Recovery of Liver Mass without Proliferation of Hepatocytes after Partial Hepatectomy in Skp2-deficient Mice¹

Cell growth (increase in cell mass) and cell proliferation (increase in cell number) are usually coordinated to ensure that cell size is maintained. This coordination is thought to be achieved as a result of the fact that cell growth is rate-limiting for progression of the cell cycle (1), and it appears to be dependent on the activity and abundance of the cyclin-dependent kinase inhibitor p27^Kip1. Mice in which the expression of p27^Kip1 or of its regulators is deficient thus exhibit an imbalance between cell growth and proliferation. For example, p27^Kip1-deficient mice manifest multiple organ hyperplasia and a decrease in the size of individual cells (2). In contrast, mice lacking Skp2, an F-box protein of the SCF ubiquitin ligase complex that targets p27^Kip1 for degradation, exhibit cellular accumulation of p27^Kip1, a reduction in both body size and the size of individual organs, an increase in the mass of individual cells, and polyploidy (3). These observations suggest that p27^Kip1 is a determinant of organ size. Furthermore, a reduced abundance of p27^Kip1 in many human cancers correlates well with poor prognosis (4, 5, 6, 7), and the loss of one p27^Kip1 allele in mice increases the sensitivity of these animals to cancer-inducing agents (8). Although hepatocytes are highly differentiated and quiescent cells, they retain the stem cell-like ability to proliferate in adults (9). In response to a reduction in liver mass, hepatocytes begin to proliferate in a relatively synchronized manner to restore hepatic mass and function. PH³ is thus a useful model with which to study the cell cycle in animals. To investigate the role of Skp2 in maintenance of the balance between cell growth and proliferation in intact animals, we examined the effects of PH in Skp2-deficient mice. Unexpectedly, liver mass recovered to almost the normal value in Skp2^−/− mice after PH, although cell proliferation was suppressed as a result of the accumulation of p27^Kip1. Our data suggest that cell growth and proliferation are independently regulated, and that the loss of the ability to proliferate can be compensated for by an increase in the size of individual cells.

We previously generated mice deficient in Skp2. Female animals at 8–12 weeks of age were subjected to PH. The experiments were approved by the Committee on the Ethics of Animal Experimentation and were performed in accordance with the Guidelines for Animal Experimentation of the Medical Institute of Bioregulation, Kyushu University.

Animals were deprived of food for 15 h before PH. Anesthesia was induced by i.p. administration of sodium pentobarbital (75 μg/g of body mass), after which resection of 70% of total liver mass (left lateral, left median, and right median lobectomy) or sham surgery was performed. For BrdUrd labeling experiments, mice were injected i.p. with BrdUrd (100 μg/g of body mass; Sigma Chemical Co.) 24, 48, 72, and 96 h after surgery. The mass of the resected liver tissue was measured after the operation, and that of the remnant liver was determined after killing of the animals 7 days after surgery. No substantial change in liver mass or in microscopic morphology was apparent after surgery in mice subjected to the sham operation in any of the experiments.

Seven days before or after PH, blood samples were obtained from the orbital venous plexus of animals subjected to anesthesia with diethyl ether. Biochemical analysis of serum was performed with a serum analyzer (Fujifilm Dri-Chem 3500V).

Resected or excised liver tissue was fixed with 4% paraformaldehyde in PBS and embedded in paraffin. Sections (3-μm thick) were stained with H&E or Feulgen solution as described (3). For p27^Kip1 staining, sections were boiled for 10 min by microwave irradiation in a solution containing 0.1 m Tris-HCl (pH 9.5) and 0.1% Tween 20, and endogenous mouse immunoglobulin was blocked with the use of a Mouse to Mouse kit (ScyTeck). Sections were subsequently incubated first with a mouse monoclonal antibody to p27^Kip1 (clone 57; Transduction Laboratories) and then with biotin-conjugated goat antibodies to mouse immunoglobulin (Chemicon). For BrdUrd staining, boiled sections were denatured with 2 m HCl containing 0.1% Triton X-100 and were then neutralized with 0.1 m sodium borate buffer (pH 8.5). After blocking of nonspecific sites with a solution containing 5% BSA and 5% rabbit serum in PBS, the sections were incubated first with a rat monoclonal antibody to BrdUrd (Harlan Sera-Lab) and then with biotinylated rabbit antibodies to rat immunoglobulin (Chemicon). Signals were amplified with a CSA kit (Dako) or a streptavidin-biotin-peroxidase-complex detection kit (Vectastain Elite; Vector) and were visualized with diaminobenzidine (Wako). Nuclei were counterstained with hematoxylin. Average diameter (r) of cells was estimated from at least 200 hepatocytes in ×400 microscopic image (33,571 μm²/view area) of liver sections prepared from resected and regenerated liver tissue of Skp2^+/+ and Skp2^−/− mice. On the assumption that the cell is a sphere, cell volume (v) was calculated as follows: v = 4πr³/3. Data are means of values obtained from five mice of each genotype.

Sections were prepared and subjected to microwave irradiation as described above. The sections were exposed to 5% goat serum in a solution containing 5% BSA in PBS and were subsequently incubated first with rabbit polyclonal antibodies to pHH3 and then with Alexa488 (green)-conjugated goat antibodies to rabbit immunoglobulin (Molecular Probes). Nuclei were stained with propidium iodide (5 μg/ml). The sections were finally covered with GelMount solution (Biomeda) and viewed with a fluorescence microscope (Eclipse E800M; Nikon). Images were captured with a chilled 3-CCD camera (C5810; Hamamatsu Photonics) and digitized with Adobe Photoshop.

Resected or excised liver tissue was homogenized in a solution containing 0.25 m sucrose, 3 mm CaCl₂, and 1 mm PMSF, and the homogenate was then centrifuged at 1000 × g for 10 min. The resulting pellet was homogenized in a solution containing 2.2 m sucrose and 3 mm CaCl₂, and the homogenate was subjected to centrifugation at 70,000 × g for 60 min. The nuclei in the resulting pellet were stained with propidium iodide (5 μg/ml) and analyzed with a flow cytometer (FACSCalibur; Becton Dickinson). DNA content was evaluated with CellQuest software (Becton Dickinson).

Frozen liver samples obtained from resected liver at the time of PH and of regenerated liver 48 h after PH in Skp2^+/+ and Skp2^−/− mice were minced into small pieces, and then each sample was homogenized in cold homogenizing buffer containing 50 mm Tris-HCl (pH 7.5), 0.25 m sucrose, 1 mm EDTA, 10 μg/ml aprotinin, 10 μg/ml leupeptin, and 1 mm PMSF for 5 min. Then the homogenate were subjected to 2× RIPA buffer containing 0.3 m NaCl, 20 mm Tris-HCl (pH 7.5), 2% NP40, 0.2% sodium deoxycholate, 0.2% SDS, 10 μg/ml aprotinin, 10 μg/ml leupeptin, and 1 mm PMSF. The lysates were incubated on ice for 30 min and were centrifuged at 20,000 × g for 20 min. The protein concentration was determined by Bradford protein assay kit (Bio-Rad) with BSA as a standard.

To evaluate expression level of p27^Kip1 in regenerating liver of Skp2^+/+ and Skp2^−/− mice, equal amounts of tissue lysate (20 μg/lane) prepared as described above were subjected to immunoblot analysis with mouse monoclonal antibodies to p27^Kip1 (clone 57; Transduction Laboratories) and to Hsp90 (clone 68; Transduction Laboratories) as a control. Immune complex was detected as described previously (3). Quantitative analysis was performed with the usage of LAS-1000 (Fujifilm) and Image Gauge software (Fujifilm).

To assess the proliferative capacity of hepatocytes in adult Skp2^−/− mice, we performed 70% PH in both mutant and wild-type animals. Mice were killed 7 days after surgery, and liver mass was determined. There was no difference in the liver:body mass ratio between Skp2^+/+ (5.10 ± 0.91%) and Skp2^−/− (5.10 ± 0.73%) mice (n = 8 for each genotype) at this time after PH (Fig. 1). In animals of both genotypes, liver mass was restored to a value similar to the preoperative mean (∼5% of body mass; data not shown). Blood biochemical indicators of hepatic function, including serum albumin concentration, also did not differ substantially between Skp2^+/+ and Skp2^−/− mice after PH (Fig. 1 B; data not shown). These results suggest that the potential of the liver of Skp2^−/− mice to recover its mass and function after PH is similar to that of the liver of Skp2^+/+ animals.

Although histopathological examination of liver sections revealed no gross structural abnormalities in Skp2^−/− mice, marked cellular and nuclear enlargement was apparent in the liver 7 days after PH compared with liver tissue resected at PH (Fig. 2). Such changes were not observed in sham-operated mice of either genotype (data not shown). In Skp2^+/+ mice, cellular enlargement in the liver was also observed 7 days after PH, but the extent of this change was much less pronounced than that apparent in Skp2^−/− mice. The estimated volume of individual cells thus increased 1.78- and 3.31-fold after PH in Skp2^+/+ and Skp2^−/− mice, respectively (Fig. 2 B). The volume of Skp2^−/− hepatocytes was 2.25 times that of Skp2^+/+ cells at the time of PH but was 4.19 times that of Skp2^+/+ cells 7 days after PH, suggesting that the difference in cell size between genotypes is enhanced by PH in adult mice.

Given that the size of hepatocyte nuclei is affected by their DNA content, we also examined the DNA content of the hepatocytes of Skp2^+/+ and Skp2^−/− mice both at the time of PH and 7 days after PH. Similar to the results obtained for cell volume, the DNA content of hepatocytes was increased markedly by PH in Skp2^−/− mice; again, a small increase in the percentage of cells exhibiting polyploidy (up to 8C) was also apparent after PH in Skp2^+/+ mice (Fig. 3). Whereas virtually no hepatocytes with a DNA content of ≥16C were detected in Skp2^+/+ mice even after PH, ∼20% of hepatocytes of Skp2^−/− mice exhibited a DNA content of ≥16C after PH. Thus, whereas both the size and DNA content of hepatocytes were increased slightly by PH in Skp2^+/+ mice, these effects were greatly exaggerated in Skp2^−/− mice.

We compared the abundance of p27^Kip1 between the livers of Skp2^+/+ and Skp2^−/− mice at the time of and 48 h after PH. Immunostaining (Fig. 4, A–D) and immunoblotting (Fig. 4, E and F) revealed that the amount of p27^Kip1 in the liver of Skp2^−/− mice 48 h after PH was increased compared with that in the same animals at the time of PH or that in Skp2^+/+ mice 48 h after PH. We have shown recently that Skp2 functions in the degradation of p27^Kip1 during S and G₂ phases of the cell cycle, but that it is not required for p27^Kip1 degradation at the G₀-G₁ transition (that is, at entry into the cell cycle; Ref. 10). Given that the abundance of p27^Kip1 is low in cells at the G₁-S boundary of the first cell cycle (10), we hypothesized that hepatocytes should be able to enter S-phase in the absence of Skp2. To test this hypothesis, we injected Skp2^+/+ and Skp2^−/− mice i.p. with BrdUrd after PH. Immunostaining of the liver with antibodies to BrdUrd 7 days after PH revealed that Skp2^−/− hepatocytes incorporated BrdUrd to an extent similar to that observed with Skp2^+/+ cells (Fig. 4, G–J) suggesting that Skp2^−/− hepatocytes are indeed able to enter S-phase. Normally, cells proceed to M-phase after completion of S-phase. Thus, a certain proportion of hepatocytes of Skp2^+/+ mice 48 h after PH stained with antibodies to pHH3, a marker for cells in M-phase (Fig. 4, K, L, O, and P). In contrast, we did not detect hepatocytes in Skp2^−/− mice 48 h after PH that stained with these antibodies (Fig. 4, M, N, Q, and R). These results thus suggested that Skp2^−/− cells are able to enter S-phase but not M-phase, a characteristic of endoreplication. The enlargement and polyploidy of Skp2^−/− hepatocytes are consistent with the repeated occurrence of S-phase without mitosis.

Organ size is usually dependent on body size, so that the organ:body size ratio is maintained relatively constant under physiological conditions. Mice deficient in p27^Kip1 exhibit an increased body size and multiple organ hyperplasia (2). In the organs of these animals, the number of cells is increased but the size of individual cells is reduced. In contrast, mice lacking Skp2, a component of the SCF ubiquitin ligase complex that targets p27^Kip1 for degradation, manifest accumulation of p27^Kip1, a reduced body and organ size, and cellular enlargement with polyploidy, the latter of which is most prominent in the liver (3). These results suggest that p27^Kip1 plays an important role in determining organ size as well as individual cell size.

To investigate whether the hepatocytes of Skp2^−/− mice are able to proliferate, we subjected these animals to PH. The gross appearance of the liver and the results of serum biochemical analysis 7 days after PH indicated that the mass and function of the liver were restored in Skp2^−/− mice to an extent similar to that observed in the wild-type animals. Unexpectedly, however, the number of hepatocytes did not appear to increase in the mutant mice in response to PH, given that no cells reactive with antibodies to the mitotic marker pHH3 were detected; rather, the volume of individual cells increased >3-fold. The restoration of liver mass and function in the mutant mice was thus achieved not by cellular proliferation but by the enlargement of individual cells. These observations suggest that organ size is not simply determined by cell number. Impaired proteolysis of p27^Kip1 as a result of the lack of Skp2 is likely responsible for the failure of hepatocytes to proliferate in the hepatectomized Skp2^−/− mice, but the fact that the organ:body size ratio was not substantially altered in these animals suggests that the final extent of liver regeneration was determined by a mechanism independent of Skp2.

Although p27^Kip1 is degraded at the G₀-G₁ transition when quiescent cells enter the cell cycle (10), the expression of Skp2 begins at S-phase (10, 11). We have shown recently that Skp2 is required for the degradation of p27^Kip1 during S and G₂ phases but not for that during the G₀-G₁ transition, suggesting the existence of a Skp2-independent mechanism for p27^Kip1 proteolysis at this transition (10). We have now shown that cells are able to enter S-phase in the absence of Skp2-mediated proteolysis of p27^Kip1. The hepatocytes of regenerating liver were thus efficiently labeled with BrdUrd in Skp2^−/− mice. However, the mitotic defect apparent in these cells suggests that p27^Kip1 accumulation during G₂ may affect directly or indirectly the kinase activity of Cdc2. Consistent with this notion, we have shown that p27^Kip1 is required for this block of mitosis, and that the kinase activity of Cdc2 is markedly reduced in the Skp2^−/− cells compared with Skp2^+/+ cells.⁴ In fission yeast, the accumulation of the cyclin-dependent kinase inhibitor Rum1 (as a result of impaired function of Pop1, an F-box protein that mediates Rum1 ubiquitination) and the consequent inhibition of the kinase activity of the Cdc2-mitotic cyclin complex leads to the development of polyploidy (12). The pop1 mutant cells thus reenter S-phase without passing through M-phase, a process known as endoreduplication, leading to polyploidy (13). Similar to Pop1 in fission yeast, Skp2 likely functions to reduce the concentration of p27^Kip1 during S and G₂ phases to avoid inhibition of the kinase activity of Cdc2 by p27^Kip1. Our present data are consistent with this notion.

Several growth factors and cytokines, including hepatocyte growth factor, epidermal growth factor, interleukin 6, tumor necrosis factor-α, and transforming growth factor-α, contribute to the regulation of liver regeneration (9). Mice deficient in these various proteins thus exhibit severe abnormalities in liver development or regeneration (14). In contrast, our present results indicate that liver regeneration does not depend on the proliferation of hepatocytes. We propose that these growth factors and cytokines primarily accelerate hepatocyte growth, and that cells begin to divide after they achieve a certain mass. Cells continue to grow even if cell division is inhibited. It has been unclear how an organ senses its size and terminates its growth.

The expression of Skp2 is up-regulated in various types of cancer, and the Skp2 gene is considered an oncogene (15, 16). The expression of Skp2 is inversely related to that of p27^Kip1 in colorectal carcinomas (17). Thus, down-regulation of p27^Kip1 by Skp2-mediated proteolysis likely contributes to cancer progression, leading to poor prognosis (4, 5, 6, 7). Adenovirus-mediated targeted expression of Skp2 in the liver resulted in increased proliferation of hepatocytes in response to PH (18). Together with these previous observations, our present study indicates that Skp2 regulates hepatocyte proliferation during liver regeneration in vivo. In addition, the phenotype of Skp2^−/− hepatocytes, including polyploidy and impaired proliferation, is similar to that of senescent cells (19, 20). Further investigation of the mechanism of Skp2-mediated regulation of p27^Kip1 might thus provide both leads to the development of new cancer therapies as well as insight into cellular senescence.

We thank S. Hatakeyama, T. Kamura, K. Kominami, and H. Nishitani for helpful suggestions; K. Shimoharada, Y. Yamada, and other laboratory members for technical assistance; and M. Kimura for help in preparation of the manuscript.