Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The process of liver regeneration after partial hepatectomy (PH)² provides a well characterized in vivo model of events related to cellular proliferation (Michalopoulos and DeFrances, 1996; Lambotte et al., 1997; Arora et al., 1998). PH is widely used to induce cell proliferation because it is fast and easy to perform, well tolerated, delivers a quantifiable stimulus, and is free of side effects associated with toxic regenerative stimuli (Bucher, 1995). Seventy percent PH is followed by rapid hyperplasia that accounts for a completely regenerated liver in 7 to 10 days (Higgins and Anderson, 1931).

The present studies were initiated to determine the role of the tumor suppressor gene p53 in cell proliferation and to elucidate its influence on the functional recovery of the drug metabolism capacity of the liver during the course of regeneration. The expression of p53 is highly up-regulated in the regenerating liver after PH (Kren et al., 1996) whereas expression of cytochrome P-450s (CYPs) is suppressed (Arora et al., 1998). A relationship between p53 and the CYP enzymes is suggested from several lines of evidence. Genetic polymorphism studies in patients with hepatocellular carcinoma show interaction between p53 polymorphic forms and CYP1A1 (Yu et al., 1999) whereas CYP2E1 polymorphism is associated with p53 mutations that predispose individuals to squamous cell carcinomas (Oyama et al., 1997). Studies done with p53-knockout mice indicate that CYP3A continues to be expressed for at least 2 weeks in primary hepatocyte cultures from these mice in contrast to homozygous primary cultures, where it rapidly disappears (Shimoji et al., 1996). Therefore, the expression of at least three CYP isoforms may be functionally linked to wild-type p53 expression.

The importance of p53 in cell proliferation is apparent from the observation that more than half of all human cancers display mutations of this gene (Levine, 1997). Roles for p53 have been postulated in G₁-S cell cycle checkpoint activity, apoptosis, differentiation, and DNA repair (Magnelli et al., 1992). G₂-M checkpoint activity has also been reported, the loss of which can lead to polyploidy (Waldman et al., 1996). p53 has been shown to directly regulate the activity of several cell cycle-related genes, including p21^waf-1, which is an inhibitor of cyclin-dependent kinases responsible for propelling cells through G₁ and S phases of cell cycle (El-Deiry, 1998). p21^waf-1 may in turn regulate the activity of proliferating cell nuclear antigen (PCNA), which is a component of the DNA polymerase complex (Chen et al., 1995).

In the present studies, the role of p53 in the regenerating liver was examined by a transient inhibition of its induction by administration of antisense oligodeoxynucleotides (ODNs). The antisense ODN sequence was named `p53T'. The control ODN was named OL(1)p53, the sequence of which is antisense to human p53 and has four mismatches with the rat sequence. Previous work in this and other laboratories has demonstrated that nuclease-resistant phosphorothioate ODNs (PS-ODNs) can be used to suppress specific gene expression in vivo (Desjardins and Iversen, 1995; Crooke, 1997; Arora et al., 1998). Pharmacokinetic data show that PS-ODNs are widely distributed but liver is a major site of accumulation (Iversen et al., 1994; Agrawal and Zhang, 1997). Thus, antisense PS-ODNs targeted at p53 mRNA present a unique opportunity for transient inhibition of the p53 gene during liver regeneration. Additionally, this study also uses cytosine C-5 propynyl (C-5-P)-modified PS-ODNs (two cytosine bases close to the 3' and 5' ends of the PS-ODN are C-5-P modified, (see Materials and Methods) and morpholine phosphorodiamidite oligomers (all riboside moieties replaced with morpholine moieties). The former analogs are thought to improve binding efficiency through better stacking interaction between bases (Wagner et al., 1996), whereas the later are uncharged molecules, appear to have less nonspecific effects, and work through a non-RNase H mechanism (Summerton et al., 1997; Giles et al., 1999).

These studies test the hypothesis that induction of p53 during liver regeneration serves as a negative regulator of proliferation by providing cell cycle checkpoint activity. Our data show that antisense suppression of p53 induction in the regenerating rat liver causes an alteration in its phenotype, resulting in loss of G₁-S cell cycle check- point activity while enhancing the weight gain of the regenerating liver, PCNA expression, p21^waf-1 expression, and functional recovery of CYP enzymes. We conclude that our hypothesis is true. The present study also highlights the use of antisense ODNs in the regenerating liver as an excellent and reproducible model for in vivo studies of cell proliferation.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals. The study was performed on male Sprague-Dawley rats (Sasco, Omaha, NE) weighing between 200 and 225 g. The animals were housed in clear plastic cages in the University of Nebraska Medical Center Association for Assessment and Accreditation of Laboratory Animal Care-approved facility with a 12-h light/dark cycle and allowed access to Purina rat chow and tap water ad libitum. All animal protocols conformed to the ethical guidelines of the 1975 Declaration of Helsinki and were approved by the Institutional Animal Care and Use Committee of the University of Nebraska Medical Center.

PH. The procedure was performed as described previously (Higgins and Anderson, 1931). Aseptic surgical techniques were used. Rats were anesthetized using methoxyflurane (Mallinckrodt Veterinary, Mundelein, IL) and positioned with their ventral surface exposed. A 3- to 4-cm-long area was shaved along the median line immediately posterior to the xiphoid process and swabbed with betadine. An incision was made along the median line to expose the liver. The medial and left lateral lobes were securely ligated and then excised. This resulted in removal of approximately 65 to 70% of the total liver. The abdominal incision was closed in two layers. Control rats that did not receive a PH underwent an identical but sham surgery; their livers were only exposed and not partly excised.

Oligonucleotide Administration. An 18-mer antisense PS-ODN with a sequence complementary to a unique portion of rat p53 mRNA (residues 1182-1199; Genbank accession number X13058) was synthesized. This molecule was named p53T and has the following sequence: 5'-TCA GTC TGA GTC AGG CCC-3'. The control PS-ODN is antisense to human p53 and does not alter p53 expression in rats as it has four mismatches with the rat sequence. It is a 20-mer named OL(1)p53 and has the following sequence: 5'-CCC TGC TCC CCC CTG GCT CC-3'. Both of these oligonucleotides have a phosphorothioate backbone. The PS-ODNs were injected i.p. at the dose of 5 mg/kg/day. The C-5-P-modified PS-ODNs and the morpholine oligomers were injected i.p. at the dose of 0.5 mg/kg/day. All rats were administered respective oligonucleotides immediately on regaining consciousness after the surgery, and every 24 h thereafter, depending on the length of the experiment. Rats were then allowed to recuperate for 1, 2, 5, or 7 days from the time of surgery.

Oligonucleotide Synthesis. All chain-extension syntheses were performed on an Applied Biosystems (ABI) model 380B DNA synthesizer (Foster City, CA) using 1-µmol column support and the cyanoethyl approach using phosphoramidate chemistry as described by ABI User Bulletin, 58, 1991. The synthesizer was programmed with the protocol in the user's manual. The ammonium hydroxide and cyanoethyl (from the phosphate links) were removed by evaporation of the PS-ODN solution in a vacuum centrifuge overnight. The dried PS-ODN was diluted in sterile saline. Purity of the PS-ODN was checked by running a sample of the PS-ODN diluted in water on a 10% polyacrylamide gel. The concentration of PS-ODN was determined by reading the absorbance at 260 nm and multiplying the absorbance by its extinction coefficient. The C-5-P cytosine-modified PS-ODN syntheses used phosphoramidate reagent from Glen Research (Sterling, VA), replacing the cytosine residues in bold type: 5'-TCAGTCTGAGTCAGGCCC-3' of p53T and 5'-CCCTGCTCCCCCCTGGCTCC-3' of OL(1)p53. The morpholine oligomers were prepared with the same sequences as p53T and OL(1)p53 and were obtained from AVI Biopharma (Corvallis, OR; www.avibio.com). Purity of morpholine ODNs was determined by reversed phase HPLC and MALDI TOF mass spectroscopy.

p53, PCNA, p21, and NADPH Reductase Levels. Levels of all three proteins were determined by Western blots with homogenates of remnant livers using standard techniques. A 10% v/v SDS/acrylamide gel with a 6% SDS/acrylamide stacking gel on top was prepared. Each sample was prepared by mixing 0.05 mg of homogenate protein in 0.01 ml SDS and 5% 2-mercaptoethanol and loaded onto the gel. All reagents for Western blot were obtained from Sigma. The gel had 20 mA constant current passed through it using a model 3000Xi electrophoresis power supply (BioRad) until the tracking dye migrated to the running gel. The current was then increased to 30 mA until the tracking dye migrated off the gel. The gels were then soaked in transfer buffer (192 mM glycine, 25 mM Tris base, and 10% methanol, pH = 8.3) for 20 min. The protein was then transferred from the gel to Immobilon-P transfer membranes (Millipore) at 480 mV for 45 min. The membranes were then soaked with blocking buffer (20 mM Tris base, 150 mM NaCl, 1% BSA, and 0.2% Tween 20) for 1 h at 37°C. The membranes were incubated at 4°C overnight with primary antibodies diluted 1:500. Monoclonal primary antibodies for p53 (Ab-1), PCNA (Ab-1), and p21^waf-1 (Ab-4) were obtained from Oncogene Research Products (Cambridge, MA). Primary antibodies and protein standard for NADPH reductase were purchased from Gentest (Woburn, MA). The membranes were washed repeatedly with blocking buffer then incubated for 30 min at 4°C with appropriate secondary antibody diluted 1:1000 (anti-rabbit IgG for p53, anti-mouse IgG for PCNA and p21^waf-1, anti-goat IgG for NADPH reductase) conjugated with horseradish peroxidase. The membranes were washed repeatedly with blocking buffer and then incubated for 1 min with ECL Western Blot Reagents (Amersham). The membranes were exposed to Kodak film for 15 to 30 s and the film was developed. Band intensities were determined by a Molecular Dynamics Personal Densitometer (Sunnyvale, CA) using ImageQuant version 3.3 software (Molecular Dynamics).

Isolation of Hepatocytes and Flow Cytometric Analysis of DNA Content. Individual hepatocytes were isolated from the whole liver by perfusing rat livers through the portal vein with 200 ml of flush buffer (8.3 g/liter NaCl, 0.5 g/liter KCl, 2.4 g/liter HEPES, and 10 mg/liter soybean trypsin inhibitor) adjusted to pH 7.4. This was followed by perfusing the liver with 150 ml of type IV collagenase (Sigma) solution at a concentration of 40 mg/100 ml. The liver capsule was then peeled and the liver tissue was gently shaken in cold PBS to liberate isolated hepatocytes. These were then filtered through 100- and 40-µm mesh successively. The cell cycle distribution of the isolated liver cells was done by the method of Telford (Fraker et al., 1995). Cells (1 × 10⁶) were suspended in 80% cold ethanol and incubated on ice for 2 h. After removal of the ethanol, cells were resuspended in 500 µl of Telford staining reagent [33 µg/ml disodium EDTA (Sigma), 124 U RNase A (93 U/mg; Sigma), 50 µg/ml propidium iodide, and 1 µl/ml Triton X-100 (Sigma) in 1× PBS] and were allowed to incubate overnight at 4°C. Cells were passed through a 40-µm filter a second time immediately before flow cytometer analysis on a FACStar Plus machine.

Microsomal Isolation. Hepatic microsomes were prepared by a modification of the method of Franklin and Estabrook (1971). Livers were perfused immediately with 10 ml of ice-cold saline via the portal vein, minced, homogenized in 10 ml 0.25 M sucrose (Sigma), and centrifuged at 8000 relative centrifugal force (rcf) for 20 min at 4°C in a Sorvall RC2-B centrifuge (DuPont, Wilmington, DE). The supernatants were centrifuged at 100,000 rcf for 45 min at 4°C in a Sorvall OTD55B ultracentrifuge (DuPont). The pellets were resuspended in 20 ml of 0.15 M KCl (Sigma) and centrifuged at 100,000 rcf for 45 min at 4°C. The final pellets were resuspended in an equivalent volume of buffer (10 mM Tris-acetate, 1 mM EDTA, 20% glycerol; Sigma; pH = 7.4) and stored at 80°C.

Measurement of CYP2E1. The activity of CYP2E1 in the microsomes was measured using p-nitrophenol hydroxylase (PNP) assay (Koop, 1986). An aliquot of 1 mg of microsomal protein was mixed with 0.01 ml of 20 mM p-nitrophenol (Sigma), and 0.05 ml of 60 mM NADPH (Sigma) and diluted to 1 ml with 0.1 M potassium phosphate buffer (pH = 7.4). The samples were incubated for 10 min at 37°C, mixed with 0.25 ml of 15% trichloroacetic acid (Sigma), and were centrifuged at 15,000 rcf for 4 min in a Hermle Z 230 M centrifuge (National Labnet Co., Woodsbridge, NJ). A 1-ml sample was added to a new set of tubes containing 0.1 ml of 10 N NaOH and mixed for 10 s. Each sample was then read at 515 nm on a Ultraspec III (Pharmacia). Enzyme activity was recorded as optical density per milligram protein per minute.

Measurement of Activities of CYP1A1 and 2B2. The activities of microsomal CYP1A1 and 2B2 were determined by using ethoxyresorufin-o-dealkylation assay (EROD) or pentoxyresorufin-o-dealkylation assay (PROD), respectively (Burke et al., 1985). An aliquot of 1 mg of microsomal protein in 1 ml of 0.1 M potassium phosphate buffer (pH = 7.4), 1 ml of 2 µM 5-ethoxy- or 5-pentoxyresorufin (Pierce, Rockford, IL), and 0.02 ml of 60 mM NADPH were mixed and incubated for 10 min at 37°C. The mixture was then added to a 2-ml cuvette and read on a RF5000U spectrofluorophotometer (Shimadzu, Columbia, MD) using an excitation wavelength of 530 nm and an emission wavelength of 585 nm. Concentrations of unknowns were calculated from a standard curve of resorufin (Pierce) standards. Results were recorded in picomoles resorufin per milligram protein per minute.

Measurement of CYP3A2. The activity of CYP3A2 was measured by using erythromycin N-demethylation (ERDEM) (Wrighton et al., 1985). The samples were prepared by mixing 1.0 mg of microsomes, 0.4 mM erythromycin, and 1.0 mM NADPH in a final volume of 1 ml in 0.1 M potassium phosphate buffer (pH = 7.4). The samples were incubated for 15 min at 37°C, mixed with 0.5 ml 17% perchloric acid (Sigma), and centrifuged on a Hermle microcentrifuge at 15,000 rcf for 5 min. Formaldehyde was measured by the colorimetric method of Nash (1953). The samples were placed in a new tube and mixed with 0.4 ml of Nash reagent (0.02 M 2,4 pentanedione, 0.6% v/v glacial acetic acid, and 3.9 M ammonium acetate) and incubated at 70°C for 20 min. The samples were read on the spectrophotometer at 412 nm. Absorbencies were compared with a standard curve generated from known concentrations of formaldehyde. Activities were recorded as micromoles formaldehyde per milligram protein per minute.

Measurement of Thiobarbituric Acid-Reactive Substances (TBARS). The changes in lipid peroxidation are a direct indicator of oxidative stress. These changes in lipid peroxidation were measured in liver homogenates as published (Esterbauer and Cheeseman, 1990). Proteins in the liver homogenate samples were precipitated with an equal volume of 10% trichloroacetic acid. After centrifugation at 1000 rpm for 10 min, 1 ml of the supernatant was incubated with 1 ml of 0.67% thiobarbituric acid (TBA, Sigma) at 100°C for 10 min. Lipid peroxides react with TBA (collectively termed TBARS) to produce a colored TBA-malondialdehyde product, which was determined at an absorbance maximum of 532. The background was corrected by applying the Allen correction: subtracting the average absorbance at 508 and 556 nm from the absorbance at 532 nm (Allen, 1950; Sunderman et al., 1985). A standard curve was generated using various concentrations of malondialdehyde (Sigma) and 0.67% TBA with a similar protocol. All experimental results were adjusted using negative controls (PBS replacing liver homogenate in reaction mixture).

Histology. Sections of liver were prepared consistently from the right lateral lobe to minimize error resulting from differential regeneration in different lobes. Liver samples were fixed in 10% formalin, embedded in paraffin, and sectioned in the Eppley Institute histology core facility. Sections were stained with hematoxylin and eosin, and photographed on a Nikon microscope with Fuji film.

Mitotic Index (MI). Hematoxilyn and eosin stained histology sections of the right lateral lobe of liver were examined under 200-fold magnification using a Nikon light microscope. The number of mitotic cells in a single field was determined based on a darker stain in the nucleus, resulting from increased chromatin content. Ten random fields were examined. The average number of mitotic cells were divided by the average number of total cells in each field and converted to a percentage. To eliminate the factor of count bias, the same slides were examined by a second investigator in a blinded manner using the same protocol. The data was pooled.

Statistical Analysis. All data were reported as mean ± S.E. as determined by the computer program InStat2 (GraphPad, San Diego). The P values were also calculated by InStat2 using Tukey's multiple comparison test and Student's t test. P values of <.05 were designated with (*) and those of <.005 were designated with (**). Standard curves and graphs were generated using Prism v2.0 (GraphPad). Prism software was also used to calculate linear regression, slope, correlation coefficient, and to test if slope of a line was significantly different from zero.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Induction of p53 Can Be Suppressed by Antisense PS-ODNs. p53 levels were evaluated in homogenates of rat livers with Western blot analysis using a monoclonal antibody. The bands on these blots were subjected to densitometry to get a quantitative estimate of variation in p53 levels. The results indicate that p53 levels are undetectable or extremely low in liver lysates of saline-treated sham PH rats (Fig. 1). Treatment of the sham PH rats with the control PS-ODN, OL(1)p53, or the test PS-ODN, p53T, did not cause any significant change in p53 expression. There was a 30- to 35-fold induction in liver p53 levels 24 h after PH in both saline- and OL(1)p53-treated rats. An induction of 42- to 50-fold was observed in p53 expression in liver lysates of saline/OL(1)p53-treated rats 48 h after PH. The treatment of rats with 5 mg/kg of p53 antisense PS-ODN, p53T, limited the p53 induction to 5- to 9-fold at the 24-h recovery time point and to 17- to 22-fold at 48 h post hepatectomy. Hence, inhibition of p53 in vivo is PS-ODN sequence-specific after 5 mg/kg/day i.p. administration.

View larger version (52K): [in this window] [in a new window]   Fig. 1.   Effect of saline, OL(1)p53, or p53T treatment on p53, PCNA, and NADPH reductase levels in rat liver tissue lysates. Rats were administered PS-ODNs at the dose of 5 mg/kg/day immediately after surgery. Representative immunoblots indicate the comparative levels of these proteins in sham hepatectomy rats and those in 24 h and 48 h after PH. The numbers at the bottom of the immunoblots indicate densitometric analysis of the representative blots in arbitrary units.

Decreased p53 Levels Alter Liver Regeneration Patterns. The wet weight of the remnant regenerating liver was used to assess the phenotypic degree of regeneration after 70% PH, in addition to the MI and the expression of PCNA. There was an increase in weight gain of the remnant-regenerating livers of the p53T-treated over OL(1)p53-treated rats both at 24 and 48 h after PH (Table 1). OL(1)p53-treated rats had remnant liver weights of 3.6 ± 0.1 g compared with 4.2 ± 0.1 g for the p53T-treated rats 24 h post PH (n = 9, P < .05). A similar but larger difference was observed when the animals were allowed to recuperate for 48 h post PH. OL(1)p53-treated rats had remnant liver weights of 4.2 ± 0.1 g compared with 4.9 ± 0.1 g for the p53T-treated rats 48 h post PH (n = 7, P < .005). There was no statistical difference in liver weights between saline- and OL(1)p53-treated animals at any of the time points studied, suggesting the inert nature of this PS-ODN in the rat liver regeneration model.

Inhibition of p53 in Regenerating Liver Is Associated with Loss of G₁-S Cell Cycle Checkpoint Activity. DNA content of isolated hepatocyte populations from the rat livers was determined by flow cytometry 24 h post PH after saline, OL(1)p53, and p53T treatments (n = 3 for each group). Data from representative OL(1)p53- and p53T-treated rats are presented in Fig. 2. Both OL(1)p53- and p53T-treated rats showed similar amounts of cell debris (DNA content <200) resulting from cell isolation procedure and suggesting uniformity in sample preparation. The major difference between OL(1)p53- and p53T-treated rats was observed in G₁ cell populations (DNA content ~200). There was about 5-fold reduction in this population in p53T-treated rat livers compared with OL(1)p53, suggesting a loss of G₁-S cell cycle checkpoint activity in the former. There was no measurable change in S or G₂ populations resulting from the release from G₁ in the p53T-treated rats. However, the p53T-treated 24 h PH rat livers displayed a greater number of polyploid cells (DNA content > 400) than OL(1)p53-treated controls consistent with p53 participation in the spindle apparatus.

View larger version (73K): [in this window] [in a new window]   Fig. 2.   Changes in cell cycle distribution of regenerating rat livers. DNA content of propidium iodide stained isolated liver cells from 5 mg/kg/day PS-ODNs OL(1)p53- and p53T-treated rats 24 h after PH (A and B show representative plots, respectively). The x-axis is DNA content in arbitrary units such that 200 would be equivalent to 2n DNA (G0/G1 phase of cell cycle), 400 would be equivalent to 4n DNA (G2/M phase of cell cycle), and so on. The y-axis represents cell counts, which remain consistent in a given session. Hematoxylin and eosin-stained histology sections from 24 h PH rats treated with OL(1)p53 and p53T are shown in (C) and (D), respectively. The darker stained nuclei indicate condensed chromatin of possible aneuploid cells.

Evaluation of Alternate Oligonucleotide Chemistries on p53 Suppression. To address the emerging concern that the thiol-rich PS-ODNs may be altering oxidative stress, as well as to further characterize the use of oligonucleotides in the liver regeneration model, two alternate oligonucleotide chemistries were also tested. The neutral backbone morpholine versions and C-5-P cytosine modified PS-ODNs of OL(1)p53 and p53T were compared with PS-ODNs with respect to wet weight gain of the remnant-regenerating livers 24 h post PH (Fig. 3). Both of the newer ODN analogs were administered at a lower dose of 0.5 mg/kg/day. The p53T sequence was active in the liver regeneration model and OL(1)p53 remained inactive with both of the alternate chemistries at the doses tested. Efficacy similar to 5 mg/kg/day dose of PS-ODN p53T was observed with 0.5 mg/kg/day of the C-5-P modified PS-ODN (n = 3) and 0.5 mg/kg/day of the morpholine p53T (n = 3), indicating that all the alternate chemistries are not equivalent in their in vivo potency.

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Remnant liver weights 24 h after PH of rats treated with three different chemistries of OL(1)p53 (open columns) and p53T (filled columns). The dose of PS-ODNs was 5 mg/kg/day whereas that of C-5-P PS-ODNs and morpholine ODNs were 0.5 mg/kg/day. *, indicates a statistically significant difference of P < .05 compared with control of similar oligonucleotide chemistry as determined by Student's t test.

Effect of p53T on Liver Regeneration Is Dose-Dependent. The 5 mg/kg/day dose for PS-ODN has been determined previously to be suitable in our laboratory for suppression of specific gene expression in the liver (Desjardins and Iversen, 1995; Arora et al., 1998). A dose response for C-5-P-modified PS-ODNs OL(1)p53 and p53T was determined in the present study in the liver regeneration model to further characterize the specificity of p53T on the suppression of p53 expression. Wet weight gain of the remnant regenerating livers was studied 24 h post PH after i.p. injections of the C-5-P versions of the PS-ODNs OL(1)p53 and p53T at the doses of 0.05, 0.5, and 5 mg/kg/day (Fig. 4). The data suggest that the weight gain response in the regenerating liver increases with increasing doses of p53T whereas OL(1)p53 has no significant effect in the tested range. Similar observations with respect to dose response of p53T were made with MIs in the regenerating livers of the same animals.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Effect of treatment of rats with increasing doses of C-5-P-modified PS-ODNs OL(1)p53 (open columns) and p53T (filled columns) 24 h after PH. Three different doses of 0.05, 0.5, and 5 mg/kg/day were used. A, represents the wet weight of the remnant regenerating livers. B, represents MIs (%). * and **, indicate statistically significant differences of P < .05 and P < .005, respectively, at similar doses as determined by Student's t test.

Long-Term Inhibition of p53 during Liver Regeneration. Longer term studies of liver regeneration were conducted for up to 7 days after PH to determine whether the effect of p53 suppression can be sustained (Fig. 5). Because the efficacy of 0.5 mg/kg/day C-5-P-modified p53T was found to be similar to 5 mg/kg/day PS-ODN p53T, the former dose was used for the longer term studies. After PH, 0.5 mg/kg/day dose of C-5-P-modified OL(1)p53 and p53T were administered i.p. OL(1)p53 and p53T-treated rats had remnant liver weights of 3.6 ± 0.2 and 4.2 ± 0.3 g at 1 day post PH and a significant weight difference was observed at 2, 5, and 7 days post PH. No apparent toxicity was observed at these time points, indicating the acute suppression of p53 is not problematic.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Long-term studies with C-5-P-modified PS-ODNs. Remnant liver weights of rats were measured on days 1, 2, 5, and 7 after PH and treatments with saline (), OL(1)p53 (), or p53T (). Both OL(1)p53 and p53T were C-5-P modified PS-ODNs administered at the dose of 0.5 mg/kg/day. *, indicates a statistically significant difference of P < .05 with respect to the saline-treated group as determined by Tukey's multiple comparison test.

Downstream Effects of p53 in Regenerating Liver Are Mediated by p21^waf-1. p21^waf-1 binds and is a potent inhibitor of the cyclin-dependent kinases, which mediate the downstream cell cycle checkpoint activity attributed to p53 (El-Deiry, 1998). To characterize the role of this inhibitor of cyclin-dependent kinases in p53 signal transduction in the liver regeneration model, the levels of this protein were assayed in liver lysates by immunoblot analysis (Fig. 6). These data indicate p21^waf-1 levels fluctuate hand in hand with the p53 levels. p21^waf-1 levels are up-regulated 24 and 48 h after PH, and this up-regulation is suppressed in p53T-treated rats along with loss of G₁-S cell cycle checkpoint activity. The higher molecular weight bands in this immunoblot are nonspecific and attributable to the secondary antibody.

View larger version (38K): [in this window] [in a new window]   Fig. 6.   p21waf-1 immunoblot analysis in regenerating livers. Representative Western blot of p21waf-1 protein levels in liver tissue lysates of sham hepatectomy and 24 h PH rats treated with saline, OL(1)p53, or p53T. The dose of PS-ODNs was 5 mg/kg/day.

Decreased p53 Levels Are Associated with Greater Oxidative Stress in the Regenerating Liver. The general degree of oxidative stress in the livers was assessed by a direct measurement of lipid peroxidation in lysates of the livers. Lipid peroxidation was measured by quantitating TBARS. The basal TBARS level in liver lysates of sham PH saline-treated rats was 3.8 ± 0.8 pM/mg lysate protein (Fig. 7). There were no statistically significant differences in TBARS levels in sham PH rats from the saline vehicle, OL(1)p53, or p53T treatment. The stress of PH is associated with increased TBARS levels (15.8 ± 2.2 pM/mg for the 24 h PH saline-treated group). However, p53T treatment after PH resulted in a additional elevation of TBARS to 39.2 ± 2.6 (P < .05) 24 h after PH. Whereas the saline-treated rats maintained their TBARS levels 48 h post PH, p53T treatment caused an even greater increase in TBARS to 47.4 ± 3.0. 

View larger version (19K): [in this window] [in a new window]   Fig. 7.   Oxidative stress in regenerating rat livers. TBARS in pM/mg of liver tissue lysate were determined in sham hepatectomy, 24 h PH, and 48 h PH rats treated with saline , OL(1)p53 (open columns), or p53T (filled columns). The dose of PS-ODNs was 5 mg/kg/day. **, indicates statistically significant difference of P < .005 with respect to saline-treated rats in the same group as determined by Tukey's multiple comparison test.

Decreased p53 Levels Prevent Loss of CYP Isoform Activities. The relevant functional capacity of regenerating liver samples was assessed in microsomal preparations of the remnant-regenerating livers. CYP 1A1/2 and 2B1/2 enzyme activities were determined by ethoxy and pentoxyresorufin O-dealkylase activities (Fig. 8A: EROD/Fig. 8B: PROD), respectively. Activities were recorded in picomoles of resorufin per milligram of microsomal protein per minute. CYP 2E1 was determined by p-nitrophenol hydroxylase (Fig. 8C: PNP) activity. Activity was recorded as optical density (A) per mg of microsomal protein per minute. Erythromycin demethylase (Fig. 8D: ERDEM) was the measure of CYP 3A2 activity and was recorded as micromoles of formaldehyde per milligram of microsome protein per minute.

View larger version (19K): [in this window] [in a new window]   Fig. 8.   Levels of various CYP enzyme activities in microsomal preparations of remnant-regenerating livers on days 1, 2, 5, and 7 after PH. C-5-P cytosine-modified OL(1)p53 () and p53T () PS-ODN analogs were used for the experiment at the dose of 0.5 mg/kg/day. A and B, represent CYP1A1 and 2B1 activities, respectively, in pM resorufin/mg protein/min. C, represents CYP2E1 activity in optical density units per milligram protein per minute. D, represents CYP3A2 activity in µM formaldehyde per milligram protein per minute. All statistical comparisons are between matching saline ()- and p53T-treated groups. *, indicates P < .05 between saline- and p53T-treated groups as determined by Student's t test. The broken lines are the mean of respective enzyme activities in untreated nonPH rat livers for comparative purposes. See Materials and Methods for description of assays.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The regenerating rat liver is an excellent model system to study the mechanisms of growth control and cellular proliferation within a normal tissue environment. The value of studying the phenomenon of liver regeneration lies in its potential value to provide detailed molecular insights into the process of neoplastic transformation. Neoplastic transformation consists of an accumulation of diverse multiple genetic and epigenetic alterations. These alterations are likely to be passed on to daughter cells, accumulating over a period of time. The rapidly proliferating cells in a regenerating liver are extremely susceptible to carcinogenesis, as there is minimal time available between cell cycles for DNA repair (Morin and Normand, 1986). Therefore, the evaluation of the role of molecular checkpoints in this constitutive in vivo system is a very relevant model to understand the dynamics of cell cycle.

We have observed significant induction of p53 during the process of regeneration (Fig. 1). Heretofore, the known inducers of p53 activity predominantly included DNA-damaging agents such as ionizing and ultraviolet irradiation (Lu and Lane, 1993) and alkylating agents (Reddy and Randerath, 1987). Hence, the endogenous damage that induces p53 during the process of liver regeneration deserves attention. Our data show that livers of PH rats treated with the antisense PS-ODN, p53T, lose their G₁-S cell cycle checkpoint activity (Fig. 2). This observation suggests that an increase in p53 provides a functional signal capable of establishing a strong checkpoint activity in the regenerating liver. The consequences of this lost checkpoint activity include increased weight gain (Table 1), MI (Table 1), and PCNA expression (Fig. 1) by these regenerating livers. We already know much about the humoral and cellular factors involved in stimulating hepatocytes to enter the cell division cycle after PH (Michalopoulos and DeFrances, 1996). The results of the present study, where we have suppressed the G₁-S brake using a PS-ODN targeted at p53 mRNA, indicate that the regenerating liver responds by increasing its proliferative activity as a likely result of decreased p21^waf-1 levels (Fig. 6). Hence, our observations in this model are consistent with what would be expected by inhibition of p53 activity.

Antisense approaches to modulation of gene expression have been tantalizing, but it is has been difficult to prove their utility. The present model system provides an in vivo setting where the functional role of transient changes in p53 expression can be evaluated. The animals are nondiseased and constitutive expressors of p53 with a normal blood supply to the liver tissue. The responses to transient suppression of p53 in the intact animal provide unique insights into the role of p53 expression and a variety of coordinate gene expression events. One such observation that was surprising was an increase in oxidative stress (Fig. 7), measured by TBARS, in the regenerating livers of p53 antisense-treated animals. It is likely that the increased TBARS levels in the p53 antisense-treated PH animals are at least partly related to faster recovery of the CYP enzymes in their regenerating livers (Fig. 8).

One of the important advantages of studying the role of p53 in the process of liver regeneration is that it allows for measurement of several important indices that can be used as markers of functional recovery of the tissue. In the present study, functional assays for activities of the major CYP enzymes 1A1, 2B1, 3A2, and 2E1 were used as indices of liver recovery (Fig. 8). Because CYP isoforms are heme-containing proteins capable of generating both hydroxyl and reactive nitrogen radicals, these observations provide an important insight into the potential relationship between endogenous DNA damage and p53 transcriptional activity. We observed that all CYP enzyme activities fall after PH and recover gradually as the liver regenerates. All of these enzymes show better functional recovery, especially over the short term (24 h), in p53T-treated PH rats. Thus, loss of p53 function is associated not only with faster structural recovery of the liver tissue, but functional metabolic recovery as well.

It has been reported previously (Hainaut and Milner, 1993; Milner, 1995) that agents that affect the conformation of the p53 polypeptide also affect its capacity to bind to a specific DNA sequence in vitro. Using p53 translated in vitro, they have shown that agents that reversibly alter the wild-type phenotype of p53 also modulate its DNA-binding capacity. These agents include metal chelators and oxidizing agents. p53 molecules whose conformations are altered by oxidation show inhibition of DNA binding. Conversely, reduction favors folding of p53 in its wild-type form and causes restoration of DNA binding. Evidence that the p53 protein is sensitive to redox conditions is particularly interesting because a conformation hypothesis proposes that normal p53 may function both to suppress and promote cell proliferation, a given function depending on the conformation of the protein. The suppressor form of p53 resembles the wild-type phenotype and may be inactivated by growth stimulation, which induces a structural change in the protein. The promoter form resembles a mutant phenotype. Redox control of p53 conformation may explain how p53 activity could be regulated by physiological processes associated with stress-induced responses and also with growth signaling, like in the case of liver regeneration.

A vast array of nonspecific activities have been reported with the use of antisense PS-ODNs. These nonspecific effects have been variously attributed to binding with plasma proteins and transcription factors, induction of interferon production, potentiation of tumor necrosis factor action, and inhibition of protein kinase C (reviewed by Neckers and Iyer, 1997). The reported number of mispairs in a putative antisense duplex required to demonstrate complete loss of sequence-specific activity is four mispairs in a 20-mer PS-ODN (Monia et al., 1996). Hence, we used a sequence complementary to human p53 mRNA, OL(1)p53, which has four mispairs in a 20-mer with the rat p53 sequence as a carefully designed inactive control. We observed that p53 expression, weight gain of regenerating livers, MI, G₁-S cell cycle checkpoint activity, as well downstream events such as PCNA and p21^waf-1 expression can all be modulated in a sequence-specific manner without significant toxicity by use of PS-ODNs in this study. The most likely explanation is that the dose of PS-ODNs required to produce the nonsequence-specific effects is significantly higher than that which results in sequence-specific effects. Furthermore, the dose-dependent effect of C-5-P modified PS-ODNs (Fig. 4, A and B) on remnant liver weight gain and MI is an additional testimony to the specificity of these ODNs in the rat PH model.

The observations that inhibition of p53 in the regenerating liver results in enhanced mitosis, elevated PCNA expression, and diminished number of cells in the G₁ phase of the cell cycle indicates that p53 acts as a functional cell cycle checkpoint in this in vivo model.