Gene expression pattern in hepatic stem/progenitor cells during rat fetal development using complementary DNA microarrays


  • Petko M. Petkov,

    1. Marion Bessin Liver Research Center, Albert Einstein College of Medicine, Bronx, NY
    2. Department of Medicine, Albert Einstein College of Medicine, Bronx, NY
    3. Jackson Laboratories, Bar Harbor, ME
    Search for more papers by this author
    • P.M.P. and J.Z. contributed equally to this work.

  • Jiri Zavadil,

    1. Biotechnology Center and Division of Nephrology, Albert Einstein College of Medicine, Bronx, NY
    2. Department of Pathology, NYU School of Medicine, New York, NY
    Search for more papers by this author
    • P.M.P. and J.Z. contributed equally to this work.

  • David Goetz,

    1. Marion Bessin Liver Research Center, Albert Einstein College of Medicine, Bronx, NY
    2. Department of Medicine, Albert Einstein College of Medicine, Bronx, NY
    Search for more papers by this author
  • Tearina Chu,

    1. Marion Bessin Liver Research Center, Albert Einstein College of Medicine, Bronx, NY
    2. Department of Medicine, Albert Einstein College of Medicine, Bronx, NY
    3. Department of Molecular Biology and Biochemistry, Mount Sinai School of Medicine, New York, NY
    Search for more papers by this author
  • Robert Carver,

    1. Marion Bessin Liver Research Center, Albert Einstein College of Medicine, Bronx, NY
    2. Department of Medicine, Albert Einstein College of Medicine, Bronx, NY
    3. Enzo Biochemical, Farmington, NY
    Search for more papers by this author
  • Charles E. Rogler,

    1. Marion Bessin Liver Research Center, Albert Einstein College of Medicine, Bronx, NY
    2. Department of Medicine, Albert Einstein College of Medicine, Bronx, NY
    Search for more papers by this author
  • Erwin P. Bottinger,

    1. Marion Bessin Liver Research Center, Albert Einstein College of Medicine, Bronx, NY
    2. Biotechnology Center and Division of Nephrology, Albert Einstein College of Medicine, Bronx, NY
    Search for more papers by this author
  • David A. Shafritz,

    1. Marion Bessin Liver Research Center, Albert Einstein College of Medicine, Bronx, NY
    2. Department of Medicine, Albert Einstein College of Medicine, Bronx, NY
    Search for more papers by this author
  • Mariana D. Dabeva

    Corresponding author
    1. Marion Bessin Liver Research Center, Albert Einstein College of Medicine, Bronx, NY
    2. Department of Medicine, Albert Einstein College of Medicine, Bronx, NY
    • Marion Bessin Liver Research Center, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461
    Search for more papers by this author
    • fax: 718-430-8975


To identify new and differentially expressed genes in rat fetal liver epithelial stem/progenitor cells during their proliferation, lineage commitment, and differentiation, we used a high throughput method—mouse complementary DNA (cDNA) microarrays—for analysis of gene expression. The gene expression pattern of rat hepatic cells was studied during their differentiation in vivo: from embryonic day (ED) 13 until adulthood. The differentially regulated genes were grouped into two clusters: a cluster of up-regulated genes comprised of 281 clones and a cluster of down-regulated genes comprised of 230 members. The expression of the latter increased abruptly between ED 16 and ED 17. Many of the overexpressed genes from the first cluster fall into distinct, differentially expressed functional groups: genes related to development, morphogenesis, and differentiation; calcium- and phospholipid-binding proteins and signal transducers; and cell adhesion, migration, and matrix proteins. Several other functional groups of genes that are initially down-regulated, then increase during development, also emerged: genes related to inflammation, blood coagulation, detoxification, serum proteins, amino acids, lipids, and carbohydrate metabolism. Twenty-eight genes overexpressed in fetal liver that were not detected in adult liver are suggested as potential markers for identification of liver progenitor cells. In conclusion, our data show that the gene expression program of fetal hepatoblasts differs profoundly from that of adult hepatocytes and that it is regulated in a specific manner with a major switch at ED 16 to 17, marking a dramatic change in the gene expression program during the transition of fetal liver progenitor cells from an undifferentiated to a differentiated state. Supplementary material for this article can be found on the HEPATOLOGY website ( (HEPATOLOGY 2004;39:617–627.)

Over the years, substantial evidence has accumulated suggesting the existence of potential liver stem cells (LSCs) in the adult liver. This evidence is indirect, because in all reports the putative LSCs were activated to proliferate and differentiate into liver progenitor cells only when the regenerative capacity of terminally differentiated hepatocytes was compromised.1–3 Because these cells are not under constant renewal—in contrast to epithelial cells of the intestine or the skin—they escape detection in the quiescent liver. For this reason, a unique specific marker for LSCs in adult liver has not yet been identified, and the LSCs have not yet been isolated.

To study liver stem/progenitor cells (LS/PCs), we took a different approach. We began to analyze the gene expression profile of hepatic cells in fetal liver. Around embryonic day (ED) 8.5 in the mouse and 1 day later in the rat, primitive epithelial cells of the ventral foregut come in contact with cardiac mesoderm and fibroblast growth factor signaling, leading to formation of the hepatic diverticulum.4–6 Subsequently, cells of the primary liver diverticulum invade the septum transversum and, under the inductive signals of bone morphogenic proteins, proliferate extensively and differentiate further.6 These cells are considered committed hepatic epithelial cells and have been termed hepatoblasts.4–6 Some of these cells are at a very early stage of their differentiation, and their expression profile could provide clues about the expression pattern of hepatic stem cells.

We speculate that the number of potential stem cells and their progeny is much higher in the fetal liver than in any of the models developed to date for expansion of oval cells. Complementary DNA (cDNA) microarrays are a promising means of studying the gene expression pattern of LS/PCs in fetal liver—as well as in different models of liver progenitor cell activation and proliferation—because this method possesses enormous capabilities for analysis of RNA samples with a large number of cDNA clones printed on the arrays. Therefore, taking into account the high homology between the mouse and rat genomes, we used rat fetal liver as a source for LS/PCs and mouse cDNA microarrays for analysis of their gene expression pattern to identify genes specifically expressed in these cells. This analysis was similar to that reported by Plescia and colleagues,7 using a mouse epithelial cell line derived from the liver diverticulum at ED 9.5 that exhibits the ability to differentiate along the hepatocytic lineage in culture in response to dimethylsulfoxide.8 Based on the expression profile of LS/PC, we identified major clusters of up- and down-regulated genes during rat liver development and a group of genes that are differentially expressed in fetal hepatoblasts, which are considered to be LS/PC-specific.


cDNA, complementary DNA; ED, embryonic day; LSCs, liver stem cells; LS/PCs, liver stem/progenitor cells; SDS, sodium dodecyl sulfate; SSC, saline sodium citrate; mRNA, messenger RNA; RT-PCR, reverse-transcriptase polymerase chain reaction; Hey1, hairy/enhancer-of-split related with YRPW motif 1; Tnc, tenascin C; EST, expressed sequence tags; Peg3, paternally expressed gene 3; Klf5, Kruppel-like factor 5; Grb10, growth factor receptor bound protein 10; Fah2, four and a half LIM domains 2; TNF, tumor necrosis factor.

Materials and Methods

Isolation of Fetal, Newborn, and Adult Liver Cells.

Cell suspensions from rat fetal liver were prepared as reported previously.9 The cells were plated on gelatin-coated dishes at a density of 20 × 106 cells per 10-cm plate and epithelial cells were allowed to attach for 16 hours, after which the hematopoietic cells, which do not attach, were removed by washing with phosphate-buffered saline (PBS).

Isolation of RNA.

RNA was isolated from fetal liver cells attached to gelatin coated dishes and from isolated adult rat hepatocytes, using TRIzol Reagent (Invitrogen/BRL, Carlsbad, CA), precipitated several times with LiCl/ethanol and purified further using the RNeasy kit (Qiagene, Valencia, CA).

cDNA Microarrays.

Mouse cDNA microarrays were obtained from the Functional Genomics Facility, Albert Einstein College of Medicine. Each chip contained 8976 clones originally obtained from Incyte Cenomics (MGEM V1.0 sequence verified set). The array platform is deposited in Gene Expression Omnibus (GEO) data repository and can be viewed or downloaded under accession number GPL409.

Labeling of cDNAs, Hybridization, and Analysis of Microarrays.

Target and reference RNAs were labeled with Cy5 and Cy3, respectively, by synthesizing first-strand cDNA using a standard protocol ( Prehybridization was carried out for 1 hour at 50° C in 35% formamide, 4 × saline sodium phosphate and EDTA (SSPE), 2.5 × Denhardt's, 0.5% sodium dodecyl sulfate (SDS), and 200 mg/mL of salmon sperm DNA. Hybridization was performed overnight at 50° C in hybridization solution as above, containing 1× blocking solution (poly-dT, tRNA, and mouse Cot1 DNA).

Following hybridization, the slides were washed consecutively in 2 × SSC/0.1% SDS, 0.2 × saline sodium citrate (SSC)/0.1% SDS, and 0.2 × SSC. A custom-built dual channel laser scanning microscope10 and the ScanAlyze Version 2.44 software (M. Eisen, Stanford University, Palo Alto, CA) were used to generate raw data files containing measurements of signal and background fluorescence emissions of Cy3 and Cy5, respectively, for each spot. Normalization of data was performed based on the 50th percentile median ratio of all arrays, using the GeneSpring 4.0 package (Silcone Genetics, Redwood City, CA). Data filtering was based on expression flags (gene expression presence or absence calls at single experimental time points and across the entire experimental set). A gene was considered expressed when the measured intensity of the spot signal channel reached a value greater than local spot background plus 2 standard deviations of the average background calculated for the entire particular slide. Significant differential messenger RNA (mRNA) abundance was established by Significance Analysis of Microarrays software (Stanford University Statistics and Biochemistry Labs, Stanford, CA), “applying a 5% false discovery rate.”

Three biological repeats were performed for each of nine experimental time points.

Quantitative Polymerase Chain Reaction.

Quantitative real-time reverse-transcriptase polymerase chain reaction (RT-PCR) was performed using the ABI Prizm 7900 Sequence Detection System (Applied Biosystems, Foster City, CA) in a two-step RT-PCR. All RNAs were treated with DNAse (Roche Applied Science, Indianapolis, IN) and purified through an RNeasy column (Qiagen). cDNA was synthesized with Superscript II reverse transcriptase (Invitrogen/BRL) as described above. Rat-specific sequences for PCR primers were designed to generate amplicons of 50 to 100 base pairs required for quantitative real-time detection using SYBR Green Master Mix (Applied Biosystems). The mRNA abundances were determined by normalization of the data to the expression levels of glyceraldehyde-3-phosphate dehydrogenase mRNA.

Reverse Northern Blot Hybridization.

Specific mouse cDNA clones of interest were propagated and plasmid DNA was purified using DNeasy (Qiagen). The plasmid DNAs were digested with EcoRI and 1 μg of DNA loaded on Zeta-Probe blotting membranes using a Bio-Dot microfiltration apparatus (Bio-Rad, Hercules, CA). cDNAs were labeled with 32P-dCTP, using the RadPrime DNA labeling system (Invitrogen/BRL). The blots were hybridized at 63° C to the high specific activity probes in QuikHyb (Stratagene, La Jolla, CA) and washed according to instructions from the manufacturer.

Results and Discussion

Fetal Liver Epithelial Progenitor Cells

To study the gene expression profile of fetal hepatoblasts, we developed a procedure to eliminate blood cells from the isolated fetal liver cell suspension by plating them on gelatin-coated dishes overnight and washing out the blood cells (see Materials and Methods section above). Hepatoblasts comprised more than 90% of the cells attached to the dishes, as judged by α-fetoprotein expression (Fig. 1). RNA was isolated 16 hours after plating the cells, at which time all genes expressed in hepatoblasts during their isolation remained active (e.g., alpha-fetoprotein, albumin). Also, genes expressed at later embryonic days did not turn on during attachment (e.g., dipeptidyl-peptidase 4 [DPP4]; see Fig. 1). [We also performed quantitative RT-PCR analysis for a dozen other genes in the cells, selected from those listed in Table 3, at the time of their isolation with that after plating for 16 hours on gelatin-coated dishes; no significant differences were found].

Figure 1.

Gene expression in hepatoblasts at the time of isolation. Hepatoblasts were isolated as described in Material and Methods and plated on gelatin coated dishes. During attachment, the cells aggregate, forming clusters. (A) Cytochemical detection of DPPIV is negative; hepatoblasts do not express DPPIV before ED 16. (B) All cells in this cluster are positive for γ-GT. (C) Negative control (primary antibody omitted) for the immunocytochemical reactions in D, E and F. (D) Albumin; (E) CK-19 and (F) AFP. Secondary antibodies are alkaline phosphatase conjugated. Vector red (Vector Biolabs, Burlingame, CA) was used as a substrate. Original magnification in A–E: 400× and F: 200×.

Table 3. Differentially Expressed Genes in Rat Fetal Liver Stem/Progenitor Cells
GB Acc. No.Gene NameGene DescriptionClassificationExpression in Adult Liver
  1. The genes in this table are expressed in fetal but show very low level of expression in adult liver determined by reverse Northern analysis.

AA260248Grb10* Growth factor receptor bound protein 10Signal transductionNo
AA387076Akap12* A kinase (PRKA) anchor protein (gravin) 12Signal transductionNo
AA049474Hey1* hairy/enhancer-of-split related with YRPW motif 1Regulation of transcriptionNo
AA432818Klf5* Kruppel-like factor 5Regulation of transcriptionNo
AA175651Casp11 Caspase 11ApoptosisNo
AA220316Ppap2c* phosphatidic acid phosphatase type 2cCell migrationNo
AA274932Gpc3* Glypican 3Histogenesis, organogenesisNo
W75532Sema3f semaphorin IV isoformMorphogenesisNo
AA153907Pcolce2 procollagen C-endopeptidase enhancer 2MorphogenesisNo
AA212256Mnat1* Menage a troit 1ProliferationNo
AA177949n/a* Murine Glvr-1 mRNA, complete cdsProtein traffickingNo
AA178155Scya4 Small inducible cytokine A4InflammationNo
AA172456Scya12 Small inducible cytokine A12InflammationNo
AA168437BaspI Brain abundant, membrane attached signal protein 1Expressed in nervous systemNo
AA259388n/a ESTsESTNo
AA466269n/a ESTESTNo
AA259552n/a* ESTESTNo
AA163624n/a* expressed sequence AI507382ESTNo
AA178792n/a ESTsESTNo
AA146022n/a* ESTESTNo
AA174941n/a* Hypothetical protein coding geneESTNo
AA050648n/a RIKEN cDNA 2810022L02 geneESTNo
AK006487n/a* RIKEN cDNA 1700029F09ESTNo
XM_24076n/a EST, similar to rat ESTESTNo
AA119887n/a* RIKEN cDNA 1700025G04 geneESTNo
AA198514n/a* RIKEN cDNA 2310061F22 geneESTNo
AA466758n/a* RIKEN cDNA 2410008J01 geneESTNo
AA276844Pkcd* Protein kinase C, deltaSignal transductionNegligible
AA254528Magoh* Mus musculus mm-Mago mRNA, complete cdsDevelopmentNegligible
AA217593Enah* enabled homolog (Drosophila)Morphogenesis; differentiationNegligible
AA060802Nfkb2* Nuclear factor of kappa light polypeptide in B-cells 2HematopoiesisNegligible
W70782Npy Neuropeptide YAngiogenesisNegligible
AA051790Adam19* A disintegrin and matalloproteinase domain 19Extracellular metallopeptidaseNegligible
AA137523Ubap2* Ubiquitin-associated protein 2Function unknownNegligible
AA286393Scya12 Small inducible cytokine A12InflammationNegligible
AA261171n/a ESTESTNegligible
AA014723n/a* ESTESTNegligible
W36046n/a* expressed sequence AW549277ESTNegligible
AA108457n/a RIKEN cDNA 2610109B12 geneESTNegligible
AA473227n/a* |expressed sequence AA589632|ESTNegligible
AA434898n/a ESTESTNegligible
W50706n/a ESTESTNegligible
AA265225Shcbp1* Shc SH2-domain binding protein 1Signal transductionVery low
AA023645Fhl2* Four and a half LIM domains 2Regulation of transcriptionVery low
AA003064Peg3* paternally expressed gene 3Transcription factor/apoptosisVery low
AA002393Orc61 Origin recognition complex, subunit 6-likeProliferationVery low
W96914Lox* Lysyl oxidaseCell adhesion and motilityVery low
AA270625Tnc* Tenascin CCell deadhesion and migrationVery low
AA051341Tnfip6 Tumor necrosis factor induced protein 6Cell adhesionVery low
AA200473Manba* Mannosidase beta A, lysosomalExoglycosidase, N-linkedVery low
BE134291n/a* Similar to sodium/potassium transporterSodium pumpVery low
BC003795Tagln* TransgelinCytoskeleton organizationVery low
BC021796Actg* Cytoplasmic gamma actinCytoskeletonVery low
AA222216Tubb4 tubulin, beta 4CytoskeletonVery low
AA197461Xpot* exportin, tRNA (nuclear export receptor for tRNAs)Cellular transportVery low
AA178286n/a* expressed sequence AI461788ESTVery low
AA146200EST RIKEN cDNA B230113H15 geneESTVery low
W14186n/a* ESTESTVery low
AA544789n/a* RIKEN cDNA 9430080K19 geneESTVery low
AA518354n/a* expressed sequence AA589382ESTVery low
AA462971n/a* RIKEN cDNA 1500012F11 geneESTVery low
W12782n/a RIKEN cDNA 2810474H01 geneESTVery low
  * Genes expressed in stem cells  

Use of Murine cDNA Microarrays to Study the Gene Expression Profile of Fetal Rat Liver Epithelial Cells

To screen a large number of expressed genes and identify new and differentially expressed genes in fetal hepatoblasts, we used murine cDNA microarrays with cDNAs from rat fetal liver RNAs at different embryological time points as targets. RNAs from ED 13, 15, 17, newborn, 7-day-old, and adult liver were isolated and labeled with Cy5, and RNA from newborn liver was labeled with Cy3 as a common reference. In these experiments, 93% of the clones spotted on the microarrays revealed hybridization signals with Cy5 or Cy3 probes. After normalization and elimination of the genes that did not pass the criteria for expression (see Materials and Methods section above), the number of expressed genes at different time points was reduced to 80 ± 4.4%. A substantial number of genes in ED13 liver are up-regulated (Fig. 2). This number decreased toward birth and increased again in adult liver. Another prominent set of genes appeared to be down-regulated in early fetal liver, while others were up-regulated in adult liver (see Fig. 2).

Figure 2.

Up- and downregulated genes during rat liver development. RNA was isolated from rat fetal liver at ED 13, 15, 17, newborn (NB), 7 days after birth (7D) and adult liver. The respective cDNAs were labeled with cy5 fluorophore. Newborn liver cDNA, labeled with cy3, was used as a reference for all time points. The left columns (closed box) represent up-regulated genes and the right columns (open box) represent down-regulated genes compared to the reference.

Global Patterns of Gene Expression During Hepatoblast Proliferation and Differentiation

To study differentially expressed genes in fetal liver in more detail and to quantitate changes in their expression during development, we studied the expression pattern of these genes on microarrays during consecutive days in rat fetal liver development at ED 13, 14, 15, 16, 17, 18, newborn, 7-day-old, and adult. We concentrated on genes that were induced or repressed in ED 13 to 14 liver and were then gradually down- or up-regulated. Using Significance Analysis of Microarrays software at a 5% false discovery rate and searching for genes that were up-or down-regulated in at least two of the nine time points, we found 511 genes represented on the cDNA microarrays to be specifically regulated in fetal hepatoblasts. These genes were further distributed in two major hierarchical clusters according to their expression pattern. As shown in Fig. 3A, three separate sets of RNA isolates from fetal liver at each of the nine experimental time points showed a reproducible pattern of expression. Genes in the first cluster—comprised of 281 members (Fig. 3B, upper panel, self-organizing map cluster 1)—were overexpressed in ED 13 through ED 16 hepatoblasts, after which their expression declined gradually. Genes in the second group—comprised of 230 members (Fig. 3B, lower panel, self-organizing map cluster 2)—were repressed in fetal liver, but their expression was induced before birth. The centroids of these two clusters obtained with self-organizing maps are shown in red (see Fig. 3B). Interestingly, most of the genes in the second cluster changed their expression pattern abruptly between ED 16 and 17.

Figure 3.

Clustering analysis of 511 genes regulated during fetal rat liver development. Rat liver gene expression profiles were determined at 9 time points (ED 13, 14, 15, 16, 17, 18, newborn rat liver, 7 days after birth and adult). 511 genes were identified as significantly regulated. (A) Hierarchical clustering of the 511 genes identifies two major patterns of regulation, up- and downregulation shown in red and green, respectively. Reproducibility of three replicates of this timed analysis (N1, N2, N3) is clearly discernible. (B) The 511 genes were distributed into two groups by self-organizing map (SOM) analysis, revealing characteristic profiles of the two distinct patterns of regulation (red line: centroid). SD of the average expression in each cluster (grey line).

The pattern of regulation of selected genes of interest in the two major clusters was verified by quantitative RT-PCR. The upper panel of Fig. 4A presents the results of microarray analysis for four genes up-regulated in fetal hepatoblasts: annexin II, a Ca(2+)-dependent phospholipid- and membrane-binding protein; the transcriptional regulator hairy/enhancer-of-split related with YRPW motif 1 (Hey1); tenascin C (Tnc), an extracellular matrix protein; and ESTD13 FL, an expressed sequence tag (EST) from 13-day fetal liver. The lower panel of Fig. 4A presents the result of quantitative RT-PCR for the same set of genes. The upper panel of Fig. 4B presents the results of microarray analysis for four genes down-regulated in fetal hepatoblasts: albumin; phenobarbital inducible Cyp2b (up-regulated after birth); insulin-like growth factor 1, the expression of which gradually increases during development; and transthyretin. The lower panel of Fig. 4B presents the results of quantitative RT-PCR for the same set of genes. When the patterns of expression obtained from microarray hybridizations panels were compared with the quantitative PCR results, the data from these two methods matched very well, confirming the validity of our microarray analysis. The same results were obtained with other overexpressed genes in fetal liver that were down-regulated during later development: proteolipid protein 2, caspase-11, four and a half LIM domain 2, paternally expressed gene 3 (Peg3), Kruppel-like factor 5 (Klf5), and nuclear factor kB2; and for genes down-regulated in early fetal liver, whose expression gradually increased during development: acute phase C-reactive protein, tryptophan dioxygenase, and dipeptidyl peptidase 4 (data not shown).

Figure 4.

Comparison of the expression pattern of selected regulated genes during fetal liver development by quantitative RT-PCR and microarrays. (A) Expression pattern of up-regulated genes: annexin 2 (Anx2), hairy/enhancer-of-split related with YRPW motif (Hey-1), tenascin C (Tnc), and EST expressed in 13 day fetal liver (ESTD13FL). Upper panels: microarray analysis; lower panels: quantitative RT-PCR. The expression pattern is compared for fetal liver at ED13, ED14, ED16 and ED18 and in newborn (NB) and adult liver. The ordinate represents relative mRNA abundance (fold difference in the expression compared to newborn liver). (B) Expression pattern of down-regulated genes: albumin, Cyp2b (phenobarbital inducible P450), insulin-like growth factor 1 (Igf1), and transthyretin (Ttr). The relative nRNA abundance for albumin, Cyp2b and Igf1 are given as log2 fold differences.

We also compared the up-regulated genes obtained in this study with those of a suppression substractive hybridization (SSH) cDNA library from rat ED13 fetal liver constructed previously in our laboratory.12 All up-regulated genes identified previously by SSH, for which a clone was present on the microchips, again showed higher expression on cDNA microarrays (Table 1).

Table 1. Rat ED13 SSH cDNA Clones Found to Be Overexpressed Using Mouse cDNA Microarrays
  1. The table lists gene previously identified as differentially expressed in fetal liver compared to adult liver (see ref. 12) and identified again in this study as overexpressed in fetal liver.

Ribosomal protein S17Mouse PZF protein
Ribosomal protein L26Bad protein
Ribosomal protein S23Rb-binding protein
Ribosomal protein S15Protein phosphatase 2A
Ribosomal protein L10AAnnexin V
Ribosomal protein L3CaB1
Ribosomal protein L37AGTPase activating protein
Ribosomal protein S8Histone H3
Ribosomal protein S12Histone H2A
Ribosomal protein L35Endonuclease G
Ribosomal protein S626 S proteasome regulatory protein 7
High mobility group HMG2 proteinDnaJ homoloque 2
Platelet factor 4Chaperonin 10
Human pescadilloSmooth muscle cell LIM protein

Using cDNA microarray analysis, Kelley-Loughnane and colleagues13 studied the expression pattern of 8,636 clones in ED 14.5 fetal mouse liver, 14-day postnatal liver, and 6 and 24 hours following partial hepatectomy and identified genes specific for each growth phase and genes showing dual overexpression during regeneration and one of the developmental phases. Some of the 61 genes overexpressed in mouse fetal liver that were not of hematopoietic origin and were present on our microarrays are also induced in rat fetal liver (e.g., glypican 3, growth factor receptor bound protein 10 (Grb10), Peg3, and several ESTs).

The two major clusters of regulated genes were distributed into functional groups (Table 2). (For a complete list of up-regulated and down-regulated genes in each of these functional groups, see the supplementary material on the HEPATOLOGY website (,Tables 1 and 2.) Many of the 281 up-regulated genes in fetal liver are related to cell growth and proliferation. This includes 50 genes encoding for proteins involved in protein synthesis, processing and degradation (ribosomal proteins, rRNA and mRNA processing, protein folding and sorting, protein degradation, etc.), 10 genes related to increased cellular metabolism, 21 genes encoding intracellular trafficking, 12 genes encoding cell structural proteins, and 11 genes related to proteins participating in DNA replication and control of the cell cycle. Other functional groups also exhibited specificity in their gene expression pattern in fetal hepatoblasts (e.g., genes related to signal transduction, development, morphogenesis and differentiation, cell adhesion, de-adhesion, migration, and matrix proteins). In contrast to genes in the up-regulated cluster, 230 genes in the second cluster showed very low if any expression in the early fetal liver but were up-regulated late in development. The largest groups of these genes were ESTs and genes related to cellular metabolism, including genes reflecting the onset of amino acid, carbohydrate, and lipid metabolism. Four other groups related to liver-specific functions also showed increased expression around or shortly after birth: inflammatory response genes, blood coagulation factors, detoxification genes, and serum proteins.

Table 2. 
Groups of Upregulated Genes in Rat Fetal Compared to Newborn Liver
ESTs and genes with unknown function101
Genes related to protein synthesis, processing and degradation50
Genes related to cell cycle and proliferation11
Genes related to transcription and regulation of transcription12
Growth factors4
Genes related to cellular metabolism10
Genes related to intracellular trafficking and translocation21
Genes related to cell structure and organization12
Genes related to signal transduction15
Genes related to development, morphogenesis and differentiation9
Cell adhesion, cell migration and matrix proteins12
Calcium and phospholipid binding proteins7
Genes related to apoptosis6
Genes related to hematopoiesis, immune respose and inflammation7
Genes preferentially expressed in nervous system4
Groups of Downregulated Genes in Rat Fetal Compared to Newborn Liver
  1. The table represents the number of up- and downregulated in fetal liver genes distributed in functional groups.

ESTs and genes with unknown function79
Genes related to cellular metabolism58
Genes related to inflammatory response17
Genes related to blood coagulation9
Genes related to detoxification8
Genes encoding serum proteins9
Genes related to hematopoiesis5
Genes related to immune response7
Genes related to cellular transport11
Genes related to transcription and regulation of transcription7
Genes encoding growth factors, receptors and signal transducers7
Genes related to cellular proliferation5
Genes related to apoptosis2
Genes encoding structural and extracellular matrix proteins6

Identification of Genes Differentially Expressed in Fetal Compared to Adult Liver

To identify genes differentially expressed in LS/PCs that might be candidates as potential markers of LS/PCs, we searched available databases to select positive clones on our chips that were not previously reported to be expressed in adult liver. To determine which of the microarray candidate genes were not expressed or were expressed at a very low level in adult liver, we obtained the respective cDNA clones from the Microarray Facility (Albert Einstein College of Medicine, Bronx, NY), sequenced them to verify their identity, and hybridized a master blot containing these clones with radioactive-labeled 32P cDNA synthesized from fetal and adult liver RNA. For most of the clones (see Table 3), there was either no expression of the respective mRNA in adult liver or the expression was negligible or very low. However, seven of the clones in the original group were clearly expressed in adult liver and have been excluded from the table. Not surprisingly, 39 of the genes in Table 3 are also expressed in embryonic, hematopoietic, or other stem cells. Among the 63 up-regulated genes in fetal liver that showed very low or no expression in adult liver, 35 are ESTs. The remainder falls into different functional categories (see below). Most interesting are those that regulate proliferation (transcriptional regulation and signal transduction), cell adhesion, survival (apoptosis) and development, morphogenesis, and differentiation of hepatic progenitor cells.

Genes Regulating Transcription.

Four of the transcription factors—Klf5, Hey1, four and a half LIM domains 2 (Fah2), and Peg3—are highly up-regulated in fetal liver but are not expressed in adult liver (Table 3). Expression of these transcription factors in fetal liver has not been reported previously, and their potential role in liver development is of significant interest. Klf5 belongs to the family of zinc finger Kruppel-like factors that act both as positive and negative regulators of transcription and are expressed in a tissue-specific manner. It is overexpressed in intestinal epithelial cells,14 and its expression is developmentally regulated. The function of Klf5 has not yet been elucidated, although it encodes a delayed early response gene to growth stimulation that positively regulates cellular proliferation.15

Hey1 is a member of the basic helix-loop-helix (bHLH) transcriptional regulators that play an important role during embryogenesis, controlling cell fate, cell proliferation, differentiation, and apoptosis. Hey proteins are effectors of Notch signaling16 and are important for neurogenesis, somatogenesis, and organogenesis. They are spatially and temporarily expressed during mouse development.17 The exact function of Hey1 is not known, and its overexpression in the liver during organogenesis has not been reported previously.

Fah2 (also known as SLIM3 or DRAL) is an LIM-only protein containing four and a half LIM domains of a zinc finger motif. Fah2 is a tissue-specific coactivator of the androgen receptor18 and is important for muscle development and myoblast differentiation.19 No function has been assigned for this transcriptional regulator in the liver. We found that Fah2 is expressed at a low level in adult liver and is overexpressed in fetal liver (see Table 3).

Genes Related to Signal Transduction.

Several signal transducers not present in adult liver are also up-regulated in fetal liver (see Table 3). Grb10 is highly expressed in fetal liver but not in adult liver, and, interestingly, is not induced in regenerating rat liver.20 It contains a proline-rich N-terminal motif, a central pleckstrin homology domain and a C-terminal Src homology region 2, and SH2 domain, through which it binds autophosphorylated growth factor receptors upon their activation. Grb10 binds activated insulin receptor, platelet-derived growth factor and epidermal growth factor receptors, and insulin-like growth factor 1 and 2 receptors and functions possibly by inhibiting signaling downstream of the receptors, sequestering the signaling kinases.21–23 Protein kinase C, delta, is the major isoform expressed in hematopoietic cells and is overexpressed in tumors. It is induced by diacylglycerol and in response to mitogenic and inflammatory stimuli and is implicated in regulation of cell growth/cell cycle progression and apoptosis.24 A kinase anchor protein 12, a kinase scaffold protein not expressed in adult liver, is also known as testis specific gene A12 or gravin.25 A kinase anchor protein 12 anchors key signal mediators tethering protein kinase A to its substrates.26 Neuropeptide Y is a neurotransmitter and neurohormone expressed in the central and peripheral nervous system during development and adulthood.27 Its overexpression in fetal liver is intriguing, considering its role in promoting proliferation of olfactory neuronal precursor cells through the extracellular signal-regulated kinase.28

Genes Related to Cell Adhesion/De-adhesion and Migration.

During rapid cell proliferation and organogenesis, hepatoblasts maintain a dynamic relationship with extracellular proteins, and the extracellular matrix has to be degraded and rebuilt. Most of the genes in this group (Table 3) are enzymes that participate in matrix synthesis and degradation. Some are not expressed in adult liver, like lysyl oxidase, an extracellular copper deaminase crosslinking collagen and elastin, and lysosomal mannosidase beta A, which is involved in the degradation of extracellular N-linked oligosaccharide moieties of glycoproteins. The “matricellular” protein Tnc, which is highly expressed during development, cell injury, and oncogenesis, deserves special attention. It is suggested that by stimulating reorganization of actin stress fibers and disassembly of focal adhesion complexes, Tnc regulates the ability of cells to migrate and differentiate.29 Its high induction in fetal liver suggests that it has an important function in liver development; its de-adhesive properties may support proliferation, protect against apoptosis, and maintain cellular differentiation.

Genes Related to Apoptosis.

Liver organogenesis requires growth and proliferation of some cells and removal of others by programmed cell death. However, only a limited number of genes related to apoptosis were present on our chips. mRNAs for tumor necrosis factor (TNF) or TNF-receptor complex proteins were not overexpressed, and it does not seem likely that TNF-induced programmed cell death is active in fetal liver cells. The imprinted Peg3, Peg3/Pw1 (see Table 3), is a multidomain and multifunctional protein and a putative zinc finger Kruppel-type transcription factor involved in the TNF/NF-κB signal transduction pathway.30 Recently, it was determined that Peg3 is a cell death mediator in p53-mediated apoptosis, inducing Bax translocation from the cytosol to mitochondria.31, 32 Peg3 is expressed in adult brain and muscle and in many embryonal tissues; however, its expression in liver was not reported previously.33 As this gene conveys both survival and death signals, it may play a multifunctional role in fetal liver. Although the function of Peg3 in fetal liver cells was not previously reported, it likely functions as a mediator of p53-dependent apoptosis.

Caspase-11 is highly induced in fetal liver (Table 3). It is not constitutively expressed in most tissues but can be induced in macrophages, lymphocytes, and hepatocytes by pro-inflammatory signals (e.g., lipopolysaccharides and interferon-γ).34 Caspase-11 appears to play a dual role as an activator of caspase-1, which in turn activates the production of pro-inflammatory cytokines interleukin-1 and interleukin-1835 and as an activator of caspase-3, which in turn triggers apoptosis.36 It is possible that caspase-11, which is not expressed in adult liver, is an initiator of apoptosis in fetal liver.

Genes Related to Development, Morphogenesis, and Differentiation.

As expected, many of the genes in this group (see Table 3) are expressed at very low levels or are not expressed in adult liver, although the function of most of these genes remains unknown. Procollagen C-proteinase enhancer is a tenfold enhancer of procollagen COOH-terminal-proteinase, processing type I collagen and bone morphogenic protein 1, and is suggested to have a pleiotropic effect on tissue morphogenesis and remodeling.37 Glypican-3 is a cell surface–linked heparin sulfate proteoglycan involved in organogenesis and cell signaling. Glypican-3–deficient mice exhibit developmental overgrowth and abnormalities typical of the Simpson-Golabi-Behmel syndrome.38, 39 This X-linked proteoglycan is expressed predominantly in tissues of mesodermal origin, controlling the cellular responses to Bmp440, but its expression in the developing mouse liver and lung has also been reported.38, 39 Adam19, a disintegrin and metalloproteinase domain 19 (meltrin beta), is a cell surface glycoprotein functioning in fertilization, heart development, neurogenesis, and protein ectodomain shedding.41, 42 Cathepsin L is a major excreted protein that is highly expressed in transformed cell lines with strong collagenolytic and elastinolytic activity and plays a role in epidermal morphogenesis and differentiation.43 Mago-nashi proliferation-associated homolog of Drosophila is a protein required for normal germ plasm assembly. It is a ubiquitously expressed, serum-inducible protein with no clear function.44, 45 Sema 3f—Sema domain, immunoglobulin domain, short basic domain, secreted (semaphorin) 3F—is a member of a semaphoring family involved in signaling through its respective receptor neurophilin and guiding axons and neuronal cell migration in the central nervous system.46 Its function in fetal liver is unknown.


The gene expression pattern of isolated fetal liver epithelial progenitor cells has been studied during their differentiation in vivo at ED 13, 14, 15, 16, 17, 18, newborn, 7 days after birth, and adult. The cluster of up-regulated genes comprised 281 clones showing high expression in fetal liver that gradually declined toward birth. The cluster of the down-regulated genes comprised 230 members. The expression of the latter increased dramatically between ED 16 and 17, marking an abrupt switch in the gene expression program during the transition from an undifferentiated to a differentiated state of fetal hepatoblasts. In addition, we have identified a group of fetal liver progenitor cell–specific genes, 39 of which are found to be expressed in cells of “stem origin.” Some of these differentially expressed genes could be of great importance because they may reveal the existence of novel, specific pathways of molecular signaling and regulation of growth, proliferation, survival, and differentiation in fetal hepatoblasts. These genes can also be considered as putative markers for identification of tissue-specific stem/progenitor cells in the adult liver.


The authors thank Ethel Hurston for technical assistance and Emily Bobe for typing the manuscript.