We have crossed mice carrying the conditional Rbtm2Brn allele with a constitutive Villin-Cre transgenic mouse. The Villin promoter in these animals is highly expressed in adult intestine and kidney proximal tubules and is expressed in the gut and nephros anlagen during embryogenesis. We report here that these mice develop tumors between 12 and 17 months old outside the gastrointestinal (GI) tract. A high penetrance of pituitary tumors and medullar carcinoma of the thyroid is observed with a lower incidence of hyperplasia of pulmonary neuroendocrine cells and aggressive liver, bile duct, stomach, oral cavity tumors, and lipomas. Rb rearrangement due to ectopic Villin promoter activity in neural crest or neural crest stem cells during embryogenesis is most likely responsible for the medullar carcinoma of the thyroid phenotype. The aggressive nature of the medullar carcinoma of the thyroid and its ability to metastasize to unusual sites make the model suitable for the study of tumor progression and mechanism of metastasis. Observed sites of metastasis include the stomach, small intestine, liver, lung, kidney, pancreas, spleen, bone marrow, salivary gland, fat, lymph nodes, and dorsal root ganglion. Because the Villin promoter is highly active throughout the GI and in the nephros anlagen during development, we find that Rb inactivation is not sufficient to initiate tumorigenesis in the GI or kidneys in mice. (Cancer Res 2006; 66(7): 3576-83)

The Rb tumor suppressor gene has long been known to control cell cycle progression, and the loss of its activity has been identified as a common event in the development of multiple human cancers (1, 2). The protein and its family members play important roles in cell cycle regulation through interaction with the E2F family of transcription factors (3). pRb can bind to and inhibit the transactivation functions of E2F. It can also alter chromatin structure through its interaction with E2F and other factors, decreasing the transcriptional program indirectly through chromatin remodeling (4). pRb exists in the hypophosphorylated state in G1 complexed to E2F. It is phosphorylated by cyclin D/cyclin-dependent kinase (cdk) 4 and cyclin E/cdk2 at the G1-S transition, resulting in decreased affinity for E2Fs and subsequent dissociation of transcriptional repression complexes (1).

In addition to cell cycle progression, Rb is now thought to be involved in many cellular processes, including irreversible exit from the cell cycle, protection from apoptosis, induction of cell type–specific gene expression, maintenance of the postmitotic state, and orderly progression through differentiation (5). Knockout mice for Rb provide direct evidence that the classic tumor suppressor protein plays a role in terminal differentiation of multiple tissue types. Rb−/− mice die during E13 to E15 days of gestation and do not develop erythroid, neuronal, and lens tissue to their fully differentiated state (6, 7). Heterozygous mice succumb to pituitary tumors and medullary carcinomas of the thyroid earlier on a mixed genetic background than on C57Bl6 background (8). Chimeric mice, designed to rescue embryonic lethality of the Rb−/− mice, develop a complex phenotype, such as pituitary tumors, medullar carcinoma of the thyroid (9, 10), and adrenal pheochromocytomas (11, 12).

The Rb pathway is implicated in the initiation of several tumor types. Its role, if any, in the progression of colorectal cancer is not defined. Mutations in Rb are not found in human gastrointestinal tumors. Rb mouse models do not develop colorectal cancer. To examine the effects of Rb as a modifier in colorectal cancer, we crossed the conditional Rb model Rbtm2Brn−/− (13) with a Villin-Cre transgenic mouse [Tg(Vil-Cre); ref. 14] before the addition of an adenomatous polyposis coli (Apc) mutation that is known to initiate gastrointestinal tumorigenesis. Villin is an actin-binding protein that contributes to the assembly of the microvillus bundle (15). In adult mice, this gene is expressed specifically in the intestinal epithelium and kidney proximal tubules. The specificity of the promoter in adult tissues has led to its use in Cre recombinase transgenic mouse models (16). Cis-acting sequences necessary to drive the Villin promoter have been studied by several groups (1719). The model used in these studies has a 9-kb fragment that drives Cre gene expression. Embryonic expression of Villin is found in primitive endoderm, gut, nephros anlagen and its derivatives in the developing embryo, and extraembryonic visceral endoderm of the yolk sac (14). It is not known to be active in nonepithelial cells at any stage of development (20, 21).

We report here that Rbtm2Brn−/−Tg(Vil-Cre) mice have a significant decrease in survival rate and succumb to an unusually malignant and metastatic tumor phenotype at age ∼1 year. All mice observed developed pituitary tumors. Ninety percent of observed Rbtm2Brn−/−Tg(Vil-Cre) mice developed medullar carcinoma of the thyroid concurrent with pituitary tumors. Tumors from the thyroid have rearranged Rb and metastasized to many sites, including the stomach, small intestine, salivary glands, liver, lung, kidney, pancreas, spleen, bone marrow, fat, lymph nodes, dorsal root ganglion. More than half of the animals developed hyperplasia of pulmonary neuroendocrine cells. There was a low penetrance of adenocarcinoma of the lung derived from type 2 pneumocytes.

None of the animals had primary intestinal tumors, although some had stomach hyperplasia, and one had an unusual stomach tumor arising from a secretory gland. A mild predisposition for liver tumors was found, including adenocarcinoma of the bile duct and hepatoma with bile duct hyperplasia. Steatitis was observed with low frequency in the abdominal fat and in one case with a lipoma. Heterozygous animals also exhibited the tumorigenic phenotype, although the median survival of these animals and their wild-type siblings was not significantly different. Molecular analysis revealed some rearrangement of the floxed allele in almost all tissues tested and corresponding decrease in wild-type Rb RNA and protein levels in intestinal tissue.

Generation of conditional Rb Villin-Cre mice. FVB;129-Rb-1tm2Brn containing a conditional mutation in the endogenous Rb gene was obtained from the Mouse Repository of the Mouse Models of Human Cancers Consortium (MMHCC; National Cancer Institute-Frederick, Frederick, MD). These mice had lox P in the introns flanking exon 19 (correlating to exon 20 of National Center for Biotechnology Information Ensemble sequence ENSMUST00000022701) in the Rb locus (13). The mice were of a mixed genetic background from FVB and 129. These animals were mated with B6;D2-Tg(Vil-Cre) 20Syr, also obtained from the MMHCC, which have Cre expressed under the control of a 9-kb regulatory region of the murine Villin promoter (14). Tg(Vil-Cre) mice were also in a mixed genetic background of C57Bl6 and D2. Pairs of mice heterozygous for the floxed Rb allele and positive for Villin-Cre were intercrossed to obtain offspring that were homozygous for the floxed allele and contained the Villin-Cre transgene. All animals were maintained in an Association for Assessment and Accreditation of Laboratory Animal Care–approved facility under barrier conditions.

Rbtm2Brn Tg(Vil-Cre) genotyping and identification of the rearranged allele. Offspring were genotyped routinely by PCR amplification as described previously (13). DNA was made from tail clippings of animals at 10 days old using DNeasy kits (Qiagen, Valencia, CA). Primers used for amplification were Rb212(149A) (5′-GAAAGGAAAGTCAGGGAATTGGG-3′), Rb18(149C) (5′-GGCGTGTGCCATCAATG-3′), and Rb19E(149B) (5′-CTCAAGAGCTCAGACTCATGG-3′). Reactions were carried out in a total of 25 μL. Reactions consisted of 0.05 units/μL Taq Gold Polymerase, 1× Taq Gold Buffer, 1.5 mmol/L MgCl2, 0.2 mmol/L deoxynucleotide triphosphate (dNTP), 0.4 μmol/L of each primer, and 25 ng DNA substrate. The PCR cycling profile consisted of 1 cycle at 94°C for 9 minutes; 35 cycles at 94°C for 30 seconds, 58°C for 30 seconds, and 72°C for 50 seconds; and 1 cycle at 58°C for 5 minutes. PCR products were analyzed on 3% agarose gels. Primers Rb19E(149B) and Rb18(149C) yield a 283-bp fragment that represents the unrecombined RbLOXP allele and a 235-bp fragment that represents the wild-type Rb allele. Primers Rb212(149A) and Rb18(149C) yield a 260-bp fragment from the recombined RbΔ19 allele and a 600-bp fragment from the wild-type Rb allele. They were also used to detect rearrangement in the DNA from various tissues and tumors made using DNeasy kits.

RNA isolation and reverse transcription-PCR. RNA from tissue samples was isolated using a RiboPure kit (Ambion, Austin, TX) according to kit instructions. Tissue samples were snap frozen in liquid nitrogen at the time of dissection and stored in a cryogenic freezer. RNA was quantitated by 260 per 280 readings from a spectrophotometer (NanoDrop, Wilmington, DE). Reverse transcription-PCR (RT-PCR) was done using the SuperScript One-Step RT-PCR with Platinum Taq System (Invitrogen, Carlsbad, CA). Actin primers (5′-GGTCACCCACACTGTGCCCATCTACG-3′ and 5′-GGATGCCACAGGACTCCATGCCCAG-3′; 0.2 μmol/L final concentration) were used for an internal control, and Rb primers (exon 18, 5′-CCTTGAACCTGCTTGTCCTC-3′ and exon 20, 5′-GAAGGCGTGCACAGAGTGTA-3′; 0.2 μmol/L final concentration) were used to amplify from cDNA. Reactions were set up as suggested in the kit insert (1× Reaction Mix, 10 ng/μL RNA substrate, 0.4 units/μL Ambion's SUPERase In, 1 μL reverse transcriptase/Platinum Taq Mix; 50 μL total volume). PCR cycling conditions were 1 cycle at 50°C for 30 minutes and 94°C for 2 minutes; 35 cycles at 94°C for 15 seconds, 56°C for 30 seconds, and 72°C for 30 seconds; and 1 cycle at 72°C for 5 minutes and at 4°C.

Identification of the Villin-Cre transgene. The presence of the Villin-Cre transgene was identified by PCR amplification with specific primers for the Villin promoter (163F, 5′-GTGTGGGACAGAGAACAAACCG-3′) and Cre recombinase gene (163R, 5′-TGCGAACCTCATCACTCGTTGC-3′). Reactions were carried out in 25 μL using 0.05 units/μL Taq Gold Polymerase, 1× Taq Gold Buffer, 1.5 mmol/L MgCl2, 0.2 mmol/L dNTP, 0.4 μmol/L primer, and 25 ng DNA substrate. The PCR cycling profile consisted of 1 cycle at 94°C for 9 minutes; 35 cycles at 94°C for 1 minute, 58°C for 45 seconds, and 72°C for 30 seconds; and 1 cycle at 58°C for 5 minutes. A diagnostic fragment of ∼900 bp obtained from the DNA of a founder mouse was isolated and eluted from a 1% agarose gel using a QIAquick gel extraction kit and sequenced to identify definitively the presence of Villin sequences adjacent to Cre recombinase.

Generation of survival curves. Birth and death dates for animals were recorded routinely. Prism software by GraphPad was used to generate survival curves. Villin-Cre-positive animals were observed as one cohort, and Villin-Cre-negative animals were observed as a second cohort for purposes of comparison.

Histopathologic analysis. Animals were euthanized by CO2 asphyxiation with subsequent cervical dislocation. Tissues were inspected on gross dissection and fixed in Bouin's fixative solution (RICCA Chemical Co., Arlington, TX). Organ samples were removed, embedded in paraffin blocks, and sectioned (5 μm) serially for histopathologic analysis. Both normal and neoplastic materials were stained with H&E and examined to determine the presence of tumor material and degree of malignancy in all tissues. Some tissues were snap frozen at the time of dissection in liquid nitrogen for RNA and DNA analysis. In sections of small intestines that were used for DNA analysis, muscularis and mucosa were present.

pRb and calcitonin detection by immunohistochemistry. Rb (M-15):sc-1538, a polyclonal antibody to a peptide mapping at the COOH terminus of murine pRb, was purchased (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). Sections of intestine from Rbtm2Brn−/−Tg(Vil-Cre) and C57Bl6 were baked at 60°C for 1 hour. Paraffin was removed from sections by four washes of xylene (4 minutes each). After hydrating through 100%, 95%, and 75% ethanol (5 minutes each) and rinsing with distilled water, endogenous peroxide was quenched with 3% H2O2 in methanol for 20 minutes. Antigen retrieval was accomplished by boiling sections in sodium citrate buffer [10 mmol/L (pH 6.0)] for 25 minutes. Slides were washed in PBS for 10 minutes, and nonspecific antigen sites were blocked using 1% normal horse serum in PBS for 30 minutes. Primary antibody Rb (M-15):sc-1538 was applied to slides (1:50, 1:100, and 1:250 dilutions) and incubated at room temperature overnight in a humidity chamber. Affinity-purified biotinylated anti-goat IgG (H+L) (Vector Laboratories, Inc., Burlingame, CA) was used as secondary antibody (1:75 dilution) and applied for 1 hour at room temperature. A Vectastain Elite ABC kit (Vector Laboratories) was used as recommended in the package insert for antibody detection. The chromogen used for visualization was 3,3′-diaminobenzidine tetrahydrochloride. Mayer's hematoxylin solution (Sigma-Aldrich, St. Louis, MO) was used for counterstaining. Slides were coverslipped with Permount mounting solution (Fisher, Fairlawn, NJ).

Calcitonin (G-18):sc-7784, an affinity-purified goat polyclonal antibody raised against a peptide mapping at the NH2 terminus of calcitonin of human origin and cross-reactive with mouse calcitonin, was purchased (Santa Cruz Biotechnology). Tissue sections containing C-cell metastases were stained for calcitonin by the method described above for Rb immunohistochemistry.

Rbtm2Brn−/−Tg(Vil-Cre) animals have decreased survival. We generated a cohort of mice (n = 124) by intercrossing Rbtm2Brn+/−Tg(Vil-Cre)-positive animals. Genotyping was done by PCR amplification as described above (Fig. 1A). The altered Rb allele was described to be functionally null (13, 22). To confirm that our PCR-based test system was accurate, we sequenced the fragment amplified from Rbtm2Brn−/− tail DNA using primers designated Rb212(149A) and Rb18(149C). The results confirmed the presence of the two lox P sites flanking exon 19. The Rbtm2Brn allele was segregated in a Mendelian fashion, the median survival time for Rbtm2Brn−/−Tg(Vil-Cre) was 12.5 months, and all animals expired at 17 months old (Fig. 1B). The survival curve for the Rbtm2Brn−/−Tg(Vil-Cre)-positive mice differed significantly from Rbtm2Brn+/−Tg(Vil-Cre)-positive and wild-type mice (P < 0.0001). Survival curves were also generated for all Rbtm2Brn Tg(Vil-Cre)-negative siblings (data not shown). As expected, this cohort (n = 134) showed no significant difference (P = 0.533) in median survival between siblings genotyped as homozygous, heterozygous, or wild-type floxed Rb. A comparison of median survival between Tg(Vil-Cre)-positive and Tg(Vil-Cre)-negative animals also did not show any significant difference (P = 0.970).

Figure 1.

Molecular characterization and Kaplan-Meier survival plot of Rbtm2Brn Tg(Vil-Cre) mice. A, structures of Rbtm2Brn and Villin-Cre loci. Rbtm2Brn allele is detected by PCR amplification of a 283-bp fragment using primers Rb19(149B) and Rb18(149C). The wild-type allele (no lox P site) is detected by amplification of a 226-bp fragment from primers Rb19(149B) and Rb18(149C). Villin-Cre transgene is detected by PCR amplification of a 900-bp fragment that spans the promoter and recombinase. B, Kaplan-Meier analysis of Rbtm2Brn Tg(Vil-Cre) mice. All animals carry the Villin-Cre transgene. Black, wild-type Rb; orange, heterozygous RbΔ19; red, homozygous RbΔ19. C, detection of Villin-Cre-mediated rearrangement of the Rbtm2Brn allele. Primers Rb212(149A) and Rb18(149C) PCR amplify a 646-bp fragment from the Rbtm2Brn allele and a 260-bp fragment from the RbΔ19 allele. Representative tissues indicate varying amounts of rearrangement. Sal. Gln., salivary gland; Liv. Tumor, liver tumor. D, aberrant Rb expression detected by RT-PCR in tissues from Rbtm2Brn−/−Tg(Vil-Cre) mice. a, liver; b, kidney; c, small intestine; d, liver tumor; e, lung. For each tissue, lane 1 of RT-PCR is made using actin primers alone (350-bp fragment), lane 2 is made using multiplexed actin and Rb primers, and lane 3 is made using Rb primers alone. Amplification from the RbΔ19 allele yields a 108-bp fragment.

Figure 1.

Molecular characterization and Kaplan-Meier survival plot of Rbtm2Brn Tg(Vil-Cre) mice. A, structures of Rbtm2Brn and Villin-Cre loci. Rbtm2Brn allele is detected by PCR amplification of a 283-bp fragment using primers Rb19(149B) and Rb18(149C). The wild-type allele (no lox P site) is detected by amplification of a 226-bp fragment from primers Rb19(149B) and Rb18(149C). Villin-Cre transgene is detected by PCR amplification of a 900-bp fragment that spans the promoter and recombinase. B, Kaplan-Meier analysis of Rbtm2Brn Tg(Vil-Cre) mice. All animals carry the Villin-Cre transgene. Black, wild-type Rb; orange, heterozygous RbΔ19; red, homozygous RbΔ19. C, detection of Villin-Cre-mediated rearrangement of the Rbtm2Brn allele. Primers Rb212(149A) and Rb18(149C) PCR amplify a 646-bp fragment from the Rbtm2Brn allele and a 260-bp fragment from the RbΔ19 allele. Representative tissues indicate varying amounts of rearrangement. Sal. Gln., salivary gland; Liv. Tumor, liver tumor. D, aberrant Rb expression detected by RT-PCR in tissues from Rbtm2Brn−/−Tg(Vil-Cre) mice. a, liver; b, kidney; c, small intestine; d, liver tumor; e, lung. For each tissue, lane 1 of RT-PCR is made using actin primers alone (350-bp fragment), lane 2 is made using multiplexed actin and Rb primers, and lane 3 is made using Rb primers alone. Amplification from the RbΔ19 allele yields a 108-bp fragment.

Close modal

Rbtm2Brn−/−Tg(Vil-Cre) animals develop aggressive tumors. All Rbtm2Brn−/−Tg(Vil-Cre) mice observed had tumors arising from the intermediate lobe of the pituitary (penetrance, 100%; Table 1). Fifty-six percent of the tumors were highly aggressive invading the bone or hypothalamus (Table 2; Fig. 2A). Forty-four percent were adenomas. Pituitary tumors have been described previously for other Rb mouse models (6, 2325).

Table 1.

Frequency of major neoplastic events in Rbtm2Brn Tg(Vil-Cre) mice

GenotypeAge (mo.)Pituitary tumors, % (n)Medullary carcinoma of thyroid, % (n)Pulmonary neuroendocrine hyperplasia, % (n)Sarcoma, % (n)Metastasis, % (n)Stomach hyperplasia, % (n)
Rb−/− 12.6 100 (10) 90 (10) 58 (12) 0 (10) 50 (12) 30 (10) 
Rb+/− 12.8 43 (7) 50 (8) 14 (7) 0 (7) 29 (7) 0 (7) 
Rb+/+ 12.3 0 (2) 0 (2) 0 (2) 50 (2) 50 (2) 0 (2) 
GenotypeAge (mo.)Pituitary tumors, % (n)Medullary carcinoma of thyroid, % (n)Pulmonary neuroendocrine hyperplasia, % (n)Sarcoma, % (n)Metastasis, % (n)Stomach hyperplasia, % (n)
Rb−/− 12.6 100 (10) 90 (10) 58 (12) 0 (10) 50 (12) 30 (10) 
Rb+/− 12.8 43 (7) 50 (8) 14 (7) 0 (7) 29 (7) 0 (7) 
Rb+/+ 12.3 0 (2) 0 (2) 0 (2) 50 (2) 50 (2) 0 (2) 

NOTE: n = sample size.

Table 2.

Summary of histopathology for Rbtm2Brn−/−Tg(Vil-Cre) mice

ID no.SexPituitary tumors*Medullary carcinomaPulmonary neuroendocrine hyperplasiaSite of metastasisOther pathology
no. no. no. no. Adenocarcinoma of bile duct cystic ovary, uterine abnormality 
Adenoma (+) Blood vessel emboli Squamous cell carcinoma in oral cavity 
no. no. − Stomach hyperplasia 
no. −  
Adenocarcinoma (+) Spleen, kidney, pancreas, stomach, fat, nerve, liver, lung, bone marrow, small intestine Stomach hyperplasia, islet cell hyperplasia 
Adenocarcinoma − Stomach hyperplasia 
Adenoma − (+) lung Lipoma with steatitus 
Adenoma − (+) liver, lung Steatitis in abdominal fat 
Adenocarcinoma −  
10 Adenocarcinoma − − −  
11 no. no. (Adenocarcinoma) (+) Liver, lung, stomach Hepatoma, bile duct hyperplasia 
12 Adenocarcinoma − − Stomach tumor 
13 Adenoma − (+) Salivary gland Focal hyperplasia of large intestine with focal colitis 
ID no.SexPituitary tumors*Medullary carcinomaPulmonary neuroendocrine hyperplasiaSite of metastasisOther pathology
no. no. no. no. Adenocarcinoma of bile duct cystic ovary, uterine abnormality 
Adenoma (+) Blood vessel emboli Squamous cell carcinoma in oral cavity 
no. no. − Stomach hyperplasia 
no. −  
Adenocarcinoma (+) Spleen, kidney, pancreas, stomach, fat, nerve, liver, lung, bone marrow, small intestine Stomach hyperplasia, islet cell hyperplasia 
Adenocarcinoma − Stomach hyperplasia 
Adenoma − (+) lung Lipoma with steatitus 
Adenoma − (+) liver, lung Steatitis in abdominal fat 
Adenocarcinoma −  
10 Adenocarcinoma − − −  
11 no. no. (Adenocarcinoma) (+) Liver, lung, stomach Hepatoma, bile duct hyperplasia 
12 Adenocarcinoma − − Stomach tumor 
13 Adenoma − (+) Salivary gland Focal hyperplasia of large intestine with focal colitis 

Abbreviation: no., not observed (organs are not evaluated pathologically).

*

Pituitary adenocarcinoma had invaded the bone or hypothalmus.

Figure 2.

Histologic analysis of Rbtm2Brn−/−Tg(Vil-Cre) mice show aggressive tumors of the pituitary and C-cell tumors of the thyroid with metastases. A, 1, section of pituitary stained with H&E showing an adenocarcinoma arising from the intermediate lobe of the pituitary with basophilic tumor cells invading into the hypothalamus (above and left of the pituitary) and the bone (below the pituitary). Magnification, ×40. 2, same H&E section. Magnification, ×100. Arrow, adenocarcinoma invading into the bone (b). Bottom right, vertical structure is growth plate cartilage. B, 1, H&E sections of C-cell carcinoma of the thyroid with metastases. Lobules of the thyroid gland are filled with tumor cells of uniform size and small slightly elliptic nuclei with scanty cytoplasm. 2, dorsal root ganglion containing four clusters of metastatic C cells (dark-stained region). Schwann and satellite cells are interspersed throughout the ganglion. 3, clusters of C cells in the mucosa of the stomach. 4, C-cell metastasis to adipose tissue. C, 1, immunohistochemical analysis of calcitonin expression in a C-cell metastasis to the stomach. Stomach section observed for calcitonin immunohistochemistry is from the same stomach that was used for metastasis identification in H&E staining in Fig. 2C,2, immunohistochemistry without primary antibody as negative control.

Figure 2.

Histologic analysis of Rbtm2Brn−/−Tg(Vil-Cre) mice show aggressive tumors of the pituitary and C-cell tumors of the thyroid with metastases. A, 1, section of pituitary stained with H&E showing an adenocarcinoma arising from the intermediate lobe of the pituitary with basophilic tumor cells invading into the hypothalamus (above and left of the pituitary) and the bone (below the pituitary). Magnification, ×40. 2, same H&E section. Magnification, ×100. Arrow, adenocarcinoma invading into the bone (b). Bottom right, vertical structure is growth plate cartilage. B, 1, H&E sections of C-cell carcinoma of the thyroid with metastases. Lobules of the thyroid gland are filled with tumor cells of uniform size and small slightly elliptic nuclei with scanty cytoplasm. 2, dorsal root ganglion containing four clusters of metastatic C cells (dark-stained region). Schwann and satellite cells are interspersed throughout the ganglion. 3, clusters of C cells in the mucosa of the stomach. 4, C-cell metastasis to adipose tissue. C, 1, immunohistochemical analysis of calcitonin expression in a C-cell metastasis to the stomach. Stomach section observed for calcitonin immunohistochemistry is from the same stomach that was used for metastasis identification in H&E staining in Fig. 2C,2, immunohistochemistry without primary antibody as negative control.

Close modal

The thyroids of 10 Rbtm2Brn−/−Tg(Vil-Cre) mice were examined by gross and/or histologic examination. Nine of these mice were found to have large medullar carcinoma of the thyroids (up to 7 mm in diameter; penetrance, 90%). Medullar carcinomas of the thyroids are malignant tumors that show parafollicular cell (C-cell) differentiation with frequent increase in calcitonin production. Observed animals (∼50%) with medullar carcinoma of the thyroid had metastases to distant sites, including the stomach (Fig. 2B), small intestine, lung, liver, kidney, pancreas, spleen, fat, lymph nodes, salivary gland, dorsal root ganglion, and bone marrow. The estimate of animals with metastases at 50% is probably somewhat low. Metastases were identified on inspection of H&E-stained sections, and representative metastases [stomach (Fig. 2C), fat, and small intestine] were later confirmed to be calcitonin positive by immunohistochemical staining.

The lungs of 12 Rbtm2Brn−/−Tg(Vil-Cre)-positive animals were examined, and 7 were found to have hyperplasia of pulmonary neuroendocrine cells in addition to thyroid and pituitary tumors (penetrance, 58%). Four of 12 (33%) animals had C-cell metastasis to the lung (Table 2). One (8%) animal had an adenocarcinoma derived from type 2 pneumocytes.

Two Rbtm2Brn−/−Tg(Vil-Cre) mice had tumors involving the liver (penetrance, 17%; Table 2): one with an enlarged gall bladder that was proven to be an adenocarcinoma of the bile duct (1 cm in diameter) and the other with a hepatoma (with necrosis; 1 cm in diameter) with bile duct hyperplasia. Another two mice showed C-cell metastasis to the liver (17%). Two mice had steatitis in abdominal fat (17%), and one had a lipoma (penetrance, 8%).

Eight moribund Rbtm2Brn+/−Tg(Vil-Cre) animals were examined histologically (Table 1). Four exhibited aggressive medullar carcinoma of the thyroid similar to Rbtm2Brn−/−Tg(Vil-Cre) siblings. C-cell metastases to the liver, kidney, lung, fat, testis, and salivary gland were observed in mice (29%). Pituitary tumors were found in 3 of 7 (43%) animals. Primary neuroendocrine hyperplasia of the lung was found in heterozygotes (14%). Wild-type animals did not have pituitary tumors. One succumbed to an aggressive undifferentiated sarcoma with metastasis to the liver and a primary benign lung tumor. None of the Rbtm2Brn Tg(Vil-Cre) homozygous or heterozygous mice had intestinal tumors. One (8%) animal developed an unusual stomach tumor that seemed to arise from a gastric secretory gland. Hyperplasia of the stomach was observed in Rbtm2Brn Tg(Vil-Cre) homozygous mice (30%).

The conditional Rb allele is rearranged in Rbtm2Brn−/−Tg(Vil-Cre) mice. Tissue sections from eight mice with histologic data and one mouse without histologic data were analyzed on a molecular level for rearrangement of the conditional Rb allele. Efficient rearrangement was observed in stomach, small and large intestines, kidney, spleen, and tumor DNAs (Fig. 1C). Wild-type Rb allele was detected at low levels in addition to an unexpected PCR amplification product (∼489 bp). Abnormal PCR amplification products have been reported previously by other groups using the Rbtm2Brn mouse (26). Because they are not present in Southern analysis, they are thought to be artifacts.

Almost all tissues showed low levels of RbΔ19 rearrangement. Extra gastrointestinal samples tested included kidney, lung, thyroid, pituitary, liver, spleen, heart, fat, eyes, reproductive organs, salivary gland, brain, tumor tissue, and tail. DNA from lung, heart, eyes, reproductive organs, and salivary glands amplified preferentially the wild-type allele (estimate made on a visual basis). Five of six thyroids with large tumors had complete rearrangement of the RbΔ19 allele. One thyroid with a tumor had both normal and deleted alleles, and one thyroid had no rearrangement and tumor. Tested liver, fat, and brain tissues had approximately equivalent ratios of wild-type to rearranged allele. Two liver tumors had complete rearrangement of the RbΔ19 allele.

We also tested for rearrangement of the RbΔ19 allele in tail DNA sampled when the mice were 10 days old. Tail clippings from 15 Villin-Cre-positive mice were found to be wild-type, heterozygous, or homozygous for the Rb allele. Rb was rearranged in low levels in all tail DNA from recessive animals and to a lesser extent in tail DNA from all heterozygous animals (Fig. 1C). One of the homozygous animals seemed to have significantly more rearranged DNA than the other four mice. The RbΔ19 allele was not rearranged in tail DNA from wild-type siblings.

The Villin-Cre transgene is intact in Rbtm2Brn−/−Tg(Vil-Cre) mice. PCR data showed that the Villin-Cre construct is intact in the Rbtm2Brn−/−Tg(Vil-Cre) mice. To confirm these data, the fragment amplified using a forward primer specific for the Villin promoter and a reverse primer specific for Cre recombinase was gel purified and sequenced. Villin promoter and Cre recombinase sequences were identified from the fragment confirming the mouse contained the transgene.

RbΔ19 RNA detected in gastrointestinal, kidney, and tumor tissue. RNA isolated from tissues stored in liquid nitrogen was reverse transcribed. PCR fragments amplified using primers from exons 18 and 20 on cDNA substrate from a C57Bl6 mouse (exon numbering; ref. 13) yielded an appropriate size fragment for all observed tissues. That fragment was absent in RT-PCR from a Rbtm2Brn−/−Tg(Vil-Cre)-positive mouse in intestinal and liver tumor tissues (Fig. 1D) and in its place was a 108-bp fragment of predicted size from the RbΔ19 allele. Liver and kidney tissues judged normal on gross inspection and amplified normal PCR fragments. Normal lung tissue from the same animal yielded an appropriate size fragment for wild-type allele.

Rbtm2Brn−/−Tg(Vil-Cre) mice do not have intestinal pRb. Immunohistochemical staining of intestinal sections of Rbtm2Brn−/−Tg(Vil-Cre) mice showed clearly an absence of pRb when compared with sections of C57Bl6 mice in almost all cells along the crypt villus axis (Fig. 3A). In some sections, the tips of the villi had trace amounts of pRb staining. C-cell metastases from the small intestine, stomach, and fat also showed an absence of pRb staining confirming the Rbtm2Brn−/− genotype of the medullary carcinoma tumors.

Figure 3.

Immunohistochemical staining of pRb in the small intestine and in a C-cell stomach lesion and polyp formation in Rbtm2Brn Tg(Vil-Cre) mice with Apc1638N mutation. A, 1, normal villi of the small intestinal from a C57Bl6 mouse show strong staining and a clear presence of pRb throughout. 2, immunohistochemistry of normal villi of the small intestine from a C57Bl6 mouse without primary antibody as negative control. 3, intestinal villi from a Rbtm2Brn−/−Tg(Vil-Cre) mouse shows no staining and effective removal of pRb from the small intestine. 4, C-cell metastasis to the stomach shows loss of pRb. Stomach section is from the same stomach observed in H&E staining (Fig. 2B,, 3) and calcitonin immunohistochemistry (Fig. 2C,, 1). B, H&E staining of intestinal polyps from three Rbtm2Brn Tg(Vil-Cre) mice with Apc1638N mutation. 1 and 2, cross-sections of polyps from the duodenum of the small intestines of two different animals. 3, two of five polyps found in the cecum from a third Rbtm2Brn Tg(Vil-Cre) mice with Apc1638N mutation. 4, stomach polyp from the same mouse with the polyps of the cecum shown in Fig. 3B , 3.

Figure 3.

Immunohistochemical staining of pRb in the small intestine and in a C-cell stomach lesion and polyp formation in Rbtm2Brn Tg(Vil-Cre) mice with Apc1638N mutation. A, 1, normal villi of the small intestinal from a C57Bl6 mouse show strong staining and a clear presence of pRb throughout. 2, immunohistochemistry of normal villi of the small intestine from a C57Bl6 mouse without primary antibody as negative control. 3, intestinal villi from a Rbtm2Brn−/−Tg(Vil-Cre) mouse shows no staining and effective removal of pRb from the small intestine. 4, C-cell metastasis to the stomach shows loss of pRb. Stomach section is from the same stomach observed in H&E staining (Fig. 2B,, 3) and calcitonin immunohistochemistry (Fig. 2C,, 1). B, H&E staining of intestinal polyps from three Rbtm2Brn Tg(Vil-Cre) mice with Apc1638N mutation. 1 and 2, cross-sections of polyps from the duodenum of the small intestines of two different animals. 3, two of five polyps found in the cecum from a third Rbtm2Brn Tg(Vil-Cre) mice with Apc1638N mutation. 4, stomach polyp from the same mouse with the polyps of the cecum shown in Fig. 3B , 3.

Close modal

Apc1638N polyps develop in a Rb−/− gastrointestinal tract. We tested the ability of the Rbtm2Brn Tg(Vil-Cre) intestine to support the development of polyps by addition of an Apc1638N mutation. These mice became occult fecal blood positive at 7 months old. Three animals with Apc1638N mutations on Rb-deficient backgrounds were observed histologically. All animals had duodenal polyps. Multiple polyps of the cecum and stomach were also found. One animal had a large hair follicle tumor concurrent with gastrointestinal polyps. In none of the animals were pituitary or thyroid tumors observed nor were liver, bile duct, or lung tumors. We draw no conclusions from these data about Rb loss on tumor progression and expect to observe a large cohort of animals in the future for this purpose.

The conditional Rb model presented in this study under the direction of Villin-Cre has been constructed primarily to remove Rb from the GI tract and to observe the effects of Rb loss directly and indirectly on colorectal cancer. We definitively show that removal of Rb from the intestine is not sufficient to initiate gastrointestinal tumorigenesis. This is possible because the Villin promoter is active throughout the GI tract, including the stomach, small and large intestines, and cecum. Villin-Cre is also expressed strongly in the proximal tubules of the kidneys of these animals and in the nephros anlagen. By these data, Rb is not sufficient to initiate kidney tumors in mice. We cannot rule out the possibility that Villin-Cre is not expressed uniformly in the nephros anlagen and that primordial cells that cause Wilms' tumor would therefore contain intact Rb. The structure of intestinal villi in the Rbtm2Brn−/−Tg(Vil-Cre) mouse appear normal on histologic inspection of H&E-stained sections and seem to undergo normal differentiation, implying that Rb has no effect on cell differentiation in intestinal villi. However, because the Notch signaling pathway that functions through Rb phosphorylation plays a role in intestinal cell lineage specification (27, 28), it would follow that Rb would have similar capabilities. We suggest that closer inspection of the Rbtm2Brn−/−Tg(Vil-Cre) intestine could offer insights into this process. We find that addition of an initiation mutation by Apc1638N mutation to the Rb−/− intestinal background leads to the production of adenomas of the stomach, duodenum, and cecum of the GIT at an early age. We look forward to observing these tumors more closely in the future. One animal developed a large but typical Apc1638N hair follicle tumor in addition to gastrointestinal polyps.

We identify a latent tumor phenotype in Rbtm2Brn−/−Tg(Vil-Cre) mice at 12 to 17 months old. Ectopic rearrangement of the Rb gene leads to generation of aggressive tumors of the pituitary, medullary thyroid, liver, and bile duct as well as to hyperplasia of pulmonary neuroendocrine cells. The Rbtm2Brn mouse has been used previously in a p53-deficient background to produce “preneoplastic” lesions using intrabronchial injection of recombinant adenovirus-expressing Cre recombinase and is suggested to be a particularly good model for small cell lung carcinoma (29, 30). Because lung tumors develop in Rbtm2Brn−/−Tg(Vil-Cre) mice without preplanned removal of p53, comparative analysis may offer interesting insights. A low-penetrance tumor phenotype is seen for adipose tissue and squamous epithelium of the oral cavity. We found that observed C-cell tumors, C-cell metastases, and liver tumors have functionally lost the wild-type Rb allele, RNA, and/or pRb. We conclude that these tumors are the result of Cre-mediated rearrangement of the Rbtm2Brn allele. These results suggest that Villin-Cre is expressed in these tissues or in their precursors during embryogenesis. Embryonic Villin expression has been reported in the primitive endoderm, gut, nephros anlagen and its derivatives, and extraembryonic visceral endoderm of the yolk sac (14). In the adult mouse, its expression is limited to the epithelial cells of the digestive and urogenital tracts (21). It is not known to be active in nonepithelial cells at any stage of development (20, 21). However, low levels of Villin transcripts have been found in liver and pancreas primordia, and the pattern of Villin expression in the adult liver and pancreas suggests that it is not expressed in epithelial cells, such as hepatocytes and pancreatic acinar cells, but is expressed in biliary and pancreatic duct cells (18). These observations correlate with the low-penetrance phenotype seen in Rbtm2Brn−/−Tg(Vil-Cre) mice of adenocarcinomas and hyperplasia of the bile duct. Because C cells become malignant in response to Rb rearrangement, we suggest the Villin promoter may be active during embryogenesis in neural crest or neural crest stem cells, the progenitor cells of the medullar thyroid. Such expression of the Villin gene has not been documented previously.

“Illegitimate” Villin transcripts have been found in normal bone marrow (31). Cre production in bone marrow followed by circulation of hematologic cells that contain the rearranged allele may explain the presence of low levels of rearranged RbΔ19 allele in tissues not expected to have Villin promoter activity during embryogenesis, such as tissue precursors for spleen, heart, and tail.

Some Cre constructs [e.g., Keratin 14 Cre (32)] have been shown to have recombinase expressed in oocytes, making transmission of the gene through males a necessity. We are unaware that Villin-Cre would have such constraints; however, to address the possibility, we assert that four of five founder Villin-Cre mice were male. Because subsequent offspring were intercrossed, if Villin is expressed in oocytes, Cre could have been expressed in the F2 generation. It is noteworthy in the Keratin 14 Cre model that early expression of Cre in oocytes leads to the complete deletion of all lox P–flanked sequences in all tissues of the offspring. Because there are tissues in Rbtm2Brn−/−Tg(Vil-Cre) mice that remain completely or predominantly unrearranged, we think it unlikely that expression of the Villin promoter in oocytes would account for the ectopic Villin expression observed in Rbtm2Brn−/−Tg(Vil-Cre) mice.

An interesting feature of these animals is the aggressive nature of the medullar carcinoma of the thyroids that were shown to have the ability to metastasize to unusual locations, including stomach, small intestine, liver, lung, kidney, pancreas, spleen, bone marrow, fat, salivary glands, lymph nodes, and dorsal root ganglion. Metastasis to the bone marrow and dorsal root ganglion is observed rarely in mouse models of cancer. Because there is a paucity of genetically engineered mouse models for metastasis, there has been a call for their development (33). Metastasis is modeled in mice in several different ways. Immune incompetent mice (e.g., nude and severe combined immunodeficient) have been used for transplantation studies involving syngeneic or xenogeneic tumor tissues. These approaches include tumor cell injection into the bloodstream and subcutaneous or transplantation of tumor material. These methods have various advantages and drawbacks. Theoretically, autochthonous tumors are more similar to tumor initiation and progression in humans. They permit the study of modifiers of tumor suppressors and oncogenes and in the presence of an intact immune system. We suggest that the Rbtm2Brn−/−Tg(Vil-Cre) model could offer insight into the process of metastasis for this reason. The ability of medullar carcinoma of the thyroid to metastasize to unusual locations must reflect significant genetic alterations in these cells. Identification of candidate genes involved in metastasis could be accomplished by microarray analysis of RNA from these tumors. Comparison of microarray expression and comparative genome hybridization patterns with patterns from human thyroid tumors known to contain Rb mutations could permit identification of novel genes involved in the human disease (34). Because medullar carcinomas of the thyroids become quite large (7-8 mm), relatively large quantities of tumor material would be available for screening.

Grant support: NIH grants CA-084301 and ES-011040 (R.S. Kucherlapati).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Dr. Mari Kuraguchi for the immunohistochemical staining procedure and Dr. Kenneth Hung for critically reading the article.

1
Weinberg RA. The retinoblastoma protein and cell cycle control.
Cell
1995
;
81
:
323
–30.
2
Weinberg RA. The retinoblastoma gene and cell growth control.
Trends Biochem Sci
1990
;
15
:
199
–202.
3
Nevins JR. Toward an understanding of the functional complexity of the E2F and retinoblastoma families.
Cell Growth Differ
1998
;
9
:
585
–93.
4
Harbour JW, Dean DC. The Rb/E2F pathway: expanding roles and emerging paradigms.
Genes Dev
2000
;
14
:
2393
–409.
5
Lipinski MM, Jacks T. The retinoblastoma gene family in differentiation and development.
Oncogene
1999
;
18
:
7873
–82.
6
Jacks T, Fazeli A, Schmitt EM, Bronson RT, Goodell MA, Weinberg RA. Effects of an Rb mutation in the mouse.
Nature
1992
;
359
:
295
–300.
7
Lee EY, Chang CY, Hu N, et al. Mice deficient for Rb are nonviable and show defects in neurogenesis and hematopoiesis.
Nature
1992
;
359
:
288
–94.
8
Leung SW, Wloga EH, Castro AF, Nguyen T, Bronson RT, Yamasaki L. A dynamic switch in Rb+/− mediated neuroendocrine tumorigenesis.
Oncogene
2004
;
23
:
3296
–307.
9
Williams BO, Schmitt EM, Remington L, et al. Extensive contribution of Rb-deficient cells to adult chimeric mice with limited histopathological consequences.
EMBO J
1994
;
13
:
4251
–9.
10
Maandag EC, van der Valk M, Vlaar M, et al. Developmental rescue of an embryonic-lethal mutation in the retinoblastoma gene in chimeric mice.
EMBO J
1994
;
13
:
4260
–8.
11
Fung KM, Chikaraishi DM, Suri C, et al. Molecular phenotype of simian virus 40 large T antigen-induced primitive neuroectodermal tumors in four different lines of transgenic mice.
Lab Invest
1994
;
70
:
114
–24.
12
Tischler AS, Powers JF, Alroy J. Animal models of pheochromocytoma.
Histol Histopathol
2004
;
19
:
883
–95.
13
Vooijs M, van der Valk M, te Riele H, Berns A. Flp-mediated tissue-specific inactivation of the retinoblastoma tumor suppressor gene in the mouse.
Oncogene
1998
;
17
:
1
–12.
14
el Marjou F, Janssen K-P, Chang BH-J, et al. Tissue-specific and inducible Cre-mediated recombination in the gut epithelium.
Genesis
2004
;
39
:
186
–93.
15
Craig SW, Powell LD. Regulation of actin polymerization by villin, a 95,000 dalton cytoskeletal component of intestinal brush borders.
Cell
1980
;
22
:
739
–46.
16
Robine S, Jaisser F, Louvard D. Epithelial cell growth and differentiation. IV. Controlled spatiotemporal expression of transgenes: new tools to study normal and pathological states.
Am J Physiol
1997
;
273
:
G759
–62.
17
Pinto D, Robine S, Jaisser F, El Marjou FE, Louvard D. Regulatory sequences of the mouse villin gene that efficiently drive transgenic expression in immature and differentiated epithelial cells of small and large intestines.
J Biol Chem
1999
;
274
:
6476
–82.
18
Madison BB, Dunbar L, Qiao XT, Braunstein K, Gumucio DL. Cis elements of the villin gene control expression in restricted domains of the vertical (crypt) and horizontal (duodenum, cecum) axes of the intestine.
J Biol Chem
2002
;
277
:
33275
–83.
19
Robine S, Sahuquillo-Merino C, Louvard D, Pringault E. Regulatory sequences on the human villin gene trigger the expression of a reporter gene in a differentiating HT29 intestinal cell line.
J Biol Chem
1993
;
268
:
11426
–34.
20
Maunoury R, Robine S, Pringault E, et al. Villin expression in the visceral endoderm and in the gut anlage during early mouse embryogenesis.
EMBO J
1988
;
7
:
3321
–9.
21
Maunoury R, Robine S, Pringault E, Leonard N, Gaillard JA, Louvard D. Developmental regulation of villin gene expression in the epithelial cell lineages of mouse digestive and urogenital tracts.
Development
1992
;
115
:
717
–28.
22
Vooijs M, Berns A. Developmental defects and tumor predisposition in Rb mutant mice.
Oncogene
1999
;
18
:
5293
–303.
23
Harrison DJ, Hooper ML, Armstrong JF, Clarke AR. Effects of heterozygosity for the Rb-1t19neo allele in the mouse.
Oncogene
1995
;
10
:
1615
–20.
24
Hu N, Gutsmann A, Herbert DC, Bradley A, Lee WH, Lee EY. Heterozygous Rb-1 δ20/+ mice are predisposed to tumors of the pituitary gland with a nearly complete penetrance.
Oncogene
1994
;
9
:
1021
–7.
25
Nikitin AY, Juarez-Perez MI, Li S, Huang L, Lee WH. RB-mediated suppression of spontaneous multiple neuroendocrine neoplasia and lung metastases in Rb/− mice.
Proc Natl Acad Sci U S A
1999
;
96
:
3916
–21.
26
Prost S, Sheahan S, Rannie D, Harrison DJ. Adenovirus-mediated Cre deletion of floxed sequences in primary mouse cells is an efficient alternative for studies of gene deletion.
Nucleic Acids Res
2001
;
29
:
E80
.
27
Fre S, Huyghe M, Mourikis P, Robine S, Louvard D, Artavanis-Tsakonas S. Notch signals control the fate of immature progenitor cells in the intestine.
Nature
2005
;
435
:
964
–8.
28
Robine S, Fre S, Huyghe M, Artavanis-Tsakonas S, Louvard D. Notch signals control the fate of immature progenitor cells in the intestine.
Med Sci (Paris)
2005
;
21
:
780
–2.
29
Minna JD, Kurie JM, Jacks T. A big step in the study of small cell lung cancer.
Cancer Cell
2003
;
4
:
163
–6.
30
Meuwissen R, Linn SC, Linnoila RI, Zevenhoven J, Mooi WJ, Berns A. Induction of small cell lung cancer by somatic inactivation of both Trp53 and Rb1 in a conditional mouse model.
Cancer Cell
2003
;
4
:
181
–9.
31
Salbe C, de Cremoux P, Bonneton C, et al. Illegitimate villin transcripts in normal bone marrow precludes detection of colon cancer micrometastases.
Int J Biol Markers
2000
;
15
:
41
–3.
32
Hafner M, Wenk J, Nenci A, et al. Keratin 14 Cre transgenic mice authenticate keratin 14 as an oocyte-expressed protein.
Genesis
2004
;
38
:
176
–81.
33
Khanna C, Hunter K. Modeling metastasis in vivo.
Carcinogenesis
2005
;
26
:
513
–23.
34
Sweet-Cordero A, Mukherjee S, Subramanian A, et al. An oncogenic KRAS2 expression signature identified by cross-species gene-expression analysis.
Nat Genet
2005
;
37
:
48
–55.