By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
This article is a US Government work and, as such, is in the public domain in the United States of America.
Signal transducer and activator of transcription 5 and 3 (Stat5 and Stat3) control pregnancy-mediated mammary development and involution-dependent remodeling, respectively. Suppressor of cytokine signaling 3 (Socs3) has been implicated in the modulation of both Stat3 and Stat5 activity. To explore the biology of Socs3 in mammary tissue, the gene was deleted using Cre-mediated recombination. Deletion of the Socs3 gene from mammary stem or early progenitor cells did not grossly alter pregnancy-mediated mammary development but resulted in impaired lactation due to attenuated proliferation. Loss of Socs3 from differentiated luminal cells did not interfere with glandular function during lactation, but resulted in accelerated tissue remodeling upon weaning. Loss of Socs3 led to enhanced and precocious Stat3 activation. Thus, Socs3 serves as a modulator of Stat3 activity to ensure controlled proliferation and apoptosis in pregnancy and involution, respectively. Developmental Dynamics 236:654–661, 2007. Published 2007 Wiley-Liss, Inc.
Establishment, function, and death of the mammary alveolar compartment are controlled by cytokines that activate defined transcription factors belonging to the family of STAT proteins (Hennighausen and Robinson,2005). Placental lactogens and prolactin activate the transcription factors Stat5a and Stat5b (Liu et al.,1996), which are essential for alveolar proliferation and differentiation during pregnancy (Liu et al.,1997; Miyoshi et al.,2001; Cui et al.,2004). Moreover, Stat5a/b are required for the maintenance of alveolar function during pregnancy (Cui et al.,2004). Upon weaning, and concomitant with the decline in prolactin signaling, LIF and IL-6 activate Stat3, which results in the death and remodeling of the mammary alveolar compartment (Chapman et al.,1999; Humphreys et al.,2002; Zhao et al.,2002,2004). Notably, inactivation of the Stat3 gene during lactation results in a failure of the gland to undergo scheduled involution (Chapman et al.,1999; Humphreys et al.,2002).
Stat3 and Stat5a/b are present throughout mammary development in their latent forms (Liu et al.,1996) and their activity is confined to specific periods of the gland's developmental cycles. Members of the SOCS family of proteins control cytokine signaling through the inhibition of STAT activation (Yoshimura et al.,2005). Ectopic expression of various SOCS proteins is able to blunt signaling by cytokines that activate Stat3 and Stat5 (Sasaki et al.,2000). Specifically, ectopic expression of cish in transgenic mice results in impaired mammary function, possibly due to compromised prolatin/Stat5 signaling (Matsumoto et al.,1999). However, the bona fide function of SOCS proteins, as judged by gene knock-out studies, appears to be more selective. In mammary tissue, Socs1 negatively regulates prolactin signaling and deletion of the respective gene results in precocious mammary development and a rescue of prolactin receptor (PrlR) haploinsuffiency (Lindeman et al.,2001). The inability of PrlR-hemizygous mice to form a functional alveolar compartment and lactate was also rescued by the deletion of both Socs2 alleles (Harris et al.,2006). In both instances, this functional rescue was paralleled by the activation of Stat5. On the other hand, targeted over-expression of Socs1 in the late pregnant and lactating mouse mammary gland appeared to have no adverse effect (Petridou et al.,2003). Moreover, Socs2 was identified in a microarray screen as a prolactin-responsive gene and expression was detected in the epithelial compartment throughout pregnancy (Harris et al.,2006). These studies demonstrate that both Socs1 and Socs2 function as negative regulators of prolactin signaling. However, since Socs2 is also a negative regulator of the growth hormone receptor, it is possible that the rescue is in part the result of GH action on mammary epithelium, which has been shown to activate Stat5 (Gallego et al.,2001).
Although Socs3 is known to negatively regulate Stat3 signaling induced by the gp130 receptor, it has also been linked to Stat5 signaling. In vitro experiments demonstrated that Socs3 can inhibit PRL induction of milk protein gene expression and Stat5 activation, probably through its direct interaction with the PrlR (Dif et al.,2001). Transcription of the Socs3 gene is stimulated by various cytokines, including GH and PRL, through Stat3 and Stat5. Low level expression of Socs3 is found throughout puberty and pregnancy. It increases sharply on day one of lactation and remains active throughout involution (Le Provost et al.,2005) suggesting a physiological significance of this negative regulator. In another study, abundant Socs3 mRNA was detected in the first 12 days of pregnancy, followed by a sharp decline throughout the remainder of pregnancy and a reappearance during involution (Tonko-Geymayer et al.,2002).
Studies aimed at understanding the involvement of Socs3 in physiological processes are complicated because both the deletion as well as the unscheduled expression of the respective gene led to embryonic lethality (Marine et al.,1999; Roberts et al.,2001). To assess the role of Socs3 in mammary epithelium at different stages of differentiation, we deleted the floxed Socs3 gene using transgenic mice expressing Cre recombinase under control of the MMTV-LTR and the WAP gene promoter (Wagner et al.,1997,2001). The MMTV-Cre transgenic mouse lines A and D were used to delete the Socs3 gene in mammary stem/early progenitors and epithelial cells during puberty, respectively, and the use of WAP-Cre mice resulted in Socs3 deletion in differentiating mammary luminal cells.
Deletion of the Socs3 Gene From Mammary Epithelium
The function of Socs3 in mammary epithelial cells was investigated through the deletion of the floxed gene using MMTV-Cre and WAP-Cre transgenic mice. Based on quantitative RT-PCR results, the level of Socs3 mRNA in mammary tissue of lactating Socs3fl/fl; MMTV-CreA mice was reduced by more than 80%. Deletion of the Socs3 gene was also determined by PCR on genomic DNA isolated from lactating tissue from the Socs3fl/fl; MMTV-CreD and Socs3fl/fl; WAP-Cre lines. Based on these results, the deletion efficiency was 65% in the MMTV-CreD line and 50% in WAP-Cre tissue (Supplementary Figure 1, which can be viewed at http://www.interscience.wiley.com/jpages/1058-8388/suppmat). The majority of mice carrying two floxed Socs3 alleles and the MMTV-Cre (line A) (Wagner et al.,1997) were unable to rear their litters to weaning age and most of the pups died within the first week of life. Only 3 out of 12 litters survived to weaning and the body weight of these pups was about 15% lower than pups reared by control dams. Whole mount (Supplementary Fig. 2) and histological evaluation of mammary glands harvested on day 13 of pregnancy (Fig. 1a–d) and on the day after delivering pups (Fig. 1e–h) revealed a sparser occupancy of the mammary fat pad in Socs3fl/fl; MMTV-CreA mice (Fig. 1b,d,f,h) compared to control tissues (Fig. 1a,c,e,g). At higher magnifications, no striking differences were observed at the level of individual alveoli between controls and Socs3fl/fl; MMTV-CreA tissues (Fig. 1c,d,g,h). On day 13 of pregnancy, tissues from control animals contained clusters of small alveoli. The secretory cells were densely packed and the lumina were small and filled with pink staining material (Fig. 1c). At the onset of lactation, the alveoli were expanded and were partially filled with milk. The cells appeared well differentiated and contained large lipid droplets (Fig, 1g). At the same stage, the alveoli in the Socs3fl/fl; MMTV-CreA glands were smaller but also appeared differentiated (Fig. 1h).
Differentiation Status of Luminal Epithelial Cells
In order to determine whether the defect in providing sufficient nutrition to their offspring was due to a differentiation defect, we next evaluated the status of epithelial differentiation during pregnancy by indirect immunostaining. The membrane proteins NKCC1 and Npt2b serve as markers for undifferentiated and differentiated secretory cells, respectively (Shillingford et al.,2003). They display complementary expression patterns. The basolateral sodium-potassium cotransporter NKCC1 is present in undifferentiated epithelial cells in virgins and is downregulated during pregnancy. The sodium-phosphate transporter Npt2b starts to be expressed on the apical surface during midpregnancy and decorates the luminal surface in lactating tissue. We opted to investigate day 13 of pregnancy, a time point at which epithelial cells are just beginning to express increasing levels of differentiation markers (Robinson et al.,1995) and the morning following delivery of pups, when the epithelial cells have achieved high levels of secretory differentiation and expression of milk proteins. On day 13 of pregnancy, NKCC1 was expressed in a small number of epithelial cells in control and Socs3fl/fl; MMTV-CreA tissues (Fig. 2a and b). Expression of Npt2b in control tissue was sporadic and found in less than 50% of the alveoli of wild type and mutant tissue (Fig. 2c and d). As evident already in the H&E stain, the size of alveoli was consistently larger in control tissue. At day 1 of lactation, only sporadic expression of NKCC1 was seen in control and mutant tissues (Fig. 2e and f) and Npt2b (Fig. 2g and h) was expressed to a similar extent in both tissues. In addition, we analyzed expression of one of the milk proteins. Whey acidic protein (WAP) was present in both tissues on pregnancy day 18; however, the levels were slightly reduced in Socs3fl/fl; MMTV-CreA samples (Supplementary Figure 3).
Activation of Stat3 and Stat5
To investigate the effect of the absence of Socs3 on the activation of Stat3 and Stat5, the most abundant STAT family members in mammary epithelium, we used immunostaining, which allows an evaluation at the cellular level. Starting after day 8 of pregnancy, the levels of Stat5a increase sharply followed by phosphorylation and nuclear localization. Stat3 activation levels are low during pregnancy and increase dramatically only at the onset of involution (Liu et al.,1996). In control tissue, phosphorylated Stat3 was not detected at day 13 of pregnancy (Fig. 3a) except for very few erratic cells. In contrast, phosphorylated Stat3 was consistently detected in a small number of cells in Socs3fl/fl; MMTV-CreA tissues (Fig. 3b). In both control and Socs3fl/fl; MMTV-CreA tissues, moderate levels of Stat5a were found on day 13 of pregnancy. While there was no difference in the pattern of Stat5a (Fig. 3c and d), the staining of phosphorylated Stat5 was reduced in Socs3fl/fl; MMTV-CreA tissues (Fig. 3f) compared to control tissue (Fig. 3e). Notably, both antibodies revealed that cells within one alveolus contained unequal levels with several cells displaying much stronger staining than others. This reflects the asynchronous initiation of the differentiation process as indicated by the staining pattern of NKCC1 and Npt2b. Similar to the pregnancy stages, phospho-Stat3 was not detected in control tissues harvested at term (Fig. 4a) but consistent staining was detected in a subset of alveoli in Socs3fl/fl; MMTV-CreA tissues (Fig. 4b). Staining was distributed unevenly and, in general, the cells that contained phosphorylated Stat3 were clustered. High levels of Stat5 and phosphorylated Stat5 were seen in control and mutant tissues (Fig. 4c–f) indicating that the reduced Socs3 levels do not influence post-partum Stat5 activity. This is in agreement with the normal differentiation of the tissue as demonstrated by the expression of Npt2b and WAP and the differentiated appearance of Socs3fl/fl; MMTV-CreA tissues in the histological staining. Furthermore, these results demonstrate that although reduction of Socs3 levels during pregnancy does not affect overall Stat5 levels, it appears to attenuate activation of Stat5. This was particularly striking in tissues harvested in mid pregnancy (Fig. 3f).
Reduced Proliferation in Socs3-Deleted Epithelium
In order to investigate whether the deletion of Socs3 affects cell proliferation, mice were subject to a defined hormone stimulus by estrogen and progesterone injection, a treatment that mimics the hormonal changes in early pregnancy and induces epithelial cell proliferation. Phosphorylation of histone3, an indicator of mitotic cells, and proliferating cell nuclear antigen (PCNA) were detected by immunostaining. Both parameters were reduced in Socs3fl/fl; MMTV-CreA compared to control tissue. While in controls 4.9% of cells were positive for phospho-histone 3 and 23.1% stained with PCNA (Fig. 5a and c), these numbers were 3.3 and 19.1% in mutant glands (Fig. 5b and d). Even though these differences were not statistically significant, they indicate a slight reduction of proliferation in the absence of Socs3. However, over the 19-day pregnancy period, these small changes will be cumulative.
Apoptosis was detected by staining for activated caspase 3 in these samples. The overall levels of apoptotic cells were very low and less than an average of 5 cells were found over an entire section in control tissues while it was possible to detect about one cell per microscopic field under 400-fold magnification in the Socs3fl/fl; MMTV-CreA samples. This suggests that the absence of Socs3 affects cell proliferation most likely by attenuating Stat5 activation.
Accelerated Involution Upon Loss of Socs3 in Differentiated Mammary Epithelium
MMTV-Cre (line D) and WAP-Cre transgenic mice were used to delete the Socs3 gene in secretory mammary epithelium during pregnancy (Wagner et al.,1997). In contrast to the MMTV-CreA line, which is active in mammary stem or early progenitor cells, the MMTV-CreD and WAP-Cre lines express Cre recombinase in committed luminal cells. Socs3fl/fl; MMTV-CreD and Socs3fl/fl; WAP-Cre mice were fertile and able to nurse their litters. Thus, mice from the MMTC-CreD lane were used to evaluate involution after lactation was fully established. On a histological level, mammary tissue from these mice at day 12 of pregnancy (data not shown) and day 10 of lactation was indistinguishable from control mice (Fig. 6a,c,e). However, upon weaning of the pups mutant mammary tissue displayed an accelerated involution. While the majority of secretory alveoli in control mice was still intact at day 2 of involution (Fig. 6b), those from Socs3fl/fl; MMTV-CreD mice had collapsed and involution had been initiated (Fig. 6f). Socs3fl/fl; WAP-Cre glands displayed a rather heterogeneous appearance (Fig. 6d) at this stage and about 50% of the alveoli had undergone remodeling, which correlates with the incomplete and mosaic degree of Socs3 gene deletion in this line (data not shown).
Enhanced Stat3 Activation in Socs3-Null Mammary Epithelium During Involution
Stat3 activation during involution correlates with tissue remodeling, and loss of Stat3 results in delayed involution (Chapman et al.,1999; Humphreys et al.,2002). To test whether the accelerated remodeling of mammary tissue in the absence of Socs3 was associated with unscheduled Stat3 activation, immunohistochemical analyses for phospho-Stat3 were performed at day 10 of lactation and throughout involution. At day 10 of lactation, Stat3 phosphorylated at tyrosine 705 was virtually absent, both in control and Socs3fl/fl; MMTV-CreD and Socs3fl/fl;WAP-Cre mice (Fig. 7a,c,e). At day two of involution, both Socs3fl/fl;MMTV-CreD and Socs3fl/fl; WAP-Cre mice (Fig. 7d and f) had a larger number of luminal cells expressing P-Stat3 than control tissue (Fig. 7b). To quantify differences in Stat3 activation, total protein was analyzed from mammary tissue from Socs3fl/fl; MMTV-CreD mice at day 10 of lactation. Low levels of P-Stat3 were detected in control tissue and Stat3 phosphorylation was enhanced approximately 2.5-fold in mutant tissue (Fig. 7g).
Confining Stat5 activity in mammary epithelium to pregnancy and lactation and Stat3 activity to the involution phase is essential for the homeostasis of mammary tissue throughout its developmental cycle. Evidence is mounting that members of the SOCS family are essential in controlling the onset, extent, and duration of cytokine-induced Stat3 and Stat5 signaling. While Socs1 (Lindeman et al.,2001) and Socs2 (Harris et al.,2006) modulate prolactin signaling during pregnancy, Socs3 is essential for the functional development of mammary tissue during pregnancy and for the timely execution of mammary tissue remodeling during involution, as shown in this study.
Although Stat3 is present throughout puberty, pregnancy, lactation, and involution, its activation is largely confined to the process of involution (Liu et al.,1996) where it controls the death and remodeling of mammary epithelium (Chapman et al.,1999; Humphreys et al.,2002). Maintaining Stat3 in a quiescent state during pregnancy and lactation could be important to ensure proliferation and differentiation of mammary epithelium instead of initiating apoptosis. Genetic studies have defined Socs3 as an essential negative regulator of Stat3 (Yoshimura et al.,2005) and it can be hypothesized that it modulates Stat3 activation at any stage of mammary development. Inactivation of the Socs3 gene in mammary stem cells and progenitors resulted in increased Stat3 activation but did not lead to overt structural changes of mammary tissue during pregnancy. Mammary epithelium expanded during pregnancy and underwent functional differentiation in these mice as judged by the expression of different marker proteins, including a milk protein. However, reduced proliferation during pregnancy, possibly due to attenuated Stat5 activation, resulted in a paucity of alveoli. In general, these alveoli were not extended, which correlated with the inability of these mice to nurse their litters.
Loss of Socs3 in differentiating mammary secretory epithelium during pregnancy did not visibly alter the histology or the function of these cells during lactation as judged by successful rearing of litters. Thus a limited activation of Stat3 during pregnancy or lactation was insufficient to trigger widespread cell death and tissue remodeling. However, upon weaning, mice lacking Socs3 in their secretory epithelium exhibited an accelerated phase of involution, which correlated with the precocious activation of Stat3. This demonstrates that Socs3 balances Stat3 to ensure a controlled involution. Prolactin-induced cell proliferation and differentiation during pregnancy and IL-6/LIF-induced cell death upon weaning is controlled by the transcription factors Stat5 and Stat3, respectively. The balance between these two developmental stages, in turn, is determined by Socs3 (Fig. 8). Since the Socs3 gene itself is a direct target of Stat5, the presence of Socs3 during pregnancy and lactation curtails the activation of Stat3, which in itself ensures that mammary epithelium does not succumb to premature apoptosis. Loss of the Socs3 gene results in unscheduled Stat3 activation and a loss of alveolar homeostasis.
Lastly, overexpression studies have demonstrated that Socs3 can suppress Stat5 signaling (Dif et al.,2001). If such negative feedback loops were of in vivo significance, the loss of Socs3 should result in an aberrant activation of Stat5 and precocious development of mammary epithelium. However, this was not observed in this study, suggesting that Socs3 is not a prominent negative regulator of Stat5 signaling in mammary epithelium during pregnancy and lactation. This is in contrast to Socs1 (Lindeman et al.,2001) and Socs2 (Harris et al.,2006), both of which curtail and regulate cytokine-induced activation of Stat5 and thus prevent inappropriate mammary differentiation prior to delivery.
After this work had been accepted for publication, Sutherland and colleagues (2006) published a study describing the deletion of the Socs3 gene in mouse mammary epithelium. In agreement with our study, these investigators observed increased apoptosis of luminal cells in the late phase of lactation and during involution, which resulted in an accelerated tissue remodeling. In addition, these authors linked apoptosis to increased c-myc expression. The WAPi-Cre transgene used in their study resulted in an efficient deletion of the Socs3 gene during lactation. Since these mutant mice were able to lactate and raise their litters, the authors concluded that Socs3 is not a key physiological regulator of prolactin signaling. However, prolactin-mediated mammary development occurs during pregnancy (Hennighausen and Robinson,2005) and deletion of Socs3 during pregnancy was not reported in this study. In contrast, our study employed an MMTV-Cre transgene that is active in mammary stem and/or progenitor cells prior to pregnancy. As shown in this study, pregnancy-mediated development of mammary tissue in the absence of Socs3 resulted in reduced proliferation and functionally impaired mammary epithelium as evidenced by a failure of most dams to nurse their litters. Thus, Socs3 has distinct contributions in the homeostasis of mammary epithelium; it modulates proliferation and functional development during pregnancy and scheduled apoptosis and tissue remodelling during involution.
Socs3fl;fl females (Yasukawa et al.,2003) were crossed with MMTV-Cre transgenic males from lines A and D and the WAP-Cre line (Wagner et al.,1997) to yield Socs3fl/fl; MMTV-Cre and Socs3fl/fl; WAP-Cre offspring, respectively. Mature females of the CreA line were mated and inspected daily for plugs to determine pregnancy stages. On day 13 of pregnancy, one of the inguinal mammary glands was harvested by biopsy and the second gland was harvested within 12 hr of delivery of a litter. A minimum of 3 mice of each genotype was analyzed. Female mice of the CreD and WAP-Cre lines with litter sizes between six and ten animals were used for analysis of involution points. At lactation day 10, pups were removed and mammary tissue of the left gland number 4 was collected by biopsy. Mice were killed 1, 2, or 3 days later and the right mammary gland number 4 was collected for analysis. For hormone stimulation, animals were injected subcutaneously twice at 24-hr intervals with 1 μg β-estradiol and 1 mg progesterone in 100 μl of sesame oil and sacrificed 24 hr later. All experiments were approved and performed according to the guidelines of the Animal Care and Use Committee of NIDDK.
Histology and Immunostaining
Tissues were fixed in neutral buffered formalin at 4°C overnight, dehydrated, and embedded in paraffin. Tissue blocks were sectioned at 5 μm and stained with hematoxylin and eosin. Images were captured using a Olympus BX51 microscope and a DXM1200 digital camera (Nikon) or a CCD Retiga Exi FAST1394 camera (Q-Imaging). For immunostaining, tissue sections were deparaffinized in xylene and rehydrated in decreasing alcohol concentrations. Antigen unmasking was performed in a Decloaking chamber (Biocare Medical) using BORG Decloaker Solution pH 9.5 (Biocare Medical no. BD1000GI) at 125°C, 18–24 PSI for 5 min. Primary antibodies were incubated for 1 hr at room temperature (NKCC1, 1:1,000; Npt2b, 1:200; WAP, 1:400; phospho-histone3, Upstate no. 06-570; PCNA, Dako no. M0879), or overnight at 4°C (Stat3, Cell Signaling Technologies no. 610182, 1:200; P-Stat3, Cell Signaling Technologies no. 9135, 1:100; Stat5a, Santa Cruz no. 1081, 1:100; P-Stat5, Cell Signaling Technologies no.9351, 1:100; E-cadherin, BD Transduction Laboratories no. 610182, 1:200). Alexafluor conjugated secondary antibodies (Molecular Probes, Eugene, OR) were used at a dilution of 1:400 for 30 min at room temperature.
Western Blot Analysis
Proteins were extracted form frozen mammary tissue (−80°C) and homogenized in lysis buffer (40 mM TrisCl pH 8, 280 mM NaCl, 20% glycerol, 2% NP-40, 4 mM EDTA pH 8, 20 mM NaF, 100 μg/μl PMSF, 400 μg/μl aprotinin, 400 μg/μl leupeptin, and 2 mM sodium vanadate) using a Polytron. The lysates were rotated for 1 hr at 4°C and were cleared from fat by centrifugation for 20 min, 16000g, 4°C. Protein concentration was determined using the Bradford protein assay. Thirty micrograms of protein per lane were separated on 8% Tris-glycine gels (Invitrogen, La Jolla, CA) and transferred to 0.45 μm pore size Nitrocellulose membranes (Invitrogen). Membranes were blocked using 5% fat-free milk powder (Bio-Rad, Richmond, CA) in TBS + 0.05% Tween-20. Primary antibodies (p-Stat3, Cell Signaling Technologies no. 9135; Stat3, Santa Cruz no. 485) were diluted 1:1,000 in TBS-T and incubated at 4°C overnight. Secondary antibodies (Amersham, no. NA934V) were diluted 1:5,000 in TBS-T 5% milk powder and the signal was detected using chemiluminescence (Pierce no. 34080). The signal was quantitated using a Kodak Image Station 440 CF with Kodak 1D Software (Version 3.6.2).
We thank Sirong Sloane Yu and Tanuj Sood for help with some experiments as part of their summer internship at the NIH. We also thank Drs. Jim Turner, NCI, and Juerg Biber, Zurich, for the kind gift of NKCC1 and Npt2b antibodies, respectively. M.P.Z. was supported by funds from the Austrian Ministry of Education, Science, and the Arts (GEN-AU).