Background
Most breast cancers originate from mammary epithelial cells (MECs) and progress through multiple steps. When a tumorigenic event such as gene amplification occurs [
1], MECs acquire the potential to proliferate in the mammary duct and develop into ductal carcinoma in situ (DCIS) [
2], which progresses to malignant carcinoma such as invasive breast carcinoma and then metastasizes into other organs through mesenchymal tissue. For in-vivo analysis of the genes involved in these processes, some genetically manipulated mouse models have been established, such as
TP53−/−, mouse mammary tumor virus (MMTV)-
neu, and MMTV-polyoma-virus middle T antigen (PyMT) mice. These models have boosted our understanding of the mechanisms by which mammary tumors develop and metastasize in vivo [
3‐
5]. We have identified several oncogene candidates from amplicons of breast cancer cell lines and showed that some of the genes exhibited tumorigenic activity when gene-introduced cells were inoculated subcutaneously into nude mice [
6‐
9]. Transgenic or knockout mouse models are useful for understanding the tumorigenic activity of those genes, but have not been practical for analyzing multiple candidate genes. Therefore, we have to establish a novel alternative method for preparing genetically engineered mice to improve the efficiency for analysis of multiple genes.
The mammary gland is a unique organ in that most mammary development occurs postnatally [
10]. A small fraction of mammary stem cells (MaSCs) and progenitor cells in the basal layers of the mammary gland have regenerative capacity and maintain organ homeostasis during estrous cycles [
11‐
17]. It has been demonstrated that the transplantation of a cell fraction with the surface markers CD49f
highCD24
+ or CD29
highCD24
+ into mouse fat pads from which the endogenous epithelium has been removed (cleared fat pads) can regenerate fully functional mammary gland [
12,
13].
Harnessing this regenerative capacity of MaSCs, gene-introduced mouse mammary gland can be produced from cells enriched in MaSCs by infection with lentiviral or retroviral vectors, which is a valuable alternative approach to producing transgenic animals [
18,
19]. However, the efficiency of viral packaging has been lower when introduced genetic elements have a longer or more complicated sequence, including promoter and transcription termination sequences. In addition, to mimic the actual mechanism of tumorigenesis, the target gene is required to be inducibly expressed after mammary reconstitution.
Here, we describe a convenient method for the genetic manipulation of mouse mammary gland using a transposon vector system and electroporation. Using this method, mCherry-expressing mammary glands were first generated. The tissues were analyzed by immunostaining to examine the transgene distribution in both luminal and basal cells. Milk production during pregnancy was also examined to verify normal differentiation and mammary gland function. In addition, a vector with long DNA (> 200 kb) derived from a bacterial artificial chromosome (BAC) clone was introduced to test its loading capacity. Moreover, using doxycycline (Dox)-inducible expression vector, we assessed the Dox-dependent expression of enhanced green fluorescent protein (EGFP) after mammary gland reconstitution. To examine whether oncogene- induced tumorigenesis is achieved with this method, PyMT was chosen for a case study and histological analyses were conducted to compare its mammary tissue with that of the MMTV-PyMT transgenic mouse model.
Methods
Mice
For a transplantation assay, rag2−/− (kindly provided by Dr. Takaki) immunocompromised albino mouse lines backcrossed with FVB or C57BL/6 J were used as recipient mice. FVB or C57BL/6 J mice were used as donors.
Vector construction and preparation
All vectors were constructed using either a ligase reaction kit (Nippon Gene, Tokyo, Japan, or Takara, Kyoto, Japan) or the In-Fusion reaction kit (Takara). In the case of long transposon donor vector loading BAC (Fig.
5a), a template vector was first constructed by joining six fragments using an In-Fusion reaction after providing each fragment with polymerase chain reaction or annealing 2-oligo DNAs. The contents of this template vector are depicted at the bottom of Fig.
5a. Here, 70-bp and 134-bp arms homologous to the BAC vector (pBACe3.6) were placed in its flanking regions. The N
otI/B
stXI (FastDigest from Thermo Scientific)-digested fragment of this template vector was introduced into an
E. coli BAC clone B6Ng01-263 N07 (RIKEN BioResource Center (BRC)) by electroporation, and a homologous recombination reaction was conducted by the RED/ET system (Gene Bridges, Heidelberg, Germany) to obtain the long transposon donor vector (Fig.
5a).
The long transposon donor vector was then purified using the NucleoBond Xtra BAC kit (Macherey-Nagel, Takara). Other vectors were purified with CsCl-gradient ultracentrifuge sedimentation after purification with an alkaline lysis solution method.
Cell culture
Mouse embryo-derived fibroblasts, C3H10T1/2 (RCB0247; RIKEN BRC), were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Wako Pure Chemical Industries, Ltd., Tokyo, Japan) containing 5% fetal bovine serum (FBS), 100 μg/mL streptomycin sulfate (Meiji Seika Pharma, Tokyo, Japan), and 100 U/mL penicillin G potassium (Meiji Seika Pharma) at 37 °C and 5% CO2. For co-culture with MEC, C3H10T1/2 cells were treated with 4 μg/mL Mitomycin C (Wako) for 3 h and incubated overnight at 5% CO2 and 37 °C for 1–2 days.
NMuMG cells were cultured in DMEM supplemented with 10% FBS, 100 μg/mL streptomycin sulfate, 100 U/mL penicillin G potassium, 10 μg/mL insulin, and 0.45% glucose.
Mammary cell preparation
The thoracic, abdominal, and inguinal mammary glands were dissected from 8- to 10-week-old female donor mice. After washing in phosphate-buffered saline (PBS), the tissues were chopped with a razor. Then, the tissues were digested for 1–2 h at 37 °C under shaking in DMEM/F12 medium (Invitrogen, Carlsbad, CA, USA) containing 5% FBS, 2 mg/mL collagenase IV (Sigma, Poole, UK), and 0.1 mg/ml hyaluronidase (Sigma). After removing the red blood cells in DMEM/F12 by repeated low-gravity centrifugation followed by suspension in ACK buffer (0.15 M NH4Cl, 10 mM KHCO3, 0.1 mM EDTA, pH 7.3), a dissociated cell suspension was yielded by pipetting for 8–10 min in cell dissociation buffer (PBS, 0.05% trypsin, 0.5 mM EDTA, 0.5% DNaseI), and then for 2 min in 5 mg/mL dispase (StemCell Technologies, Vancouver, Canada)–0.5% DNaseI, followed by filtration through a 42-μm mesh.
Cell sorting
A dissociated single-cell suspension was labeled with biotinylated CD45, Ter119, CD31, and BP-1 antibodies contained in EasySep Mouse Epithelial Cell Enrichment Cocktail (StemCell Technologies) for 30 min on ice, and after washing, incubated with anti-CD49f-PE (eBioscience, San Diego, CA, USA), anti-CD24-perCP-cy5.5 (eBioscience), streptavidin-ECD (Beckman-Coulter, Brea, CA, USA), and 7-amino-actinomycin D (7-AAD; Calbiochem, San Diego, CA, USA) for 30 min on ice. Cells were suspended in 2% FBS–Hank’s balanced salt solution including DNaseI before sorting. Cell sorting was carried out using fluorescence-activated cell sorting (FACS) SH800 (Sony, Tokyo, Japan). CD45− Ter119− CD31− BP-1− (Lin(−)) 7-AAD− CD49fhigh CD24+ cells were collected to represent the MaSC-enriched fraction.
Transfection
A total of 1 × 10
6 cells of the sorted MaSC-enriched fraction from about eight mice were suspended and washed in Opti-MEM (Life Technologies, Carlsbad, CA, USA) twice. Then, 10 μg of donor and helper vector DNA were mixed at a ratio of 3:1 with the MECs in Opti-MEM, and subjected to electroporation using NEPA21 (NEPAGENE, Chiba, Tokyo) with a 2-mm gap electrode cuvette (NEPAGENE, EC-002S). The settings for this electroporation are shown in Additional file
1 (Table S1). After electroporation, the cell suspension was immediately suspended in culture medium.
For transfection into NMuMG, dissociated cells were plated on six-well plates at 20–30% confluence a day before transfection. Donor and helper vectors were introduced at an OD260 ratio of 3:1 by pouring the mixture of 4 μg of DNA and 8–12 μg of polyethylenimine (Polysciences, Warrington, PA) into 200 μL of Opti-MEM.
Culture system for maintaining stemness of MaSCs
MECs were co-cultured with mitomycin C-treated C3H10T1/2 in DMEM/F12 supplemented with 10% FBS, 10 ng/mL human epidermal growth factor (hEGF; BD Biosciences, Tokyo, Japan), 5 μg/mL insulin (Wako), 0.5 μg/mL hydrocortisone (Sigma), 5 μM forskolin (Wako), 1.8 × 10
−4 M adenine (Sigma), 100 μg/mL streptomycin (Meiji Seika Pharma), 100 U/mL penicillin G (Meiji Seika Pharma), 50 μg/mL gentamycin (Nakalai Tesque, Tokyo, Japan), 10 μM Rho-associated coiled-coil-forming kinase inhibitor (ROCKi; Y-27632; LC Laboratories, New Boston, MA, USA) [
20,
21], and 10% Matrigel (growth factor reduced; BD Biosciences) at 5% CO
2 and 37 °C for 7 days. The numbers of MECs and C3H10T1/2 were 5000 and 6.25 × 10
4, respectively, in 250 μL of culture medium in a 48-well plate. When wells of different sizes were used, these numbers were changed proportionately relative to the area of the well.
Transplantation assay
The transfected MECs under culture were treated with dispase for 1–2 h at 37 °C and then approximately 3% of all transfected cells were suspended in 4–10 μL of DMEM/F12 including 10% Matrigel per transplantation site and transplanted into the cleared fat pads of the inguinal mammary glands of rag2−/− mice from which the endogenous epithelium had been removed using a 50-μL syringe equipped with a 30-G needle (ITO, Shizuoka, Japan). Mammary repopulation was analyzed by detecting bioluminescence using an in-vivo imaging system (IVIS Lumina XR, PerkinElmer, Waltham, MA, USA) and mCherry fluorescence using a stereomicroscope (Leica, Wetzlar, Germany).
Carmine alum staining
The dissected mammary glands were spread on a glass slide and fixed using Carnoy’s fixative (60% ethanol, 30% chloroform, and 10% glacial acetic acid) overnight at room temperature. Fixed tissues were washed in 70% ethanol, gradually rehydrated to distilled H2O, and then incubated in carmine alum solution (0.2 wt% of carmine (Sigma) and 0.5 wt% of aluminum potassium sulfate (Wako) in distilled H2O) for 2 h to overnight at room temperature. After gradual dehydration from 70% ethanol to 100% ethanol, fat pads were cleared overnight in xylene and mounted in MGK-S (Matsunami, Osaka, Japan).
Hematoxylin and eosin staining
After washing in PBS, the dissected mammary glands and tumors were fixed in 4% paraformaldehyde–PBS overnight and for 2 days, respectively. After washing in PBS, samples were dehydrated gradually from 70% ethanol to 100% ethanol and then from 50% xylene in ethanol to 100% xylene, following paraffin replacement using a Leica ASP300 fully automatic closed tissue processor and paraffin-embedding using a Leica EG1160. The paraffin-embedded tissues were then cut into 5-μm thick sections using a Leica SM200R sliding microtome. The sections were gradually deparaffinized in xylene and then with ethanol, gradually decreasing from 100% to 50% ethanol, and washed with distilled H2O, and then stained in hematoxylin solution (0.25% hematoxylin (Nacalai Tesque), 0.05% sodium iodate (Nacalai Tesque), 12.5% potassium alum (Wako), and 0.25% citric acid (Wako)) for 10 min. After washing in distilled H2O, sections were blued in 0.1% saturated lithium carbonate at 37 °C for 5 min and then washed in distilled H2O and stained in eosin solution (1% eosin (Wako) and 0.02% glacial acetic acid) for 10 min. After washing in 90% and 100% ethanol, sections were soaked in xylene for 5 min and then mounted in MGK-S (Matsunami).
Immunohistochemistry
The inguinal mammary glands were dissected, cut into ~ 1–5-mm
3 fragments, and pre-fixed for 10–15 min in 4% paraformaldehyde–PBS on ice. Tissues were washed in cold PBS and incubated for about 1 h in 10% sucrose–PBS at 4 °C, for about 1 h in 20% sucrose–PBS at 4 °C, and overnight in 40% sucrose–PBS at 4 °C. Tissues were frozen in cryoembedding medium [
22] in liquid nitrogen. The frozen blocks were then cut into sections of 5 or 10 μm thickness using a Leica CM1850 cryostat. Sections were dried for 1–10 min at room temperature, placed for more than 10 min in PBS, and post-fixed for 10–15 min in 3% paraformaldehyde at room temperature. After washing for more than 10 min in PBS or 0.1% Tween-PBS, sections were incubated in blocking buffer (0.1% Triton/10% goat serum in PBS) for 1 h at room temperature or overnight at 4 °C. Staining with primary antibodies was performed overnight at 4 °C or for 1 h at room temperature. Then, sections were washed three times in PBS or 0.1% Tween-PBS for 10 min, and staining with secondary antibody solutions containing 4′,6-diamidino-2-phenylindole (DAPI) was performed for 1 h at room temperature. In the case of milk fat globule (MFG) staining, sections were incubated in BODIPY 493/503 (3 μg/mL; Molecular Probes) containing DAPI in PBS for 10 min at room temperature. Then, sections were washed twice in PBS for 10 min. Finally, slides were mounted in MOWIOL DABCO.
The following primary antibodies were used: anti-KRT14 (mouse, 1:1000; Novocastra, Newcastle, UK), anti-KRT8 (rat, 1:250; Developmental Studies Hybridoma Bank, University of Iowa), anti-mCherry (rabbit, 1:250; Abcam, Cambridge, UK), and anti-PyMT (rat, 1:500; Santa Cruz, CA, USA). The following secondary antibodies were also used: anti-mouse, anti-rabbit, and anti-rat conjugated to AlexaFluor 488 (1:1000; Molecular Probes, Eugene, OR, USA) and to AlexaFluor 568 (1:1000; Molecular Probes). Staining of nuclei was performed with DAPI (1:4000; DOJINDO, Tokyo, Japan). Adjustment for the variations in the brightness and/or contrast of the entire area of images was performed using Photoshop (Adobe, San Jose, CA, USA).
Immunoblotting
Cells were collected in RIPA buffer (10 mM Tris-HCl (pH 8.0), 1% (w/v) NP40, 0.1% (w/v) sodium deoxycholate (Wako), 0.1% (w/v) SDS (Wako), 0.15 M NaCl (Wako), 1 mM EDTA, 10 mM NaF (Wako), 1.5 mM Na3VO4 (Wako), and cOmplete™ Protease Inhibitor Cocktail (Roche)). Protein concentrations were titered using the BCA protein assay kit (Thermo Fisher Scientific). Collected protein lysate was mixed with SDS-PAGE loading buffer (0.15 M Tris-HCl, 6% (w/v) SDS, 0.003% (w/v) bromophenol blue (Wako), 30% (w/v) glycerol (Wako), and 15% (w/v) β-mercaptoethanol (Wako)) and then boiled at 95 °C for 5 min, followed by SDS-PAGE and immunoblotting. Antibodies used for immunoblotting were as follows: anti-milk (rabbit, 1:1000; Nordic-MUbio, Susteren, the Netherlands), anti-PyMT (rat, 1:500; Santa Cruz), anti-mCherry (rabbit, 1:500; Abcam), anti-histone H3 (rabbit, 1:2500; Cell Signaling Technology, MA, USA), and anti-α-tubulin (mouse, 1:5000; Calbiochem).
Discussion
The mouse mammary gland provides a unique model for the study of MaSCs and differentiation pathways and pathogenesis of breast cancer [
38‐
40]. Recent studies of cellular hierarchical issues using mammary glands have contributed to better understanding of “cells of origin” and “cancer stem cells” with regard to breast cancer [
5,
40‐
43]. Breast cancer occurs when genetic mutations occur during tissue development. Transgenic and knockout mouse models have been used in previous studies involving genetic analysis. By several in vitro, allograft, and xenograft screenings and analyses of microarray data, we found multiple candidate genes and hoped to evaluate their oncogenic or malignant phenotypes in vivo on mammary epithelium at specific differentiation phases [
6‐
9]. For practicality of this situation, an alternative method needs to be developed for producing genetically manipulated mammary glands with ease and rapidity. Combined with
piggyBac transposon vector system and electroporation, we obtained genetically manipulated mammary gland in 2–3 months. Although, the reconstitution efficiency of the first genetically manipulated mammary gland was low (3/18 to 7/18) (Table
1), secondary transplantation after cutting the first reconstituted mammary glands into small pieces and selecting gene-introduced mammary gland using the mCherry fluorescence marker yielded successful expansion with high efficiency (11/16 to 12/14 from one regenerated mammary gland). This property is highly useful for statistical analyses or maintenance of mammary tissue lines. Doxycycline-dependent change in expression levels was successfully achieved in a case where an EGFP- or PyMT-inducible expression cassette was used in the vector. In future studies, we hope to apply this tool to candidate genes that may confer malignant phenotypes and evaluate their in-vivo function during the development of mammary gland. By using IVIS, events such as metastasis can be monitored by our vector system. In addition, the expression promoter presented here can be modified depending on the purpose of the particular research. For example, by utilizing the K8 or K14 promoter, lineage-specific expression control in luminal or basal cells, respectively, can be achieved. In addition, it is also possible to establish a gene suppression system by introducing CRISPR-Cas9 or a knockdown approach. These modifications are easily achieved thanks to the use of the transposon donor vector backbone, which has almost no limitation concerning DNA complexity (termination or promoter sequence) and length. Actually, our system enabled us to produce a mammary gland into which a BAC vector was integrated whose vector size was > 200 kb (Fig.
5), suggesting the possibility of using genomic DNA elements of almost unlimited length for gene transduction. Thus, the presented methods and tools may be broadly applicable and open a new avenue for breast cancer research.
Acknowledgments
We thank all laboratory members who supported the maintenance and breeding of mice, without whom the experiments would not have been successful. We are also grateful to Dr. Allan Bradley and Dr. Kosuke Yusa (the Wellcome Trust Sanger Institute) for providing
piggyBac transposon’s backbone donor vectors [
44] and a hyPBase vector [
29], Dr. Keiji Miyazawa (University of Yamanashi in Japan) for NMuMG cells, Dr. Satoshi Takaki (National Center for Global Health and Medicine in Japan) for the
rag2−/− mouse line, and Dr. Takeharu Sakamoto (IMSUT) for tissue samples of MMTV-PyMT transgenic mice. The authors would like to thank Enago (
https://www.enago.jp) for the English language review.