Background
The serine/threonine kinase Akt, also known as Protein Kinase B (PKB) plays a key role in multiple cellular processes. To date, three human isoforms of Akt have been identified: Akt1/PKBα, Akt2/PKBβ and Akt3/PKBγ [
1]. Akt1 has been shown to play a key role in cellular survival pathways by inhibiting apoptotic processes, inducing cell proliferation, and protein synthesis [
1‐
3]. The PKB/Akt pathway can be activated by a variety of growth factors, hormones, and cytokines [
4,
5] and can participate in oncogenic transformation of mammalian cells through its regulation of key biological process [
6,
7]. Akt is frequently constitutively active in many types of human cancers, such as breast, ovarian, and prostate [
8‐
10]. It has been demonstrated by our group and others that activated Akt1 in tumor cells are associated with high grade tumors, late stage of diagnosis, and poor outcome in patients [
9,
10]. Cell lines derived from breast cancer patients who were resistant to trastuzumab treatment have also shown upregulation of Akt [
11]. Furthermore, patients who have poor outcome and resistance to endocrine therapy have also been found to have activated Akt1 in their tumor cells [
9]. We have recently demonstrated that activation of Akt1 in breast tumor cells
in vitro, leads to inactivation of FOXO1 and a decrease in response or resistance to trastuzumab-induced growth inhibition in HER2 overexpressing breast cancer cells [
12]. Significant data from
in vivo and
in vitro studies suggest that activation of Akt regulates cell survival signaling and plays a key role in tumorigenesis. Although a few
in vivo studies have explored the role of Akt in normal mammary development and tumorigenesis using transgenic mouse models [
13‐
16], the mechanisms associated with the oncogenic role of Akt remains to be further elucidated.
In this study, a transgenic mouse model with activated Akt was developed, such that the in vivo mechanisms by which Akt induces tumor development and progression in the mammary gland can be better understood. This transgene model also explores the mechanisms by which a carcinogen, DMBA, may further enhance the induction of Akt1 driven mammary gland tumors in virgin and post-lactation mice. In addition, this MMTVmyr-Akt1 transgenic model could serve as a preclinical model for studying ER-positive breast cancers that become resistant to endocrine therapy and develop into metastatic disease. Specifically, this model could be used to develop novel target therapies in breast cancer research and treatment.
Methods
This study was approved by Institutional Animal Care and Use Committee (IACUC) at Charles R. Drew University of Medicine and Science.
Generation of transgenic mice
The DNA structures of constitutively active Akt1 (myr-Akt1) (#21-151) were obtained from the Upstate Biotechnology Company. The activation of Akt1 was accomplished by the presence of the 11 N-terminal amino acids of avian c-
src that were required for protein myristoylation at the amino terminus of Akt1. The promoter in the original myr-Akt1 DNA structure (#21-151, Upstate Biotechnology) was a human cytomegalovirus immediate-early promoter (CMV). To create a mammary tissue specific mouse transgene, the CMV promoter was replaced by the MMTV-LTR (American Type Culture Collection). The MMTV-LTR promoter was inserted into plasmid DNA containing myr-Akt1 sequences (Additional file
1: Figure S1A). The construct was digested and the fragments containing the MMTV-LTR promoter, myr-Akt1 sequences, and SV40 polyadenylation signal were isolated and purified (Additional file
1: Figure S1B). The transgenic mice were generated by microinjection of DNA directly into one of the pronuclei of C57BL/6 mouse-fertilized zygotes at the University of California, at Los Angeles (UCLA) transgenic facility. The genotyping of the transgenic MMTVmyr-Akt1 mice was performed by polymerase chain reaction (PCR) with genomic DNA extracted from the tails of mice. Integration of the transgene and the transgene copy number were determined by real-time quantitative PCR [
17,
18]. The standard curve was generated by mixing non-transgenic tail DNA with transgene plasmid DNA. The single copy transgene was determined using the calculation method provided by the University of Michigan Transgenic Animal Model Core facility available at
http://www.med.umich.edu/tamc. An example of the copy number standard curve is shown in Additional file
1: Figure S1C. The Ct numbers from the genomic DNA of positive transgenic and wild-type (WT) litters were filled into a linear regression model and the copy number of transgenes was estimated. As shown in Additional file
1: Figure S1D, myr-Akt1+/+was defined as mice with 2 copy numbers of the transgene, such as mice 2-11 and 2-23. Myr-Akt1+/- was defined as mice with 1 copy number of the transgene, such as mice 2-10, 2-12, and 2-33. Mice 2-14 and 2-21 were transgene negative.
DMBA Treatment
The carcinogen 7, 12 dimethyl-1,2-benzanthracene (DMBA, Sigma) was dissolved in olive oil at a concentration of 10 mg/ml. This mixture was heated at 37°C and shaken vigorously to fully dissolve the DMBA. Virgin and post-lactation female transgenic (MMTVmyr-Akt1+) and wild-type (WT) mice were treated with 1 mg doses of DMBA via oral gavage weekly for a total of 6 weeks. Mice were maintained in the absence of males and were checked by palpation for tumor formation after completion of the DMBA treatments. The mice were sacrificed by CO2 inhalation either when tumors reached 225 mm2 or if the mice became moribund either during or at the end of the 34 week observation period. As indicated before, this study was approved by Institutional Animal Care and Use Committee (IACUC) at Charles R. Drew University of Medicine and Science.
Histopathology and immunohistochemistry (IHC)
The mammary glands, tumor tissues, and other organs from the mice were formalin-fixed and embedded in paraffin. The tissue sections were stained with either hematoxylin and eosin (H&E) or underwent immunohistochemistry (IHC) with specific antibodies. The antibodies used were as follows: anti-phospho-Akt1(ser473) (Cell Signaling, MA), anti-pan-keratins (pan-CKs), anti-ERα, anti-EGFR (Santa Cruz Biotechnology, CA), and anti-ERBB2 (Vector Laboratories, Inc, CA). The dilution of each antibody was according to the manufactures’ instructions, and positive staining was detected using diaminobenzidene (DAB) (Vector Lab, CA) according to the manufacturer's instructions. The data was independently reviewed by two pathologists.
Reverse transcription-PCR (RT-PCR) and quantitative real-time PCR (Q-PCR)
The total RNA from tumor tissues or organs was isolated using the RNeasy micro kit (#74004, QIAGEN) according to the manufacturer’s instructions. Subsequently, cDNA was synthesized by reverse transcription (RT) with ThermoScript™ RT-PCR system (Invitrogen) per the manufacturer’s instructions and followed by PCR. The primers used were as follows: AKT: 5’TCCTCAAGAAGGAAGTCATCGT-3’ (forward) and 5’-CGTACTCCATGACAAAGCAGAG3’ (reverse); myr-Akt1: 5’- ACCACTTGTCTCACATCCTTGTT -3’ (forward) and 5’TTCTAGACTTGGGCTTGCTCTT-3’ (reverse); β-actin: 5’- GTCTTCCCCTCCATCGT-3’ (forward) and 5’- CGTACATGGCTGGGGTGT-3’ (reverse). The PCR products were separated on 2% agarose gels with ethidium bromide. The final results for each gene were quantified using GelQuant.NET software (BiocheLabSolution.com) and adjusted for β-actin. Q-PCR analysis was performed with iCycle iQ real-time PCR detection system (Bio-Rad Lab, Hercules, CA) using SYBR Green Master Mix (#204143, QIAGEN). The primers used were as follows: SLUG: 5- 5’-AGAGCATTTGCAGACAGGTCA-3’ (forward) and 5’AGCAGCCAGATTCCTCATGTT-3’ (reverse), 18S, 5'-GATCCATTGGAGGGCAAGTC-3' (forward) and 5'-TCCCAAGATCCAACTACGAG-3' (reverse). The mRNA levels of SLUG were quantified by measuring the threshold cycle (Ct) and were adjusted for the level of 18S.
Mammary gland whole mounts
Mammary glands near the hind legs were harvested from WT and transgenic mice at the indicated times and fixed in 10% formalin overnight. The fixed mammary glands were then placed in 75% ethanol for 2 hours and placed in acetone overnight at -20°C. The tissues were then rehydrated with successive incubation with different concentrations of ethanol, stained with hematoxylin for 1.5 hours, and then rinsed in crude tap water overnight. Tissues were then placed in 50% ethanol with 12 N of HCl for 30 minutes for destaining followed again with successive incubation in graded ethanol for dehydration. Subsequently, the tissues were incubated in xylene, then mounted in Permount (Fisher Scientific). Under 10x magnification, the total number of buds (terminal end buds and alveoli buds) were counted and divided by the total number of ducts as follows: pre-puberty and pubertal mice 2-4 ducts, 12-week virgin mice 7-11 ducts, 16-week and 26-week mice 11-17 ducts. Two mice per condition and per genotype were used for analysis.
Western blot analysis
The total protein from snap-frozen tumor tissues were extracted by homogenization using the Powergen Homogenizer 125 (Fisher Scientific) in 1 ml extraction buffer (25 mM Tris-HCl pH 7.6, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS and 10 μl Protease Inhibitor Cocktail - Cat#78410, Thermo Scientific). Homogenates were centrifuged at 1000 g for 10 minutes at 4°C, and 80 μg of proteins for each sample were used for Western Blot analysis. The following antibodies were used: anti-phospho-Akt1 (ser473), anti-AKT, anti-Cyclin D1 (Cell Signaling, MA); anti-ERα, anti-CK18, anti-Twist, and anti-β-actin (Santa Cruz Biotechnology, CA).
Measurement of estradiol levels
Blood was collected from both WT and transgenic mice from each of the following conditions: 3 month-old virgin mice, ~4 month-old lactating mice 3 days after giving birth, and ~ 5 month-old post-lactating mice one month after giving birth. The blood samples were collected immediately after euthanization (two mice for each group). Serum estradiol (E2) levels were measured by ELISA assay (Cat# ES180S-100, Calbiotech Inc, CA) per the manufactures’ instructions. The E2 level was determined based on the principle of competitive binding between E2 in serum samples and E2 enzyme conjugate for a constant amount of anti-estradiol antibody. The sensitivity of the assay is <3 pg/ml.
Statistical analysis
Statistical analysis was performed with the SPSS statistical package (Version 11.0 for windows, IBM). Kaplan-Meier survival curve analysis with the log-rank test was used to determine the tumor-free survival between transgenic and wild-type mice treated with DMBA. The Spearman rank test was used to examine the differential incidences of different tumors types between different groups of DMBA treated mice. P-values < 0.05 were considered statistically significant.
Discussion
The goal of this study was to examine the oncogenic role of Akt in the development of mammary tumors. Myristoylation of Akt is a well-accepted mechanism to produce constitutively activated Akt in cells. Hence, we selected an established DNA structure of constitutively active Akt1 (myr-Akt1) that was generated by Upstate Biotechnology and tested
in vitro in our previous study [
12]. Myr-Akt1 was driven by MMTV-LTR to target the expression of the transgene specifically to the mammary gland tissue. Similar strategies have been used by other groups to develop Akt1-transgenic mouse models [
13,
15].
The myr-Akt1 transgene was expressed most significantly in the mice mammary glands in our model. This is consistent with the findings that there is an increase in the MMTV-LTR transcriptional activity throughout mammary development and during pregnancy [
21]. Increased phosphorylation of Akt at ser473 was detected in the mammary glands of 9 week old transgenic mice in the model developed by Blanco-Aparicio and colleagues [
15]. Similarly, we detected phosphorylated Akt at ser473 in our transgenic mice at 12 weeks old. Myr-Akt1 was also detected in the brain, salivary glands, ovaries, and uterus in the model developed by Blano-Aparicio group. We also found expression of myr-Akt1 in the ovaries and uterus of our transgenic mice. We did not examine expression in the brain or salivary glands in our study. In addition, the mRNA level of total AKT was higher in most organs from our transgenic mice compared to wild-type mice examined at pre-puberty, puberty, during lactation, and at post-lactation stages.
In our model, a significantly high level of circulating estradiol remained in the transgenic mice at during the later stages of lactation and decreased at the post-lactation stage. Overall transgenic mice had higher estradiol levels compared to WT mice. Estradiol is mainly produced by the granulosa cells of the ovaries. Under normal physiological conditions during pregnancy, circulating estrogen increases and decreases during the lactation period. However, activation of Akt1 in the ovaries may increase the proliferation of granulosa cells and lead to an increased production of estradiol even in the later-lactation stage. The higher circulating estrogen levels could further induce higher AKT expression. The increased estradiol by the activation of Akt1 in the ovaries could partially contribute to the mammary tumorigenesis in the transgenic mice in our model. Further studies examining mice with removed ovaries need to be conducted for clarification. However, the mammary tumorigenesis is more likely to be driven by the expression of myr-Akt1 in the mammary glands since we did not observe any ovarian or uterine cancers in our model. In addition, local estrogen biosynthesis such through Akt1 mediated aromatase expression in the mammary glands or myr-Akt1 induction of prolactin [
22‐
24] could all contribute to mammary tumorigenesis. High levels of prolactin have been associated with postmenopausal breast cancer risk [
25].
In our model, the transgene was maintained in C57BL6 mice which are genetically resistant to spontaneous mammary tumor formation [
15,
26] in order to reduce background tumor activity in the wild-type mice. Therefore, there is no spontaneous tumor formation in the wild-type mice in this study.
The role of Akt1 in mammary gland development and function has been examined in the postnatal stage of mice. The activation of Akt1 or overexpression of total AKT has been shown to delay mammary gland involution, and to induce hyperplasia in similar Akt1-transgenic mouse models [
13,
14]. In this study, we further examined the role of Akt1 in mammary gland development from pre-puberty, puberty, post-puberty (virgin), lactating to post-lactation stages. Our data showed for the first time that myr-Akt1 played a crucial role in the development of the early adult mammary gland. Myr-Akt1+ mice had mammary ductal outgrowths though the fat pads and extending alveolar epithelium during the post-puberty stage. Consistent with other Akt1 transgenic models, a delayed mammary gland involution was observed in our model as well. The activation of Akt has been shown to accelerate mammary tumorigenesis in MMTV-c-ErbB2 mice [
27]. However, overexpression or activation of Akt1 in the mammary glands alone is not enough to induce dysplasia or neoplasia [
28‐
31]. Activation of Akt1 induces oncogenic transformation upon exposure to carcinogens [
15]. Similarly in our model, we demonstrated that exposure to the carcinogen DMBA in transgenic mice with active Akt1 expression led to a significantly higher incidence of mammary carcinoma. In addition we compared tumor formation between transgenic mice and WT mice at different ages.
Epidemiological studies have suggested that lactation may have some protective effect on breast cancer risk in humans [
29‐
31]. Pregnancy and breast feeding have been associated with a reduction of lifetime risk of breast cancer among premenopausal women and women with family history of breast cancer [
32‐
34]. Mechanism studies using mouse models suggest that the secretory mammae and low rate of DNA synthesis in lactating mammary glands may account for the elimination of carcinogens [
35‐
37]. Studies have found that lactating mice were relatively refractory to the effects of chemical carcinogens. Conversely, the levels of estrogen and prolactin are increased during pregnancy and lactation, and the high levels of estrogen and prolactin could be associated with the development of breast cancer [
25,
35]. Normally, the mammary gland regresses to its pre-pregnancy status by a coordinated process of apoptosis, remodeling, and inflammation following pregnancy and lactation [
38]. Activation of Akt in mammary tissue could result in de-regulation of the survival signal [
31]. The normal mammary gland involution after pregnancy and lactation could be interrupted by active Akt1 expression in mammary tissue. The irregular involution of the mammary gland might increase susceptibility to carcinogen-induced tumor formation. Data from our current study demonstrated that the mice carrying the myr-Akt1 transgene in the mammary glands had accelerated carcinogen-induced tumorigenesis during the virgin and post-lactation stages. The protection against carcinogen-induced mammary malignancy by lactation was eliminated by expressing active Akt1 in the mammary tissues of the transgenic mice. We found that there was no DMBA induced tumor formation in the wild-type mice throughout pregnancy and lactation. However, DMBA induced mammary carcinomas were observed in 45.5% of mice with the MMTVmyr-Akt1 transgene.
Our studied identified that the DMBA induced mammary carcinomas in both transgenic and wild-type mice were poorly differentiated. However, the tumor tissues from transgenic mice had increased phosphorylated Akt1 at ser473 and the tumors showed strong positive expression of nuclear ERα. These findings are similar to data reported from another group [
15]. The tumors from our model were also positive for EGFR expression. Moreover, the tumors with low ERα expression concurrently had membranous expression of EGFR. ERα positive tumors with increased pAkt were associated with strong cytoplasmic expression levels of EGFR. However, the function of cytoplasmic expression of EGFR is unclear, although both membranous and cytoplasmic expression of EGFR has been reported before in cancers such as pancreatic cancer. Tissues resected from pancreatic cancer patients with EGFR over-expression in the cell membrane showed worse clinical outcomes compared to patient tissues with low EGFR expression [
39]. Other studies demonstrated that EGFR expression in the membrane is more strongly associated with malignant tumors, while cytoplasmic EGFR expression is associated with normal cells or benign tumors [
40,
41]. Cytoplasmic EGFR over-expression was not significantly associated with recurrence or survival in our study.
It has been suggested that the activation of Akt might cooperate with ERα activity in the development of mammary tumors induced by DMBA [
15]. The observations from our current study are in agreement with this Akt - ERα cooperation concept. In addition, our data implies that the Akt-ERα cooperation may also have interactions with the EGFR pathway. Cyclin D1 plays role in tumor cell proliferation and is inversely correlated with tumor size in human breast cancer [
42]. Overexpression of Cyclin D1 has been reported in 40% to 90% of invasive breast cancers, while gene amplification is seen in about 5–20% of tumors [
43]. Increase in the protein levels of Cyclin D1 is apparent in hyperplasias and is further increased in malignancies [
43,
44]. Cyclin D1 has also been shown as an essential oncogenic intermediary for the
Neu pathway by inducing mammary carcinoma in transgenic mice [
28,
45]. In our model, we observed that Cyclin D1 expression was higher in DMBA-induced myr-Akt1 driven mammary carcinomas. However, Cyclin D1 expression was not simply associated with activation of Akt1. It was also not essential for ER-positive tumor formation since the levels of Cyclin D1 varied in tumors from the transgenic mice. Similar observations were reported by Blanco-Aparicio et al. [
15].
In humans, more than 60% of breast cancer is ERα-positive. Endocrine therapy, tamoxifen, is the most important systemic treatment of ER-positive breast cancer. Unfortunately, patients with advanced ER-positive disease eventually acquire resistance to tamoxifen treatment. The MMTVmyr-Akt1 transgenic mice tumor formation induced by carcinogens could mimic ER-positive human breast cancer, as shown with increases in pAkt, Cyclin D1, and EGFR. Furthermore, the transgenic mice in our model exhibited characteristics of advanced ER-positive mammary tumors with bone metastasis and increased expression of the mesenchymal markers, twist and slug. The data from this study showed that mammary tumors with metastatic characteristics were associated with increasing expression of twist and slug. The bone metastatic lesions were also positive for ERα. Therefore, the MMTV-myr-Akt1 transgenic mice could be useful for exploring the underlying mechanisms of ER-positive breast cancer acquiring resistance to tamoxifen and developing metastatic disease. EGFR
-targeted therapy, such as cetuximab and lapatinib in combination with hormonal therapy may also be explored as methods to overcome resistance. Patients with ERα-negative or triple negative breast tumors frequently demonstrate increased EGFR expression [
46,
47]. These patients are more likely to have poor disease outcomes. Similar to the aforementioned patient-based studies, we observed that wild-type mice with mammary tumors from DMBA induction showed an inverse relationship between nuclear expression of ERα and membranous EGFR. Thus, the carcinogen-induced mammary tumors from virgin mice in this study might also provide an
in vivo model for better understanding the etiology and tumor biology of this type of breast cancer. From a translational perspective, this will facilitate the design of better target therapies for treating ER/HER2-/EGFR + breast cancer.
Competing interests
The authors declare that they have no financial and non-financial competing interests.
Authors’ contributions
JVV made contributions to conception, design, overall guidance for the study, data interpretation and manuscript editing, revising for important content; YW made contributions to design the transgene and experiments; data acquisition, analysis, interpretation, and drafting the manuscript; JK made contributions to acquisition, analysis of the data and also involve the manuscript preparation, MS made contributions to the preparation of the manuscript; and YE made contributions to analysis and interpretation of the data, as well as preparation of the manuscript. All authors read and approved the final version of the manuscript to be published.