c‐myc as a mediator of accelerated apoptosis and involution in mammary glands lacking Socs3

Kate D Sutherland, François Vaillant, Warren S Alexander, Tim M Wintermantel, Natasha C Forrest, Sheridan L Holroyd, Edward J McManus, Gunther Schutz, Christine J Watson, Lewis A Chodosh, Geoffrey J Lindeman, Jane E Visvader

Author Affiliations

  1. Kate D Sutherland1,2,
  2. François Vaillant1,
  3. Warren S Alexander3,
  4. Tim M Wintermantel4,
  5. Natasha C Forrest1,
  6. Sheridan L Holroyd1,
  7. Edward J McManus3,
  8. Gunther Schutz4,
  9. Christine J Watson5,
  10. Lewis A Chodosh6,
  11. Geoffrey J Lindeman1 and
  12. Jane E Visvader*,1
  1. 1 VBCRC Laboratory, The Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria, Australia
  2. 2 Department of Medical Biology, University of Melbourne, Victoria, Australia
  3. 3 Cancer & Haematology Division, The Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria, Australia
  4. 4 Molecular Biology of the Cell I, German Cancer Research Centre, Heidelberg, Germany
  5. 5 Mammary Apoptosis and Development Group, Department of Pathology, University of Cambridge, Cambridge, UK
  6. 6 Abramson Family Cancer Research Institute, University of Pennsylvania School of Medicine, Philadelphia, PA, USA
  1. *Corresponding author. VBCRC Laboratory, The Walter and Eliza Hall Institute of Medical Research, 1G Royal Parade, Parkville, Victoria 3050, Australia. Tel.: +61 3 9345 2494; Fax: 61 3 9347 0852; E‐mail: visvader{at}
View Full Text


Suppressor of cytokine signalling (SOCS) proteins are critical attenuators of cytokine‐mediated signalling in diverse tissues. To determine the importance of Socs3 in mammary development, we generated mice in which Socs3 was deleted in mammary epithelial cells. No overt phenotype was evident during pregnancy and lactation, indicating that Socs3 is not a key physiological regulator of prolactin signalling. However, Socs3‐deficient mammary glands exhibited a profound increase in epithelial apoptosis and tissue remodelling, resulting in precocious involution. This phenotype was accompanied by augmented Stat3 activation and a marked increase in the level of c‐myc. Moreover, induction of c‐myc before weaning using an inducible transgenic model recapitulated the Socs3 phenotype, and elevated expression of likely c‐myc target genes, E2F‐1, Bax and p53, was observed. Our data establish Socs3 as a critical attenuator of pro‐apoptotic pathways that act in the developing mammary gland and provide evidence that c‐myc regulates apoptosis during involution.


Morphogenesis of the mammary gland occurs in distinct stages that are governed by the concerted action of hormones and locally produced cytokines (reviewed in Hennighausen and Robinson, 2005). The mammary gland primarily develops after birth, with ductal branching and elongation resulting in an epithelial tree that fills the mammary fat pad by the end of puberty. During pregnancy, lobuloalveolar units develop and undergo terminal differentiation before parturition, thus allowing milk production during lactation. During involution, structural remodelling of the gland occurs in a two‐phase process (Lund et al, 1996). The first phase involves extensive apoptosis of the alveolar epithelial cells and is reversible, whereas the second phase involves remodelling of the mammary gland through degradation of the extracellular matrix by matrix metalloproteinases (MMPs), together with macrophages, and is irreversible. In mice, the number of apoptotic cells peaks around days 2–3 of involution and remodelling of the gland is complete by days 6–8 of this process (Lund et al, 1996; Li et al, 1997).

Signalling mediated by glycoprotein 130 (gp130), the common receptor chain for interleukin (IL)‐6‐type cytokines, plays an important role in regulating programmed cell death during involution. Involution is delayed in IL‐6‐ and gp130‐deficient glands, accompanied by reduced Stat3 activation in the latter model (Zhao et al, 2002, 2004). Conditional deletion of Stat3 in the mammary epithelium also results in a dramatic delay in mammary gland involution (Chapman et al, 1999). LIF appears to be a key regulator of Stat3‐dependent apoptosis in the mammary gland as LIF−/− female mice exhibit delayed involution accompanied by a decrease in Stat3 phosphorylation (Kritikou et al, 2003). Recent studies have shown that one mechanism by which Stat3 induces apoptosis is by downregulating PI3‐K activity via the modulation of its regulatory subunits, p50α and p55α (Abell et al, 2005). These findings suggest that crosstalk between multiple signalling pathways determines cell death during involution.

The suppressor of cytokine signalling (SOCS) proteins are critical regulators of cytokine‐mediated signalling in multiple organs. These proteins are cytokine‐inducible and act in a classical negative feedback loop to inhibit signal transduction along the JAK/STAT pathway (reviewed in Kubo et al, 2003; Alexander and Hilton, 2004). Functionally, SOCS proteins interact with JAK kinases and/or cytokine receptors, targeting them for ubiquitination and proteasomal degradation. The eight members of the SOCS family are characterised by an N‐terminal region of variable length and limited homology, a central SH2 domain and a conserved SOCS box at the C‐terminus. Physiologically, Socs proteins have been shown to have distinct roles. In the mammary gland, both Socs1 and Socs2 attenuate prolactin receptor (PRLR) signalling in vivo. Mice carrying targeted deletions of Socs1 and Ifn‐γ exhibited precocious lobuloalveolar development, and loss of a single Socs1 allele was sufficient to rescue lactation in PRLR‐deficient mice (Lindeman et al, 2001). In the case of Socs2, removal of both alleles was necessary to restore lactogenesis in PRLR heterozygous females (Harris et al, 2006). Overexpression of Cis in transgenic mice led to impaired terminal differentiation and failure of lactation (Matsumoto et al, 1999), although Cis‐null mice appeared to lactate normally (Marine et al, 1999).

Socs3 has the greatest sequence homology to Socs1 but binds the receptor rather than directly interacting with JAKs to gain access to the JAK activation loop (Yasukawa et al, 1999; Nicholson et al, 2000). Within the mammary epithelium, Socs3 expression is inducible by prolactin (Prl) and epidermal growth factor (EGF) (Tam et al, 2001; Tonko‐Geymayer et al, 2002), and by activation of Stat3 in KIM‐2 mammary epithelial cells (Clarkson et al, 2006). Overexpression studies in 293T and HC11 mammary epithelial cells have shown that Socs3 inhibits Stat5 activation in response to Prl and EGF signalling (Helman et al, 1998; Pezet et al, 1999; Tonko‐Geymayer et al, 2002), and inhibits β‐casein expression in response to Prl in SCp2 cells (Lindeman et al, 2001). Although the expression of Socs3 as well as other family members is induced by a plethora of cytokines in vitro, gene targeting experiments have revealed that they have more specific roles in vivo.

Socs3 nullizygous mice die at mid‐gestation owing to placental insufficiency (Marine et al, 1999; Roberts et al, 2001). However, cell‐specific targeting of the Socs3 locus has established that Socs3 is a key negative regulator of IL‐6 signalling in hepatocytes and macrophages (Croker et al, 2003; Lang et al, 2003; Yasukawa et al, 2003). We sought to elucidate the physiological role of Socs3 in the developing mammary gland and to identify signalling pathways regulated by Socs3 in mammary epithelium. Deletion of the Socs3 locus in alveolar epithelial cells using an improved WAPiCre transgene resulted in a marked increase in epithelial apoptosis and accelerated involution. This was accompanied by increased levels of activated Stat3, as well as a marked increase in the level of c‐myc and its target genes. We conclude that Socs3 functions as a negative regulator of gp130 signalling during involution and that c‐myc, a target of Stat3, plays a role in mediating apoptosis in the mammary gland.


WAPiCre‐mediated excision of Socs3 in the mammary gland

To examine the physiological role of Socs3 in the adult mammary gland, we utilised WAPiCre transgenic mice (Wintermantel et al, 2002). The specificity and timing of WAPiCre‐mediated recombination during mammary gland development was confirmed by intercrossing with GtRosa26 reporter mice (Soriano, 1999), which express lacZ following Cre‐mediated deletion. Intense lacZ staining was observed in the majority of alveolar cells in WAPiCre;GtRosa26 bi‐transgenic mice at day 10 of lactation and day 2 of involution, but no staining was detectable in the mammary glands of control GtRosa26 mice (Figure 1A).

Figure 1.

Targeted deletion of Socs3 in alveolar cells by WAPiCre. (A) Small portions of the thoracic and inguinal mammary glands from GtRosa26 mice or WAPiCre;GtRosa26 bitransgenic mice were analysed at day 10 of lactation (L) and day 2 of involution (I) by staining for β‐galactosidase activity. Scale bar, 100 μm. (B) Genomic DNA derived from Socs3+/fl and WAPiCre;Socs3−/fl mammary glands at day 10 of lactation (L) and days 1, 2 and 3 of involution (I) was digested with BamHI. Membranes were hybridised with a 5′ Socs3 probe that distinguishes the wild‐type (+, 20 kb), conditional (fl, 9 kb), null (−, 5 kb) and excised (Δ, 4.9 kb) Socs3 alleles. (C) The expression of Socs3 transcripts relative to Hprt was determined by real‐time PCR analysis of RNA derived from Socs3+/fl and WAPiCre;Socs3−/fl (depicted as Socs3−/Δ) mice (2–3) at day 10 of lactation (L) and days 1–3 of involution (I) performed in duplicate. (D) Mammary gland protein lysates (50 μg) were analysed by Western blotting using an anti‐Socs3 antibody and reprobed with anti‐tubulin to control for protein loading.

Socs3+/− mice bearing a single copy of the WAPiCre transgene were crossed with Socs3fl/fl mice (Croker et al, 2003) to generate offspring in which the conditional (fl) allele was deleted in lobuloalveolar cells. The frequency of excision by WAPiCre was determined by Southern blot analysis of genomic DNA, demonstrated by loss of the floxed allele (fl, 9 kb) and appearance of the excised allele (Δ, 4.9 kb). As expected, this was highest in mammary glands from WAPiCre;Socs3−/fl mice during lactation and involution (Figure 1B). As the mammary gland comprises a heterogeneous cell population that includes stroma, in which the WAPiCre transgene is inactive, the extent of Socs3 inactivation in mammary epithelium is likely to be underestimated (Figure 1B). To confirm the absence of Socs3 expression in the mammary glands of WAPiCre;Socs3−/fl mice, we analysed Socs3 transcript levels by real‐time PCR (Figure 1C and Supplementary Figure 1) and Socs3 protein levels (Figure 1D) by Western blot analysis. Taken together, these data demonstrate efficient deletion of Socs3 in alveolar cells during lactation and involution.

Conditional deletion of Socs3 during lactation leads to increased apoptosis

We examined the effect of deleting Socs3 in the mammary gland during established lactation using WAPiCre transgenic mice. At day 4 of lactation, no differences were observed in the ability of either Socs3+/fl or WAPiCre;Socs3−/fl mothers to maintain their litters, and both Socs3+/fl and WAPiCre;Socs3−/fl mammary glands exhibited well‐rounded alveoli with large distended lumens (Figure 2A). At day 10 of lactation, the alveoli in WAPiCre;Socs3−/fl glands appeared smaller than in control Socs3+/fl glands (Figure 2A), but there were no TUNEL‐positive cells in either these or control glands (data not shown). By day 14, the lumens had partly collapsed and adipocytes were more conspicuous throughout WAPiCre;Socs3−/fl mammary glands (Figure 2A). This was accompanied by an increase in the number of TUNEL‐positive cells (1.0 versus 0%; Figure 3A). These findings suggest that apoptosis and tissue remodelling are accelerated in Socs3‐deficient mammary glands during established lactation.

Figure 2.

Conditional deletion of Socs3 in alveolar cells results in accelerated involution. (A) Sections of mammary glands from control Socs3+/fl and Socs3‐deficient mice at days 4, 10 and 14 of lactation (L). Scale bar, 100 μm. (B) Sections from Socs3+/fl and WAPiCre;Socs3−/fl mice at days 1, 2, 3 and 6 of involution (I). Scale bar, 100 μm.

Figure 3.

Loss of Socs3 results in increased epithelial apoptosis during involution. (A) TUNEL assays were performed on mammary glands from three mice for each time point. Error bars represent standard error of the mean; **P<0.05. (B) Top: Western blot analysis of protein lysates derived from Socs3+/fl and WAPiCre;Socs3−/fl mammary glands at days 1, 2, 3 and 4 of involution (I), using antibodies specific for Mcl‐1, Bak and Bax. Expression of tubulin served as a control for protein loading. Bottom: The level of Bim transcripts was determined by reverse transcription–PCR, and products resolved by electrophoresis. (C) Gelatin zymography was used to detect MMP‐2 and MMP‐9 activity. The latent (68 kDa) and activated (60 kDa) forms of MMP‐2, and total MMP‐9 protein (109 kDa) are indicated. (D) Decreased Timp‐3 mRNA expression in Socs3‐deficient mammary glands during the first stage of involution, as determined by real‐time PCR analysis of Timp‐3 mRNA expression in glands from Socs3+/fl and WAPiCre;Socs3−/fl (Socs3−/Δ) mice, at days 1, 2 and 3 of involution. Results represent the mean data from at least two mice. Error bars represent standard error of the mean.

Mammary gland involution is accelerated by loss of Socs3 and is accompanied by increased apoptosis

Given that Socs3 expression peaks at the onset of mammary gland involution (Supplementary Figure 2), we examined the effect of loss of Socs3 during involution. Pups were removed from Socs3+/fl and WAPiCre;Socs3−/fl mice after 10 days of lactation, and glands were collected for histological analysis at days 1, 2, 3 and 6 of involution (Figure 2B). WAPiCre;Socs3−/fl glands exhibited a dramatic increase in apoptotic bodies as early as 12 h (data not shown) and 24 h of involution (Figure 2B). By 48 h, remodelling of the gland had already commenced. Collapsed lobuloalveolar structures, concomitant with the reappearance of stroma, were readily apparent in WAPiCre;Socs3−/fl mammary glands but not in the glands from Socs3+/fl (Figure 2B) or Socs3−/fl mice (data not shown), both of which displayed well‐ordered alveolar structures with apoptotic cells visible within their lumens. By day 3 of involution, the lobuloalveolar structures in the mutant mammary glands had largely been remodelled, resulting in glands that were typical of mice that had completed involution (Figure 2B).

To determine whether the accelerated involution evident in Socs3‐deficient mammary glands was due to increased apoptosis, TUNEL assays were performed in mammary gland sections. A significant increase in TUNEL‐positive cells was seen at day 1 (4.6±1.6% in WAPiCre;Socs3−/fl versus 1.0±0.2% in control mammary glands) and day 2 (7.9±1.1% in WAPiCre;Socs3−/fl versus 1.9±0.1% in control mammary glands) of involution (Figure 3A). In late involution (days 4 and 6), no difference in the number of TUNEL‐positive cells was detected between WAPiCre;Socs3−/fl and Socs3+/fl glands. Thus, Socs3 deficiency profoundly accelerates the rate of apoptosis of alveolar cells during the initial phase of involution.

Mammary epithelial cell proliferation was assessed in involuting mammary glands following bromodeoxyuridine (BrdU) injection and immunostaining. Cell proliferation was elevated at least six‐fold in WAPiCre;Socs3−/fl glands at day 4 of involution compared to control glands (5.5±0.9% in WAPiCre;Socs3−/fl versus 0.7±0.4% in Socs3+/fl glands), whereas no difference was found at day 2 (4.3±0.6% in WAPiCre;Socs3−/fl versus 4.9±1.9% in Socs3+/fl glands) and day 6 (5.0±1.2% in WAPiCre;Socs3−/fl versus 3.4±1.1% in Socs3+/fl glands) of involution (Supplementary Figure 3). These results indicate that Socs3 deficiency results in increased epithelial cell proliferation during the second but not the first phase of involution. This may represent a compensatory response to the precocious apoptosis that ensues in Socs3‐deficient glands or reflect a more direct activity of a Socs3‐regulated pathway that acts in late involution.

Increased levels of Bax and Bak in Socs3‐deficient glands at the onset of involution

Given that Socs3 deficiency promotes apoptosis of lobuloalveoli, we examined the expression of members of the Bcl‐2 family (Cory et al, 2003). Expression of the pro‐survival family members Mcl‐1 (Figure 3B), Bcl‐XL and Bcl‐2 (data not shown) was not altered in Socs3‐deficient glands relative to Socs3+/fl control glands, as detected by Western blot analysis. It is noteworthy that Bcl‐XL levels also did not change in Stat3‐deficient mammary glands (Chapman et al, 1999). In contrast, expression of the pro‐apoptotic regulators Bax and Bak was significantly increased in WAPiCre;Socs3−/fl mammary glands at days 1 and 2 of involution (Figure 3B), relative to control glands. In addition, the level of Bim mRNA was elevated in Socs3‐deficient glands at day 1 of involution (Figure 3B). It is therefore likely that Bax, Bak and Bim directly contribute to the increased epithelial apoptosis in Socs3‐deficient glands.

Perturbed activation of matrix metalloproteinases in Socs3‐deficient mammary glands

The second phase of mammary gland involution is characterised by the activation of MMPs, which remodel the gland through degradation of the extracellular matrix (ECM) (reviewed in Mott and Werb, 2004). Gelatinase A (MMP‐2) and gelatinase B (MMP‐9) are activated during involution and degrade collagen, gelatin and type IV collagen within the ECM. Gelatin zymography of control glands revealed activation of MMP‐2 (60 kDa) at day 3 of involution, with high levels of both the latent (pro‐MMP‐2; 68 kDa) and active forms evident by day 4 (Figure 3C). In comparison, WAPiCre;Socs3−/fl glands exhibited activated MMP‐2 at day 2 of involution, with high levels of the latent and activated forms at day 3 (Figure 3C). By day 4 of involution (Figure 3C), similar levels were apparent in mutant and control glands. An increase in MMP‐9 protein (109 kDa), representing both pro‐ and active forms, was also observed in Socs3‐deficient glands at days 1 and 2 of involution (Figure 3C).

MMP activity is inactivated by a small family of inhibitors known as tissue inhibitors of metalloproteinases (TIMPs) (Lafleur et al, 2003a), and the ratio of MMP to TIMP expression is critical for tissue remodelling during involution. To examine whether TIMP levels changed in Socs3‐deficient mammary glands, we determined Timp‐3 expression by real‐time PCR analysis. A two‐ to four‐fold reduction in Timp‐3 expression was observed in WAPiCre;Socs3−/fl mammary glands throughout involution relative to that in Socs3+/fl mice (Figure 3D). The premature activation of MMP‐2 and MMP‐9 and decrease in Timp‐3 mRNA expression are compatible with the accelerated tissue remodelling that occurs in Socs3‐deficient mammary glands (Figure 2B). Furthermore, accelerated apoptosis was observed in Timp‐3‐deficient mammary glands (Fata et al, 2001).

Socs3 deficiency leads to increased Stat3 activation in involuting mammary glands and hyper‐responsiveness to LIF in Socs3‐null mammary epithelial cells

Stat3 phosphorylation is negligible during lactation, but increases at the onset of involution (Liu et al, 1996; Philp et al, 1996). WAPiCre;Socs3−/fl mammary glands displayed substantially elevated levels of Stat3 phosphorylation at days 10 and 14 of lactation, relative to that in control mammary glands (Figure 4A). During involution, higher levels of activated Stat3 were evident in Socs3‐deficient glands at days 2 and 3, whereas by day 4 they were comparable to that seen in control mammary glands. These data suggest that the increased levels of activated Stat3 in Socs3‐deficient glands during established lactation (day 10) and early involution trigger apoptosis.

Figure 4.

Increased levels of activated Stat3 but not Akt in Socs3‐deficient mammary glands. (A) The levels of phospho‐Stat3 and total Stat3 in Socs3+/fl control and WAPiCre;Socs3−/fl glands at days 4, 10 and 14 of lactation (L), and days 2, 3 and 4 of involution (I) were determined by Western blot analysis. (B) Socs3+/fl and Socs3−/Δ immortalised mammary epithelial cell (MEC) clones were pulsed with 100 ng/ml leukaemia inhibitory factor (LIF) for 15 min and protein lysates (10 μg) were analysed by immunoblotting using antibodies specific for phospho‐Stat3, total Stat3, phospho‐ERK1/2 and total ERK1/2. (C) Levels of serine136 phosphorylated Akt and total Akt were assessed by Western blot analysis at the indicated time points. Two mice of each genotype are shown.

To explore whether the kinetics of Stat3 activation by cytokines was altered in the absence of Socs3, we generated mammary epithelial cell (MEC) lines lacking Socs3. Primary MEC cultures were established from Socs3−/fl mid‐pregnant mammary glands and then transduced with an E6/E7‐expressing retrovirus to immortalise the cells, followed by a Cre‐recombinase‐expressing retrovirus (Krempler et al, 2002) to excise the conditional allele (Supplementary Figure 4A). Socs3+/fl and Socs3−/Δ cells were pulsed with LIF and Stat3 activation was analysed by Western blotting. Stimulation with LIF resulted in the rapid activation of Stat3 in both Socs3‐deficient and control cells, but the intensity and duration of Stat3 phosphorylation were significantly enhanced in Socs3−/Δ cells (Figure 4B). No Socs3 mRNA was detectable in the Socs3‐null cells in contrast to control lines (Supplementary Figure 4B). Although LIF has also been shown to signal through the MAP kinase pathway, similar patterns of ERK1/2 phosphorylation were observed in Socs3+/fl and Socs3−/Δ cells (Figure 4B). Prolonged Stat3 activation was also observed in Socs3−/Δ MECs when stimulated with IL‐6 (data not shown). Taken together, these data suggest that Socs3 is a critical negative regulator of Stat3‐mediated apoptosis in the mammary gland.

Akt activation is unaltered in glands lacking Socs3

Stat3 has been proposed to inhibit Akt‐mediated survival by inducing expression of the PI3‐K regulatory subunits p50α and p55α (Abell et al, 2005). We found that the level of activated Akt in Socs3+/fl and WAPiCre;Socs3−/fl mammary glands was similar at days 1 and 2 of involution (Figure 4C; data not shown), suggesting that Socs3 does not act on the Akt pathway in early involution. The increase in phosphorylated Akt in WAPiCre;Socs3−/fl glands around days 2–3 of involution (Figure 4C) may reflect a feedback response to the accelerated epithelial apoptosis that occurs in these mammary glands. In any event, no inhibition of Akt activation was evident in Socs3‐deficient glands, indicating that Stat3 is acting through other pathways.

Elevated levels of c‐myc and its potential target genes in Socs3‐deficient glands

The c‐myc gene is a direct target of Stat3 and is a critical regulator of both cell proliferation and apoptosis (Eisenman, 2001). We therefore examined whether increased c‐myc expression accompanied enhanced Stat3 activation in Socs3‐deficient mammary glands. Substantially higher levels of c‐myc protein were apparent in Socs3‐deficient mammary glands at days 1, 2, 4 and 6 of involution (Figure 5A) relative to Socs3+/fl glands. Higher levels of c‐myc mRNA were also observed in Socs3‐deficient glands using real‐time PCR analysis (Figure 5B). Expression of c‐myc was also elevated during lactation (days 10 and 14) in glands lacking Socs3 (data not shown). Notably, c‐myc mRNA expression was acutely induced following LIF treatment of Socs3‐deficient mammary epithelial cells relative to that in control cells, as determined by real‐time PCR analysis (Figure 5C). This observation is consistent with c‐myc being a target of Stat3. The mechanism by which c‐myc is downregulated at 1 h following LIF treatment is yet to be determined. Interestingly, levels of the pro‐apoptotic proteins E2F‐1 and p53, whose genes are either direct or indirect targets of c‐myc, were markedly elevated in involuting mammary glands lacking Socs3 (Figure 5D).

Figure 5.

Increased expression of c‐myc in the absence of Socs3. (A) The levels of c‐myc in Socs3+/fl and WAPiCre;Socs3−/fl glands at days 1, 4 and 6 of involution (I) were assessed by Western blot analysis using a polyclonal anti‐c‐myc antibody (AbCam). Two mice for each genotype are shown for day 4 of involution. (B) The expression of c‐myc transcripts relative to 18S rRNA was determined in individual Socs3+/fl control and Socs3−/Δ mammary glands by real‐time PCR analysis. Error bars represent the standard error of the mean for two to three independent experiments. (C) The expression of c‐myc transcripts relative to 18S rRNA was determined in Socs3+/fl and Socs3−/Δ immortalised mammary epithelial cell clones by real‐time PCR analysis, following induction with 100 ng/ml LIF. Error bars represent standard error of the mean for two independent experiments. (D) The expression of likely c‐myc apoptotic targets, E2F‐1 and p53, was assessed in Socs3+/fl and WAPiCre;Socs3−/fl glands by Western blot analysis. The higher molecular weight bands represent phosphorylated forms of E2F‐1 and the lower band corresponds to free E2F‐1. Expression of tubulin served as a control.

c‐myc promotes alveolar apoptosis and premature involution in the mammary gland

The physiological role of the c‐myc proto‐oncogene in mammary gland involution could not be addressed in MMTV‐myc or in MMTV‐rtTA/TetO‐MYC (MTB/TOM) bi‐transgenic mice chronically induced with doxycycline, as they fail to nurture their pups (Stewart et al, 1984; Andres et al, 1988; Blakely et al, 2005). To circumvent this problem, we induced expression of the c‐MYC transgene in bitransgenic MTB/TOM females (D'Cruz et al, 2001) with doxycycline at day 8 of lactation, before initiating involution on day 10 of lactation via forced weaning, as transient c‐MYC expression has been shown to be compatible with maintenance of lactation (Blakely et al, 2005). Involution was found to be dramatically accelerated in MTB/TOM mice compared to glands from doxycycline‐treated MTB mice, as shown histologically in Figure 6A. Furthermore, the number of TUNEL‐positive cells in MTB/TOM mice at day 2 of involution was increased at least three‐fold compared to control glands from MTB mice (Figure 6B).

Figure 6.

c‐myc promotes apoptosis of mammary epithelial cells during involution. (A) Mammary gland sections from MTB and MTB/TOM mice at 2 days of involution (I) following treatment with doxycycline (1 mg/ml) for 4 days (day 8 of lactation to day 2 of involution). Two mice of each genotype are shown. Scale bar, 100 μm. (B) TUNEL assays were performed on mammary glands from 2 to 4 mice of each genotype. Error bars represent standard error of the mean. (C) Northern blot analysis of total RNA from mammary glands taken from independent MTB (#26, 29) and MTB/TOM mice (#66, 65, 69, 133, 122) following treatment with doxycycline for 4 days (day 8 of lactation to day 2 of involution). The filter was hybridised sequentially with a human c‐MYC cDNA probe, specific for the transgene, and an 18S rRNA oligonucleotide probe, to control for the level of RNA. (D) Mammary gland protein lysates from doxycycline‐treated MTB and MTB/TOM mice analysed at day 2 of involution were subjected to Western blot analysis using specific antibodies against p53, E2F‐1, Bax, phospho‐Stat3 and total Stat3. Expression of tubulin served as a control for protein loading.

The level of c‐MYC induction following 4 days of doxycycline administration was determined by Northern blot analysis. Notably, the expression of c‐myc was comparable between Socs3‐deficient mammary glands during involution and MTB/TOM bitransgenic glands following doxycycline induction (Supplementary Figure 5), indicating that MYC is not highly overexpressed in the transgenic glands and providing further support that c‐myc functions as a mediator of apoptosis in the post‐lactational mammary gland. At day 2 of involution, MTB/TOM mice exhibited a marked yet variable increase in c‐MYC expression compared with MTB control mice (Figure 6C). Moreover, elevated protein levels of Bax, E2F‐1 and p53 were apparent in doxycycline‐induced MTB/TOM glands at day 2 of involution (Figure 6D), consistent with their corresponding genes being potential downstream targets of c‐MYC. No change in activated Stat3 was apparent in c‐MYC‐expressing mice.


Here we demonstrate that Socs3 is an important attenuator of pro‐apoptotic pathways in the mammary gland during lactation and involution. Targeted deletion of Socs3 in alveolar epithelial cells led to increased apoptosis during established lactation with evidence of tissue remodelling. More profoundly, deletion of Socs3 resulted in increased apoptosis and accelerated involution. Premature degradation and remodelling of the extracellular matrix, which normally occurs in the second phase of involution, was also evident in Socs3‐deficient mammary glands in early involution. Levels of activated Stat3 as well as activated MMP‐2 were substantially elevated in Socs3‐mutant glands. Our data implicate c‐myc, a target gene of Stat3, as an effector of apoptosis induced by the Jak1–Stat3 pathway during involution.

Socs3 is not a key attenuator of Prl signalling in the mammary gland

Prl signal transduction via the JAK2/STAT5 pathway is critical for the establishment and maintenance of alveolar cells during pregnancy and lactation (Hennighausen and Robinson, 2005). Given that Prl induces the expression of Socs3 in the mammary gland and that overexpression of Socs3 attenuates PRLR‐mediated responses in vitro, we examined the consequences of Socs3 loss during pregnancy using β‐lactoglobulinCre (Selbert et al, 1998) and MMTVCre (Wagner et al, 1997) transgenic mice. No histological differences were apparent between control and Socs3‐deleted glands at days 10.5, 12.5, 16.5 and 18.5 of pregnancy, indicating that Socs3 is not a key physiological attenuator of Prl signalling in the mammary gland (data not shown). Thus, Socs3 appears to have a distinct role from its closest SOCS family member (Socs1) in the mammary gland.

Socs3 is an important regulator of gp130 signalling in the mammary gland

Socs3 is a critical negative regulator of signalling mediated by the gp130 common receptor subunit in different cell types. Ablation of Socs3 in hepatocytes and macrophages led to prolonged Stat3 activation and altered gp130 signalling in response to IL‐6 (Croker et al, 2003; Lang et al, 2003; Yasukawa et al, 2003), and Socs3‐null mice show aberrant LIF signalling (Takahashi et al, 2003; Robb et al, 2005). Our data support a role for Socs3 as an important regulator of gp130 signalling in the mammary gland. The current model suggests that Socs3 binds with high affinity to the phosphorylated Y757 residue of the gp130 receptor chain, thus diminishing Stat3 activation (Nicholson et al, 2000; Schmitz et al, 2000). In agreement, augmented levels of phosphorylated Stat3 were observed in Socs3‐deficient mammary glands, and prolonged Stat3 activation occurred in MECs lacking Socs3 upon stimulation with either LIF or IL‐6.

Other IL‐6‐type cytokines may contribute to accelerated involution

There is considerable evidence for gp130–Stat3 signalling playing a major role in mammary gland involution, but it is likely that more than one cytokine is involved in this process. Mammary glands in gp130‐ and Stat3‐deficient mice, and LIF‐ and IL‐6‐null mice, all exhibit a delay in involution, but differences exist between the affected pathways. It seems unlikely that hyper‐responsiveness to LIF fully accounts for the phenotype observed in mammary glands lacking Socs3, as Akt signalling plays a prominent role in LIF‐mediated activation of Stat3 and apoptosis (Zhao et al, 2002, 2004; Kritikou et al, 2003), but not in Socs3‐deficient glands undergoing involution. These data suggest that Socs3 may attenuate signalling by other cytokine‐family members during involution. Oncostatin M (OSM) induces heterodimerisation of gp130 with the OSM‐specific β receptor (OSMR) subunit, resulting in activation of the JAK/STAT and MAP kinase signalling pathways (Heinrich et al, 2003). Interestingly, Socs3 is induced by OSM, and OSMR has recently been identified as a Stat3 target in KIM‐2 cells (Clarkson et al, 2006; Stross et al, 2006). OSMR signalling has also been implicated in tissue remodelling (Richards et al, 1993), suggesting that it may regulate mammary gland remodelling during involution. Indeed, OSMR‐knockout mice exhibit increased liver destruction, in part due to augmented MMP‐9 activity and reduced expression of the TIMP‐1 and TIMP‐2 genes (Nakamura et al, 2004).

c‐myc appears to be a central effector of apoptosis in mammary glands lacking Socs3

The c‐myc gene has emerged as a key target of Stat3 in the involuting mammary gland and a central pro‐apoptotic regulator during this process. Stat3 has been shown to directly activate c‐myc gene transcription in a number of different contexts (Kiuchi et al, 1999; Shirogane et al, 1999; Bowman et al, 2001; Cartwright et al, 2005). In the mammary gland, c‐myc expression is induced at the onset of involution, consistent with it representing a Stat3 target gene in mammary epithelium (Blakely et al, 2005). Mammary glands from both Socs3‐deficient and c‐MYC transgenic mice exhibited a marked increase in apoptosis and accelerated involution. No change in the level of activated Stat3 occurred in c‐MYC transgenic mice, compatible with Stat3 lying upstream of c‐MYC. Premature tissue remodelling was more pronounced in Socs3‐deficient glands, consistent with Stat3 regulating multiple target genes in addition to c‐myc.

A number of potential pro‐apoptotic targets of the c‐myc transcription factor have been identified, such as Bax, E2F‐1, p53 and Bim (Reisman et al, 1993; Mitchell et al, 2000; Tanaka et al, 2002; Egle et al, 2004). The expression of Bax was considerably elevated in both Socs3‐deficient mammary glands and c‐MYC transgenic mice. These observations are consistent with bax‐deficient mice exhibiting reduced mammary epithelial apoptosis (Schorr et al, 1999). Elevated levels of E2F‐1 and p53 were also evident in involuting mammary glands from Socs3‐deficient and c‐MYC transgenic mice.

In summary, our data implicate Socs3 in the regulation of a signalling network that controls the survival of mammary epithelium (Figure 7). Socs3 acts as a negative regulator of Jak1–Stat3 signalling during involution, with Stat3 as a central mediator. The c‐myc proto‐oncogene, a target gene of Stat3, may act at a nodal point in controlling apoptosis during involution. Perturbed c‐myc expression, in turn, is likely to activate (either directly or indirectly) the expression of pro‐apoptotic genes including E2F‐1, p53 and Bax, which leads to premature apoptosis and involution in mammary glands lacking Socs3.

Figure 7.

Proposed model by which Socs3 mediates precocious cell death during involution in mammary glands lacking Socs3. Under normal conditions, Socs3 serves to attenuate Stat3 signal transduction and therefore c‐myc induction. There may be other Stat3‐regulated pathways that act during involution in Socs3‐deficient mice.

Materials and methods

Mouse strains

All animal experiments were conducted according to the Melbourne Health Research Directorate Animal Ethics Committee guidelines. The Socs3‐null and conditionally targeted Socs3 mice have been described previously (Roberts et al, 2001; Croker et al, 2003). Mice expressing the WAPiCre transgene (Wintermantel et al, 2002) were initially bred onto a background heterozygous for the Socs3 null allele (Socs3+/−) and then mated with Socs3fl/fl mice to produce mice containing one Cre‐excised allele (Socs3Δ) with either a wild‐type or a null Socs3 allele. Mouse tail DNA was genotyped for wild‐type (+) null (−) or conditional (fl) Socs3 alleles by PCR using the following primers:



5′‐GATAACTGCCGTCACTCCAACG‐3′. WAPiCre transgenic mice were maintained on an FVB genetic background. GtRosa26 mice (Soriano, 1999) were generously provided by Dr P Soriano. The generation of the MMTV‐rtTA (MTB) transactivator and TetO‐MYC (TOM) responder lines has been described previously (D'Cruz et al, 2001). Bitransgenic MTB/TOM and littermate MTB control mice were administered doxycycline (1 mg/ml; Sigma) in their drinking water to induce expression of the MYC transgene. For the involution studies, adult female mice with litters of at least 6 pups were maintained. Pups were removed after 10 days to initiate involution.

Histology and whole‐mount staining for β‐galactosidase activity

For histological examination, mammary glands were fixed in 4% paraformaldehyde in phosphate‐buffered saline (PBS), embedded in paraffin and sections (1.5 μm) prepared and stained by haematoxylin and eosin. Mammary tissue from WAPiCre;GtRosa26 and GtRosa26 mice was stained for β‐galactosidase activity as described (Sum et al, 2005).

Northern blot and reverse transcription–PCR

Total RNA was isolated from mouse mammary glands using TRIzol reagent (Life Technologies) according to the manufacturer's instructions, and then treated with DNA‐free™ Kit (Ambion). A human c‐MYC cDNA restriction fragment was used to probe the Northern, followed by a 32P‐labelled 18S oligonucleotide. Primers used for RT–PCR of Bim from cDNA: forward 5′‐TCAGGAACCTGAAGATCTG‐3′ and reverse


Derivation of Socs3‐deficient mammary epithelial cells and cytokine stimulation

The thoracic and inguinal mammary glands were excised from day 12.5 pregnant Socs3+/fl and Socs3−/fl mice and MEC suspensions were prepared as described (Shackleton et al, 2006). MECs were cultured for 72 h prior to infection with a retrovirus expressing the human papilloma virus 16 proteins E6/E7 and a neomycin resistance cassette for immortalisation (kindly provided by D Galloway). Twenty‐four hours post‐infection, neomycin (200 μg/ml) was added to the cells, and selection continued for 14 days. To excise the floxed allele, immortalised epithelial cells were infected with a retrovirus expressing Cre‐ and puromycin‐resistance cassette, generously provided by K‐U Wagner (Krempler et al, 2002). Cells were selected in 1.5 μg/ml puromycin for 5 days. Socs3 MEC clones were genotyped by PCR as described (Croker et al, 2003), as shown in Supplementary Figure 4.

Socs3+/fl and Socs3−/Δ MEC clones (2 × 105) were plated in 6 cm2 dishes and starved of serum overnight. The following day, cells were stimulated with LIF (100 ng/ml). After harvesting, cells were lysed directly in 100 μl RIPA buffer supplemented with Complete protease inhibitor (Roche), 10 mM NaF, 1 mM Na3VO4 and 1 mM PMSF.

Western blot analysis and antibodies

Mammary gland protein lysates were prepared by grinding the tissue in liquid nitrogen and solubilizing in 1% TEB (150 mM NaCl; 5 mM EDTA, 50 mM Tris (pH 7.5); 0.1% NP‐40) supplemented with Complete protease inhibitor (Roche), 10 mM NaF, 1 mM Na3VO4 and 1 mM PMSF. Protein (50 μg) was resolved on polyacrylamide gels (Novex), before transfer to polyvinylidene difluoride membranes (Millipore). Non‐specific binding of proteins to membranes was blocked by incubation in PBS containing 0.1% Tween 20 and either 5% skim milk or 5% skim milk and 1% casein, before incubation with the primary antibody. These included α‐phospho‐STAT3 (Tyr705), α‐STAT3, α‐phospho‐p42/44 (Thr2020/Tyr204) and α‐p42/44 from Cell Signaling Technologies, α‐Bak (Ab‐1; Calbiochem), α‐Bax (2D2 and 5B7; Sigma), α‐Mcl‐1 (Rockland), α‐c‐myc (AbCam, ab11917; Santa Cruz, N‐262), α‐E2F‐1 (Zymed Laboratories), α‐SOCS‐3 (Immuno‐Biological Laboratories), α‐p53 (DO‐1, Santa Cruz), and α‐tubulin mAb (Sigma). Membranes were then incubated with horseradish peroxidase‐coupled secondary antibodies (Amersham Biosciences, Inc.) and developed by ECL (Amersham Biosciences, Inc.).


Mammary gland protein lysate (40 μg) was mixed with an equal volume of SDS sample buffer (50 mM Tris–HCl pH 6.8, 1% SDS, 0.025% bromophenol blue and 10% glycerol) and resolved on a 10% polyacrylamide gel containing 1 mg/ml gelatin (Sigma) (Lafleur et al, 2003b). Following electrophoresis, the gels were washed twice in buffer A (50 mM Tris–HCl, pH 8.0; 5 mM CaCl2; 2.5% Triton X‐100), before incubation in buffer B (50 mM Tris–HCl, pH 7.5; 5 mM CaCl2) overnight at 37°C. Gels were stained with 2.5 mg/ml Coomassie brilliant blue (R) dye (Bio‐Rad) in 10% acetic acid; 10% isopropanol for 2–4 h, then destained in 10% acetic acid; 10% isopropanol for 30–60 min. Gelatinolytic activity appeared as a clear band on a blue background.

TUNEL and bromodeoxyuridine immunodetection

Apoptotic nuclei were detected in paraformaldehyde‐fixed paraffin sections by TUNEL analysis using Terminal deoxynucleotidyl transferase (TdT; Promega) in the presence of biotinylated dUTP (Roche). Incorporated biotinylated dUTP was revealed by HRP‐conjugated streptavidin (LSAB2; Dako), followed by detection with DAB. TUNEL‐positive cells were scored by counting greater than 1000 epithelial nuclei in 10 random fields (400 × magnification) from each gland.

Mice were injected with BrdU Cell Labelling Reagent (0.5 mg/10 g body weight; Amersham Biosciences Inc.) 1 h before tissue collection and immunodetection was carried out as described (Sum et al, 2005).

Supplementary data

Supplementary data are available at The EMBO Journal Online (

Supplementary Information

Supplementary Figures [emboj7601455-sup-0001.pdf]

Supplementary Figure Legends [emboj7601455-sup-0002.doc]


We are grateful to D Smith for invaluable advice, D Krebs for expert help with AKT Western blots, B Duscio for excellent assistance, M Le Fleur for advice on zymography, M‐L Asselin‐Labat and H Barker for discussions and S Mihajlovic for histology. We also wish to thank M Ernst for critical review of the manuscript, as well as B Croker, C Ormandy, D Huang, D Smith, K‐U Wagner, D Galloway and P Humbert for generous gifts of mice, antibodies and plasmids. This work was supported by the Victorian Breast Cancer Research Consortium Inc., The Cancer Council Victoria and the National Health and Medical Research Council (Australia).


View Abstract