Insulin-like growth factors (IGFs) are important survival signals that can protect a range of cell types from apoptosis. Although IGF bioavailability is modulated by high affinity interactions with IGF-binding proteins (IGFBPs), there is currently no experimental evidence that IGFBPs regulate the survival function of IGFs in the mammary gland. We have examined IGFBP expression during mammary gland development and studied the effects of IGFBPs on IGF-mediated survival and signalling in mammary epithelial cells in culture. IGFBP-5 protein was greatly increased during days 1-3 of mammary gland involution, when levels of apoptosis are dramatically elevated to remodel the gland after lactation. Primary cultures of mammary epithelial cells (MECs) expressed IGFBP-5 from their basal surface suggesting that IGFBP-5 is suitably located to inhibit IGF signalling. Addition of exogenous IGFBP-5 and IGFBP-3 to MECs suppressed IGF-I-mediated survival, resulting in threefold greater apoptosis in cells incubated with IGF-I and IGFBP-5 compared with IGF-I alone. Examination of signalling pathways involved in apoptosis revealed that phosphorylation of PKB and the forkhead transcription factor, FKHRL1, was induced by IGFs, but that phosphorylation was blocked by IGFBP-5 and IGFBP-3. This study provides evidence that IGFBP-5 plays an important role in the regulation of apoptosis in the mammary gland.
Insulin-like growth factors (IGFs) play a pivotal role in tissue homeostasis, regulating cell proliferation, differentiation and migration during development and in the adult. In addition, IGFs are critical cellular survival factors. IGFs protect cells from apoptosis induced by a wide variety of conditions in culture including growth factor withdrawal, chemotherapy and oncogene expression (Muta and Krantz, 1993; Harrington et al., 1994; Sell et al., 1995; Kulik et al., 1997).
The actions of IGFs are modulated by high affinity interactions with a family of structurally related IGF-binding proteins (IGFBPs), IGFBP-1 to IGFBP-6 (Jones and Clemmons, 1995). IGFBPs are themselves subject to modification by proteolysis, post-translational alterations such as phosphorylation and interactions with cell surface and extracellular matrix components (Jones et al., 1993; Westwood et al., 1997). IGFBPs are known to regulate the bioavailability of IGFs in the circulation; however, their functions at the cellular level are not fully understood. IGFBPs have been reported to both inhibit and enhance IGF-I action depending on the system under investigation (Jones and Clemmons, 1995).
We have demonstrated that IGF-I suppresses apoptosis of primary mammary epithelial cells in culture (Farrelly et al., 1999). Mammary gland models also indicate the importance of IGFs as survival factors in vivo. In studies where transgenic mice overexpress IGF-I or II specifically in the mammary gland, apoptosis is reduced (Neuenschwander et al., 1996; Hadsell et al., 1996; Moorehead et al., 2001). Involution, the rapid induction of apoptosis that occurs to remodel the gland after lactation, was delayed as a result of reduced apoptosis in these animals. Regulating the availability of IGFs to mammary epithelial cells may therefore represent a physiological mechanism for initiating apoptosis during the process of involution.
All six IGFBP mRNAs are present in the mouse mammary gland, although IGFBP-3 and -5 are expressed most prominently in the stroma and also in ductal and alveolar epithelium in the pregnant gland (Wood et al., 2000). An increase in IGFBP-5 protein expression has been observed in rat milk after 48 hours of involution (Tonner et al., 1997). From these studies, it has been proposed that IGFBPs may regulate apoptosis in the mammary gland by blocking IGF-mediated survival. In this work, we therefore aimed to determine whether IGFBP-5 can directly regulate apoptosis by interfering with IGF signalling.
We provide conclusive evidence that IGFBP-5 is expressed during mammary gland involution when apoptosis levels are markedly increased, and demonstrate that IGFBPs are secreted into a cellular compartment where they can modulate IGF actions. We also observed that IGFBPs inhibit IGF-mediated survival signalling to cause apoptosis of mammary epithelial cells. This study therefore demonstrates a functional role for IGFBP-5 in the control of mammary cell survival. Additionally, our results provide a possible mechanism for the initiation of involution in the mammary gland.
Materials and Methods
Human recombinant IGF-I and IGF-II were obtained from Peprotech (London, UK) and R&D Systems (Abingdon, UK), respectively. Recombinant IGFBP-1 was obtained from Genentech (San Francisco, CA) and IGFBP-3 and -5 were obtained from Gropep (Adelaide, Australia). Polyclonal antibodies to phospho-PKBα (Ser 473) and PKB were obtained from New England Biolabs (Hitchin, UK). Monoclonal antibodies to E-cadherin and phosphotyrosine (clone 4G10) were purchased from BD Pharmingen (Oxford, UK). Polyclonal antibodies to IRS-1, IGFBP-3 and FKHRL1 were obtained from Upstate Biotechnology (Buckingham, UK). IGFBP-5 polyclonal antibody was purchased from Gropep. A monoclonal antibody to casein was kindly donated by C. S. Kaetzel, Case Western Reserve University, USA (Kaetzel and Ray, 1984). Unless otherwise stated, all other chemicals were obtained from Sigma (Poole, UK).
Protein and DNA extraction from mouse mammary gland tissue
Mammary tissue from ICR mice (Harlan Sera-Labs Ltd, Loughborough, UK) was collected from 8-week-old virgin animals (V) and mice pregnant for 9 and 15 days after detection of a vaginal plug (P9 and P15). At birth, pup numbers were normalized to 10 per dam. Mammary tissue was then removed from mice in full lactation for 9 days (L9), and from mice where the pups had been removed for up to 5 days following a 9-day lactation period (involution, I1-5). For immunoblot analysis, tissue was ground to a powder in liquid nitrogen and homogenized in lysis buffer containing 50 mM Tris (pH 7.4), 150 mM NaCl, 1% (w/v) Nonidet-P40, 1% (w/v) sodium deoxycholate, 0.1% SDS, 2 mM EDTA, 1 mM Na3VO4, 10 mM NaF, and protease inhibitors (Calbiochem, Nottingham, UK).
For analysis of DNA integrity, tissue was ground to a powder in liquid nitrogen and homogenized in 100 mM NaCl, 10 mM EDTA, 300 mM Tris, 200 mM sucrose, pH 8.0, 0.65% SDS before incubation for 30 minutes at 60°C. Following proteinase-K (Roche, Lewes, UK) digestion for 1 hour at 55°C, DNA was extracted using phenol chloroform as described previously (Pullan et al., 1996). Equal amounts of DNA were separated on a 1.2% agarose gel, stained with ethidium bromide and photographed.
Substrata and cell culture
Collagen-I-coated dishes were prepared by incubating dishes overnight at 4°C with rat tail collagen at 8 μg/cm2. The plates were washed extensively with phosphate-buffered saline before use. Growth-factor-reduced Matrigel (Becton Dickinson, Oxford, UK) was used to coat dishes at 14 mg/ml.
Primary mammary epithelial cells were prepared from 14.5-18.5-day pregnant ICR mice as previously described (Pullan and Streuli, 1996) and plated on different substrata in nutrient mixture F-12 (Biowhittaker, Wokingham, UK) containing 10% heat-inactivated fetal calf serum (Biowhittaker), 1 mg/ml fetuin (Sigma), 5 ng/ml EGF (Harlan Sera-Labs) and 1 μg/ml hydrocortisone (Sigma). After 48 hours, the medium was removed and fresh F-12 medium was added. After a further 24 hours, cells were serum-starved overnight in Dulbecco's modified Eagle's medium/nutrient mixture F-12 medium (DF-12; Biowhittaker).
To investigate the IGFBP expression profile of primary mammary epithelial cells in vitro, the serum-free medium removed after the 24-hour starvation process was concentrated (tenfold) and prepared for ligand blot analysis. To investigate whether IGFBP was secreted basally or apically, cells were grown on basement membrane (growth-factor-reduced Matrigel) in DF-12 differentiation medium containing 1 μg/ml hydrocortisone, 5 μg/ml insulin and 3 μg/ml prolactin for 4 days. After this time, the medium was replaced with serum-free medium or serum-free medium containing 2.5 mM EGTA to open tight junctions (Barcellos-Hoff et al., 1989). After 10 minutes, the medium was removed and was concentrated for ligand and immunoblot analysis of casein expression.
FSK-7 mouse mammary epithelial cells (Kittrell et al., 1992) were cultured in DMEM:F12 medium supplemented with 2% fetal calf serum. For signalling experiments, FSK-7 or primary mammary epithelial cells were treated with 1 nM IGF-I±IGFBP as appropriate for 15 minutes after serum-starvation. Following treatment, cells were harvested in lysis buffer as above.
Primary cultures of mammary epithelial cells grown on collagen I were serum-starved overnight and then were treated for 24 hours with 1 nM IGF-I±IGFBP as appropriate. Following treatment, the medium was removed and detached cells were pelleted by centrifugation. Attached cells were trypsinised and combined with detached cells. Cells were cytospun onto polysine slides (Merck, Poole, UK), and fixed in 3.7% paraformaldehyde for 10 minutes. Apoptosis was quantified by examining the nuclear morphology of cells stained with 4 μg/ml Hoechst 33258 (Molecular Probes, Leiden, The Netherlands). Each experiment was repeated at least three times and in each experimental condition more than 1000 single cells were scored for apoptosis.
For immunoprecipitations with mammary gland tissue, tissue lysate was incubated with 4 μl IGFBP-5 antibody for 16 hours at 4°C and then 50μ l protein A-Sepharose beads (Zymed Laboratories, Cambridge Bioscience, Cambridge, UK) were added for a further 2 hours. For primary mammary epithelial cells, cell lysate was incubated with 1 μg IRS-1 antibody for 1 hour at 4°C then 50 μl protein A-Sepharose beads were added for a further 4 hours.
Immunoblot and ligand blot analysis
Immunoprecipitates or tissue/whole cell lysates were subjected to SDS-PAGE and transferred to 0.2 μm nitrocellulose membrane (BioRad, Hemel Hempstead, UK). Membranes were blocked in 3% nonfat milk (Marvel) for 1 hour and incubated with antibodies to IGFBP-3 (1:1000), IRS-1 (1 μg/ml), phospho-PKB (1:1000), PKB (1:1000), FKHRL1 (1:750), E-cadherin (1:3000) or casein (1:4000) overnight at 4°C. Membranes were then incubated with secondary antibody (Amersham, Buckinghamshire, UK; 1:3000) for 1.5 hours at room temperature. Proteins were visualized using Supersignal West Femto (Perbio, Tattenhall, UK) or ECL (Amersham). In each of the studies presented, the results are typical of at least three independent experiments.
For ligand blot analysis, proteins were subjected to SDS-PAGE under non-reducing conditions. Following transfer to nitrocellulose membrane, the membrane was incubated in 0.5% (w/v) sodium azide and 1% (v/v) Nonidet-40 in phosphate-buffered saline (PBS) for 30 minutes at 4°C. The membrane was then blocked in 1% bovine serum albumin in PBS/0.5% Tween for 1 hour at room temperature before incubation with 125I-IGF-I (150,000 cpm/ml) for 3 hours. The ligand blot was then washed and exposed to film for 72 hours at -70°C.
Analysis was performed using the statistical software package Instat2 (GraphPad Prism, San Diego, USA). Comparisons of apoptosis values gained by each method were analysed using Student's t-test. A P-value less than 0.05 was considered to indicate statistical significance.
IGFBP-5 expression increases during mouse mammary gland involution in vivo
IGFBP-5 has been detected in rat milk (Tonner et al., 1997), but its levels through the pregnancy cycle have not been examined. To determine if IGFBP expression is increased in a manner that would support a role for regulation of IGF-mediated survival and initiation of post-lactational involution, the pattern of IGFBP expression was examined during mouse mammary gland development (Fig. 1). Ligand blot analysis using radiolabelled IGF-I indicated that an IGF-I binding protein with a similar molecular mass to IGFBP-5 (∼30 kDa) was expressed with low abundance during early pregnancy and increased slightly in late pregnancy. During lactation, protein expression was low, but increased dramatically during early involution (Fig. 1A). The identity of this upregulated binding protein was confirmed as IGFBP-5 by immunoprecipitating the tissue lysate with an antibody against IGFBP-5 prior to ligand blot analysis (Fig. 1B). To ensure that tissue samples contained equal epithelial cell content, the epithelial cell marker E-cadherin was used for normalization. IGFBP-1 and -3 could not be detected by ligand (Fig. 1A) or immunoblotting (data not shown) indicating that these IGFBPs are not highly expressed in the mouse mammary gland.
To confirm that IGFBP-5 expression is increased coincidently with apoptosis, DNA integrity was also assessed in samples from the same mammary glands used for protein analysis (Fig. 1C). DNA ladders were not observed in samples taken from virgin, pregnant and lactating mice; however, DNA internucleosomal fragmentation was marked at involution day 1, increasing at involution day 2. Thus, the profile of DNA fragmentation mirrors IGFBP-5 expression during mammary involution suggesting a link between IGFBP-5 secretion and apoptosis induction.
Mammary epithelial cells secrete IGFBPs in culture from the basal surface
In vivo, IGFBP-5 must have ready access to IGFs if it is to inhibit IGF-mediated survival. We therefore examined whether IGFBPs are secreted basally or into the lumen of cultured mammary pseudo-alveoli. Mouse mammary epithelial cells (MECs) grown on basement membrane in the presence of prolactin form partially differentiated alveolar-like structures with their basal cell surfaces orientated outwards towards the medium and their apical surfaces towards a sequestered lumen (Barcellos-Hoff et al., 1989). This model attempts to recapitulate mammary epithelial morphogenesis. Secretion of proteins from the basal surface can be detected in the medium surrounding these cells, and apical secretion into the lumen can be determined when medium is analyzed after the tight-junctions are opened using EGTA (Barcellos-Hoff et al., 1989).
Medium from differentiated MECs grown on basement membrane was replaced with EGTA-containing medium or medium alone before the medium was removed, and concentrated. Ligand blotting revealed the predominance of an IGF-I binding species that co-migrated with recombinant IGFBP-5 (Fig. 2). A faint band corresponding to the molecular weight of glycosylated IGFBP-3 was also detected. After EGTA-treatment, when the lumenal contents were released into the medium, levels of IGFBP-5 were not elevated indicating that MECs are able to secrete IGFBPs basally. To act as a positive control for tight junction opening, the distribution of casein was also measured and a protein of the same molecular weight as β-casein was found to be present only in EGTA-treated-primary cultures.
These results provide evidence that not only is IGFBP-5 expressed in mouse mammary gland and that its level of expression is increased at involution, but also that IGFBP-5 is synthesized by mammary epithelial cells and secreted basally into a tissue compartment where it could physiologically regulate IGF-I-mediated survival in vivo.
IGF-I-mediated survival is inhibited by IGFBP-5 and -3
To determine if IGFBPs act as physiological regulators of IGF function, we investigated whether IGF-mediated survival of primary MECs could be modulated by exogenous IGFBPs. Since IGFBP-5 was the most prominent IGFBP expressed in the mammary gland in vivo, we examined its effect on survival and apoptosis. Although expression of IGFBP-3 could not be detected in vivo, it was secreted by primary cultures so the effects of IGFBP-3 on IGF-mediated survival were also investigated. Apoptosis, determined by changes in nuclear morphology, was quantified in primary MECs exposed to IGF-I in the presence or absence of IGFBPs for 24 hours (Fig. 3). Serum withdrawal resulted in a fourfold increase in the number of cells undergoing apoptosis; but this was reversed by inclusion of 1 nM IGF-I (28.5±4% vs 7.9±1.6%; P=0.001). This confirms the importance of IGF-I as a survival factor in MECs (Fig. 3A,B) (Farrelly et al., 1999).
Recent studies have indicated that IGFBPs may act independently of IGFs to induce apoptosis in other cell systems (Nickerson et al., 1997; Perks et al., 2000); however, when MECs were incubated with 10 nM IGFBP-5 or -3 alone apoptosis was not significantly increased compared to serum-starved cells. A modest survival effect was seen although this was not statistically significant. Apoptosis was not maximal in these cells since apoptosis in the absence of serum can be further induced by other pharmacological agents (data not shown).
When IGFBP-5 or -3 was co-incubated with IGF-I, apoptosis was significantly higher than in cells incubated with IGF-I alone (21.3±5% vs 7.9±1.6%; P≤0.01) and the percentage of apoptotic cells was not significantly different from cells grown in serum-free medium (Fig. 3A,B). The concentrations of IGFBPs and IGFs used are physiologically relevant given that there is greater than tenfold molar excess of IGFBP-5 compared with IGF-I in the involuting mammary gland (Tonner et al., 2000). Inclusion of 5 nM IGF-II suppressed apoptosis and this was completely inhibited by 10 nM IGFBP-5 or -3 (data not shown).
These results demonstrate that although IGFBP-5 and IGFBP-3 do not induce apoptosis themselves, they can abrogate IGF-I and IGF-II-mediated survival.
IGF-initiated signalling is inhibited by IGFBP-5 and -3
To understand the mechanism responsible for IGFBP effects on IGF-initiated survival, we examined changes in phosphorylation of key signalling molecules. We concentrated on the insulin-receptor substrate/phosphatidylinositol 3-kinase/protein kinase B pathway because 1 nM IGF-I induced phosphorylation of this pathway but did not induce mitogen-activated protein kinase phosphorylation at this concentration (data not shown). In FSK-7 mouse MECs, IGF-I (1 nM, 15 minutes) stimulated a dramatic increase in insulin-receptor substrate 1 (IRS-1) phosphorylation (Fig. 4A, lane 2) which was inhibited in cells co-incubated with 5 nM IGFBP-5 or 5 nM IGFBP-3 (Fig. 4A, lanes 5,6). Similarly, IGFBP-5 inhibited IGF-I-mediated IRS-1 phosphorylation in primary cultures of MECs (Fig. 4B). Neither IGFBP-5 nor -3 had independent effects on IRS-1 phosphorylation (Fig. 4A, lanes 3,4, and data not shown).
To determine if these signalling changes were reflected by alterations in downstream effectors, phosphorylation of the serine/threonine kinase protein kinase Bα (PKB) at Ser473 was investigated (Fig. 5). PKB phosphorylation was stimulated when MECs were incubated with 1 nM IGF-I for 15 minutes (Fig. 5A,B, lane 3) to a similar level seen with cells grown in the presence of serum (data not shown). IGF-initiated PKB phosphorylation was however inhibited in the presence of 5-10 nM IGFBP-5 and 1-10 nM IGFBP-3 (Fig. 5A,B, lanes 4-6). PKB was not phosphorylated under control conditions or when cells were exposed to 10 nM IGFBP-5 or -3 alone (Fig. 5A,B, lanes 1,2).
We also determined if IGFBP-5 and -3 could modulate downstream signals mediated by IGF-II (Fig. 5C). PKB phosphorylation was observed with 5 nM IGF-II alone (Fig. 5C, lane 2) and was inhibited by a twofold molar excess of IGFBP-3 and -5 (10 nM) (Fig. 5C, lanes 3,4). Together, the results indicate that IGFBPs can potently modulate intracellular signals mediated by both IGF-I and IGF-II.
A number of substrates for PKB have been proposed; however, the Forkhead transcription factor FKHRL1, has been particularly implicated in the modulation of apoptosis (Brunet et al., 1999; Zheng et al., 2000). Phosphorylation by PKB results in FKHRL1 inhibition and retention in the cytoplasm. In contrast, dephosphorylation leads to FKHRL1 activation, translocation to the nucleus and subsequent transcription of genes encoding death-activating proteins (Brunet et al., 1999). We therefore examined whether IGFs could mediate the phosphorylation of FKHRL1 in mammary epithelial cells and if this is inhibited by IGFBPs. In contrast to untreated cells and cells exposed to 10 nM IGFBP-5, a shift in FKHRL1 mobility, indicative of phosphorylation, was observed when cells were exposed to 1 nM IGF-I (Fig. 6A, lanes 1-3). A similar shift in FKHRL1 mobility has been attributed to phosphorylation in IGF-I-treated fibroblasts (Brunet et al., 1999). When MECs were incubated with IGF-I in the presence of IGFBP-5, this mobility shift was not observed (Fig. 6A, lane 4). Similarly, IGF-II induced a shift in FKHRL1 mobility that was inhibited by 10 nM IGFBP-5 or IGFBP-3 (Fig. 6B).
Together, these results demonstrate that, in common with IRS-1 and PKB, FKHRL1 phosphorylation is modulated by IGFBP-5 and -3 and suggest that FKHRL1 may be an important mediator of the actions of IGFs in the mammary gland.
This study provides evidence that IGFBP-5 is a physiological regulator of epithelial cell survival in the mammary gland. Previous reports indicate that IGFBP-5 may be associated with increased apoptosis. For example, IGFBP-5 co-localizes with areas of apoptosis in the interdigital zone of the developing mouse limb bud (van Kleffens et al., 1998); however, we now provide experimental evidence that IGFBP-5 can directly regulate apoptosis by interfering with IGF signalling in the mammary gland.
The implications of this finding are significant given that IGFs are known survival factors for several cell types. IGF-I is able to protect cells from apoptosis under a wide variety of circumstances, including growth factor withdrawal in haematopoietic and neuronal cells; overexpression of myc in fibroblasts; chemotherapy, and UV-B irradiation (Muta and Krantz, 1993; Harrington et al., 1994; Sell et al., 1995; Kulik et al., 1997; Datta et al., 1997). IGF-II can also act as a survival factor; for example, IGF-II suppresses hepatocyte apoptosis triggered by deregulated N-myc expression and can block cell death during oncogenesis in vivo (Christofori et al., 1994; Ueda and Ganem, 1996).
IGFBP-5 expression is increased during involution when apoptosis is maximal
Previous studies have shown that IGFBP-5 increases in rat milk after 48 hours of involution (Tonner et al., 1997). We now reveal its pattern of tissue expression through a more extensive developmental analysis in a murine model. Levels of IGFBP are not greatly altered during the transition from ductal epithelium in the non-pregnant mammary gland to alveolar epithelium in pregnancy and lactation. However, IGFBP-5 is dramatically upregulated at day 1 of involution, and was maximal at days 2 and 3, decreasing at day 5. This profile is consistent with the pattern of DNA fragmentation, providing correlative evidence for the involvement of IGFBP-5 in mammary apoptosis. This observation has also been supported by recent in vivo data (Tonner et al., 2002).
IGFBPs are appropriately located to inhibit the actions of IGFs
The previous observation that IGFBP-5 is present in large quantities in milk (Tonner et al., 1997) could suggest that the IGFBP-5 expressed in involution may not be able to access the cellular compartment basal to epithelial cells where the actions of IGFs take place. We show experimentally that IGFBP-5 and -3 are secreted basally from the alveolar-like structures formed by MECs in culture. This result leads us to suggest that IGFBP-5 synthesized and secreted by mammary epithelial cells, binds to IGF-I within the mammary stroma and prevents an interaction between IGF-I and the type I IGF receptor (IGF-IR) on the basal cell surface.
IGFBPs block IGF-mediated survival signalling leading to increased apoptosis
IGFBP-5 and -3 inhibit IGF-mediated survival in our model. Studies using cultured preovulatory follicles, osteosarcoma cells and neuronal cells indicate that IGFBP-3 can inhibit IGF-mediated survival (Chun et al., 1994; Niikura et al., 2001; Schmid et al., 2001). To our knowledge, this is the first report demonstrating that IGFBP-5 can also regulate IGF-mediated survival.
Recently, a number of studies indicate that IGFBPs may act not only via sequestration of IGF-I but independently of IGF-I to regulate apoptosis. IGFBP-3 induces apoptosis in MCF-7 breast cancer cells (Nickerson et al., 1997) and accentuates ceramide-induced apoptosis in Hs578T breast cancer cells (Perks et al., 1999; Perks et al., 2000). However, the situation appears to be different in normal mammary epithelial cells since treatment with either IGFBP-3 or -5 alone does not trigger apoptosis beyond that induced by serum-withdrawal. These differences may alternatively reflect the time course and IGFBP concentration used.
Our results reveal a possible mechanism for the effects of IGFBPs on mammary epithelial apoptosis (Fig. 7). IRS-1 phosphorylation is inhibited when IGFBP-5 and -3 and IGFs are co-incubated. This is in contrast to studies in MCF-7 human breast carcinoma cells where IGFBP-3, but not IGFBP-5, inhibited IRS-1 phosphorylation (Ricort and Binoux, 2001). We also observed that IGFBPs inhibit both IGF-I- and IGF-II-mediated signalling via PKB (Fig. 7). PKB has a proven role in promoting survival (Datta et al., 1997) and has been implicated in the regulation of involution. Expression of a constitutively active form of PKB in the mammary gland of transgenic mice delays onset of apoptosis after weaning (Hutchinson et al., 2001; Schwertfeger et al., 2001). Additionally, sustained phosphorylation of PKB correlates with the reduced apoptosis and delayed involution observed in transgenic mice overexpressing IGF-II (Moorehead et al., 2001). Thus, modulation of PKB phosphorylation via IGFBP-5 inhibition of IGF-I or -II signalling may be an important mechanism in involution.
IGF-I has been shown to phosphorylate FKHRL1 via PKB in other cell types such as haematopoietic cells (Brunet et al., 1999), but this has not been studied in the mammary gland. Other growth factors, such as epidermal growth factor (Jackson et al., 2000) and vascular endothelial growth factor (Price et al., 2001) have been shown to activate FKHRL1 in human breast cancer cells, but this is the first observation that IGF-I and -II can phosphorylate FKHRL1 in non-transformed mammary epithelial cells. We also demonstrate that FKHRL1 phosphorylation is blocked by IGFBP-5 and -3 (Fig. 7). Studies concerning the effect of FKHRL1 on apoptosis have mainly used ectopically expressed FKHRL1. The Fas ligand gene was shown to be a target for FKHRL1 transcription in fibroblasts, cerebellar granule neurons and Jurkat T lymphocytes (Brunet et al., 1999); however, in murine Ba/F3 haematopoietic cells, FKHRL1 induces apoptosis through a death-receptor-independent pathway that involves transcriptional upregulation of the pro-apoptotic Bcl-2 family member, Bim (Dijkers et al., 2000; Dijkers et al., 2002). Although we cannot rule out other PKB substrates as additional effectors, activation of FKHRL1 provides a potential mechanism to explain the increase in apoptosis observed when IGF signalling is inhibited by IGFBPs.
Our studies indicate that IGFBP-5 is expressed in mouse mammary gland in a time frame corresponding to apoptosis and is synthesized and secreted basally where it may modulate IGF-actions. Additionally, our culture experiments reveal that IGFBPs inhibit IGF-mediated signalling through PKB and FKHRL1 and stimulate apoptosis. These studies provide experimental evidence that IGFBP-5 regulates apoptosis by modulating the survival function of IGFs.
The support of the BBSRC is gratefully acknowledged. We would also like to thank Salford Royal Hospital NHS Trust for financial support.
- Accepted November 5, 2002.
- © The Company of Biologists Limited 2003