Premature Expression of the Winged Helix Transcription Factor HFH-11B in Regenerating Mouse Liver Accelerates Hepatocyte Entry into S Phase

Generation of TTR–HFH-11B transgenic mice.

The transthyretin (TTR) minigene construct (Fig. 2) consists of the −3-kb TTR promoter region, the first and second TTR exons fused to the simian virus 40 (SV40) 3′ end and poly(A) sequences (6, 55). The ATG sequence in the TTR first exon is mutated and allows insertion of the HFH-11B cDNA in the StuI site of the TTR second exon. This TTR transgene therefore provides a 5′ splice site for more reliable expression in transgenic mice (53). The 2.6-kb human HFH-11B cDNA was excised with EcoRI-HindIII, blunt ended with Klenow fragment of DNA polymerase I, and ligated into a unique StuI site located in the second exon of the pTTR-exv3 minigene vector (see Fig. 2). The 7.0-kb HindIII fragment containing the −3-kb TTR promoter driving HFH-11B cDNA expression was separated from the vector fragment by regular agarose gel electrophoresis (FMC) and then purified from the agarose by using the GeneClean II kit (Bio 101, Inc.). The transgene was injected into the pronuclei of CD1 mouse eggs at a concentration of 3 ng/μl, and transgenic mice were generated as described by Hogan et al. (24). After microinjection of the TTR–HFH-11B transgene construct into CD-1 fertilized mouse eggs and transfer to surrogate mothers, 61 mice were born and 3 of these carried the transgene as identified by PCR. Identification of mice carrying the HFH-11B transgene was performed by PCR analysis of genomic DNA extracted from the tails of 2- to 4-week-old mice. The primers used were 5′-AAAGTCCTGGATGCTGTCCGAG-3′ (sense TTR exon two 5′ primer) and 5′-CAGACATGATAAGATACATTGATG-3′ (antisense SV40 3′ primer). The 5′ and 3′ primers for an internal control thyroid stimulating hormone gene were 5′-TCC TCA AAG ATG CTC ATT AG-3′ and 5′-GTA ACT CAC TCA TGC AAA GT-3′.

Surgical procedure for mouse partial hepatectomy.

For liver regeneration studies, 10- to 12-week-old male CD-1 mice (30 to 35 g [body weight]) were anesthetized with methoxyflurane (Metofane; Schering-Plough Animal Health Corp., Union, N.J.) and subjected to midventral laparotomy with a two-third liver resection (left lateral, left median, and right median lobes) (23). Extra care was taken to avoid excision of the gallbladder situated at the cleft between the left and right median lobes. One subcutaneous injection of ampicillin (50 μg/g [body weight]) in saline was given to the animal after the surgical procedure. Two hours before the remnant liver was harvested, animals were injected intraperitoneally with 5-bromo-2′-deoxyuridine (BrdU; 50 μg/g [body weight]; 0.2% solution in phosphate-buffered saline [PBS]). Four to six mice from each group were sacrificed at 4-h intervals from 24 to 52 h and at 68 h posthepatectomy by CO₂ asphyxiation. A portion of liver tissue was used to prepare total RNA; another portion was fixed overnight in 4% paraformaldehyde (in PBS solution at pH 7.4) at 4°C, paraffin embedded the second day, and sectioned as 6- to 8-μm sections by using a microtome. The sectioned tissues were used for (i) routine microscopy with hematoxylin and eosin stain, (ii) evaluation of the fraction of hepatocytes undergoing DNA synthesis by BrdU incorporation, (iii) in situ hybridization with antisense ³³P-labeled riboprobes, and (iv) immunohistochemistry analyses with HFH-11 antibody (57).

RNA extraction and RNase One protection assay.

Total RNA was extracted from mouse liver by an acid guanidium thiocyanate-phenol-chloroform extraction method with RNA-STAT-60 (Tel-Test “B” Inc., Friendswood, Tex.). The HFH-11B antisense RNA probe was generated from the PCR-derived BamHI-EcoRI HFH-11A cDNA pGEM1 clone (409-bp fragment) that contained exon AII and sequences encoding the 58 amino acids N-terminal to exon AII (amino acids 366 to 469) as described previously (57). The TTR transgene probe was made from an EcoRI-digested pTTRExV3 template DNA with SP6 RNA polymerase (55) and detects expression of both the endogenous TTR and the transgene. The protected fragments of TTR and transgene were 90 and 310 nucleotides, respectively. The C/EBPβ antisense RNA probe was made from an EcoRI-digested rat C/EBPβ cDNA pGEM-1 template (867 to 1,392 bp) with SP6 RNA polymerase. pTRI vector containing cDNA templates for mouse p34cdc2, cyclin B1, and cyclophilin were purchased from Ambio, Inc. (Austin, Tex.). The protected fragments of p34cdc2, cyclin B1, and cyclophilin were 250, 214, and 113 nucleotides, respectively.

RNase protection assay was performed with [³²P]UTP-labeled antisense RNA as previously described (57). Approximately 1 ng of each antisense riboprobe was hybridized at 55°C to 20 μg of total RNA in a solution of 20 mM PIPES (pH 4.6), 400 mM NaCl, 1 mM EDTA, and 80% formamide for 16 h. As the internal control, 1 ng of antisense cyclophilin riboprobe was included in each hybridization reaction, except for the analysis of the TTR transgene. After hybridization, the samples were digested for 1 h at 30°C by using 5 U of RNase One and processed according to manufacturer’s protocol (Promega). Protected fragments were electrophoresed on an 8% polyacrylamide–8 M urea gel, followed by either autoradiography or scanning with a Storm 860 PhosphorImager (Molecular Dynamics). Quantitation of expression levels was determined with ImageQuant program (Molecular Dynamics) and/or with scanned X-ray films by using the BioMax 1D program (Kodak).

Immunohistochemistry and in situ hybridization.

Livers from BrdU-injected mice (50 μg/g [body weight]; 2 h before harvesting) were fixed in 4% paraformaldehyde overnight, embedded in paraffin, prepared for histological analysis, and immunohistochemically stained with an anti-BrdU monoclonal antibody according to the manufacturer’s protocol (Boehringer Mannheim) or with the HFH-11-specific antibody (57). For immunohistochemistry, the paraffin wax was removed from sections with xylenes, after which they were rehydrated in ethanol and then placed in PBS plus 0.25% Triton X-100 (PBT). We used a microwave-based antigen retrieval method to enhance the antigenic reactivity of antibodies with paraformaldehyde-fixed dewaxed paraffin-embedded sections as described previously (59). Primary antibodies were detected by using secondary anti-mouse immunoglobulin G coupled to horseradish peroxidase staining with the appropriate substrates (Vector Laboratories). To detect nuclear localization of the HFH-11 protein during mouse liver regeneration, we used the HFH-11 antibody (57) for immunohistochemistry staining of liver sections (see Fig. 2) with AEC as the peroxidase substrate (stains red), followed by counterstaining with hematoxylin (stains nuclei blue). The number of BrdU-positive hepatocytes undergoing DNA synthesis was determined by randomly counting the positive BrdU staining nuclei per 1,000 in a total of at least 3,000 hepatocytes. Mitotic figures in hepatocytes were counted and quantitated as a percentage of at least a total of 1,500 cells under 10 high-power fields at the indicated time after hepatectomy. The data from BrdU labeling experiments, mitotic figures, and fold induction of cdc2 expression levels are represented as the means ± the standard error of the mean. The Student t test was used to compare the different parameters between the two groups by using the Analysis ToolPak in Microsoft Excel 98. A P value of <0.05 was considered significant.

In situ hybridization was performed with ³³P-labeled antisense riboprobes hybridized to tissue sections and rinsed at high stringency, and hybridization signals detected by autoradiography by using procedures described by Rausa et al. (42). ³³P-labeled antisense HFH-11 riboprobes were generated from the two rat HFH-11 cDNA EcoRI-PstI pGEM1 subclones (450 bp) by using the appropriate RNA polymerase enzyme as described previously (57). After hybridization and rinsing, the in situ hybridization slides were dipped in a photographic emulsion (NTB2; 1:1 dilution with water) from Kodak. The sections were stored in a light-tight desiccated box for exposure of the in situ hybridization signals at 4°C for 2 to 3 weeks, developed, fixed, and then counterstained with standard hematoxylin and eosin solution.

Mouse cDNA array analysis.

RNA samples were isolated from liver tissues at 24, 28, 32, and 40 h posthepatectomy. ³²P-labeled cDNA probe was prepared by using mouse liver poly(A)⁺ RNA from wild-type and transgenic animals according to the user manual (Clontech, Palo Alto, Calif.). The labeled cDNA was hybridized to Atlas Mouse cDNA array membranes at 65°C overnight, and the blots were washed in 0.1× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) at 65°C. The arrays were then exposed to phosphor screens overnight and scanned with a Storm 860 PhosphorImager. Quantitation of expression levels was determined by utilizing the ImageQuant analysis program. All measurements were stored in a computer database for analysis and interpretation by using Microsoft Excel 98. Each array blot contained 588 well-characterized genes, such as cytokines and their receptors, transcription factors, apoptosis-related genes, and cell cycle regulators. Each blot also contained nine housekeeping genes for normalizing the hybridization signals.

Induced expression and nuclear translocation of HFH-11 occurs prior to DNA replication in regenerating mouse hepatocytes at 32 h post-PH.

Previous studies of regenerating rat liver demonstrated that induction of HFH-11 expression precedes hepatocyte entry into DNA replication (S phase), but it follows immediate-early gene expression at 4 h post-PH (57). HFH-11 thus exhibits the expression kinetics of a delayed-early transcription factor gene. In regenerating mouse liver, DNA replication initiates at 40 h post-PH and is delayed compared to regenerating rat liver. In order to determine the expression pattern of HFH-11 during mouse liver regeneration, we performed in situ hybridization of paraffin sections of regenerating mouse liver by using ³³P-labeled antisense HFH-11 RNA probe. In regenerating mouse liver, HFH-11 hybridization signals are barely visible by 24 h after PH (Fig. 1A and B), and its expression increases to maximal levels by 32 h post-PH (Fig. 1C to F). Furthermore, we observed more intense HFH-11 labeling in regenerating hepatocytes surrounding the periportal region (Fig. 1C to F), which are the first hepatocytes to initiate DNA replication during liver regeneration (19, 50). To detect nuclear localization of the HFH-11 protein during mouse liver regeneration, we used the HFH-11 antibody (57) for immunohistochemistry staining of liver sections (Fig. 2) with AEC as the substrate (stains red), followed by hematoxylin counterstaining (stains nuclei blue). The nuclear localization studies reveal cytoplasmic HFH-11 protein staining at 24 h post-PH (Fig. 2A and B), perinuclear HFH-11 staining by 28 h post-PH (Fig. 2C) and, ultimately, nuclear translocation of the HFH-11 protein is observed by 32 h after PH (Fig. 2D to E, red nuclei). These expression studies of regenerating mouse liver indicate that nuclear translocation of the HFH-11 protein occurs during the G₁-to-S-phase transition.

Hepatocyte nuclear localization of the HFH-11B transgene protein requires mitotic signalling induced during liver regeneration.

In order to ectopically express the transcriptionally active human HFH-11B isoform in hepatocytes (57), we generated transgenic mice by using the −3-kb TTR minigene construct (Fig. 3A), which effectively targets transgene expression to hepatocytes (53, 55). Three founder transgenic mice were obtained (T38, T69, and T70) that expressed the HFH-11B transgene at different levels (Fig. 3B). None of the transgenic lines exhibited aberrant hepatocyte proliferation or defects in the liver architecture, nor was there a significant difference in liver/body weight ratio (data not shown). Furthermore, liver function was not affected in the TTR–HFH-11B transgenic mice as assessed by determining the normal serum levels of glucose, alkaline phosphatase, albumin, liver aminotransferase enzymes, and bilirubin (Table 1). We used the T38 TTR–HFH-11B transgenic line for further liver regeneration studies because it displayed the highest ectopic expression of the HFH-11B transgene (Fig. 3) and its expression was observed throughout the parenchyma of nonregenerating transgenic liver, albeit, it exhibited a more intense periportal expression (Fig. 1G and H). Sustained expression of the HFH-11B transgene was also observed throughout the parenchymal cells in regenerating livers of transgenic mice (Fig. 1I and J).

Because aberrant proliferation was not observed in nonregenerating hepatocytes of the T38 transgenic mouse line, we subjected transgenic mice to two-thirds PH and examined hepatocyte subcellular localization of the HFH-11B protein by immunohistochemical staining by using the AEC substrate to detect the HFH-11 antibody and nuclear counterstaining with hematoxylin. Consistent with the absence of aberrant hepatocyte proliferation in nonregenerating transgenic livers, the HFH-11B transgene protein was diffusely distributed throughout the cytoplasm of hepatocytes (Fig. 2F). In contrast, perinuclear and nuclear staining of the HFH-11B protein was observed by 16 h post-PH and the transgene protein becomes localized to hepatocyte nuclei by 24 h post-PH (Fig. 2G, blue-red nuclei). Nuclear localization of the HFH-11B protein was sustained in hepatocytes throughout the period of proliferation (Fig. 2I to K, red nuclei). These studies demonstrated that mitotic signalling induced during liver regeneration is required for hepatocyte nuclear localization of the HFH-11B transgene protein. Furthermore, premature availability of the HFH-11B transgene protein in regenerating liver allowed its nuclear localization to occur at 24 h post-PH, which is 8 h earlier than the time observed with regenerating wild-type liver (Fig. 2, compare panels D and H).

Premature hepatocyte expression of HFH-11B accelerates the onset of DNA synthesis and mitosis in regenerating transgenic mouse liver.

To determine whether earlier nuclear translocation of the HFH-11B protein in regenerating liver would alter the timing of hepatocyte DNA replication, liver regeneration studies were performed with wild-type and TTR–HFH-11B transgenic mice, and DNA synthesis was monitored by immunohistochemical staining of BrdU incorporation into DNA (8, 54). We used three to six regenerating livers per time point, and the results of the BrdU labeling studies are shown graphically in Fig. 4G. Consistent with previous reports (8, 14, 21, 44, 54), only a few of the wild-type hepatocytes exhibit BrdU incorporation at 36 h post-PH, and hepatocyte replication reaches a maximum by 40 h and is diminished by 48 h after surgery (Fig. 4A, C, and E). In contrast, maximal hepatocyte DNA replication in regenerating transgenic mouse liver occurs at 32 h post-PH (Fig. 4B, D, F, and G), a result consistent with an 8-h-earlier nuclear translocation of the HFH-11B transgene protein (Fig. 2). Consistent with regenerating wild-type livers (19, 50), the initiation of hepatocyte proliferation is restricted to the periportal region of regenerating transgenic livers but becomes more diffuse at later time points (Fig. 4D to F). Furthermore, initiation of mitosis occurred 8 h earlier in regenerating transgenic hepatocytes, as evidenced by an increase in the number of mitotic bodies present at 40 h post-PH (Fig. 4H). The total numbers of hepatocytes undergoing DNA synthesis and mitosis, however, were not significantly different between transgenic and wild-type regenerating livers (Fig. 4G and H). These studies suggest that earlier nuclear localization of the HFH-11B protein accelerates the timing of hepatocyte entry into DNA replication and mitosis.

Earlier expression of DNA repair enzymes and cyclin genes in regenerating TTR–HFH-11B transgenic mouse livers.

To identify hepatocyte proliferation genes that are differentially expressed in regenerating transgenic and wild-type mouse livers, Atlas Expression cDNA Array Blots (Clontech) were hybridized with radioactive cDNA prepared from various stages of regenerating mouse liver (24, 32, and 40) h post-PH; see Materials and Methods). The Atlas Expression cDNA Array Blot contains 588 distinct cDNAs spotted in duplicate and organized in quadrants containing genes participating in similar cellular pathways. In regenerating transgenic mouse liver, a threefold increase in expression of DNA repair genes was observed by 24 h post-PH (Fig. 5, cDNAs 1 to 3). This level of increase is normally induced during hepatocyte DNA replication at 40 h post-PH. These DNA repair genes included the double-stranded break repair genes XRCC1 and mHR21spA (mouse homolog of yeast rad21 gene) genes and the nucleotide excision repair mHR23B (mouse homolog of yeast rad23 gene) gene. The earlier induction of DNA repair genes in the TTR–HFH-11B regenerating livers is consistent with the accelerated hepatocyte DNA replication phenotype.

In agreement with premature DNA replication and mitosis in the regenerating transgenic hepatocytes, our cDNA array analysis reveals an earlier induction of cyclin genes (Fig. 5). In regenerating transgenic livers, we observed elevated expression of cyclin D3 and p58/GTA (galactosyltransferase-associated protein kinase or cdc2-related protein kinase) by 24 h post-PH, and their increased expression is sustained during transgenic hepatocyte DNA replication (Fig. 5, cDNAs 4 and 5). Compared to the same time period of wild-type liver regeneration, S phase-promoting cyclin D1 hybridization signals are elevated at the peak of transgenic hepatocyte replication at 32 h post-PH (Fig. 5, cDNA 6). A more detailed RNase protection assay demonstrates that cyclin D1 expression exhibits a biphasic induction profile in regenerating wild-type liver, with a second peak of expression during DNA replication at 40 h post-PH (Fig. 6A, cyclin D1 WT). By contrast, regenerating transgenic liver displays a more protracted expression of cyclin D1 between 28 and 36 h post-PH and therefore prolongs the induction of cyclin D1 through transgenic hepatocyte DNA replication (Fig. 6A, cyclin D1 Tg). We also observed earlier induction of M phase promoting cyclin B1 and B2 at 32 h post-PH (Fig. 5, cDNAs 7 and 8), in that these factors are normally induced during hepatocyte DNA replication at 40 h post-PH (44, 48). RNase protection assays also confirm this 8-h-earlier induction of cyclin B1 and cyclin-dependent kinase p34cdc-2 (cdc2) expression in regenerating transgenic livers (Fig. 6A and B). Analysis of cDNA expression arrays also showed, in comparison to regenerating wild-type livers, increased expression of cyclin A, A1, G, and G2 in regenerating transgenic livers at 24 h post-PH (Fig. 5 and data not shown). However, we observed no differences in cyclin D2, E, or F expression between regenerating wild-type and transgenic livers (data not shown). These studies demonstrate that the early onset of DNA replication and mitosis in regenerating transgenic livers correlates with premature induction of DNA repair and cyclin gene expression.

A protracted increase in C/EBPβ expression is observed prior to transgenic hepatocyte DNA replication.

Liver regeneration studies with C/EBPβ null mice demonstrate a 75% reduction in hepatocyte DNA replication and reduced expression of proliferation and metabolic homeostasis genes (21). Because C/EBPβ plays an important role in mediating hepatocyte replication after partial PH, we examined its expression pattern in regenerating wild-type and transgenic livers. Regenerating wild-type livers exhibit a biphasic expression pattern with a sharp increase at 24 h post-PH and a second increase during hepatocyte replication at 40 h after PH (Fig. 6C). A more protracted induction of C/EBPβ expression was observed prior to transgenic hepatocyte DNA replication (24 and 28 h post-PH), and elevated C/EBPβ mRNA levels were maintained throughout the period of hepatocyte replication (Fig. 6C). The C/EBPβ expression profile was similar to that observed with cyclin D1, suggesting that their protracted expression contributes to the early onset of hepatocyte DNA replication in regenerating transgenic liver.