbers 16 and 17 in Table SI in Supplemental Material) and components of protein transport machineries (numbers 1 and 24 in Table SI in Supplemental Material). This suggests that machinery for protein synthesis, folding, and transport were prepared at the early stage of liver regeneration, presumably to synthesize and deliver proteins of altered population with modified efficiencies. Importance of adequate protein synthesis and processing for cells to pass the restriction point and enter the S phase has been repeatedly noted. When the cellular ribosome content decreases, the translation of CLN3 mRNA is inhibited, presumably leading to growth arrest (12Polymenis M. Schmidt E.V. Curr. Opin. Genet. Dev. 1999; PubMed Scopus Google Scholar). of translation factors in growth regulation has been by the that overexpression of the mRNA protein cell growth and of and cells Y. A. J. 1999; Full Text Full Text PDF PubMed Scopus Google Scholar). Importance of and posttranslational protein control in cell has been also of proteins in translation and results in cell arrest Proc. Natl. Acad. Sci. U. S. A. 1999; PubMed Scopus Google Scholar). After PHx, of the hepatocytes in the liver in one or two and the of DNA synthesis is about at 24–40 h (1Michalopoulos G.K. DeFrances M.C. Science. 1997; 276: 60-66Crossref PubMed Scopus (2917) Google Scholar, 3Fausto N. J. Hepatol. 2000; 32: 19-31Abstract Full Text PDF PubMed Google Scholar). of hepatocytes have to pass the restriction point and enter the S phase this by early activation of genes for of Genes for and for two major functions, i.e. metabolism and plasma protein secretion exhibited cluster The category “metabolism” is mainly composed of involved in metabolism of compounds such as amino and Genes of this category were significantly enriched in the cluster G, which exhibited general features of gene a of mRNAs were under the control level even at 24 and 48 h after PHx (Fig. in the liver have to a two-thirds of the liver by PHx, presumably also by of of the enzyme proteins during the regeneration. In the of protein in the liver R.W. J. Full Text PDF PubMed Google Scholar). the other the category “secreted proteins” plasma proteins as a large and a number of proteins involved in liver Genes of this category were enriched in the clusters C and F, which are characterized by gradual in mRNA levels in the early of mRNAs were the control level at 24 and 48 h following PHx (Fig. This at least two in regulation of plasma protein is of phase proteins such as (numbers and in Table SI in Supplemental Material), α-1 acid glycoprotein (number in Table SI in Supplemental Material), protein (number in Table SI in Supplemental Material) and components (number in Table SI in Supplemental Material), in to injury following PHx. The other is of plasma proteins that are under the two-thirds PHx to of plasma protein synthesis on the remnant liver. Thus, of and plasma proteins to be regulated in the acids and energy by synthesis of enzyme proteins are to be plasma protein synthesis during liver regeneration. Genes with Expression Levels during Liver (number in Table SI in Supplemental Material) in the cluster C is an membrane protein of M. K. T. K. S. J. PubMed Scopus Google Scholar). a factor was to be during liver regeneration Y. S. N. Y. T. M. A. K. K. S. T. S. Res. PubMed Scopus Google Scholar). of during cell is very protein (number in Table SI in Supplemental also in the cluster C is a of the and to the tumor necrosis factor D. 1995; PubMed Scopus (62) Google Scholar), possibly regulation of the by the Y. M. Kato S. M. M. A. N. K. J. Biochem. 1996; PubMed Scopus Google Scholar). protein (number in Table SI in Supplemental also and in the cluster C was previously shown to be induced during liver regeneration K. Res. Google Scholar), and was to also in tumor necrosis activation of 2002; Full Text Full Text PDF PubMed Scopus Google Scholar). nuclear factor (number in Table SI in Supplemental also was the four members of the cluster E, which exhibited a striking expression of rapid transient and are members of the head transcription factor and a number of genes V.R. Genes Dev. PubMed Scopus Google Scholar, V.R. Genes Dev. PubMed Scopus Google Scholar). HNF-3/fork head was shown to be during the regeneration U. S. 1997; PubMed Scopus Google Scholar). In the of hepatocyte DNA and during liver regeneration was 1999; PubMed Scopus Google Scholar). Thus, the head transcription factor including to roles in liver regeneration. (number in Table SI in Supplemental Material) in the cluster F was as a protein induced enter a in growth regulation S. J. G.K. Arch. Biochem. Biophys. 2002; PubMed Scopus Google Scholar). Recently, this protein was shown to be a that of in proteins A. S. Y. J. 276: Full Text Full Text PDF PubMed Scopus (62) Google Scholar). was in the and as as the cell and to be involved in of proteins including secreted S. J. G.K. Arch. Biochem. Biophys. 2002; PubMed Scopus Google Scholar). in the cluster F that a number of secreted proteins, the of the The that mRNA level of is at 48 h after PHx (Table SI in Supplemental Material) is also with the that this enzyme a in growth arrest of the and remains to be investigated. In the present comprehensive with cDNA microarray analysis and following statistical analysis revealed features of changes in gene expression during liver regeneration after PHx. is that early activation of genes for protein synthesis and processing is for hepatocytes to enter the S activation of plasma protein genes and of enzyme genes changes in the flow of amino acids and energy in protein synthesis. A number of genes were also results on mechanism and pathophysiology of the liver regeneration. We M. K. T. M. and for and with

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.