Hyperproliferation and p53 status of lens epithelial cells derived from alphaB-crystallin knockout mice.

alphaB-Crystallin, a major protein of lens fiber cells, is a stress-induced chaperone expressed at low levels in the lens epithelium and numerous other tissues, and its expression is enhanced in certain pathological conditions. However, the function of alphaB in these tissues is not known. Lenses of alphaB-/- mice develop degeneration of specific skeletal muscles but do not develop cataracts. Recent work in our laboratory indicates that primary cultures of alphaB-/- lens epithelial cells demonstrate genomic instability and undergo hyperproliferation at a frequency 4 orders of magnitude greater than that predicted by spontaneous immortalization of rodent cells. We now demonstrate that the hyperproliferative alphaB-/- lens epithelial cells undergo phenotypic changes that include the appearance of the p53 protein as shown by immunoblot analysis. Sequence analysis showed a lack of mutations in the p53 coding region of hyperproliferative alphaB-/- cells. However, the reentry of hyperproliferative alphaB-/- cells into S phase and mitosis after DNA damage by gamma-irradiation were consistent with impaired p53 checkpoint function in these cells. The results demonstrate that expression of functionally impaired p53 is one of the factors that promote immortalization of lens epithelial cells derived from alphaB-/- mice. Fluorescence in situ hybridization using probes prepared from centromere-specific mouse P1 clones of chromosomes 1 and 9 demonstrated that the hyperproliferative alphaB-/- cells were 30% diploid and 70% tetraploid, whereas wild type cells were 83% diploid. Further evidence of genomic instability was obtained when the hyperproliferative alphaB-/- cells were labeled with anti-beta-tubulin antibodies. Examination of the hyperproliferative alphaB-/- mitotic profiles revealed the presence of cells that failed to round up for mitosis, or arrested in cytokinesis, and binucleated cells in which nuclear division had occurred without cell division. These results suggest that the stress protein and molecular chaperone alphaB-crystallin protects cells from acquiring impaired p53 protein and genomic instability.

␣B-Crystallin, a major protein of lens fiber cells, is a stress-induced chaperone expressed at low levels in the lens epithelium and numerous other tissues, and its expression is enhanced in certain pathological conditions. However, the function of ␣B in these tissues is not known. Lenses of ␣B؊/؊ mice develop degeneration of specific skeletal muscles but do not develop cataracts. Recent work in our laboratory indicates that primary cultures of ␣B؊/؊ lens epithelial cells demonstrate genomic instability and undergo hyperproliferation at a frequency 4 orders of magnitude greater than that predicted by spontaneous immortalization of rodent cells. We now demonstrate that the hyperproliferative ␣B؊/؊ lens epithelial cells undergo phenotypic changes that include the appearance of the p53 protein as shown by immunoblot analysis. Sequence analysis showed a lack of mutations in the p53 coding region of hyperproliferative ␣B؊/؊ cells. However, the reentry of hyperproliferative ␣B؊/؊ cells into S phase and mitosis after DNA damage by ␥-irradiation were consistent with impaired p53 checkpoint function in these cells. The results demonstrate that expression of functionally impaired p53 is one of the factors that promote immortalization of lens epithelial cells derived from ␣B؊/؊ mice. Fluorescence in situ hybridization using probes prepared from centromere-specific mouse P1 clones of chromosomes 1 and 9 demonstrated that the hyperproliferative ␣B؊/؊ cells were 30% diploid and 70% tetraploid, whereas wild type cells were 83% diploid. Further evidence of genomic instability was obtained when the hyperproliferative ␣B؊/؊ cells were labeled with anti-␤-tubulin antibodies. Examination of the hyperproliferative ␣B؊/؊ mitotic profiles revealed the presence of cells that failed to round up for mitosis, or arrested in cytokinesis, and binucleated cells in which nuclear division had occurred without cell division. These results suggest that the stress protein and molecular chaperone ␣B-crystallin protects cells from acquiring impaired p53 protein and genomic instability.
␣〉-crystallin (␣B) 1 is a stress-inducible chaperone expressed in lens fiber cells and in the lens epithelium. It has also been detected at varying levels in numerous tissues, particularly heart, kidney, skeletal muscle, lung, and brain (1)(2)(3). However, the function of ␣B in these tissues remains elusive. A number of recent studies suggest a broad cellular function of ␣B associated with growth (4 -13). ␣Aand ␣B-crystallins have autokinase activity (4). ␣B expression increases in mitotically arrested NIH 3T3 fibroblasts, and in Chinese hamster ovary cells induced to express ␣B, the localization of ␣B changes from the cytoplasm to nucleus during interphase (5,6). Nearly 20% of ␣B polypeptides isolated from the lens are phosphorylated in vivo, suggesting an association with signal transduction pathways (8,12). Phosphorylated forms of ␣B have been detected in the centrosomes and midbodies of dividing cells (7). ␣B also accumulates in the developing central nervous system (particularly in glial cells), in fibroblasts transformed by the v-mos and Ha-ras oncogenes, and in brain tumors of neuroectodermal origin (9 -16). Enhanced expression of ␣B is also observed in numerous neurodegenerative diseases, such as Alexander's, Alzheimer's, Creutzfeldt-Jakob, multiple sclerosis, and Parkinson's diseases (13)(14)(15).
␣BϪ/Ϫ mice develop muscle abnormalities but do not develop cataract (17). The major phenotype in these mice, lacking both ␣B and small heat shock protein HSPB2 (encoded by an adjacent gene, which was also disrupted), is a progressive muscular dystrophy, which destroys a specific subset of skeletal muscles, particularly those in the head and perivertebral regions. Because HSPB2 is not expressed in the lens (18), its deletion is not expected to contribute to lens phenotypes. Lens epithelial cells derived form ␣BϪ/Ϫ mice demonstrate hyperproliferation and genomic instability (19). In contrast, cells derived from mice lacking ␣A-crystallin (␣A), a protein normally associated with ␣B in the lens, have a diminished growth potential (20). Additional studies indicate that the mutation of a conserved arginine residue (Arg-120) to cysteine in ␣B results in desmin-related myopathy, and is associated with congenital cataracts in humans (21). The mutated protein has a higher oligomeric mass, an altered tertiary structure, and a diminished chaperone-like activity in vitro (22).
A number of studies suggest that molecular chaperones play important roles in cell growth and differentiation (5)(6)(7)(23)(24)(25)(26)(27). Heat shock protein 70 (HSP70), HSP90, and TriC may control the quality of proteins during cell cycle progression (24,25). TriC is strongly up-regulated during cell growth and may contribute to the folding of tubulin and other proteins (24). HSP70 and HSP90 have been detected in centrosomes of mitotic cells, and antibodies to ␣B recognize midbodies and centrosomes (7). ␣B expression is also up-regulated in mitotic fibroblasts (6). ␣A and ␣B are detected during the early stages of lens development (28) and may also be important for proper lens growth.
The tumor suppressor protein p53 plays a central role in protecting cells from deregulated growth (29 -32). It does so by acting as a transcription factor that induces the expression of genes that induce cell cycle arrest, DNA repair, and/or apoptosis after one of several stress signals has led to its activation. p53 has a very short half-life in unstressed cells, because of a negative feedback mechanism in which binding to the protein Mdm2 results in its degradation. Stress signals that can activate p53 are initiated by agents that produce DNA damage, hypoxia, oncogene activation, deregulation of microtubule assembly, and other forms of cellular trauma (31)(32)(33)(34). Nearly half of all human cancers carry missense mutations in the p53 gene. Mutant p53 proteins do not drive the negative feedback loop, and are therefore present at abnormally high levels in cancer cells. A proper balance of p53 function is essential for normal growth and homeostasis in the lens. The overexpression of wild type p53 in mouse lenses leads to apoptosis and microophthalmia (35).
Other studies indicate that p53 plays a critical role in cell cycle arrest after DNA damage and prevents the reentry of cells into the cycle after spindle disruption (36 -40). Following DNA damage, cells arrest at the G 1 -S or G 2 -M transition depending upon cell type, growth conditions, and the checkpoint controls operative in the cell (36). In cells with wild type p53, DNA damage usually results in arrest at the G 1 -S boundary (36,(41)(42)(43)(44). However, cells arrest in G 2 in the absence of functional p53. This arrest results from the inactivation of a protein kinase that is required for mitotic entry (41)(42)(43)(44).
The induction of p53 protein is manifested by increased levels and activity of the protein through post-translational mechanisms such as phosphorylation, which inhibit its interaction with Mdm2, thereby stabilizing it in cells (30 -32). Numerous other mechanisms by which cells regulate the p53-Mdm2 feedback loop have also been identified (31,32,45). Recently it has been shown that compounds such as hydroxyurea and aphidicolin, which arrest cells in S phase, increase p53 levels in cells and concomitantly trigger a process by which p53 activity is functionally impaired (33). Other studies show that a functionally inactive p53 accumulates by DNA damage or cellular differentiation in mouse teratocarcinoma cells (46). High concentrations of functionally inactive p53 are also found in adenovirus-transformed human cells (47).
Our previous work with lens epithelial cells cultured from ␣B knockout mice showed that deletion of the ␣B gene permits the emergence of lens epithelial cells with vastly increased proliferative ability (19). In the present work we elucidated mechanistically how ␣B interacts with proteins involved in cell cycle regulation. We carried out experiments that were a direct extension of our previous study, and confirmed and extended the primary conclusions of that study using a second independently derived line of ␣BϪ/Ϫ mice. Furthermore, we used specific mouse chromosome probes and fluorescent in situ hybridization, and demonstrate the presence of tetraploid cells in the ␣BϪ/Ϫ cultures. We also show that hyperproliferative ␣BϪ/Ϫ cells contain a detectable level of p53 protein, suggesting that the protein has been stabilized in these cells. We also assessed the checkpoint function of p53 in mouse wild type and hyperproliferative ␣BϪ/Ϫ lens epithelial cells after exposure to DNA damaging agents. We demonstrate here that hyperproliferative ␣BϪ/Ϫ cells reenter the S phase and mitosis after treatment with ␥-radiation, which may be the result of the loss of p53 checkpoint function in these cells. We also demonstrate that the ␣BϪ/Ϫ cells infected with Ad12-SV40 hybrid virus to express the SV40 large T antigen undergo apoptosis and cell extinction, in contrast to wild type lens epithelial cells, which remain viable.

EXPERIMENTAL PROCEDURES
Animals-The wild type mouse strain 129Sv was obtained from Taconic Laboratories. ␣B52Ϫ/Ϫ and ␣B168Ϫ/Ϫ mice were derived from two independent embryonic stem cell clones, which resulted from the targeted disruption of the ␣B gene and the adjacent HSPB2 gene (17). The absence of HSPB2 transcripts by Northern blotting (17) and RT-PCR, and of the HSPB2 protein by Western immunoblotting, 2 leaves little doubt that the latter gene is not expressed in the lens and is not expected to give lens phenotypes (17).
Lens Morphology-Wild type, ␣B52Ϫ/Ϫ, and ␣B168Ϫ/Ϫ mice were euthanized, and eyes were removed. A small incision was made with a sharp scalpel, near the ciliary body, to promote penetration of the fixative. The eyes were fixed in phosphate-buffered 4% formaldehyde, embedded in methacrylate, sectioned, and stained with hematoxylin and eosin.
Cell Culture-Mouse lens epithelial cell cultures were utilized in this study. Lenses were obtained from 6 -12-week-old mice. To obtain primary lens epithelial cells, capsule/epithelial explants of wild type (129Sv) and ␣BϪ/Ϫ mouse lenses were dissected and cultured in 20% fetal bovine serum-minimal essential medium in 35-mm tissue culture plates as described previously (19,20), and cultured for 1-2 weeks until growth of cells was detectable. Cells were passaged using trypsin-EDTA, and plated in 35-mm plates. In some experiments, cells were passaged and plated on glass coverslips to facilitate microscopic examination. Cultures were fed twice weekly. Unless indicated otherwise, all experiments were performed using lens epithelial cells derived from both ␣B52Ϫ/Ϫ and ␣B168Ϫ/Ϫ lines of mice.
To determine cell growth rates, primary cultures of wild type or ␣B52Ϫ/Ϫ lens epithelial cells in passage 1 were subcultured in 24-well plates (2 ϫ 10 3 cells/well) for a period of 2 weeks. Attachment efficiency determined 2 h after plating was found to be 85%, and was the same for wild type and ␣BϪ/Ϫ cells. Cell numbers were measured with a Coulter counter (19). Hyperproliferative ␣BϪ/Ϫ cells were plated at low densities to examine clonal growth (19). To determine if the hyperproliferative ␣B52Ϫ/Ϫ cells were immortal, they were subcultured every week, the number of cells was counted, and the population doubling level was determined as described previously (48).
Two control cell cultures were also used. The human lung carcinoma cell line H1299, which expresses a mutant form of p53, lacking functional p53, was utilized as a positive control for mutant p53 expression (49). In addition, lens epithelial cells were cultured from p53Ϫ/Ϫ mice (Jackson Laboratories). These mice carry a mutation in p53 that eliminates the production of any p53 protein in its cells (50). Cells were grown to confluence, and lysates were prepared for SDS-PAGE and Western blot analysis as described previously (48).
Immunofluorescence-Cells were fixed in 4% paraformaldehyde/PBS for 30 min at room temperature, permeabilized in 0.1% Triton X-100 for 30 min, and blocked with 10% goat serum. To visualize the distribution of ␣B, cells were incubated overnight at 4°C in a 1:50 dilution of a polyclonal antibody raised against bovine ␣B (Novocastra Laboratories). This primary antibody was used because it gave a very low background in immunocytochemistry with the ␣BϪ/Ϫ mouse lens slices. An Alexa 488 -conjugated goat anti-rabbit IgG was used as the secondary antibody. To visualize the organization of f-actin, fixed and permeabilized cells were incubated in a 1:50 dilution in PBS of a Texas Red-phalloidin (Molecular Probes Inc.) methanolic stock solution (100 units/ml of methanol). Cells were stained with Texas Red-phalloidin for 20 min, washed three times for 5 min each time in PBS, and viewed. Lens epithelial cells were viewed using a Zeiss LSM 410 confocal microscope equipped with an argon-krypton laser (19).
The distribution of ␣A was also visualized in wild type and ␣B52Ϫ/Ϫ cells as described previously (19). Briefly, cells were incubated overnight at 4°C in a 1:40 dilution of a monoclonal antibody against bovine ␣A. This primary antibody was used because it has been shown to react with a high specificity with ␣A in immunocytochemical detection (20), and gave no background with ␣AϪ/Ϫ mouse lens slices. An Alexa 568conjugated goat anti-mouse IgG was used as the secondary antibody.
␤-Tubulin was visualized by immunofluorescence using a monoclonal antibody (Sigma). DNA was labeled with the dye TOTO-1 (Molecular Probes, Inc.).
Fluorescence in Situ Hybridization (FISH) Analyses-Labeled probes were prepared by Incyte Genomics, Inc. (St. Louis, MO) from centromere-specific mouse P1 clones for chromosome 1 and chromosome 9 (51). Chromosome 1 probe was labeled with digoxigenin and detected with anti-digoxigenin FITC, and chromosome 9 probe was labeled with biotin and detected with Texas Red-avidin. The conditions for hybridization and post-hybridization washes were standard procedures as provided by the vendor. Slides were denatured at 70°C for 2 min in 70% formamide and 2ϫ SSC. Hybridization was performed for 24 h. Chromosomes were counterstained with 4,6-diamidino-2-phenylindole, and slides were examined using the Zeiss Axiophot epifluorescence microscope equipped with the appropriate filter combinations. Acquired images were digitized with a fluorescence microscope/CCD camera interfaced to a computer work station. In some experiments, stained nuclei were viewed using a Zeiss LSM 410 confocal microscope equipped with an argon-krypton laser. FISH analyses were performed at the Children's Hospital of Michigan and in the authors' laboratory at Washington University.
Ad12-SV40 Hybrid Virus Infection-Extended life span mouse lens epithelial cell lines were generated by infection of primary mouse lens epithelial cultures with Ad12-SV40 hybrid virus using a published procedure (48,52). Briefly, wild type, normally growing ␣B52Ϫ/Ϫ or hyperproliferative ␣B52Ϫ/Ϫ cells were infected with Ad12-SV40 hybrid virus in normal growth medium overnight. The medium was replaced with fresh growth medium. Cells were subcultured after 2-4 days. In our studies, we found that ␣BϪ/Ϫ cells expressing the SV40 T-antigen undergo widespread cell death (Ͼ90%) after attaining confluence. Therefore, to prevent cell death, SV40 T antigen-expressing cultures were observed daily and passaged before they reached confluence.
Analysis of Cell Death-The procedure for measuring cell death has been described previously (52). Briefly, annexin V-FITC (Pharmingen), a Ca 2ϩ -dependent phospholipid-binding protein that binds to the plasma membrane of cells in early stages of apoptosis, was used. Propidium iodide (PI) binds to cells that have ruptured their membranes and died. To distinguish apoptotic cells (annexin-positive/PI-negative) from necrotic cells (annexin-negative/PI-positive) or those cells which had already died via apoptosis (annexin-positive/PI-positive), attached cells were trypsinized, combined with cells floating in the medium, and labeled with annexin V-FITC and propidium iodide according to the guide from the manufacturer (Pharmingen). Cells were washed with PBS and resuspended in 500 l of annexin-binding buffer. Annexin V-FITC (5 l) and propidium iodide (10 l of a 50 g/ml stock solution) were added, and flow cytometry was performed using a BD Biosciences FACScan as described previously (52). The flow cytometer had a laser excitation beam at 488 nm. Band pass filters at 530 and 685 nm were used to collect the fluorescence emission, and data collected in each window were designated FL-1 and FL-2 fluorescence, respectively. Unlabeled cells, cells labeled with annexin only, and cells labeled with PI only were used as controls to adjust the compensation between the flow cytometer and the fluorescence detectors and to set the quadrants. Data analyses were performed with CellQuest software.
Live cells were not labeled with either annexin or PI, and were identified in the lower left quadrant of the fluorescence-activated cell sorting data, as shown in Fig. 7. Annexin-positive but PI-negative cells were interpreted as cells undergoing apoptosis. These cells were identified in the lower right quadrant. Cells that had already died by apoptosis were both annexin-positive and PI-positive, and these cells were identified in the upper right quadrant. PI-positive but annexinnegative cells, identified in the upper left quadrant, were defined as cells that had already died by necrosis.
␥-Irradiation-Lens epithelial cells were plated in 35-mm plates and grown to confluence. Cells were washed with PBS at 37°C and irradiated with a 137 Cs ␥-irradiator for times ranging from 0.28 to 2.2 min to give doses between 5 and 40 Gy (500 -4,000 rads). After irradiation, cells were incubated in normal growth medium for 24 h, and then prepared for cell cycle distribution analysis. In other experiments, ␥-irradiated cells were labeled with BrdUrd or treated with nocodazole, and the ability of the cells to reenter the cell cycle was determined as described below.
Cell Cycle Distribution-Wild type or ␣BϪ/Ϫ cells (10 6 cells) were washed with PBS, and cell pellets were labeled for 30 min on ice with propidium iodide (50 g/ml) in 0.1% sodium citrate containing 0.3% Nonidet P-40 and 20 g/ml ribonuclease A (pH 8.3). The percentage of cells in each phase of the cell cycle was determined using a flow cytometer (BD Biosciences FACScan) and analyzed using the Cell Quest software (19,20).
Detection of p53-Immunoblot analysis was used to examine the expression of p53 in cell cultures. The antibody used for immunoblot analysis was monoclonal antibody to human p53 (clone 240 from Novocastra) at a dilution of 1:100. This antibody recognizes both wild type and mutant forms of p53 under denaturing conditions. Lysates from 10 6 wild type, ␣BϪ/Ϫ, hyperproliferative ␣BϪ/Ϫ lens epithelial cells, the human lung carcinoma cell line H1299, and p53Ϫ/Ϫ primary mouse lens epithelial cells were prepared and analyzed by SDS-PAGE using 15% gels. Proteins were transferred to polyvinylidene difluoride membranes and probed with the antibody to p53. The secondary antibody was horseradish peroxidase-labeled goat anti-mouse IgG. Blots were incubated with Luminol reagent (Santa Cruz Biotechnology) and exposed to Kodak film to visualize the protein bands.
Sequencing of p53 cDNA-RNA from wild type and hyperproliferative ␣BϪ/Ϫ lens epithelial cells was reverse transcribed into cDNA. PCR was performed by using the primers 5Ј-ATATCAGCCTCGAGCTC-CCT-3Ј and 3Ј-CTGTAGCATGGGCATCCTTT-5Ј. These primers are based on the p53 cDNA sequence (accession no. AB017815), targeting a 1032-bp sequence. These PCR products were cloned into the cloning vector TOPO (Invitrogen) and sequenced by an ABI sequencer at the Protein and Nucleic Acid Chemistry Laboratory at Washington University. Sequences were aligned using the Sequencher program. The cloned sequence allowed us to screen for possible mutations between codon 11 and codon 353 of mouse p53 cDNA. Hyperproliferative lens epithelial cell lines derived from both lines of mice (␣B52Ϫ/Ϫ and ␣B168Ϫ/Ϫ) were used.
Detection of Retinoblastoma Protein Rb-Immunoblot analysis was used to examine the expression of Rb protein in cell cultures. The antibody used for immunoblot analysis was monoclonal antibody to Rb (NCL-Rb from Novocastra) at a dilution of 1:100. Lysates from 10 6 wild type, normally growing ␣BϪ/Ϫ, hyperproliferative ␣BϪ/Ϫ lens epithelial cells, and p53Ϫ/Ϫ primary mouse lens epithelial cells were prepared and analyzed by SDS-PAGE using 6% gels. Whole cell lysates from the K562 cells (Santa Cruz Biotechnology) was used as a positive control for Rb. Proteins were transferred to polyvinylidene difluoride membranes and probed with the antibody to Rb. The secondary antibody was horseradish peroxidase-labeled goat anti-mouse IgG. Blots were incubated with Luminol reagent (Santa Cruz Biotechnology) and exposed to Kodak film to visualize the protein bands.
BrdUrd Labeling-Wild type p53 has been shown to prevent cells in G 1 from entering S phase following DNA damage (36 -38, 41-43). Our data suggested that hyperproliferative ␣BϪ/Ϫ cells accumulated functionally inactive forms of p53, which may decrease the ability of the cells to maintain cell cycle arrest after ␥-irradiation (41)(42)(43)(44). We therefore used wild type or hyperproliferative ␣BϪ/Ϫ cultures to determine the fraction of cells that entered S phase after exposure to the DNA damaging agent ␥-radiation. Wild type or hyperproliferative ␣BϪ/Ϫ cells were grown to 80 -90% confluence as monolayers in 35-mm plates and exposed to a 137 Cs source (GammaCell 40) for a dose of 4,000 rads (40 Gy). Cells were untreated or treated with ␥-radiation, incubated for 72 h, and then pulsed with BrdUrd for 30 min and fixed. BrdUrd was detected by immunofluorescence as described previously (40,53). The DNA stain TOTO-1 was used to count the total number of cells. Cultures were examined by confocal microscopy. The BrdUrd labeling index was determined by dividing the number of BrdUrd labeled cells by the total number of cells.
Mitotic Index-Wild type or ␣BϪ/Ϫ cells were grown to 80 -90% confluence as monolayers in 35-mm plates, and exposed to ␥-radiation from a 137 Cs source (GammaCell 40) for a dose of 40 Gy. Mitotic trapping experiments were done, by adding nocodazole (0.2 g/ml) to the culture medium as described (36). Cells were collected by incubation with trypsin containing EDTA, centrifuged, and fixed in 4% formaldehyde, and TOTO-1 was added in PBS. TOTO-1-stained nuclei were visualized by confocal microscopy. Nuclei with condensed, uniformly stained chromosomes were scored as mitotic. In each culture, 100 cells were examined for each determination.

RESULTS
Morphology of ␣BϪ/Ϫ Mice-Lines of ␣B/HSPB2-deficient mice were generated from two independent embryonic stem cell clones, clones 52 and 168, both exhibiting the correctly targeted gene locus. Mice in these two lines are similar, both losing body fat and weight, developing a hunched posture, and having life spans approximately half that of wild type mice. Moreover, the gross morphology of 49-week-old ␣B52Ϫ/Ϫ lenses (Fig. 1) re-sembles that of similarly aged ␣B168Ϫ/Ϫ and wild type mice (17), indicating that these two independently derived mouse lines have, as expected, similar morphology.
Expression of ␣B in Cultured Lens Epithelial Cells-Primary lens epithelial cells were cultured from wild type and ␣BϪ/Ϫ mice. The expression of ␣B was examined in wild type and ␣B52Ϫ/Ϫ mouse lens epithelial cells (Fig. 2). ␣B was readily detectable in the cytoplasm of wild type mouse lens epithelial cells by immunofluorescence but was absent in the cultures derived from ␣B52Ϫ/Ϫ lenses (Fig. 2, A and B). The organization of the actin cytoskeleton in wild type and ␣BϪ/Ϫ lens epithelial cells was also examined. There were no significant differences in the actin cytoskeleton of wild type and ␣BϪ/Ϫ cultured lens epithelial cells. The lack of ␣B in the cultures derived from ␣BϪ/Ϫ lenses was also confirmed by immunoblot analysis (Fig. 2D).
In the lens, ␣B is normally associated with ␣A in large heteromeric complexes (19). Therefore, we also examined the expression of ␣A in cultured lens epithelial cells. ␣A expression was detected in the cytoplasm of the wild type and ␣B52Ϫ/Ϫ cells, and the levels of ␣A were similar in wild type and ␣B52Ϫ/Ϫ lens epithelial cells, as shown by immunoblot analysis (Fig. 2C). Moreover, immunofluorescence studies showed that the distribution of ␣A in the wild type and ␣B52Ϫ/Ϫ cells was similar (data not shown), indicating that disruption of the ␣B gene does not affect the expression and distribution of ␣A in lens epithelial cells.
Previous work in our laboratory showed that primary lens epithelial cells derived from ␣B168Ϫ/Ϫ mice demonstrated an increased frequency of hyperproliferation (19). Fast growing colonies were also observed in the lens epithelial cultures derived from the second line of mice, ␣B52Ϫ/Ϫ (Fig. 3). ␣B52Ϫ/Ϫ lens epithelial cells, like ␣B168Ϫ/Ϫ cells, produced hyperproliferative clones at a high frequency (10 Ϫ1 ), 4 orders of magnitude greater than that predicted for spontaneous immortalization of rodent cells, 10 Ϫ5 to 10 Ϫ6 (54). The hyperproliferative cells had a doubling time of 1.5 days as compared with 3 days for the wild type or normally growing ␣B52Ϫ/Ϫ cells. Primary cultures of wild type cells and normally growing ␣B52Ϫ/Ϫ cells could only be propagated up to ϳ7 population doublings. However, hyperproliferative ␣B52Ϫ/Ϫ cells grew at an undiminished rate for Ͼ100 population doublings. These results indicate that the hyperproliferative ␣B52Ϫ/Ϫ cells, like ␣B168Ϫ/Ϫ cells (19), were truly immortal.
Although ␣B was primarily expressed in the cell cytoplasm during interphase ( Fig. 2A), its distribution during mitosis in lens epithelial cells is not known. We therefore examined the subcellular localization of ␣B during cell division by immunofluorescence and confocal microscopy. In these studies, the chromosomes were stained with the DNA-binding dye TO-PRO-3. The mitotic spindle was examined by the immunofluo- Cells were cultured in minimal essential medium containing 20% serum, trypsinized, and counted in a Coulter counter. The hyperproliferative ␣B52Ϫ/Ϫ cells maintained their growth potential over Ͼ100 population doubling levels, whereas the wild type and normal-growing ␣B52Ϫ/Ϫ cells had a low growth potential, and did not grow beyond population doubling levels Х7.
During our examination of mitotic profiles in wild type and hyperproliferative ␣BϪ/Ϫ cells, we noted a number of unusual mitotic profiles in the hyperproliferative cells (Fig. 5B). Aberrant mitotic profiles have been reported in various cells lacking central spindle components or other proteins (55)(56)(57). In these studies, wild type and hyperproliferative ␣BϪ/Ϫ cells were labeled with an antibody to ␤-tubulin to label microtubules, and chromosomes were stained with the DNA binding dye TOTO-1. Mitotic stages from prometaphase, metaphase, early and late anaphase, cytokinesis, and telophase were observed in both wild type and ␣BϪ/Ϫ cells (Fig. 5A, a-f). However, a significant proportion of mitotic cells in the hyperproliferative ␣BϪ/Ϫ cultures were found to be arrested in anaphase or cytokinesis (Fig. 5B, h and i), and had adopted a G 1 -like state with a flattened appearance. Some cells did not round up at prometaphase as chromosome condensation began (Fig. 5B, g). A significant proportion (5%) of the cells were also binucleate (Fig. 5B, j) with decondensed chromosomes, indicating that nuclear division had occurred without cell division. The binucleate cells were examined in 20 different fields in three independent cell lines of hyperproliferative ␣BϪ/Ϫ cells. Importantly, binucleated cells or cells arrested in cytokinesis were not detected in wild type cultures.
FISH Analysis-Previous studies showed that ␣BϪ/Ϫ cultures contained a significant proportion of tetraploid and polyploid cells (19). To confirm and extend this result, we used FISH analysis of wild type and ␣BϪ/Ϫ cells. The mouse ␣B gene is on chromosome 9. The distribution of signals for chromosome 9 was analyzed in nuclei and metaphase chromosome spreads of wild type and ␣BϪ/Ϫ cells. Chromosome 1 signals were also analyzed as a nonspecific control. The distribution of signals observed in nuclei and metaphases of lens epithelial cells is shown in Fig. 6A. It can be readily noted that, in the wild type mouse lens epithelial cells, a majority of the nuclei were diploid, and a minor proportion were tetraploid (Ն4N). Of 53 nuclei examined, 83% cells were diploid and 17% were tetraploid. In contrast, hyperproliferative ␣BϪ/Ϫ cells were 30% diploid and 70% tetraploid. These results confirmed our previous results obtained using karyotypic analysis of metaphase cells of mouse ␣BϪ/Ϫ primary lens epithelial cultures (19).
Ad12-SV40 Hybrid Virus Infection-Wild type and ␣BϪ/Ϫ cells (both normally growing and hyperproliferative cultures) were infected with Ad12-SV40 hybrid virus to increase their growth potential (48,52,58). The infected cells expressed the SV40 T antigen in their nuclei (data not shown). Cell lines derived from wild type or ␣B52Ϫ/Ϫ cells were either grown to the subconfluent state (80 -90% confluence) or allowed to grow to confluence and cultured for 2 days after confluence. The level of apoptosis was analyzed by annexin labeling and flow cytometry, at each confluence state. The ␣B52Ϫ/Ϫ cells that were allowed to remain in confluent culture died by apoptosis (Fig.   FIG. 4. Distribution of ␣B in mitotic cells. Wild type mouse lens epithelial cells were fixed and stained with anti-␣B-crystallin and anti-␤-tubulin antibodies, and DNA was stained with TOPRO-3. In panels a1-c1, staining of ␤-tubulin (red) and DNA (blue) is shown. In panels a2-c2, staining of ␣B (green) is shown. Cells in metaphase (a1 and a2), mid-cytokinesis (b1 and b2), and late cytokinesis (c1 and c2) are shown. Note that ␣B is distributed throughout the cell cytoplasm but is excluded from the chromosomes at all stages of mitosis, and is detected in the midbody during late cytokinesis (panel c2). Note also that the chromosomes have begun decondensing in one of the daughter cells during late cytokinesis (panel c1). Scale bar (a1-c2) ϭ 10 m.

FIG. 5. Visualization of mitotic figures in mouse primary wild type and hyperproliferative ␣B؊/؊ lens epithelial cells reveals abnormal mitosis in ␣B؊/؊ cells.
Cells were fixed and immunostained with anti-␤-tubulin antibody and visualized using a fluorescent secondary antibody (red). DNA was stained with TOTO-1 (green). Wild type cells are shown in panels a-f and ␣B Ϫ/Ϫ cells in panels g-j. A, wild type cells. Cells in prometaphase (a), metaphase (b), anaphase (c), mid-cytokinesis (d), late cytokinesis (e), and telophase (f) are shown. B, hyperproliferative ␣BϪ/Ϫ cells. Cells did not round up for mitosis (g). Additionally, a significant number of cells that had arrested in cytokinesis and then adopted an interphase (G 1 ) appearance were observed (h and i). Further, the central spindle was less orderly in these cells than in the wild type cells. A significant fraction of the cells in the ␣BϪ/Ϫ culture also became binucleate (j). Aberrant mitotic profiles were observed in hyperproliferative lens epithelial cell lines derived from both ␣B52Ϫ/Ϫ and ␣B168Ϫ/Ϫ lines of mice. Scale bars for a-f and g, h, and j ϭ 10 m; panel i was visualized at a slightly lower magnification. 7), in contrast to wild type cells at any stage of confluence (Fig.  7A). SV40 T antigen induced an apoptosis mode of cell death in ␣BϪ/Ϫ cells only, and not in wild type or ␣AϪ/Ϫ cells. These results further indicate that the lack of expression of ␣B triggers an unusual cell death response in the ␣BϪ/Ϫ cells only, and this response may be related to the intrinsic genomic instability of these cells. When infected with Ad12-SV40 hybrid virus, the hyperproliferative ␣BϪ/Ϫ cells also died, but in a more protracted manner, 6 -7 days after reaching confluence, rather than 2 days for the normally growing ␣BϪ/Ϫ cells.
p53 and Rb Expression in ␣BϪ/Ϫ Cells-The transcription factor p53 plays an important role in maintaining genomic integrity of cells, and monitors cell cycle checkpoints and apoptosis. In normal cells, levels of p53 protein are very low (31)(32)(33)(34). In tumor cells or cells exposed to DNA damaging agents, p53 levels increase. Tumor cells accumulate mutant forms of p53 that are functionally inactive and can counteract the negative growth controls exerted by wild type p53 (31)(32)(33)(34)59). This loss of p53 function gives a growth advantage to tumor cells. Because lens epithelial cells derived from ␣BϪ/Ϫ mice demonstrate a great deal higher frequency of hyperproliferation than wild type cells, it was logical to consider that p53 protein may be involved. We hypothesized that the hyperproliferation observed in ␣BϪ/Ϫ cells could be a result of loss of FIG. 6. Fluorescence in situ hybridization in mouse wild type and hyperproliferative ␣B؊/؊ lens epithelial cells. Wild type and ␣BϪ/Ϫ cells were examined by fluorescence in situ hybridization using probes prepared from centromere-specific P1 clones for mouse chromosome 1 and chromosome 9. A, distribution of loci for chromosome 1 in wild type (top) and ␣BϪ/Ϫ (bottom) nuclei. The number of loci was predominantly 2 in the wild type cells. In ␣BϪ/Ϫ cells, the proportion of loci for chromosome 1 increased to 4. Similar results were obtained when chromosome 9 probe were used (data not included). B, visualization of chromosomes in mouse lens epithelial cells. a, dual immunofluorescence detection of a control slide showing the specificity of the two probes in a preparation of wild type mouse chromosomes. Chromosome 1 probe was labeled with digoxigenin and detected with antidigoxigenin-FITC (green). Chromosome 9 probe was labeled with biotin and detected with Texas Red-avidin (red). Note that two copies of each chromosome are detected in this preparation of a wild type diploid nucleus. b and c, wild type mouse lens epithelial cells were stained with a probe prepared from the centromere-specific P1 clone of mouse chromosome 1 (b) or chromosome 9 (c). The individual probes were labeled with digoxigenin and detected with antidigoxigenin-FITC on two different slides. Note that, in 83% of the nuclei labeled, two copies of each of the chromosomes were detected, indicative of diploid nature of these cells. d and e, ␣BϪ/Ϫ mouse lens epithelial cells were stained with a probe prepared from the centromere-specific P1 clone of mouse chromosome 1 (d) or chromosome 9 (e). The individual probes were labeled with digoxigenin and detected with antidigoxigenin-FITC on two different slides. Note that a significant number of the ␣BϪ/Ϫ cells had four or more copies of each chromosome, indicating that these cells are tetraploid or polyploid. p53 function and acquisition of mutant p53 in these cells. To determine the p53 levels in wild type and ␣BϪ/Ϫ cells, immunoblot analysis was performed. As shown in Fig. 8A, wild type cells or ␣BϪ/Ϫ cells with normal growth potential had very low or undetectable levels of p53, whereas hyperproliferative ␣BϪ/Ϫ cells had a detectable level of p53 (Fig. 8A, lane 4). The H1299 cell line, which is a human lung cancer cell line, lacking normal p53 function, was used as a positive control for mutant p53, and showed a pattern of bands significantly labeled with the p53 antibody. In contrast, lens epithelial cells derived from p53Ϫ/Ϫ mice (50) had no detectable p53 protein. These results clearly indicate that hyperproliferative ␣BϪ/Ϫ cells have accumulated p53, possibly because of the acquisition of mutations. The distribution of p53 protein appears to be nuclear, because it could be detected in the detergent-insoluble fraction (Fig.  8A), but was not detected in cytoplasmic fractions of the hyperproliferative ␣BϪ/Ϫ lens epithelial cells (Fig. 8B).
To determine whether the increased expression of p53 observed in the hyperproliferative ␣BϪ/Ϫ cultures was the result of the acquisition of mutations in the p53 gene, we focused our efforts on detecting mutations in the coding region of p53 using RT-PCR-based methods and gene-specific primers. These primers allowed us to screen for mutations between codon 11 and codon 353 of mouse p53 cDNA, a region encompassing the entire DNA-binding domain of p53 (60). Sequence analysis of the cloned PCR target sequences showed no changes to the sequence of the p53 gene, between wild type cells and two independently derived hyperproliferative ␣BϪ/Ϫ cell lines.
These results suggest that the increased expression of the p53 protein observed in hyperproliferative ␣BϪ/Ϫ cells was not the result of its stabilization by mutations in the p53 gene.
To investigate whether the cell cycle regulatory protein Rb plays a role in the hyperproliferation of ␣BϪ/Ϫ lens epithelial cells, we analyzed its expression in lens epithelial cells using immunoblot analysis. Rb blocks the G 1 to S phase transition, and is inhibited by phosphorylation (31). Hypophosphorylated Rb is active, and arrests cells in G 1 phase, whereas hyperphosphorylated Rb is inactive, and allows cells to progress into S phase. We used an antibody to Rb to look for multiple bands indicative of different phosphorylated forms of Rb. As shown in Fig. 8C, the majority of the Rb protein appears to be in a partially phosphorylated state in lens epithelial cells. No significant differences were observed between the molecular weight of the Rb protein in wild type, normally growing ␣BϪ/Ϫ cells and the hyperproliferative ␣BϪ/Ϫ lens epithelial cells.
These results indicate the Rb pathway was not significantly affected in the hyperproliferation of ␣BϪ/Ϫ lens epithelial cells.
p53 Function in Hyperproliferative ␣BϪ/Ϫ Cells-We then focused our efforts on determining whether p53 protein detected in hyperproliferative ␣BϪ/Ϫ cells was functionally active. p53 has been shown to prevent cells in G 1 phase from entering S phase following DNA damage by ionizing radiation (36 -38, 41-43). We analyzed the effect of increasing doses of ␥-irradiation on the cell cycle arrest of wild type and hyperproliferative ␣BϪ/Ϫ mouse lens epithelial cells. Flow cytometric analysis of propidium iodide-stained cells showed that wild type cells arrested in G 1 phase after ␥-irradiation (Fig. 9C). Dose-response studies were performed between 5 and 20 Gy of ␥-radiation. The extent of cell cycle arrest was proportional to the dose of ␥-radiation (Fig. 9A). Wild type cells had 50% cells in G 1 phase before treatment, but 24 h after ␥-irradiation (20 Gy), the proportion of cells in the G 1 phase increased to 70%. These observations confirm previous reports describing a p53dependent cell cycle arrest in G 1 following ␥-radiation (39 -43).
In contrast to cells with normal p53 function, cells that express impaired p53 protein have been shown to arrest in the G 2 phase of the cell cycle after exposure to ␥-irradiation (39 -44). Hyperproliferative ␣BϪ/Ϫ lens epithelial cells were exposed to 5-20 Gy of ␥-radiation (Fig. 9B). The ␣BϪ/Ϫ hyperproliferative cells showed a dose-dependent cell cycle arrest in G 2 -M after treatment with ␥-radiation, with 10% of the cells in G 2 -M before treatment, but 25% of the cells in G 2 -M 24 h after 20 Gy of ␥-irradiation (Fig. 9, B and D). Similar G 2 -M arrest has been shown for numerous cell lines with mutant and functionally impaired p53, consistent with the observation (Fig. 8A) that the hyperproliferative ␣BϪ/Ϫ cells have accumulated functionally inactive p53 (39 -44). In previous studies, ␥-irradiation has been used as a specific regulator of p53 expression in cells (36 -38, 61, 62). Our studies showed that, after ␥-radiation, the levels of p53 increased in wild type and normally growing lens epithelial cells, but were weakly increased or unaffected in hyperproliferative ␣BϪ/Ϫ lens epithelial cells (Fig. 9E).
An important function of p53 is to prevent reentry of cells containing damaged DNA into the S phase of the cell cycle. We therefore examined whether hyperproliferative ␣BϪ/Ϫ cells arrested after DNA damage can reenter the S phase. Seventytwo hours after treatment with ␥-radiation, cells were pulsed with 10 M BrdUrd for 30 min. Immunofluorescence of BrdUrd was performed to detect BrdUrd incorporation, and nuclear staining was done with TOTO-1. Fig. 10 shows that, after ␥-irradiation, BrdUrd positive cells were not present in the wild type lens epithelial population, consistent with their abil- Note that a strong immunopositive band for p53 was detected in the hyperproliferative ␣BϪ/Ϫ cells. No p53 band was detected in the p53Ϫ/Ϫ lens epithelial cells. Note also that the p53 protein was weakly or not detected in the wild type controls or the normal growing ␣BϪ/Ϫ cells (lanes 1 and 2). C, analysis of Rb. Lysates were prepared from 10 5 cells and analyzed by SDS-PAGE using 6% gels. The detergent-soluble fractions are shown. The detection of Rb protein was done by immunoblotting. A monoclonal antibody to Rb was used.

␣B-Crystallin Knockout Lens Epithelial Cells
ity to induce cell cycle arrest at the G 1 -S transition after ␥-irradiation. However, many of the hyperproliferative ␣BϪ/Ϫ cells incorporated BrdUrd after ␥-irradiation, indicating that they were able to reenter the cell cycle and initiate DNA synthesis. These results further suggest that hyperproliferative ␣BϪ/Ϫ cells lack normal p53 function.
Next we tested whether hyperproliferative ␣BϪ/Ϫ cells, containing impaired p53 protein, entered M phase after ␥-irradiation (Fig. 11). Morphological examination of TOTO-1 stained cells was used to determine the proportion of cells in mitosis. A nocodazole trapping experiment was done with mouse wild type and hyperproliferative ␣BϪ/Ϫ lens epithelial cells (36). Cells were treated with ␥-irradiation and then with nocodazole, and examined at 12-h intervals for 4 days. Morphological examination revealed that wild type lens epithelial cells remained mitotically inactive for the duration of the experiment. However, hyperproliferative ␣BϪ/Ϫ cells, which accumulate p53, entered mitosis within 24 h after irradiation. This indicates that G 2 arrest initiated in hyperproliferative ␣BϪ/Ϫ cells was not sustained, and suggests that the p53 protein in these cells is functionally impaired (36). The effect of nocodazole trapping after ␥-radiation on hyperproliferative ␣BϪ/Ϫ cells was similar to that on the cell line H1299 lacking functional p53 (data not included).
Differences in wild type and hyperproliferative ␣BϪ/Ϫ cells were also observed when cells were treated with doxorubicin. This drug induces double-strand DNA breaks in cells, resulting in cell cycle arrest and apoptotic cell death. Wild type mouse lens epithelial cells were arrested in G 1 phase 24 h after treatment with doxorubicin. In contrast, hyperproliferative ␣BϪ/Ϫ cells exhibited G 2 cell cycle arrest after treatment with doxorubicin, similar to the H1299 cells lacking functional p53 (data not included). In our studies, we tested two lines of mice that carry the ␣BϪ/Ϫ gene deletion (␣B168Ϫ/Ϫ and ␣B52Ϫ/Ϫ). Our previous study used only the ␣B168Ϫ/Ϫ line of mice (19). The present study with the two independent lines gave identical results, thus extending our previous study and confirming the conclusions of that study. FIG. 9. Cell cycle arrest and p53 expression in wild type and hyperproliferative ␣B؊/؊ lens epithelial cells. A and B, wild type lens epithelial cells; C and D, hyperproliferative ␣BϪ/Ϫ lens epithelial cells. Cells were exposed to ␥-irradiation, cultured in normal growth medium for 24 h, stained with propidium iodide, and analyzed by flow cytometric analysis. The proportion of cells in each cell cycle phase was determined. A, percentage of wild type cells in G 1 phase following exposure to 5-20 Gy of ␥-irradiation. B, relative number of wild type cells in the different phases of cell cycle after 20 Gy of ␥-irradiation. Note that the wild type cells arrested in the G 1 phase after exposure to ␥-irradiation. C, percentage of hyperproliferative ␣BϪ/Ϫ cells in G 2 -M phases following exposure to 5-20 Gy of ␥-irradiation. D, relative number of hyperproliferative ␣BϪ/Ϫ cells in the different phases of cell cycle after 20 Gy of ␥-irradiation. E, -fold change in the expression of p53 protein by 20 Gy of ␥-irradiation of wild type, normally growing ␣BϪ/Ϫ, and hyperproliferative ␣BϪ/Ϫ lens epithelial cells as analyzed by quantitative immunoblotting. Note that the p53 protein was weakly or not detectable in wild type cells before ␥-irradiation, but its expression increased 2-3-fold after irradiation. p53 protein was highly detectable in hyperproliferative ␣BϪ/Ϫ cells before ␥-irradiation, but increased weakly (ϳ1.2-fold) after irradiation.

DISCUSSION
␣B is a small heat shock protein expressed in the lens and numerous other tissues, and its expression increases under certain pathological conditions (9,(13)(14)(15). To understand the function of ␣B protein in lens epithelial cells, we have used primary lens epithelial cells derived from ␣BϪ/Ϫ mice. When we cultured primary lens epithelial cells from ␣BϪ/Ϫ mice, we noted that a small but significant proportion of cells produce hyperproliferative clones, whereas the other cultures appear to have normal growth (19). As compared with wild type cells, the ␣BϪ/Ϫ cells (both normally growing and hyperproliferative) had a significant proportion of cells that were tetraploid and aneuploid, indicative of chromosome instability (19). Furthermore, the ␣BϪ/Ϫ cells died by apoptosis when infected with the hybrid virus Ad12-SV40. Our current work focused on understanding the mechanism involved in the hyperproliferation of ␣BϪ/Ϫ cells. We investigated the status of two key cell cycle regulators, p53 and Rb, in the hyperproliferative ␣BϪ/Ϫ lens epithelial cells, and compared them with wild type cells. Our results demonstrate that there were no significant differences in Rb phosphorylation between the wild type, normally growing ␣BϪ/Ϫ and hyperproliferative ␣BϪ/Ϫ cells, but the levels of p53, undetectable in wild type or normally growing ␣BϪ/Ϫ cells, accumulated to detectable levels in the hyperproliferative ␣BϪ/Ϫ cells. These results suggest an association between hyperproliferation and expression of p53.
The increase in immunostaining for p53 in the hyperproliferative ␣BϪ/Ϫ cells suggested that the protein was stabilized by mutations (31)(32)(33)(34)59). We therefore determined whether the p53 gene had mutated in these cells. However, by sequencing the p53 cDNA in wild type and hyperproliferative ␣BϪ/Ϫ lens epithelial cells using RT-PCR-based methodologies, we did not detect any mutations in the coding region of p53, in two hyperproliferative ␣BϪ/Ϫ cell lines independently derived from lens epithelial cells of ␣B52Ϫ/Ϫ and ␣B168Ϫ/Ϫ mice.
We next determined whether the p53 was functional in hyperproliferative ␣BϪ/Ϫ lens epithelial cells. Our data suggest that wild type lens epithelial cells express normal p53, whereas hyperproliferative ␣BϪ/Ϫ cells had lost endogenous p53 function and accumulated p53 that is functionally impaired. p53 is kept in a repressed state in normal cells, but is activated by post-translational modification in response to genotoxic stress such as ionizing radiation (31)(32)(33)(34). p53 controls and coordi-nates the anti-proliferative responses to prevent DNA replication from occurring when cells are exposed to genotoxic stress. It acts as a multifunctional transcription factor to enhance or repress the expression of genes involved in cell proliferation, cell cycle arrest, and apoptosis. When wild type lens epithelial cells were exposed to ionizing radiation, p53 suppressed cell cycle progression in the G 1 phase. Upon ␥-irradiation, wild type lens epithelial cells underwent cell cycle arrest in G 1 , but like other cells containing mutant and functionally impaired p53 protein, hyperproliferative ␣BϪ/Ϫ lens epithelial cells arrested in the G 2 phase after treatment with ␥-irradiation, suggesting that p53 function is impaired in hyperproliferative ␣BϪ/Ϫ cells. The G 1 arrest after ␥-irradiation results in part from the p53-regulated synthesis of the cell cycle inhibitor p21, which leads to inhibition of cyclin-Cdk complexes required for the transition from G 1 to S phase (36,(41)(42)(43). However, arrest in G 2 after DNA damage occurs, in both mouse and human cells, in the absence of p53 or p21 (36,(41)(42)(43)(44). Arrest at DNA damage checkpoints G 1 -S or G 2 -M of the cell cycle prevents DNA replication and mitosis in the presence of unrepaired chromosomal alterations. Our data show that wild type cells did not enter S phase or mitosis after DNA damage by ␥-irradiation, whereas hyperproliferative ␣BϪ/Ϫ cells, presumably lacking functional p53, underwent DNA synthesis (BrdUrd labeling) and mitosis after ␥-irradiation, typical of cells with impaired p53 function. An additional assay for p53 function was the effect of doxorubicin, an anticancer drug that induces double-strand DNA breaks and causes cell cycle arrest, which gave the same result as ␥-irradiation.
The nature of the signaling event that triggers p53 accumulation in hyperproliferative ␣BϪ/Ϫ cells is not known. p53 accumulates in cells in which DNA replication is blocked by drugs such as hydroxyurea or aphidicolin (33), and by DNA damage (46) or adenoviral transfection (47). DNA strand breaks are common when replication is blocked, and are likely to be the main trigger activating p53 (33). It has also been demonstrated that the p53 protein, which accumulates when DNA replication is blocked, is impaired in its transcriptional activity (33). We hypothesize that impairment of p53 transcriptional activity may be one explanation for the inability of hyperproliferative ␣BϪ/Ϫ cells to arrest cells in G 1 phase after ␥-irradiation. Improper chromosome segregation and arrest in anaphase may also trigger p53. Further studies are necessary to determine whether the loss of ␣B can stabilize p53, and yet actively block its transcriptional activating program.
Further assessment of the genomic integrity of the ␣BϪ/Ϫ cells was obtained from the effect of SV40 T antigen on cell death. Normally a growth-promoting protein, the expression of SV40 T antigen in ␣BϪ/Ϫ cells had an unusual effect, massive cell death after attaining confluence. Wild type lens epithelial cells did not demonstrate the cell extinction observed in ␣BϪ/Ϫ cells by Ad12-SV40 infection. These observations underscore that the large T antigen does not always promote proliferation, but depending on cell type and conditions, may induce a cell death response. A similar cell death response has been reported for the small t antigen in U2OS and H1299 cells (63), and a large T antigen-mediated sensitization of rat embryo fibroblasts has also been reported recently (64). In the current work, normally growing T-antigen positive ␣BϪ/Ϫ cells died soon after reaching confluence, whereas hyperproliferative ␣BϪ/Ϫ cells died in a more protracted manner. The apoptosis in ␣BϪ/Ϫ cells after large T antigen expression may occur through a p53-mediated pathway, because hyperproliferative ␣BϪ/Ϫ cells that have impaired p53 function are resistant to apoptosis.
Our studies indicate that, in some of the hyperproliferative FIG. 11. Mitotic entry after ␥-irradiation in mouse wild type and hyperproliferative ␣B؊/؊ lens epithelial cells. Wild type (E-E) or hyperproliferative ␣BϪ/Ϫ (X-X) cells were tested for their response to DNA damage. After treatment with ␥-radiation, cells were treated with nocodazole to trap the cells in mitosis. DNA was stained with TOTO-1 at different time points after irradiation. Morphological examination revealed mitotic cells in the ␣BϪ/Ϫ cell population within 24 h after ␥-irradiation, whereas wild type cells remained mitotically inactive for the duration of the experiment. Thus, although cell cycle arrest was initiated in both cell types, this arrest was overcome in only the hyperproliferative ␣BϪ/Ϫ cells.
␣BϪ/Ϫ lens epithelial cells, the mitotic spindle was not properly assembled. Wild type cells showed normal mitotic profiles from prometaphase to telophase, but some of the cells in hyperproliferative ␣BϪ/Ϫ cultures showed various abnormalities such as failure to round up, cell cycle arrest during anaphase, and binucleated cells. These aberrant mitotic profiles were not visualized in the wild type cells. Because the mitotic spindle performs a number of well documented functions, including the capture and segregation of the duplicated parental chromosomes to daughter cells (65), and molecular chaperones such as HSP70 and HSP90 are associated with the mitotic spindle, assisting in tubulin folding (27,66,67), these results suggest a role of ␣B in the spindle assembly. Cells defective in assembly of the mitotic spindle undergo cell cycle arrest or cell death (68,69). It is also interesting to note that antibodies against a phosphorylated form of ␣B (at serine 59) label the centrosomes and midbodies of dividing cells (7), suggesting that, like other chaperones, ␣B may also play a role in the assembly of microtubules. ␣B has also been suggested to be a chaperone for tubulin (70). The relationship between these aberrant mitotic profiles, impaired p53 function, and the genomic instability of epithelial cells derived from ␣BϪ/Ϫ lenses needs further investigation.
Although the hyperproliferation of ␣BϪ/Ϫ lens epithelial cells was observed in culture, loss of ␣B did not produce a phenotype in vivo, as shown by the lack of a significant difference in size and morphology of lenses. Some studies suggest that this lack of hyperproliferation of ␣BϪ/Ϫ cells is a result of the negative effect exerted by lens fibers on mitotic activity of lens epithelial cells (71). It remains to be determined whether the expression of p53 is enhanced in the ␣BϪ/Ϫ mouse lens epithelium in vivo.
Preservation of genomic integrity is important for cell survival, and p53 was the first protein demonstrated to be involved in this function (72). Genomic maintenance functions include preservation of primary DNA sequence, cell cycle checkpoint control, and accurate chromosome segregation during cell division, a function essential for preservation of chromosome ploidy and structure (73,74). In normal cells, chromosome rearrangements are actively suppressed by a series of extensive and redundant pathways (75). The stress protein ␣B may be necessary for chromosome stability. We propose that the genomic instability of ␣BϪ/Ϫ cells is likely to lead to p53dependent cell death, and that the accumulation of functionally inactive p53 is a survival mechanism, conferring an immortal phenotype.
We can also propose that ␣B, like its aggregation partner in the lens ␣A, is important to ensure proper completion of mitosis. The absence of ␣A increases cell death in the in vivo mouse lens epithelium (53) and decreases proliferation in vitro (20). The hyperproliferation, abnormal mitotic profiles, and ploidy changes of ␣BϪ/Ϫ lens epithelial cells demonstrated here suggest that both ␣A and ␣B, like other chaperones, may be important for the proper assembly of the mitotic spindle. Further studies are needed to investigate this possibility.