Polyploidy associated with oxidative injury attenuates proliferative potential of cells

Polyploid cells accumulate large amounts of DNA compared with diploid cells, due to DNA synthesis without cell division (Hieter and Griffiths,1999EF22). Polyploidy is encountered in plants, lower life forms, as well as all mammalian tissues. Advanced polyploidy in mammalian cells accompanies terminal differentiation and cell senescence (Brodsky and Uryvaeva,1977EF6). If polyploidy were utilized as a mechanism to cull senescent cell subpopulations, then its circumvention would provide survival advantage to cells and either facilitate continuation of developmental programs or favor the onset of neoplasia under appropriate circumstances. Analysis of ontogeny and organ regeneration following injury, in organs such as the liver, indicates that parenchymal cells lose their replication potential after the onset of terminal differentiation (Smith and Pereira-Smith,1966EF45). Even subpopulations of cancer cells exhibit differences in proliferative activity in the context of polyploidy (Fujikawa-Yamamoto et al.,1997EF13). However, regulation of polyploidy and its biological significance are incompletely understood.

Hepatic polyploidy accompanies late fetal development and postnatal maturation (Sigal et al.,1999EF42) and its onset in the adult liver is well recognized (Brodsky and Uryvaeva,1977EF6; Carriere,1969EF8; Alison and Wright,1985EF2). Moreover, weaning and commencement of feeding (Dallman et al.,1974EF10; Barbason and Houbrechts,1974EF4), compensatory liver hypertrophy following partial hepatectomy (Bucher et al., 1963; Brodsky and Uryvaeva, 1977EF6; Sigal et al.,1999EF42), toxin and drug-induced liver disease (Bohm and Noltemeyer,1981EF5; Madra et al.,1995EF29; Kato et al.,1996EF24; Aardema et al.,1998EF1), as well as administration of specific growth factors and hormones (Printseva et al.,1989EF38; Cruise et al.,1989EF9; Torres et al.,1999EF50) may induce hepatic polyploidy. Although liver growth control has long been studied, whether the replication potential of polyploid hepatocytes is altered remains unresolved,in part, owing to difficulties in distinguishing between cellular DNA synthesis and generation of daughter cells (Simpson and Finckh,1963EF43, Solopaev and Bobyleva,1981EF47). More recently,transplanted cells were shown to integrate into the liver parenchyma and then to repopulate the host liver (Gupta et al.,1995EF17). Regulated proliferation in transplanted cells, with no proliferation within the normal liver, and extensive proliferation in animals where endogenous host hepatocytes were lost selectively, permitted establishment of clonogenic type assays in intact animals (Rhim et al., 1994EF40;Overturf et al., 1997EF37; Mignon et al., 1998EF32; Laconi et al.,1998EF26; Gupta et al., 1999a;Guha et al., 1999b).

Here we provide evidence from studies utilizing a variety of systems,including a genetically defined cell transplantation system in F344 rats, to demonstrate the proliferation capacity of polyploid rat hepatocytes. We found evidence of oxidative DNA injury in polyploid hepatocytes isolated from rats subjected to two-thirds partial hepatectomy (PH). Moreover, we were able to recapitulate polyploidy in cultured primary rat hepatocytes following mitogenic stimulation in the setting of oxidative DNA injury.

Male F344 rats were obtained from the National Cancer Institute (Bethesda,MD). The Special Animal Core of the Liver Research Center provided dipeptidyl peptidase IV (DPPIV)-deficient F344 rats. The animals were maintained under 14-hour light and 10-hour dark cycles and provided pelleted Rodent Chow 5001(PMI Feed, Richmond, VA) ad libitum. All animals received humane care in compliance with National Research Council criteria outlined in the Guide for the Care and Use of Laboratory Animals (NIH Publication 85-23, Revised 1985). The Animal Care and Use Committee at the Albert Einstein College of Medicine approved the animal protocols.

Two-thirds PH was performed under ether anesthesia between 8 and 10 AM according to Higgins and Anderson (Higgins and Anderson,1931), without restrictring food or water intake. Several DPPIV-deficient rats were treated with 30 mg/kg retrorsine (Sigma Chemical Co., St. Louis, MO) intraperitoneally at 6 and 8 weeks of age as described previously (Laconi et al.,1998). Four weeks after the final retrorsine dose, animals were subjected to two-thirds PH followed by injection of 5×10⁶ F344 rat hepatocytes into the spleen as described previously (Gupta et al.,1995; Gupta et al., 1999a). Intrasplenic injection deposits hepatocytes into liver sinusoids and cells integrate subsequently in the liver parenchyma (Gupta et al.,1997; Gupta et al.,1999b).

The liver was perfused in situ with collagenase to isolate hepatocytes, as described previously (Gupta et al.,1997). Hepatocytes were plated on collagen-coated dishes at 4×10⁴ cells/cm² in RPMI 1640 medium containing penicillin, streptomycin, amphotericin B and 10%fetal bovine serum (FBS; Atlanta Biologicals, Norcross, GA). The cell viability was estimated by exclusion of Trypan Blue dye. The culture medium was changed 3 hours after cell attachment. Cells were cultured for up to 5 days and harvested by trypsinization and centrifugation under 500 g for 5 minutes at 4°C. Cells were washed twice with cold phosphate-buffered saline (PBS) pH 7.4.

A Gamma Cell-40 Irradiator (Atomic Energy of Canada, Ottawa, Canada) with Cesium-137 as the source was used. Radiation dose was administered at 84 cGy/minute and cells were radiated to a total of 10-30 Gy. For transplantation studies, hepatocytes were suspended in RPMI 1640 medium at 2×10⁶ cells/ml, radiated to 30 Gy immediately after cell isolation and transplanted into animals within 2 hours of isolation.

After specified periods, [³H]thymidine (3 μCi, 70 Ci/mmole,ICN Radiochemicals, CA) was added to cultured hepatocytes for 1 hour. Cells were washed twice with cold PBS and DNA was extracted as described (Gupta et al., 1992). ³H activity was measured by liquid scintillation counting in an aliquot, and DNA microquantitation by fluorimetry. All experiments were in triplicate at least. To induce DNA synthesis, 20 ng/ml transforming growth factor α(TGFα), 0.1 μM norepinephrine and 25 μM vasopressin (Sigma) were added to the culture medium in some experiments.

Changes in cell profiles were analyzed in isolated nuclei as described previously (Sigal et al.,1999EF42). Cells were harvested by trypsinization, washed with PBS once and resuspended in citrate buffer (250 mM sucrose, 40 mM trisodium citrate.2H₂O and 5% DMSO). Cells stored at-80°C were rapidly thawed to 37°C and lyzed in 100 μl citrate buffer containing 900 μl trypsin, 3.4 mM trisodium citrate.2H₂O,0.1% NP-40 and 1.5 mM spermine tetrahydrochloride, pH 7.6, at room temperature for 10 minutes with occasional mixing. The nuclei were stained with 0.04%propidium iodide on ice. Flow cytometry was carried out with the FACSTAR plus machine (Becton Dickinson, San Jose, CA). Typically, approximately 10,000 events were collected. Data were analyzed with Lysis II software. A laser scanning cytometer (Compucyte, Cambridge, MA) was used to document cell ploidy in parallel. Each condition was examined at least in triplicate and experiments were reproduced on more than five occasions.

Unless specified, all chemicals below were from Sigma. Harvested cells were stored at -80° in 6% salicylic acid. Cells were thawed to 4°C and disrupted by ultrasonication, as described previously (Anderson,1985). Cell debris were eliminated by pelleting at 10,000 g for 10 minutes at 4°C. To 100 μl supernatant, 800 μl 0.3 mM nicotinamide adenine dinucleotide phosphate, reduced form (NADPH), 100 μl of 6 mM 5.5′-dithiobis(2-nitrobenzoic acid) and 0.5 U glutathione reductase were added. Changes in absorbance at 412 nm were measured for 2 minutes. Glutathione standards were prepared in a linear range with 100 μM stock solution. Protein content was assayed in aliquots using the Bradford assay. All experiments were done in triplicate.

Frozen cells were thawed on ice, ultrasonicated and then centrifuged at 10,000 g for 10 minutes at 4°C according to the methood of Luck (Luck, 1965). To 10-40μl supernatant, 3 ml H₂O₂-phosphate buffer was added in a silica cuvette. The time (T) for change in optical density at 240 nm from 0.45 to 0.40 was measured at room temperature. The catalase activity was derived from the formula 17/T=units/assay mixture. Protein content was determined in aliquots using the Bradford assay. Data were expressed as catalase units/μg protein using triplicate conditions.

Harvested cells were suspended in 100 μl deionized water and peroxidation was measured using a calorimetric kit, according to the manufacturer's instructions (Calbiochem-Novabiochem Corp., San Diego, CA). All conditions were in triplicate.

Cultured cells were fixed in 0.5% glutaraldehyde in PBS (pH 7.4) at room temperature for 10 minutes. Cells were washed and incubated at 37°C with 5-bromo-4-chloro-3-indolyl β-galactoside (X-gal) in either PBS or citric acid/sodium phosphate buffer, pH 6.0, according to the method of Dimri et al.(Dimri et al., 1995EF12) for 18 hours.

Samples from the median liver lobe were frozen in methylbutane at -70°C and 5 μm thick cryostat sections were prepared.

To demonstrate oxidative DNA injury, sections were fixed in 70% ethanol and stained with anti-8-oxo-2′ dG (Trevigen, Gaithersburg, MD, Catalog No. 4355-MC-100) was according to the manufacturer's instructions. Briefly,sections were digested with RNase (100 μg/ml) for 1 hour at 37°C. DNA was denatured by immersion in 4 M HCl for 7 minutes and neutralized in 50 mM Tris-base. After blocking with 10% fetal bovine serum, tissues were incubated with antibody diluted 1:300, at 37°C for 16 hours. Antibody binding was localized with a biotinylated multilink secondary antibody (Biogenex, San Ramon, CA). Endogenous peroxidase was quenched with 3%H₂O₂ in methanol for 30 minutes. The Vectastain ABC system (Vector Laboratories, Burlingame, CA) was used for visualization of antibody binding and color was developed with the enhanced diaminobenzidine substrate (Dako Corporation, Carpinteria, CA).

Transplanted cells were identified in DPPIV-deficient F344 rats by detecting DPPIV activity histochemically as described previously (Gupta et al., 1995EF17). Tissue sections were fixed in chloroform-acetone (1:1 vol/vol) at 4°C. Liver from normal F344 rats was included as a positive control and liver from DPPIV-deficient rats, which were not subjected to cell transplantation, served as a negative control.

For morphometric analysis, multiple photomicrographs of all conditions were obtained. The number of transplanted cell foci in 44-56 consecutive high-magnification fields was determined. Individual transplanted cells in each of these foci were identified as described previously (Sokhi et al.,2000EF46). To establish the number of transplanted cells composing each cell focus, 43-66 cell foci were scored in each animal and the data tabulated for comparison.

The data were analyzed with the SigmaStat software (Jandel Scientific, San Rafael, CA), and expressed as mean±s.d. The significance of differences was analyzed by the Student's t-test, χ² test,Mann-Whitney rank correlation test for non-parametrically distributed data,Kruskall-Wallis one way analysis of variance (ANOVA), and Dunn's test to isolate groups differing from others. P values <0.05 were considered significant.

The viability of isolated hepatocytes was 80-90%. Flow cytometry showed fewer diploid cells and more polyploid hepatocytes in rats subjected to partial hepatectomy, which was similar to previous studies (Sigal et al.,1999EF42). Also, hepatocytes isolated from livers after PH showed greater autofluorescence due to cytoplasmic complexity on flow cytometry. This reflects intracellular changes,such as accumulation of the pigment lipofuscin, which is an end product of intracellular oxidation (Harman,1989EF21). This prompted us to hypothesize that PH induced oxidative injury in hepatocytes.

Depletion of cellular glutathione content and catalase activity and increase in lipid peroxidation were used to assess the onset of oxidative injury. Hepatocytes were isolated from three normal rats and three rats on which two-thirds PH had been performed 5 days previously. The total cellular glutathione content, as well as catalase activity, declined significantly in hepatocytes from animals subjected to PH(Table 1). Also, hepatocytes from animals subjected to PH exhibited greater lipid peroxidation. Our lipid peroxidation assay demonstrated malonaldehyde and 4-hydroxyalkenal accumulation in cells. The findings were in agreement with induction of oxidative stress in hepatocytes after PH.

Immunostaining with an antibody that recognizes 8-hydroxyguanine adducts established that PH caused oxidative DNA injury. The portion of the liver that was removed served as an internal control for each animal (n=3). Liver samples were obtained from animals at 5 days after PH for the analysis. In control livers, only occasional hepatocytes showed 8-hydroxyguanine adducts(Fig. 1). Additional biliary and arteriolar cells also showed immunostaining. In contrast, 5 days after PH,8-hydroxyguanine adducts became apparent in most hepatocytes. The nuclei of these cells with oxidative DNA injury often had a `stippled' appearance,possibly indicating extensive adduct formation. On occasion, cells with evidence for oxidative DNA adduct formation were found to be undergoing apoptosis. Morphometric analysis showed that in control livers, 8±6 nuclei per high-magnification field (×400) had evidence for oxidative DNA adduct formation. After PH, this increased several-fold, with 90±20 nuclei per high-magnification field showing oxidative DNA damage, P<0.001, t-test.

To further establish the consequences of oxidant stress in hepatocytes, we exposed cells to ionizing radiation, which induces release of free radicals. Hepatocytes were irradiated after attachment to culture dishes and they were then cultured for 1, 3 and 5 days. After radiation, hepatocytes from normal control rats showed dose-dependent decreases in glutathione and catalase activity, along with increased lipid peroxidation, which was in agreement with oxidative injury (Fig. 2). Hepatocytes from livers after PH showed a similar pattern of oxidative injury,albeit with greater changes in glutathione, catalase and lipid peroxidation after radiation. At 5 days after 30 Gy radiation, glutathione content in hepatocytes from livers after PH was only 50% of that in control hepatocytes,19±0.6 nmol/μg protein versus 10±0.3 nmol/μg protein, P<0.005, t-test. Catalase activity decreased in control hepatocytes after radiation. This decrease was more pronounced in hepatocytes from livers after PH, where catalase activity following radiation became undetectable by the third day of culture. Finally, when lipid peroxidation was examined in cells cultured for 5 days after radiation, hepatocytes from PH livers showed 2-fold greater lipid peroxidation compared with cells from the normal liver, P<0.05. These findings indicated that hepatocytes exposed to oxidative injury after PH were more susceptible to radiation-induced oxidative injury.

If oxidative injury were responsible for cellular DNA damage, induction of polyploidy would be affected. Nuclear DNA was analysed with flow cytometry after cell culture for 5 days. Hepatocytes were studied on four occasions from unperturbed rats and rats subjected to PH.

Culture for 5 days altered the initial ploidy classes of hepatocytes. In cells from the unperturbed liver, the diploid fraction decreased from 29±1% to 25±0.6% after 5 days, P<0.001, t-test, the tetraploid cell fraction was unchanged at 67±3%and 68±2%, p=n.s., and the octaploid cell fraction increased from 1±0.5% to 6±0.5%, P<0.001. Ploidy of cells from the normal liver began to resemble initial ploidy distributions of cells from liver 5 days after PH, where the diploid cell fraction constituted 19±2%, the tetraploid cell fraction 75±3% and the octaploid cell fraction 6±1%. However, after 5 days of culture, hepatocytes from PH livers showed more diploid cells, 46±2%, P<0.001, t-test, and fewer tetraploid cells, 48±3%, P<0.001, t-test, while the fraction of the polyploid cells remained unchanged, 5±0.3%. Such diploid cell enrichment could have arisen from transition of cells to the next ploidy class with depletion of the most highly polyploid cells from the culture. Alternatively, diploid cells could have proliferated selectively.

The response of cultured cells to radiation provided further information. For these assays, cells were attached to tissue culture plates for 24 hours,exposed to either 10 Gy or 30 Gy radiation and then cultured for an additional 4 days. Changes in cell ploidy are summarized in Table 2. In hepatocytes from unperturbed livers, diploid cell fractions decreased modestly after radiation,albeit significantly, without much change in the tetraploid or the octaploid cell fractions. In contrast, after radiation, hepatocytes from PH animals showed a two-fold increase in the proportion of octaploid cells to approx. 10%. SABG activity was used to demonstrate whether oxidative injury induced senescence-type changes in cells from PH livers. The fraction of cells with pancytoplasmic SABG staining was analyzed in at least 1000 cells in 10 random microscopic fields for each cell type. Radiation increased SABG expression by 2.6±0.5-fold (2.7±0.3% versus 7.2±1.5%, P<0.001, t-test). These findings suggested that oxidative injury caused cell ploidy to increase with activation of senescence-type changes.

To demonstrate whether polyploidy perturbed the ability of cells to proliferate, we used an in vivo assay in DPPIV-deficient F344 rats. Treatment of rats with retrorsine and two-thirds PH prior to cell transplantation induces extensive proliferation in transplanted cells (Laconi et al.,1998EF26).

Groups of four to six DPPIV-deficient rats were established as follows. Group I received hepatocytes from unperturbed F344 rats; Group II received hepatocytes from F344 rats 5 days after two-thirds PH; Group III received hepatocytes radiated to 30 Gy from unperturbed F344 rats; and Group IV received hepatocytes irradiated to 30 Gy from F344 rats 5 days after two-thirds PH. Two rats in each group were sacrificed at 2 days after cell transplantation to determine whether cells had engrafted in the liver. All rats were sacrificed 10 days after cell transplantation to analyze cell proliferation.

Two days after cell transplantation, similar numbers of transplanted cells were observed in the liver of all DPPIV-deficient recipients. Transplanted cells were distributed in periportal areas at this time with 1-3 cells per portal area, which indicated that engraftment of the cells was unperturbed(Fig. 3). Ten days after cell transplantation, Group I animals showed large foci of transplanted cells, with 53±30 cells per focus. In contrast, Group II recipients showed far smaller foci of transplanted cells, with 18±11 cells per focus,representing a 2.9±1.7-fold decrease in cell numbers, P<0.001, Mann Whitney rank sum correlation. Group III recipients showed smaller foci of transplanted cells, similar to Group II recipients,with 22±13 cells per focus, which indicated a decline in the number of transplanted cells by 2.4±1.4-fold, P<0.001. The size of transplanted cell foci was the smallest in Group IV animals, which received hepatocytes from PH liver and radiation, with only 4±3 transplanted cells per focus, which was 13.3±7.5-fold less than Group I cell recipients, P<0.001. Additional data from these studies are shown in Table 3.

If oxidative injury were causally involved in polyploid induction, this change should be reproducible in suitable systems. To demonstrate this, we studied primary hepatocytes from unperturbed rats. We reasoned that as polyploid cells contain greater amounts of DNA, advancement of ploidy would require induction of cellular DNA synthesis by adding growth factors, such as TGFα, to the culture medium. It was hoped that this would mimic some aspects of PH-induced hepatic DNA synthesis. Also, in some experiments, we added norepinephrine (NE) and vasopressin (VP), which are released after PH and modulate growth factor responses in cultured hepatocytes. To induce oxidative injury, cultured cells were exposed to 30 Gy radiation. Assays involving cells cultured for 48 hours, to verify the bioactivity of TGFα, showed 3- to 5-fold greater DNA synthesis rates in stimulated cells, P<0.001, t-test. NE and VP increased TGFα-induced DNA synthesis further by between 1.4 and 1.6 fold, P<0.01, t-test. As expected, radiation of cells before adding TGFα abrogated hepatocellular DNA synthesis. The final experimental protocol involved addition of TGFα to hepatocytes within 2 hours of culture, radiation 24 hours after the start of culture and analysis of cells after 4 more days in culture with TGFα.

The findings from one of two experiments providing similar results are shown in Fig. 4. Exposure to TGFα by itself increased enrichment of octaploid cells to approx. 10%, P<0.001, χ² test. After cell culture with TGFα and either NE or VP, cellular ploidy advanced further, with octaploid hepatocytes constituting up to 15%, P<0.001. Moreover,radiation of cells cultured with TGFα alone or with the addition of either NE or VP advanced cell ploidy even more with decline in diploid cell fractions and rise in octaploid cell fractions. The advancement in cell ploidy was verified to be associated with the appearance of megalonuclei in our cultured cells.

These data indicate that induction of hepatic polyploidy following two-third PH, which is an excellent model of compensatory hepatic hypertrophy,was associated with oxidative DNA injury. Polyploid cells showed susceptibility to further radiation-induced oxidant injury with depletion of anti-oxidant defenses, exemplified by cellular glutathione and catalase, and induction of senescence-type changes shown by SABG expression. Hepatic polyploidy in cells with oxidative injury attenuated cellular proliferative capacity. The findings advance insights into cellular polyploidy by further establishing that oxidative DNA damage in the setting of mitogenic stimulation induces polyploidy.

The findings provide direct evidence for impaired postmitotic activity in polyploid cells, which has implications in liver growth control (Gupta,2000EF15). Moreover, our findings should have wider biological implications because polyploidy is an ubiquitous process in tissues and organs, particularly after various types of hypertrophic stimuli, e.g., hypertension (myocardial cells), lactation(mammary glands), or pregnancy (endometrial cells), microbial infection(lymphocytes), oncogenesis (multiple tumors and cell lines) etc. (Brodsky and Uryvaeva, 1977EF6; Gupta,2000EF15). In other situations,reactive oxygen species are generated after exposure to xenobiotics, metals,such as iron and copper, mineral dusts and chemotherapeutic drugs. The hyroxyl-free radicle reacts aggressively with other molecules leading to 8-hydroxy adducts of guanine (Toraason et al.,1999EF49). Consequential DNA repair results in 8-hydroxy-2′-deoxyguanosine adducts, as shown above. Further evidence for the role of oxidative injury in polyploidy is provided by studies showing that in transgenic mice overexpressing copper-zinc-superoxide dismutase and glutathione peroxidase, which are antioxidants, PH-induced hepatic polyploidization is decreased (Nakatani et al.,1997EF34). Similarly, treatment with aminoguanidine, which attenuates oxidative stress, decreased polyploidy(Diez-Fernandez et al.,1998EF11).

If one considers models of cell lineage-dependent organ development, where replacement of epithelial cells helps maintain structure-proliferation units(Slack, 2000EF44), extreme polyploidy would be expected to be associated with cell senescence and eventual cell loss. Sigal et al. (Sigal et al.,1999EF42) showed that SABG expression, p21 expression and apoptosis were induced by PH in remnant hepatocytes, which are associated with terminal cell differentiation. Although genetic regulation of polyploidy has not been fully defined, it is noteworthy that in p21 transgenic mice hepatocytes become polyploid and have smaller liver lobules indicating the presence of fewer hepatocytes in the liver (Wu et al., 1996EF53). Studies using colon carcinoma cells showed that polyploidy, induced by p21, increases spontaneous apoptosis rates, as well as susceptibility to ionizing radiation(Waldman et al., 1997EF51). Although we did not study regulation of p21 or other activities in our cells,greater SABG activation in cultured hepatocytes and enhanced polyploidy after multiple oxidative injuries, i.e., partial hepatectomy and radiation, were in agreement with such models. We chose to study the effects of radiation in our cells for two specific reasons. First, ionizing radiation is well known to induce oxidative stress, through physical changes, without metabolic activation and the possible subsequent perturbations in metabolic pathways as with chemical inducers of oxidative stress. Secondly, radiation could be administered equally to all cells, irrespective of their ploidy class distributions, whereas use of chemical agents to induce oxidative stress could potentially have been affected by differences in the metabolic properties of various hepatocyte subpopulations (Rajvanshi et al.,1998EF39). Of course, radiation had the disadvantage of inducing some DNA damage directly rather than solely through the activation of oxidative stress. Therefore, additional analysis of polyploidy induction with drugs or chemicals known to cause oxidant stress through specific mechanisms should be helpful.

Use of rats prepared with retrorsine-PH to induce proliferation in transplanted hepatocytes established that polyploid hepatocytes from PH livers were less capable of producing daughter cells. Moreover, cell proliferation ceased almost completely when hepatocytes from PH livers were treated additionally with radiation. It was noteworthy that some cells subjected to PH alone, as well as radiated cells from the unperturbed normal liver, showed proliferation in our in vivo assay. Previous work by Overturf and colleagues(Overturf et al., 1997EF37)established that adult mouse hepatocytes may possess stem-cell-like properties with indefinite proliferation in intact mice. The studies involved transplantation of unfractionated hepatocytes from normal mice into the liver of genetically diseased FAH mice, in which tyrosinemia leads to selective and progressive loss of endogenous hepatocytes. However, it was unclear whether indefinite replication capacity was a feature of all transplanted hepatocytes or whether replication capacity was restricted to hepatocyte subsets. Our studies suggest that polyploid hepatocytes were significantly less capable of proliferating, although cellular proliferation capacity was lost only after more than one injury (PH plus radiation) and presumably extensive DNA injury. Attenuation of proliferative capacity in hepatocyte subpopulations is further substantiated by studies using diploid and polyploid rat hepatocytes fractionated from the normal liver (Rajvanshi et al.,1998EF39). Also, studies using`small hepatocytes' from the rat liver, which constitute diploid cell populations, indicate that small cells proliferate far more than larger hepatocytes (Mitaka et al.,1992EF33; Tateno et al.,2000EF48). An exception to these findings was reported by Grompe and colleagues using mouse hepatocytes, in which transplantation of diploid cells was less effective in clonogenic liver repopulation assays (Overturf et al.,1999EF36); however,cotransplantation of unfractionated cells could have confounded their analysis. Another set of studies from Sandgren and colleagues using mouse hepatocytes fractionated with flow cytometry was interpreted as showing no difference in the behavior of diploid and higher ploidy cells (Weglarz et al.,2000EF52). However, these studies did not exclude the possibility of DNA damage from laser energy during flow cytometric cell separations, which could have altered cell proliferation capacity, as suggested by limited proliferation in their diploid cells subjected to this manipulation. Therefore, our studies offer new information by clearly establishing restriction in the replication potential of polyploid cells subjected to oxidative DNA injury.

We found it of interest that polyploidy was induced when mitogenically stimulated hepatocytes were exposed to oxidative injury. We used TGFα to stimulate DNA synthesis in our cells, although other hepatic growth factors,such as hepatocyte growth factor, epidermal growth factor, etc., could also have been utilized (Gupta et al.,1992EF16; Sigal et al.,1999EF42). Hepatic stimulation with TGFα increases hepatic polyploidy in intact transgenic animals (Lee et al., 1992EF27). Costimulation of these cells with soluble signals (NE and VP) that mediate `stress'responses greatly amplified cellular polyploidization. These findings are especially relevant in the setting of partial hepatectomy, because NE is released after partial hepatectomy and VP is known to play significant roles in liver regeneration after partial hepatectomy (Russell and Bucher,1983EF41; Knopp et al.,1999EF25). In view of the wide distribution in tissues of NP, VP and related hormones and exposure of virtually all cells to oxidizing events, it is likely that our findings will be of physiological relevance to nonhepatic epithelia. Again, in structure-proliferation models of acinar renewal, interactions with extracellular matrix components, stromal cells, growth factors and ambient soluble signals, are critical elements in activating cell renewal versus terminal differentiation (Slack,2000EF44).

Models were recently proposed to demonstrate how senescence-type changes in polyploid hepatocytes would be relevant in diseases (Gupta,2000EF15). One consideration is that if an organ contains excessive numbers of polyploid cells with depletion of renewing cell units, organ failure may occur in the setting of continuing liver injury because polyploid cells will exhibit survival disadvantage. For instance, liver failure and death occur when hepatocytes fail to survive in the setting of hepatic polyploidy, such as in mutant mice with impaired nucleotide excision and repair (McWhir et al.,1993EF31). Secondly, to escape this fate, nonpolyploid cell clones with resistance to ongoing disease processes may emerge and confer greater propensity for oncogenesis. Our findings of impaired proliferation capacity in polyploid cells are in agreement with these possibilities and should thus provide further conceptual frameworks in areas concerning organ development, regeneration and oncogenesis.

From a translational perspective, induction of polyploidy with oxidative DNA injury will have potential for therapeutic liver repopulation with transplanted cells. Work from our laboratory and other laboratories has shown that selective proliferation of transplanted cells is needed for significant liver repopulation (Rhim et al.,1994EF40; Overturf et al.,1997EF37; Mignon et al.,1998EF32; Laconi et al.,1998EF26; Gupta et al.,1999EF18; Guha et al.,1999EF14). Extensive polyploidy is induced in rats treated with retrorsine and PH (Gupta,2000EF15), which as shown here,permits proliferation of transplanted cells in the liver. Retrorsine can be combined with repeated tri-iodothyronine (T3) instead of PH for inducing transplanted cell proliferation (Oren et al.,1999EF35). We consider that T3 may be effective because thyroid hormones regulate polyploidy following PH (Torres et al., 1999EF50). Similarly, the combination of radiation and PH, which induces hepatic polyploidy, as shown here, permits extensive transplanted cell proliferation (Guha et al.,1999EF14). Finally, when we extended our findings and conditioned the host rat liver with oxidative hepatic injury using radiation and ischemia-reperfusion, it became possible to repopulate virtually the entire liver (Malhi et al.,2000EF30; manuscript in preparation). Therefore, additional strategies to induce polyploidy in endogenous cells by oxidative DNA injury should facilitate organ repopulation with unaffected normal cells and help obtain further insights into biological mechanisms in the context of cellular polyploidy.

The work was supported in part by the Irma T. Hirschl Trust and NIH grants R01 DK46952, P30-DK-41296, P30-CA13330.