A bioluminescence reporter mouse that monitors expression of constitutively active β-catenin

This short technical report describes the generation and characterization of a bioluminescence reporter mouse that is engineered to detect and longitudinally monitor the expression of doxycycline-induced constitutively active β-catenin. The new responder transgenic mouse contains the TetO-ΔN89β-CatTMILA transgene, which consists of the tet-operator followed by a bicistronic sequence encoding a stabilized form of active β-catenin (ΔN89β-catenin), an internal ribosome entry site, and the firefly luciferase gene. To confirm that the transgene operates as designed, TetO-ΔN89β-CatTMILA transgenic mouse lines were crossed with an effector mouse that harbors the mouse mammary tumor virus-reverse tetracycline transactivator (MMTV-rtTA) transgene (termed MTB hereon), which primarily targets rtTA expression to the mammary epithelium. Following doxycycline administration, the resultant MTB/CatTMILA bigenic reporter exhibited precocious lobuloalveologenesis, ductal hyperplasia, and mammary adenocarcinomas, which were visualized and monitored by in vivo bioluminescence detection. Therefore, we predict that the TetO-ΔN89β-CatTMILA transgenic responder mouse—when crossed with the appropriate effector transgenic—will have wide-applicability to non-invasively monitor the influence of constitutively active β-catenin expression on cell-fate specification, proliferation, differentiation, and neoplastic transformation in a broad spectrum of target tissues.

Not surprisingly, aberrant β-catenin signaling is causal for numerous embryonic and postnatal developmental abnormalities-including tumorigenesis-in diverse anatomic sites [1]. Mutations that prevent phosphorylation and turnover of the cytoplasmic pool of β-catenin lead to the accumulation of a constitutively active form of β-catenin that can inappropriately induce downstream Wnt target genes in a Wnt-independent manner. Importantly, the genetically engineered mouse was instrumental in validating many of the above findings in an in vivo context.
Because the engineered mouse has been a pivotal in vivo experimental model to study Wnt signaling in general and β-catenin signaling in particular, it is imperative that the full potential of the mouse is attained to further advance our understanding of β-catenin action in vivo. Accordingly, this brief technical report describes the generation and first-line characterization studies of a new TetO-ΔN89β-Cat TMILA transgenic mouse that enables noninvasive in vivo detection of constitutively active β-catenin expression using a bioluminescence reporter. Given the importance of β-catenin signaling in development, adult tissue homeostasis, and tumorigenesis, we believe this bioluminescence reporter mouse will have wide-applicability.

Materials and methods
Creation of the TetO-ΔN89β-Cat TMILA transgenic mouse Mice were housed in an AAALAC accredited vivarium at Baylor College of Medicine, which operates a 12h-light: 12h-dark recurrent photocycle in temperature-controlled mouse rooms (22 ± 2˚C). Mice were fed irradiated Teklad global soy protein-free extruded rodent diet (Harlan Laboratories, Inc., Indianapolis, IN) and fresh water ad libitum when not treated with doxycycline (see below). The TetO-ΔN89β-Cat TMILA transgene was generated by inserting a cDNA (2.1kb) encoding Xenopus ΔN89β-catenin [8][9][10] into unique EcoR1/Cla1 restriction sites in the TMILA reporter vector (7.4kb (Chodosh plasmid #652)) [11]. The 89 amino acid N-terminal deletion in ΔN89β-catenin renders the truncation mutant constitutively active [12,13]. The insertion of the ΔN89β-catenin cDNA into the TMILA reporter vector positions the ΔN89β-catenin cDNA downstream of the cytomegalovirus (CMV) minimal promoter and a tandem repeat of seven Tet operator (TetO) sequences derived from the pTetSplice vector [14]. With this cloning approach, the ΔN89β-catenin cDNA insertion is sequentially followed by an internal ribosome entry site (IRES) and the codon optimized firefly luciferase 2 gene cassette from Photinus pyralis (Promega, Madison, USA). A strong Simian virus 40 (SV40) splicing/polyadenylation sequence is included at the 3' end of the transgene. The resultant TetO-ΔN89β-Cat TMILA transgene (6.6kb) was released from pTetSplice vector sequences (2.9kb) with Not1 digestion, isolated from vector sequences, and then purified for microinjection into pronuclei of single-cell embryos of the FVB/N inbred strain. Founder mice (F0) and their progeny were identified by PCR genotyping of genomic DNA isolated from tail snips. The PCR primers that were used to detect the TetO-ΔN89β-Cat TMILA transgene were previously described [10]. Each of the four TetO-ΔN89β-Cat TMILA responder transgenic lines (F1) was crossed with the MTB effector transgenic [14] to generate MTB/Cat TMILA bigenics that were maintained in the FVB/N background strain. All studies described herein were conducted with nulliparous mice that were hemizygous for the transgene.

Bioluminescence imaging
Twenty four hours prior to mouse imaging, the ventral area of the skin was depilated using a commercially available depilating cream. On the day of imaging, isoflurane-anesthetized mice were intraperitoneally injected with RediJect D-luciferin bioluminescent substrate (PerkinElmer, Waltham, MA (150mg/kg)) in sterile 0.9% saline. After 5 minutes, bioluminescence was detected and recorded using the Bruker FX Pro Imager (Bruker, Billerica, MA) equipped with an isoflurane manifold for continuous anesthesia; mice were placed in the ventral recumbent position for bioluminescence detection. Bioluminescence was captured within a 30 second exposure time with 4x4 pixels binning followed by X-ray image capture (10 second exposure). Using Bruker Molecular Imaging software (v.7.1.3.20550), bioluminescence images were exported in pseudo color format with matched rainbow-colored bar scales (minimum and maximum photons/second). For final presentation purposes, bioluminescence images were overlaid upon the corresponding grey-scale X-ray image.

Whole mount, immunohistochemical, and western immunoblot analysis
Carmine-red stained mammary gland whole-mounts were performed as previously reported [10,15,16]. The antibodies and conditions used for immunohistochemical detection of the myc-epitope tag and 5'-bromo-2'-deoxyuridine (BrdU) incorporation have been described [10]. To determine the percentage of mammary epithelial cells that is immunopositive for The ΔN89β-catenin cDNA (2.1kb) was cloned into a single EcoR1 restriction site downstream of the TetO sequence in the TMILA (7.4kb) cloning vector [11]. The ΔN89β-catenin cDNA encodes the truncated Xenopus β-catenin protein with a myc-epitope tag fused in-frame at its Nterminus (black box). The location of the PCR primers for genotyping (black arrowheads) as well as the 13 centrally located Armadillo repeats (Arm repeats) is indicated. The inserted ΔN89β-catenin cDNA is followed by an IRES and a cDNA encoding the firefly luciferase protein. A SV40 polyadenylation signal (PA) serves as a strong transcriptional termination signal. The TetO-ΔN89β-Cat TMILA transgene was linearized with Not1, isolated from vector sequences, and purified prior to pronuclear microinjection. (B) Schematic depicts the breeding strategy to generate the MTB/Cat TMILA bigenic mice by crossing the MTB effector transgenic [14] with TetO-ΔN89β-Cat TMILA responder transgenic. (C) Typical western immunoblot of isolated mammary epithelial cell protein. Lane 1, 2, and 3 denote mammary epithelial protein isolated from wild type ((WT) or non-transgenic) control (without doxycycline), MTB/ Cat TMILA bigenic (without doxycycline), and MTB/Cat TMILA bigenic mice on food and water supplemented with doxycycline for 1-month respectively. Using antibodies to full-length β-catenin and the myc-epitope tag, the transgene-derived ΔN89β-catenin protein (75kDa) is only detected in the MTB/Cat TMILA bigenic treated with doxycycline (lane 3); β-actin serves as a loading control. Each lane represents a protein isolate pooled from four individual mice per genotype and treatment. doi:10.1371/journal.pone.0173014.g001 BrdU incorporation, 6 control monogenic and 5 MTB/Cat TMILA bigenics were used. Note: ductal and alveolar epithelial cells were counted for the MTB/Cat TMILA bigenic whereas only ductal epithelial cells could be counted for the control monogenic gland since there are very few alveolar cells in the adult virgin mammary epithelium to get an equivalent count. Only intensely stained (dark brown) nuclei for BrdU incorporation were included in the cell count. The average number of immunopositive cells was calculated from a total of 500 mammary epithelial cells from three separate mammary gland sections per mouse. Final counts were expressed as an average percentage mean of cells counted. Antibodies and conditions used for western immunoblot detection of β-catenin, the myc-epitope tag, cyclin D1, and β-actin in protein isolates from mammary epithelial cells and tumor tissue have been previously detailed [10]. Mammary epithelial cell isolation was followed according to established methods [17,18].

Mammary tumor induction
Mice chronically administered doxycycline for mammary tumor induction were checked twice weekly by manual palpation. Mice were euthanized when mammary tumors reached approximately 1.0cm in diameter as measured by Vernier calipers. For each mouse, tumor size, number, and ventral location were recorded prior to euthanasia. GraphPad Prism 6 software (GraphPad Software, Inc., La Jolla, CA) was used to generate and statistically analyze tumor-free Kaplan-Meier plot. The two-sided log-rank test was used to determine significance of the difference in tumor-free rate between virgin bigenic mice that chronically received food and water supplemented with doxycycline (n = 21) and bigenic mice maintained on regular mouse chow (n = 28); a p-value <0.05 was considered significant.

Results and discussion
Design and generation of the TetO-ΔN89β-Cat TMILA responder transgenic mouse The design of the TetO-ΔN89β-Cat TMILA transgene is schematically shown in Fig 1A. The Xenopus ΔN89β-catenin cDNA (2.1kb) was inserted between EcoR1 (5') and Cla1 (3') restriction sites in the multiple cloning cassette of the TMILA vector (7.4kb) [11]. Note: the Xenopus and human β-catenin protein sequences share 98% homology. To enable specific immunodetection, the ΔN89β-catenin cDNA was engineered to express in-frame a myc-epitope tag at the N-terminus of the ΔN89β-catenin protein [10]. Due to deletion of its first 89 amino acids, the ΔN89β-catenin protein is constitutively active. The insertion of the ΔN89β-catenin cDNA into  upstream of (in sequential order) the IRES, luciferase reporter, and the SV40 intron/polyA cassette. With standard transgenic methodology [10,15,16,19], four out of six mice positive for the TetO-ΔN89β-Cat TMILA transgene transmitted the transgene through the germline (#G4704; #G4715; #G4717; and #G4720). To confirm that ΔN89β-catenin is expressed by these transgenics in response to doxycycline administration, each of the four transgenic lines was crossed with the MTB effector transgenic [14] to generate the MTB/Cat TMILA bigenic ( Fig  1B). Unless otherwise stated, data derived from the #G4715 are shown herein which are representative of the other transgenic lines. Western immunodetection for β-catenin and the myc-epitope tag demonstrates that the transgene-derived ΔN89β-catenin protein (75kDa) is specifically expressed in mammary epithelial protein isolates derived from the MTB/Cat TMILA bigenic in response to doxycycline administration (Fig 1C).

Doxycycline induction of ΔN89β-catenin expression in the mammary epithelium of the virgin MTB/Cat TMILA bigenic is detected by bioluminescence
Using the luciferase reporter as a surrogate for transgene-derived ΔN89β-catenin expression, the bioluminescence emission signal was detected as early as 24 hours following doxycycline administration in the MTB/Cat TMILA bigenic (Fig 2A). The bioluminescence signal was detected in the majority of mammary glands (thoracic and inguinal) in the adult virgin MTB/ Cat TMILA bigenic; as expected, the bioluminescence signal was not detected in the control mouse (Fig 2A). Within a short time-period, only the mammary gland of the doxycyclinetreated MTB/Cat TMILA bigenic exhibited precocious lobuloalveologenesis and ductal sidebranching (Fig 2B and 2C). This result demonstrates that the TetO-ΔN89β-Cat TMILA responder transgene both operates as a bioluminescent reporter and causes β-catenin-dependent disruption of epithelial growth homeostasis. Immunohistochemical staining for the mycepitope tag and BrdU incorporation demonstrated that ΔN89β-catenin expression was restricted to the nucleus and cytoplasm of the mammary epithelium and that this expression pattern was coincident with mammary epithelial hyperplasia (Fig 2D-2I). Collectively, these data show that bioluminescence detection can forecast ΔN89β-catenin induced mammary epithelial abnormalities in the MTB/Cat TMILA bigenic reporter mouse.

Longitudinal monitoring of mammary tumor progression in the MTB/ Cat TMILA bigenic reporter mouse
With continuous doxycycline administration, the MTB/Cat TMILA bigenic reporter mouse develops palpable mammary tumors that can be detected and longitudinally monitored by bioluminescence (Fig 3A). As expected these mammary tumors are mostly adenocarcinomas that exhibit a strong immunopositive signal for the transgene-derived myc-epitope tag and are highly proliferative (Fig 3B-3F). The mammary tumor phenotype is 100% penetrant and is completely dependent on doxycycline administration (Fig 4A); many of the MTB/Cat TMILA bigenic reporters exhibit multifocal mammary tumors that are readily visualized by bioluminescence detection (Fig 4B and 4C). Although most of the MTB/Cat TMILA mammary tumors are adenocarcinomas (Fig 3B-3F), approximately 15% of tumors are squamous metaplasias arrowhead  [20,21] that are strongly immunopositive for the transgene-derived myc-epitope tag and BrdU incorporation (Fig 4D-4G). Interestingly, approximately 10% of MTB/Cat TMILA palpable mammary tumors do not regress following doxycycline withdrawal (S2A Fig). Bioluminescence monitoring clearly reveals that transgene expression activity is rapidly attenuated following doxycycline withdrawal but mammary tumor volume does not decrease (S2A and Doxycycline-induction of ΔN89β-catenin expression in the murine salivary gland can be temporally monitored using bioluminescence detection Bioluminescence monitoring also reveals that progeny from one transgenic line (#G4704) exhibit particularly strong ΔN89β-catenin expression in the submandibular salivary gland with low expression in the mammary epithelium (Fig 5A). Following chronic doxycycline exposure, histological analysis of all bigenic mice from this transgenic line revealed that the salivary gland exhibited epithelial hyperplasia that was strongly positive for myc-epitope tag expression (Fig 5B-5G). This result is not surprising as the MTB effector mouse has been shown to express rtTA activity in the salivary gland [14], an exocrine tissue developmentally similar to the mammary gland. However, we did not observe strong transgene expression in both the salivary and mammary gland in any transgenics examined. Because dysregulation of β-catenin signaling has been shown to elicit salivary gland tumorigenesis [22][23][24], this model may be useful as a non-invasive tool to explore further the involvement of β-catenin signaling in the ontogenesis of this poorly understood head and neck cancer.

Conclusion
We and others previously generated variations of the TetO-ΔN89β-Catenin transgenic effector mouse that was designed to conditionally induce constitutively activated ΔN89β-catenin protein in response to doxycycline [10,25]. In this short technical report, we describe an important improvement to this model with the generation of the TetO-ΔN89β-Cat TMILA transgenic, which has the added capability of bioluminescence detection. Because of the obvious advantages of optical bioluminescence imaging-a rapid, sensitive, and user-friendly optical modality for temporal assessment of tumor progression and therapy-along with the pleiotropic role of β-catenin signaling in normal and abnormal tissue homeostasis, we anticipate that this new bioluminescence reporter mouse will prove to be invaluable for future investigations to discern the normal and pathogenic role of β-catenin in vivo. A representative MTB/Cat TMILA bigenic with a palpable mammary tumor is shown (black arrowhead) in the top two panels (low and high magnification). For this representative mouse, 136 days of doxycycline administration was required to induce a palpable thoracic (#3) mammary tumor with a~1cm diameter. Following 6-days on a standard diet without doxycycline (or de-induction), the size of the same tumor rapidly reduced (middle two panels). By 12-days without doxycycline in the diet, the thoracic mammary tumor in the MTB/Cat TMILA bigenic is undetectable by manual palpation. Of the MTB/Cat TMILA bigenic mice in this study (n = 20), 18 mice showed rapid mammary tumor regression within 14-days whereas 2 mice did not show mammary tumor regression within this time period (S2 Fig). (TIF)