Introduction
The serine/threonine kinase Akt is a critical downstream effector in multiple signal transduction pathways and regulates cellular proliferation, survival, and metabolism. Consistent with this, Akt is inappropriately activated in a wide range of human cancers [
1,
2]. Three Akt isoforms (Akt1, Akt2, and Akt3) are present in mammals, and targeted deletion of each gene has revealed distinct as well as overlapping functions in cellular physiology.
Akt1-/- mice exhibit increased perinatal mortality and a modest growth defect.
Akt2-/- mice are viable but develop insulin resistance and a diabetes-like phenotype, whereas
Akt3-/- mice have normal glucose homeostasis but decreased brain size [
3‐
7].
Given the high degree of homology among Akt isoforms, the possibility that these three proteins play redundant roles has been addressed by generating mice deficient for multiple isoforms. This has revealed that
Akt1-/-;
Akt2-/- mice display perinatal lethality, reduced growth, and defects in skin and bone development, as well as adipogenesis [
8].
Akt1-/-;
Akt3-/- mice die during embryonic development at embryonic day 12, and
Akt1-/-;
Akt3+/- mice exhibit developmental abnormalities in multiple organs that result in the death of 90% of mice shortly after birth [
9]. In contrast,
Akt2-/-;
Akt3-/- mice are born in Mendelian ratios, but are significantly smaller than wild-type littermates [
10]. These results indicate that individual Akt isoforms play both unique and overlapping roles in development and physiology, and further suggest that a critical threshold of Akt activity may be required to produce a given cellular output.
Similar to Akt, the Stat5 pathway plays a central role in regulating cellular function and is activated in response to a wide range of stimuli, including growth factors, cytokines, and hormones. Stat5 regulates cell proliferation, survival, and differentiation through direct activation of target genes. Stat5 is encoded by two closely related genes, Stat5a and Stat5b. While constitutive deletion of both genes leads to embryonic lethality [
11], conditional knockouts have revealed tissue-specific functions of Stat5 [
11‐
14].
The best-characterized role of Stat5 in normal physiology is the regulation of pregnancy-induced mammary epithelial development, where it is essential for the proliferation, differentiation, and survival of mammary epithelial cells [
11]. Analysis of a variety of Stat5 mutant alleles in mice has revealed that Stat5 deficiency in the mammary gland leads to a near-complete loss of lobuloalveolar development, a reduction in expression of milk protein genes during pregnancy, and lactation failure [
11,
15‐
18]. Stat5 deletion during pregnancy results in the death of differentiated mammary epithelial cells, with further experiments suggesting that the requirement for Stat5 is required for the survival of alveolar luminal progenitor cells during pregnancy-induced lobuloalveolar development [
11,
19].
Constitutive activation of Akt in the mammary epithelium promotes the precocious accumulation of intracellular lipid droplets during pregnancy and delays post-weaning involution by inhibiting apoptosis [
20‐
22]. Consistent with its upstream role as a negative regulator of Akt activity,
Pten-deficient mice exhibit delayed mammary involution and reduced apoptosis [
23], whereas forced expression of Pten in the mammary gland results in impaired lactation due to decreased mammary epithelial proliferation and increased apoptosis during pregnancy [
24].
We recently investigated the role of Akt in mammary development by examining mice bearing targeted deletions in either
Akt1 or
Akt2 [
25]. We found that loss of both alleles of
Akt1 results in failure of the coordinated metabolic response required for the establishment of lactation at parturition, including increased glucose uptake and lipid synthesis, which in turn results in decreased milk production. In contrast, deletion of both alleles of
Akt2 had no discernible effect on lactation. Notably, despite the requirement for Akt1 in the metabolic control of the lactating mammary gland, mammary epithelial differentiation, proliferation, and survival were unaffected during pregnancy in
Akt1-/- mice.
In light of the overlapping functions of Akt isoforms, we considered the possibility that Akt may play an essential role in regulating mammary epithelial development that is not evident in Akt1-/- mice due to compensation by other Akt isoforms. To address this issue, we interbred mice bearing targeted deletions in Akt1 and Akt2 in order to determine the effect on mammary differentiation. We find that deletion of one allele of Akt2 in Akt1-/- mice results in a severe defect in terminal mammary epithelial differentiation and lactation failure due to a loss of prolactin-mediated Stat5 activation. Notably, this defect occurred in the absence of changes in pregnancy-induced lobuloalveolar development, including proliferation, apoptosis or acinar formation. As such, the defects observed in Akt1-/-;Akt2+/- mice reflect the abrogation of Stat5 signaling and the molecular program of terminal differentiation in the mammary gland, a program that is intact in Akt1-deficient mice. Our observations demonstrate an unexpected requirement for Akt in Prlr-Jak-Stat5 signaling in the mammary gland, and establish Akt as an essential regulator of differentiation, metabolism and lactation in the mammary gland.
Materials and methods
Animals
Akt1-/- and
Akt2-/- mice of C57BL/6 genetic background were generated and provided by Dr Morris Birnbaum (Howard Hughes Medical Institute, University of Pennsylvania, Philadelphia, PA, USA) [
5,
6]. Mice were housed and maintained according to Institutional Animal Use and Care Committee guidelines. For timed pregnancies, the morning of the observed virginal plug was counted as day 0.5. At given time points, animals were killed by carbon dioxide asphyxiation and the mammary gland tissues were harvested and snap-frozen on dry ice, fixed in 4% paraformaldehyde in 1 × PBS (4% paraformaldehyde (PFA)) or frozen in Optimal Cutting Temperature (OCT) compound for further analysis.
The determination of pup weight, milk volume and pup mortality among various knockout mice has been described previously [
25,
26]. All experiments and experimental methods related to the use of animals were approved by the University of Pennsylvania Institutional Animal Use and Care Committee.
Antibodies
The following rabbit polyclonal antibodies were used in the study: phospho-S6 (Ser235/236), S6, and Akt (Cell Signaling Technology, Danvers, MA, USA), suppressor of cytokine signaling 2(Socs2) (Zymed Laboratories, South San Francisco, CA, USA), inhibitor of DNA binding 2 (Id2) (Santa Cruz Biotechnology, Santa Cruz, CA, USA), and mouse milk-specific proteins (Nordic Immunological Laboratories, Tilburg, Netherlands). The following mouse monoclonal antibodies were used: phospho-Stat5a/b (Tyr694/Tyr699) and Stat5a/b (Upstate, Lake Placid, NY, USA), β-tubulin (BioGenex, San Ramon, CA, USA), caveolin-1 (BD Biosciences, San Diego, CA, USA) and Gata-3 (Santa Cruz Biotechnology). The goat anti-Elf5 polyclonal antibody was purchased from Santa Cruz Biotechnology.
The rabbit anti-Npt2b polyclonal antibody was generously provided by Dr Jim Turner (National Institute of Dental and Craniofacial Research, National Institute of Health, Bethesda, MD, USA). The rat anti-cytokeratin 8 (CK8) was purchased from the Developmental Studies Hybridoma Bank (University of Iowa, Iowa City, IA, USA).
Mammary gland whole-mount and histological analysis
The abdominal mammary glands fixed in 4% PFA were either stained in carmine/aluminum potassium sulfate for whole mounts as previously described [
27] or embedded in paraffin wax. For histological analysis, 5 μm tissue sections were cut, dewaxed in xylene, rehydrated, and stained with H & E.
Immunofluorescence analysis
After paraffin-embedded 5 μm tissue sections were cleared and rehydrated, sections were subjected to antigen retrieval performed by heating treatment in an antigen unmasking solution (Vector Laboratories, Burlingame, CA, USA). Subsequently, sections were incubated in blocking solution consisting of 5% BSA and 10% (v/v) normal goat serum in PBS at room temperature for 1 hour. The primary antibodies phospho-Stat5 (1:100), Npt2b (1:300) and anti-CK8 (1:100) were then applied and incubated at 4°C overnight. Appropriate fluorescein-conjugated secondary antibodies (Molecular Probes, Eugene, OR, USA) were applied for 1 hour, counterstained with 5 mg/ml Hoechst 33258 (1:10,000), and mounted with Immu-mount (Thermo Scientific, Pittsburgh, PA, USA).
Stained sections were examined using a Leica microscope (Model DM5000B; Leica Microsystems, Bannockburn, IL, USA) equipped with a mercury lamp and FITC (L5), Texas Red (TX2) and Hoechst/DAPI (A4) filter cubes. Images were acquired by a digital camera (Leica DFC350FX) operated and analyzed with Image-Pro Express software and processed with the Photoshop program.
Northern blot, in situhybridization and quantitative RT-PCR analysis
Total RNA isolation from snap-frozen abdominal and inguinal mammary tissues without lymph nodes, preparation of radioactively labeled cDNA probes, northern blots and
in situ hybridization were performed as previously described [
28]. The cDNA probes for northern hybridization correspond to
β-casein (nucleotides 181 to 719), whey acidic protein (
WAP) (nucleotides 131 to 483),
ε-casein (nucleotides 83 to 637),
α-lactalbumin (nucleotides 174 to 700) and
CK18 (nucleotides 589 to 1287). The sequence of an oligomer used to detect 18S rRNA is CGGAACTACGACGGTATCTG.
Single-stranded cDNA for quantitative RT-PCR analysis was generated by the high-capacity cDNA reverse transcription kit (Applied Biosystems, Carlsbad, CA, USA). Quantitative RT-PCR was performed using the TaqMan-based ABI 7900HT fast real-time PCR system according to the manufacturer's instructions (Applied Biosystems). The probes used were Aldoc Mm01298111_g1, Fads1 Mm00507605_m1, Elovl5 Mm00506717_m1, Prlr Mm00599957_m1, and cytokeratin 18 Mm01601706_g1.
Western blot analysis
Frozen abdominal and inguinal mammary tissues (lymph nodes removed) were homogenized in lysis buffer (1% Triton X-100, 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM ethylenediamine tetraacetic acid (EDTA), 50 mM sodium fluoride (NaF), 3 mM sodium pyrophosphate, and 5 mM β-glycerolphosphate) supplemented with protease inhibitor cocktail (Roche, Indianapolis, IN, USA) and were subjected to western blot analysis, performed as described previously [
25]. The following primary antibodies were used: phospho-S6 (Ser235/236) (1:1,000), S6 (1:1,000), Akt (1:1,000), phospho-Stat5a/b (Tyr694/Tyr699) (1:500), Stat5a/b (1:500), mouse milk-specific proteins (1:20,000), β-tubulin (1:1,000), caveolin-1 (1:1,000), Socs2 (1:150), Id2 (1:100), and Elf5 (1:100). Chemiluminescence was detected with horseradish peroxidase-conjugated goat anti-rabbit or mouse secondary antibodies at a dilution of 1:5,000, and was developed in the ECL plus system based on the manufacturer's protocol (Amersham Biosciences, Piscataway, NJ, USA) followed by exposure to X-ray film (Kodak, Rochester, NY, USA). All experiments were independently repeated three times. Only the representative images are shown. Densitometry for western blots was carried out with the Photoshop program.
Lactose analysis
Lactose in the mammary gland was measured by subjecting 20 μg tissue lysates to the lactose assay kit as per the manufacturer's instructions (MBL, Woburn, MA, USA). The lactose level was calculated by subtracting the free galactose level from the total galactose level.
Mammary gland culture
The whole-organ culture of the mammary gland was as described previously but with slight modifications [
26]. Briefly, the lymph node-free abdominal and inguinal glands of
Akt1+/+;
Akt2+/+ mice and
Akt1-/-;
Akt2+/- mice at day 18.5 of pregnancy were aseptically collected and minced into fine pieces (~2 mm). Tissues were incubated in Waymouth's serum-free medium (Invitrogen, Carlsbad, CA, USA) supplemented with 20 mM HEPES, 4 mM glutamine, 5 μg/ml insulin, and 1 μg/ml hydrocortisone. The tissues were replaced with fresh culture medium daily to remove hormones carried from mice, and the mammary tissues were cultured for 5 days followed by growth-factor starvation overnight. The tissues treated with 0.2 μg/ml prolactin for the indicated periods were collected.
Statistical analysis
Data are presented as the mean ± standard error of the mean (SEM). Statistical analysis was calculated by Student's t test unless otherwise indicated.
Discussion
The prolactin-Jak-Stat5 pathway has long been recognized as a central mediator of pregnancy-induced lobuloalveolar development and lactation, which together constitute a developmental transition that is essential for the survival of mammals. Accordingly, the role of this pathway in mammary development has been intensively studied. In the present manuscript we describe a previously unrecognized requirement for Akt in Prlr-Jak-Stat5 signaling. Mice lacking one allele of Akt2 and both alleles of Akt1 displayed a severe lactation defect due to the global impairment of alveolar epithelial cell differentiation. Consistent with their failure to terminally differentiate, pregnant Akt1-/-;Akt2+/- mice fail to upregulate Npt2b or phospho-Stat5a/b and display markedly reduced synthesis of each of the three major components of milk during lactation. Notably, epithelial cell proliferation, cell survival and the formation of architecturally normal acini during pregnancy were unaffected by Akt deletion, reinforcing that the lactation defect observed in Akt-/-;Akt2+/- mice results from a defect in differentiation, rather than from a failure to form acinar structures. In aggregate, these findings establish an essential but heretofore unrecognized role for Akt in epithelial differentiation.
Despite the fact that both
Akt1-/-;
Akt2+/+ mice and
Akt1-/-;
Akt2+/- mice exhibit defects in lactation, the molecular basis of their lactation phenotypes is strikingly different. The isoform-specific defect in lactation observed in
Akt1-deficient mice occurs in the absence of defects in differentiation and results from a defect in metabolism [
25]. This metabolic defect is due to the failure of
Akt1-deficient mammary epithelial cells to upregulate key Akt1 target genes, most notably the glucose transporter Glut1 and three lipid synthetic enzymes
Acly,
Scd2, and
Scd3. These molecular defects result in a profound inability of terminally differentiated mammary epithelial cells to take up glucose or to synthesize normal amounts of lipid. Nevertheless, mammary epithelial differentiation is normal in lactating
Akt1-/- mice, as demonstrated by physiologically normal levels of Stat5 activation, normal upregulation of the terminal differentiation marker Npt2b, normal downregulation of the virgin-specific transporter NKCC1, normal expression of all major milk proteins, and normal intraepithelial lactose levels.
In contrast, our current study demonstrates that deletion of one allele of Akt2 in Akt1-deficient mice results in a severe defect in mammary epithelial differentiation that is due to a failure to activate Stat5. In contrast to Akt1-deficient mice, late-pregnant Akt1-/-;Akt2+/- mice exhibit dramatically reduced Prlr-Jak-Stat5 signaling, as well as markedly reduced milk protein expression, lactose levels, lipid synthesis, expression of the terminal differentiation marker Npt2b, and mTOR activity.
The lactation defect observed in Akt1-/- mice is thus due to a metabolic defect that results from a failure to upregulate Glut1 and other Akt1-specific target genes. This defect occurs in the context of normal Prlr-Jak-Stat5 signaling and normal mammary epithelial differentiation. In contrast, the lactation defect in Akt1-/-;Akt2+/- mice is due to a profound defect in mammary epithelial differentiation that results from a failure to activate Stat5. While the outward phenotypes (that is, lactation defect) of Akt1-/- mice and Akt1-/-;Akt2+/- mice are similar at a superficial level, the molecular phenotypes as well as the molecular basis for these phenotypes are profoundly different.
The defect in Stat5 activation observed in Akt1-/-;Akt2+/- mice is due, at least in part, to a failure to upregulate the positive regulator of Prlr-Jak-Stat5 signaling, Id2, or to downregulate the negative regulators of prolactin-Jak2-Stat5 signaling, caveolin-1 and Socs2. In addition, our findings suggest that Akt most likely regulates the expression or activity of other molecules that modulate Prlr-Jak-Stat5 pathway activity. Together, these findings provide a molecular basis for this previously unrecognized connection between the Akt and Stat5 pathways.
Notably, mammary epithelial proliferation and apoptosis rates were unaffected in Akt1-/-;Akt2+/- mice during pregnancy, suggesting that Akt is essential for Stat5-dependent secretory differentiation of mammary epithelium, but possibly not for Stat5-dependent alveolar development or acinar formation; that is, whereas Stat5-deficiency in the mammary gland results in a failure of lobuloalveolar development as well as secretory differentiation, Akt1-/-;Akt2+/- mice exhibit only a defect in secretory differentiation. Nevertheless, it is possible that deletion of all four Akt1 and Akt2 alleles would result in a defect in prolactin-Stat5-mediated epithelial proliferation and, thereby, lobuloalveolar development similar to that observed in Stat5-deficient mice. Given the perinatal lethality of combined germline deletion of Akt1 and Akt2, however, mammary-specific deletion of these genes may be required to determine the role of the remaining Akt2 allele in mammary epithelial proliferation and differentiation.
Elf5 has been shown recently to regulate alveolar cell differentiation by acting downstream of the prolactin receptor [
41]. Since Elf5 expression was unchanged in the mammary glands of
Akt1-/-;
Akt2+/- mice despite their dramatic reduction in Stat5 activity, our data suggest that the impact of Akt deletion on Prlr-Jak-Stat signaling is not mediated by Elf5. Rather, our findings suggest that Akt and Elf5 may act via parallel pathways, that Elf5 alone is not sufficient to compensate for loss of Prlr signaling, and that factors other than Prlr signaling may regulate Elf5 [
19,
41].
Maroulakou and colleagues have reported that
Akt1 deficiency delayed, whereas
Akt2 deficiency facilitated, mammary epithelial differentiation during pregnancy and lactation [
42]. In contrast, consistent with our prior study, the findings described here confirm that deletion of either
Akt1 or
Akt2 alone has no appreciable effect on secretory differentiation during pregnancy or lactation. Moreover, our finding that deletion of one allele of
Akt2 in
Akt1-deficient mice results in a pronounced defect in mammary epithelial differentiation that is not observed in mice deficient for Akt1 alone strongly suggests that Akt2 synergizes with - rather than antagonizes - the pro-differentiation effects of Akt1 on mammary epithelial cells. Whether this discrepancy is explained by the mixed genetic background of mice used in the former study or other factors remains to be determined.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
CCC participated in the design, execution and analysis of experiments and participated in drafting the manuscript. RBB participated in the execution and analysis of mouse experiments. DBS participated in the execution and analysis of immunofluorescence experiments. CPP participated in the execution of mouse experiments. RHH participated in the execution of mouse and molecular experiments. JVA participated in the analysis of data and drafting of the manuscript. MJB participated in the mouse experiments and drafting of the manuscript. LAC participated in the design and analysis of experiments and drafting of the manuscript.