Cdk2 and Cdk4 activities are dispensable for tumorigenesis caused by the loss of p53

Abstract

The loss of p53 induces spontaneous tumors in mice, and p53 mutations are found in approximately 50% of human tumors. These tumors are generally caused by a number of events, including genomic instability, checkpoint defects, mitotic defects, deregulation of transcriptional targets, impaired apoptosis, and G(1) deregulation or a combination of these effects. In order to determine the role of proteins involved in G(1) control in tumorigenesis, we focused on Cdk2 and Cdk4, two cyclin-dependent kinases that in association with cyclin E and cyclin D promote the G(1)/S phase transition. We analyzed the consequence of loss of Cdk2 in p53-null animals by generating Cdk2(-/-) p53(-/-) mice. These mice are viable and developed spontaneous tumors, predominantly lymphoblastic lymphomas, similar to p53(-/-) mice. In contrast, the genotypes Cdk4(-/-) p53(-/-) were mostly lethal, with few exceptions, and Cdk2(-/-) Cdk4(-/-) p53(-/-) mice die during embryogenesis at embryonic day 13.5. To study the oncogenic potential, we generated mouse embryonic fibroblasts (MEFs) and found that p53(-/-), Cdk2(-/-) p53(-/-), Cdk4(-/-) p53(-/-), and Cdk2(-/-) Cdk4(-/-) p53(-/-) MEFs grew at similar rates without entering senescence. Ras-transformed MEFs of these genotypes were able to form colonies in vitro and induce tumors in nude mice. Our results suggest that tumorigenicity mediated by p53 loss does not require either Cdk2 or Cdk4, which necessitates considering the use of broad-spectrum cell cycle inhibitors as a means of effective anti-Cdk cancer therapy.

FIG. 1.

Cdk2^−/− p53^−/− and Cdk2^−/− mice are sterile. (A) Representative photograph comparing testes from WT, Cdk2^−/− p53^−/−, Cdk2^−/−, and p53^−/− mice. The pictures indicate that testes of both Cdk2^−/− and Cdk2^−/− p53^−/− mice are smaller than WT testis. (B to D) Histology of testis sections indicates that the seminiferous tubules of Cdk2^−/− p53^−/− mice are atrophic like the ones from Cdk2^−/− mice. Bars, 40 μm.

FIG. 2.

Loss of p53 improves growth of Cdk2^−/− MEFs. (A) Analysis of spontaneous immortalization of MEFs using a 3T3 assay. The five genotypes plotted are Cdk2^−/−, Cdk2^−/− p53^−/−, p53^−/−, p53^+/−, and Cdk2^−/− p53^+/−. The number of passages is plotted on the x axis, and cumulative growth is plotted on the y axis. (B) Proliferation assay performed on passage 6 for a period of 7 days for four genotypes: WT, Cdk2^−/− p53^−/−, Cdk2^−/−, and p53^−/−. Cdk2^−/− MEFs display a decreased growth rate, which is rescued by loss of p53. (C) Western blot (WB) analysis of protein extracts from MEFs using antibodies against Cdk2, p53, Cdk4, Emi1, Cdk1^pY15, and actin.

FIG. 3.

Cdk2^−/− p53^−/− mice develop spontaneous tumors. (A) Survival curve generated after following 25 p53^−/− mice and 30 Cdk2^−/− p53^−/− mice for a period of 275 days. (B and C) Representative thymic lymphoma found in one of the Cdk2^−/− p53^−/− mice. (D) Thymic lymphoma found in Cdk2^−/− p53^−/− mice compared with the normal thymus of a WT animal of the same age. (E) Comparison of the spleens of a WT animal and Cdk2^−/− p53^−/− mice. (F to I) Histological sections of tumors found in Cdk2^−/− p53^−/− mice (Table 2). Sections have been stained with hematoxylin and eosin. Magnifications: panels F, G, and I, ×40; panel H, ×10. Bars: 100 μm (H) and 300 μm (F, G, and I).

FIG. 4.

Ras transforms MEFs lacking p53^−/−. (A to D) p53^−/−, Cdk2^−/− p53^−/−, Cdk4^−/− p53^−/−, and Cdk2^−/− Cdk4^−/− p53^−/− MEFs were infected with empty vector (left) or activated Ras^G12V (right). Colonies were stained with Giemsa after 10 days of culture. The numbers of colonies are the averages from three 10-cm plates.

FIG. 5.

Loss of p53 improves growth of Cdk2^−/− Cdk4^−/− cells. (A) 3T3 assay performed for 16 passages using WT, Cdk2^−/− Cdk4^−/−, Cdk2^−/− Cdk4^−/− p53^−/−, Cdk2^−/− Cdk4^−/− p53^+/−, and p53^−/− MEFs. The x axis shows the number of passages, and the y axis indicates the cumulative cell number for each passage. (B) Western blot (WB) analysis. One hundred micrograms of protein lysate from passage 3 MEFs of WT (lane 1), Cdk2^−/− Cdk4^−/− (lane 2), Cdk2^−/− Cdk4^−/− p53^−/− (lane 3), and p53^−/− (lane 4) genotypes was separated on a 12.5% gel, and proteins were transferred to a polyvinylidene difluoride membrane. Antibodies against Hsp90, Cdk2, Cdk4, p53, p21, and p27 were used for detection. (C to F) β-Galactosidase staining. Shown are representative images from WT (C), Cdk2^−/− Cdk4^−/− (D), p53^−/− (E), and Cdk2^−/− Cdk4^−/− p53^−/− (F) MEFs stained with β-galactosidase as a marker for senescence. (G) Graph depicting the percentage of proliferating cells and senescent cells at passage 3 in WT, Cdk2^−/− Cdk4^−/−, Cdk2^−/− Cdk4^−/− p53^−/−, and p53^−/− MEFs, counting at least 200 cells for each genotype.

FIG. 6.

Cdk2^−/− Cdk4^−/− p53^−/− MEFs display a defect in S phase entry. (A) Histogram of FACS analysis of cells in S phase. MEFs were collected after release from serum starvation at different time points (x axis) after 1 h of pulse-labeling with BrdU and were stained with propidium iodide followed by FACS analysis. The y axis represents the percentage of cells in S phase in the different MEF genotypes: WT, Cdk2^−/− Cdk4^−/−, Cdk2^−/− Cdk4^−/− p53^−/−, and p53^−/−. (B) Kinase assay. Two hundred fifty micrograms of protein lysate prepared from Cdk2^−/− Cdk4^−/− p53^−/− and p53^−/− MEFs 0, 12, 16, 20, and 24 h after serum stimulation was used. Cyclin A2 (cycA2), cyclin B1 (cycB1), and Cdk1 antibodies coupled to protein A-agarose beads or suc1 beads were used for immunoprecipitation (IP) followed by an in vitro kinase assay using the substrate histone H1.

FIG. 7.

Cdk1 binds to all cyclins. (A) Western blot analysis. One hundred micrograms of protein extract from WT, Cdk2^−/− Cdk4^−/−, Cdk2^−/− Cdk4^−/− p53^−/−, and p53^−/− MEFs was separated on a 12.5% Tris-HCl gel, and proteins were transferred to a polyvinylidene difluoride membrane. Antibodies against Cdk1, cyclin A2 (cycA2), and cyclin B1 (cycB1) were used to probe the membranes. (B) Kinase assays. Two hundred fifty micrograms of protein lysate from WT, Cdk2^−/− Cdk4^−/−, Cdk2^−/− Cdk4^−/− p53^−/−, and p53^−/− MEFs was used for immunoprecipitation (IP) using protein A-agarose beads coupled to Cdk2, Cdk1, cyclin B1, and cyclin A2 antibodies or suc1 beads. The immunoprecipitated complexes were used to perform an in vitro kinase assay using histone H1 as a substrate and radiolabeled ATP. (C) Immunoprecipitation followed by Western blot (WB) analysis. Protein A-agarose beads coupled to cyclin A2, cyclin D1, cyclin B1, and cyclin E antibodies were used to immunoprecipitate 250 μg protein lysate. The immunoprecipitated proteins were probed with anti- bodies against Cdk2, Cdk4, Cdk1, and p21. ND, not determined. (D) Cycloheximide treatment. Cdk2^−/−Cdk4^−/− and Cdk2^−/−Cdk4^−/−p53^−/− MEFs at passage 4 were treated with cycloheximide for 3 or 6 h. Western blot analysis was performed on these cell extracts using antibodies specific for HSP90, Cdk1, and Cdk2.

FIG. 8.

Loss of p21 does not immortalize Cdk2^−/− Cdk4^−/− MEFs. (A) 3T3 assay of WT, Cdk2^−/− Cdk4^−/−, Cdk2^−/− Cdk4^−/− p21^−/−, and p21^−/− MEFs. The y axis depicts the cumulative cell number for each passage. (B) Western blot (WB) analysis. One hundred micrograms of protein lysate from passage 3 WT (lanes 1), Cdk2^−/− Cdk4^−/− (lanes 2), Cdk2^−/− Cdk4^−/− p21^−/− (lanes 3), and p21^−/− (lanes 4) MEFs were separated on a 12.5% Tris-HCl gel. Proteins were transferred to a polyvinylidene difluoride membrane and probed with antibodies against Cdk2, Cdk4, p53, and p21. (C to H) Representative pictures depicting senescent cells in WT, Cdk2^−/− Cdk4^−/−, Cdk2^−/− Cdk4^−/− p21^−/−, p21^−/−, Cdk2^−/− Cdk4^−/− p21^−/− (expressing Ras), and p21^−/− (expressing Ras) MEFs after β-galactosidase and neutral red staining. (I) Histogram generated after random counting of at least 200 cells for each genotype. The senescent cells are indicated with a blue bar, and nonsenescent cells are indicated with pink bars.