Accumulating evidence indicates that heregulins, EGF (epidermal growth factor)-like ligands, promote breast cancer cell proliferation and are involved in the progression of breast cancer towards an aggressive and invasive phenotype. However, there is limited information regarding the molecular mechanisms that mediate these effects. We have recently established that HRG (heregulin β1) promotes breast cancer cell proliferation and migration via cross-talk with EGFR (EGF receptor) that involves the activation of the small GTPase Rac1. In the present paper we report that Rac1 is an essential player for mediating the induction of cyclin D1 and p21Cip1 by HRG in breast cancer cells. Inhibition of Rac function by expressing either the Rac-GAP (GTPase-activating protein) β2-chimaerin or the dominant-negative Rac mutant N17Rac1, or Rac1 depletion using RNAi (RNA interference), abolished the cyclin D1 and p21Cip1 induction by HRG. Interestingly, the proliferative effect of HRG was impaired not only when the expression of Rac1 or cyclin D1 was inhibited, but also when cells were depleted of p21Cip1 using RNAi. Inhibition of EGFR, PI3K (phosphoinositide 3-kinase; kinases required for Rac activation by HRG) or MEK [MAPK (mitogen-activated protein kinase)/ERK (extracellular-signal-regulated kinase) kinase] also blocked the up-regulation of cyclin D1 and p21Cip1 by HRG. In addition, we found that HRG activates NF-κB (nuclear factor κB) in a Rac1- and MEK-dependent fashion, and inhibition of NF-κB abrogates cyclin D1/p21Cip1 induction and proliferation by HRG. Taken together, these findings establish a central role for Rac1 in the control of HRG-induced breast cancer cell-cycle progression and proliferation through up-regulating the expression of cyclin D1 and p21Cip1.
Heregulins (also called neuregulins), widely expressed EGF (epidermal growth factor)-like ligands, have been implicated in the progression of different types of cancers, particularly breast cancer [1–3]. Heregulins are expressed in approx. 30% of breast tumours [4,5] and have been associated with the progression of breast cancer towards an aggressive and invasive phenotype [2,6]. Expression of heregulin β2 in MCF-7 breast cancer cells induces oestrogen-independent tumour formation in ovariectomized nude mice . Moreover, blockade of heregulin β2 expression inhibits tumorigenicity and metastasis of breast cancer cells . Heregulins bind to ErbB3 and ErbB4 receptors to promote their dimerization with distinct members of the ErbB receptor family. Indeed, ErbB2/ErbB3 and ErbB2/ErbB4 dimers can be readily detected upon heregulin stimulation [8–10]. Despite their well-established roles in tumour development and progression, the mechanisms by which heregulins exert their tumorigenic effects are not fully understood. Recent studies have shown that HRG (heregulin β1)-induced breast cancer cell proliferation and migration are mediated not only by ErbB3 and ErbB2, but also by EGFR (EGF receptor), as judged by the ability of EGFR inhibitors and EGFR RNAi (RNA interference) to impair HRG mitogenic and motogenic effects . The involvement of multiple ErbB receptors as mediators of HRG responses predicts a complex array of signalling pathways activated by this ligand.
The small Rho GTPases have been established as important mediators of responses elicited by tyrosine kinase and G-protein-coupled receptors. Rac1, a member of the Rho GTPase family, is known to play crucial roles in actin cytoskeleton organization, cell proliferation, transformation, migration, invasion and metastasis [12–14]. Rac1 becomes activated by Rac-GEFs (guanine-nucleotide-exchange factors) and inactivated by Rac-GAPs (GTPase-activating proteins), and a misbalance in the expression or function of these Rac regulators may contribute to cancer progression. In breast cancer, for example, overexpression of Rac, up-regulation of Rac-GEFs (such as Tiam1) and down-regulation of Rac-GAPs (such as β2-chimaerin) have been reported [15–20], arguing for the involvement of Rac and its effectors in breast cancer progression and/or the maintenance of the malignant phenotype. The Rac effector Pak1 (p21-activated kinase 1) is also overexpressed in breast tumours and plays an important role in breast cancer metastasis [21,22]. We have recently established an essential role for Rac1 in the mitogenic response of HRG in breast cancer cell models . Stimulation of breast cancer cells with HRG promotes a strong activation of Rac1, an effect mediated by ErbB3 and ErbB2 and independent of ErbB4. Unlike EGF, HRG causes a sustained Rac activation. Moreover, EGFR and PI3K (phosphoinositide 3-kinase) are also required for Rac activation by HRG. Despite these observations, there is a limited knowledge of the mechanisms downstream of Rac1 that mediate the HRG proliferative response.
In the present study we demonstrate a key role for Rac1 in regulating cyclin D1 induction in response to HRG stimulation in breast cancer cells, and more interestingly, we found that HRG caused a sustained elevation in the levels of the cell-cycle inhibitor p21Cip1 that is required for the mitogenic effect of HRG. Like cyclin D1, p21Cip1 induction by HRG is also dependent on Rac1. In addition, we established a role for the NF-κB (nuclear factor κB) pathway in the proliferative effect of HRG via Rac1. Our results support a central role for Rac1 in the regulation of HRG-mediated breast cancer cell mitogenicity and argue for the involvement of a complex set of signalling events regulated by this small G-protein that contribute to the HRG response.
MATERIALS AND METHODS
Cell culture, chemicals and plasmids
Human breast cancer T-47D cells were purchased from ATCC and cultured in RPMI 1640 medium supplemented with 10% FBS (foetal bovine serum), 2 μg/ml insulin, 1 mM sodium pyruvate and 2 mM glutamine, at 37 °C in a humidified 5% CO2 atmosphere. HRG was purchased from R&D Systems. Insulin, actinomycin D and cycloheximide were obtained from Sigma. Wortmannin and AG1478 were from LC Laboratories. U0126 was from Calbiochem. A pcDNA3 vector encoding for the NF-κB super suppressor (IκBαM; mutant inhibitory κBα)  was a gift from Dr Han-Ming Shen (Department of Community, Occupational and Family Medicine, National University of Singapore, Singapore).
Generation of AdVs (adenoviruses) and infection
The generation of AdVs for β2-chimaerin and LacZ has been described previously . N17Rac1-AdV was a gift from Dr Anne Ridley (Ludwig Institute for Cancer Research, University College London, U.K.). Serum-starved (for 8 h) T-47D cells were infected with the corresponding AdVs for 16 h, and then AdVs were removed by extensive washing. Maximum protein expression was generally observed 24 h after removal of the AdV and remained stable for at least 3 additional days. Experiments were performed 48 h after infection.
siRNA (small interfering RNA) duplexes were purchased from Dharmacon Research. The target sequences were AAACACGCGCAGACCTTCG (cyclin D1) and AACATACTGGCCTGGACTGTT (p21Cip1). The Rac1 and control siRNA sequences have been described previously . siRNA duplexes (100 nM) were transfected into T-47D cells using Lipofectamine™ 2000 (Invitrogen) in serum-free medium, and 24–48 h later cells were stimulated with HRG (10 ng/ml) for different times.
RT (reverse transcriptase)-PCR and Q-PCR (quantitative PCR)
Total RNA was prepared using TRIzol® according to the manufacturer's protocol and reverse transcribed using SuperScript™ II RT (Invitrogen). The resulting cDNA was used for PCR amplification using the following primers: cyclin D1 (forward), 5′-TGTGCTGCGAAGTGGAAACC-3′; cyclin D1 (reverse), 5′-CCATTTGCAGCAGCTCCTCG-3′; p21Cip1 (forward), 5′-GCGATGGAACTTCGACTTTGT-3′; p21Cip1 (reverse), 5′-GGGCTTCCTCTTGGAGAAGAT-3′; GAPDH (glyceraldehyde-3-phosphate dehydrogenase; forward), 5′-CCCTTCATTGACCTCAACTACATGG-3′; and GAPDH (reverse), 5′-CATGGTGGTGAAGACGCCAG-3′.
For Q-PCR, RNA (75 ng) was reverse transcribed in a 10 μl reaction mixture with Multiscribe RT (Applied Biosystems) using random hexamer primers. Taqman Q-PCR was carried out using 2 μl of cDNA in a 22 μl reaction mixture (Applied Biosystems). Primers and probes were as follows: cyclin D1 (forward primer), 5′-TGTTCGTGGCCTCTAAGATGAAG-3′; cyclin D1 (reverse primer), 5′-AGGTTCCACTTGAGCTTGTTCAC-3′; cyclin D1 (probe), 6FAM-AGCAGCTCCATTTGCAGCAGCTCCT-TAMRA (where FAM is 6-carboxyfluorescein and TAMRA is 6-carboxytetramethylrhodamine); 18S (forward primer), 5′-CCTGGTTGATCCTGCCAGTAG-3′; 18S (reverse primer), 5′-CCGTGCGTACTTAGACATGCA-3′; and 18S (probe), VIC®-TGCTTGTCTCAAAGATTA-MGBNFQ. For p21Cip1, the Assay-on-Demand primer–probe mix (Applied Biosystems) was used. cDNA amplification was carried out in an ABI Prism 7000 Sequence Detection System using the standard curve method. Cyclin D1 and p21Cip1 mRNA levels were normalized to 18S rRNA.
Western blot analysis
Cells were lysed and subjected to SDS/PAGE as described previously  using 10–30 μg of protein/lane. The following antibodies were used: anti-Rac and anti-cyclin D1 (Upstate Biotechnology); anti-p27 and anti-p53 (Santa Cruz Biotechnology); anti-p21Cip1, anti-HA (haemagluttinin) tag, anti-Myc tag, antiphospho-Rb (retinoblastoma; Ser780), anti-phospho-IκBα (inhibitory κBα; Ser32), anti-total-IκBα, and anti-phospho-ERK1/2 (extracellular-signal-regulated kinase; Thr202/Tyr204) (Cell Signaling Technology); anti-phospho-Rb (retinoblastoma; Thr821) (Abcam); anti-β2-chimaerin ; and anti-β-actin (Sigma).
Rac-GTP pulldown assays
After 48 h of serum starvation, cells were stimulated with HRG (10 ng/ml) for 5 min. Maximum Rac stimulation was observed at this time . Rac-GTP levels were determined with a PBD (Pak p21-binding domain) pulldown assay, as described previously , using an anti-Rac antibody for Western blot detection.
T-47D cells in 60 mm dishes were transfected with 0.5 μg of a plasmid encoding the NF-κB super suppressor pIκBαM  or empty vector (pcDNA3) using Lipofectamine™ 2000 (Invitrogen) in serum-free medium, and 24–48 h later cells were stimulated with HRG (10 ng/ml) for different times, as indicated in the corresponding experiments.
NF-κB luciferase reporter assay
T-47D cells in 6-well plates were co-transfected with a NF-κB-luciferase vector (pNFκB-luc, 0.5 μg) and a Renilla-luciferase vector (pRL-CMV, 10 ng), together with a vector expressing NF-κB super-suppressor (pIκαM, 0.5 μg) or an empty vector (pcDNA3, 0.5 μg). After 24 h serum starvation, cells were stimulated with HRG (10 ng/ml) for 24 h and luciferase activity was measured using the Promega dual-luciferase reporter assay system. Results are presented as means±S.D. (n=3).
Cells were serum-starved for 24 h and stimulated with 10 ng/ml HRG twice every 24 h (a protocol that maximized the HRG response). Cells were harvested 48 h later by trypsinization, and the cell number was determined with a haemocytometer.
Results are presented as means±S.D. or means±S.E.M., and were analysed using either a Student's t test or ANOVA with a Scheffe's test. A P value of <0.05 was considered statistically significant.
Cyclin D1 and p21Cip1 induction are required for HRG-induced breast cancer cell proliferation
We have recently established that HRG induces breast cancer cell proliferation via ErbB3/ErbB2 cross-talk with EGFR . To initiate the characterization of the relevant cell-cycle players involved in this mitogenic effect, T-47D breast cancer cells were serum-starved, treated with HRG (10 ng/ml), and at different times cell lysates were subjected to Western blot analysis for various cell-cycle regulators. Figure 1(A) shows that HRG caused a significant elevation in cyclin D1 levels, starting at 6 h after treatment. In addition, HRG caused a marked increase in p21Cip1 levels, but it did not appreciably affect the levels of other negative cell-cycle regulators expressed in T-47D breast cancer cells, such as p53 and p27. p21Cip1 up-regulation was noticeable at 3 h and persisted for at least 12 h after HRG treatment. Increased pRb phosphorylation by HRG in Ser780 (the cyclin D-CDK4/6 site, where CDK is cyclin-dependent kinase) and Thr821 (the cyclin E-CDK2 site) was observed at 12 and 24 h after treatment respectively (Figure 1A). These results suggest that p21Cip1 up-regulation in this context was not inhibitory for the activities of CDK4/6 and CDK2. HRG stimulation also markedly increased cyclin D1 and p21Cip1 mRNA levels in a time-dependent manner, as determined by RT-PCR (Figure 1B). These observations were confirmed by Q-PCR (see Figure 3 and results not shown). Elevations in cyclin D1 and p21Cip1 protein levels were abolished by inhibitors of transcription (actinomycin D) and translation (cycloheximide), suggesting that the effects may be at a transcriptional level in either case (Figure 1C). Next, to determine whether a causal relationship existed between cyclin D1 and p21Cip1 up-regulation and HRG-induced cell proliferation, we used a RNAi knockdown approach. As shown in Figure 1(D), upon delivery of specific dsRNAs (double stranded RNAs) for either cyclin D1 or p21Cip1, which caused 60±5% and 71%±4% depletion respectively, HRG failed to cause a significant elevation in either cell-cycle regulator. No effect was observed upon delivery of an unrelated (control) RNAi duplex. The specificity of cyclin D1 and p21 RNAi was demonstrated by RT-PCR (Supplementary Figure 1 at http://www.biochemj.org/bj/410/4100167add.htm). As predicted, the proliferative response of HRG was markedly reduced in cyclin D1-depleted cells. Notably, a similar inhibition of proliferation was observed upon p21Cip1 RNAi (Figure 1E), suggesting that up-regulation of both cyclin D1 and p21Cip1 was required for the mitogenic activity of HRG in breast cancer cells.
HRG promotes breast cancer T-47D cell proliferation via cyclin D1 and p21Cip1
Induction of cyclin D1 and p21Cip1 by HRG is Rac1-dependent
Our recent studies showing an essential role for Rac1 in the mitogenic response of HRG in breast cancer cells  prompted us to examine whether this small G-protein plays any role in the induction of cyclin D1 and p21Cip1 by HRG. Three approaches were used to address this issue. First, we ectopically expressed the Rac-specific GAP β2-chimaerin, which inhibits Rac activation and proliferation by HRG in breast cancer cells . β2-Chimaerin inhibits Rac without interfering with RhoA or Cdc42 activities, both in vitro and in cellular models, including breast cancer cells [11,18,24]. β2-Chimaerin was delivered into T-47D cells by adenoviral means, using a range of MOIs (multiplicities of infection). Figure 2(A) shows that expression of β2-chimaerin (MOI=3–30 pfu/cell; where pfu is plaque-forming unit) dose-dependently inhibited Rac activation by HRG, as previously described . A control LacZ AdV (MOI=30 pfu/cell) showed no effect. Interestingly, induction of either cyclin D1 or p21Cip1 by HRG was markedly impaired in β2-chimaerin AdV-infected cells. The effect was proportional to the expression levels of the Rac-GAP. Moreover, a striking correlation between the inhibition of Rac activation and the induction of either cyclin D1 or p21Cip1 by HRG was observed (R=0.97 and 0.99 respectively) (Figure 2B). As a second approach we used an AdV encoding a dominant-negative Rac1 mutant (Myc-tagged N17Rac1). Upon delivery of this mutant (MOI=1–10 pfu/cell), expression of N17Rac1 could be readily detected by Western blot analysis, and it impaired dose-dependently the activation of Rac in response to HRG (Figure 2C). As observed with the Rac-GAP, N17Rac1 also inhibited cyclin D1 and p21Cip1 induction by HRG (Figure 2C). Both β2-chimaerin and N17Rac1 inhibited HRG-induced proliferation in a MOI-dependent manner (Figure 2D). Lastly, to confirm further the involvement of Rac1 in cyclin D1 and p21Cip1 up-regulation we used Rac1 RNAi, which reduced Rac1 expression by 78±6% (n=3) (Figure 3A). HRG failed to up-regulate cyclin D1 and p21Cip1 in Rac1-depleted T-47D cells (Figure 3A). Figures 3(B) and 3(C) show that Rac1 RNAi significantly reduced the elevations of cyclin D1 and p21Cip1 mRNA levels. As observed with the Rac-GAP and dominant negative Rac1, depletion of Rac1 using RNAi also significantly inhibited HRG-induced cell proliferation (Figure 3D). Taken together, these results strongly suggest that Rac activation is required for cyclin D1 and p21Cip1 induction by HRG.
Induction of cyclin D1 and p21Cip1 by HRG is Rac-dependent: correlation with proliferation
Rac1 RNAi impairs HRG-induced up-regulation of cyclin D1 and p21Cip1 and cell proliferation
EGFR, PI3K and MEK [MAPK (mitogen-activated protein kinase)/ERK kinase] mediate cyclin D1 and p21Cip1 induction by HRG
In a recent study we showed that HRG-induced Rac activation requires PI3K. Rac1 mediates ERK1/2 activation by HRG in breast cancer cells and inhibition of Rac1 impairs ERK1/2 activation and proliferation by HRG . Figure 4(A) shows that the MEK inhibitor UO126 efficiently blocked the induction of cyclin D1 and p21Cip1 by HRG. Likewise, the PI3K inhibitor wortmannin impaired HRG-induced up-regulation of cyclin D1 and p21Cip1 (Figure 4B).
Inhibition of MEK, PI3K or EGFR impairs cyclin D1 and p21Cip1 up-regulation by HRG
We have recently shown that HRG-induced activation of Rac and proliferation involves cross-talk with the EGFR, as both responses can be impaired by the EGFR inhibitors AG1478 and gefitinib (Iressa), as well as by EGFR RNAi . Figure 4(C) shows that AG1478 markedly inhibited the induction of cyclin D1 and p21Cip1 by HRG, further supporting the mechanistic link between HRG stimulation and the induction of cyclin D1 and p21Cip1 via Rac.
Involvement of NF-κB in HRG-induced proliferation and signalling in breast cancer cells
Previous studies have shown that HRG activates NF-κB in breast cancer cells [25,26]. We hypothesized that this pathway could be involved in the induction of cyclin D1 and p21Cip1 by HRG. It was found that HRG stimulation elevates NF-κB activity, as determined using a luciferase reporter assay (Figure 5A). Figure 5(B) shows that HRG activates NF-κB in T-47D breast cancer cells in a time-dependent manner, as judged by Western blot analysis using an anti-phospho-IκBα antibody. Phosphorylation of IκBα became evident 30 min after HRG stimulation and lasted for at least 3 h. A time-course analysis in the same samples revealed that IκBα phosphorylation preceded cyclin D1 and p21Cip1 up-regulation (Figure 5B). To determine the involvement of NF-κB, we expressed the NF-κB super suppressor IκBαM  in T-47D cells, and found that it markedly inhibited cyclin D1 and p21Cip1 up-regulation by HRG (Figure 5C), and importantly, it also inhibited the proliferative response of HRG (Figure 5D). These findings strongly argue for the functional involvement of NF-κB in cyclin D1 and p21Cip1 up-regulation by HRG.
NF-κB is involved in HRG-induced cyclin D1 and p21Cip1 up-regulation and cell proliferation
Next, to establish further the sequence of events leading to cyclin D1 and p21Cip1 induction by HRG in breast cancer cells, we examined whether Rac1 was involved in NF-κB activation. Figure 6(A) shows that T-47D cells subject to Rac1 RNAi have impaired IκBα phosphorylation in response to HRG stimulation. The MEK inhibitor UO126 also significantly impaired IκBα phosphorylation by HRG (Figure 6B). Furthermore, the PI3K inhibitor wortmannin, which blocks Rac and ERK activation by HRG, also abrogated IκBα phosphorylation (Figure 6C) and reduced cell proliferation by ∼30% (results not shown). Taken together, these results indicate that HRG activates NF-κB via a Rac1/ERK-dependent pathway.
HRG activates NF-κB via Rac/ERK
Although accumulating evidence indicates that heregulins promote breast cancer cell proliferation and breast tumorigenesis, the detailed mechanisms involved in these effects are not well understood. Our recent studies have established an essential role for the small GTPase Rac1 in HRG mitogenic signalling, which involves the activation of EGFR and is independent of ErbB4 . HRG triggers a sustained Rac activation in breast cancer cells, which is dependent on ErbB3, ErbB2, EGFR and PI3K. Rac activation by HRG promotes breast cancer cell proliferation mainly through the MEK/ERK pathway . In the present study, we demonstrate that in addition to stimulating cyclin D1 expression, HRG also elevates the levels of p21Cip1, a well-established negative regulator of the cell cycle [27,28]. We found that HRG-induced proliferation via Rac1 depends not only on cyclin D1 but, paradoxically, also on p21Cip1.
Cyclin D1 is overexpressed in approx. 30–50% of breast tumours and plays important roles in the development of breast cancer [29,30]. Cyclin D1 is one of the essential cyclins that regulate G1- to S-phase transition during normal cell-cycle progression, and it is also an essential component of hormone- and growth factor-induced mitogenesis in breast epithelial cells. In many breast cancers, overexpression of cyclin D1 could not be explained by gene amplification, and possibly involves epigenetic mechanisms. ErbB receptors and their ligand-triggered oncogenic signals are known to up-regulate cyclin D1 both at transcriptional and post-transcriptional levels . Although studies have shown that cyclin D1 protein levels are elevated in response to HRG in NIH 3T3  or breast cancer T-47D cells , the mechanisms involved are not well understood. In the present study we demonstrate that elevation in cyclin D1 expression both at mRNA and protein levels is dependent on Rac1, as evidenced by the ability of a dominant-negative Rac1 mutant (N17Rac1), a Rac-GAP (β2-chimaerin) and Rac1 RNAi to impair cyclin D1 up-regulation by HRG. Moreover, inhibition of EGFR or PI3K, which are both required for Rac1 activation by HRG, or inhibition of ERK1/2, whose activity is enhanced by Rac , abrogate cyclin D1 induction by HRG. Consistent with its role in cell-cycle progression, cyclin D1 up-regulation is required for HRG-induced breast cancer cell proliferation, as depletion of cyclin D1 using RNAi or inhibition of cyclin D1 induction through Rac1 inactivation/depletion impairs the proliferative activity of HRG. HRG also up-regulates p21Cip1, in agreement with previous studies [32–35]. The present study provides the first evidence that this effect is mediated by the small GTPase Rac. Indeed, Rac1 depletion as well as inhibition of Rac function by β2-chimaerin or N17Rac1 blocks p21Cip1 induction by HRG. As observed with cyclin D1, inhibition of EGFR or PI3K, which are both required for Rac activation by HRG, abrogates p21Cip1 induction. p21Cip1 up-regulation is also impaired by the MEK inhibitor UO126, thus supporting that, like cyclin D1, p21Cip1 induction by HRG is mediated by a Rac/ERK-dependent pathway. A recent study showed that heregulin activates Brk (breast tumour kinase), and Brk siRNA impaired Rac and p38 activation . Both Brk and p38 activities are involved in cyclin D1 induction by heregulin; however, it is not known whether p38 activation by heregulin is Rac-dependent or independent . Remarkably, the mitogenic response of HRG is lost when p21Cip1 is knocked-down from T-47D cells using RNAi. The role for p21Cip1 as a negative regulator of the cell cycle is well established [27,28]. However, increasing evidence suggests that p21Cip1 can also have positive roles in cell proliferation and tumorigenesis [37–40]. p21Cip1 was found to be essential for promoting the assembly of the cyclin D1–CDK4 complex in mammary epithelial cells , murine fibroblasts [41,42] and vascular smooth muscle cells . p21Cip1 binds to the active cyclin D1–CDK4 complex leading to increased stability of cyclin D1 [37,42]. We speculate that p21Cip1 up-regulation in response to HRG stimulation stabilizes the cyclin D1–CDK complex and promotes pRb phosphorylation, therefore favouring the progression through G1-/S-phase and the mitogenic response. However, although increased pRb phosphorylation was observed upon HRG stimulation (Figure 1A), depletion of p21Cip1 by RNAi did not markedly affect pRb phosphorylation (Supplementary Figure 2 at http://www.biochemj.org/bj/410/4100167add.htm), suggesting that additional mechanisms may also take place.
The NF-κB complex consists of different homodimers and heterodimers of the Rel/NF-κB family of transcription factors. In its inactive state, the NF-κB complex is retained in the cytoplasm in a latent form by the inhibitory protein IκB. Many stimuli rapidly activate NF-κB by triggering the phosphorylation and proteolytic degradation of IκB, freeing the dimer complex, and subsequently enabling the translocation of NF-κB into the nucleus where it transcriptionally regulates target genes [44,45]. Heregulins were known to activate NF-κB in breast cancer cells [25,26], and our results showing an increase in NF-κB luciferase reporter activity and time-dependent phosphorylation of IκBα in response to HRG support this concept. In addition, we determined that NF-κB plays a role in cyclin D1 and p21Cip1 induction by HRG via Rac1 in breast cancer cells, as these effects were impaired by expression of the NF-κB super-suppressor IκBαM. Cyclin D1 is a well-known NF-κB target gene. Activated Rac1 induces cyclin D1 expression in NIH 3T3 cells in an NF-κB-dependent manner, and inhibition of NF-κB reduces cyclin D1 levels induced by activated Rac1 . The reported activation of NF-κB by Rac in fibroblast cells [47,48] fits well with our results in breast cancer cells. We also found that activation of NF-κB by HRG is dependent on ERK, as revealed by the ability of the MEK inhibitor UO126 to block IκBα phosphorylation. Thus Rac serves as a point of hierarchical integration that transmits the HRG signal via an MEK/ERK/NF-κB pathway. Interestingly, in a recent study we found that Rac can also up-regulate cyclin D1 expression in an NF-κB-independent manner in mouse embryonic fibroblasts, arguing for a more complex regulatory mechanism . While p21Cip1 is a major target of p53 [27,28], p53-independent up-regulation of p21Cip1 can also occur in response to various stimuli [38,50]. Studies have shown that p21Cip1 expression is also subject to regulation by NF-κB. For example, Javelaud et al.  found that induction of p21Cip1 by TNFα (tumour necrosis factor α) in Ewing tumour cells requires NF-κB activity. Wuerzberger-Davis et al.  reported that G2/M arrest in leukaemic T-cells in response to ionizing radiation or the chemotherapeutic agent etoposide is mediated by NF-κB-dependent induction of p21Cip1.
Our studies may have significant clinical implications for breast cancer progression and treatment. HRG is up-regulated in approx. 30% of breast tumours. It would be interesting to determine whether those tumours also present cyclin D1 up-regulation. p21Cip1 is also frequently overexpressed in breast carcinomas and is strongly correlated with cyclin D1 overexpression  and ErbB2 overexpression [54,55]. Recent studies have shown that p21Cip1 depletion using antisense causes apoptosis in breast cancer cells and suppresses breast cancer growth and angiogenesis in animal models [39,56,57]. Given the fact that Rac and Rac effectors are overexpressed or hyperactivated in human breast tumours [15–17,22], our findings provide a strong rationale for designing breast cancer therapeutic strategies targeting the Rac signalling pathway.
This work was supported by grants RO1-CA74197 (National Institutes of Health) to M.G.K. and RO1-GM069064 (National Institutes of Health) to R.K.A.
breast tumour kinase
epidermal growth factor
MAPK (mitogen activated protein kinase)/ERK kinase
multiplicity of infection
nuclear factor κB
small interfering RNA
Present address: Department of Physiology and Center for Integrative Toxicology, Michigan State University, 4171 Biomedical Physical Sciences, East Lansing, MI 48864, U.S.A.