Function of Slit/Robo signaling in breast cancer

Abstract

Slit and Robo are considered tumor suppressors because they are frequently inactivated in various tumor tissue. These genes are closely correlated with CpG hypermethylation in their promoters. The Slit/Robo signaling pathway is reportedly involved in breast cancer development and metastasis. Overexpression of Slit/Robo induces its tumor suppressive effects possibly by inactivating the β-catenin/LEF/TCF and PI3K/Akt signaling pathways or by altering β-catenin/E-cadherin-mediated cell-cell adhesion in breast cancer cells. Furthermore, loss of Slit proteins or their Robo receptors upregulates the CXCL12/CXCR4 signaling axis in human breast carcinoma. In addition, this pathway regulates the distant migration of breast cancer cells not only by mediating the phosphorylation of the downstream molecules of CXCL12/CXCR4 and srGAPs, such as PI3K/Src, RAFTK/ Pyk2, and CDC42, but also by regulating the activities of MAP kinases. This review includes recent studies on the functions of Slit/Robo signaling in breast cancer and its molecular mechanisms.

Keywords

Slit/Robo / hypermethylation / β-catenin / CXCL12/CXCR4 / migration

Introduction

Breast cancer is a common malignant tumor that affects women worldwide and has become the top cause of cancer-related death among women. The incidence of breast cancer is increasing, and the age of patients suffering from this disease is becoming younger. Approximately 180 000 women are diagnosed with breast cancer and 40 000 die from the disease each year in America [ 1]. In addition, 6%−10% of newly diagnosed breast cancer patients have locally advanced or metastatic disease. Metastatic breast cancer patients have a median survival period of 2−3 years [ 2], with only few (2%) surviving up to 20 years after the diagnosis of metastasis [ 3].

Metastasis, the formation of secondary tumors in organs distant from the site of the primary cancer, is the main cause of treatment failure and death in patients with breast cancer. Distant metastasis is difficult to detect using conventional methods. Breast cancer metastasis is induced not only by the activation of various cancer-promoting genes but also by the inactivation of tumor suppressor genes. Slit proteins and their Robo receptors are abnormally expressed in various tumor tissue, including breast cancer. Recent studies have reported that the Slit/Robo signaling pathway functions in breast cancer metastasis. The current review focuses on the regulation and molecular mechanisms of Slit/Robo signaling in breast cancer.

Structures of Slit and Robo

Slit is a large extracellular matrix-secreted glycoprotein that was first identified in Drosophila [ 4]. Slit proteins regulate the repulsive cues on axons and growth cones during central nervous system development and suppress tumor growth in various human cancers [ 5, 6]. Slit proteins contain five highly conserved structures, namely, an N terminus signal peptide, four tandem arrays of leucine-rich repeats (LRR), seven or nine epidermal growth factor (EGF)-like repeats, a laminin G domain, and a cytosine-rich C-terminal region [ 7]. The Slit family has three members, namely, Slit1 (1q23.3-q24), Slit2 (4p15.2), and Slit3 (5q35-34). Recent studies have detected Slit1, Slit2, and Slit3 in the central nervous system and Slit2 and Slit3 in other tissue and organs [ 8, 9].

The Robo receptor family consists of the following four members: Robo1/Dutt1 (3p12.3), Robo2 (3p12.3), Robo3/Rig-1 (11q22.4), and Robo4/Magic Roundabout (11q24.2). These transmembrane proteins are candidate receptors for the repulsive guidance ligand members of the Slit family. Robo1, Robo2, and Robo3 share the same architecture, which comprises five Ig-like (immunoglobulin-like) domains, three fibronectin (Fn) type III repeats in the extracellular domain, and four conserved linear motifs in the intracellular domain. Robo4 consists of only two Ig and two Fn domains in the extracellular domain [ 10]. Structure-function studies have shown that Robo1, Robo2, and Robo3 can bind to a single active site located at the second LRR domain of Slit [ 11].

Expression of Slit and Robo in breast cancer

Recent studies have considered Slit and Robo as candidate tumor suppressor genes because of their frequent inactivation during tumorigenesis [ 12]. Slit2 or Slit3 expression is silenced in approximately 50% of sampled human breast tumors. Moreover, primary breast cancer and several breast cancer cell lines exhibit reduced or absent Slit2 expression. The expression patterns of Robo receptors often differ in various breast cancer cell lines. Robo1 is more highly expressed than Robo2 in DU4475 cells, whereas the opposite phenomenon can be observed in MDA-MB-231 cells [ 13].

The Slit and Robo genes are mainly inactivated through loss of heterozygosity (LOH) and promoter region hypermethylation. Analysis of the Slit and Robo promoter region showed the presence of extensive hypermethylation of their cytosine-guanine (CpG) island in invasive cervical cancer, breast cancer, and lung carcinoma [ 14, 15]. Silenced or reduced expression of Slit2 correlates with CpG hypermethylation, and further treatment with the demethylating agent 5-azacytidine could restore Slit2 expression. Slit2 promoter hypermethylation in tissue and serum samples from breast cancer patients was proposed to be a possible marker for the early detection of the disease [ 16]. The Slit2 gene has been mapped to 4p15.2 [ 17], and the chromosome 4p15.1-15.3 region shows frequent allele loss in approximately 63% of breast tumors [ 18]. The putative receptor Robo1 is methylated in<20% of breast tumors; however, the LOH of Robo1 frequently occurs in lung, breast, and kidney tumors [ 19]. Therefore, Slit and Robo possibly function as novel tumor suppressor genes.

Slit/Robo signaling inhibits the growth and invasion of breast cancer via β-catenin

Breast cancer cells that overexpress Slit2 exhibit decreased Akt activity and hence decreased glycogen synthase kinase 3 β (GSK-3β) phosphorylation. GSK-3β dephosphorylation enhances β-catenin phosphorylation, which is recognized by the ubiquitin ligase complex that mediates β-catenin degradation. Therefore, Slit2-overexpressing breast cancer cells possess a decreased total amount of β-catenin but an increased amount of β-catenin on the membrane. β-catenin on the cell membrane promotes intercellular adhesion [ 20], a catenin- and cadherin-mediated process that participates in the development of aggressive breast cancer [ 21]. Thus, Slit2 inhibits the invasion of breast cancer cells possibly by regulating the expression and distribution of β-catenin, which interacts with the E-cadherin/actin cytoskeletal system during cell-cell adhesion.

Aside from promoting intercellular adhesion, Slit2 also inhibits the growth of breast cancer cells via β-catenin/LEF/TCF. Loss of Slit2 or Robo1 in the mammary epithelium leads to the formation of hyperplastic and disorganized lesions [ 22]. Slit2 overexpression significantly decreases the proliferation of breast cancer MCF-7 cells compared with the vector control and inhibits the growth of breast cancer in animal models in vivo [ 13]. Thus, low Slit2 expression possibly triggers the proliferation of breast cancer cells. Slit2-overexpressing cells exhibit decreased β-catenin nuclear translocation and inhibited interaction with the TCF/LEF family of DNA binding proteins, which regulate the expression of the Cyclin D1, MMPs, and c-myc genes. These downstream genes of β-catenin/LEF/TCF are critical mediators of proliferation [ 23, 24] (Fig. 1). Moreover, a recent study has indicated that the activation of Robo1 signaling in breast cancer cells by Slit2 from stromal fibroblasts could restrain tumorigenesis by blocking the phosphoinositide 3-kinase (PI3K)/Akt/β-catenin pathway [ 25]. These data collectively suggest that Slit/Robo signaling suppresses tumor growth by regulating the β-catenin/LEF/TCF and PI3K/Akt signaling pathways or by altering β-catenin/E-cadherin-mediated cell-cell adhesion in breast cancer.

Slit/Robo signaling pathway inhibits breast cancer metastasis by regulating the CXCL12/CXCR4 axis

The CXCR4/CXCL12 axis is involved in breast cancer cell chemotaxis, invasion, adhesion, and metastasis to target organs, such as lymph nodes, lung, and bone. CXCR4 is highly expressed in breast cancer but lowly expressed or absent in normal breast tissue [ 26]. CXCL12, which is the only ligand of CXCR4, exhibits peak levels of expression in primary organ sites (bone marrow, lymph nodes, and lungs) of breast cancer metastasis [ 26]. Therefore, a simple hypothetical CXCR4/CXCL12 axis-dependent model of breast cancer metastasis can be proposed; that is, CXCR4-expressing breast cancer cells, which are easily attracted by CXCL12-expressing target organs, shed from the primary tumor mass and circulate the body. The combination of CXCR4 and CXCL12 induces the movement of cancer cells to target organs and promotes cell proliferation, angiogenesis, and secondary tumor formation.

Recent studies have reported an association between Slit/Robo signaling and CXCL12/CXCR4 in breast cancer. Loss of Slit expression upregulates both CXCL12 and CXCR4; by contrast, Slit overexpression in human breast carcinoma suppresses CXCR4 expression. Ubiquitin-specific protease 33 (USP33)/von Hippel-Lindau tumor suppressor protein-interacting deubiquitinating enzyme 1 (VDU1) is a protein that interacts with the Robo1 intracellular domain, suggesting that Slit/Robo signaling negatively regulates CXCL12/CXCR4 via USP33/VDU1 [ 27].

Slit inhibits the migration and adhesion of breast cancer cells via the phosphorylation of downstream CXCL12/CXCR4 molecules, such as PI3K, FAK, RAFTK/Pyk2, and p44/42 MAP, which are involved in tumor development. Focal adhesion kinase (FAK) and the related adhesion focal tyrosine kinase/praline-rich tyrosine kinase 2 (RAFTK/Pyk2) regulate the morphology, motility, adherence, and migration of breast cancer cells [ 28, 29]. Cells with Slit stimulation inhibit the CXCL12-induced activation of PI3K and Src, which regulate the tyrosine phosphorylation of FAK and RAFTK/Pyk. Aside from inhibiting the phosphorylation of RAFTK and FAK, Slit selectively blocks p44/42 MAPK, which can be involved in cell motility and adhesion by regulating the downstream gene expression or the activation of myosin light chain kinase. Matrix metalloproteinases (MMPs) degrade the extracellular matrix and facilitate tumor invasion [ 30]. The present study further supported that Slit inhibits the expression and activities of MMP2 and MMP9 by regulating the CXCL12/CXCR4 axis in breast cancer cells [ 13] (Fig. 2).

The Slit/Robo complex possibly inhibits Cdc42 activity through CXCL12/CXCR4 and Slit/Robo guanine triphosphatase (GTPase)-activating proteins (srGAPs), thereby regulating cell motility and cycling. A Cdc42 activation cycle is organized by PI3K, which is induced by the CXCL12/CXCR4 axis [ 31]. Moreover, the intracellular domain of Robo strongly interacts with a novel family of srGAPs that can regulate the activities of the Rho subfamily of small GTPases, particularly Rho, Cdc42, and Rac [ 32, 33]. Cdc42 can stimulate the activity of p21-activated kinase, which also regulates actin dynamics and cell cycle [ 34]. These results collectively suggest that the Slit/Robo complex inhibits the activity of Cdc42 and influences the mobility and proliferation of breast cancer cells.

Slit/Robo signaling pathway and breast cancer brain metastases

The migration and metastasis of breast cancer cell share many similarities to leukocyte trafficking, which is critically regulated by chemokines and their receptors. Slit2, which is expressed in vascular endothelial cells in the brain, is a ligand of glypican-1 and facilitates cancer cell adhesion in systemic circulation [ 35, 36]. In vitro experiments showed that MDA-MB-231BR (MDA-MB-231 brain-seeking clone) cells enhance MMP9 expression and tumor cell migration in response to Slit2 stimulation for 30 min. In addition, the mRNA expression level of vascular EGF (VEGF) is higher in MDA-MB-231BR than in MDA-231BO (MDA-MB-231 bone-seeking clone) and MDA-MB-231P (MDA-231 parental line) [ 37]. VEGF promotes the transendothelial migration of breast cancer cells by regulating the permeability of brain microvascular endothelial cells [ 38]. However, the mRNA expression levels of Slit2, Robo1, and MMP9 are lower in MDA-MB-231BR than in MDA-MB-231P. Therefore, current studies suggest that the Slit/Robo pathway potentially functions in the target brain metastasis of breast cancer. However, the exact functions and underlying mechanisms of the Slit/Robo pathway in breast cancer metastasis warrant further investigation.

Conclusions

Metastasis, the major malignant character of tumor, is the main cause of treatment failure, poor prognosis, and death in breast cancer patients. During the growth and metastasis of cancer cells, Slit/Robo collaborates with the β-catenin/TCF/LEF, PI3K/AKT, KAK/RFAT, and CXCL12/CXCR4 signaling pathways. β-catenin/TCF/LEF serves critical functions in breast cancer metastasis. Initial findings from previous studies warrant further elucidation of the mechanisms by which Slit/Robo signaling regulates β-catenin-E-cadherin interaction and β-catenin distribution to the plasma membrane.

CXCL12/CXCR4 expression is closely associated with local recrudescence and tumor metastasis. Slit/Robo signaling inhibits the distant metastasis of breast cancer via the phosphorylation of the downstream molecules of CXCL12/CXCR4. However, whether or not Slit/Robo determines the variant expression of CXCR4 in different breast cancer patients remains to be determined. Over the last decade, CXCL12/CXCR4 axis antagonists have garnered considerable interest as promising treatments for cancer metastases, and significant progress has been achieved from the results of in vivo experiments.

Slit proteins and their Robo receptors are abnormally expressed in various cancer tissue and have been associated with LOH, hypermethylation, and mutation in their promoters. Therefore, understanding the exact molecular mechanism of Slit/Robo signaling may improve disease prognosis and provide new ideas and strategies for targeted therapy in breast cancer.

References

[1] Nelson HD, Tyne K, Naik A, Bougatsos C, Chan BK, Humphrey L. Screening for breast cancer: an update for the U.S. Preventive Services Task Force. Ann Intern Med 2009; 151: 727–737

[2] Landis SH, Murray T, Bolden S,Wingo PA. Cancer statistics, 1999. CA Cancer J Clin 1999; 49: 8–31

[3] Greenberg PA, Hortobagyi GN, Smith TL, Ziegler LD, Frye DK, Buzdar AU. Long-term follow-up of patients with complete remission following combination chemotherapy for metastatic breast cancer. J Clin Oncol 1996; 14(8): 2197–2205

[4] Rothberg J M, Hartley DA, Walther Z,Artavanis-Tsakonas S. slit: an EGF-homologous locus of D. melanogaster involved in the development of the embryonic central nervous system. Cell 1988; 55: 1047–1059

[5] Qiu H, Zhu J, Yu J, Pu H, Dong R. SLIT2 is epigenetically silenced in ovarian cancers and suppresses growth when activated. Asian Pac J Cancer Prev 2011; 12(3): 791–795

[6] Huang Z, Wen P, Kong R, Cheng H, Zhang B, Quan C, Bian Z, Chen M, Zhang Z, Chen X, Du X, Liu J, Zhu L, Fushimi K, Hua D, Wu JY. USP33 mediates Slit-Robo signaling in inhibiting colorectal cancer cell migration. Int J Cancer 2015; 136(8): 1792–1802

[7] Morlot C, Thielens NM, Ravelli RB, Hemrika W, Romijn RA, Gros P, Cusack S, McCarthy AA. Structural insights into the Slit-Robo complex. Proc Natl Acad Sci USA 2007; 104(38): 14923–14928

[8] Wu JY, Feng L, Park HT, Havlioglu N, Wen L, Tang H, Bacon KB, Jiang Zh, Zhang Xc, Rao Y. The neuronal repellent Slit inhibits leukocyte chemotaxis induced by chemotactic factors. Nature 2001; 410(6831): 948–952

[9] Holmes GP, Negus K, Burridge L, Raman S, Algar E, Yamada T, Little MH. Distinct but overlapping expression patterns of two vertebrate slit homologs implies functional roles in CNS development and organogenesis. Mech Dev 1998; 79(1-2): 57–72

[10] Hohenester E. Structural insight into Slit-Robo signalling. Biochem Soc Trans 2008; 36( 2): 251–256

[11] Howitt JA, Clout NJ, Hohenester E. Binding site for Robo receptors revealed by dissection of the leucine-rich repeat region of Slit. EMBO J 2004; 23(22): 4406–4412

[12] Ma WJ, Zhou Y, Lu D, Dong D, Tian XJ, Wen JX, Zhang J. Reduced expression of Slit2 in renal cell carcinoma. Med Oncol 2014; 31(1): 768

[13] Prasad A, Fernandis AZ, Rao Y, Ganju RK. Slit protein-mediated inhibition of CXCR4-induced chemotactic and chemoinvasive signaling pathways in breast cancer cells. J Biol Chem 2004; 279(10): 9115–9124

[14] Alvarez C, Tapia T, Cornejo V, Fernandez W, Muñoz A, Camus M, Alvarez M, Devoto L, Carvallo P. Silencing of tumor suppressor genes RASSF1A, SLIT2, and WIF1 by promoter hypermethylation in hereditary breast cancer. Mol Carcinog 2013; 52(6): 475–487

[15] Dallol A, Da Silva NF, Viacava P, Minna JD, Bieche I, Maher ER, Latif F. SLIT2, a human homologue of the Drosophila Slit2 gene, has tumor suppressor activity and is frequently inactivated in lung and breast cancers. Cancer Res 2002; 62(20): 5874–5880

[16] Kim GE, Lee KH, Choi YD, Lee JS, Lee JH, Nam JH, Choi C, Park MH, Yoon JH. Detection of Slit2 promoter hypermethylation in tissue and serum samples from breast cancer patients. Virchows Arch 2011; 459(4): 383–390

[17] Shivapurkar N, Maitra A, Milchgrub S, Gazdar AF. Deletions of chromosome 4 occur early during the pathogenesis of colorectal carcinoma. Hum Pathol 2001; 32(2): 169–177

[18] Shivapurkar N, Sood S, Wistuba II, Virmani AK, Maitra A, Milchgrub S, Minna JD, Gazdar AF. Multiple regions of chromosome 4 demonstrating allelic losses in breast carcinomas. Cancer Res 1999; 59(15): 3576–3580

[19] Dallol A, Forgacs E, Martinez A, Sekido Y, Walker R, Kishida T, Rabbitts P, Maher ER, Minna JD, Latif F. Tumour specific promoter region methylation of the human homologue of the Drosophila Roundabout gene DUTT1 (ROBO1) in human cancers. Oncogene 2002; 21(19): 3020–3028

[20] Prasad A, Paruchuri V, Preet A, Latif F, Ganju RK. Slit-2 induces a tumor-suppressive effect by regulating beta-catenin in breast cancer cells. J Biol Chem 2008; 283(39): 26624–26633

[21] Takeichi M. Cadherin cell adhesion receptors as a morphogenetic regulator. Science 1991; 251(5000): 1451–1455

[22] Marlow R, Strickland P, Lee JS, Wu X, Pebenito M, Binnewies M, Le EK, Moran A, Macias H, Cardiff RD, Sukumar S, Hinck L. SLITs suppress tumor growth in vivo by silencing Sdf1/Cxcr4 within breast epithelium. Cancer Res 2008; 68(19): 7819–7827

[23] Papkoff J, Rubinfeld B, Schryver B, Polakis P. Wnt-1 regulates free pools of catenins and stabilizes APC-catenin complexes. Mol Cell Biol 1996; 16(5): 2128–2134

[24] Tetsu O, McCormick F. Beta-catenin regulates expression of cyclin D1 in colon carcinoma cells. Nature 1999; 398(6726): 422–426

[25] Chang PH, Hwang-Verslues WW, Chang YC, Chen CC, Hsiao M, Jeng YM, Chang KJ, Lee EY, Shew JY, Lee WH. Activation of Robo1 signaling of breast cancer cells by Slit2 from stromal fibroblast restrains tumorigenesis via blocking PI3K/Akt/β-catenin pathway. Cancer Res 2012; 72(18): 4652–4661

[26] Müller A, Homey B, Soto H, Ge N, Catron D, Buchanan ME, McClanahan T, Murphy E, Yuan W, Wagner SN, Barrera JL, Mohar A, Verástegui E, Zlotnik A. Involvement of chemokine receptors in breast cancer metastasis. Nature 2001; 410(6824): 50–56

[27] Yuasa-Kawada J, Kinoshita-Kawada M, Rao Y, Wu JY. Deubiquitinating enzyme USP33/VDU1 is required for Slit signaling in inhibiting breast cancer cell migration. Proc Natl Acad Sci USA 2009; 106(34): 14530–14535

[28] Avraham H, Park SY, Schinkmann K, Avraham S. RAFTK/Pyk2-mediated cellular signalling. Cell Signal 2000; 12: 123–133

[29] Park SY, Avraham H, Avraham S. Characterization of the tyrosine kinases RAFTK/Pyk2 and FAK in nerve growth factor-induced neuronal differentiation. J Biol Chem 2000; 275: 19768–19777

[30] Nabeshima K, Inoue T, Shimao Y, Sameshima T. Matrix metalloproteinases in tumor invasion: role for cell migration. Pathol Int 2002; 52(4): 255–264

[31] Beemiller P, Zhang Y, Mohan S, Levinsohn E, Gaeta I, Hoppe AD, Swanson JA. A Cdc42 activation cycle coordinated by PI 3-kinase during Fc receptor-mediated phagocytosis. Mol Biol Cell 2010; 21(3): 470–480

[32] Wong K, Ren XR, Huang YZ, Xie Y, Liu G, Saito H, Tang H, Wen L, Brady-Kalnay SM, Mei L, Wu JY, Xiong WC, Rao Y. Signal transduction in neuronal migration: roles of GTPase activating proteins and the small GTPase Cdc42 in the Slit-Robo pathway. Cell 2001; 107(2): 209–221

[33] Ghose A, Van Vactor D. GAPs in Slit-Robo signaling. BioEssays 2002; 24(5): 401–404

[34] Cau J, Faure S, Comps M, Delsert C, Morin N. A novel p21-activated kinase binds the actin and microtubule networks and induces microtubule stabilization. J Cell Biol 2001; 155(6): 1029–1042

[35] Mertens G, Cassiman JJ, Van den Berghe H, Vermylen J, David G. Cell surface heparan sulfate proteoglycans from human vascular endothelial cells. Core protein characterization and antithrombin III binding properties. J Biol Chem 1992; 267(28): 20435–20443

[36] Liang Y, Annan RS, Carr SA, Popp S, Mevissen M, Margolis RK, Margolis RU. Mammalian homologues of the Drosophila slit protein are ligands of the heparan sulfate proteoglycan glypican-1 in brain. J Biol Chem 1999; 274(25): 17885–17892

[37] Schmid BC, Rezniczek GA, Fabjani G, Yoneda T, Leodolter S, Zeillinger R. The neuronal guidance cue Slit2 induces targeted migration and may play a role in brain metastasis of breast cancer cells. Breast Cancer Res Treat 2007; 106(3): 333–342

[38] Lee TH, Avraham HK, Jiang S, Avraham S. Vascular endothelial growth factor modulates the transendothelial migration of MDA-MB-231 breast cancer cells through regulation of brain microvascular endothelial cell permeability. J Biol Chem 2003; 278(7): 5277–5284

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg