SLIT2, a Human Homologue of the Drosophila Slit2 Gene, Has Tumor Suppressor Activity and Is Frequently Inactivated in Lung and Breast Cancers1

[CANCER RESEARCH 62, 5874–5880, October 15, 2002] SLIT2, a Human Homologue of the Drosophila Slit2 Gene, Has Tumor Suppressor Activity and Is Frequently Inactivated in Lung and Breast Cancers1

Ashraf Dallol, Nancy Fernandes Da Silva, Paolo Viacava, John D. Minna, Ivan Bieche, Eamonn R. Maher, and Farida Latif2 Section of Medical and Molecular Genetics, Division of Reproductive and Child Health, University of Birmingham, Birmingham B15 2TT, United Kingdom [A. D., N. F. D. S., E. R. M., F. L.]; CRC Renal Molecular Oncology Research Group, University of Birmingham, Birmingham B15 2TG, United Kingdom [N. F. D. S., E. R. M., F. L.]; Department of Oncology, Division of Pathology, University of Pisa, Pisa, Italy [P. V.]; Hamon Center for Therapeutic Oncology Research, University of Texas Southwestern Medical Center, Dallas, Texas 75390 [J. D. M.]; and Laboratoire d’Oncoge´ne´tique-INSERM E0017, Centre Rene´Huguenin, F-92210 St-Cloud, France [I. B.]

ABSTRACT knot (Fig. 1; Ref. 4). These motifs are also found in other proteins and have been implicated in mediating protein-protein and ECM-protein Slit2 plays a vital role in axon guidance by signaling through Robo interactions. The SLIT2 protein can be present in three forms. The receptors. Recent evidence suggests that Slit2 protein may function in full-length protein (M ϳ200,000) is proteolytically cleaved into an other settings because human and Xenopus Slit2 has been shown to inhibit r leukocyte chemotaxis. SLIT2 protein is a putative ligand for the ROBO NH2-terminal fragment and a smaller COOH-terminal fragment. It is receptors. We recently demonstrated that ROBO1 is inactivated by pro- thought that different axons have specific responses to particular moter region hypermethylation in <20% of human cancers; furthermore, SLIT2 fragments (5). tumor suppressor activity has not been shown. Thus, the importance of SLIT2 is expressed in neuronal as well as nonneuronal tissues. Rat ROBO1 inactivation in human cancer is uncertain. Therefore, we inves- Slit2 is expressed in lung, breast, kidney, and heart (6). In situ tigated the status of SLIT2 located at 4p15.2 in lung and breast cancers. hybridization of human kidneys revealed the expression of SLIT2 in Although somatic SLIT2 mutations were not detected, epigenetic inacti- mesangial and epithelial cells in the glomeruli, in epithelial cells in the vation was common. SLIT2 promoter methylation was detected in 59% of tubules, and in endothelial cells of both arterioles and venules in the breast cancer, 77% of non-small cell lung cancer, and 55% of small cell kidney (6). In addition, Unigene cluster Hs.29802 shows that SLIT2 is lung cancer cell lines. In these tumor lines, SLIT2 expression was restored expressed in adult lung and kidney in addition to heart. This broad by treatment with a demethylating agent. SLIT2 promoter methylation was detected in 43% of breast cancer, 53% of non-small cell lung cancer, pattern of SLIT2 expression suggests that it may have other functions. and 36% of small cell lung cancer primary tumors. The majority of Slit2 has been shown to inhibit leukocyte chemotaxis (6), possibly methylated tumors demonstrated allelic loss at 4p15.2. In addition, SLIT2 through the Slit2 effect on actin cytoskeleton organization mediated expression was down-regulated in methylated breast tumors, relative to through cdc42 (7). Slit protein triggers Robo-DCC (deleted in colon normal control, as demonstrated by quantitative real-time reverse tran- cancer) interaction with the subsequent loss of responsiveness of DCC scription-PCR. Overexpression of SLIT2 suppressed >70% of colony to its ligand, netrin-1 (8). This loss may activate apoptotic pathways growth in each of three breast tumor lines (with either absent or low through caspase 3 and caspase 9 (9). Exon 2 of ROBO1 is homozy- SLIT2 expression). Because SLIT2 is primarily a secreted protein, SLIT2- gously deleted in two lung tumor cell lines and one breast tumor cell conditioned medium suppressed the growth of several breast cancer lines line (10). Most mice with deletion of exon 2 of Robo1 receptor die (with absent or weak SLIT2 expression) by 26–51% but had no significant effect on a breast tumor cell line that expresses normal levels of SLIT2. because of delayed lung maturation, whereas the surviving mice These findings demonstrate that SLIT2 is frequently inactivated in lung develop bronchial hyperplasia (11). and breast cancer by promoter region hypermethylation and allele loss The SLIT2 gene has been mapped to chromosome 4p15.2 (12). and is an excellent candidate for the lung and breast tumor suppressor Microsatellite markers from this region show LOH in 63% of me- gene previously mapped to 4p15.2. sothelioma, 60% of SCLC, and 25% of NSCLC (13). Around 63% of breast tumors also show LOH at 4p15.1–15.3 (14). Deletion of the INTRODUCTION 4p15.1–15.3 region has also been shown to occur in colorectal carcinoma (15), invasive cervical cancer (16), head and neck squamous The Slit family comprises large ECM3 secreted and membrane- cell carcinoma (17), and bladder cancer (18). associated glycoproteins. Human Slits (SLIT1, SLIT2, and SLIT3) are Along with mutation screening of candidate TSGs, methylation candidate ligands for the repulsive guidance receptors, the ROBO analysis of promoter region CpG islands (if present) is required to gene family. The Slit-Robo interactions mediate the repulsive cues on study tumor-specific inactivation. Aberrant methylation of promoter axons and growth cones during neural development. The Slit2-Robo1 region CpG islands has been largely associated with gene silencing in interaction has been shown to be enhanced by cell surface heparan human cancers. This has been reported for several TSGs, most notably sulfates including glypican-1 (1, 2). The Slit proteins are highly for the new TSG RASSF1A located at 3p21.3 (19, 20). conserved in evolution and contain a putative signal peptide, four We have shown that the putative SLIT2 receptor, DUTT1/ROBO1, tandem arrays of leucine-rich repeats required for the Slit-Robo in- is methylated in Ͻ20% of breast tumors and clear cell renal cell teraction and axon repulsion (3), seven to nine EGF repeats, an carcinomas. Methylation of the putative DUTT1 promoter was a rare agrin-laminin-perlecan-Slit conserved spacer motif, and a cysteine event in lung carcinomas (21). The coding region of the DUTT1 gene lacked any inactivating mutations, and the gene was expressed in the Received 5/29/02; accepted 8/20/02. majority of tumor cell lines examined. However, LOH of DUTT1 The costs of publication of this article were defrayed in part by the payment of page occurred frequently in lung, breast, and kidney tumors, prompting us charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. to hypothesize that haploinsufficiency of DUTT1 coupled with hap- 1 Supported in part by Cancer Research UK (E. R. M. and F. L.) and by Specialized loinsufficiency/inactivation of its ligand could be detrimental for Programs of Research Excellence P50 CA70907 (to J. D. M.). tumor progression and metastasis. Thus ROBO1 remains a candidate 2 To whom requests for reprints should be addressed. Phone: 121-627-2741; Fax: 121-627-2618; E-mail: [email protected]. for the lung and breast 3p12 TSG (22). We reasoned that if ROBO1 3 The abbreviations used are: ECM, extracellular matrix; NSCLC, non-small cell lung inactivation was a significant event in human breast and lung cancers, cancer; SCLC, small cell lung cancer; RT-PCR, reverse transcription-PCR; EGF, epider- mal growth factor; LOH, loss of heterozygosity; TSG, tumor suppressor gene; 5-aza-dC, then other components of the signaling pathway might also be in- 5-aza-2Ј-deoxycytidine; MSP, methylation-sensitive PCR; Ct, control. volved in cancer. Therefore, we undertook genetic, epigenetic, and 5874

Fig. 1. Genomic structure of SLIT2. The SLIT2 gene spans a genomic region of approximately 300 kb and is composed of 37 coding exons. The CpG island fulfilled the criteria of GC content Ͼ 0.5 and an observed:expected CpG ratio Ͼ 0.6 (CpG Plot program; http://www.ebi.ac.uk/Tools/). The exon sizes ranged from 24 to 383 bp. Thirteen exons share the same size of 72 bp. The exons are colored according to the domains they code for. Green indicates a cysteine-rich NH2-terminal flanking region; blue indicates a cysteine-rich COOH-terminal flanking region. Gray indicates the leucine-rich tandem repeats. Dark purple indicates the EGF-like domains, whereas light purple indicates the laminin G domains. Orange indicates the cysteine knot. The arrows indicate the relative position of the proteolytic cleavage site and the relative position of the marker D4S1546. functional analysis of the SLIT2 gene in lung and breast cancers. We GGGATTGTTTAGATATTT-3Ј) and Sli2MOD4R2 (5Ј-CAAAAACTCCT- demonstrate that the promoter region of SLIT2 is frequently hyper- TAAACAACTTTAAATCCTAAAA-3Ј). We used 0.02 volume of the first methylated in breast and lung cancer and that this mode of gene PCR reaction (with primers Sli2MOD4F and Sli2MOD4R2) in another PCR inactivation is the main cause of loss of function in these tumors. We reaction using Sli2MOD4F and the nested primer, Sli2MOD4R (5Ј-ACTA- Ј also demonstrate that overexpression of exogenous wild-type SLIT2 AAACTTCCAACAACTACTAAAATACAAAAA-3 ). The PCR conditions causes loss of colony formation ability in vitro and that conditioned for both the first and second PCR were 95°C for 10 min, followed by 30–40 cycles of 1 min of denaturation at 95°C, 1 min of annealing at 54°C, and 2 min medium from COS-7 cells transfected with SLIT2 also suppresses of extension at 74°C. The PCR products were cloned into pGEM T-Easy breast cancer growth. Our results suggest that SLIT2 should be re- vector, and at least five clones were sequenced. garded as a candidate TSG from chromosome region 4p15.2 and that The methylation-sensitive primers were designed based on the sequencing it is inactivated frequently in lung and breast cancers by promoter data of the PCR-amplified bisulfite-modified cell line genomic DNA. The region hypermethylation and allelic loss. methylated alleles were amplified using Sli2-MSP-F (5Ј-CGGTTTAGGTT- GCGGCGGAGTCGAGGGC-3Ј) and Sli2-MSP-R (5Ј-CGCGAAAAC- Ј MATERIALS AND METHODS CCAACGAACCCGTAACAAAACGCG-3 ). The PCR cycling conditions for this reaction consisted of 10 min at 95°C followed by 30 cycles of 1 min of Patients and Samples. DNA from a total of 36 SCLC and 30 NSCLC denaturation at 95°C, 1 min of annealing at 68°C, and 2 min of extension at tumors (6 squamous cell tumors, 11 adenocarcinomas, 3 large cell tumors, 1 74°C. The unmethylated allele was amplified using primers Sli2-USP-F (5Ј- adenosquamous tumor, and 9 tumors of unknown histopathology) and corre- TGGTTTAGGTTGTGGTGGAGTTGAGGGT-3Ј) and Sli2-USP-R (5Ј-CA- sponding normal tissue and 31 SCLC and 17 NSCLC tumor cell lines was CAAAAACCCAACAAACCCATAACAAAACACA-3Ј). The PCR cycling analyzed. The primary tumors and corresponding normal tissue were obtained conditions were similar to those above, but annealing was done at 59°C. by either autopsy or operation (20). Cell Lines and 5-aza-dC Treatment. Lung and breast carcinoma cell lines A total of 37 invasive ductal breast carcinomas and corresponding normal were routinely maintained in RPMI 1640 (Invitrogen, San Diego, CA) sup- breast tissue were analyzed. Formalin-fixed, paraffin-embedded sections were plemented with 10% FCS at 37°C, 5% CO2. The demethylating agent 5-aza-dC obtained as described previously (23). They were either detected by mammo- (Sigma) was freshly prepared in dd H2O and filter sterilized. Cells (5– graphic screening or had presented symptomatically. None of the tumors were 10 ϫ 105) were plated in a 25-cm2 flask in RPMI 1640 supplemented with 10% from women with a known family history of breast or other cancers. All tissue FCS. Twenty-four h later, cells were treated with 10 ␮M 5-aza-dC. The histological assessment was performed by P. V. In addition, DNA obtained medium was changed 24 h after treatment and then changed every 3 days. from an additional 16 breast tumors (frozen tissue) and 27 breast tumor cell RNA was prepared at 10 days after treatment using the RNeasy kit (Qiagen) lines was also analyzed for SLIT2 methylation. according the manufacturer’s guidelines. SLIT2 expression was detected by Bisulfite Modification and Methylation Analysis. Bisulfite DNA se- RT-PCR using the primers 5Ј-GGTGTCCTCTGTGATGAAGAG-3Ј and 5Ј- quencing was performed as described previously (20). Briefly, 0.5–1.0 ␮gof GTGTTTAGGAGACACACCTCG-3Ј. The expected product size is 387 bp. genomic DNA was denatured in 0.3 M NaOH for 15 min at 37°C, and then Expression of GAPDH was used as a control. The GAPDH primers were unmethylated cytosine residues were sulfonated by incubation in 3.12 M 5Ј-TGAAGGTCGGAGTCAACGGATTTGGT-3Ј and 5Ј-CATGTGGGC- sodium bisulfite (pH 5.0; Sigma)/5 mM hydroquinone (Sigma) in a thermocy- CATGAGGTCCACCAC-3Ј. The expected product size is 982 bp. cler (Hybaid) for 20 cycles of 30 s at 99°C and 15 min at 50°C. The sulfonated Mutation Analysis of Primary Tumors and Cell Lines. Intron-exon DNA was recovered using the Wizard DNA cleanup system (Promega) in boundaries were determined by matching the cDNA sequence for SLIT2 accordance with the manufacturer’s instructions. The conversion reaction was (GenBank accession number AB017168) with the working draft sequence of 5 completed by desulfonating in 0.3 M NaOH for 10 min at room temperature. human genome, the UCSC assembly. Mutation screening of 37 exons was The DNA was ethanol-precipitated and resuspended in water. performed on 22 SCLC, 9 NSCLC, and 7 breast cell lines by the PCR-single- The SLIT2 putative promoter region was predicted by Promoter Inspector strand conformational polymorphism method using intronic primers. Aber- software.4 This region is from Ϫ761 to Ϫ212 relative to the translation start rantly migrating bands were sequenced on an ABI377 automated sequencer. site. The region was amplified from breast and lung cell lines not expressing The sequences of the primers used, along with the annealing temperature and SLIT2 or expressing SLIT2 using the primers Sli2MOD4F (5Ј-GGGAGGT- expected product size, are available upon request.

4 http://www.genomatix.de. 5 http://www.genome.ucsc.edu. 5875

Microsatellite Repeat Analysis. PCR amplification of the dinucleotide Western blotted by standard methods with a monoclonal 9E10 anti-myc repeat microsatellite sequence of marker D4S1546 was performed on 36 antibody (Clontech) or anti-SLIT2 antibody (Ref. 5; gift from M. Tessier- SCLC, 30 NSCLC, and 37 breast carcinomas. The D4S1546 marker was the Lavigne, Stanford University, Stanford, CA). Total protein levels were quan- nearest possible marker (within 100 kb) available to the SLIT2 gene on 4p15.2 tified using the Bradford assay. according to GDB and the UCSC assembly of the human genome sequence. Plasmid Constructs and Growth Suppression Analysis. The SLIT2 ex- The PCR conditions were 96°C for 5 min, followed by the addition of 0.2 unit pression construct was made by cloning the full-length human Slit2 coding of Taq polymerase (Invitrogen), followed by 35 cycles of 96°C for 30 s, 52°C region into the BamHI-XbaI sites of pSecTagB vector (Invitrogen); thus, it for 30 s, and 72°C for 30 s. A final extension step of 5 min at 72°C was contains both myc and His epitope tags at its COOH terminus (5). Transfection performed. with FuGENE 6 reagent (Roche Molecular Biochemicals) used 2 ␮g of each Real-Time RT-PCR. Quantitative real-time RT-PCR was performed as plasmid per 25-cm2 flask containing 5 ϫ 105 cells seeded 24 h before described previously (24). Briefly, cDNA was made from total RNA extracted transfection. Forty-eight h after transfection, 5 ϫ 104 transfected cells were from frozen normal and tumor breast tissues. We quantified transcripts of the seeded and maintained in RPMI 1640 and 10% fetal bovine serum supple- TBP (TATA box-binding protein) gene as the endogenous RNA control. Each mented with 50 ␮g/ml zeocin (Invitrogen). Surviving colonies were counted sample was normalized on the basis of TBP content. Results, expressed as 14–21 days later, after staining with crystal violet. SLIT2-conditioned media N-fold differences in target gene expression relative to the TBP gene (termed were prepared by transient transfection of COS-7 cells with SLIT2 or vector target target ϭ ⌬ N ), were determined by the following formula: N 2 Ctsample, where only plasmids. Forty-eight h after transfection, the media were cleared by the ⌬Ct value of the sample was determined by subtracting the average centrifugation, and a total of 10 nM SLIT2 protein was added to 5 ϫ 104 cells Ct value of the target gene from the average Ct value of the TBP gene. The seeded 24 h earlier in each well of 6-well plates. After 4 days, the cells in each Ntarget values of the samples were subsequently normalized such that the mean indicated conditioned medium were counted using a hemocytometer. ratio of the normal breast samples would equal a value of 1. The nucleotide sequences of the primers used for PCR amplification were as follows: (a) RESULTS SLIT2-U (5Ј-CGTTTGGAAAATGTGCAGCATAA-3Ј) and SLIT2-L (5Ј- TTCGATTGCTTCTCAACATCAAAGT-3Ј) with a SLIT2-specific product SLIT2 Is Silenced by Hypermethylation in Lung and Breast size of 76 bp; and (b) TBP-U (5Ј-TGCACAGGAGCCAAGAGTGAA-3Ј) and Cancer Cell Lines and Is Down-Regulated in Breast Cancer. We TBP-L (5Ј-CACATCACAGCTCCCCACCA-3Ј) with a TBP-specific product analyzed 15 breast and lung tumor cell lines for SLIT2 expression size of 132 bp. PCR was performed using the SYBR Green PCR Core using RT-PCR and found that the majority (62%) had weak or absent Reagents kit (Perkin-Elmer Applied Biosystems). The thermal cycling condi- SLIT2 RNA expression (Fig. 3A). We analyzed these cell lines for tions comprised an initial denaturation step at 95°C for 10 min and 50 cycles at 95°C for 15 s and 65°C for 1 min. Experiments were performed with SLIT2 promoter region hypermethylation (as determined by the pres- duplicates for each data point. ence of a CpG island and analysis of the SLIT2 gene with Promoter- 4 Antibodies and Western Blotting. SLIT2- or vector only-transfected Inspector software; Fig. 2A). After bisulfite modification of cell line COS-7 cells or conditioned media were solubilized in SDS-PAGE sample DNA and PCR amplification of the SLIT2 promoter region, we loading buffer and run out on a SDS-PAGE gel (7.5%). The samples were sequenced at least five alleles for each of the nine cell lines (Fig. 2B).

Fig. 2. Methylation analysis of SLIT2. A, sequence of the SLIT2 promoter region showing the relative positions of the primers used for amplification. The primer sequences are given in “Materials and Methods.” The positions of the CpG dinucleotides analyzed are numbered 1–35. B, a composite representation of the methylation status of each CpG site (labeled 1–35 in A) in the region Ϫ761 to Ϫ212 relative to the translation start site determined by sequencing bisulfite-treated genomic DNA from at least 5 alleles/tumor cell line (breast tumor lines are shown, with the exception of NCI-H460, which is a NSCLC cell line) or primary tumor samples. White and black squares represent unmethylated and methylated CpG, respectively. Partially filled squares represent partially methylated CpG. 5876

Fig. 3. Promoter region methylation silences SLIT2 expression. A, expression of SLIT2 mRNA as detected by RT-PCR in a panel of human breast and lung cancer cell lines and its relation with methylation status. Completely methylated cell lines are indicated by ϩ. Partially methylated cell lines are indicated by ϩ/Ϫ. Unmethylated cell lines are indicated by Ϫ. B, expression of SLIT2 in three breast cancer cell lines after treatment with the demethylating agent 5-aza-dC (10 ␮M) for 10 days. C, expression of SLIT2 in five methylated breast tumors relative to normal breast controls. Results are expressed as N-fold differences for SLIT2 expression relative to TBP expression and relative to the mean of three normal breast tissue samples. Arbitrarily, we have given a value of 1 for this normal mean value (see “Materials and Methods” and Ref. 24).

We also sequenced normal/tumor pairs from breast and lung cancer primary NSCLC tumors (Table 1; Fig. 4). SLIT2 methylation was not and found that they exhibited tumor-specific SLIT2 hypermethylation detected in normal blood samples from healthy controls or in normal (Fig. 2B). SLIT2 promoter region hypermethylation correlated with tissue samples from patients without SLIT2 tumor methylation. Nor- expression status, and all cell lines with promoter methylation dem- mal tissue methylation was detected (but at a lower degree than the onstrated reduced (methylated as well as unmethylated alleles de- corresponding tumor samples) in 14%, 13%, and 8% of methylated tected) or absent (only methylated alleles detected) mRNA expres- breast, NSCLC, and SCLC tumors. We could detect the unmethylated sion. Cell lines with apparently normal SLIT2 mRNA levels had no allele in almost all tumors tested. This could be attributed to contam- detectable methylation (Fig. 3A). To establish that methylation was ination by normal tissue because none of the tumor tissues analyzed responsible for silencing SLIT2 gene expression, we treated three were microdissected. There was no correlation between SLIT2 meth- breast cancer cell lines with SLIT2 methylation with the demethylat- ylation and grade of tumors, and there was no correlation between ing agent 5-aza-dC. SLIT2 expression was restored in these lines after SLIT2 methylation and histology of NSCLCs. SLIT2 methylation was 5-aza-dC treatment (Fig. 3B). Using quantitative real-time RT-PCR, seen even in grade 1 breast tumors, SCLCs, and NSCLCs. we found that SLIT2 expression was significantly down-regulated in The SLIT2 Gene Is Not Commonly Mutated in Lung and Breast breast tumors, in which it is methylated compared with SLIT2 expres- Cancers. We analyzed the 37 coding exons of SLIT2 by PCR-single- sion in normal breast tissues (Fig. 3C). strand conformational polymorphism in the primary lung tumors and SLIT2 Methylation Is Widespread in Primary Lung and Breast Cancer. We used the sequence information obtained from the meth- Table 1 Incidence of methylation in breast and lung tumors, cell lines, and ylated and unmethylated cell lines, as well as tumors, to design a MSP nonmalignant control tissues assay. To establish the frequency of SLIT2 methylation in breast Samples Total no. No. methylated (%) cancer, we analyzed a total of 27 breast tumor cell lines and 37 Breast primary invasive ductal breast carcinomas with matched histologi- Cell lines 27 16 (59%) Primary tumors 53 23 (43%) cally normal tissues. We also analyzed an additional 16 primary Nonmalignant breast 37 5 (14%) invasive ductal breast carcinomas without matched normal tissues NSCLC Cell lines 17 13 (77%) (Table 1; Fig. 4). SLIT2 methylation was detected in 16 of 27 (59%) Primary tumors 30 16 (53%) breast cancer cell lines and 23 of 53 (43%) primary breast tumors. Nonmalignant lung 30 4 (13%) SLIT2 methylation was also detected in 17 of 31 (55%) SCLC cell SCLC Cell lines 31 17 (55%) lines and 13 of 36 (36%) primary SCLC tumors (Table 1; Fig. 4). We Primary tumors 36 13 (36%) also analyzed NSCLC for SLIT2 methylation. SLIT2 methylation was Nonmalignant lung 36 3 (8%) detected in 13 of 17 (77%) NSCLC cell lines and 16 of 30 (53%) Peripheral blood lymphocytes 20 0 (0%) 5877

Fig. 4. Detection of aberrant hypermethylation of the SLIT2 promoter region using the MSP assay. Shown are the results of the MSP assay of the putative SLIT2 promoter using primers specific for amplification of either methylated (M) or unmethylated (U) templates from cancer cell lines or matched normal (N) or tumor (T) tissues.

Table 2 LOH and retention at the SLIT2 locus in methylated (M) and unmethylated harvested from COS-7 cells transfected with SLIT2 on the T-47D, (U) breast, NSCLC, and SCLC cancers MCF-7, and HTB-19 breast cancer cell lines. SLIT2 expression in LOH Retention UIa Total COS-7 cells and conditioned media was confirmed by Western blot- Breast M 2 3 2 7 ting (Fig. 6A). The previously reported proteolytic cleavage of SLIT2 U1 4 3 8 NSCLC M 7 9 0 16 was confirmed (4, 5). SLIT2 induced up to 51% growth suppression or U 1 12 1 14 cell death compared with conditioned medium from vector control- SCLC M 4 5 4 13 transfected COS-7 cells (Fig. 6B). This effect was consistent through- U 4 16 3 23 out three independent experiments. This effect corresponded to the a UI, uninformative cases. level of inhibition of colony formation ability in the three different tumor cell lines. To check whether SLIT2 overexpression has a gen- breast tumor lines without SLIT2 methylation. We detected five in- eral cytotoxic rather than growth-suppressive effect, we analyzed the tronic and two exonic polymorphisms and two missense substitutions response of the unmethylated HCC38 breast tumor cell line (normal (W983R in a breast tumor line and H1090G in a SCLC tumor sample). level of expression) to SLIT2 conditioned medium. SLIT2 had no The H1090G missense substitution was also found in the correspond- significant effect on this line, thus supporting the growth-suppressive ing normal DNA, hence it is a germ-line missense change. The role of SLIT2 (Fig. 6B). W983R missense change was in a breast tumor cell line, and DNA from corresponding normal tissue was not available; hence, this change could be either somatic or germ line. The W983R and H1090G changes occur in evolutionarily conserved tryptophan and histidine, and both of these amino acids are located in an EGF-like domain. SLIT2 Inactivation Is Consistent with Knudson’s Two-hit TSG Hypothesis. LOH analysis at D4S1546 (which maps within 100 kb of SLIT2) demonstrated allele loss in 28% and 30% of informative lung and breast cancers, respectively (Table 2). SLIT2 methylation was present in 88% (seven of eight) NSCLC tumors, 50% (four of eight) of SCLCs, and 67% (two of three) of breast tumors with allele loss at D4S1546. Overexpression of SLIT2 Induces Growth Suppression. Having demonstrated that epigenetic inactivation of SLIT2 is frequent in two major human cancers, we investigated tumor suppressor activity of SLIT2 in in vitro colony formation assays. A wild-type SLIT2 expression plasmid or the pSecTagB empty vector was transfected into three breast cancer cell lines with absent (HTB-19) or low (T-47D and MCF-7) SLIT2 expression. Both SLIT2 and empty constructs expressed a zeocin resistance gene, but transfection with SLIT2 reduced the number of zeocin-resistant colonies by Ͼ70% compared with transfection with empty vector in all three cell lines tested (Fig. 5, A and B). This effect was consistent through five independent experiments and using three independent plasmid DNA preparations. There were no significant differences between the colony formation abilities of the three cell lines tested after SLIT2 transfection (P Ͼ 0.05). Fig. 5. Growth-suppressive effect of SLIT2. A, summary of the results of the colony Growth Suppression by Conditioned Medium from COS-7 formation assays. In each case, the vector-transfected number of zeocin-resistant colonies was set at 100%. Values are the mean Ϯ SD of at least three separate experiments, each Cells Transfected with SLIT2. Because SLIT2 is primarily a se- calculated from triplicate plates. B, crystal violet-stained zeocin-resistant MCF-7 colonies creted ECM protein, we analyzed the effect of conditioned medium after 3 weeks of zeocin selection. 5878

Fig. 6. Growth-suppressive effect of secreted SLIT2. A, detection by Western blot of the different SLIT2 fragments in COS-7 cells or conditioned medium (CM). The antibody against ϳ human SLIT2 detects the full-length protein (Mr 190,000) in addition to the NH2-terminal fragment (Mr 140,000), both of which are indicated by the arrowheads. The anti-MYC antibody detects the full-length protein in addition to the COOH terminus fragment (Mr 50,000–60,000) indicated by the arrowheads. Both fragments could be detected with this antibody from cellular extracts; however, the full-length fragment is undetectable in cleared conditioned medium. B, phase-contrast image of T-47D cells after treatment with conditioned media from vector- or SLIT2-transfected COS-7 cells. Values are the mean Ϯ SD of at least three separate experiments, each calculated from triplicate plates. In each case, the vector control was set at 100%. SLIT2 conditioned medium contains, on average, 10 nM SLIT2.

DISCUSSION located at 3p12 (approximately 5 Mb distal to ROBO1). Homology searches identify ROBO3 as the mouse Rig-1 homologue, located at Tumor-acquired promoter hypermethylation has been recently 11q24. ROBO4 was recently cloned from 11q24 and is thought to be shown as a common inactivation mechanism for TSGs in human involved in angiogenesis (28). cancers (25). We and others have shown that the new TSG RASSF1A The SLIT family is one of four conserved families of axonal is inactivated almost exclusively by promoter region hypermethyla- guidance cues that have prominent developmental effects. The other tion in lung, breast, and kidney cancers (19, 20, 26) and that muta- three are the netrins, the semaphorins, and the ephrins (29). Evidence tions, apart from infrequent germ-line missense substitutions, are rare. is growing for the involvement of these guidance cues and their In this study we report that SLIT2 is a target gene for methylation and receptors in carcinogenesis. SEMA3B, a member of the semaphorin inactivation in lung and breast cancers. The SLIT2 promoter region is family, suppresses growth of adenocarcinoma cancer cell lines (30). methylated in 36–53% of primary lung and breast tumors. SLIT2 SEMA3B also induces apoptosis and causes growth suppression of locus methylation was associated with loss of gene expression in lung lung cancer cell lines (31), possibly through the p53 pathway (32). and breast cancer cell lines, and this loss was reversed by treatment SEMA3B expression is frequently lost in lung cancer, and this loss is with a demethylating agent. Furthermore, SLIT2 expression was also reduced in methylated primary breast tumors. Methylation of one caused by hypermethylation of its promoter region (31). The netrin-1 allele of SLIT2 is mostly associated with deletion of the other allele. receptor, DCC (deleted in colorectal cancer), is frequently inactivated This observation is in agreement with Knudson’s model of biallelic in colorectal cancer (33). DCC loss of expression is also associated inactivation of TSGs (27). The detection of SLIT2 methylation in with promoter region hypermethylation in primary gastric cancer (34). 8–14% of histologically normal matched tissues may indicate that We have recently shown that the putative SLIT2 receptor, DUTT1/ SLIT2 methylation occurs very early in cancer progression. However, ROBO1, is epigenetically inactivated in a subset of kidney and breast it is equally possible to suggest that this detectable methylation is cancers (21). Most mice with an absence of DUTT1/ROBO1 exon 2 caused by the presence of invading tumor cells in the neighboring die because of delayed lung maturation, whereas the surviving mice normal tissue. Mutational analysis of the 37 coding exons of SLIT2 develop bronchial epithelial hyperplasia (11). The finding of SLIT2 has failed to reveal any inactivating changes. However, the two inactivation further strengthens the concept that ROBO1 loss or meth- missense changes we found could have a detrimental effect on the ylation will promote tumorigenesis. protein function because they occur in highly conserved amino acids. In conclusion, aberrant promoter methylation and associated tran- Further analysis of these missense changes is warranted. scriptional silencing are now recognized as a major mechanism of Ectopic expression of SLIT2 in several breast cancer cell lines TSG inactivation. SLIT2 resembles other TSGs such as RASSF1A in suppressed growth and reduced colony formation abilities. Addition- that epigenetic inactivation appears to be much more frequent than ally, secreted SLIT2 in conditioned medium from SLIT2-transfected mutational mechanisms. SLIT2 methylation appears to be a frequent COS-7 cells reduced the cell number and growth of these breast event in two leading cancer types. SLIT2 inactivation provides an cancer cell lines. The mechanism of this growth suppression is un- additional example of the significance of human developmental sig- known. Slit2 protein is reported to be proteolytically cleaved into two naling pathways for cancer biologists and suggests that additional fragments. The NH2 terminus fragment is longer and is primarily studies of SLIT2 in other tumor types are indicated. In addition, other membrane-associated. The shorter, COOH terminus fragment is pri- ROBO and SLIT human homologues may also prove to be implicated marily secreted (5). In the conditioned medium assay, the inhibitory in the pathogenesis of human cancers. Finally, we note that Slit2 effect is presumed to be mediated by the COOH-terminal fragment proteins function as secreted chemorepellants (6), suggesting that (Fig. 6A). ROBO1 is thought to be the putative receptor for SLIT2 (8), SLIT2 inactivation in human cancers might be amenable to therapeutic but there are at least three more ROBO receptors. ROBO2 is also intervention by reversal of epigenetic inactivation and by local ad- 5879

Downloaded from cancerres.aacrjournals.org on September 25, 2021. © 2002 American Association for Cancer Research. SLIT2 HAS TUMOR SUPPRESSOR ACTIVITY ministration of SLIT2 (or functional homologues of SLIT2). In addi- 16. Sherwood, J. B., Shivapurkar, N., Lin, W. M., Ashfaq, R., Miller, D. S., Gazdar, A. F., tion to opportunities to develop novel therapeutic products, the de- and Muller, C. Y. Chromosome 4 deletions are frequent in invasive cervical cancer and differ between histologic variants. Gynecol. Oncol., 79: 90–96, 2000. tection of SLIT2 methylation in sputum or breast aspirates may 17. Pershouse, M. A., El Naggar, A. K., Hurr, K., Lin, H., Yung, W. K., and Steck, P. A. provide novel diagnostic markers for lung and breast cancer. Deletion mapping of chromosome 4 in head and neck squamous cell carcinoma. Oncogene, 14: 369–373, 1997. 18. Polascik, T. J., Cairns, P., Chang, W. Y., Schoenberg, M. P., and Sidransky, D. ACKNOWLEDGMENTS Distinct regions of allelic loss on chromosome 4 in human primary bladder carcinoma. Cancer Res., 55: 5396–5399, 1995. We thank Dr. Marc Tessier-Lavigne for providing the SLIT2 construct and 19. Dammann, R., Li, C., Yoon, J. H., Chin, P. L., Bates, S., and Pfeifer, G. P. Epigenetic the antibody. inactivation of a RAS association domain family protein from the lung tumour suppressor locus 3p21.3. Nat. Genet., 25: 315–319, 2000. 20. Agathanggelou, A., Honorio, S., Macartney, D. P., Martinez, A., Dallol, A., Rader, J., REFERENCES Fullwood, P., Chauhan, A., Walker, R., Shaw, J. A., Hosoe, S., Lerman, M. I., Minna, J. D., Maher, E. R., and Latif, F. Methylation associated inactivation of RASSF1A 1. Hu, H. Cell-surface heparan sulfate is involved in the repulsive guidance activities of from region 3p21.3 in lung, breast and ovarian tumours. Oncogene, 20: 1509–1518, Slit2 protein. Nat. Neurosci., 4: 695–701, 2001. 2001. 2. Ronca, F., Andersen, J. S., Paech, V., and Margolis, R. U. Characterization of Slit 21. Dallol, A., Forgacs, E., Martinez, A., Sekido, Y., Walker, R., Kishida, T., Rabbitts, P., protein interactions with glypican-1. J. Biol. Chem., 276: 29141–29147, 2001. Maher, E. R., and Latif, F. Tumour specific promoter region methylation of the 3. Battye, R., Stevens, A., Perry, R. L., and Jacobs, J. R. Repellent signaling by Slit human homologue of the Drosophila Roundabout gene DUTT1 (Robo1) in human requires the leucine-rich repeats. J. Neurosci., 21: 4290–4298, 2001. cancers. Oncogene, 21: 3020–3028, 2002. 4. Brose, K., Bland, K. S., Wang, K. H., Arnott, D., Henzel, W., Goodman, C. S., 22. Huebner, K. Tumor suppressors on 3p: a neoclassic quartet. Proc. Natl. Acad. Sci. Tessier-Lavigne, M., and Kidd, T. Slit proteins bind Robo receptors and have an evolutionarily conserved role in repulsive axon guidance. Cell, 96: 795–806, 1999. USA, 98: 14763–14765, 2002. 5. Nguyen Ba-Charvet, K. T., Brose, K., Ma, L., Wang, K. H., Marillat, V., Sotelo, C., 23. Viacava, P., Naccarato, A. G., and Bevilacqua, G. Different proliferative patterns Tessier-Lavigne, M., and Chedotal, A. Diversity and specificity of actions of Slit2 characterize different preinvasive breast lesions. J. Pathol., 188: 245–251, 1999. proteolytic fragments in axon guidance. J. Neurosci., 21: 4281–4289, 2001. 24. Bieche, I., Parfait, B., Le Doussal, V., Olivi, M., Rio, M. C., Lidereau, R., and 6. Wu, J. Y., Feng, L., Park, H. T., Havlioglu, N., Wen, L., Tang, H., Bacon, K. B., Vidaud, M. Identification of CGA as a novel estrogen receptor-responsive gene in Jiang, Z., Zhang, X., and Rao, Y. The neuronal repellent Slit inhibits leukocyte breast cancer: an outstanding candidate marker to predict the response to endocrine chemotaxis induced by chemotactic factors. Nature (Lond.), 410: 948–952, 2001. therapy. Cancer Res., 61: 1652–1658, 2001. 7. Wong, K., Ren, X. R., Huang, Y. Z., Xie, Y., Liu, G., Saito, H., Tang, H., Wen, L., 25. Baylin, S. B., and Herman, J. G. DNA hypermethylation in tumorigenesis: epigenetics Brady-Kalnay, S. M., Mei, L., Wu, J. Y., Xiong, W. C., and Rao, Y. Signal joins genetics. Trends Genet., 16: 168–174, 2000. transduction in neuronal migration: roles of GTPase activating proteins and the small 26. Morrissey, C., Martinez, A., Zatyka, M., Agathanggelou, A., Honorio, S., Astuti, D., GTPase Cdc42 in the Slit-Robo pathway. Cell, 107: 209–221, 2001. Morgan, N. V., Moch, H., Richards, F. M., Kishida, T., Yao, M., Schraml, P., Latif, 8. Stein, E., and Tessier-Lavigne, M. Hierarchical organization of guidance receptors: F., and Maher, E. R. Epigenetic inactivation of the RASSF1A 3p21.3 tumor suppres- silencing of netrin attraction by slit through a Robo/DCC receptor complex. Science sor gene in both clear cell and papillary renal cell carcinoma. Cancer Res., 61: (Wash. DC), 291: 1928–1938, 2001. 7277–7281, 2001. 9. Forcet, C., Ye, X., Granger, L., Corset, V., Shin, H., Bredesen, D. E., and Mehlen, P. 27. Knudson, A. G., Jr., Hethcote, H. W., and Brown, B. W. Mutation and childhood The dependence receptor DCC (deleted in colorectal cancer) defines an alternative cancer: a probabilistic model for the incidence of retinoblastoma. Proc. Natl. Acad. mechanism for caspase activation. Proc. Natl. Acad. Sci. USA, 98: 3416–3421, 2001. Sci. USA, 72: 5116–5120, 1975. 10. Sundaresan, V., Chung, G., Heppell-Parton, A., Xiong, J., Grundy, C., Roberts, I., 28. Huminiecki, L., Gorn, M., Suchting, S., Poulsom, R., and Bicknell, R. Magic James, L., Cahn, A., Bench, A., Douglas, J., Minna, J., Sekido, Y., Lerman, M., Latif, roundabout is a new member of the roundabout receptor family that is endothelial F., Bergh, J., Li, H., Lowe, N., Ogilvie, D., and Rabbitts, P. Homozygous deletions specific and expressed at sites of active angiogenesis. Genomics, 79: 547–552, 2002. at 3p12 in breast and lung cancer. Oncogene, 17: 1723–1729, 1998. 29. Yu, T. W., and Bargmann, C. I. Dynamic regulation of axon guidance. Nat. Neurosci., 11. Xian, J., Clark, K. J., Fordham, R., Pannell, R., Rabbitts, T. H., and Rabbitts, P. H. 4 (Suppl.): 1169–1176, 2001. Inadequate lung development and bronchial hyperplasia in mice with a targeted 30. Tse, C., Xiang, R. H., Bracht, T., and Naylor, S. L. Human semaphorin 3B deletion in the Dutt1/Robo1 gene. Proc. Natl. Acad. Sci. USA, 98: 15062–15066, (SEMA3B) located at chromosome 3p21.3 suppresses tumor formation in an adeno- 2001. carcinoma cell line. Cancer Res., 62: 542–546, 2002. 12. Georgas, K., Burridge, L., Smith, K., Holmes, G. P., Chenevix-Trench, G., Ioannou, 31. Tomizawa, Y., Sekido, Y., Kondo, M., Gao, B., Yokota, J., Roche, J., Drabkin, H., P. A., and Little, M. H. Assignment of the human slit homologue SLIT2 to human chromosome band 4p15.2. Cytogenet. Cell Genet., 86: 246–247, 1999. Lerman, M. I., Gazdar, A. F., and Minna, J. D. Inhibition of lung cancer cell growth 13. Shivapurkar, N., Virmani, A. K., Wistuba, I. I., Milchgrub, S., Mackay, B., Minna, and induction of apoptosis after reexpression of 3p21.3 candidate tumor suppressor J. D., and Gazdar, A. F. Deletions of chromosome 4 at multiple sites are frequent in gene SEMA3B. Proc. Natl. Acad. Sci. USA, 98: 13954–13959, 2001. malignant mesothelioma and small cell lung carcinoma. Clin. Cancer Res., 5: 17–23, 32. Ochi, K., Mori, T., Toyama, Y., Nakamura, Y., and Arakawa, H. Identification of 1999. semaphorin 3B as a direct target of p53. Neoplasia, 4: 82–87, 2002. 14. Shivapurkar, N., Sood, S., Wistuba, I. I., Virmani, A. K., Maitra, A., Milchgrub, S., 33. Shibata, D., Reale, M. A., Lavin, P., Silverman, M., Fearon, E. R., Steele, G., Jr., Minna, J. D., and Gazdar, A. F. Multiple regions of chromosome 4 demonstrating Jessup, J. M., Loda, M., and Summerhayes, I. C. The DCC protein and prognosis in allelic losses in breast carcinomas. Cancer Res., 59: 3576–3580, 1999. colorectal cancer. N. Engl. J. Med., 335: 1727–1732, 1996. 15. Shivapurkar, N., Maitra, A., Milchgrub, S., and Gazdar, A. F. Deletions of chromo- 34. Sato, K., Tamura, G., Tsuchiya, T., Endoh, Y., Usuba, O., Kimura, W., and some 4 occur early during the pathogenesis of colorectal carcinoma. Hum. Pathol., Motoyama, T. Frequent loss of expression without sequence mutations of the DCC 32: 169–177, 2001. gene in primary gastric cancer. Br. J. Cancer, 85: 199–203, 2001.

5880

Downloaded from cancerres.aacrjournals.org on September 25, 2021. © 2002 American Association for Cancer Research. SLIT2, a Human Homologue of the Drosophila Slit2 Gene, Has Tumor Suppressor Activity and Is Frequently Inactivated in Lung and Breast Cancers

Ashraf Dallol, Nancy Fernandes Da Silva, Paolo Viacava, et al.

Cancer Res 2002;62:5874-5880.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/62/20/5874

Cited articles This article cites 34 articles, 15 of which you can access for free at: http://cancerres.aacrjournals.org/content/62/20/5874.full#ref-list-1

Citing articles This article has been cited by 23 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/62/20/5874.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/62/20/5874. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.