BRCA1-IRIS activates cyclin D1 expression in breast cancer cells by downregulating the JNK phosphatase DUSP3/VHR

Authors

  • Lu Hao,

    1. Department of Pharmacology and Experimental Therapeutics, Stritch School of Medicine and Cardinal Bernardin Cancer Center, Loyola University Chicago, Maywood, IL
    Search for more papers by this author
  • Wael M. ElShamy

    Corresponding author
    1. Department of Pharmacology and Experimental Therapeutics, Stritch School of Medicine and Cardinal Bernardin Cancer Center, Loyola University Chicago, Maywood, IL
    • Department of Pharmacology and Experimental Therapeutics, Stritch School of Medicine and Cardinal Bernardin Cancer Center, Loyola University Chicago, 2160 S. First Avenue Bldg. 112, Rm. 234, Maywood, IL 60153, USA
    Search for more papers by this author
    • Fax: +708-327-2245.


Abstract

Cyclin D1 plays an important role in cell cycle progression. In breast cancer, Cyclin D1 expression is deregulated by several mechanisms. We previously showed that in breast cancer cells, overexpression of BRCA1-IRIS induces Cyclin D1 overexpression and increases cell proliferation. BRCA1-IRIS alone or in complex with steroid receptor co-activators was targeted to the cyclin D1 promoter pre-bound by the c-Jun/AP1 and activated its transcription, which could explain the co-overexpression of BRCA1-IRIS and Cyclin D1 in breast cancer cells coupled with their increased proliferation. We report here an alternate or a complementary pathway by which BRCA1-IRIS activates Cyclin D1 expression. BRCA1-IRIS overexpression decreases the expression of the dual specificity phosphatase, DUSP3/VHR, an endogenous inhibitor of several MAPKs, including c-Jun N-terminal kinase. Although, the mechanism by which BRCA1-IRIS overexpression accomplishes that is not yet known, it is sufficient to induce Cyclin D1 overexpression in a human mammary epithelial cell model. Cyclin D1 overexpression could be blocked by co-overexpression of VHR in those cells. Furthermore, in 2 breast cancer cell lines that overexpress both BRCA1-IRIS and Cyclin D1 (MCF-7 and SKBR3) depletion of BRCA1-IRIS by RNA interference attenuated the expression of Cyclin D1 by elevating the expression level of VHR. These data demonstrate a critical role for BRCA1-IRIS in human breast cancer cell-cycle control and suggest that deregulated expression of BRCA1-IRIS is likely to reduce dependence on normal physiological growth stimuli, thereby providing a growth advantage to tumor cells and a potential mechanism of resistance to endocrine therapy. © 2007 Wiley-Liss, Inc.

Morphogenesis and development of mammary glands result from a complex interplay between genetic and epigenetic pathways that are controlled by hormones, growth factors, and other signaling molecules. Indeed, cyclin D1−/− mice show profound defects in mammary lobular-alveolar development, especially during pregnancy, because the breast epithelial cell compartment is unresponsive to ovarian steroids and fails to undergo the massive proliferative changes that these agents normally induce.1, 2

Cyclin D1 plays an important role in regulating mammary gland proliferation through its control of the G1 phase of the cell cycle, where it binds to cyclin-dependent kinase 4 (CDK4) or CDK6 to promote phosphorylation of the retinoblastoma protein. Overexpression of cyclin D1 has been reported in humans with breast cancer,3, 4 and it distinguishes invasive and in situ breast carcinomas from nonmalignant lesions.5 Interestingly, 50% of breast tumors overexpress Cyclin D1, even though Cyclin D1 gene amplification can be found in only 13–15% of tumors.6 These data indicate that epigenetic mechanisms, such as promoter deregulation or aberrant hormonal signaling, contribute significantly to this cancer-specific phenotype. Furthermore, Cyclin D1 overexpression in murine mammary epithelial cells induces breast cancer.6

Cyclin D1 is a growth regulatory gene activated by 17β-estradiol (E2) in MCF7 cells and in rodent breast epithelial tissue, ovary and uterine endometrium, and it is not reported to contain an estrogen receptor element (ERE)-related sequence in its proximal promoter.1, 7, 8 The addition of E2 to G0/G1-arrested MCF7 cells resulted in a 2 to 3-fold greater induction of Cyclin D1 mRNA within 1–3 hr and a 3 to 5-fold increase in protein synthesis and steady-state levels of Cyclin D1 within 3–6 hs compared with controls.8 Given that Cyclin D1 knockout mice are fertile, the physiologic relevance of Cyclin D1 as a direct target of estrogen-receptor (ER) actions is not clear.1 Additionally, prior studies demonstrated that E2 increased AP1 activity without increased synthesis of c-jun or c-fos and without the synthesis of any new proteins.9, 10, 11 Together, these findings suggest that the activation of AP1, a critical regulator of E2-induced cell growth, does not depend entirely on ER-mediated actions.

AP1 is implicated in diverse cellular processes, including differentiation, cell proliferation and transformation.12 AP1 consist of various combinations of Jun (c-Jun, JunB and JunD) and Fos (c-Fos, Fra-1, Fra-2 and FosB) proteins.9, 10, 11 Jun proteins can form homodimers or more stable heterodimers with proteins of the Fos family that do not homodimerize.9, 10, 11 Higher constitutive levels of c-jun and c-fos mRNA and AP1 levels have been reported for drug-resistant cells (such as etoposide-resistant human leukemia K562 cells) compared with drug-sensitive parental lines.13 Increased c-jun or c-fos expression and AP1 activity results from mitogenic stimulation by insulin or other insulin-like growth factors in the human breast adenocarcinoma cell line MCF-7.12 Furthermore, c-Jun transcriptional activity is enhanced by amino-terminal phosphorylation on Ser 63/73 (S63/73).14 This inducible phosphorylation is mediated by members of the Jun amino-terminal kinase (JNK, also known as stress-activated protein kinase, SAPK) subfamily14, 15 and is required to recruit the transcriptional co-activator CBP. The JNK/c-Jun pathway is a critical component of the proliferative response and induces G0–G1 cell cycle progression in many cell types.

In response to extracellular stimuli, MAP kinases (MAPK), such as ERK, JNK and p3816 are activated by their specific upstream dual-specificity kinases (MAP-kinase kinases).15 MAPK downregulatory mechanisms are less known; however, several members of the dual-specificity phosphatase (MKP) family have been shown to dephosphorylate and inactivate various MAPKs.16 For instance, the dual-specificity phosphatase 3 (DUSP3/VHR) targets and inactivates ERK and JNK.17 Most MKPs are immediate-early genes and are induced by various mitogens, growth factors and stress. Several of these phosphatases are constitutively expressed, and their phosphatase activity appears to be highly regulated through an unknown mechanism.18, 19

BRCA1-IRIS is a recently discovered, 1399-residue BRCA1 locus splice variant.20 Although BRCA1-IRIS and the full-length product of this tumor suppressor gene, BRCA1/p220 (hereafter, p220)21, 22 share 1,365 residues, BRCA1-IRIS-unlike p220-promotes cell proliferation during S-phase.20 Furthermore, wtBRCA1-IRIS is overexpressed in several human breast and ovarian cancer cell lines, as well as known BRCA1 mutant cell lines; for example, HCC1937 and SNU25123, 24 both have an intact BRCA1-IRIS but a mutant p220 cistron, and wtp220 cannot be detected in these cells.23, 24 Moreover, when overexpressed, BRCA1-IRIS induces epithelial-to-mesenchymal transition/migration/invasion abilities in normal breast cancer cells (ElShamy and Livingston, unpublished data).

In breast cancer cell lines, co-overexpression of BRCA1-IRIS and Cyclin D1 was observed, which made us wonder whether BRCA1-IRIS affects Cyclin D1 expression.20 We found earlier that BRCA1-IRIS, either alone or in complex with steroid receptor co-activators such as SRC1 or SRC3, transcriptionally induces Cyclin D1 expression through binding and activating c-jun/AP1 bound to the Cyclin D1 promoter.25 In the present study we explored whether BRCA1-IRIS can also activate Cyclin D1 expression in breast cancer cell line in a nontranscriptional fashion. We found that BRCA1-IRIS downregulates the expression of the dual specificity phosphatase DUSP3/VHR, an inhibitor of JNK in normal (e.g. human mammary epithelial (HME)) and cancerous (e.g. MCF-7 and SKBR3) breast cell lines, leading to activation of JNK and subsequently activation of c-Jun/AP1 and Cyclin D1 expression.

Taken together, the data suggest that BRCA1-IRIS can activate the expression of Cyclin D1 in a non-transcriptional fashion regardless of ERα status of the cell, raising the interesting possibility that BRCA1-IRIS overexpression might be an important step on the road to acquiring a more aggressive/endocrine resistance breast cancer phenotype.

Material and methods

Cell culture and drugs

All cells were cultivated in RPMI medium 1640 (Gibco) containing 10% fetal bovine serum (Gemini), at 37°C in a 10% CO2-containing atmosphere. The JNK inhibitor (L)-JNKI1 was from Calbiochem (Cat. No. 420116)

Immunoblotting

Cells at ∼75% confluence were washed several times with PBS, scraped and collected in PBS. Then they were sonicated 3 times for 10 sec each using a Fisher/Sonic Dimembrator Model 100. After sonication, whole cell lysates (including chromatin) were centrifuged twice at high speed for 10 min at 4°C, and the supernatants were retained. The following Abs were used: purified polyclonal rabbit (rAb) or mouse monoclonal antibody (mAb) α-BRCA1/IRIS,20 mAb α-Cyclin D1 (Oncogene, Ab-3), mAb α-Actin (Oncogene, Ab-1), rAb α-Sp1 (Santa-Cruz, sc-14027), mAb α-ERα (Novous Biologicals, ab7822), c-Jun and phospho-c-Jun (PhosphoPlus Antibody Kit, Cell Signaling, 9260), mAb α-VHR (BD Transduction Lab, 610546), α-ERK (New England Biolabs), mAb α-p-ERK (Santa-Cruz, sc-7383), rAb α-p38 (Santa-Cruz, sc535-G), rAb α-p-p38 (Sigma, P1491), mAb α-β-catenin (BD Transduction, 610153), a-LacZ (Abcam, ab 616), mAb a-V5 (Invitrogen, R960-25), mAb α-His (Novagen, 70796-3), mAb α-ErbB2 (Oncogene, Ab-3) and mAb α-EGFR (Abcam, ab-5368), mAb α-β-catenin (BD transduction Lab, Clone 14), mAb α-V5 (Invitrogen, R960-25), mAb α-ERα (Novus Biologicals, ab7822) and mAb α-His (Novagen, 70796-3).

Plasmids and transfection

Full-length BRCA1-IRIS or LacZ cDNAs were cloned into the pIND/V5-His-TOPO plasmid (Invitrogen, CA). BRCA1-IRIS cDNA was also cloned into the pRevTRE plasmid (retrovirus version, Clontech). ErbB2 full-length cDNA was cloned in the pBabe retroviral expression plasmid.26 Finally, VHR full-length cDNA generated with PCR was cloned into the pcDNA3.1 mammalian expression plasmid. All transfections were performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions.

Generation of cell lines using transfection and infection

U2OS VgRXR-expressing cell line (Invitrogen) was transfected with pIND-IRIS or pIND-LacZ. On Day 2, to select for transfected cells culture medium was supplemented with 500 μg of G418 and the selection continued for 2 weeks. Retroviral particles were generated by transfecting 293 T cells with a desired vector (e.g. pRevTRE-IRIS or pBabe-ErbB2) together with all necessary packaging plasmids. On Days 2 and 3 the supernatant was collected, polled and used as-is to infect HME cells (or HME cells stably expressing pTet-On plasmid, Clontech in the case of IRIS). Infected HME cells were of selected using Hygromycine for 2 weeks, and cell lines were obtained. Finally, to induce the expression of these ectopic genes, cells were incubated either with ponasterone (5 μM, Invitrogen in the case of pIND cell lines) or 2 μg/ml of doxycycline (Invitrogen, in the case of pRevTre clones).

Cell synchronization and FACS analysis

U2OS cell lines stably expressing LacZ or BRCA1-IRIS under the control of the ponasterone-inducible promoter was first grown for 24 hr in the presence of 250 μg/ml of nocodazole to arrest the cells in the G2/M stage. Sixteen hours later a sample from each cell line was collected, and the rest of the samples were washed 2 times with PBS and then incubated for 2 hr with medium containing 20% serum to release cells from G2/M into the G0/G1 phase, followed by incubation in serum-free medium for another 24 hr, after which sample from each cell line was collected. While still in serum-free medium, cells were induced with ponasterone (see above). At 24 and 48 hr after induction, samples of each cell line were collected. All the samples were then fixed in 70% ethanol, stained with propidium iodide (PI; 2.5 μg/ml; Sigma, St Louis, MO), supplemented with 19 μg/ml RNase A, and incubated at 37°C for 1–2 hr.

Immortalization assay

To study the ability of these genes to induce immortalization of HME cells, we first transfected RevTreIRIS- and ErbB2-expressing cells with the pcDNA4.1 plasmid. This vector carries a Zeocin selection marker. After transfection cells were trypsinized and plated at 25,000 cells/10-cm dish. Cells were grown in the presence of 250 μM Zeocin to ensure that all cells were cultured under similar conditions and that colonies indeed produced from single cell.

Small interfering RNA reagents and transfection

A dAdA-N19 double-stranded siRNA BRCA1-IRIS oligomer19 was used, which corresponded to a specific segment of the intron 11 sequence. A double-stranded, 19-nucleotide luciferase-specific siRNA was used as a negative control (Dharmacon). MCF-7 or SKBR3 cells were switched to serum-free medium and transfected with control or with BRCA1-IRIS siRNA reagent using oligofectamine as the reagent (Invitrogen). Twenty-four hours later, the cells were switched back into complete medium and incubated therein for 48 hr before collection.

Real time RT-PCR

Trizol (Gibco, Life Technologies)-isolated, DNaseI-treated total RNA was routinely used in real-time RT-PCR experiments (as specified in the Results section) using the QuantiTect SYBR Green RT-PCR kit. Routinely, 0.1 μg of total RNA or 0.2 μg of poly A+ RNA was used as template in each reaction for amplification of ∼100–200 bp of RNA species. The following primer sets were used to amplify BRCA1-IRIS (forward: GGTCTGAGTGACAAGGAATTGGTTTCAGATGATGAAGAA and reverse: TTA ACTATACTTGGAAATTTGTAAAATGTG), ERa (nt 609-781 of NM_021689; forward: CCCTACTACCTGGAGAACGAG CCCA, reverse: GCACAGTAGCGAGTCTCCTTGGCAGA), and cyclin D1: (nt 274-440 of NM_171992; forward: GGGAGATTGT GCCATCCATGCGGAAAATCGT, reverse: GGTGGCCCCCAGAAGCTGCAGGCGGCTCTTCTT). GAPDH was used according to published data.

Results

Previously we showed that increasing the level of BRCA1-IRIS in (HME) or U2OS cell lines leads to an increase in the rate of their DNA replication.20 More recently, we showed that BRCA1-IRIS regulates Cyclin D1 expression in the breast cancer cell line MCF7.25 BRCA1-IRIS is recruited to the Cyclin D1 promoter through binding to c-Jun pre-bound to that promoter and induces its transcriptional activity,25 which could account for the ability of BRCA1-IRIS to induce cell proliferation when overexpressed.20, 25 Conversely, depletion of BRCA1-IRIS from MCF7 decreased Cyclin D1 expression and cell proliferation.25 In the present study we explored whether BRCA1-IRIS affects Cyclin D1 expression and cell proliferation in a non-transcriptional manner.

BRCA1-IRIS overexpression induces cell cycle progression in the absence of serum

To determine whether forced BRCA1-IRIS overexpression induces cell cycle progression of cells arrested in G0/G1 in the absence of serum or other stimuli, we generated inducible U2OS cell lines expressing BRCA1-IRIS or LacZ (about the same size protein, as a control) upon addition of ponasterone.

Both cell lines were synchronized in G0/G1 according to the protocol described in Material and Methods and induced with 5 μM of ponasterone to express BRCA1-IRIS or LacZ for 24 or 48 hr in the absence of serum or other stimuli. Samples were collected at each time point (see Material and Methods). All of the samples were then stained with PI and analyzed using FACScan.

Both cell lines were uniformly synchronized until they reached the G0/G1 border before induction (see I and II in Fig. 1, upper). However, after induction of the proteins in the presence of 5 μM ponasterone (for expression at 24 hr, see Fig. 1 lower panels) cells expressing LacZ remained locked in G0/G1 (Fig. 1, III and IV), whereas cells overexpressing BRCA1-IRIS initiated their cell cycle for at least 2 rounds (Fig. 1, III and IV), suggesting that overexpression of BRCA1-IRIS and not the control protein induces cell cycle progression. Alternatively, BRCA1-IRIS overexpression might induce the expression of cell cycle-inducing agents such as Cyclin D1.

Figure 1.

BRCA1-IRIS overexpression induces cell cycle progression. Stable cell lines overexpressing BRCA1-IRIS or LacZ (lower panel) after ponasterone addition. Cells were first made G2/M by incubating them for 16 hr in a medium containing 250 μg/ml of nocodazole. The next day cells were washed 2–3 times, and 20% serum was added for 2 hr. Cells were next washed again 2–3 times with PBS and medium with no serum added for 22 hr. At this point most cells accumulated in the G0/G1 phase of the cell cycle. Ponasterone (5 μM) was then added to the serum-free culture medium. Cell cycle profiles were measured at the intervals shown using FACScan. Lower panel show western analysis on these cell lines at 24 hrs after ponasterone addition. Actin western shown is from IRIS cell line; however, similar results were obtained for LacZ cell line.

Co-overexpression of BRCA1-IRIS and cyclin D1 mRNAs in ERα-positive and negative cells

Breast cancer cell lines arrested in G0/G1 when stimulated by addition of E2 showed transcriptional activation of c-Myc and Cyclin D1.27, 28, 29 However, both Cyclin D1 and c-Myc are overexpressed in many ERα-positive as well as ERα-negative breast cancer cell lines (Fig. 2a, and data not shown). How then is Cyclin D1 (and c-Myc) expression induced in ERα-negative breast cancer cell lines? One possibility is that in ERα-positive cells Cyclin D1 expression is induced by other pathways as well, some are probably overexpressed and amplified in ERα-negative cells to compensate for the loss of ERα-dependent activation.

Figure 2.

Loss of BRCA1-IRIS blocks Cyclin D1 expression in mammary epithelial cells. (a) Comparison of RNA levels of cyclin D1 (open boxes), ERα (gray boxes), and BRCA1-IRIS (closed boxes) in 6 breast cancer cell lines (MCF7, T47D, HS587T, MDAMB231, MDAMB453 and SKBR3) and a normal mammary epithelial cell line (HME) using real time RT/PCR. All data were first normalized against GAPDH and then against the level of each of these RNAs found in HME cells, therefore the level of all RNAs in HME cells is taken as 1. (b) MCF-7 or SKBR3 cells were transfected with siRNA against Luc (as control) or BRCA1-IRIS (against the intron 11 segment). Seventy-two hours later cells were collected, washed, sonicated, centrifuged twice and whole cell extracts (including chromatin) were tested for the expression of BRCA1-IRIS, cyclin D1 (b), β-catenin, and Actin as loading control (c), ErbB2 and (d)EGFR.

Interestingly, BRCA1-IRIS RNA is overexpressed in ERα-positive as well as ERα-negative breast cancer cell lines (Fig. 2a) compared with HME cell line. More importantly, BRCA1-IRIS RNA overexpression level differed between cancer cell lines with higher overexpression level observed in ERα-negative cell lines (Fig. 2a). Interestingly, the level of overexpression of Cyclin D1 RNA did not change between ERα-positive (e.g. MCF7) and -negative (e.g. SKBR3) cell lines when compared with HME (Fig. 2a). Thus, MCF7 and SKBR3 were chosen for the subsequent analysis.

BRCA1-IRIS loss blocks cyclin D1 expression in mammary epithelial cells

Depletion of BRCA1-IRIS (using RNA interference) from MCF7 or SKBR3 dramatically reduced Cyclin D1 expression with no effect on the expression of Actin or β-catenin (Figs. 2b and 2c). To determine whether this effect was caused by reduced ERα or alternatively receptor tyrosine kinases, such as EGFR or ErbB2, that are also overexpressed and amplified in breast cancer cells, MCF7 or SKBR3 cells depleted of BRCA1-IRIS were tested for the expression of these 3 proteins using western blot analysis. As can be seen from Figure 2d, EGFR and ErbB2 expression was dramatically reduced in cells depleted of BRCA1-IRIS compared with cells treated with control siRNA. However, ERα expression was not affected (at least in the ERα-positive cell line, MCF7) (Fig. 3b), suggesting that the effect of BRCA1-IRIS on Cyclin D1 depends on the expression and activation of growth factor or cell surface receptors, which are known to affect Cyclin D1. Taken together it is clear that BRCA1-IRIS, in addition to its nuclear pathway, which transcriptionally affects Cyclin D1 expression,25 is also able to regulate the expression of this protein through an extracellular or cytoplasmic pathway.

Figure 3.

BRCA1-IRIS activates Cyclin D1 gene expression in a c-Jun-dependent manner. MCF-7 or SKBR3 cells were transfected with siRNA against Luc (as control) or BRCA1-IRIS. Seventy-two hours later cells were collected, washed, sonicated, centrifuged twice and whole cell extracts (including chromatin) were tested for the expression of IRIS, Sp1, or Actin as loading control (a) or ERα, c-Jun, p-c-Jun (S63/73) or (b) Actin as loading control. Also in (b) is shown data from cells treated with JNK inhibitors for 2 hr.

cJun-dependent activation of cyclin D1 by BRCA1-IRIS

Activated JNK phosphorylates c-Jun on Ser-63/73 (S63/73) and activates it to interact with fos family members to form an active AP1 transcriptional complex,30, 31 which in turn is recruited to the Cyclin D1 promoter and activates its transcription (see Introduction). To study whether depletion of BRCA1-IRIS blocked Cyclin D1 expression through downregulation of JNK and c-Jun expression or activity, we used siRNA to deplete BRCA1-IRIS (control is siLuc) from MCF7 and SKBR3 cells, and the levels of 3 different transcription factors (Sp1, ERα, and c-Jun) were then measured using western blotting.

Depletion of BRCA1-IRIS did not affect the expression of these proteins in MCF7 cells (ERα-positive) and SKBR3 cells (ERα-negative) (Fig. 3), suggesting that mechanism(s) other than gene expression might be the principle pathway by which BRCA1-IRIS depletion affects Cyclin D1 (see above). In contrast, activation of c-Jun (i.e. phosphorylation on S63/73), which was detected by using immunoblotting with a phospho-specific antibody, was dramatically decreased in MCF7 or SKBR3 cells depleted of BRCA1-IRIS (Fig. 3b), suggesting that in the absence of BRCA1-IRIS c-Jun/AP1 is inactive. This finding raises the interesting possibility that the effect is at the upstream kinase (JNK) level.

To study this possibility, we also incubated MCF7 or SKBR3 cells with JNK inhibitor for 2 hr. In both cell lines treatment with this inhibitor, as with siIRIS, abolished c-Jun activation (Fig. 3b) with no effect on the protein expression level (Fig. 3b). Taken together, we suspected that BRCA1-IRIS is targeting JNK in cells.

BRCA1-IRIS depletion inactivated JNK through induction of its upstream phosphatase DUSP3/VHR

Again using specific siRNA we depleted BRCA1-IRIS (or Luc) from MCF7 or SKBR3 cells and studied the expression or activation (using phospho-specific antibodies) of all 3 MAPKs (ERK, JNK and p38). In MCF7 or SKBR3 cells, depletion of BRCA1-IRIS did not change the levels of the 3 MAPK proteins (Fig. 4a) compared with cells treated with Luc siRNA (Fig. 4a). However, activation (i.e. phosphorylation) of 2 (ERK and JNK) of these 3 MAPKs was dramatically reduced in BRCA1-IRIS depleted cells (Fig. 4a). In contrast, p38 phosphorylation, which was minimal in control siRNA (Luc)-treated cells, increased significantly upon depletion of BRCA1-IRIS (Fig. 4a). The significance of this result is discussed below. Taken together these data suggest that in the absence of BRCA1-IRIS, an upstream signaling pathway possibly connected to EGFR and ErbB2 (see above), is reducing JNK and possibly ERK activation and thus affecting Cyclin D1 transcriptional activation. A mechanism that might inactivate JNK after depletion of BRCA1-IRIS is the activation of an upstream inactivating phosphatase.

Figure 4.

BRCA1-IRIS depletion induces the expression of JNK phosphatase, DUSP3/VHR. MCF-7 or SKBR3 cells were transfected with siRNA against Luc (as control) or BRCA1-IRIS. Seventy-two hours later cells were collected, washed, sonicated, centrifuged twice and whole cell extracts (including chromatin) were tested for the expression of ERK, p-ERK, JNK, p-JNK, p38, pp38 or Actin as loading control (a) or DUSP3/VHR or (b) Actin as loading control.

JNK and ERK (but not p38) are targeted for inactivation by VHR.32 Thus, we found that depletion of BRCA1-IRIS increased the levels of VHR significantly in both cell lines (MCF7 and SKBR3) compared with cells treated in parallel with siLuc (Fig. 4b). It is important to note that independently VHR RNA was found to increase by ∼150-fold in HME cells depleted of BRCA1-IRIS compared with the same cell line treated with siLuc as measured on an expression microarray (Affymetrix array) (data not shown) where it topped the list of BRCA1-IRIS depletion-induced targets. Furthermore, the fact that VHR targets JNK and ERK and not p3832 aids our data (Fig. 4a).

BRCA1-IRIS overexpression induces ErbB2, EGFR and cyclin D1 overexpression and downregulation of VHR in HME cells

A possible scenario is that overexpression of BRCA1-IRIS induces EGFR and ErbB2 expression and activation, and hence these kinases activate JNK (and ERK) possibly through inactivation of VHR. Thus JNK will activate c-Jun/AP1 and hence Cyclin D1 transcription. Indeed, compared with control HME cells BRCA1-IRIS overexpression (induced by 2 μg/ml doxycycline for 3 days) in HME cells increased the expression of tyrosine kinase receptors EGFR and ErbB2 as well as Cyclin D1 (Fig. 5a), whereas the same treatment dramatically reduced VHR expression (Fig. 5). Controls such as β-catenin or Actin (Fig. 5a) were not affected (Fig. 5a).

Figure 5.

VHR overexpression suppresses Cyclin D1 overexpression even in the presence of high levels of BRCA1-IRIS. (a) Parental human mammary epithelial cell lines (HME) or HME cells expressing inducible BRCA1-IRIS were incubated with 2 μg/ml of doxycycline. Seventy-two hours later cells were collected, washed, sonicated, centrifuged twice and whole cell extracts (including chromatin) were tested for the expression of IRIS, EGFR, ErbB2, cyclin D1, VHR, β-catenin, or Actin as loading control. (b) BRCA1-IRIS over-expressing cells were maintained in 2 μg/ml doxycycline and transfected with an empty vector or a mammalian plasmid expressing full-length VHR. Seventy-two hours later cells were collected, washed, sonicated, centrifuged twice and whole cell extracts were tested for the expression of IRIS, EGFR, ErbB2, Cyclin D1, VHR, β-catenin or Actin as loading control.

VHR overexpression overcomes BRCA1-IRIS activation of cyclin D1 expression

To determine whether this relationship between BRCA1-IRIS and VHR is indeed exclusive, we overexpressed VHR in BRCA1-IRIS-overexpressing HME cells (same clone as in Fig. 5a), and the levels of the proteins aforementioned were again tested using western blotting. Although, VHR overexpression did not affect the expression of BRCA1-IRIS, EGFR, ErbB2, β-catenin, or Actin, there was a dramatic reduction in Cyclin D1 expression (Fig. 5b), suggesting that a combination of overactivation of JNK through overexpression of ErbB2 and EGFR by BRCA1-IRIS as well as reduction in VHR expression also through BRCA1-IRIS probably accounted for the effects we measured on Cyclin D1 expression in cells with altered BRCA1-IRIS levels when compared with control cells.

BRCA1-IRIS overexpression transforms HME cells

Cyclin D1 by itself can induce G0/G1-arrested cells to enter the cell cycle as well as shorten their G1 phase,30, 33 suggesting that factors other than the E2/ERα-signaling pathway account for Cyclin D1 overexpression in breast cancer cells and to their transformation. The fact that BRCA1-IRIS overexpression promotes cell cycle progression (Fig. 1) suggests that it could be a transforming agent. To study that possibility, we generated HME clones carrying ErbB2 cDNA and compared them together with HME cells inducibly overexpressing BRCA1-IRIS for colony formation. Three BRCA1-IRIS clones were chosen for further analysis in part because they expressed increasing levels of BRCA1-IRIS (Fig. 6, lower panels). In preliminary experiments, we noticed that BRCA1-IRIS overexpression from a constitutive promoter such as that used with ErbB2 in HME cells induced increased cell death in the initial stages of selection. For that reason we selected to compare inducible (BRCA1-IRIS) and constitutively (ErbB2) expressing cells. However, our data should be representative, because the expression of Actin and several control proteins such as β-catenin did not change in the different clones compared with control cells (data not shown). We selected cells expressing Zeocin as opposed to using clones alone because we wanted another way of assuring the growth of a single cell into a colony, which is an important parameter in studying immortalization/transformation.

Figure 6.

Transforming ability of BRCA1-IRIS. Several stable cell lines expressing ErbB2 or BRCA1-IRIS (inducible in the presence of 2 μg/ml doxycycline) at different levels (see lower panel) were all transfected with pcDNA4.1 (containing the Zeocine selection marker) plasmid. Twenty-four hours later 250 μg/ml Zeocine and 2 μg/ml doxycycline were added to the culture medium and maintained there for 2 weeks. Images show the colonies formed 2 weeks later.

As expected, ErbB2 overexpression induced HME cell transformation (as measured using this assay after 2 weeks) (compare parental with ErbB2 in Fig. 6). Interestingly, clone #3 (that expressed the highest level of BRCA1-IRIS) (see Fig. 6 lower panel) induced transformation of HME cells to a similar if not higher level than that seen with the ErbB2 cell line (compare #3 to parental and ErbB2-expressing clone in Fig. 6) also after 2 weeks, suggesting that BRCA1-IRIS overexpression does indeed transform HME cells. Furthermore, the number of colonies formed using clone #2 was lower than those seen with clone #3 (Fig. 6) (overexpresses lower levels of BRCA1-IRIS than #3, see Fig. 6, lower panel), yet greater than those seen with clone #1 (expresses the lowest level of BRCA1-IRIS) (Fig. 6). To ensure that these effects indeed depend on BRCA1-IRIS expression, clone #2 was grown in the presence or absence of doxycycline. As anticipated, only in the presence of doxycycline, clone #2 was able to form colonies (Fig. 6). Collectively these data suggest that BRCA1-IRIS, like ErbB2, behaves more like a transforming agent inducing cell cycle entry and cellular proliferation or survival when overexpressed in human mammary epithelial cells.

Discussion

The data presented here show that when BRCA1-IRIS is overexpressed in a mammary epithelial cell, it induces its transformation, in part, by inducing Cyclin D1 overexpression.32, 34, 35 Many patients with advanced breast cancer initially respond to hormonal therapy (e.g. tamoxifen) but in cases of recurrence, hormone-unresponsive tumors develop,36 and they tend to be more aggressive, invasive and resistant to chemotherapy. Thus, it is increasingly important to determine the mechanisms by which breast cancer cells display ERα-dependent and ERα-independent Cyclin D1 transcription. Our data suggest a potential role for BRCA1-IRIS in this independent pathway.

We propose that whereas most E2-dependent cells undergo cell death after E2 withdrawal in vivo, a subset of cells with adequate levels of BRCA1-IRIS/ErbB2 (and/or EGFR) and thus Cyclin D1 survive and expand. This subset could be the seeds for E2-independent, hormonally unresponsive breast cancer. It is possible that these 2 events might not occur simultaneously in tumor cells and that the immediate induction of BRCA1-IRIS after E2 withdrawal may be sufficient over the short term to maintain the tumor, while the proliferative rate exceeds the rate of apoptosis. This could provide adequate time for activation of other signaling pathways, such as ErbB2 overexpression. Interestingly, upon E2 withdrawal ErbB2 expression is increased in MCF-7 cells (data not shown). Furthermore, using chromatin immunoprecipitation we recently found that BRCA1-IRIS binds and activates the EGFR and ErbB2 promoters (ElShamy and Livingston, unpublished data). Thus it is possible that loss of E2 dependency and BRCA1-IRIS overexpression together generate this E2-independent phenotype. This observation suggests that continued ERα expression in such cells might not be necessary once tumors progress to a more aggressive phenotype, as suggested by others.36 Thus, derangements of the regulatory pathway described here could be a possible pathogenic mechanism for breast carcinogenesis and for tumor progression to a hormone-independent phenotype that is resistant to endocrine therapy.

Two of the promoters to which BRCA1-IRIS binds, c-Myc25 and ErbB2 (ElShamy and Livingston, unpublished data), contain origins of replication within them. Since, we have already shown that BRCA1-IRIS is involved in pre-replication complex formation on origin of replication,20 as well as transcription of at least some cancer-inducing genes, one wonders whether the 2 events are related. One possibility is that replication and transcription, to some extent, go hand in hand, especially for important growth and survival genes such as c-Myc, Cyclin D1 and ErbB2.

In summary, we have shown that when JNK is activated through downregulation of VHR expression by BRCA1-IRIS/EGFR or ErbB2, Cyclin D1 is expressed and HME cells are transformed. However, because forced VHR overexpression in cells overexpressing BRCA1-IRIS blocked Cyclin D1 expression with no change in the expression of BRCA1-IRIS, EGFR and ErbB2 suggests that in breast cancer cells that overexpress BRCA1-IRIS, EGFR or ErbB2 restoration of VHR expression might be sufficient to overcome Cyclin D1 overexpression and over-proliferation or transformation.

BRCA1-IRIS downregulation reduced JNK and ERK activation levels in a VHR-dependent manner. Activation of p38, the third MAPK, was instead increased. The explanation for this unexpected result might come from the fact that p38; (i) not a target of VHR, (ii) more importantly, activation of p38 after cellular stress might be a mechanism to arrest the growth of cells as well (through induction of Cdc25A degradation37), which probably decreases Cyclin D1 expression as well.34

Taken together, our studies identify BRCA1-IRIS as a component of the diverse signaling pathways that integrate multiple stimuli into an appropriate cellular response. Through MAPK and other downstream effectors, BRCA1-IRIS plays critical roles in many biological processes, especially in regulating cell growth. However, much still remains to be explored to determine the mechanisms involved. Thus, in addition to representing a potential therapeutic target in ERα-negative breast cancer, inhibiting BRCA1-IRIS might also effectively obviate progression.

Our findings collectively suggest that wild type BRCA1-IRIS might induce mammary epithelial cells or tumor proliferation by inducing transcription of cell transformation or proliferation-inducing genes. Hence, inhibiting or inactivating it's (i.e. BRCA1-IRIS) ability to perform such functions may contribute to reversing the carcinogenic phenotype of those mammary cells.

Acknowledgements

We thank Dr. Tomas Mostelin for supplying VHR antibody and for invaluable discussion. We also thank Ms. J. Lane and Ms. E. Nakuci for superb technical assistance. W. ElShamy was supported in part by a European Molecular Biology Organization (EMBO) Fellowship and in part by a “Massachusetts Department of Public Health Breast Cancer Research Grant” #30481126109 and an NCI grant “Specialized Program of Research Excellence (SPORE)” in breast cancer research at the Dana-Farber/Harvard Cancer Center.

Ancillary