Identification of Transcriptional Networks during Liver Regeneration

The molecular analysis of mammalian cellular proliferation in vivo is limited in most organ systems by the low turnover and/or the asynchronous nature of cell cycle progression. A notable exception is the partial hepatectomy model, in which quiescent hepatocytes reenter the cell cycle and progress in a synchronous fashion. Here we have exploited this model to identify regulatory networks operative in the mammalian cell cycle. We performed microarray-based expression profiling on livers 0-40 h post-hepatectomy corresponding to G0, G1, and S phases. Differentially expressed genes were identified using the statistical analysis program PaGE (Patterns from Gene Expression), which was highly accurate as confirmed by quantitative reverse transcription-PCR of randomly selected targets. A shift in the transcriptional program from genes involved in lipid and hormone biosynthesis in the quiescent liver to those contributing to cytoskeleton assembly and DNA synthesis in the proliferating liver was demonstrated by biological theme analysis. In a novel approach, we employed computational pathway analysis tools to identify specific regulatory networks operative at various stages of the cell cycle. This allowed us to identify a large cluster of genes controlling mitotic spindle assembly and checkpoint control at the 40-h time point as regulated at the mRNA level in vivo. The molecular analysis of mammalian cellular proliferation in vivo is limited in most organ systems by the low turnover and/or the asynchronous nature of cell cycle progression. A notable exception is the partial hepatectomy model, in which quiescent hepatocytes reenter the cell cycle and progress in a synchronous fashion. Here we have exploited this model to identify regulatory networks operative in the mammalian cell cycle. We performed microarray-based expression profiling on livers 0-40 h post-hepatectomy corresponding to G0, G1, and S phases. Differentially expressed genes were identified using the statistical analysis program PaGE (Patterns from Gene Expression), which was highly accurate as confirmed by quantitative reverse transcription-PCR of randomly selected targets. A shift in the transcriptional program from genes involved in lipid and hormone biosynthesis in the quiescent liver to those contributing to cytoskeleton assembly and DNA synthesis in the proliferating liver was demonstrated by biological theme analysis. In a novel approach, we employed computational pathway analysis tools to identify specific regulatory networks operative at various stages of the cell cycle. This allowed us to identify a large cluster of genes controlling mitotic spindle assembly and checkpoint control at the 40-h time point as regulated at the mRNA level in vivo. The liver is one of few organs that retains the capacity to rapidly respond to changes in mass and/or function in both humans and animals. This property has significant implications for a variety of clinical situations, including surgical removal of a portion of the liver, such as that which occurs following tumor resection or living-related liver transplantation and for recovery from fulminant liver failure. In contrast to the synchronous and extensive loss of liver mass that occurs in surgical contexts and acute liver failure, the majority of human liver diseases are characterized by repetitive inflammatory or toxic insults to the liver that are associated with loss of liver cells as the result of necrotic or apoptotic cell death. As a consequence of repetitive injury, proliferative signals to hepatocytes are associated with increased risk of hepatocellular carcinoma. However, the mechanisms that regulate both normal and pathologic proliferation are still poorly understood. A widely used experimental model of hepatic growth is the partial hepatectomy model in rodents in which ∼70% of the liver is removed and restoration of liver mass occurs within 10-14 days (1Higgins G.M. Anderson R.M. Arch. Pathol. 1931; 12: 186-202Google Scholar). Within 30 min of the surgery, cytokine and growth factor stimulation leads to the activation of preexisting transcription factors, including NFκB, Stat3, AP-1, and C/EBPβ, and induction of a large number of genes responsible for stimulating normally quiescent hepatocytes and nonparenchymal liver cells to reenter the cell cycle and ultimately restore the liver mass (2Fausto N. J. Hepatol. 2000; 32: 19-31Abstract Full Text PDF PubMed Google Scholar, 3Michalopoulos G.K. DeFrances M.C. Science. 1997; 66: 60-66Crossref Scopus (2918) Google Scholar, 4Koniaris L.G. McKillop I.H. Schwartz S.I. Zimmers T.A. J. Am. Coll. Surg. 2003; 197: 634-659Abstract Full Text Full Text PDF PubMed Scopus (231) Google Scholar, 5Taub R. Greenbaum L.E. Semin. Liver Dis. 1999; 19: 117-127Crossref PubMed Scopus (133) Google Scholar). The genes that are either increased from basal levels of expression or are induced de novo encode proteins involved in maintaining homeostasis and stimulating cells to reenter the cell cycle and proliferate. In a model proposed by Fausto (2Fausto N. J. Hepatol. 2000; 32: 19-31Abstract Full Text PDF PubMed Google Scholar), there appear to be at least two distinct phases of liver regeneration that are regulated via different mechanisms: a priming phase in which quiescent hepatocytes are induced to reenter the cell cycle as a result of tumor necrosis factor α and interleukin 6 stimulation, followed by a second phase in which hepatocytes respond to growth factors and progress through the G1 stage of the cell cycle (2Fausto N. J. Hepatol. 2000; 32: 19-31Abstract Full Text PDF PubMed Google Scholar). The priming phase in the rodent hepatectomy model in which immediate-early genes are activated corresponds to the first 4 h post-hepatectomy (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, 7Weglarz T.C. Sandgren E.P. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 12595-12600Crossref PubMed Scopus (79) Google Scholar). The peak of DNA synthesis occurs ∼40 h after resection in the mouse, and the period immediately preceding this is associated with the induction of cell cycle genes, including cyclin D1, fox M1b, cyclin-dependent kinases, and the cyclin-dependent kinase inhibitor p21 (also known as CDKN1A) (8Albrecht J.H. Meyer A.H. Hu M.Y. Hepatology. 1997; 25: 557-563Crossref PubMed Scopus (105) Google Scholar, 9Wang X. Kiyokawa H. Dennewitz M.B. Costa R.H. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 16881-16886Crossref PubMed Scopus (277) Google Scholar, 10Abrecht J.H. Hansen L.K. Cell Growth 10: 397-404PubMed Google Scholar, 11Loyer P. Glaise D. Cariou S. Baffet G. Meijer L. Guguen-Guillouzo C. J. Biol. Chem. 1994; 269: 2491-2500Abstract Full Text PDF PubMed Google Scholar, 12Albrecht J.H. Poon R.Y. Ahonen C.L. Rieland B.M. Deng C. Crary G.S. Oncogene. 1998; 16: 2141-2150Crossref PubMed Scopus (167) Google Scholar). Many of the individual genes involved in the priming phase of liver regeneration have been identified using mouse genetic models and more recently through the use of high density oligonucleotide arrays (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, 13Li W. Liang X. Kovalovich K.K. Poli V. Taub R. J. Biol. Chem. 2002; 277: 28411-28417Abstract Full Text Full Text PDF PubMed Scopus (264) Google Scholar, 14Strey C.W. Markiewski M. Mastellos D. Tudoran R. Spruce L.A. Greenbaum L.E. Lambris J.D. J. Exp. Med. 2003; 198: 913-923Crossref PubMed Scopus (343) Google Scholar, 15Liao Y. Shikapwashya O.N. Shteyer E. Dieckgraefe B.K. Huruz P.W. Rudnick D.A. J. Biol. Chem. 2004; 279: 43107-43116Abstract Full Text Full Text PDF PubMed Scopus (87) Google Scholar, 16Yamada Y. Webber E.M. Kirillova I. Peschon J.J. Fausto N. Hepatology. 1998; 28: 959-970Crossref PubMed Scopus (216) Google Scholar, 17Yamada Y. Kirillova I. Peschon J.J. Fausto N. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 1441-1446Crossref PubMed Scopus (842) Google Scholar, 18Fukuhara Y. Hirasawa A. Li X.-K. Kawasaki M. Fujino M. Funeshima N. Katsuma S. Shiojima S. Yamada M. Okuyama T. Suzuki S. Tsujimoto G. J. Hepatol. 2003; 38: 784-792Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar, 19Arai M. Yokosuka O. Chiba T. Imazeki F. Kato M. Hashida J. Ueda Y. Sugano S. Hashimoto K. Saisho H. Takiguchi M. Seki N. J. Biol. Chem. 2003; 278: 29813-29818Abstract Full Text Full Text PDF PubMed Scopus (70) Google Scholar, 20Locker J. Tian J. Carver R. Concas D. Cossu C. Leda-Columbiano G.M. Columbano A. Hepatology. 2003; 38: 314-325Crossref PubMed Scopus (74) Google Scholar, 21Cressman D.E. Greenbaum L.E. DeAngelis R.A. Ciliberto G. Furth E. Poli V. Taub R. Science. 1996; 274: 1379-1383Crossref PubMed Scopus (1325) Google Scholar, 22Greenbaum L.E. Li W. Cressman D. Peng Y. Ciliberto G. Poli V. Taub R. J. Clin. Investig. 1998; 102: 996-1007Crossref PubMed Scopus (237) Google Scholar). However, information regarding the pattern of gene expression in the periods following this immediate early priming phase is still limited. In addition, as a consequence of the complexity and robust nature of the transcriptional changes that occur in response to partial hepatectomy, there remains much to be learned regarding the integration of the products of these genes into the complex signaling and transcriptional hierarchies that result in this highly synchronized regenerative response. In this study, we utilized a cDNA microarray enriched for genes expressed in hepatocytes to profile changes in gene expression in mouse liver cDNA samples isolated 0, 2, 16, and 40 h post-hepatectomy (23Kaestner K.H. Lee C.S. Scearce L.M. Brestelli J.E. Arsenlis A. Le P.P. Lantz K.A. Crabtree J. Pizarro A. Mazzarelli J. Pinney D. Fischer S. Manduchi E. Stoeckert Jr., C.J. Gradwohl G. Clifton S.W. Brown J.R. Inoue H. Cras-Meneur C. Permutt M.A. Diabetes. 2003; 52: 1604-1610Crossref PubMed Scopus (50) Google Scholar, 27Scearce L.M. Brestelli J.E. McWeeney S.K. Lee C.S. Mazzarelli J. Pinney D.F. Pizarro A. Stoekert Jr., C.J. Clifton S.W. Permutt M.A. Brown J. Melton D.A. Kaestner K.H. Diabetes. 2002; 51: 1997-2004Crossref PubMed Scopus (73) Google Scholar). We selected these time points to detect differentially expressed genes during the priming phase, mid-G1, and peak of S phase in the mouse partial hepatectomy model (see Fig. 1A).Table IPAGE analysis of significantly differentially expressed genes between time points Two class, unpaired data analysis with 20% false discovery rate.ComparisonTotalUp-regulated (>1.5)Down-regulated (1.5-fold up or down.View Large microarray performed using models of liver regeneration have of temporally regulated genes have to these genes regulatory networks (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, 18Fukuhara Y. Hirasawa A. Li X.-K. Kawasaki M. Fujino M. Funeshima N. Katsuma S. Shiojima S. Yamada M. Okuyama T. Suzuki S. Tsujimoto G. J. Hepatol. 2003; 38: 784-792Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar, 19Arai M. Yokosuka O. Chiba T. Imazeki F. Kato M. Hashida J. Ueda Y. Sugano S. Hashimoto K. Saisho H. Takiguchi M. Seki N. J. Biol. Chem. 2003; 278: 29813-29818Abstract Full Text Full Text PDF PubMed Scopus (70) Google Scholar, 20Locker J. Tian J. Carver R. Concas D. Cossu C. Leda-Columbiano G.M. Columbano A. Hepatology. 2003; 38: 314-325Crossref PubMed Scopus (74) Google Scholar). these have identified factors and highly expressed genes that are through microarray analysis. In the study, we novel to the analysis of microarray on time points during the hepatocyte cell cycle that have been We novel computational tools to identify and regulatory networks that hepatic proliferation in vivo. We demonstrated that this those genes in which expression levels are the of microarray including those growth factors and DNA This be to expression data of or of were to partial hepatectomy as L.E. Li W. Cressman D. Peng Y. Ciliberto G. Poli V. Taub R. J. Clin. Investig. 1998; 102: 996-1007Crossref PubMed Scopus (237) Google Scholar). The were and was from the liver from at 0, 2, 16, and 40 h post-hepatectomy by in and by R. Taub R. Biol. PubMed Scopus Google Scholar). samples were using to the and of the samples this control with a of the S of were used for expression analysis. and of was using and to reverse was to the cDNA and to the cDNA microarray (23Kaestner K.H. Lee C.S. Scearce L.M. Brestelli J.E. Arsenlis A. Le P.P. Lantz K.A. Crabtree J. Pizarro A. Mazzarelli J. Pinney D. Fischer S. Manduchi E. Stoeckert Jr., C.J. Gradwohl G. Clifton S.W. Brown J.R. Inoue H. Cras-Meneur C. Permutt M.A. Diabetes. 2003; 52: 1604-1610Crossref PubMed Scopus (50) Google Scholar, 27Scearce L.M. Brestelli J.E. McWeeney S.K. Lee C.S. Mazzarelli J. Pinney D.F. Pizarro A. Stoekert Jr., C.J. Clifton S.W. Permutt M.A. Brown J. Melton D.A. Kaestner K.H. Diabetes. 2002; 51: 1997-2004Crossref PubMed Scopus (73) Google Scholar). liver from one of the time points was with a control with a different The control of a of from both and I. L. P. A. E. D. C. S. D. G. F. L. R. A. Ciliberto G. F. Poli V. J. PubMed Scopus Google from quiescent 2, 16, 30 and h This of expression and was used to the of in the of expression for the time analysis. biological were used for of the and 40-h time a was such that for time two of the biological were as control and the two biological samples as control The analysis the of or on the of gene expression at a time The of on the were by using the and the of expression for on the was in of and A The data were by the using the in the S. J. Biol. 2002; Google Scholar). statistical the genes were identified as differentially expressed using the from Gene used from Gene quantitative reverse transcription as E. S. Stoeckert Jr., C.J. 2000; 16: PubMed Scopus Google Scholar). unpaired data were performed to identify genes that were differentially expressed by more the different time and In a false discovery of was to identify differentially expressed The microarray data are through the a data data E. H. J. Pizarro Stoeckert Jr., C.J. 2004; PubMed Scopus Google at gene expression was confirmed using Liver was reverse at for h with of and reverse were using the of and the at a were to The were performed using the program on the were for min and 40 of or and followed by a analysis. were performed with biological and with The cycle was used for and cycle were to the expression of the gene to be differentially expressed with microarray expression are The was used to that the was significant and the of the by the unpaired using the table with a of was used for genes was used to in the of genes identified as significantly differentially expressed at 2, 16, and 40 h post-hepatectomy D.A. Jr., R.A. Biol. 2003; PubMed Google Scholar). first Gene for genes using the by the gene Within various systems by the gene the is for of gene biological or molecular and the are for within those systems hormone or In a second the of genes differentially expressed at time point are to which are the of the number of genes in a pathway in the of differentially expressed genes is with the of the pathway the genes on the microarray and is as the This is used to the known as the which is a to the that in of by more genes and more robust be to the the of the expressed on the were with gene using assembly from the of The for of and tools at the for and Jr., G. D.A. J. R.A. Biol. 2003; PubMed Google Scholar). The expressed were from the as novel and these on the be with a gene false of a gene were removed from the a of gene and and gene genes differentially expressed by >1.5-fold at 2, 16, and 40 The data were for using the gene cellular molecular function and biological by M. D. H. K. M.A. L. A. S. J.E. M. G.M. G. 2000; 25: PubMed Scopus Google Scholar). The of and genes at time point were with the function of and the most significantly were identified using significant from the biological in networks were from the of genes that were differentially expressed at the and 40-h time points with the time data were through the use of a that the and analysis of networks to and data gene gene and corresponding expression as changes were as a gene was to corresponding gene in the The genes identified as differentially expressed by statistical analysis using PaGE and which a between of the time points of were in the analysis. genes, were used as the point for biological the program the for between genes and gene in the and a of The program a for to the of the to the of The the of the genes in a to A of that there is a in that the genes were randomly into a to of during Liver identify genes that were differentially expressed at specific phases of the regenerative liver samples were from 0, 2, 16, and 40 h after partial time points to quiescent liver, early G1, mid-G1, and S phase of the hepatocyte cell cycle in this significant to high analysis of large changes in gene expression is the of statistical that both the and of of differentially expressed The PaGE statistical analysis is a for microarray gene expression data for expressed genes between two and for E. S. Stoeckert Jr., C.J. 2000; 16: PubMed Scopus Google Scholar). PaGE that using a false discovery on In the study, we used a false discovery of 20% to significantly identify genes involved in early growth as as phases of liver we which genes were differentially expressed between time unpaired were performed to identify genes that were differentially expressed by >1.5-fold at 2, 16, and 40 h post-hepatectomy with quiescent these a of genes were to be differentially expressed and Fig. A in the number of genes differentially expressed was with time after partial hepatectomy, with the between and the in gene the genes by up or for of the the of genes is in The of differentially expressed genes are in the Venn diagram of Fig. the that the number of genes were differentially regulated at 40 h post-hepatectomy or genes identified in of the in gene gene h 2 h h 16 h h 40 h Open table in a new tab PaGE of in Gene was used to the PaGE statistical analysis identified differentially expressed genes were randomly from the 40 h these genes, were to be differentially expressed between and 40 h post-hepatectomy by PaGE is a highly accurate for the analysis of microarray The of the PaGE statistical analysis program was by the between the 20% false discovery and the that of the genes significant in gene expression by of for and in the changes in the expression of networks of genes involved in growth or are for restoration of liver mass and during the post-hepatectomy The program (see and was used to identify biological genes that were differentially expressed at 2, 16, and 40 h post-hepatectomy D.A. Jr., R.A. Biol. 2003; PubMed Google Scholar). Fig. of the most significant that were identified using this a of with significant is within involved in and lipid were as early as the post-hepatectomy time point and the of the time the post-hepatectomy time which corresponds to hepatocyte phase, genes involved in and as as were and 40 h 40 h the peak of S phase is biological on were corresponding to the increased of the hepatocytes during DNA the of the analysis a shift to those transcriptional that encode proteins involved in DNA synthesis and with a corresponding in the expression of genes proteins involved in most which is in the in Fig. of the analysis information regarding changes in large of biological we were to individual genes were into specific regulatory and signaling This of analysis has been in microarray of the liver and that have been for the partial hepatectomy model. networks were from the of genes that were differentially expressed at the and 40-h time points through the use of (see and for time were and we have the most significant for stage of liver regeneration 2 h the pathway in Fig. 4 was identified as the most genes, with a highly significant of This pathway a of this and analysis that genes involved in were differentially expressed 2 h post-hepatectomy However, of the known growth response genes that are induced in early G1 phase were as of analysis. of these growth response genes are on the were of the of this and microarray As in Fig. the pathway analysis identified and as in early growth on the expression of genes in the by this we utilized the more analysis to that and were and The of these data confirmed the of the induction of these in the partial hepatectomy model and allowed us to identify a between and inhibitor of kinase signaling L.A. S. J. 1996; PubMed Scopus Google Scholar). this is the first of expression and post-hepatectomy in the 16 h the in Fig. genes, with a highly significant of The of the expression for at this time point identified regulatory to genes that were differentially expressed at the post-hepatectomy time The that were differentially expressed encode proteins involved in including cytokine signaling X. F. J. 2003; PubMed Scopus Google Scholar, X. F. J. Biol. Chem. 2003; 278: Full Text Full Text PDF PubMed Scopus Google Scholar), M. T. M. P. J. Biol. Chem. 2003; 278: Full Text Full Text PDF PubMed Scopus Google Scholar), and cell cycle J.H. Poon R.Y. Ahonen C.L. Rieland B.M. Deng C. Crary G.S. Oncogene. 1998; 16: 2141-2150Crossref PubMed Scopus (167) Google Scholar). The 40-h time point corresponds to the peak of hepatocyte S phase in the partial hepatectomy model. of genes differentially expressed at this point identified regulatory including involved in DNA C.J. 2004; 12: PubMed Scopus Google Scholar), Y. J. Biol. Chem. 2004; 279: Full Text Full Text PDF PubMed Scopus Google Scholar), G. U. J. Cell Sci. 2003; PubMed Scopus Google mitotic spindle assembly L.K. Y. PubMed Scopus Google Scholar), M. L. Biol. 2003; Google Scholar), M. V. U. M. 2004; PubMed Scopus Google Scholar), T. T. Y. Y. K. T. Cell PubMed Scopus Google Scholar, de T. A. O. V. M. A. V. M. D. 102: PubMed Scopus Google and mitotic checkpoint control E. M.B. J. 2004; Full Text Full Text PDF PubMed Scopus Google Scholar), L.G. J. Cell Biol. 2003; PubMed Scopus Google Scholar), M. Cell Biol. 2003; PubMed Scopus Google Scholar), D. G. M. M. A. Cell Biol. PubMed Scopus Google Scholar), M. K. Inoue H. H. T. T. K. T. S. K. A. H. T. 2002; PubMed Scopus Google Scholar), J. X. J. Biol. Chem. 2004; 279: Full Text Full Text PDF PubMed Scopus Google the for of and of mitotic to the phase M. Cell Biol. Full Text Full Text PDF PubMed Scopus Google Scholar, C.J. J. Scopus Google Scholar). We demonstrated that mRNA levels for a large number of genes for phase at the S phase peak are regulated the that either transcriptional activation and/or a in phases of cell cycle the G1 analysis levels of mRNA and between these two genes identified on including and are transcriptional to growth be responsible for the of these mitotic checkpoint genes in the liver E. G. E. M. M. L. V. W. R. S.W. C. 2004; PubMed Scopus Google Scholar, S. Y. E. D. Oncogene. 2002; Scopus Google Scholar). to detect changes in gene expression at the 40-h post-hepatectomy time point is to be a function of the highly synchronous nature of this in vivo model of cell cycle a to the hepatectomy model. In this study, we the changes in gene expression that occur following partial hepatectomy at selected time points with the of the biological and specific regulatory networks that are during early priming stages of the regenerative and time points corresponding to and S the use of a variety of computational analysis we have identified genes that are differentially expressed during various stages of liver regeneration and have characterized biological and regulatory networks that both and specific information regarding this highly synchronized proliferative and response. the between individual genes identified by pathway analysis are from the to this is the first of the of of these regulatory networks in the partial hepatectomy model. In addition, we have demonstrated that the use of pathway analysis to two of microarray expression which are both limited and the of genes on a transcriptional is a by which function is pathway analysis genes in a regulatory that are or as a result of that be regulatory The for the analysis of the liver regeneration as a model for the of complex biological