Restoring expression of wild-type p53 suppresses tumor growth but does not cause tumor regression in mice with a p53 missense mutation

The transcription factor p53 is a tumor suppressor. As such, the P53 gene is frequently altered in human cancers. However, over 80% of the P53 mutations found in human cancers are missense mutations that lead to expression of mutant proteins that not only lack p53 transcriptional activity but exhibit new functions as well. Recent studies show that restoration of p53 expression leads to tumor regression in mice carrying p53 deletions. However, the therapeutic efficacy of restoring p53 expression in tumors containing p53 missense mutations has not been evaluated. Here we demonstrate that restoring wild-type p53 expression halted tumor growth in mice inheriting a p53^R172H missense mutation that is equivalent to a P53 missense mutation detected in approximately 6% of human cancers. However, it did not lead to tumor regression, as was observed in mice lacking p53. We further showed that the dominant-negative effect of the mutant p53 encoded by p53^R172H dampened the activity of the restored wild-type p53. We therefore conclude that in a mutant p53 background, p53 restoration has the therapeutic potential to suppress tumor progression. Our findings support using p53 restoration as a strategy to treat human cancers with P53 missense mutations and provide direction for optimizing p53 restoration in cancer therapy.

The tumor suppressor p53 is a transcription factor. Upon activation by signals, such as DNA damage, oncogenic stimuli, and hypoxia, wild-type p53 activates the transcription of genes involved in apoptosis, cell cycle arrest, differentiation, and senescence (1, 2). These potent antitumor activities prevent cells with aberrant growth signals from proliferating. Approximately, half of human cancers have P53 gene alterations that result in loss of p53 activity. While a few of these alterations are P53-null mutations, over 80% are P53 missense mutations that lead to expression of mutant p53 proteins (3, 4). Many p53 missense mutants lack p53 transcriptional activity and show gain-of-function activities.

In particular, the arginine-to-histidine mutation at codon 175 of the P53 gene (corresponding to p53^R172H in mice) occurs in about 6% of human cancers (5). The p53^R172H mutation has gain-of-function properties, manifested as a tumor metastasis phenotype in p53^R172H heterozygous mice that is lacking in p53^+/– mice (6, 7). Another property of the p53R172H mutant is its dominant-negative effect that silences wild-type p53 under some circumstances (8). Thus, for example, in response to γ-irradiation, mutant p53R172H inactivates wild-type p53 activities (9). Additionally, mutant p53 binds and suppresses the activities of the related proteins, p63 and p73 (7). However, the p53^R172H heterozygous mice that express equal amounts of wild-type and mutant p53 have survival curves identical to those of p53^+/– mice, indicating that mutant p53 does not inactivate wild-type p53 (8). One possible explanation for this discrepancy is that once stabilized, mutant p53 has a longer half-life than wild-type p53 and thus overwhelms wild-type p53 function (10).

Since p53 inactivation is common in human cancers, p53 restoration offers a very promising therapy for cancer treatment. The replenished wild-type p53, if properly activated, confers cell cycle arrest and cell death, both of which represent effective mechanisms of tumor suppression (11). Recent mouse tumor studies have provided direct evidence to support the rationale for using p53 restoration in cancer therapy. Three research groups independently generated genetically modified mice that allowed restoration of wild-type p53 expression in tumors lacking p53 (12–14). The lymphomas, sarcomas, and carcinomas that developed in these mice regressed after p53 restoration by initiating apoptosis or senescence in a tumor-specific manner. While exciting, these studies do not address the effects of p53 restoration in tumors with mutant p53, as would be the case in many human tumors. The activity of the restored wild-type p53 might be compromised by the stabilized mutant p53 in tumors. We hypothesized that in a mutant p53 background, the restored wild-type p53 may be less effective as a tumor suppressor.

The purpose of this study was to evaluate the therapeutic efficacy of p53 restoration in spontaneous mouse tumors with mutant p53. We generated a mouse model that allowed wild-type p53 restoration in vivo and demonstrated that p53 restoration led to tumor regression in mice lacking p53, as previously published (13). More importantly, p53 restoration halted tumor growth in mice containing the p53^R172H mutation, although it did not lead to tumor regression. We further showed that the dominant-negative effect of the p53R172H mutant dampened the activity of the restored wild-type p53. Our results support the use of p53 restoration in the treatment of human cancers with p53 missense mutations.

Generation of a latent p53 allele. We generated a latent p53 allele using gene targeting. This allele, designated the p53neo allele, contained a floxed PGKneo cassette that was inserted into intron 4 of the p53 locus (Figure 1A). p53 was confirmed to be wild type by sequencing (data not shown). Southern blot analysis was used to verify correct targeting and germline transmission of the p53^neo allele in mice (Figure 1B). The PGKneo cassette was removed from the p53^neo allele by Cre recombination, and the resultant allele was designated p53^Δneo (Figure 1A). The p53 expression from the p53^neo and p53^Δneo alleles was examined in mouse embryonic fibroblasts (MEFs) (Figure 1C). After doxorubicin treatment, the levels of p53 expression in p53^neo/– and p53^neo/neo MEFs were 7.4% and 16.1%, respectively, of the level of p53 expression in p53^+/+ MEFs (Figure 1C), while p53^+/– MEFs were 50.5% of the level of p53 expression in p53^+/+ MEFs (Supplemental Figure 1; supplemental material available online with this article; doi: 10.1172/JCI44504DS1). p53^neo/– and p53^neo/neo MEFs activated p21, a direct target of p53, at 33.7% and 68.9% that of p53^+/+ MEFs, respectively. Equivalent p53 and p21 protein levels were observed in p53^Δneo/Δneo and p53^+/+ MEFs. Thus, the PGKneo cassette downmodulated endogenous p53 expression from the p53^neo allele, and deletion of PGKneo allowed restoration of normal p53 levels.

Figure 1

Generation of a latent p53^neo allele. (A) Targeting strategy to generate p53^neo and p53^Δneo alleles. The p53^neo allele, with insertion of the PGKneo gene flanked by loxP sites (triangles) in intron 4 (introns are represented by lines and exons by rectangles), was generated by homologous recombination. The p53^Δneo allele was generated by crossing p53^neo/+ mice with ZP3-Cre transgenic mice. (B) Southern blot analysis of tail DNA samples of p53 wild-type (+/+), p53^neo/+ (neo/+), and p53^neo/neo (neo/neo) mice after digestion with EcoRI and SalI. (C) Western blot analysis of p53 and p21 expression in doxorubicin-treated MEFs with p53^–/– (–/–), p53^neo/– (neo/–), p53^neo/neo, p53^Δneo/Δneo (Δ/Δ), and p53 wild type genotypes. Levels of p53 and p21 were normalized to actin. The numbers under lanes indicate the p53 and p21 protein levels (percentages) relative to those for wild-type mice. (D) Survival curves of p53^–/– (n = 35), p53^R172H/– (H/–, n = 34), p53^neo/R172H (neo/H, n = 40), p53^neo/– (n = 37), p53^neo/neo (n = 30), p53^+/– (+/–, n = 40), and p53^Δneo/Δneo (n = 38) mice.

Transformed mouse 3T3 cell double minute 2 (Mdm2) is a potent p53 inhibitor, and the embryonic lethality of Mdm2-null mice is rescued by concomitant deletion of p53 (15, 16). Thus, the Mdm2-null phenotype offers a unique opportunity to examine p53 activity in vivo. To examine whether the p53^neo allele was functionally null, we crossed p53^–/–Mdm2^–/– mice with p53^neo/neoMdm2^+/– mice. Although p53^neo/–Mdm2^–/– and p53^neo/–Mdm2^+/– mice were born at the correct ratios, the p53^neo/–Mdm2^–/– mice were smaller and less mobile than their p53^neo/–Mdm2^+/– siblings and died by weaning age (Supplemental Figure 2 and Supplemental Table 1). The p53^neo/neo genotype did not rescue the embryonic lethality of Mdm2^–/– mice (data not shown). Taken together, these data argue that the p53^neo allele retained some wild-type activity and was thus hypomorphic. In crosses with p53^R172H/R172H mice, the p53^neo/R172HMdm2^–/– mice were completely rescued (Supplemental Figure 2 and Supplemental Table 1), indicating that the p53^R172H allele masked the wild-type activity of the p53^neo allele in vivo.

Analysis of survival and tumorigenesis. To monitor survival and tumor development, we generated cohorts of p53 mutant mice in the C57BL/6 background. Compared with p53^–/– and p53^+/– mice with median survival time of 160 and 578 days, respectively, p53^neo/– and p53^neo/neo mice had median survival times of 251 and 395 days, respectively (Figure 1D). This gene-dosage effect on survival confirmed that the p53^neo allele was hypomorphic. In contrast, the survival curve of p53^neo/R172H mice was identical to those of p53^–/– and p53^R172H/– mice (Figure 1D). Therefore, in p53^neo/R172H mice, the presence of p53^R172H eliminated the potential increase in survival allowed by the p53^neo allele. The p53^Δneo/Δneo mice survived normally, with no tumor development by 25 months, similar to wild-type mice. Most p53^neo/–, p53^neo/neo, and p53^neo/R172H mice died of tumor development. The most common tumor types in these mice were lymphomas and sarcomas (Table 1). p53^neo/R172H mice developed 59.6% lymphomas and 36.5% sarcomas, which was similar to the 56% and 40% observed frequency, respectively, in p53^–/– mice (7). In contrast, the p53^neo/– and p53^neo/neo mice showed an increased frequency of sarcomas (61.4% and 61.3%, respectively) with a decreased frequency of lymphoma development (34.1% and 32.2%, respectively), as compared with that in p53^–/– mice. Thus, expression of a small amount of wild-type p53 from the p53^neo allele led to a shift in tumor spectrum with an increase in the number of sarcomas, except in combination with the p53^R172H allele.

Table 1

Tumor spectra of p53^neo/R172H, p53^neo/–, and p53^neo/neo mice

p53 restoration in vivo. To temporally restore the p53 gene in vivo, we generated p53^neo/–Cre-ER^TM and p53^neo/R172HCre-ER^TM mice. The Cre-ER^TM transgene constitutively expresses a Cre-ER fusion protein that becomes active only after tamoxifen exposure (17). Six-week-old p53^neo/–Cre-ER^TM mice were treated with tamoxifen by weekly intraperitoneal injections. Southern blot analysis confirmed the conversion of the p53^neo allele to p53^Δneo in various tissues of mice treated for 1 or 3 weeks (Supplemental Figure 3). To determine how efficient p53 restoration was in activation of p53, we performed immunohistochemistry experiments with p53 and cleaved caspase-3 antibodies after Cre-mediated recombination, followed by whole-body γ-irradiation (Figure 2). Increased p53 and cleaved caspase-3 staining were observed in the spleens and thymi of irradiated, tamoxifen-treated p53^neo/–Cre-ER^TM mice compared with that in mice of the same genotype without tamoxifen treatment. As a control, irradiated p53^+/– mice showed extensive p53 activity (Figure 2 and Supplemental Figure 4). These data demonstrated that tamoxifen treatment worked effectively to trigger p53 restoration in vivo.

Figure 2

In vivo p53 restoration in a mouse model. (A) Immunohistochemistry of p53 and cleaved caspase-3 (CC3) in the spleens of vehicle- or tamoxifen-treated (Tam-treated) p53^neo/–Cre-ER^TM mice (0×, vehicle treatment; 1×, tamoxifen treatment once; 3×, tamoxifen treatment 3 times) exposed to 6 Gy γ-radiation (IR). Tissues were collected 5 hours after γ-radiation. The irradiated p53^+/– mice were used as a positive control, and tamoxifen-treated p53^neo/–Cre-ER^TM mice without γ-radiation (no IR) were used as a negative control. Scale bar: 100 μm. (B) Percentages of cells staining positive for p53 and cleaved caspase-3 in A. Positive cells were counted in 4 random fields (original magnification, ×40) and presented as the mean ± SEM.

Tumor suppression after p53 restoration. To characterize tumor response after p53 restoration, we established cohorts of p53^neo/R172H and p53^neo/– mice with and without Cre-ER^TM. These mice underwent MRI periodically to screen for spontaneous tumor development. Upon detection of a tumor, mice were treated 4 times weekly with tamoxifen by intraperitoneal injection. If a mouse survived for 4 weeks, it was then imaged again. Otherwise, a second image was taken before the mouse was euthanized. Tumor volumes were calculated from a sequence of MRI images, and the change in tumor volume was calculated as a percentage of the original tumor volume at the time of detection. In the p53^neo/–Cre-ER^TM mice, tamoxifen treatment caused tumor regression, as previously reported for other p53 restoration models (12–14). All 8 tamoxifen-treated p53^neo/–Cre-ER^TM mice (4 lymphomas and 4 sarcomas) showed tumor regression, ranging from 18.3% to 98.9% (Figure 3, A and D, and Supplemental Table 2). We confirmed that this effect was caused by p53 restoration, not tamoxifen itself, as control p53^neo/– mice (3 lymphomas and 5 sarcomas) showed rapid progression within 3 weeks despite tamoxifen treatment (Figure 3D and Supplemental Table 2). Tumor volumes increased from as little as 67.3% to as much as 387.9%.

Figure 3

p53 restoration led to tumor regression in p53^neo/–Cre-ER^TM mice and suppressed tumor progression in p53^neo/R172HCre-ER^TM mice. (A) Tumor MRI images and tumor volume calculations in representative tamoxifen-treated p53^neo/–Cre-ER^TM mice (1 sarcoma and 1 lymphoma). (B) Tumor MRI images and tumor volume calculations in representative tamoxifen-treated p53^neo/R172HCre-ER^TM mice (2 lymphomas and 1 sarcoma). (C) Tumor MRI images and tumor volume calculations in representative tamoxifen-treated p53^neo/R172 mice (1 lymphoma and 1 sarcoma). (A–C) Tumor volume measurements were obtained in cm³ from T2-weighted MRI image stacks. For each time point, measurements (n = 4) from image sequences were used to calculate a mean volume ± SEM. Day 0 (0d) indicates the day before tamoxifen treatment; day 8, 14, 18, and 28 indicate 8, 14, 18, and 28 days after tamoxifen treatment, respectively. Green dashed lines outline representative tumor images (coronal view, no magnification). (D) Summary of changes in tumor volumes after tamoxifen treatment in p53^neo/R172HCre-ER^TM (neo/HER), p53^neo/R172H, p53^neo/–Cre-ER^TM (neo/–ER), and p53^neo/– mice. L1–17, lymphomas numbered 1–17; S1–15, sarcomas numbered 1–15 (see Supplemental Table 1). (E) Survival of p53^neo/R172HCre-ER^TM (n = 8), p53^neo/R172H (n = 8), p53^neo/–Cre-ER^TM (n = 8), and p53^neo/– (n = 8) mice after tamoxifen treatment, regardless of tumor types.

In p53^neo/R172HCre-ER^TM mice that express mutant p53, p53 restoration did not cause tumor regression but halted tumor growth. All 8 p53^neo/R172HCre-ER^TM tumors (5 lymphomas and 3 sarcomas) showed little or no growth (<45%) after weekly injections of tamoxifen (Figure 3, B and D, and Supplemental Table 2). In contrast, the p53^neo/R172H control tumors (5 lymphomas and 3 sarcomas) showed rapid growth, with tumor volumes ranging from 145.3% to 1,036.4% within 2 weeks (Figure 3, C and D, and Supplemental Table 2). Moreover, after tamoxifen treatment, 62.5% of p53^neo/R172HCre-ER^TM mice and 75.0% of p53^neo/–Cre-ER^TM mice survived to the fourth week, whereas all p53^neo/R172H and p53^neo/– control mice died within 3 weeks (Figure 3E). This effect on survival occurred regardless of tumor size at the onset of treatment.

To exclude the possibility that the phenotypic differences observed in p53^neo/R172HCre-ER^TM and p53^neo/–Cre-ER^TM mice might simply be explained as differences in efficiency of p53^neo recombination, we examined the levels of the remaining p53^neo allele in tumors after p53^neo recombination, using a genomic DNA real-time PCR strategy. Since these tumors contained 1 p53^neo allele originally, the measurement of the remaining level of the p53^neo allele after tamoxifen treatment will denote the efficiency of the p53^neo recombination in these tumors. The average levels of p53^neo remaining in the p53^neo/–Cre-ER^TM and p53^neo/R172HCre-ER^TM lymphomas after tamoxifen treatment were similar (52.1% and 53.5%, respectively; P = 0.87), suggesting no significant difference in p53^neo recombination in these tumors (Supplemental Figure 5). Another concern is that the presence of the Cre-ER^TM transgene may promote tumorigenesis in mice. We therefore compared the survival of the p53^neo/R172H and the p53^neo/R172HCre-ER^TM mice without treatment. The overlapping survival curves between these 2 cohorts suggest that the Cre-ER^TM transgene did not contribute to tumorigenesis in the p53^neo/R172HCre-ER^TM mice (Supplemental Figure 6).

Apoptosis and senescence contribute to tumor suppression upon p53 restoration. To determine the mechanism(s) responsible for tumor suppression, we compared tumor response with and without p53 restoration. After 4 weekly tamoxifen injections, apoptosis was elevated in both p53^neo/–Cre-ER^TM (3 samples) and p53^neo/R172HCre-ER^TM lymphomas (5 samples) compared with that in lymphomas without Cre-ER^TM (8 samples) (Figure 4, A and B). The p53^neo/R172HCre-ER^TM tumors appeared to have an increased apoptosis response to tamoxifen as compared with p53^neo/–Cre-ER^TM tumors. However, p53^neo/– mice with lymphomas lived significantly longer than p53^neo/R172H mice with lymphomas (Supplemental Figure 7). Remarkably, in addition to apoptosis, 3 out of the 5 p53^neo/R172HCre-ER^TM lymphomas, but none of the p53^neo/–Cre-ER^TM lymphomas, showed induction of senescence, as detected by senescence-associated β-galactosidase (SA-β-gal) activity (Figure 4A). This phenotype was also observed in lymphomas that were collected 3 days after a single dose of tamoxifen. Four out of these six p53^neo/R172HCre-ER^TM lymphomas showed induction of senescence, as detected by SA-β-gal assays and immunohistochemistry with senescence markers PML and DcR2 (Supplemental Figure 8). Therefore, the restored p53 activated both apoptosis and senescence programs in lymphomas from p53^neo/R172HCre-ER^TM mice. Additionally, immunohistochemistry revealed lower levels of proliferation markers Ki-67 and phospho-histone 3 in Cre-ER^TM–positive lymphomas than in Cre-ER^TM–negative controls (Figure 4, A and B), indicating that p53-induced apoptosis and/or senescence decreased tumor cell proliferation.

Figure 4

p53 restoration resulted in elevated apoptosis and induction of senescence in lymphomas and decreased tumor cell proliferation. Tamoxifen-treated lymphomas collected in the MRI study were subjected to H&E staining; immunohistochemical staining for cleaved caspase-3, Ki-67, and phospho-histone 3 (PH3); and SA-β-gal assays. (A) Images of analyses of the representative lymphomas from the tamoxifen-treated p53^neo/–, p53^neo/–Cre-ER^TM, p53^neo/R172H, and p53^neo/R172HCre-ER^TM mice. Scale bar: 100 μm. (B) Percentages of cells staining positive for cleaved caspase-3 and phospho-histone 3 in A. Positive cells were counted in 4 random fields (original magnification, ×40) and presented as the mean ± SEM.

In p53^neo/–Cre-ER^TM (3 samples) and p53^neo/R172HCre-ER^TM (3 samples) angiosarcomas, the major tumor response after tamoxifen treatment was senescence, as detected by SA-β-gal staining (Figure 5A). However, in other sarcomas, such as giant-cell sarcoma and spindle-cell sarcoma, both apoptosis and senescence were detected (Supplemental Figure 9 and data not shown). The tamoxifen-treated p53^neo/–Cre-ER^TM angiosarcomas also showed less tumor cell proliferation than the Cre-ER^TM–negative control tumors (7 samples) (Figure 5, A and B). Thus, the p53-induced antitumor activities suppressed tumor growth in different tumors by different and sometimes overlapping mechanisms. We next addressed the role of the innate immune system in tumors after p53 reactivation in angiosarcomas. Inflammatory cell infiltration in tumors of different genotypes was determined by counting the number of polymorphonuclear cells and lymphocytes in areas devoid of focal inflammatory aggregates. p53^neo/– and p53^neo/R172H tumors harbored a slightly higher (although insignificant) number of interstitial inflammatory cells compared with that in their respective p53^neo/–Cre-ER^TM and p53^neo/R172HCre-ER^TM tumor counterparts (Supplemental Figure 10). Histological analysis also revealed that p53^neo/– and p53^neo/R172H tumors had occasional focal inflammatory aggregates that were rare or absent in p53^neo/–Cre-ER^TM and p53^neo/R172HCre-ER^TM tumors.

Figure 5

p53 restoration resulted in induction of senescence in angiosarcomas and decreased tumor cell proliferation. Tamoxifen-treated angiosarcomas collected in the MRI study were subjected to H&E staining; immunohistochemical staining for cleaved caspase-3 (CC3), Ki-67, senescence-associated β-galactosidase (SA-β-gal assays). (A) Images of analyses of the representative angiosarcomas from tamoxifen-treated p53^neo/–, p53^neo/–Cre-ER^TM, p53^neo/R172H, and p53^neo/R172HCre-ER^TM mice. Scale bar: 100 μm. (B) Percentages of cells staining positive for SA-β-gal and Ki-67 in A. Positive cells were counted in 4 random fields (original magnification, ×40) and presented as the mean ± SEM.

p53 response upon restoration in tumors. To monitor p53 reactivation in tumors upon initial tamoxifen treatment, we performed biopsies to harvest tumor tissue from mice bearing visible subcutaneous tumors. We compared the effect of wild-type p53 restoration in tumors from p53^neo/–Cre-ER^TM mice with those from p53^neo/R172HCre-ER^TM mice. In p53^neo/–Cre-ER^TM spindle-cell sarcomas (2 samples), p53 and p21 levels were elevated as early as 1 day after tamoxifen treatment. The highest levels of p53-induced apoptosis were observed 3 days after tamoxifen treatment (Figure 6, A and C). In p53^neo/R172HCre-ER^TM spindle-cell sarcomas (2 samples), immunohistochemistry revealed increased p53 levels 1 day after tamoxifen treatment (we cannot distinguish between wild-type and mutant p53 staining). Moreover, in the same tumors, p21 levels were elevated after tamoxifen treatment, suggesting wild-type p53 activity. Apoptosis was also induced in these tumors but to a lesser degree than in p53^neo/–Cre-ER^TM tumors (Figure 6, B and C). In a p53^neo/–Cre-ER^TM angiosarcoma, the tumor showed elevated p53 and p21 levels when examined at day 3 after tamoxifen treatment, and the restored p53 also activated senescence as revealed by SA-β-gal staining (Supplemental Figure 11). In a p53^neo/R172HCre-ER^TM angiosarcoma, p53 and p21 were also elevated 3 days after tamoxifen treatment. However, fewer cells were senescent upon p53 restoration in the p53^neo/R172HCre-ER^TM angiosarcoma than in the p53^neo/–Cre-ER^TM angiosarcoma (Supplemental Figure 11). These data suggest that the restored wild-type p53 was activated in tumors after the initial tamoxifen injection and that p53 restoration led to lower level of p53 activation in the p53^neo/R172HCre-ER^TM tumors than that in the p53^neo/–Cre-ER^TM tumors.

Figure 6

Activation of wild-type p53 in tumors after tamoxifen treatment. Tumor biopsy specimens harvested from the p53^neo/–Cre-ER^TM and p53^neo/R172HCre-ER^TM mice, before (day 0) and 1, 3, and 5 days after tamoxifen treatment (day 1, 3, and 5), were subjected to H&E staining and immunohistochemical staining for p53, p21, and cleaved caspase-3. (A) Images of analyses of biopsies from a representative spindle-cell sarcoma in a p53^neo/–Cre-ER^TM mouse (day 0, 1, 3, and 5). Scale bar: 100 μm. (B) Images of analyses of biopsies from a representative spindle-cell sarcoma in a p53^neo/R172HCre-ER^TM mouse (day 0, 1, 3, and 5). Scale bar: 100 μm. (C) Percentages of cells staining positive for p53, p21, and cleaved caspase-3 in A and B. Positive cells were counted in 4 random fields (original magnification, ×40) and presented as the mean ± SEM. Statistical significance between tumors with different genotypes was calculated using the t test, and the P value is less than 0.05 for p21 and cleaved caspase-3 at day 3 and day 5.

p53R172H dampens the activity of the restored wild-type p53. To explore the antagonizing effect of p53R172H on the restored wild-type p53, we collected angiosarcomas from the p53^neo/R172HCre-ER^TM and p53^neo/–Cre-ER^TM mice 3 days after tamoxifen treatment. First, we compared recombination at the p53^neo allele in these tumors using genomic DNA and real-time PCR. Recombination of p53^neo was comparable in the p53^neo/R172HCre-ER^TM and the p53^neo/–Cre-ER^TM tumors (P = 0.97; Supplemental Figure 12). Despite similar recombination frequencies, the p53^neo/–Cre-ER^TM angiosarcomas showed more cells with a senescent phenotype than the p53^neo/R172HCre-ER^TM angiosarcomas, as detected by SA-β-gal assays and immuno­histochemistry of senescence markers p16 and Dec1 (P = 0.019, P = 0.032, and P = 0.043, respectively; Figure 7, A and B). Expression of p21 and cyclin G1 (Ccng1), 2 p53 target genes, was also enhanced in the p53^neo/–Cre-ER^TM angiosarcomas relative to that in the p53^neo/R172HCre-ER^TM angiosarcomas (P = 0.0033 and P = 0.030, respectively; Figure 7, C and D).

Figure 7

p53R172H dampened the transcription activity of the restored wild-type p53. (A) Representative images of SA-β-gal assays and p16 and Dec1 staining of tamoxifen-treated p53^neo/–Cre-ER^TM and p53^neo/R172HCre-ER^TM angiosarcomas. Scale bar: 100 μm. (B) Percentage of SA-β-gal–, p16-, and Dec1-positive staining cells in p53^neo/–Cre-ER^TM and p53^neo/R172HCre-ER^TM angiosarcomas. (C) Comparison of mRNA levels of p21 and cyclin G1 (Ccng1) in p53^neo/–Cre-ER^TM and p53^neo/R172HCre-ER^TM angiosarcomas using real-time RT-PCR. p53^neo/R172H cells without Cre-ER^TM were used as controls. (B and C) Horizontal bars indicate the mean, and symbols indicate individual angiosarcomas. (D) Western blot analyses of p53, p21, and cyclin G1 (Ccng1) protein levels in control and tamoxifen-treated tumors. Protein levels were normalized to actin. (E) Recombination in p53^neo/–Cre-ER^TM and p53^neo/R172HCre-ER^TM tumor cell lines after 4-hydroxy-tamoxifen treatment. The results are presented as the mean ± SEM of 3 assays. (F) Comparison of DNA binding by p53 to p21 and puma promoters in p53^neo/–Cre-ER^TM and p53^neo/R172HCre-ER^TM tumor cells using ChIP assays. The acetylcholine receptor (AchR) promoter was used as a control, and the results are presented as percentage of input. Data are representative of 2 ChIP experiments performed in triplicate. (G) Comparison of mRNA levels of p21 and puma in p53^neo/–Cre-ER^TM and p53^neo/R172HCre-ER^TM tumor cells with (+) and without (–) 4-hydroxy-tamoxifen treatment, using real-time RT-PCR analysis. The results are presented as the mean ± SEM of 3 assays.

We also compared levels of p53 activation in isogenic p53^neo/R172HCre-ER^TM and p53^neo/–Cre-ER^TM MEFs treated with 4-hydroxy-tamoxifen (to recombine the p53^neo allele) and with doxorubicin (to activate p53). Cells from both genotypes had similar p53^neo recombination frequencies, but p53^neo/–Cre-ER^TM MEFs have a higher level of p53 target gene expression (p21, Ccng1, and p53 up-regulated modulator of apoptosis [puma]) than p53^neo/R172HCre-ER^TM MEFs (Supplemental Figure 13). In both tumors and MEFs, p53 levels were also upregulated, although we cannot distinguish between wild-type and mutant p53 proteins using Western blots (Figure 7D and Supplemental Figure 13). To determine whether the p53R172H mutant binds to wild-type p53, we transfected His-tagged wild-type p53 in p53^neo/–Cre-ER^TM and p53^neo/R172HCre-ER^TM sarcoma tumor cell lines that were established from untreated mice and then immunoprecipitated the p53 complex with the mutant p53-specific antibody, PAb240, that recognizes the mutant p53 conformation. As shown in Supplemental Figure 14, the p53R172H mutant bound His-tagged wild-type p53. To examine direct p53 binding to p21 and puma promoters, we performed ChIP experiments in the above p53^neo/R172HCre-ER^TM and p53^neo/–Cre-ER^TM sarcoma tumor cell lines. After 4-hydroxy-tamoxifen treatment, both cell lines showed similar p53^neo recombination frequencies (Figure 7E). Reproducibly, p53 bound to the p21 but not to the puma promoter (Figure 7F). More importantly, the p53^neo/–Cre-ER^TM tumor cells showed more p53 binding to the p21 promoter than p53^neo/R172HCre-ER^TM tumor cells (Figure 7F). These data were consistent with increased expression of p21 but not puma (Figure 7G). Thus, p53R172H dampened the activity of restored wild-type p53 through inhibition of wild-type p53 transcriptional activity.

In this study, we engineered a mouse model in which we restored wild-type p53 in tumors that contained a p53 missense mutation, since most human cancers harbor P53 missense mutations with gain-of-function and dominant-negative activity. In these mice, p53 restoration led to delayed tumor growth. The delay in tumor growth predicts a positive therapeutic effect of p53 restoration in p53 mutant tumors.

Tumor regression in our p53^neo/–Cre-ER^TM mice was consistent with that in previous reports, even though only 1 copy of the wild-type p53 allele was restored in our mice (13). The restored wild-type p53 induced apoptosis and senescence in tumors, indicating that the p53 pathway is reactivated in tumors and that absence of p53 is required for maintenance of the tumor phenotype. We also observed that tumor response to p53 restoration was tumor-type specific. p53 restoration mainly induced apoptosis in lymphomas and senescence in angiosarcomas. However, p53 restoration induced both apoptosis and senescence in giant-cell and spindle-cell sarcomas. An understanding of the mechanisms that trigger apoptosis or senescence should lead to better therapies.

Unlike the tumor regression phenotype observed in p53^neo/–Cre-ER^TM mice, we found arrested tumor growth in the p53^neo/R172HCre-ER^TM mice after p53 restoration, suggesting a dominant-negative effect of the p53^R172H mutation in vivo. The dominant-negative effect was first shown in in vitro cotranslation experiments, in which mutant p53 drove wild-type p53 into an inactive conformation after hetero-oligomerization (18). Hetero-oligomerization prevents wild-type p53 from binding to its target DNA sequence (19, 20). In the present study, the p53^neo/R172H mice had survival times similar to those of the p53^–/– and p53^R172H/– mice, indicating that the p53^R172H allele completely dominated the hypomorphic p53^neo allele. Therefore, the p53R172H mutant effectively abrogates the activity of the small amount of wild-type p53 expressed by the p53^neo allele but can not completely abolish normal levels of wild-type p53 activity, as indicated by the fact that p53^+/– and p53^+/R172H mice share similar survival curves.

Overexpression of mutant p53 shows a dominant-negative effect in cultured tumor cell lines (21–24). Here we demonstrated that the p53R172H mutant partially inhibited the activity of the restored wild-type p53 in tumors. Unlike the short-lived wild-type p53 protein, p53R172H is stabilized and accumulates to a high level in tumors (10). The stabilized mutant p53 increases the probability that hetero-oligomers form between the wild-type and the mutant p53, which would result in loss of the wild-type p53 activities. However, in our study, the restored wild-type p53 was not overwhelmed, as evidenced by the fact that in p53^R172H tumors, p53 retained some activity. We hypothesize that some wild-type p53 escaped from binding the mutant p53 and formed functional homo-tetramers.

Inactivation of wild-type p53 by the p53R172H mutant reduced levels of p53 activation in tumors after p53 restoration, as shown by the lower levels of apoptosis and senescence in p53^neo/R172HCre-ER^TM tumors than in p53^neo/–Cre-ER^TM tumors. Interestingly, we observed senescence in p53^neo/R172HCre-ER^TM lymphomas but not in p53^neo/–Cre-ER^TM lymphomas. This may be dose dependent, as stochastic assembly of fewer wild-type p53 tetramers may lead to senescence not apoptosis. In vitro studies suggest that the intracellular level of activated p53 affects promoter binding and transactivation (25, 26). A high level of active p53 initiates apoptosis because low-affinity p53 binding sites present in promoters of apoptosis-related genes are activated. On the other hand, p53, when present at low levels, binds only high-affinity sites present in promoters of genes involved in cell cycle arrest (27–29). Therefore, in p53^neo/–Cre-ER^TM lymphomas, the restored wild-type p53, in the absence of mutant p53, insured the activation of apoptosis. However, in p53^neo/R172HCre-ER^TM lymphomas, the decreased level of p53 activation caused by the dominant-negative effect of mutant p53 may have promoted the transactivation of genes involved in cell cycle arrest and senescence.

Our studies have revealed a dosage-dependent effect of p53-induced tumor suppression in vivo. First, the small amount of wild-type p53 expressed by the hypomorphic p53^neo allele delayed tumor onset and resulted in a shift in tumor spectrum toward sarcomas. Second, the decreased level of the functional wild-type p53 in the p53^R172H tumors led to arrested tumor growth but not tumor regression. We therefore speculate that excessive wild-type p53 will increase the level of functional p53 over that of mutant p53. The goal of p53 gene transfer in clinical trials is to achieve overexpression of wild-type p53 in human cancers (30). Our findings in mice lead us to speculate that this strategy could lead to tumor regression in humans. Likewise, other strategies, such as the use of the chemical drug PRIMA to restore the mutant p53 to a wild-type competent conformation, may synergize with restoration of wild-type p53, leading to tumor regression (31).

The use of genetically engineered mouse models bypasses some difficulties confronted by clinicians in designing p53 gene therapy. First, p53 restoration in mouse models is a one-hit genetic event that is guaranteed to happen. Second, the wild-type p53 is produced endogenously by a modified p53 locus, ignoring any delivery blockades that hinder P53 gene transfer into tumor cells of human patients. Additionally, some human cancers containing no P53 mutations may tolerate wild-type p53 restoration. These may account for the limited effectiveness of P53 gene therapy that has been thus far realized in clinical trials (32–36). Nevertheless, our results, together with those from previous studies, emphasize the potential therapeutic efficacy of p53 restoration in human cancer treatment. Specifically, the most important implication of the current findings is that p53 restoration has therapeutic potential in human cancers with P53 missense mutations, not just loss of P53.

View Supplemental data

This study was supported by NIH grants CA82577 and CA34936 to G. Lozano. G. Lozano is the recipient of the Mattie Allen Fair Research Chair at MD Anderson Cancer Center. We thank Sean Post for comments on the manuscript; Doug Webb, Jorge DeLacerda, and Charlie Kingsley for help with MRI; Angelito De Villa and Maurice J. Dufilho IV for help with histology; and Peirong Yang for help with animal care.

Address correspondence to: Guillermina Lozano, Department of Genetics, Unit 1010, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd., Houston, Texas 77030, USA. Phone: 713.834.6386; Fax: 713.834.6380; E-mail: gglozano@mdanderson.org.

Conflict of interest: The authors have declared that no conflict of interest exists.

Reference information: J Clin Invest. 2011;121(3):893–904. doi:10.1172/JCI44504.

Tamara Terzian’s present address is: Department of Dermatology, University of Colorado, Aurora, Colorado, USA.

1. Levine AJ, Finlay CA, Hinds PW. P53 is a tumor suppressor gene. Cell. 2004;116(2 suppl):S67–S69.
   View this article via: PubMed Google Scholar
2. Lane DP. Exploiting the p53 pathway for the diagnosis and therapy of human cancer. Cold Spring Harb Symp Quant Biol. 2005;70:489–497.
   View this article via: PubMed CrossRef Google Scholar
3. Bartek J, et al. Aberrant expression of the p53 oncoprotein is a common feature of a wide spectrum of human malignancies. Oncogene. 1991;6(9):1699–1703.
   View this article via: PubMed Google Scholar
4. Greenblatt MS, Bennett WP, Hollstein M, Harris CC. Mutations in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis. Cancer Res. 1994;54(18):4855–4878.
   View this article via: PubMed Google Scholar
5. Petitjean A, et al. Impact of mutant p53 functional properties on TP53 mutation patterns and tumor phenotype: lessons from recent developments in the IARC TP53 database. Hum Mutat. 2007;28(6):622–629.
   View this article via: PubMed CrossRef Google Scholar
6. Olive KP, et al. Mutant p53 gain of function in two mouse models of Li-Fraumeni syndrome. Cell. 2004;119(6):847–860.
   View this article via: PubMed CrossRef Google Scholar
7. Lang GA, et al. Gain of function of a p53 hot spot mutation in a mouse model of Li-Fraumeni syndrome. Cell. 2004;119(6):861–872.
   View this article via: PubMed CrossRef Google Scholar
8. Lozano G. The oncogenic roles of p53 mutants in mouse models. Curr Opin Genet Dev. 2007;17(1):66–70.
   View this article via: PubMed Google Scholar
9. Liu G, et al. High metastatic potential in mice inheriting a targeted p53 missense mutation. Proc Natl Acad Sci U S A. 2000;97(8):4174–4179.
   View this article via: PubMed CrossRef Google Scholar
10. Terzian T, et al. The inherent instability of mutant p53 is alleviated by Mdm2 or p16INK4a loss. Genes Dev. 2008;22(10):1337–1344.
    View this article via: PubMed CrossRef Google Scholar
11. Lowe SW, Cepero E, Evan G. Intrinsic tumour suppression. Nature. 2004;432(7015):307–315.
    View this article via: PubMed CrossRef Google Scholar
12. Xue W, et al. Senescence and tumour clearance is triggered by p53 restoration in murine liver carcinomas. Nature. 2007;445(7128):656–660.
    View this article via: PubMed CrossRef Google Scholar
13. Ventura A, et al. Restoration of p53 function leads to tumour regression in vivo. Nature. 2007;445(7128):661–665.
    View this article via: PubMed CrossRef Google Scholar
14. Martins CP, Brown-Swigart L, Evan GI. Modeling the therapeutic efficacy of p53 restoration in tumors. Cell. 2006;127(7):1323–1334.
    View this article via: PubMed CrossRef Google Scholar
15. Montes de Oca Luna R, Wagner DS, Lozano G. Rescue of early embryonic lethality in mdm2-deficient mice by deletion of p53. Nature. 1995;378(6553):203–206.
    View this article via: PubMed CrossRef Google Scholar
16. Jones SN, Roe AE, Donehower LA, Bradley A. Rescue of embryonic lethality in Mdm2-deficient mice by absence of p53. Nature. 1995;378(6553):206–208.
    View this article via: PubMed CrossRef Google Scholar
17. Hayashi S, McMahon AP. Efficient recombination in diverse tissues by a tamoxifen-inducible form of Cre: a tool for temporally regulated gene activation/inactivation in the mouse. Dev Biol. 2002;244(2):305–318.
    View this article via: PubMed CrossRef Google Scholar
18. Milner J, Medcalf EA. Cotranslation of activated mutant p53 with wild type drives the wild-type p53 protein into the mutant conformation. Cell. 1991;65(5):765–774.
    View this article via: PubMed CrossRef Google Scholar
19. Cho Y, Gorina S, Jeffrey PD, Pavletich NP. Crystal structure of a p53 tumor suppressor-DNA complex: understanding tumorigenic mutations. Science. 1994;265(5170):346–355.
    View this article via: PubMed CrossRef Google Scholar
20. Prives C. How loops, beta sheets, and alpha helices help us to understand p53. Cell. 1994;78(4):543–546.
    View this article via: PubMed CrossRef Google Scholar
21. Sun Y, Dong Z, Nakamura K, Colburn NH. Dosage-dependent dominance over wild-type p53 of a mutant p53 isolated from nasopharyngeal carcinoma. FASEB J. 1993;7(10):944–950.
    View this article via: PubMed Google Scholar
22. Park DJ, et al. Transactivational and DNA binding abilities of endogenous p53 in p53 mutant cell lines. Oncogene. 1994;9(7):1899–1906.
    View this article via: PubMed Google Scholar
23. Williams AC, Miller JC, Collard TJ, Bracey TS, Cosulich S, Paraskeva C. Mutant p53 is not fully dominant over endogenous wild type p53 in a colorectal adenoma cell line as demonstrated by induction of MDM2 protein and retention of a p53 dependent G1 arrest after gamma irradiation. Oncogene. 1995;11(1):141–149.
    View this article via: PubMed Google Scholar
24. Davis P, Bazar K, Huper G, Lozano G, Marks J, Iglehart JD. Dominance of wild-type p53-mediated transcriptional activation in breast epithelial cells. Oncogene. 1996;13(6):1315–1322.
    View this article via: PubMed Google Scholar
25. Liebermann DA, Hoffman B, Vesely D. p53 induced growth arrest versus apoptosis and its modulation by survival cytokines. Cell Cycle. 2007;6(2):166–170.
    View this article via: PubMed CrossRef Google Scholar
26. Das S, Boswell SA, Aaronson SA, Lee SW. P53 promoter selection: choosing between life and death. Cell Cycle. 2008;7(2):154–157.
    View this article via: PubMed CrossRef Google Scholar
27. Nakano K, Vousden KH. PUMA, a novel proapoptotic gene, is induced by p53. Mol Cell. 2001;7(3):683–694.
    View this article via: PubMed Google Scholar
28. Kaeser MD, Iggo RD. Chromatin immuno­precipitation analysis fails to support the latency model for regulation of p53 DNA binding activity in vivo. Proc Natl Acad Sci U S A. 2002;99(1):95–100.
    View this article via: PubMed Google Scholar
29. Weinberg RL, Veprintsev DB, Bycroft M, Fersht AR. Comparative binding of p53 to its promoter and DNA recognition elements. J Mol Biol. 2005;348(3):589–596.
    View this article via: PubMed CrossRef Google Scholar
30. Roth JA. Adenovirus p53 gene therapy. Expert Opin Biol Ther. 2006;6(1):55–61.
    View this article via: PubMed CrossRef Google Scholar
31. Bykov VJ, et al. Restoration of the tumor suppressor function to mutant p53 by a low-molecular-weight compound. Nat Med. 2002;8(3):282–288.
    View this article via: PubMed CrossRef Google Scholar
32. Gottesman MM. Cancer gene therapy: an awkward adolescence. Cancer Gene Ther. 2003;10(7):501–508.
    View this article via: PubMed CrossRef Google Scholar
33. McNeish IA, Bell SJ, Lemoine NR. Gene therapy progress and prospects: cancer gene therapy using tumour suppressor genes. Gene Ther. 2004;11(6):497–503.
    View this article via: PubMed CrossRef Google Scholar
34. Schuler M, et al. Adenovirus-mediated wild-type p53 gene transfer in patients receiving chemotherapy for advanced non-small-cell lung cancer: results of a multicenter phase II study. J Clin Oncol. 2001;19(6):1750–1758.
    View this article via: PubMed Google Scholar
35. Schuler M, et al. A phase I study of adenovirus-mediated wild-type p53 gene transfer in patients with advanced non-small cell lung cancer. Hum Gene Ther. 1998;9(14):2075–2082.
    View this article via: PubMed CrossRef Google Scholar
36. Zeimet AG, Marth C. Why did p53 gene therapy fail in ovarian cancer? Lancet Oncol. 2003;4(7):415–422.
    View this article via: PubMed Google Scholar
37. Evans SC, Viswanathan M, Grier JD, Narayana M, El-Naggar AK, Lozano G. An alternatively spliced HDM2 product increases p53 activity by inhibiting HDM2. Oncogene. 2001;20(30):4041–4049.
    View this article via: PubMed CrossRef Google Scholar
38. Dimri GP, et al. A biomarker that identifies senescent human cells in culture and in aging skin in vivo. Proc Natl Acad Sci U S A. 1995;92(20):9363–9367.
    View this article via: PubMed CrossRef Google Scholar
39. Terzian T, Wang Y, Van Pelt CS, Box NF, Travis EL, Lozano G. Haploinsufficiency of Mdm2 and Mdm4 in tumorigenesis and development. Mol Cell Biol. 2007;27(15):5479–5485.
    View this article via: PubMed CrossRef Google Scholar
40. Barboza JA, Iwakuma T, Terzian T, El-Naggar AK, Lozano G. Mdm2 and Mdm4 loss regulates distinct p53 activities. Mol Cancer Res. 2008;6(6):947–954.
    View this article via: PubMed CrossRef Google Scholar
41. Jackson JG, Pereira-Smith OM. Primary and compensatory roles for RB family members at cell cycle gene promoters that are deacetylated and downregulated in doxorubicin-induced senescence of breast cancer cells. Mol Cell Biol. 2006;26(7):2501–2510.
    View this article via: PubMed CrossRef Google Scholar