Introduction
Cell division cycle 42 (Cdc42) is a member of the Ras homolog (Rho) family of small guanine nucleotide triphophatases (GTPases) that is overexpressed and hyperactivated in human breast invasive ductal carcinomas [
1‐
3]. A number of studies in other cell and tissue types have demonstrated that Cdc42 regulates cell cycle progression, polarity, migration, cell fate determination, and differentiation, which are essential for mammary gland development and become disrupted during tumor formation [
4‐
6]. We previously showed that Cdc42 is required for primary mammary epithelial cell (MEC) morphogenesis
in vitro and that it regulates polarity establishment, proliferation, and survival of MECs during acinar morphogenesis [
7]. Studies investigating the importance of Cdc42 during postnatal mammary gland development
in vivo are currently ongoing in our laboratory, and our data indicate that Cdc42 knockout MECs are outcompeted by wild-type neighboring MECs (Bray and Vargo-Gogola, unpublished results). Thus, these loss-of-function studies indicate that Cdc42 is important for normal mammary gland morphogenesis
in vitro and
in vivo. However, our understanding of the mechanisms by which aberrant expression of Cdc42 disrupts mammary gland morphogenesis and facilitates tumor formation and progression
in vivo has been limited until now due to a paucity of
in vivo mouse models of Cdc42 overexpression and hyperactivation.
Postnatal mammary gland development is initiated in mice at 3 weeks of age when rising levels of ovarian hormones and local growth factors stimulate MEC proliferation and motility within terminal end buds (TEBs). The concerted action of MEC proliferation and motility drives TEB invasion and branching into the mammary fat pad, giving rise to an arborized ductal tree [
8]. Mammary gland branching morphogenesis also requires reiterative interactions between the epithelial and stromal compartments, and both secreted factors and mechanical signals between the two compartments contribute to branch formation and patterning [
9‐
11].
Rho family GTPases function within epithelia to integrate and transduce bidirectional soluble and mechanical signals between the epithelial and stromal compartments [
12]. Studies indicate that Rho signaling coordinates cell proliferation and motility with changes in cell shape and intracellular contractility that are important for outgrowth and patterning of the branched mammary ductal tree. For example, real-time imaging studies using an
in vitro model of mammary gland branching have revealed that Ras-related C3 botulinum toxin substrate 1 (Rac) and actomyosin contractility are required for mammary gland branch formation, whereas Rho-associated coiled-coil containing protein kinase (ROCK) functions to suppress hyperbranching and promote reorganization of the bilayered mammary epithelium [
9,
13]. Although Rac and ROCK have been implicated in mammary gland branching morphogenesis, the contribution of Cdc42 is not well understood.
To investigate the effects of Cdc42 overexpression during normal and neoplastic mammary gland development in vivo we developed a Cdc42 conditional overexpression mouse model and examined the effects of Cdc42 overexpression during postnatal mammary gland morphogenesis. We show that Cdc42 overexpression disrupts TEB morphogenesis and induces hyperbranching. Intriguingly, Cdc42 overexpression does not impact cell cycle progression to drive increased branching. Rather, it enhances MEC contractility and migration potentially by activating mitogen-activated protein kinase (MAPK) signaling. Our studies also suggest that Cdc42 overexpression in the mammary epithelium alters epithelial-stromal interactions, leading to stromal alterations that likely contribute to the epithelial phenotypes.
Methods
Mouse husbandry and care
Mice were housed in the University of Notre Dame Freimann Life Science Center. All studies were approved by the Institutional Animal Care and Use Committee at the University of Notre Dame and Indiana University School of Medicine (Protocol Number 14-015) and were conducted in accordance with the guidelines of the US Public Health Service Policy for Humane Care and Use of Laboratory Animals. All efforts were made to minimize suffering of the mice. Mice were fed a conventional diet or doxycyline (dox) containing chow ad libitum and were maintained at 21 to 22°C with a 12-hour light and 12-hour dark cycle. To induce transgene expression or control for any effects of dox, bitransgenic and control female mice were fed dox-containing chow (2 g/kg) (Bio-Serv Inc, Frenchtown, NJ, USA, S3893) for the specified number of weeks beginning at 4.5 weeks of age.
Transgenic mice
To generate the tetracycline (tet)-regulatable Cdc42 transgenic mouse model we created the following construct. The approximately 600 bp wild-type human Cdc42 cDNA was subcloned from the pCMV-Sport6 vector (Thermo Scientific Open BioSystems, Waltham, MA, USA) using PCR. Primers used to create and subclone the insert were as follows: Forward-5′-CCG GAA TTC ATG TAC CCA TAC GAT GTT CCA GAT TAC GCT CAG ACA ATT AAG TGT GTT GTT GTG GGC GAT G-3′ and reverse-5′-CCG GAA TTC GAT GTT CAT AGC AGC ACA CAC CTG-3′. The forward primer contained an EcoRI site and an ATG start site. The reverse primer contained an EcoRI restriction site. The PCR product was gel purified and digested with EcoRI. The Cdc42 insert was then ligated into the TMILA tetracycline operator (TetO)-IRES-luciferase vector downstream of the TetO [
14]. The construct was sequenced and then tested in MCF7 Tet-On cells [
15], which confirmed inducibility. A 5.2 Kb fragment containing the TetO-Cdc42-IRES-luciferase cassette was microinjected into the pronuclei of fertilized FVB/N oocytes by the Transgenic and Knockout Mouse Core at the Indiana University School of Medicine, Indianapolis, IN, USA. PCR was used to identify founder mice, which were bred with mouse mammary tumor virus-reverse tetracycline transactivator mice (MMTV-rtTA) [
16] already in use in our laboratory [
17]. Mice were maintained on an inbred FVB/N background.
Genotyping
Primer pairs used to genotype mice for the presence of the MMTV-rtTA transgene were: 5′-TCC AAG GGC ATC GGT AAA CA-3′ and 5′-GCA TCA AGT CGC TAA AGA AG-3′. Reaction conditions were 94°C for 3 min, followed by 30 cycles of 94°C for 30 s, 60°C for 45 s, and 72°C for 45 s followed by one cycle of 72°C for 5 min. Glycerol, 3 μl of a 50% solution, was added as part of a 25 μl total reaction volume to enhance product yield. Primer pairs used to genotype mice for the presence of the TetO-Cdc42 transgene were: 5′-CGT CAG ATC GCC TGG AGA CG-3′ and 5′-GAT GTT CAT AGC AGC ACA CAC CTG CGG-3′. Reaction conditions were 94°C for 3 min, followed by 30 cycles of 94°C for 1 min, 55°C for 30 s, and 72°C for 30 s, followed by one cycle of 72°C for 5 min.
Wholemount mammary gland preparation and morphometric analysis
Number 4 left mammary glands were fixed immediately after dissection in 4% PFA in PBS for 2 h while rocking on ice. Glands were rinsed in PBS and stained in carmine overnight with gentle rocking followed by destaining and dehydration in a series of ethanols for 1 h each with gentle rocking. Glands were cleared in xylenes overnight and stored long term in methyl salicylate. A StereoImager (Carl Zeiss, Inc, Oberkochen, Germany) was used to image the carmine-stained glands at low magnification (16×) for quantification of gland size and at high magnification (80×) for quantification of branching, ductal dilation, and TEB morphology. Branching was quantified by counting the total number branch nodes found in three 4 × 106 μm2 defined areas adjacent to the sides and leading edge of the lymph node. Ductal tree area was measured using the outline measurement tool in AxioVision 4.6 software (Zeiss). Ductal dilation was quantified using high magnification images (80×) and scoring the number of independent dilated regions found within the ductal tree from the back of the lymph node toward the leading end of the fat pad. Abnormal TEBs were defined as TEBs that were trifurcated or had multiple buds on the neck.
Histological methods and image quantification
Glands were fixed immediately after dissection in 4% PFA in PBS for 2 h while rocking on ice, rinsed in PBS, and stored in 70% ethanol at 4°C until paraffin embedding. Five-μm sections were cut, deparaffinized in xylenes, and rehydrated. Antigen retrieval was performed by boiling the sections for 20 min in 10 mM sodium citrate buffer. Sections were blocked and antibodies were diluted in 5% BSA/Tween or M.O.M.™ block reagent and M.O.M.™ antibody diluents (Vector Laboratories, Burlingame, CA, USA, BMK2202) for mouse-derived primary antibodies. Sections were incubated with blocking buffer for 1 h, with primary antibodies overnight, and secondary antibodies for 1 h. Elite ABC Reagent (Vector Laboratories, PK7100) and DAB (Vector Laboratories, SK-4100) were used to develop staining for immunohistochemistry (IHC) and the slides were counterstained with hematoxylin (Thermo Fisher Scientific, CS400-1D). For immunofluorescence staining Alexa Fluor™ 488 anti-mouse (Molecular Probes, Eugene, OR, USA, A11001), Alexa Fluor™ 488 anti-rat (A11006), Alexa Fluor™ 555 anti-rabbit (A21428), and Texas Red™-X anti-rabbit (T-6391) secondary antibodies were used. Masson’s Trichrome staining to detect collagen was done and quantified as previously described [
18]. At least five animals per group and a minimum of three TEBs per animal were analyzed for each experiment. Images were taken using an AxioImager (Zeiss). Images were quantified using the ImageJ cell counter plugin.
Bromodeoxyuridine (BrdU) injection of mice
Control and Cdc42-overexpressing mice treated with dox-containing chow for three weeks were given intraperitoneal injections of 3 mg/ml BrdU (Sigma-Aldrich, St Louis, MO, USA, B-5002) in saline at 10 μl per gram bodyweight 2 h prior to euthanasia and dissection of the mammary glands. The glands were fixed and stained as described above.
Organoid and fibroblast isolation for luciferase assays, GLISA, and qRT-PCR
Four-and-a-half-week-old Cdc42-overexpressing and control mice were treated with the dox diet for 1 week prior to euthanasia and mammary gland dissection. The 2, 3, and 4 mammary gland pairs were dissected, and lymph nodes were removed from the number 4 glands. Organoids and fibroblasts were isolated as previously described [
18]. Briefly, the glands were manually minced and incubated in DMEM/F-12 (Thermo Fisher Scientific, SH30272) with 2 mg/ml collagenase A (Roche, Geneva, Switzerland, 10103578001), 100 units/ml hyaluronidase (Sigma-Aldrich, H3506), and 1 x antibiotic-antimycotic (Invitrogen, Carlsbad, CA, USA, 15240-062) for 1 h at 37°C with 200 rpm rotation at a 45° angle. The tissues were shaken manually at 30 min and 60 min during the digestion to aid in breaking apart the tissues. The cells were washed with DMEM/F-12 and centrifuged twice at 450 g for 10 min. The cells were incubated at room temperature for 3 min with manual shaking in DMEM/F-12 with 2 units/ml DNase I (Sigma-Aldrich, D2463) and centrifuged at 450 g for 10 min. Differential centrifugation was used to separate fibroblasts from organoids, which consisted of pulse centrifugation to 450 g with the supernatant from the first spin containing the fibroblasts. Fibroblasts and organoids for luciferase assays were immediately frozen. Fibroblasts for qRT-PCR were frozen in Trizol (Invitrogen, 15596-018) for RNA isolation.
Single mammary epithelial cell isolation
Primary MECs used in in vitro assays were isolated from the 2, 3, and 4 mammary glands from mice treated for 1 week with dox. Initial steps for single cell isolation were identical to those used for organoid isolation as described above. Cells were washed in PBS and then digested in 0.05% trypsin + EDTA in PBS for 5 min at 37°C with 200 rpm rotation. An equivalent volume of wash buffer was added and cells were triturated at least 50 times with a p1000 pipette. Cells were spun at 600 g for 3 min. Cells were counted and used immediately for in vitro experiments or were frozen in 90% FBS + 10% DMSO freezing media and later thawed for some of the in vitro experiments.
Luciferase assays
Lysates used for luciferase assay were prepped from pulverized frozen whole glands, organoids, or fibroblasts in passive lysis buffer (Promega, Madison, WI, USA, E1941). After a 10-min incubation on ice, the lysates were centrifuged at 13,000 rpm at 4°C for 10 min to remove debris. Lysates were allowed to warm to room temperature before luciferase substrate (Promega, E148) was added. A GloMax 20/20 Luminometer (Promega) was used to read luciferase activity. Values were normalized to total protein determined by BCA assay (Pierce, Rockford, IL, USA, 23225).
Western blot analysis
Lysates used for western blot analysis were derived from whole glands snap frozen in liquid nitrogen immediately after dissection. Frozen glands were pulverized with a mortar and pestle followed by lysis in ice-cold RIPA buffer plus protease and phosphatase inhibitors (Pierce, 78440) for 10 min on ice. The lysates were then cleared by centrifugation at 10,000 rpm for 10 min at 4°C. BCA assay was used to determine lysate protein concentrations. Lysates were electrophoresed on 10% SDS-polyacrylamide gels and transferred onto PVDF membranes. Five percent milk in TBST was used for blocking and primary antibodies were diluted in 5% milk/TBST and incubated with the membrane for 2 h or overnight. Blots were probed with secondary HRP-conjugated antibodies (Jackson Immunologicals, West Grove, PA, USA) for 1 h. Phosphorylated primary antibodies were diluted in 5% BSA in TBST. Blots were developed using a GE Healthcare (Little Chalfont, UK) ImageQuant and ImageJ was used to calculate densitometry values.
Antibody concentrations
The following antibodies were used at the indicated dilutions for the specified applications. Western analysis: β-actin 1:5000 (Sigma-Aldrich, A5441); Cdc42 1:1000 (BD Transduction Laboratories, San Jose, CA, USA, 610929); phosphorylated MLC ser19 1:1000 (Cell Signaling, Beverly, MA, USA, 3671); phosphorylated ERK 1:1000 (Cell Signaling, 4370); total Erk 1:1000 (Cell Signaling, 4695); phosphorylated p38 1:1000 (Cell Signaling, 4511); β-tubulin (Sigma-Aldrich, T5201). IHC/IF: Ki67 1:5000 (Abcam, Cambridge, UK, ab15580); BrdU 1:1000 (Thermo Fisher Scientific, MA3-071); CC3 1:1000 (Cell Signaling, 9661); phosphorylated histone-H3 1:5000 (Merck Millipore, Darmstadt, Germany, 06-570); F4/80 1:50, no antigen retrieval (Invitrogen, MF48000); phosphorylated ERM 1:1000 (Cell Signaling, 3141); E-cadherin 1:250 (BD Transduction, 610181); K14 1:400 (Covance, Leeds, UK, PRB-155P); K8 1:250 (Developmental Studies Hybridoma Bank, TROMA-I). The K8 monoclonal antibody developed by Philippe Brulet and Rolf Kemler was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biology, Iowa City, IA, USA, 52242.
RNA isolation and qRT-PCR
RNA was isolated from control and Cdc42-associated fibroblasts from three mice per genotype pooled using Trizol and an RNeasy RNA purification column (Qiagen, Valencia, CA, USA, 91355) according to manufacturer’s recommendations. One μg of RNA was converted to cDNA using the RT2 First Strand Kit (Qiagen, 330401) and amplified using RT2 Profiler PCR Array Mouse Extracellular Matrix and Adhesion Molecules (PAMM-013A) per manufacturer’s instructions. Data were analyzed using the web-based software RT2 Profiler PCR Array Data Analysis from SABiosciences (Frederick, MD, USA). To validate gene expression changes identified by the array, cDNA was amplified using RT2 SYBR Green qPCR Master Mix (Qiagen, 330522), the StepOnePlus Real-Time PCR System (Applied Biosystems, Foster City, CA, USA), and the following primers: Col1a1 primers: 5′-GCT CCT CTT AGG GGC CAC T-3′ and 5′-CCA CGT CTC ACC ATT GGG G-3′; Fn1 primers: 5′-TTC AAG TGT GAT CCC CAT GAA G-3′ and 5′-CAG GTC TAC GGC AGT TGT CA-3′; Mmp2 primers: 5′-ACC TGA ACA CTT TCT ATG GCT G-3′ and 5′-CTT CCG CAT GGT CTC GAT G-3′; Mmp3 primers: 5′-ACA TGG AGA CTT TGT CCC TTT TG-3′ and 5′-TTG GCT GAG TGG TAG AGT CCC-3′; Mmp9 primers: 5′-CTG GAC AGC CAG ACA CTA AAG-3′ and 5′-CTC GCG GCA AGT CTT CAG AG-3′; and Gapdh primers: 5′-CCA ATG TGT CCG TCG TGG ATC-3′ and 5′-GTT GAA GTC GCA GGA GAC AAC-3′. The reaction was setup in triplicate and conditions were as follows: 95°C for 10 min then 40 cycles of 95°C for 15 s and 60°C for 1 min followed by a melting curve. Col1a1, Fn1, Mmp2, Mmp3, and Mmp9 mRNA levels were normalized to Gapdh mRNA levels and the data was analyzed using comparative CT.
Cdc42, RhoA, and Rac1 activity assays on isolated organoids
GLISA Cdc42 Activated Assay Biochem Kit (Cytoskeleton, Denver, CO, USA, BK127), GLISA RhoA Activation Assay Biochem Kit (Cytoskeleton, BK121) and Rac1 Activation Assay Biochem Kit (Cytoskeleton, BK126) were used to measure levels of activated Cdc42, RhoA, and Rac1 according to the manufacturer’s instructions. Mammary organoid lysates were prepared using the kit lysis buffer. Organoids isolated from two to five mice were pooled per group after 1 week and 3 weeks of dox treatment and the assays were run in triplicate. All lysates were prepared within 10 min prior to snap freezing.
Contraction assays
Primary MEC contractility was analyzed using the Cell Contraction Assay (Cell Biolabs, Inc., San Diego, CA, USA, CBA201) according to the manufacturer’s instructions. Growth media (MEGM BulletKit, Lonza, Walkersville, MD, USA, CC-3150) with 2 μg/ml dox was added once the gels solidified and changed when the gels were released and after each time point measurement. ROCK inhibitor, 25 μM Y27632 (Tocris Bioscience, Bristol, UK, 1254), or an equal volume of vehicle was added when the gels were released. Quantification of gel contraction was done using images of the gels taken immediately after their release and after 24 and 48 h post release to measure the difference in gel area from time of release. Imaging and quantification was done with a Zeiss Axioimager A1 epifluorescence microscope. Individual assays were conducted in duplicate or triplicate and averaged. Data without inhibitor are representative of four independent experiments and data with the ROCK inhibitor are representative of two independent experiments.
In vitromigration assays
Cryopreserved primary MECs were used for these studies. Approximately 500,000 MECs were plated onto a 6-cm dish and allowed to adhere to the plate and form characteristic epithelial cobblestone patches in MEGM Bullet Kit Media + dox. The media was replaced with serum-free F12 + dox and the cells were serum starved for 24 h. The cells were washed with PBS, trypsinized with 0.05% trypsin for 15 min and removed. Cells were then spun at 600 g for 3 min and resuspended in F12 media + dox and plated onto 8 μm-pore transwell filters (BD Transduction Laboratories) into 24-well plates (75,000 cells per well). Eight hundred μl of serum containing MEGM media was added to each well below the filter. The cells were allowed to migrate through the filter for 24 h at which time the upper surface of the filter was scraped twice with a cotton swab and media was suctioned off to remove any cells that did not migrate through the filter. The filters were then fixed in 4% PFA for 20 min and permeabilized for 10 min with 0.05% Triton-X 100 (Thermo Fisher Scientific, BP151). The filters were then removed from the well, transferred to a glass slide, and mounted with Vectashield + DAPI (Vector Laboratories, H-1200). A minimum of nine, 200× fields per filter were quantified and the total number of migrated cells was recorded per experiment. The fold changes of total migrated cells between control and Cdc42-overexpressing MECs were averaged from four independent experiments. Seven control mice and 11 Cdc42 overexpressing mice are represented in the data.
Three-dimensional (3D) culture assays
Primary MECs were isolated and plated on tissue culture plastic plates. MECs from at least three mice were pooled per group for each experiment. Plates were treated with 2% Matrigel containing MEGM media for at least 1 h at 37°C prior to plating of the cells. Cells were allowed to adhere to the plate and form characteristic epithelial cobblestone patches. After 48 to 72 h, the cells were washed with PBS, trypsinized with 0.05% trypsin for 15 min and removed. Cells were then spun at 600 g for 3 min and resuspended at 15,000 or 30,000 cells per well in 40 μl Matrigel per well of an 8-well chamber slide. The gel was allowed to solidify for 20 min at 37°C and 400 μl of warm MEGM + 2% Matrigel + 2 μg/ml dox was added to each well. The media was replaced every 3 days and the cultures were analyzed after 5 days using immunostaining and a Zeiss LSM 7 confocal microscope. Entire wells were quantified for each experiment. Invasive acini were defined as structures made up of five or more cells that had an invasive protrusion or at least one cell actively migrating away from the acinus. Data represent the average fold change between control and Cdc42-overexpressing MECs in three independent experiments. Dysmorphic acini were defined as acini with nonspherical morphologies with or without invasive protrusions or cells migrating away from the acinus. Data represent the average fold change between control and Cdc42 overexpressing acini in a total of three wells per group from three independent experiments.
For the spindle orientation three-dimensional culture assays, cryopreserved primary MECs were used and plated as described above. After 48 h, the cultures were fixed, immunostained with antibodies to tubulin and α6-integrin to identify the spindle and basal surface, respectively, and quantified using confocal microscopy. Acini were defined as structures with three or more clustered cells as previously described [
7], and the first 25 acini identified with a mitotic spindle were quantified. Data represent the percentages of structures with the spindle parallel, perpendicular, or angled relative to the basal surface of the forming acinus out of the total number of structures. Two independent experiments were performed and MECs from at least four mice were pooled per group for each experiment.
Three-dimensional culture immunofluorescence (IF) staining
Staining was performed using methods adapted from Debnath
et al., [
19]. Three-dimensional acini were fixed with 2% or 4% PFA for 20 min at room temperature, permeabilized with 0.5% Triton-X 100 in PBS for 10 min, washed in 7.5 mg/ml glycine (Thermo Fisher Scientific, G46-1) in PBS. IF buffer consisted of 7.7 mM NaN3, 0.1% BSA, 0.2% Triton X-100, and 0.05% Tween-20 in PBS. Invasive three-dimensional culture assay wells were stained with Alexa Fluor™ 488 phalloidin (Molecular Probes, Carlsbad, CA, USA, A12379) diluted 1:50 in IF buffer for 1 h and TO-PRO3 (Molecular Probes, T3605) diluted 1:200 in PBS for 10 to 20 min. Mitotic spindle orientation culture wells were stained overnight with α6-integrin (Millipore, MAB1378) diluted 1:200 in IF buffer + 10% goat serum (Sigma-Aldrich, G9023) and α-tubulin (Abcam, Ab1825) diluted 1:400 in IF buffer + 10% goat serum. Wells were stained with Alexa Fluor™ 488 goat anti-rat (Invitrogen, A11006) and Alexa Fluor™ 555 goat anti-rabbit (Molecular Probes, A21428) secondary antibodies diluted 1:200 in IF buffer + 10% goat serum for 1 h and TO-PRO3 diluted 1:200 in PBS for 10 to 20 min. All slides were mounted with 4 μl Vectashield + DAPI per well with coverslips and allowed to dry in the dark for 24 to 72 h before sealing coverslips with nail polish and imaging.
Flow cytometry analysis
Single MECs isolated as described above were suspended in 1 ml of PBS and fixed by adding 2.5 ml of 100% ethanol. Ethanol was added 500 μl at a time while gently vortexing to prevent clumping, and cells were fixed on ice for 15 min and stored at 4°C until analysis. Cells were pelleted by centrifugation at 600 g for 5 min and resuspended in propidium iodide (PI) staining solution (50 μg/ml PI; 0.1 mg/ml RNAse A; 0.05% Triton X-100) and incubated for 30 min in a 37°C water bath. The cells were transferred using a 26 G syringe through a cell strainer cap of a flow tube (BD Falcon, Franklin Lakes, NJ, USA, 352235) to break up clumps. At least 10,000 events were analyzed using a Beckman Coulter (Brea, CA, USA) FC500 Flow Analyzer for PI fluorescence intensity. MECs from two to three mice were pooled for each experiment. Data are representative of two independent experiments.
Statistical analysis
Unpaired Student’s t test was used for all statistical tests. P values less than 0.05 were considered significant. Error bars represent the standard error of the mean.
Discussion
Cdc42 is overexpressed and hyperactivated in human breast tumors [
2,
3]. However, the effects of Cdc42 overexpression on mammary gland development and the stochastic process of tumor formation
in vivo have not been previously investigated due to a lack of Cdc42-overexpressing mouse models. Here we describe the generation of a novel conditional Cdc42-overexpressing mouse model and the effects of Cdc42 overexpression during postnatal mammary gland development. We show that overexpression of Cdc42 disrupts TEB morphogenesis and promotes hyperbranching of the ductal tree in association with stromal alterations. Furthermore, our data suggest that increased MEC contractility and migration, potentially via upregulation of MAPK signaling, contributes to these phenotypes.
Although Cdc42 is overexpressed and hyperactivated in breast tumors [
2,
3], mutations in Cdc42 have not been found [
32]. Thus, regulation of Cdc42 protein translation, stability, and GTP hydrolysis and exchange rates likely underlie Cdc42’s altered expression and activity in tumors. When we designed the TetO-Cdc42 model, we aimed to accurately recapitulate this regulation, and therefore, chose to overexpress wild-type Cdc42 rather than a constitutively active mutant. Using this approach, we showed that Cdc42 and RhoA activities are differentially elevated in the mammary epithelium of the Cdc42-overexpressing mice during the developmental time course that we examined.
Cdc42 is upregulated early in MECs undergoing morphogenesis [
33], and it is essential for polarity establishment and proliferation during the early stages of MEC morphogenesis [
7,
34,
35]. These studies together with our data demonstrating that Cdc42 overexpression disrupts MEC acinus morphology in three-dimensional cultures and TEB architecture
in vivo, suggest that precise regulation of Cdc42 activity is important for proper MEC morphogenesis and branching. Normal MECs likely have multiple mechanisms to ensure that proper levels of Cdc42 activity are maintained. One potential mechanism could involve RhoGDI1, which functions to sequester and maintain Rho GTPases in their inactive state in the cytoplasm [
36]. Levels of RhoGDI1 are limited, and overexpression of one Rho GTPase can displace another family member, resulting in increased activation of the displaced family member [
37]. This mechanism could account for the absence of elevated Cdc42 activity and concomitant increase in RhoA that we observed at the early developmental time. At the mid-developmental time point after the early stages of MEC morphogenesis and initiation of branching are complete, precise control of Cdc42 activity may no longer be crucial and elevated Cdc42 activity may facilitate branch elongation by enhancing MEC migration and invasion. Increased stromal deposition affects Rho GTPase activity in MECs [
26,
38], and it is possible that the stromal alterations detected in the Cdc42-overexpressing mammary glands also impact RhoA and Cdc42 activation. Further studies will be required to define the mechanisms regulating Rho GTPase activity in normal and transformed MECs.
It is interesting to note that overexpression of Cdc42 and p190B RhoGAP in the developing postnatal mammary gland have similar phenotypes, including abnormal TEBs, hyperbranching, and increased stromal deposition [
17]. Similar to the Cdc42-overexpressing MECs, p190B-overexpressing MECs also display increased contractility [
18]. Thus, disruption of the Rho signaling network in MECs, leading to increased intracellular contractility may play an important role in driving these phenotypes. Indeed, actomyosin-mediated contractility regulates mammary organoid branching
in vitro[
9].
One distinction between the p190B- and Cdc42-overexpressing mammary gland phenotypes is that p190B-overexpressing MECs exhibited both increased migration and proliferation [
17,
39], whereas Cdc42-overexpressing MECs displayed only increased migration. We anticipated that Cdc42 overexpression might enhance proliferation as our published studies investigating the effects of Cdc42 knockout during primary MEC morphogenesis
in vitro demonstrated that Cdc42 is a critical regulator of MEC proliferation [
7]. Furthermore, mammary gland branching is driven by both MEC proliferation and migration [
9]. However, our studies suggest that Cdc42 overexpression facilitates aberrant branching morphogenesis by promoting increased MEC contractility and migration in the absence of any effects on proliferation.
The promigratory effects of Cdc42 overexpression in MECs are supported by the literature describing other Cdc42 gain of function models. A migration phenotype was reported in mouse embryonic fibroblasts isolated from Cdc42GAP knockout mice in which elevated Cdc42 activity disrupted directional migration [
40]. In addition, neutrophils isolated from the Cdc42GAP knockout mice had an increased ability to migrate, but the direction of migration was disrupted [
41]. Interestingly, MAPK signaling contributed to the migration phenotype in the Cdc42GAP knockout neutrophils, which showed changes in ERK and p38 phosphorylation that were similar to those detected in the Cdc42-overexpressing mammary glands. MAP kinases, including ERK, p38, and JNK, have been broadly implicated as regulators of cell proliferation and migration in diverse cell types in response to a variety of stimuli [
42‐
44]. Our detailed analysis of cell cycle progression and apoptosis in the Cdc42-overexpressing mammary glands did not reveal any alterations in cell proliferation or survival. Thus, elevated MAPK activity in the Cdc42-overexpressing mammary glands may regulate MEC migration and invasion to promote hyperbranching.
Disruption of epithelial architecture is an important hallmark of breast cancer initiation, it contributes to invasion and metastasis, and it can be used to help predict survival [
45‐
48]. The abnormal TEB morphologies detected in the Cdc42-overexpressing mammary glands together with our reported loss-of-function studies demonstrating the requirement for Cdc42 during the early stages of MEC acinus formation [
7], suggest that Cdc42 is a key regulator of mammary epithelial architecture. Thus, Cdc42 overexpression may cooperate with initiating oncogenes to facilitate the disruption of epithelial architecture during the early stages of tumorigenesis. The increased migratory and invasive capacity of the Cdc42-overexpressing MECs suggests that Cdc42 overexpression may facilitate mammary tumor cell invasion and metastasis
in vivo, and indeed, studies investigating the effects of Cdc42 knockdown in breast cancer xenografts have shown that Cdc42 regulates tumor cell invasion and metastasis
in vivo[
49]. Furthermore, an intriguing possibility is that Cdc42 overexpression may function during the early stages of tumor formation to induce protumorigenic and proinvasive stromal alterations. Future studies will be aimed at using this novel mouse model to determine the contribution of Cdc42 during different stages of tumor formation and progression and to define the molecular mechanisms by which aberrant Cdc42 expression facilitates these processes. A better understanding of how Cdc42 overexpression impacts the development and progression of breast cancer will help to pinpoint when targeting Cdc42 would be most effective and will define how best to target its aberrant actions.
Competing interests
The authors declare no competing interests.
Authors’ contributions
KB designed and performed experiments and analyzed data, including analysis of Cdc42 expression and signaling, mammary gland phenotypes, immunostaining, and three-dimensional culture experiments. KB also helped draft the manuscript. MG performed contraction assays, immunostaining, q-RT-PCR, and analyzed data. JY performed migration assays, GLISA assays, spindle analysis, analyzed mammary gland phenotypes, and analyzed and interpreted results. EL performed and analyzed immunostaining on tissue sections. MH and JCS generated and tested the construct to make the TetO-Cdc42-overexpressing mice. MH performed initial characterization to identify inducible Cdc42-overexpressing transgenic lines. TVG conceived of the study and designed and directed the experiments, analyzed data, and drafted the manuscript. All authors have read and approved the final manuscript.