Development and translational imaging of a TP53 porcine tumorigenesis model

Cancer is the second deadliest disease in the United States, necessitating improvements in tumor diagnosis and treatment. Current model systems of cancer are informative, but translating promising imaging approaches and therapies to clinical practice has been challenging. In particular, the lack of a large-animal model that accurately mimics human cancer has been a major barrier to the development of effective diagnostic tools along with surgical and therapeutic interventions. Here, we developed a genetically modified porcine model of cancer in which animals express a mutation in TP53 (which encodes p53) that is orthologous to one commonly found in humans (R175H in people, R167H in pigs). TP53^R167H/R167H mutant pigs primarily developed lymphomas and osteogenic tumors, recapitulating the tumor types observed in mice and humans expressing orthologous TP53 mutant alleles. CT and MRI imaging data effectively detected developing tumors, which were validated by histopathological evaluation after necropsy. Molecular genetic analyses confirmed that these animals expressed the R167H mutant p53, and evaluation of tumors revealed characteristic chromosomal instability. Together, these results demonstrated that TP53^R167H/R167H pigs represent a large-animal tumor model that replicates the human condition. Our data further suggest that this model will be uniquely suited for developing clinically relevant, noninvasive imaging approaches to facilitate earlier detection, diagnosis, and treatment of human cancers.

The need for immediate and rapid progress in cancer detection, diagnosis, and treatment is apparent, with 1 of every 4 deaths in the United States resulting from cancer (1). Despite recent technological and pharmaceutical advancements, effective treatment strategies for many forms of cancer remain elusive. Ideally, methods to detect the very early stages of tumorigenesis would allow rapid, effective therapeutic intervention to ablate existing cancer cells and prevent tumor progression. Medical imaging modalities (i.e., computed tomography [CT], magnetic resonance imaging [MRI], and positron emission tomography [PET]) play important roles not only in detection of tumors, but also in discrimination of benign from malignant disease, treatment response evaluation, and screening for metastatic disease. Advancements in medical imaging technology, applications, and comprehensive validation of clinical utility are needed to significantly improve cancer mortality rates. Unfortunately, there are challenges associated with research in human cancer subjects, including varied genetics and population demographics, differing treatment strategies, and limited access to tissues for histopathological validation.

Nonhuman tumor models are the most feasible platform to develop new and improved imaging approaches for noninvasive cancer detection and monitoring. Murine models are the most commonly used species in cancer research; however, they typically do not reflect the heterogeneity of disease observed in humans (2), and the microimaging systems developed to study murine models lack the capabilities of current clinical imaging systems (3). Large animals, such as dogs and pigs, have been used to develop new CT imaging protocols for translation to human subjects (4–6) and improving the clinical capabilities to characterize early stages of disease (7). However, use of large animal species in cancer research has been limited, with most examples being spontaneous, incidental cancers in companion and livestock animals. Efforts have begun to move cancer-modeling technology previously developed in mice to new species. For example, xenograft models are being tested in immunodeficient pigs (8), and several groups, including our own in the present study, are using gene engineering technology to introduce cancer-related mutations to the porcine genome (9, 10).

TP53 (which encodes p53) is the most commonly inactivated gene in sporadic human cancers (11), and it is also mutated in the germline of Li-Fraumeni patients who are strongly predisposed to developing multiple types of cancers. It is estimated that p53 function is compromised in the vast majority of human tumors, through either TP53 gene mutation or alterations targeting the numerous regulators of p53 signaling (11–14). The primary role of p53 is to transcriptionally regulate genes involved in cell proliferation, apoptosis, DNA repair, autophagy, metabolism, and other critical processes that maintain genomic integrity of cells following genotoxic insults, thereby preventing tumorigenesis and metastasis (11). The majority of the cancer-associated mutations in p53 impair its binding to DNA, including the R175H “hot-spot” mutation, which occurs in Li-Fraumeni patients and is one of the most frequent missense mutations in sporadic tumors (15–17). Various TP53 mutations, including R175H, have been modeled in mice, which develop the expected range of tumors, including carcinomas, soft tissue and bone sarcomas, leukemia, and glioblastoma multiforme (18–21). However, there is a lack of genetically defined large-animal models of cancer that would be more ideally suited to longitudinal studies of tumor initiation, progression, and metastasis, as well as tumor remission in response to treatment, involving technology designed for human application (e.g., medical imaging, radiation oncology, and surgery).

Here, we generated and characterized a novel gene-targeted porcine model of cancer in which the orthologous R175H mutation (R167H in pigs) was introduced into the endogenous TP53 alleles. These TP53^R167H/R167H animals developed a spectrum of tumor lineages that mimicked those seen in humans with Li-Fraumeni syndrome and in mice expressing the same TP53 mutation. Importantly, the potential value and feasibility of clinical imaging approaches, including CT and MRI, to detect and monitor tumor formation was validated.

Targeted mutation of TP53 in porcine fetal fibroblasts. Gene targeting was used to introduce the R167H missense mutation in the endogenous TP53 gene. The TP53 R167H targeting vector was delivered into male and female fibroblasts harvested from Yucatan miniature pig fetuses using recombinant adeno-associated virus (rAAV) (Figure 1A, left). A similar approach was previously used to target porcine CFTR (22). After rAAV infection and antibiotic selection, a PCR screen identified 29 TP53-targeted cell lines (Figure 1A, middle). An allele-specific Southern blot confirmed that 24 of the PCR-positive clones also contained the intended R167H mutation (Figure 1A, middle), and overall gene targeting efficiencies of approximately 1% were achieved in both male and female cells (Table 1). A subset of PCR-positive, R167H-containing cells was selected for further characterization. Southern blots hybridized with TP53- and Neo^R-specific probes confirmed proper TP53 targeting and absence of random integration (Figure 1A, right). Furthermore, the intended recombination event was verified by DNA sequencing.

Figure 1

Production of TP53-targeted pigs. (A) Gene targeting via homologous recombination was used to introduce the R167H mutation in the endogenous TP53 gene (left). AflII (A) and EcoNI (E) sites and TP53 Southern blot probe are shown. PCR screen identified potentially targeted cell clones, and a biotin-labeled, allele-specific probe identified those contained the R167H mutation (middle). Amplified genomic DNA Southern blots confirmed properly targeted TP53^R167H/+ cell clones (right). (B) TP53^R167H/+ pigs were produced using somatic cell nuclear transfer (left). Genotypes were confirmed by genomic Southern blot (middle) and DNA sequencing (right). (C) A cross of TP53^R167H/+ males and females produced TP53^R167H/R167H pigs (left), with genotypes being confirmed by PCR (middle) and genomic Southern blot (right).

Table 1

Gene targeting and somatic cell nuclear transfer summary

Production of TP53-targeted pigs. TP53^R167H/+ cells were used as nuclear donors for somatic cell nuclear transfer (23). We performed a total of 7 embryo transfers (3 with male embryos, 4 with female embryos). These produced 5 full-term pregnancies resulting in 16 male and 29 female TP53^R167H/+ pigs (Figure 1B, left). PCR, Southern blot, and DNA sequencing were used to confirm the expected genotypes (Figure 1B, middle and right).

Upon reaching sexual maturity, TP53^R167H/+ males and females were mated, producing several litters whose genotypes were confirmed by PCR and genomic Southern blot (Figure 1C). We studied the offspring from 3 of the resulting litters, which contained 2 TP53^+/+, 7 TP53^R167H/+, and 6 TP53^R167H/R167H pigs. While this limited number of litters did not contain the expected Mendelian inheritance pattern, χ² analysis of genotypes from subsequent litters (now totaling 68 pigs) fit a 1:2:1 ratio.

Functional inactivation of the porcine p53-R167H mutant protein. Wild-type p53 is kept at low levels by one of its own transcriptional targets, the Mdm2 E3 ubiquitin ligase, to allow for normal cell survival and proliferation (24, 25). In response to various cellular stresses, such as DNA damage or oncogene activation, the p53 protein becomes stabilized and transactivates genes that induce growth arrest or apoptotic cell death (11, 26). This ensures the repair of damaged DNA before cells divide or the removal of genetically compromised cells by apoptosis (Figure 2A). By comparison, TP53 mutants such as human R175H (mouse R172H) are defective for wild-type p53 transcriptional function and accumulate to high levels in tumor cells due to loss of Mdm2 regulation (17, 18, 27). Failed induction of cell cycle–inhibitory genes, such as the cyclin-dependent kinase inhibitor CDKN1A (encoding p21; also known as Cip1 and WAF1), also abolishes normal checkpoint arrest by mutant p53. As a consequence, p53 mutant cells continue to cycle improperly and divide in the face of DNA damage. This initially causes greater genetic mutation and cell death, but ultimately leads to the outgrowth of genetically altered survivors that promote tumorigenesis (Figure 2A).

Figure 2

Defective molecular and biological activity of the porcine p53-R167H mutant protein. (A) Known differential response of cells expressing wild-type p53 (+/+) and mutated p53-R167H alleles (+/m or m/m) to DNA damage. Figure adapted with permission from Nature (26). (B) Western blots of p53, p21, and GAPDH (loading control) levels in p53 wild-type or R167 mutant pig fibroblasts. Cells were treated with DMSO vehicle (–) or 400 ng/ml doxorubicin (Dxn) for 3 days. Lanes were run on the same gel but were noncontiguous (black line). (C) Representative histograms of cell cycle distributions for doxorubicin-treated cells from B. Red, percent of cells arrested in G1 phase (2N DNA content); blue, cells undergoing apoptosis (<2N DNA content). (D) Mutant RAS induces cellular senescence in TP53^+/+ (+/+) cells, whereas TP53^R167H/+ (+/m) and TP53^R167H/R167H (m/m) cells escape mutant RAS-induced senescence and proliferate, during which time they sustain chromosomal alterations that foster cell transformation. (E and F) Pig fibroblasts were infected with pBabe-puro vector or pBabe-H-RAS^G12V-puro viruses and selected with puromycin for 7–10 days. Shown are (E) phase-contrast microscopy and (F) counts of SA-βgal–stained cells. Cells expressing mutant p53 evaded senescence and proliferated, whereas those with wild-type p53 adopted the enlarged, flattened morphology of senescent cells and stained positively for SA-βgal. Data in F are mean ± SD of 3 independent experiments. Scale bars: 100 μm. *P < 0.05, **P < 0.001 vs. vector, 2-way ANOVA.

We tested whether porcine p53-R167H behaves like its human and mouse orthologs by analyzing its expression and activity in primary pig fetal fibroblasts in the absence or presence of drug-induced DNA damage. As expected, mutant p53-R167H protein was more stable and highly expressed than wild-type protein under basal conditions (Figure 2B). DNA damage by doxorubicin treatment caused robust induction of p21 by wild-type p53 that coincided with G1 and G2 phase arrest of cells (Figure 2C). In contrast, dose-dependent loss of p21 induction and G1 phase arrest, along with an increase in apoptotic cell death (<2N DNA content), was seen in doxorubicin-treated cells expressing 1 or 2 alleles of p53-R167H (Figure 2, B and C). Similar in vitro findings have been reported for independently generated porcine fibroblasts expressing inducible TP53^R167H/R167H alleles (10).

In normal cells expressing wild-type p53, mutagenic activation of RAS or other oncogenes provokes a permanent p53-mediated cell cycle arrest called senescence that protects cells from chromosomal instability and oncogenic transformation (Figure 2D and refs. 11, 28). In contrast, TP53 mutation impairs its induction of senescence genes, such as CDKN1A, and allows cells to continue proliferating and acquiring genetic alterations in the face of mutant RAS expression. Consistent with those findings in mouse and human cells, pig fibroblasts expressing 1 or 2 TP53^R167H alleles effectively escaped mutant H-RAS–induced senescence (Figure 2, E and F, and Supplemental Figure 1; supplemental material available online with this article; doi:10.1172/JCI75447DS1). Wild-type cells infected with human H-RAS^G12V viruses expressed increased p21, became fully arrested, adopted the flattened and enlarged morphology of senescent cells, and expressed high levels of senescence-associated β-galactosidase (SA-βgal) relative to vector controls, as expected (Figure 2, E and F, and Supplemental Figure 1). By comparison, cells expressing 2 alleles of mutant TP53 continued to proliferate, failed to induce p21, were devoid of SA-βgal activity, and displayed no morphological features of senescent cells in response to mutant H-RAS expression. R167H-heterozygous fibroblasts exhibited an intermediate phenotype. Notably, H-RAS^G12V expression still had the ability to slow down the proliferation of p53 mutant cells (Supplemental Figure 1C), likely due to induction of the cell cycle inhibitor p16^INK4a (28). Together, these analyses confirmed that the porcine p53-R167H mutant protein was functionally defective in mediating DNA damage and oncogene checkpoints, similar to its human and mouse counterparts.

Tumor development in TP53^R167H/R167H pigs. TP53^R167H/R167H, TP53^R167H/+, and TP53^+/+ pigs were monitored with a combination of clinical evaluation, peripheral blood analysis, in vivo CT and MRI, histopathology, and genetic characterization. The human-like size of this Yucatan miniature pig model enabled the use of clinical imaging technology and protocols (CT and MRI) for the longitudinal monitoring of tumorigenesis in this cohort. Supplemental Table 1 provides a summary of the pigs that underwent necropsy, including age at necropsy, clinical signs, and type of tumors detected. In addition to these necropsied animals, 7 TP53^R167H/+ and 5 TP53^+/+ pigs were monitored with in vivo imaging for tumor development. No tumors were detected via in vivo imaging or necropsy in the TP53^R167H/+ or TP53^+/+ cohort over 30 months of observation. Continued monitoring of the cohort is being conducted to determine whether tumors do develop in TP53^R167H/+ pigs at an advanced age.

All TP53^R167H/R167H pigs that reached sexual maturity developed neoplastic lesions, including lymphomas, osteogenic tumors, and a renal tumor. However, 1 pig (case 1) in the first litter of TP53^R167H/R167H pigs died 2 hours after birth. The necropsy did not reveal any discreet tumors, although the liver did have microscopic evidence of hepatocellular atypia (Supplemental Figure 2). Although no other pigs in our cohort displayed similar atypia or overt tumors in any livers, TP53 mutations are associated with hepatocellular carcinoma (29); therefore, future phenotypic studies of the model will help determine whether this represents a precursor lesion for hepatocellular tumors.

In the TP53^R167H/R167H cohort, 3 cases of lymphoma were identified and confirmed histopathologically (cases 2–4), with necropsy occurring over a range of 27–43 weeks of age. In vivo CT and MRI data collected for case 4 showed lymphadenomegaly consistent with lymphoma 16 weeks prior to necropsy (27 weeks of age; Figure 3). We observed enlargement of the liver and spleen, both indicators of possible diffuse infiltration of the liver with neoplastic lymphocytes (30, 31); thus, further examination of these organs was performed. The volume of the liver in TP53^R167H/R167H case 4, as assessed by CT, was found to be much larger than its weight-matched TP53^R167H/+ control (923 ml vs. 659 ml, respectively; Supplemental Figure 3I). The liver volume/animal weight ratio in TP53^R167H/R167H case 4 was also larger compared with that in a TP53^R167H/+ cohort (26.4 ml/kg vs. 15.6 ml/kg, respectively). A mild elevation in quantitative CT–determined spleen volume was noted in the TP53^R167H/R167H lymphoma case compared with the weight-matched TP53^R167H/+ pig (532 ml vs. 486 ml, respectively) and evidence of spleen enlargement was also reflected in the corresponding values of the spleen volume/animal weight ratio for the TP53^R167H/R167H lymphoma case compared with the TP53^R167H/+ cohort (15.2 ml/kg vs. 13.5 ml/kg, respectively; Supplemental Figure 3I).

Figure 3

Lymphoma in TP53^R167H/R167H pigs. Early-stage enlargement in inguinal lymph nodes were detected. (A) Coronal CT cross-sections illustrating the difference in lymph node short axis size: 26 mm in TP53^R167H/R167H (case 3) relative to 9 mm in TP53^+/+ (case 7). Image data were acquired 4 months prior to necropsy at 27 weeks of age. (B and C) Lymphadenopathy was a consistent finding upon external examination (arrows, B; case 3) and at necropsy (C; case 6). (D) Lymph nodes were consistently effaced by sheets of neoplastic lymphocytes (asterisks; case 3). (E) Microscopically, the neoplastic lymphocytes had numerous mitotic figures (arrows) and frequent tingible body macrophages (arrowheads). Original magnification, ×100 (D); ×600 (E).

Elevated levels of contrast enhancement and delayed contrast clearance in both liver and spleen were observed by 3D fast gradient echo gadolinium-enhanced MRI imaging. Furthermore, quantitative assessment of the contrast-enhanced MRI study was performed comparing the liver/muscle and spleen/muscle MRI signal ratios. Both the liver/muscle ratio (1.13 vs. 0.89, respectively) and the spleen/muscle ratio (1.02 vs. 0.95 respectively) were increased in the TP53^R167H/R167H lymphoma case compared with the weight-matched TP53^R167H/+ pig. Heterogeneity in signal enhancement in the organs in the TP53^R167H/R167H lymphoma case compared with a weight-matched TP53^R167H/+ pig was reflected through calculation of the coefficient of variation in signal intensity in the spleen (27% vs. 14%, respectively) and liver (15% vs. 10%, respectively). Heterogeneous organ perfusion and delayed clearance of contrast media in the TP53^R167H/R167H pig supported an imaging diagnosis of lymphoma.

At necropsy, enlargement of the lymph nodes and spleen were prominent and consistent features in affected TP53^R167H/R167H pigs (Figure 3, B–E, and Supplemental Figure 3, A and B). Microscopically, sheets of neoplastic lymphocytes effaced tissue architecture in these organs (Figure 3, D and E, and Supplemental Figure 3B), with extension into liver and lung (Supplemental Figure 3, E–H). In case 2, the animal died suddenly and had hemo-abdomen secondary to splenic rupture (Supplemental Figure 3, C and D). The lungs (and other organs) had vessels filled with abundant cellular and nuclear debris (Supplemental Figure 3, E and F), histologically consistent with tumor lysis syndrome, seen in many mouse models (32, 33) and in humans undergoing cytolytic cancer therapy (34, 35). Case 2 had routine blood work evaluated 2 days prior to its sudden death. At that time, the absolute lymphocyte count of case 2 was similar to or marginally increased compared with control littermates, but all these values were within the normal reference range (3.3–11.5 × 10³ cells/μl). This suggests that leukemia was not present and that the necrosis/rupture of the spleen in the hours prior to death was likely the mechanism for tumor cells and cellular debris to enter into the vasculature as emboli. All 3 TP53^R167H/R167H lymphoma cases had changes in selected peripheral blood parameters compared with control animals in the final days prior to necropsy (Figure 4, A–D). Interestingly, a routine blood sample collected 2 days prior to necropsy did not distinguish case 2 from the other lymphoma cases (cases 3 and 4), which suggests that the tumor lysis syndrome was associated with acute splenic rupture as opposed to tumor burden.

Figure 4

Peripheral blood parameters. Nondiseased control (cases 7, 9, and 10) and lymphoma cases (cases 2–4) in TP53^R167H/R167H pigs were examined ~4–6 weeks prior to (time period 1) and the week of (time period 2) euthanasia/death. TP53^R167H/R167H pigs were similar to nondiseased controls at time period 1, but had significant decreases compared with controls in albumin (P < 0.001; A), red blood cells (P < 0.001; B) and platelets (P < 0.05; C) at time period 2. (D) Lymphocyte counts were similar in TP53^R167H/R167H and control pigs at time period 1, but by time point 2 there was a nonsignificant trend toward increased numbers in the TP53^R167H/R167H animals. *P < 0.05, 2-way repeated-measures ANOVA with Bonferroni post-test.

Osteogenic tumors were also identified through in vivo imaging in 2 TP53^R167H/R167H pigs (cases 5 and 6). CT and MRI detected a 28 mm right parasagittal, calvarial lytic (137 HU) tumor in case 5 at 53 weeks of age. The tumor expanded into the intracranial cavity and also caused destruction of adjacent bone tissue (Supplemental Figure 4). In vivo CT and MRI data collection was completed for case 6 at 2 time points. At 53 weeks of age (time point 1), osteogenic tumors in the long bones were detected, but no skull/calvarial tumor was present; at 60 weeks of age (time point 2; 51 days later), a 39 mm intracranial tumor was detected (Figure 5). The tumor infiltrated through the bony calvarium to invade both the frontal sinuses and the intracranial compartment, with evidence of extrinsic compression of the brain parenchyma. The tumor contained bony matrix, as indicated by increased CT attenuation (305 HU). This was supported by the heterogeneous signal enhancement present in the SPACE MRI data for both case 5 and case 6 calvarial tumors, indicative of areas of both soft tissue and calcification/ossification. Microscopically, the 2 calvarial tumors were composed of spindled to epitheloid cells that produced osteoid trabeculae (Figure 5 and Supplemental Figure 4). Both tumors had invasion, with the case 5 tumor infiltrating into the frontal sinus(es) (Figure 5, A–D). The sum of the imaging, histological, and biological features were consistent with osteosarcoma.

Figure 5

Osteogenic tumor a in a TP53^R167H/R167H pig. In vivo imaging of case 6 with CT, showing coronal and sagittal cross-sectional views. (A) Whereas views from time point 1 showed no evidence of tumor lesion, views from time point 2 in CT and MRI (51 days later) revealed aggressive tumor growth (39 mm). (B and C) The extradural mass (arrows, B) expanded into the cranial vault (arrows, C), compressing the brain. (D) The mass (asterisks) invaded into the frontal sinus cavity. (E) The calvarial mass was composed of round to spindle cells that produced irregular trabeculae of osteoid (asterisks). (F) The neoplastic cells had varying sized nuclei with a prominent nucleoli (arrows). Original magnification, ×200 (E); ×600 (F).

The microscopic lesions in the calvarial tumors were similar in appearance, whereas the other osteogenic tumors in these animals had a varied microscopic appearance, even within the same pig. These changes ranged from osteoid/bone-rich to osteoid-poor lesions with proliferative mesenchymal cells (Figures 6 and 7 and Supplemental Figures 5 and 6). These other osteogenic tumors were located in the tibia, femur, and sacrum, ranging in size from 8 to 18 mm (cases 5 and 6; Table 2). Lytic lesions presented with mean CT attenuations of 87–156 HU. Heterogeneous and hyperdense lesions ranged 307–927 HU in mean CT attenuation. The osteogenic tumors in the long bones of case 6 were all present at time point 1, prior to development of the calvarial tumor. No significant change in tumor size was detected for any of the osteogenic tumors of the long bones, while some subtle changes in CT attenuation were observed, such as the increased solid component of the tumor in Figure 6. Importantly, no osteogenic tumors were seen in long bones of the pigs without osteosarcoma tumors (littermate TP53^R167H/R167H, TP53^R167H/+, and TP53^+/+ pigs).

Figure 6

Osteogenic tumor in a TP53^R167H/R167H pig. (A) Imaging of case 6 showed an 18 mm lesion in the left distal femur with heterogeneous CT density. Views at time point 2 showed growth in the solid component of the lesion compared with time point 1. (B) The tumor was composed of irregular trabeculae of osteoid and bone extending from the edge of the cortex (asterisks). (C) The bony trabeculae were at times lined by loose mesenchymal to fibrous tissue and uncommon rims of osteoblasts. Original magnification, ×20 (B); ×400 (C).

Figure 7

Osteogenic tumor in a TP53^R167H/R167H pig. (A) In case 5, a 13 mm lytic tumor was identified by noncontrast CT in the left distal femur. (B and C) The tumor (arrows, B) effaced bone marrow (note absence of trabeculae at right edge, B) and was composed of sheets of pleomorphic round to spindle cells with scant osteoid deposition and multinucleate osteoclasts (arrows, C) were prominent within the tumor. Original magnification, ×40 (B); ×400 (C).

Table 2

Examples of imaging techniques linking to pathology findings in the TP53^R167H/R167H pigs

In one of the TP53^R167H/R167H pigs with osteogenic tumors (case 6), a renal tumor was also detected via CT and MRI imaging (Figure 8, A–D). The tumor (25.8 mm; 8.9 ml) was located in the right cranial pole of the kidney, with an accompanying focus of hemorrhage (23.4 ml), at time point 1. Image data of the tumor at time point 2 showed nominal changes in tumor size (26.3 mm; 7.1 ml), but had obvious resolution of the hemorrhage component. Contrast-enhanced CT attenuation of the tumor decreased from time point 1 to time point 2 (62 HU to 55 HU, respectively), reflecting an increase in necrotic tissue component, which was supported by histopathology findings (Figure 8, E and F).

Figure 8

Renal tumor in a TP53^R167H/R167H pig. In case 6, the change in renal tumor size over time was assessed with iodinated contrast-enhanced CT at time points 1 (A) and 2 (B); sagittal views are shown. (C) The renal tumor (arrow) was elevated from the cortex at the cranial pole of the right kidney. (D) Cut surface of the tumor (T) showed an encapsulated (arrow) mass that was well-demarcated from normal kidney (K). (E) The tumor was composed of necrosis and hemorrhages (red patches) that were distinct from the unaffected kidney. (F) The necrotic tumor tissue had irregular tubules separated by cords of connective tissue (asterisks). (G) At the edges of the tumor, invading through the tumor capsule were small nests (see E, arrows) of poorly differentiated tubules with a high mitotic rate (arrows). Original magnification, ×20 (E); ×100 (F); ×600 (G).

Signal intensity in the tumor relative to the unaffected renal cortex was assessed within the MRI data. Precontrast, T1-weighted (VIBE) MRI data revealed a shift from time point 1 to time point 2 from isointense to hypointense signal relative to the renal cortex. The hemorrhage in time point 1 was clearly evident as a hyperintense component. Postcontrast, T1-weighted (VIBE) MRI demonstrated a heterogeneous enhancement at time point 1 and a decrease in signal intensity relative to the renal cortex at time point 2. The resolution of the hemorrhage component and increase in necrosis between the time points is likely responsible for the relative decrease in signal intensity in the follow-up postcontrast MRI data.

Microscopically, the renal tumor was composed of necrotic tubules and intervening cords of connective tissue and foci of hemorrhage (Figure 8, E and F). Within and extending through the tumor capsule were small, invading nests of poorly differentiated tubules that had a high mitotic rate (Figure 8, E and G). The imaging and pathological findings were consistent with a Wilms tumor (nephroblastoma), a malignancy commonly associated with p53 pathway perturbations in which TP53 mutation correlates with a less favorable prognosis (36–38).

In addition, a mesenteric tumor was detected for case 6 (Supplemental Figure 6). The tumor shape was highly irregular and spiculated, with high mean CT attenuation compared with the surrounding soft tissues (472 HU vs. 30–50 HU, respectively). A significant growth in tumor volume from 2.07 ml to 3.08 ml (48% increase) was measured from imaging time point 1 to time point 2, with minimal change in mean CT attenuation (+2 HU). At necropsy, the irregularly shaped bony tumor was located within the mesentery (Supplemental Figure 6). The location of the bony mesenteric tumor was near the site of hemorrhage from the renal tumor (Figure 8); this and its general appearance were consistent with mesenteric myositis ossificans (Supplemental Table 2).

Molecular genetic analyses of TP53^R167H/R167H tumors. Several excised lymph nodes and osteogenic tumors were subjected to Western blot analyses to confirm p53-R167H mutant protein expression and activity. Compared with the essentially undetectable expression of nonmutated p53 from wild-type pig brain tissue, mutant p53-R167H protein was expressed at higher levels in each malignant sample (Figure 9A), as anticipated. The porcine p53-R167H mutation, like its human (R175H) and mouse (R172H) counterparts, is expected to behave as a gain-of-function mutant that fails to regulate normal p53 transcriptional targets (such as p21; see Figure 2 and Supplemental Figure 1B) and instead transactivates a new set of oncogenic gene targets, including cyclin B1 (17, 39). Indeed, tumors and lymph nodes expressing p53-R167H showed marked upregulation of cyclin B1 protein (Figure 9A). Human tumors expressing mutant p53 primarily sustain chromosomal alterations as opposed to genomic sequence mutations (40). We conducted cytogenetic analyses of metaphases from a lymph node and osteosarcoma isolated from TP53^R167H/R167H pigs, which revealed significant abnormalities in both chromosome number and structure (Figure 9, C and D). In contrast, a normal karyotype was seen in the skin tissue of a wild-type pig control (Figure 9B). These findings confirm the altered expression and activity of mutant p53-R167H in the TP53^R167H/R167H tumors, as well as the characteristic chromosomal instability known to accompany TP53 mutation in human tumors.

Figure 9

Molecular changes and cytogenetic abnormalities in TP53^R167H/R167H pig tumors. (A) Western blots show increased expression of mutant p53-R167H protein and its transcriptional target, cyclin B1, in lymph nodes (LN) and osteogenic tumor (Os) from TP53^R167H/R167H pigs (m/m), relative to low levels of each wild-type protein in the brain (Br) of normal TP53^+/+ pigs. GAPDH levels served as the loading control. The case number for each specimen is indicated. (B–D) Representative karyotypes from wild-type TP53^+/+ pig skin tissue (B), TP53^R167H/R167H lymph node (C; case 5), and TP53^R167H/R167H osteosarcoma (D; case 5) showing abnormal chromosome number and structures in the malignant cells.

Large animals are becoming increasingly relevant for modeling human diseases, especially for cases in which body size and lifespan are important. This is particularly true for cancer, for which imaging technologies and surgical procedures are instrumental in current detection and treatment plans. Conventional small-animal models such as mice have inherent limitations, including a body size approximately 1:1,000 that of humans and a lifespan of only 1–2 years. To address the lack of a suitable large-animal model of cancer, we mutated the porcine TP53 gene and generated pigs with lymphoma, renal, and osteogenic tumors. Importantly, clinically relevant imaging techniques were used to identify each tumor type and direct the postmortem collection of samples for histopathological and molecular analyses. During the course of the present study, another group reported a similar gene modification in domestic pigs; however, no cancer was described (10). To our knowledge, the TP53 mutant pigs described here are the first gene-targeted large animals with confirmed cancer development.

Advantages of the porcine cancer model. Specific attributes of the porcine model offer new research opportunities that are not feasible in current cancer models; 2 prime examples are size and lifespan. We chose to develop the mutant TP53 model in the Yucatan miniature pig, a breed that reaches a full-grown weight of 60–75 kg at 2 years of age, similar to that of an adult human. This makes it better suited than small-animal models for studies involving the optimization and validation of imaging technologies and surgical procedures. It is important to note that a model can also be too large. For example, domestic pig breeds, such as those being used in other cancer models (8–10), are 2–3 times bigger. This would exclude domestic pigs from most longitudinal monitoring/treatment studies with clinical imaging technologies, as they would quickly outgrow the size capacity of a typical clinical imaging scanner with a bore diameter of 60–70 cm. Also related to size, as opposed to small-animal models, tissues from Yucatan miniature pigs are available in sufficient quantities to develop porcine tissue repositories containing cancer and matched normal tissues for future studies. Likewise, the larger circulatory volume of Yucatan miniature pigs compared with mice permits more repeated sampling of peripheral blood for potential biomarkers of cancer, enabling a broader spectrum of disease parameters to be monitored over time. Finally, the lifespan of pigs (10–15 years) is quite long compared with other animal models, facilitating investigations of the natural history of various cancers, long-term treatment effects, and successive surgical interventions.

A large-animal cancer model presents opportunities to improve clinical cancer outcomes in humans by serving as a controlled surrogate for human cancer patients. Consequently, research studies may be designed with a small target population of animals that may be studied extensively, with unrestricted access to validating tissue samples. Thus, large amounts of data, with tightly controlled cofactors and low population variation, may be achieved, requiring fewer subject numbers for equivalent statistical significance than a human cancer cohort. Furthermore, many cancer types are heterogeneous in content, composed not only of cancerous cells, but also of inflammatory mediators (41), fibrotic stroma (42), necrotic tissue, and complex vascularity (43). Our porcine model does not rely on an immunocompromised host for tumor development, unlike xenograft models (8). The mutant TP53 pig opens new lines of investigation to explore the interplay between a functional immune response and the tumor microenvironment as it relates to tumor progression, diagnosis, treatment, and monitoring (44).

Applications of the porcine cancer model. The porcine cancer model provides new opportunities to improve cancer detection, monitoring, and treatment approaches in ways not possible with murine models or human patients. Medical imaging technologies (e.g., CT, MRI, and PET) play a key role in the clinical management of cancer patients in facilitating cancer detection, staging, and treatment response monitoring. However, the advanced capabilities of these systems are in many cases underutilized clinically, due to significant challenges in protocol optimization and validation within the cancer patient population (45). A large-animal cancer model allows for the collection of many imaging datasets with varying protocols and modalities in the same cancerous tumor. Thus, image data features can be directly compared, and optimal protocols and/or modalities may be identified. This capability would be highly useful in the optimization of radiation dose, contrast delivery rate, and scan delay timing for dynamic dual-energy CT (46). Comparison of alternate PET radiotracers, along with improved acquisition strategies to optimize PET resolution, would also benefit from a human-sized cancer model (47, 48).

While microimaging technology exists for CT, PET, and MRI data acquisition in murine cancer models, these systems do not contain the same capabilities of the equivalent clinical systems (3, 49). These differences in acquisition technology — combined with large differences in subject size as well as heart, respiratory, and metabolic rates — make rodents ineffective models for translatable protocol development. In a human cancer cohort for research, there is often a high degree of variability in cancer type and stage, clinical treatment approach, subject comorbidities, and access to endpoint data (e.g., mortality and comprehensive histopathology). There are also restrictions that limit the ability to acquire multiple data points, such as minimizing exposure to medical ionizing radiation, study time limitations, and primary concern for the patient’s clinical treatment and survival. The porcine model overcomes these constraints.

Another important capability provided by the porcine cancer model is to use medical imaging technology to monitor cancer development over time, with or without intervention. Studying cancer progression in the absence of treatment is not possible in humans. In this porcine model, data may be collected prior to tumor development, during tumor growth, and through to metastatic disease. This will provide insight into cancer pathogenesis and the ability of current medical imaging technologies to capture and classify the changing states of cancer progression. Medical imaging is key in planning clinical interventions. The similar anatomy and physiology to humans makes this porcine model ideal to test image-guided, minimally invasive surgical techniques as well as intensity-modulated radiation therapy strategies (50). Comprehensive postintervention evaluation in the porcine model is also feasible, using multimodality imaging and tissue resection to determine whether effective destruction of all cancer cells is achieved, whether early indicators of metastatic disease can be found in other target organs, and/or whether toxicity occurs in nondiseased tissue.

TP53 and cancer. TP53 mutations, which occur in the majority of sporadic human cancers and some inherited cancer-prone disorders, compromise protective checkpoints in cells that normally ensure genomic integrity, thereby facilitating cellular transformation and tumorigenesis (11, 17). In the present study, all TP53^R167H/R167H pigs that reached maturity developed some type of cancer, including lymphoma, Wilms tumor (nephroblastoma), and osteogenic tumors (osteosarcoma). Each tumor type is associated with mutations of TP53 and/or its pathways in humans and is found in Li-Fraumeni patients expressing the p53-R175H mutant modeled in our TP53^R167H/R167H pigs (51–53). The complete tumor penetrance obtained in TP53^R167H/R167H pigs, and the corresponding lack of tumor development in wild-type control animals, indicates that the cancers arising in these pigs are due to TP53 gene targeting and not spontaneous tumorigenesis, whose incidence for the tumor types observed is exceedingly low (between 2–20 per 100,000; refs. 54, 55).

The osteogenic tumors seen in TP53^R167H/R167H pigs may be of particular clinical interest and relevance. While Li-Fraumeni patients develop this type of tumor (51, 53), most osteosarcomas seen in humans actually arise sporadically and are associated with somatic TP53 mutations (56), which correlates with greatly reduced event-free survival (57, 58). Importantly, while the clinical aspects of osteosarcoma are well defined (Supplemental Table 2 and ref. 59), the early pathogenesis of this malignancy is not well understood. A prior study of transformation from benign to malignant osteogenic tumors suggests that tumor progression is associated with an increased frequency of genetic alterations (60). Chromosome clonality (61) and TP53 alterations (62) have been described in benign osteogenic tumors, and subsequent osteosarcoma arising from a benign osteogenic tumor has been reported (63). Here, the TP53^R167H/R167H pigs had a wide spectrum of osteogenic tumors ranging from osteosarcoma of the calvaria to various low-grade/benign tumors (e.g., giant cell tumor of bone) of long bones. Metastases could explain the multiple osteogenic tumors in a single animal; however, the timing and biological appearance of tumors at calvarial versus appendicular sites, together with the lack of lymph node and pulmonary metastases, would suggest that these tumors are de novo in origin. It will be informative to study the clonality of these individual tumors and determine whether their biology and genetics change over time to match their progression to osteosarcoma, as is seen in humans. The ability to conduct broader interspecies comparisons among humans, mice, and now pigs should strengthen the probability of identifying common and potentially causative molecular events required for malignant conversion to osteosarcoma (64).

Limitations. Although the initial diagnostic characterization is promising, this model does possess some limitations. First, the spectrum of tumors thus far was restricted, in part, due to the experimental design of this study that prioritized detection and phenotypic assessment of the initial tumors in the model. Mouse models with nonfunctional p53 commonly have early clinical disease due to lymphoma, with an expanded range of sarcoma and carcinoma tumor types that develop over time. As more TP53-targeted pigs are studied over a longer time, we expect to observe progression of tumors already seen in this study, as well as development of additional tumor types frequently seen in humans and mice with TP53 mutations. For example, although breast carcinomas are the most frequent tumor observed in Li-Fraumeni patients, we did not detect any in this initial cohort. Second, since the timing and location of tumor presentation in these pigs was random, longitudinal monitoring (e.g., imaging) may be required for some studies. Third, our detection and phenotypic evaluation of early tumorigenesis in the TP53-targeted pig model likely precluded development of metastases that occur in later stages of disease. Those issues notwithstanding, the mutant TP53-targeted pigs represent a powerful platform for developing other cancer models with conditional, tissue-specific tumors. Of great interest is modeling cancer types such as lung, ovarian, and pancreatic adenocarcinomas, in which earlier detection by innovative imaging and/or biomarker analyses should translate into marked improvements in clinical management and patient survival. In that regard, the mutant TP53 heterozygous pigs were 30 months old at the end of the present study and lacked evidence of tumor development. We anticipate these animals will get tumors at a slower rate than the mutant TP53 homozygotes. This delay provides an ideal window of time for examining cooperative tumorigenic effects of mutations to other cancer genes in the heterozygote TP53 model. For example, we are currently pursuing models of several tumor types by adding conditional oncogene activation (e.g., KRAS) to the heterozygote TP53 background. This strategy has been extremely successful in murine models (65, 66), and we anticipate that success will extend to pigs, as well.

Conclusion. In summary, we have generated a new large-animal model of cancer by targeting porcine TP53. Using translational protocols and techniques, we demonstrated that the porcine model recapitulated many aspects of human cancer development. We anticipate that porcine cancer models will provide a comparatively low-cost platform (relative to human studies) for optimizing tumor diagnostic methodologies, including novel imaging approaches that can be directly applied in the clinic. Indeed, these long-lived tumor models should address several unmet needs and provide new opportunities for longitudinal studies of tumorigenesis (e.g., initiation, progression, and response to treatment) that will ultimately improve our ability to detect and treat cancers earlier in humans.

View Supplemental data

This research was funded in part by Exemplar Genetics and the NIH via the University of Iowa Environmental Health Sciences Research Center (grant P30 ES005605), the Institute for Clinical and Translational Science (grant UL1 TR000442), and the Holden Comprehensive Cancer Center (grant P30 CA086862). This publication’s contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIH. We thank Daniel Thedens, Jered Sieren, Nicholas Stoyles, David Stoltz, Mahmoud Abou, Mark Hoegger, Brian Huedepohl, and Cara Haden as well as the Comparative Pathology Laboratory and the Shivanand R. Patil Cytogenetics and Molecular Laboratory for technical assistance. We also thank John Buatti, Frederick Domann, Aliasger Salem, Chuck Sherr, Martine Roussel, Lezlee Coghlan, Donna Kusewitt, and John Swart for helpful discussions.

Address correspondence to: Christopher S. Rogers, Exemplar Genetics, 2500 Crosspark Rd., Ste E126, Coralville, Iowa 52241, USA. Phone: 319.665.2664; Email: chris.rogers@exemplargenetics.com.

Conflict of interest: Xiao-Jun Wang, Bryan T. Davis, Judy A. Rohret, Jason T. Struzynski, Frank A. Rohret, and Christopher S. Rogers are employees of Exemplar Genetics, a company that has applied for a patent related to the work reported herein.

Reference information: J Clin Invest. 2014;124(9):4052–4066. doi:10.1172/JCI75447.

1. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA Cancer J Clin. 2012;62(1):10–29.
   View this article via: PubMed CrossRef Google Scholar
2. Cheon DJ, Orsulic S. Mouse models of cancer. Annu Rev Pathol. 2011;6:95–119.
   View this article via: PubMed CrossRef Google Scholar
3. Ritman EL. Current status of developments and applications of micro-CT. Annu Rev Biomed Eng. 2011;13:531–552.
   View this article via: PubMed CrossRef Google Scholar
4. Won C, et al. CT-based assessment of regional pulmonary microvascular blood flow parameters. J Appl Physiol (1985). 2003;94(6):2483–2493.
   View this article via: PubMed Google Scholar
5. Chon D, Beck KC, Larsen RL, Shikata H, Hoffman EA. Regional pulmonary blood flow in dogs by 4D-X-ray CT. J Appl Physiol (1985). 2006;101(5):1451–1465.
   View this article via: PubMed Google Scholar
6. Dakin JH, Evans TW, Hansell DM, Hoffman EA. Regional pulmonary blood flow in humans and dogs by 4D computed tomography. Acad Radiol. 2008;15(7):844–852.
   View this article via: PubMed CrossRef Google Scholar
7. Alford SK, van Beek EJ, McLennan G, Hoffman EA. Heterogeneity of pulmonary perfusion as a mechanistic image-based phenotype in emphysema susceptible smokers. Proc Natl Acad Sci U S A. 2010;107(16):7485–7490.
   View this article via: PubMed CrossRef Google Scholar
8. Basel MT, et al. Human xenografts are not rejected in a naturally occurring immunodeficient porcine line: a human tumor model in pigs. Biores Open Access. 2012;1(2):63–68.
   View this article via: PubMed Google Scholar
9. Flisikowska T, et al. A porcine model of familial adenomatous polyposis. Gastroenterology. 2012;143(5):1173–1175 e1171.
   View this article via: PubMed Google Scholar
10. Leuchs S, et al. Inactivation and inducible oncogenic mutation of p53 in gene targeted pigs. PLoS One. 2012;7(10):e43323.
    View this article via: PubMed Google Scholar
11. Levine AJ, Oren M. The first 30 years of p53: growing ever more complex. Nat Rev Cancer. 2009;9(10):749–758.
    View this article via: PubMed Google Scholar
12. Vousden KH, Prives C. Blinded by the Light: The growing complexity of p53. Cell. 2009;137(3):413–431.
    View this article via: PubMed Google Scholar
13. Vogelstein B, Lane D, Levine AJ. Surfing the p53 network. Nature. 2000;408(6810):307–310.
    View this article via: PubMed CrossRef Google Scholar
14. Lowe SW, Sherr CJ. Tumor suppression by Ink4a-Arf: progress and puzzles. Curr Opin Genet Dev. 2003;13(1):77–83.
    View this article via: PubMed Google Scholar
15. Nichols KE, Malkin D, Garber JE, Fraumeni JF, Fraumeni JF Jr, Li FP. Germ-line p53 mutations predispose to a wide spectrum of early-onset cancers. Cancer Epidemiol Biomarkers Prev. 2001;10(2):83–87.
    View this article via: PubMed Google Scholar
16. 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
17. Freed-Pastor WA, Prives C. Mutant p53: one name, many proteins. Genes Dev. 2012;26(12):1268–1286.
    View this article via: PubMed Google Scholar
18. 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
19. Kuperwasser C, et al. Development of spontaneous mammary tumors in BALB/c p53 heterozygous mice. A model for Li-Fraumeni syndrome. Am J Pathol. 2000;157(6):2151–2159.
    View this article via: PubMed CrossRef Google Scholar
20. 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
21. Hill KA, et al. Most spontaneous tumors in a mouse model of Li-Fraumeni syndrome do not have a mutator phenotype. Carcinogenesis. 2006;27(9):1860–1866.
    View this article via: PubMed CrossRef Google Scholar
22. Rogers CS, et al. Disruption of the CFTR gene produces a model of cystic fibrosis in newborn pigs. Science. 2008;321(5897):1837–1841.
    View this article via: PubMed Google Scholar
23. Walker SC, et al. A highly efficient method for porcine cloning by nuclear transfer using in vitro-matured oocytes. Cloning Stem Cells. 2002;4(2):105–112.
    View this article via: PubMed Google Scholar
24. Wu X, Bayle JH, Olson D, Levine AJ. The p53-mdm-2 autoregulatory feedback loop. Genes Dev. 1993;7(7A):1126–1132.
    View this article via: PubMed CrossRef Google Scholar
25. Senturk E, Manfredi JJ. Mdm2 and tumorigenesis: evolving theories and unsolved mysteries. Genes Cancer. 2012;3(3-4):192–198.
    View this article via: PubMed CrossRef Google Scholar
26. Lane DP. p53, guardian of the genome. Nature. 1992;358(6381):15–16.
    View this article via: PubMed CrossRef Google Scholar
27. Dittmer D, et al. Gain of function mutations in p53. Nat Genet. 1993;4(1):42–46.
    View this article via: PubMed CrossRef Google Scholar
28. Serrano M, Lin AW, McCurrach ME, Beach D, Lowe SW. Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a. Cell. 1997;88(5):593–602.
    View this article via: PubMed Google Scholar
29. Hussain SP, Schwank J, Staib F, Wang XW, Harris CC. TP53 mutations and hepatocellular carcinoma: insights into the etiology and pathogenesis of liver cancer. Oncogene. 2007;26(15):2166–2176.
    View this article via: PubMed CrossRef Google Scholar
30. Lee WK, Lau EW, Duddalwar VA, Stanley AJ, Ho YY. Abdominal manifestations of extranodal lymphoma: spectrum of imaging findings. AJR. AJR Am J Roentgenol. 2008;191(1):198–206.
    View this article via: PubMed CrossRef Google Scholar
31. Leite NP, et al. Cross-sectional imaging of extranodal involvement in abdominopelvic lymphoproliferative malignancies. Radiographics. 2007;27(6):1613–1634.
    View this article via: PubMed CrossRef Google Scholar
32. Treuting PM, Albertson TM, Preston BD. Case series: acute tumor lysis syndrome in mutator mice with disseminated lymphoblastic lymphoma. Toxicol Pathol. 2010;38(3):476–485.
    View this article via: PubMed Google Scholar
33. Vogel P, Pletcher JM, Liang Y. Spontaneous acute tumor lysis syndrome as a cause of early deaths in short-term carcinogenicity studies using p53^+/– mice. Vet Pathol. 2010;47(4):719–724.
    View this article via: PubMed CrossRef Google Scholar
34. McBride A, Westervelt P. Recognizing and managing the expanded risk of tumor lysis syndrome in hematologic and solid malignancies. J Hematol Oncol. 2012;5:75.
    View this article via: PubMed CrossRef Google Scholar
35. Schifter T, Cohen A, Lewinski UH. Severe tumor lysis syndrome following splenic irradiation. Am J Hematol. 1999;60(1):75–76.
    View this article via: PubMed Google Scholar
36. Gylys-Morin V, Hoffer F, Kozakewich H, Shamberger R. Wilms tumor and nephroblastomatosis: imaging characteristics at gadolinium-enhanced MR imaging. Radiology. 1993;188(2):517–521.
    View this article via: PubMed Google Scholar
37. Stanhope-Baker P, Kessler PM, Li W, Agarwal ML, Williams BR. The Wilms tumor suppressor-1 target gene podocalyxin is transcriptionally repressed by p53. J Biol Chem. 2004;279(32):33575–33585.
    View this article via: PubMed CrossRef Google Scholar
38. Ordonez N, Rosai J. Urinary tract. In: Rosai J, ed. Rosai and Ackeman’s Surgical Pathology. New York, New York, USA: Mosby; 2004.
39. Brachova P, Thiel KW, Leslie KK. The consequence of oncomorphic TP53 mutations in ovarian cancer. Int J Mol Sci. 2013;14(9):19257–19275.
    View this article via: PubMed CrossRef Google Scholar
40. Ciriello G, Miller ML, Aksoy BA, Senbabaoglu Y, Schultz N, Sander C. Emerging landscape of oncogenic signatures across human cancers. Nat Genet. 2013;45(10):1127–1133.
    View this article via: PubMed CrossRef Google Scholar
41. Coussens LM, Werb Z. Inflammation and cancer. Nature. 2002;420(6917):860–867.
    View this article via: PubMed Google Scholar
42. Kalluri R, Zeisberg M. Fibroblasts in cancer. Nat Rev Cancer. 2006;6(5):392–401.
    View this article via: PubMed CrossRef Google Scholar
43. Sieren JC, et al. An automated segmentation approach for highlighting the histological complexity of human lung cancer. Ann Biomed Eng. 2010;38(12):3581–3591.
    View this article via: PubMed Google Scholar
44. Shiao SL, Ganesan AP, Rugo HS, Coussens LM. Immune microenvironments in solid tumors: new targets for therapy. Genes Dev. 2011;25(24):2559–2572.
    View this article via: PubMed CrossRef Google Scholar
45. Frangioni JV. New technologies for human cancer imaging. J Clin Oncol. 2008;26(24):4012–4021.
    View this article via: PubMed Google Scholar
46. Simons D, Kachelriess M, Schlemmer HP. Recent developments of dual-energy CT in oncology. Eur Radiol. 2014;24(4):930–939.
    View this article via: PubMed CrossRef Google Scholar
47. Sharma R, Aboagye E. Development of radiotracers for oncology — the interface with pharmacology. Br J Pharmacol. 2011;163(8):1565–1585.
    View this article via: PubMed Google Scholar
48. Shiga T, et al. A new PET scanner with semiconductor detectors enables better identification of intratumoral inhomogeneity. J Nucl Med. 2009;50(1):148–155.
    View this article via: PubMed Google Scholar
49. Weissleder R. Scaling down imaging: molecular mapping of cancer in mice. Nat Rev Cancer. 2002;2(1):11–18.
    View this article via: PubMed CrossRef Google Scholar
50. Bowen SR, et al. Challenges and opportunities in patient-specific, motion-managed and PET/CT-guided radiation therapy of lung cancer: review and perspective. Clin Transl Med. 2012;1(1):18.
    View this article via: PubMed Google Scholar
51. Birch JM, et al. Prevalence and diversity of constitutional mutations in the p53 gene among 21 Li-Fraumeni families. Cancer Res. 1994;54(5):1298–1304.
    View this article via: PubMed Google Scholar
52. Kleihues P, Schauble B, zur Hausen A, Esteve J, Ohgaki H. Tumors associated with p53 germline mutations: a synopsis of 91 families. Am J Pathol. 1997;150(1):1–13.
    View this article via: PubMed Google Scholar
53. Porter DE, Holden ST, Steel CM, Cohen BB, Wallace MR, Reid R. A significant proportion of patients with osteosarcoma may belong to Li-Fraumeni cancer families. J Bone Joint Surg Br. 1992;74(6):883–886.
    View this article via: PubMed Google Scholar
54. Jacobs R, Messick J, VAlli V. Tumors of the hematopoietic system. In: Tumors In Domestic Animals. Ames, Iowa, USA: Iowa State Press; 2002:159–160.
55. Migaki G, Nelson LW, Todd GC. Prevalence of embryonal nephroma in slaughtered swine. J Am Vet Med Assoc. 1971;159(4):441–442.
56. Miller CW, et al. Frequency and structure of p53 rearrangements in human osteosarcoma. Cancer Res. 1990;50(24):7950–7954.
    View this article via: PubMed Google Scholar
57. Tsuchiya T, Sekine K, Hinohara S, Namiki T, Nobori T, Kaneko Y. Analysis of the p16INK4, p14ARF, p15, TP53, and MDM2 genes and their prognostic implications in osteosarcoma and Ewing sarcoma. Cancer Genet Cytogenet. 2000;120(2):91–98.
    View this article via: PubMed Google Scholar
58. Wunder JS, et al. TP53 mutations and outcome in osteosarcoma: a prospective, multicenter study. J Clin Oncol. 2005;23(7):1483–1490.
    View this article via: PubMed Google Scholar
59. Luetke A, Meyers PA, Lewis I, Juergens H. Osteosarcoma treatment — where do we stand? A state of the art review. Cancer Treat Rev. 2014;40(4):523–532.
    View this article via: PubMed Google Scholar
60. Radig K, Schneider-Stock R, Mittler U, Neumann HW, Roessner A. Genetic instability in osteoblastic tumors of the skeletal system. Pathol Res Pract. 1998;194(10):669–677.
    View this article via: PubMed CrossRef Google Scholar
61. Dal Cin P, Sciot R, Samson I, De Wever I, Van den Berghe H. Osteoid osteoma and osteoblastoma with clonal chromosome changes. Br J Cancer. 1998;78(3):344–348.
    View this article via: PubMed Google Scholar
62. Oliveira CR, et al. Classical osteoblastoma, atypical osteoblastoma, and osteosarcoma: a comparative study based on clinical, histological, and biological parameters. Clinics. 2007;62(2):167–174.
    View this article via: PubMed CrossRef Google Scholar
63. Unni K. Dahlin’s Bone Tumors: General Aspects and Data on 11,087 Cases. Philadelphia, Pennsylvania, USA: Lippencout-Raven; 1996.
64. Walkley CR, et al. Conditional mouse osteosarcoma, dependent on p53 loss and potentiated by loss of Rb, mimics the human disease. Genes Dev. 2008;22(12):1662–1676.
    View this article via: PubMed Google Scholar
65. Hingorani SR, et al. Trp53R172H and KrasG12D cooperate to promote chromosomal instability and widely metastatic pancreatic ductal adenocarcinoma in mice. Cancer Cell. 2005;7(5):469–483.
    View this article via: PubMed CrossRef Google Scholar
66. Jackson EL, et al. The differential effects of mutant p53 alleles on advanced murine lung cancer. Cancer Res. 2005;65(22):10280–10288.
    View this article via: PubMed CrossRef Google Scholar
67. Lai L, Prather RS. Production of cloned pigs by using somatic cells as donors. Cloning Stem Cells. 2003;5(4):233–241.
    View this article via: PubMed Google Scholar
68. Pomp D, Good BA, Geisert RD, Corbin CJ, Conley AJ. Sex identification in mammals with polymerase chain reaction and its use to examine sex effects on diameter of day-10 or -11 pig embryos. J Anim Sci. 1995;73(5):1408–1415.
    View this article via: PubMed Google Scholar
69. McCreath KJ, Howcroft J, Campbell KHS, Colman A, Schnieke AE, Kind AJ. Production of gene-targeted sheep by nuclear transfer from cultured somatic cells. Nature. 2000;405(6790):1066–1069.
    View this article via: PubMed Google Scholar
70. Muniz VP, et al. RABL6A promotes oxaliplatin resistance in tumor cells and is a new marker of survival for resected pancreatic ductal adenocarcinoma patients. Genes Cancer. 2013;4(7–8):273–284.
    View this article via: PubMed Google Scholar
71. Modestou M, Puig-Antich V, Korgaonkar C, Eapen A, Quelle DE. The alternative reading frame tumor suppressor inhibits growth through p21-dependent and p21-independent pathways. Cancer Res. 2001;61(7):3145–3150.
    View this article via: PubMed Google Scholar
72. Huang J, et al. RGS6 suppresses Ras-induced cellular transformation by facilitating Tip60-mediated Dnmt1 degradation and promoting apoptosis. [published online ahead of print September 2, 2013]. Oncogene. doi:10.1038/onc.2013.324.
    View this article via: PubMed CrossRef Google Scholar
73. Rokudai S, Laptenko O, Arnal SM, Taya Y, Kitabayashi I, Prives C. MOZ increases p53 acetylation and premature senescence through its complex formation with PML. Proc Natl Acad Sci U S A. 2013;110(10):3895–3900.
    View this article via: PubMed Google Scholar
74. Adam SJ, Rund LA, Kuzmuk KN, Zachary JF, Schook LB, Counter CM. Genetic induction of tumorigenesis in swine. Oncogene. 2007;26(7):1038–1045.
    View this article via: PubMed CrossRef Google Scholar
75. Pfaffl MW. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res. 2001;29(9):e45.
    View this article via: PubMed CrossRef Google Scholar
76. Rosset A, Spadola L, Ratib O. OsiriX: an open-source software for navigating in multidimensional DICOM images. J Digit Imaging. 2004;17(3):205–216.
    View this article via: PubMed Google Scholar
77. Eisenhauer EA, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer. 2009;45(2):228–247.
    View this article via: PubMed Google Scholar
78. Guo J, Fuld MK, Alford SK, Reinhardt JM, Hoffman EA. Pulmonary Analysis Software Suite 9.0: Integrating quantitative measures of function with structural analyses. In: Brown M, et al., eds. 1th International Conference on Medical Image Computing and Computer Assisted Intervention, Pulmonary Image Analysis Workshop. New York, New York, USA: CRC Press; 2008:283–292.