ITGBL1 Is a Runx2 Transcriptional Target and Promotes Breast Cancer

Published OnlineFirst June 9, 2015; DOI: 10.1158/0008-5472.CAN-15-0240

Cancer Molecular and Cellular Pathobiology Research

ITGBL1 Is a Runx2 Transcriptional Target and Promotes Breast Cancer Bone Metastasis by Activating the TGFb Signaling Pathway Xiao-Qing Li1,2, Xin Du1, Dong-Mei Li1, Peng-Zhou Kong1, Yan Sun3, Pei-Fang Liu4, Qing-Shan Wang1,2, and Yu-Mei Feng1,2

Abstract

Bone metastasis affects more than 70% of advanced breast the TGFb signaling pathway as a downstream effector of cancer patients, but the molecular mechanisms of this process ITGBL1 and the transcription factor Runx2 as an upstream remain unclear. Here, we present clinical and experimental activator of ITGBL1 expression. In support of these ﬁndings, evidence to clarify the role of the integrin b-like 1 (ITGBL1) we also found that ITGBL1 was an essential mediator of Runx2- as a key contributor to bone metastasis of breast cancer. In an in induced bone metastasis of breast cancer. Overall, our results vivo model system and in vitro experiments, ITGBL1 expression illuminate how bone metastasis occurs in breast cancer, and promoted formation of osteomimetic breast cancers, facilitat- they provide functional evidence for new candidate biomarkers ing recruitment, residence, and growth of cancer cells in bone andtherapeutictargetstoidentifyrisk,toprevent,andtotreat microenvironment along with osteoclast maturation there to this dismal feature of advanced breast cancer. Cancer Res; 75(16); form osteolytic lesions. Mechanistic investigations identiﬁed 3302–13. 2015 AACR.

Introduction EGF-like stalk fragment of integrin b (4), but contains neither a transmembrane domain nor an RGD (Arg–Gly–Asp)-binding Breast cancer cells preferentially metastasize to the bone, lead- domain, suggesting that ITGBL1 performs functions distinct ing to osteolytic lesions. To form visible metastatic bone lesions, from those of integrin. In our previous gene-expression proﬁl- breast cancer cells undergo processes, including recruitment to the ing dataset of breast cancer tissues, ITGBL1 was coexpressed bone, survival, and clonal expansion in the bone microenviron- with genes encoding proteins involved in bone remodeling and ment, as well as osteoclast activation to resorb the bone matrix bone metastasis (5). Garcia and colleagues (6) found that (1). Increasing evidence has demonstrated that the ectopic expres- ITGBL1 is overexpressed in bone metastatic subclone cells sion of bone-remodeling genes or gene signatures in primary compared with their parental cells and is included in the breast cancer cells increases the risk of bone metastasis (2); "osteoblast-like gene expression signature" and "bone meta- however, the underlying molecular mechanisms remain largely static gene signature." These evidences implied that ITGBL1 unknown. might contribute to the development of the osteoblast-like Integrin beta-like 1 (ITGBL1), which was ﬁrst cloned and (osteomimetic) phenotype and bone metastasis of breast can- characterized from an osteoblast cDNA library (3), encodes a cer cells. Currently, information regarding the role and molec- ten integrin EGF-like repeat domain-containing protein (TIED). ular mechanism of ITGBL1 expressed by breast cancer cells in The ITGBL1 protein is highly homologous to the N-terminal bone metastasis remains limited. Runx2 is a critical transcription factor for osteogenic lineage

1 commitment and bone formation. It switches on the expression of Department of Biochemistry and Molecular Biology, Tianjin Medical – University Cancer Institute and Hospital, National Clinical Research several bone matrix remodeling genes by binding to the osteo- Center of Cancer, Tianjin, China. 2Key Laboratory of Breast Cancer blast-speciﬁc cis-acting element 2 (OSE2; refs. 7, 8). Runx2 and its Prevention and Treatment of the Ministry of Education,Tianjin Medical target genes are highly expressed in breast cancer tissues and play University Cancer Institute and Hospital, National Clinical Research pivotal roles in breast cancer bone metastasis (9–12). On the basis Center of Cancer, Tianjin, China. 3Department of Pathology, Tianjin Medical University Cancer Institute and Hospital, National Clinical of the evidence that ITGBL1 is coexpressed with RUNX2 in breast Research Center of Cancer, Tianjin, China. 4Department of Radiology, cancer cells (6) and the fact that there are OSE2 motifs in the Tianjin Medical University Cancer Institute and Hospital, National ITGBL1 promoter, we hypothesized that ITGBL1 might be a Clinical Research Center of Cancer, Tianjin, China. transcriptional target of Runx2 and mediate Runx2-driven bone Note: Supplementary data for this article are available at Cancer Research metastasis. Online (http://cancerres.aacrjournals.org/). In this study, we present both in vivo and in vitro evidence to Corresponding Author: Yu-Mei Feng, Tianjin Medical University Cancer Institute clarify the roles and the underlying molecular mechanisms of and Hospital, Huan-Hu-Xi Road, He-Xi District, Tianjin 300060, China. Phone: ITGBL1 in breast cancer bone metastasis. Moreover, we identify 86-22-2334-0123, ext. 6002; E-mail: [email protected] ITGBL1 as a transcriptional target of Runx2, mediating doi: 10.1158/0008-5472.CAN-15-0240 the Runx2-driven bone metastatic potential of breast cancer 2015 American Association for Cancer Research. cells.

3302 Cancer Res; 75(16) August 15, 2015

ITGBL1 in Breast Cancer Bone Metastasis

Materials and Methods Statistical analysis Kaplan–Meier survival curves and the log-rank test were used to Patients and tissue samples evaluate the bone metastasis outcomes of patients with different A total of 88 primary breast cancer tissue specimens were ITGBL1 expression profiles. The Wilcoxon rank-sum test was used collected from breast cancer patients who developed distant to determine the differences in the ITGBL1 expression values metastases within a 5-year follow-up period. Of these patients, among different groups. A repeated measure ANOVA was used 10 patients developed bone-only metastasis, and the other 78 to compare the differences in the proliferation capabilities of patients suffered other organ metastases with or without bone cancer cells. All other comparisons were determined using the metastasis. The use of these tissues was approved by the two-tailed Student t test. Institutional Review Board and the Research Ethics Committee of Tianjin Medical University Cancer Institute and Hospital (TMUCIH), and written consent was obtained from all Results participants. ITGBL1 is coexpressed with a set of genes related to bone remodeling and bone metastasis in primary breast cancer Cells and treatment tissues The human breast cancer MDA-MB-231 and T47D cells, oste- On the basis of our previous gene expression profiling dataset oblast-like MG-63 cells, immortalized lung epithelial BEAS-2B (5) of breast cancer tissues and the GOBO (15) dataset (Fig. 1A), cells, and mouse osteoclast precursor RAW 264.7 cells were we found that ITGBL1 was coexpressed with a set of genes related b obtained from the ATCC. To block the TGF signaling pathway, to bone remodeling and bone metastasis encoding the osteoblast- m cells were treated with 1.0 mol/L SB-431542 (Santa Cruz Bio- specific transcription factor Runx2; osteoblast-specific adhesion technology) diluted in DMSO or with an equal amount of DMSO molecules OB-Cadherin/CDH11 and integrins (ITG); bone as control. matrix proteins, including collagens (COL), osteoblast-specific factor 2 (OSF-2) and osteoglycin (OGN); and the growth factors In vitro chemotaxis assays TGFb3 and bone morphogenetic protein 1 (BMP1). These results In vitro cell chemotaxis assays were performed using Boyden suggest that ITGBL1 may contribute to the formation of "osteo- 4 chamber inserts (BD Biosciences). A total of 2.5 10 cancer cells mimetic" breast cancer, which is characterized by the ectopic were seeded in the upper chambers and allowed to migrate toward expression of bone remodeling-related factors and the properties 80% confluent MG-63 or BEAS-2B cells in the bottom chambers of bone metastasis (16, 17). for 12 hours (MDA-MB-231) or 30 hours (T47D). Then, the fi migrated cells were xed, stained, and counted. Breast cancers with high ITGBL1 expression levels are prone to metastasize to bone In vitro osteoclastogenesis assays Because of the coexpression of ITGBL1 with a set of genes Primary preosteoclasts were isolated from bone marrow cells related to bone remodeling and bone metastasis, we questioned flushed from the tibias of 6-week-old wild-type Balb/c mice and whether ITGBL1 expression is involved in bone metastasis. We 5 cultured overnight in aMEM with 10% FBS. A total of 5 10 used RT-qPCR to quantify the mRNA expression levels of ITGBL1 nonadherent cells were plated in a 24-well plate supplemented in 88 primary breast cancers that developed distant metastases with 50 ng/mL macrophage colony stimulating factor (M-CSF; within a 5-year follow-up period. The optimized cutoff value was PeproTech) for 2 days and then cultured with cancer cell used to group the patients into a low ITGBL1 mRNA group conditioned media (CM) containing 50 ng/mL receptor acti- (ITGBL1low) and a high ITGBL1 group (ITGBL1high). The inci- k vator of NF B ligand (RANKL; PeproTech) for 5 additional dence of bone-only metastasis in the ITGBL1high group (17.0%; days. Osteoclast precursor RAW 264.7 cells were induced for 9/53) was significantly higher than in the ITGBL1low group (2.9%; fi the rst 4 days with 50 ng/mL RANKL and then cultured with 1/35). Patients in the ITGBL1high group had higher risks of bone- cancer cell CM containing RANKL for another 3 days. Multi- only metastasis than patients in the ITGBL1low group [OR, 5.4; nuclear cells that stained positive for tartrate-resistant acid 95% confidence interval (CI), 0.7–42.8; P ¼ 0.072; Fig. 1B]. Next, phosphatase (TRAP) were scored as mature osteoclasts. the combined Wang–Minn microarray dataset (18, 19) revealed that the tumors developing bone metastasis expressed significant- Animal experiments ly higher ITGBL1 mRNA levels than those with non-bone metas- ITGBL1–GFP-overexpressing MDA-MB-231 cells or control tasis (Supplementary Fig. S1A and S1B). The incidences of both cells were injected into the left cardiac ventricle (1.0 106; ref. bone metastasis and bone-only metastasis (25.4% and 22.4%, 6 13), the cortex of the right tibia (1.0 10 ; ref. 14), or the left respectively) were significantly higher in the ITGBL1high group 6 lower abdominal mammary fat pad (5.0 10 ) of 5-week-old than in the ITGBL1low group (14.4% and 8.5%, respectively). In female SCID mice. The formation of tumors and metastases was addition, patients in the ITGBL1high group had higher risks of observed and assessed by bioluminescence imaging using a bone metastasis (OR, 1.8; 95% CI, 1.0–3.0; P ¼ 0.033) and bone- Xenogen IVIS 200 Imaging System (Caliper Life Sciences) at weeks only metastasis (OR, 2.6; 95% CI, 1.3–5.0; P ¼ 0.005) than 4 and 8. At week 8, the mice were sacrificed to observe osteolytic patients in the ITGBL1low group (Fig. 1B and C). Interestingly, lesions using an X-ray and SkyScan microCT system (Bruker, we found a negative link between ITGBL1 mRNA levels and non- Eschborn, GER). Metastases in the lung, liver, and bone were bone metastases (lung and brain metastases; Fig. 1B and C; observed by hematoxylin and eosin (H&E) staining. TRAP stain- Supplementary Fig. S1C–S1E). Taken together, these results pro- ing was used to identify the mature osteoclasts in metastatic bone vide clinical evidence for the role of ITGBL1 in facilitating lesions. The animal experiment protocols were approved by the the bone-specific metastasis of breast cancer. Moreover, on Animal Ethics Committee of TMUCIH. the basis of the Zhang and colleagues dataset (20), the ITGBL1

www.aacrjournals.org Cancer Res; 75(16) August 15, 2015 3303

Li et al.

Figure 1. ITGBL1 is coexpressed with a set of bone metastasis–related genes, and high ITGBL1 expression facilitates breast cancer bone metastasis. A, genes coexpressed with ITGBL1 based on our previous microarray data and the GOBO dataset. Genes marked in red are overlapped in the two datasets. B, Kaplan–Meier survival analysis representing the probability of bone-only metastasis (BM only)–free survival, bone metastasis (BM)–free survival, or non-bone metastasis (NBM)–free survival for patients carrying tumors with different ITGBL1 levels based on our RT-qPCR dataset and the Wang–Minn gene-expression proﬁling dataset. C, heat maps showing the distribution of patients with BM-only and NBM according to tumors with different ITGBL1 levels. D and E, box-and-whisker plot (D) and c2 test (E) comparing ITGBL1 expression levels in bone metastatic tumors with those in metastatic tissues of other organs based on the Zhang and colleagues dataset (20).

mRNA expression levels in bone metastatic tissues were signiﬁ- although ITGBL1-overexpressing cells lead to smaller tumors than cantly higher than in metastatic tissues of other organs (Fig. 1D). the control cells (Fig. 2A). Then, considering the high correlation All of the bone metastatic tumor samples (18/18) were sorted into of ITGBL1 and bone metastasis in clinical breast cancer cases, we the ITGBL1high group, and none of these samples was included in speculated that ITGBL1 in breast cancer cells could endow tumor the ITGBL1low group (Fig. 1E). Thus, this clinical evidence further cells with the ability to seed bones as these cells disseminated in suggests that high ITGBL1 expression may help breast cancer cells the circulation. We inoculated ITGBL1–GFP-overexpressing to survive in the bone microenvironment and facilitate the bone MDA-MB-231 cells or control cells into the left cardiac ventricle metastasis of breast cancer. of SCID mice. The incidence of bone metastasis generated by ITGBL1-overexpressing cells was 75% (3/4) at 90 days (Fig. 2B ITGBL1 contributes to the seeding of bone metastases of cancer and C), which is higher than the incidences generated by the cells and osteolytic bone lesions control cells (0/3; Fig. 2C) and the parental MDA-MB-231 cells Taking into account of the potential effect of tumorigenicity on (31% at 100 days) reported by Kang and colleagues (21). More- the metastatic capability of cancer cells, ITGBL1-overexpressing over, ITGBL1-overexpressing cancer cells led to serious osteolytic þ MDA-MB-231 cells and control cells were injected into the mam- bone lesions with a greater number of TRAP osteoclasts, but mary fat pad of mice. We observed a similar tumorigenicity of reduced lung metastasis (1/4; Fig. 2B–E), whereas all mice (3/3) in ITGBL1-overexpressing cells (5/5) with the control cells (5/5), the control group died of serious lung metastasis within 40 days

3304 Cancer Res; 75(16) August 15, 2015 Cancer Research

ITGBL1 in Breast Cancer Bone Metastasis

A B

Intracardiac Intratibial Orthotopic injection injection injection 1.0 1.0 Tumor incidence ITGBL1 ITGBL1 Group 0.8 0.8 1 Week 2 Weeks 0.6 0.6 Control 5/5 (100%) 5/5 (100%) Control Control 0.4 0.4 ITGBL1 4/5 (80%) 5/5 (100%) 0.2 0.2 P = 0.010 P = 0.078 Overall survival Overall Overall survival Overall 0 0 Control 40 0123 0123 ITGBL1 30 Time (months) Time (months) ) 3 C 20 Injection Metastasis site (mm Group 10 * site Bone Lung Liver Tumor volume volume Tumor 0 Intracardiac Control 0/3 (0%) 3/3 (100%) 1/3 (33%) 1 week 2 weeks ITGBL1 3/4 (75%) 1/4 (25%) 1/4 (25%) Intratibial Control 5/5 (100%) 5/5 (100%) 1/5 (20%) D Lung metastasis ITGBL1 4/4 (100%) 0/4 (0%) 0/4 (0%) Bone metastasis E 12 Control )

11 ITGBL1 * Bioluminescence X ray Bone Lung 9 0 days 4 weeks H&E TRAP H&E 6 BM BM 3 BM BLi (x10 0

Control B B 160 Control ITGBL1 120 T T 80 B B

ITGBL1 40 *

No. of LM Nodules No. of 0 F G Lung metastasis 20 Control Bone osteolytic lesion ) *** 11 ITGBL1 15 Bioluminescence Bone Lung X ray Micro CT 10 4 weeks 8 weeks H&E TRAP H&E 5 ***

B BM BLi (x 10 0 B 4 weeks 8 weeks

Control TT

) 30 Control

3 * ITGBL1 T T 20 B B 10 ITGBL1

T T lesions (mm 0 Volume of osteolytic Volume

Figure 2. ITGBL1 enhances breast cancer seeding to the bone. A, tumor incidence and tumor volume in the mammary fat pad following orthotopic injection of ITGBL1- overexpressing MDA-MB-231 cells (n ¼ 5) or control cells (n ¼ 5). B, Kaplan–Meier survival analysis representing the overall survival of mice with intracardiac injection or intratibial injection of ITGBL1-overexpressing MDA-MB-231 cells or control cells. C, table showing the incidence of metastases in the bone, lung, and liver generated by the indicated cells. D and E, ITGBL1-overexpressing MDA-MB-231 cells (n ¼ 4) or control cells (n ¼ 3) were injected into the left cardiac ventricle of SCID mice. The histograms (E) display the quantitative bioluminescence of BM at the limbs and the number of lung metastatic (LM) nodules. F and G, ITGBL1- overexpressing MDA-MB-231 cells (n ¼ 4) or control cells (n ¼ 5) were injected into the cortex of the right tibia of SCID mice (F). The histograms display the quantitative bioluminescence and volume of osteolytic lesions in the right tibia (G). Representative bioluminescence imaging, X-ray and micro-CT imaging, H&E staining, and TRAP staining show the bone, lung metastatic lesions, and activated osteoclasts. T, tumor cells; B, bone tissue; and BM, normal bone marrow; , P < 0.05; , P < 0.001.

www.aacrjournals.org Cancer Res; 75(16) August 15, 2015 3305

Li et al.

(Fig. 2B–E) without detectable bone metastasis by either X-ray fied, including 19 bone metastasis-related genes identified by imaging or H&E staining. Combining with the evidence that the Kang and colleagues (21), 13 osteoblast-like breast cancer gene ITGBL1-overexpressing cells and control cells are similarly tumor- markers identified by Bellahcene and colleagues (16), 4 receptors igenic, these observations exclude the possibility that overexpres- for chemotactic factors, 3 proliferation-related genes, and several sion of ITGBL1 reduces lethal lung metastasis by decreasing genes involved in ECM remodeling and bone metastasis-related tumorigenicity of cancer cells and indicate a selective and specific signaling pathways (Supplementary Table S2). Then, we demon- role for ITGBL1 in the seeding of breast cancer cells in bone. We strated that ITGBL1 KD reversed the osteomimetic phenotype and also observed more serious bone lesions in the right tibia and a further weakened the bone metastatic potential of MDA-MB-231 þ greater number of TRAP osteoclasts along the bone–tumor cells by clustering the expression profiling of ITGBL1 KD sub- interface of bone lesions in the ITGBL1-overexpressing group clones, control cells and parental cells based on Bellahcene and with intratibial injection (Fig. 2F and G), indicating activated colleagues osteoblast-like signature (16) and Kang and colleagues osteoclasts induced by ITGBL1-overexpressing tumor cells. Sur- bone metastasis signature (21), as shown in Supplementary Fig. prisingly, all mice in the control group with intratibial injection S2B. Using immunoblotting and RT-qPCR, we identified adhe- developed lung metastasis, whereas no mice in the ITGBL1-over- sion molecules involved in bone metastasis, including integrin a5 expressing group developed non-bone metastases (lung or liver (ITGA5; refs. 22, 23) and CD44 (24), and the osteomimetic metastasis; Fig. 2C and F). These results suggest that ITGBL1- markers osteopontin (OPN), bone sialoprotein (BSP), osteonec- overexpressing cancer cells were more likely to arrest in the bone tin (OSN), and OSF-2 that decreased in ITGBL1 KD cells and and that they had a greater survival advantage in the bone increased in ITGBL1-overexpressing cells compared with their microenvironment than control cells. However, some of the control cells (Fig. 3G and Supplementary Fig. S2C and S2D). control cancer cells entered the circulating blood rather than Taken together, these results provide a suitable molecular basis for arresting in bone and then disseminated to the lung and liver. ITGBL1 activity in promoting the bone metastasis of breast cancer.

ITGBL1 facilitates breast cancer cell recruitment to osteoblasts ITGBL1 contributes to osteoclastogenesis in the bone in vitro microenvironment We tested the ability of two ITGBL1 knockdown (KD) MDA- Because of the osteolytic characteristics of breast cancer, we MB-231 subclones (ITGBL1 KD1 and ITGBL1 KD2), two ITGBL1- further investigated whether ITGBL1 in breast cancer cells con- overexpressing MDA-MB-231 subclones (ITGBL1-1 and ITGBL1- tributes to osteoclast formation and differentiation. We isolated 2), ITGBL1-overexpressing T47D cells (ITGBL1), and their respec- primary preosteoclasts from bone marrow cells of mice and tive control cells to migrate toward osteoblast-like MG-63 cells or induced them with cancer cell CM in the presence of RANKL. No þ lung epithelial BEAS-2B cells using a Transwell chemotaxis assay significant changes in the number of TRAP cells were observed in in vitro to investigate whether ITGBL1 facilitates cancer cells the primary osteoclast cultures with cancer cell CM with different þ homing to bone. As shown in Fig. 3A, ITGBL1 KD reduced the ITGBL1 expression levels. However, TRAP osteoclasts induced by chemotactic migration of MDA-MB-231 cells toward MG-63 cells. the CM from ITGBL1 KD cells were significantly smaller than Conversely, ITGBL1 overexpression increased the chemotactic those induced by the CM from control cells, and the CM of þ migration of both MDA-MB-231 and T47D cells toward MG-63 ITGBL1-overexpressing cells induced larger TRAP osteoclasts cells (Fig. 3A). However, ITGBL1 KD increased and ITGBL1 over- than the CM of the control vector cells (Fig. 4A and B). We also expression decreased the chemotactic migration of cancer cells observed the osteoclastogenesis from osteoclast precursor cell line toward BEAS-2B cells (Fig. 3B). Considering the potential role of RAW264.7 cultured with the CM of cancer cells with different ITGBL1 in the formation of osteomimetic breast cancer, we ITGBL1 expression levels in the presence of RANKL. TRAP staining þ speculated that osteomimetic breast cancer with high ITGBL1 showed significantly decreased numbers of TRAP multinuclear expression might adapt to survive in the bone rather than in the cells among the RAW 264.7 cells cultured with the CM of ITGBL1 lung. Therefore, we assessed the colony formation and prolifer- KD cells compared with those cultured with CM from control cells þ ation capabilities of ITGBL1 KD and ITGBL1-overexpressing cells, (Fig. 4C and D). The TRAP osteoclasts induced by CM from as well as their respective control cells, in CM from MG-63 and ITGBL1 KD cells were significantly smaller and contained fewer BEAS-2B cells, which were used to mimic the bone and lung nuclei than those induced by the CM from control cells (Fig. 4E microenvironment. ITGBL1 KD reduced both of the anchorage- and F). Conversely, CM from ITGBL1-overexpressing cells þ independent and anchorage-dependent survival and proliferation induced more TRAP osteoclasts with larger sizes and more nuclei capacity of cancer cells in MG-63 CM, but increased the anchor- than the CM of the control vector cells (Fig. 4C–F). Taken together, age-independent survival in BEAS-2B CM (Fig. 3C–F and Sup- these results suggest that cancer cells with high ITGBL1 expression plementary Fig. S2A). In contrast, ITGBL1 overexpression promoted osteoclast formation and differentiation. Next, we enhanced the survival and proliferation of cancer cells in MG- sought to identify osteoclast-activating cytokines affected by 63 CM and weakened their survival in BEAS-2B CM (Fig. 3C–F and ITGBL1 in breast cancer using a human cytokine antibody array Supplementary Fig. S2A). These results provide in vitro evidence and gene expression profile microarray. The mRNA expression that ITGBL1 plays important roles in cancer cell recruitment to levels of interleukins and the secreted protein levels of granulo- bone and in survival and growth in the bone microenvironment. cyte-macrophage colony stimulating factor (GM-CSF), growth- To further investigate the molecular mechanisms of ITGBL1 in regulated alpha protein (GRO), IL1a, monocyte chemotactic the bone metastasis of breast cancer cells, we performed gene protein (MCP) 1, MCP-3, and macrophage stimulating protein expression profiling of two ITGBL1 KD subclones, control cells alpha (MSP-a) were decreased in the CM of ITGBL1 KD cells and parental MDA-MB-231 cells by cDNA microarray analysis. compared with the CM of control cells (Supplementary Table S2 Differentially expressed genes with more than 2-fold changes in and Fig. 4G). The expression levels of RANKL, a pivotal activator ITGBL1 KD subclones compared with control cells were identi- of osteoclastogenesis, decreased in the ITGBL1 KD cells and

3306 Cancer Res; 75(16) August 15, 2015 Cancer Research

ITGBL1 in Breast Cancer Bone Metastasis

A MDA-MB-231 B MDA-MB-231

Tumor cells Control KD ITGBL1 KD1 ITGBL1 KD2 Tumor cells Control KD ITGBL1 KD1 ITGBL1 KD2 MG-63 BEAS-2B

T47D T47D Vector ITGBL1 Vector ITGBL1-1 ITGBL1-2 Vector ITGBL1 Vector ITGBL1-1 ITGBL1-2

20 T47D-Vector 50 231-Control KD 250 231-Vector 30 T47D-Vector 125 231-Control KD 40 231-Vector T47D-ITGBL1 231-ITGBL1 KD1 231-ITGBL1-1 T47D-ITGBL1 231-ITGBL1 KD1 231-ITGBL1-1 100 15 *** 50 231-ITGBL1 KD2 200 231-ITGBL1-2 231-ITGBL1 KD2 30 231-ITGBL1-2 20 *** *** 30 150 *** 75 10 *** 20 50 20 100 10

Cell count 5 Cell count 10 10 *** *** 50 *** 25 *** *** 0 0 0 0 0 0

C MDA-MB-231 D MDA-MB-231 Tumor cells Tumor cells Control KD ITGBL1 KD1 ITGBL1 KD2 Control KD ITGBL1 KD1 ITGBL1 KD2 Soft agar with Soft agar with MG-63 CM BEAS-2B CM T47D T47D Vector ITGBL1 Vector ITGBL1-1 ITGBL1-2 Vector ITGBL1 Vector ITGBL1-1 ITGBL1-2

15 T47D-Vector 20 231-Control KD 40 231-Vector T47D-Vector 60 231-Control KD 231-Vector T47D-ITGBL1 231-ITGBL1 KD1 231-ITGBL1-1 30 T47D-ITGBL1 231-ITGBL1 KD1 30 231-ITGBL1-1 231-ITGBL1 KD2 231-ITGBL1-2 231-ITGBL1 KD2 231-ITGBL1-2 15 30 10 * *** *** * * *** 20 40 ** 20 ** 10 20 ** 5 10 ** 20 10 5 10 No. of colonies No. of No. of colonies No. of 0 0 0 0 0 0 )

E 4 8 231-Control KD 10 231-Vector *** G MDA-MB-231 T47D 231-ITGBL1 KD1 231-ITGBL1-1 *** 8 6 231-ITGBL1 KD2 231-ITGBL1-2 6 4 *** 4 ITGBL1-2 Vector Vector ITGBL1-1 ITGBL1 ITGBL1 KD2 Control KD 2 ITGBL1 KD1

In MG-63 CM *** 2 b-Actin Cell counts (x10 0 0 01234 5 01234 5 Endogenous ITGBL1 Time (day) Time (day) Exogenous ITGBL1 F 231-Control KD 231-Vector

) 12 231-ITGBL1 KD1 8 CD44 4 231-ITGBL1-1 231-ITGBL1 KD2*** 231-ITGBL1-2 9 *** 6 CDH11 ** 6 4 *** ITGA5

3 2 OPN

In BEAS-2B In BEAS-2B CM BSP Cell counts (x10 0 0 012345 012345 OSN Time (day) Time (day)

Figure 3. ITGBL1 promotes the attraction of breast cancer cells to osteoblasts and their survival in osteoblast CM. Two ITGBL1 knockdown MDA-MB-231 subclones (ITGBL1 KD1 and ITGBL1 KD2) and two ITGBL1-overexpressing MDA-MB-231 subclones (ITGBL1-1 and ITGBL1-2) were established by stably transfecting MDA-MB-231 cells with shITGBL1 and pEGFP-ITGBL1 plasmids; T47D cells were transfected with pEGFP-ITGBL1 plasmid (ITGBL1). A and B, Transwell assay for chemotactic migration of cancer cells toward osteoblast-like MG-63 cells (A) or lung epithelial BEAS-2B cells (B). C and D, soft agar colony formation assay of cancer cells in MG-63 CM (C) or BEAS-2B CM (D). E and F, growth curves of cancer cells in MG-63 CM (E) or BEAS-2B CM (F). G, ITGBL1 and proteins related to bone metastasis were detected by Western blot analysis. These assays were performed in triplicate; , P < 0.05; , P < 0.01; , P < 0.001.

www.aacrjournals.org Cancer Res; 75(16) August 15, 2015 3307

Li et al.

A B 231-Control KD MDA-MB-231 CM 1.5 231-ITGBL1 KD1 Control KD ITGBL1 KD cells 1.0 ** + Tumor cells CM 0.5 Diameter of TRAP BM-derived OC 0 T47D CM 4 Vector Vector ITGBL1 Vector ITGBL1 ITGBL1

cells ***

+ 2 *** Diameter of TRAP 0 T47D 231

C Tumor cells CM D E F + RAW264.7 30 231-Control KD 15 231-Control KD 2.0 231-Control KD + 231-ITGBL1 KD1 231-ITGBL1 KD1 231-ITGBL1 KD1 MDA-MB-231 CM cells 1.5 20 + 10 Control ITGBL1 KD1 1.0 10 5 *** *** 0.5 No. of TRAP No. of ** * multinuclear cells 0 0 multinuclear cells 0 Diameter of TRAP Diameter of

No. of TRAP No. of 3–5 >105–10 Number of nuclei + 40 231-Vector 20 231-Vector 4 231-Vector + 231-ITGBL1-1 231-ITGBL1-1 231-ITGBL1-1

Vector ITGBL1-1 cells 30 15 3 ** ** + 20 2 10 *** ** 10 5 1 No. of TRAP No. of 0 multinuclear cells 0 0 multinuclear cells Diameter of TRAP Diameter of

No. of TRAP No. of >105–103–5 Number of nuclei

G H MDA-MB-231 T47D 1.5 231-Control KD 231-ITGBL1 KD1 231-ITGBL1 KD2 1.0 ITGBL1-1 Vector ITGBL1-2 Vector ITGBL1 KD1 Control KD ITGBL1 KD2 ITGBL1 0.5 β-Actin

RANKL 0 Relative secretion level GM-CSF GRO IL1a MCP-1 MCP-3 MSP-a OPG Cytokines

Figure 4. ITGBL1 contributes to osteoclastogenesis in the bone microenvironment. A and B, primary preosteoclasts isolated from bone marrow cells of mice were preinduced þ by 50 ng/mL M-CSF for 2 days, followed by culture with the CM of cancer cells containing 50 ng/mL RANKL for additional 5 days. The diameters of TRAP multinuclear cells (B) were determined in six fields of two independent experiments. C–F, osteoclast precursor RAW 264.7 cells were induced for the first 4 days with þ 50 ng/mL RANKL and then cultured with the CM of MDA-MB-231 transfectant cells containing 50 ng/mL RANKL for another 3 days. The number of TRAP multinuclear cells (D), the number of TRAPþ cells with different numbers of nuclei (E), and the diameters of TRAPþ multinuclear cells (F) were determined in six fields of two independent experiments. G, levels of macrophage-activating cytokines secreted by ITGBL1 KD cells and control cells were detected using a human cytokine antibody array. H, proteins related to osteoclastogenesis in MDA-MB-231 transfectant cells were detected by Western blot analysis. OPG, osteoprotegerin. , P < 0.05; , P < 0.01; , P < 0.001.

increased in the ITGBL1-overexpressing cells. In contrast, the ther indicated that breast cancer cells with high ITGBL1 levels, expression levels of osteoprotegerin, a key inhibitor of osteoclas- after spreading and surviving in the bone microenvironment, togenesis, increased in the ITGBL1 KD cells and decreased in the could stimulate osteoclastogenesis, and thereby contribute to ITGBL1-overexpressing subclones (Fig. 4H). These ﬁndings fur- osteolytic lesions.

3308 Cancer Res; 75(16) August 15, 2015 Cancer Research

ITGBL1 in Breast Cancer Bone Metastasis

A B Smad2 Merged C

Vector+DMSO 3.0 ITGBL1+DMSO TGFb1 ITGBL1+SB Vector Figure 5. 2.0 Blockade of TGFb signaling reduces TGFb2 the function of ITGBL1 in promoting 1.0

the chemotactic migration of cancer TGFb3 Smad Luc. cells toward osteoblasts and 0 osteoclastogenesis. To block the ITGBL1 TGFb pathway, cells were treated with 1.0 mmol/L SB431542; DMSO was used D in the control cells. A, the levels of Vector+DMSO Vector+DMSO ITGBL1+DMSO ITGBL1+SB 75 ITGBL1+DMSO secreted TGFb1, TGFb2, and TGFb3in ITGBL1+SB the CM of ITGBL1-overexpressing cells 50 *** and the vector control cells were detected by dot blot. B, 25 immunoﬂuorescence staining # # #

showing the subcellular localization of Cell counts Smad2 in the above cells. C, the 0 activity of the TGFb–Smad pathway was evaluated by transiently E Vector CM+DMSO transfecting a Smad luciferase 40 Vector CM + DMSO ITGBL1 CM + DMSO ITGBL1 CM + SB ITGBL1 CM+DMSO reporter into the above cells. ITGBL1 CM+SB D, chemotactic migration assay of 30 ** cancer cells toward MG-63 cells. 20 E, osteoclastogenesis derived from RAW 264.7 cells was detected by 10 # # # TRAP staining. F, the expression of No. of TRAP + TRAP No. of 0 bone metastasis-related molecules multinuclear cells was detected by Western blot analysis. OPG, osteoprotegerin. Data, F means SD in six ﬁelds of two Vector CM+DMSO Vector CM+DMSO 4 independent experiments; , P < 0.05; 15 ITGBL1 CM+DMSO ITGBL1 CM+DMSO , P < 0.01; , P < 0.001 compared ITGBL1 CM+SB ITGBL1 CM+SB b-Actin ## P < 3 with the control cells; , 0.01; 10 * ### P < CDH11 , 0.001 compared with the ** ** 2 ITGBL1-overexpressing cells. # # # 5 CD44 1 # # # # # # # # OPG 0 0 multinuclear cells Control CM ITGBL1 CM ITGBL1 CM+SB No. of TRAP cells + TRAP No. of 3–5 5–10 >10 + TRAP Diameter of RANKL Number of nuclei

TGFb mediates the function of ITGBL1 in bone metastasis of RANKL and mRNA transcripts of CDH11, OPN, OSN were breast cancer upregulated in ITGBL1-overexpressed cells, and these changes Considering the positive correlation between the expression were reversed by blocking the TGFb signaling pathway (Fig. 5F, levels of ITGBL1 and TGFB3 (Fig. 1A) and the downregulation of Supplementary Fig. S3A and S3B). These results suggest that the six genes encoding TGFb signaling pathway molecules after TGFb signaling pathway mediates the role of ITGBL1 in breast ITGBL1 KD (Supplementary Table S2), we speculated that TGFb cancer bone metastasis. signaling mediated the ITGBL1-induced promotion of breast cancer bone metastasis. Indeed, we observed increased secretion ITGBL1 is a transcriptional target of Runx2 of TGFb1 and TGFb3 by ITGBL1-overexpressing cells (Fig. 5A). Because Runx2 is a key transcription factor during bone devel- Further immunoﬂuorescent staining showed that Smad2 trans- opment and ITGBL1 is coexpressed with RUNX2 and Runx2- located into the nuclei of ITGBL1-overexpressing cells (Fig. 5B), regulated bone remodeling-related genes in primary breast cancer and a Smad2 luciferase reporter assay conﬁrmed that Smad2 tissues, we investigated whether ITGBL1 was regulated by Runx2. transcriptional activity was elevated by ITGBL1 overexpression Indeed, through a sequence search of the ITGBL1 promoter, we (Fig. 5C), suggesting that the TGFb/Smad signaling pathway was found three OSE2 elements in the proximal promoter region of activated in ITGBL1-overexpressing cells. Then, we used the ITGBL1: 793 to 786, 1208 to 1201 and 1320 to 1295 TGFb receptor inhibitor SB-431542 to block the TGFb/Smad (Fig. 6A). We performed a ChIP assay to enrich the Runx2-bound signaling pathway (Fig. 5C). Consistent with TGFb/Smad sig- DNA fragments and found that Runx2 bound the ITGBL1 naling pathway activity, the chemotactic responses of MDA- promoter region containing the 793/786 site but not the MB-231 cells toward MG-63 cells (Fig. 5D) and the stimulation 1320/1295 or 1208/1201 site (Fig. 6B). of osteoclastogenesis (Fig. 5E) were increased by ITGBL1 over- We further investigated the regulatory activity of Runx2 on the expression and inhibited by SB-431542 treatment. Molecularly, ITGBL1 promoter using dual-luciferase reporter assays and found thebonemetastasis–related proteins CDH11, CD44, and that ITGBL1 promoter activity decreased in the MDA-MB-231 cells

www.aacrjournals.org Cancer Res; 75(16) August 15, 2015 3309

Li et al.

A B ITGBL1 promoter

–1320/–1295 –793/–786 ITGBL1 –1003/–695 –1208/–1201 +1 –1387/–1153 Putative Runx2 binding site

C E 1.5 Control –825 Runx2 KD –825/+36 luc+ 1.0 –740 *** –740/+36 luc+

*** 0.5 –825 mu

-825/+36 mu luc+ level mRNA ** 0 ** RUNX2 ITGBL1 Mutated Runx2 binding sites 00.51.01.5 ITGBL1 promoter activity Genes

D ** 40 Vector Control KD luc+ Runx2 ** 30 Runx2 KD luc+ 2 * Vector luc+ ** 1

Runx2 luc+ level mRNA 0 0 0.5 1.0 1.5 2 RUNX2 ITGBL1 ITGBL1 promoter activity Genes

Figure 6. ITGBL1 is a target gene of the bone-speciﬁc transcription factor Runx2. A, schematic showing the potential Runx2-binding sites in the ITGBL1 promoter. B, recruitment of Runx2 to the promoter of ITGBL1 in Runx2-His–overexpressing MDA-MB-231 cells was detected by ChIP using anti-His tag antibody. C, ITGBL1 promoter activities with/without the wild-type Runx2-binding motif and with a mutated Runx2-binding motif were evaluated using a dual-luciferase reporter assay. D, ITGBL1 promoter activities in Runx2 KD cells and Runx2-overexpressing cells were compared with their control cells. E, the mRNA expression levels of RUNX2 and ITGBL1 in Runx2 KD cells and Runx2-overexpressing cells compared with their respective control cells. Assays were performed in duplicate; , P < 0.05; , P < 0.01; , P < 0.001.

transfected with pGL3-ITGBL1 740/þ36 and in cells transfected ITGBL1 mediates Runx2-driven bone metastasis of breast with pGL3-ITGBL1 825/þ36 with a mutated Runx2-binding cancer cells motif compared with the control cells transfected with pGL3- Once ITGBL1 was confirmed as a Runx2 target gene that plays ITGBL1 825/þ36 with the wild-type Runx2-binding motif pivotal roles in breast cancer bone metastasis, we further inves- (Fig. 6C). Moreover, we also performed dual-luciferase reporter tigated whether ITGBL1 mediates the Runx2-driven bone metas- assays to compare the ITGBL1 promoter activity upon the tran- tasis of breast cancer cells. The chemotactic migration assay sient transfection of pGL3-ITGBL1 825/þ36 in a stable subclone showed that Runx2 overexpression significantly increased the of MDA-MB-231 with that in shRNA-mediated Runx2 KD, Runx2- chemotactic migration of MDA-MB-231 cells toward MG-63 cells. overexpressing cells, and their control cells. Runx2 KD weakened ITGBL1 KD in Runx2-overexpressing cells reversed the Runx2- and Runx2 overexpression enhanced the activity of the ITGBL1 induced chemotactic migration phenotype (Fig. 7A). We also promoter compared with their respective control cells (Fig. 6D). observed that the chemotactic migration of MDA-MB-231 cells Together, these results demonstrate that Runx2 regulates the toward MG-63 cells was significantly reduced by Runx2 KD and activity of the ITGBL1 promoter by binding to the OSE2 element was rescued by the accompanying ITGBL1 overexpression at 793 to 786. (Fig. 7A). The culture of preosteoclast RAW 264.7 cells with CM þ To further confirmtheregulatoryeffectofRunx2onITGBL1 from Runx2-overexpressing cells induced more and larger TRAP transcription levels, the mRNA expression levels of ITGBL1 in osteoclasts with more nuclei than did control medium (Fig. 7B). Runx2 KD cells, Runx2-overexpressing cells, and their corre- Moreover, osteoclast differentiation induced by the Runx2-over- sponding control cells were measured by RT-qPCR. The ITGBL1 expressing cells was attenuated by ITGBL1 KD (Fig. 7B). Consis- mRNA expression level was downregulated in Runx2 KD cells tently, the CM of Runx2 KD cancer cells reduced osteoclast and upregulated in Runx2-overexpressing cells compared with differentiation in the primary osteoclast culture system, and their respective control cells (Fig. 6E). These results confirm that ITGBL1 overexpression recovered the capability of Runx2 KD ITGBL1 is a positively regulated transcriptional target of Runx2. cancer cells to activate osteoclasts (Fig. 7C). These phenotypes

3310 Cancer Res; 75(16) August 15, 2015 Cancer Research

ITGBL1 in Breast Cancer Bone Metastasis

A Control C Control CM Runx2 + Runx2 + 120 Runx2+control KD Runx2 KD + Runx2 KD + 3 Runx2 KD+vector Runx2+ITGBL1 KD Control control KD ITGBL1 KD Control CM vector CM ITGBL1 CM Runx2 KD+ITGBL1 80 *** 2 # #

40 # # # 1 ** Cell counts Diameter of TRAP + cells + TRAP 0 0 Control Runx2 KD+vector Runx2 KD + Runx2 KD + 90 Runx2 KD+ITGBL1 Control vector ITGBL1 # # # D 60

ontrol ontrol Runx2+controlRunx2+ITGBL1 KD KD 30 *** C C Runx2 KD+vectorRunx2 KD+ITGBL1 b-Actin Cell counts 0 RANKL

BSP

B Control CM Control CM Control CM Runx2 + Runx2 + 30 Runx2+control KD CM 15 ** Runx2+control CM 4 Runx2+control KD Runx2+ITGBL1 KD Control CM control KD CM ITGBL1 KD CM Runx2+ITGBL1 KD CM Runx2+ITGBL1 CM 3 20 * 10 # # * 2 ## * 10 5 # # 1 # #

No. of TRAP+ TRAP+ No. of # multinuclear cells multinuclear cells 0 0 0 Diameter of TRAP+ TRAP+ Diameter of No. of TRAP+ TRAP+ cells No. of 3–5 5–10 >10 Number of nuclei

Figure 7. ITGBL1 mediates Runx2-driven bone metastasis. MDA-MB-231 cells were transfected with the pcDNA vector combined with pSilencer-control (Control), pcDNA-RUNX2 mixed with pSilencer-control (Runx2 þ Control KD), pcDNA-RUNX2 and pSilencer-ITGBL1 (Runx2 þ ITGBL1 KD), pSilencer-Runx2 with pcDNA vector (Runx2 KD þ vector), and pSilencer-Runx2 combined with pcDNA-ITGBL1 (Runx2 KD þ ITGBL1). A, chemotactic migration assay of cancer cells toward MG-63 cells. B, osteoclastogenesis derived from RAW 264.7 cells was detected by TRAP staining. C, osteoclastogenesis derived from primary preosteclasts was detected by TRAP staining. D, markers for bone metastasis were detected by Western blot analysis. Data, means SD in six ﬁelds of two independent experiments; , P < 0.05; , P < 0.01; , P < 0.001 compared with the control cells; #, P < 0.05; ##, P < 0.01; ###, P < 0.001 compared with the Runx2-overexpressing/Runx2 KD cells.

were also confirmed by the expression of mRNA markers and tropism of breast cancer cells. The chemokine (C-X-C motif) protein markers related to bone metastasis and osteoclastogenesis ligand 12 (CXCL12/SDF-1a) and CXCR4 axis is a well-known as detected by Western blot analysis (Fig. 7D and Supplementary chemotactic mechanism for breast cancer cell metastasis to bone Fig. S3C–S3E). Taken together, these results demonstrate that (21, 25). CXCL12 is expressed at high levels by osteoblasts and ITGBL1 mediates the role of Runx2 in promoting the capability of bone marrow stromal cells and recruits CXCR4-positive cancer breast cancer cells to form bone metastasis. cells to bone marrow (26). Integrin av on the surface of bone metastatic breast cancer cells binds to RGD domain-containing Discussion bone matrix proteins such as OPN, BSP, COLs, thus leading to breast cancer cell adhesion to bone (27–29). CD44, a transmem- The present study provides clinical and experimental evidence brane glycoprotein, is a major adhesion molecule for the extra- for a role of ITGBL1 in conferring a bone-specific metastasis cellular matrix (ECM) that binds primarily to the extracellular capability to breast cancer cells. ITGBL1 is highly expressed in glycosaminoglycan hyaluronan (HA). HA is a major non-protein bone metastatic tumor cells, as well as in primary cancer cells with glycosaminoglycan component of the ECM in human bone bone metastasis capabilities. ITGBL1 facilitates the acquisition of matrix, and the interaction of CD44 and HA facilitates tumor tumor cell advantages in recruiting, residing, and growth in bone cell arrest and colonization in bone, leading to increased bone and further stimulates osteoclast maturation in the bone micrometastasis (24, 30). environment to form bone metastatic lesions. We also identified a Metastatic cancer cells that home and localize to the bone fundamental mechanism of ITGBL1 in bone metastasis via TGFb marrow can remain dormant (31, 32) until they are activated to pathway activation. Moreover, ITGBL1, as a transcriptional target colonize and grow into visible metastatic bone lesions. In this of Runx2, mediates the role of Runx2 in promoting breast cancer study, we found that ITGBL1 was highly expressed in clinical bone bone metastasis. metastatic tumors and that ITGBL1 enhanced the colony forma- Tropism and adhesion are crucial steps in the communication tion and proliferation capability of cancer cells in a bone-mim- and interaction between breast cancer cells and the bone matrix or icking microenvironment, emphasizing its roles in the potential host cells in bone. Both clinical data and animal experiments reprogramming to activate quiescent cancer cells and survival demonstrated the role of ITGBL1 in organ selectivity to the bone advantage in bone microenvironment. One explanation for this during breast cancer metastasis. Furthermore, our evidence that survival advantage is the osteomimetic aspect of cancer cells. ITGBL1 regulates the expression of chemokine receptor 4 Osteomimetic properties of breast cancer cells include the ectopic (CXCR4), integrin av, and CD44 on the surface of breast cancer expression of bone matrix proteins and osteoblast-specific adhe- cells supports the function of ITGBL1 in promoting the bone sion molecules, and even mineralization under appropriate

www.aacrjournals.org Cancer Res; 75(16) August 15, 2015 3311

Li et al.

conditions (16, 17). These osteoblast-like features allow cancer Runx2 is regarded as a pivotal transcription factor for the cells to reside in bone due to their escape from immune surveil- formation of the osteomimetic phenotype (39, 40) and contri- lance in the bone microenvironment (6, 16, 17). Combining the butes to cancer cell dissemination into the blood, survival in the report that ITGBL1 was identified as a gene in the osteoblast-like bone microenvironment, and stimulation of bone resorption (41, gene-expression signature (6) with our finding that ITGBL1 42). Expression of Runx2 target genes, including OPN, BSP, and increased the expression of osteomimetic markers, we propose osteocalcin, appears to confer osteomimetic features upon breast that ITGBL1 contributes to osteomimetic breast cancer, and cancer cells, and is reported to play important roles in breast therefore facilitates cancer cells preferentially residing in and cancer bone metastasis (39). In this study, we identified ITGBL1 as colonizing the bone microenvironment. Moreover, ITGBL1 pos- a novel Runx2 target gene and demonstrated that ITGBL1 exhibits itively regulated the cell proliferation-related genes KI67 and properties similar to Runx2 in promoting the formation of osteo- CCNE, suggesting that ITGBL1 can also confer a proliferation mimetic breast cancer and mediates the Runx2-induced conferral advantage to tumor cells in the bone microenvironment. of bone metastasis capability on cancer cells. Most breast cancer metastases in bone cause osteolytic lesions. In conclusion, our study revealed a new molecular mechanism Continuous expansion of osteolytic bone metastasis is driven by that promotes breast cancer metastasis to the bone. We also the "vicious cycle" of tumor-dependent activation of bone- identified an activator of the TGFb signaling pathway and a key degrading osteoclasts and bone stroma-dependent stimulation mediator of Runx2 in bone metastasis. Importantly, we provided of tumor malignancy (22, 31). Osteoclasts are derived from preclinical evidence for ITGBL1 as a molecular marker to predict precursors in the mononuclear-phagocyte lineage and are respon- bone metastasis risk and as a therapeutic target against breast sible for bone resorption. RANKL, primarily produced by osteo- cancer bone metastasis. blasts and cancer cells, binds to its cognate receptor RANK on osteoclast precursors and plays a critical role in promoting oste- Disclosure of Potential Conflicts of Interest oclast differentiation and activation, leading to bone resorption No potential conflicts of interest were disclosed. (33). Osteoprotegerin is a soluble decoy receptor for RANKL that blocks osteoclast formation by inhibiting the binding of RANKL Authors' Contributions to RANK (33). Indeed, we observed that high ITGBL1 expression Conception and design: X.-Q. Li, Y.-M. Feng in cancer cells promoted osteoclast activation in vitro and osteo- Development of methodology: X.-Q. Li, D.-M. Li Acquisition of data (provided animals, acquired and managed patients, lytic lesions in vivo. Consistent with that phenotype, ITGBL1- provided facilities, etc.): X.-Q. Li, D.-M. Li, P.-Z. Kong, Y. Sun, P.-F. Liu, overexpressing cancer cells showed elevated levels of RANKL and Q.-S. Wang decreased expression of osteoprotegerin. Moreover, ITGBL1 pos- Analysis and interpretation of data (e.g., statistical analysis, biostatistics, itively regulated the secretion of the osteoclast differentiation computational analysis): X.-Q. Li, X. Du, Y. Sun, P.-F. Liu activators GM-CSF (34), GRO (35), IL1a (36), MCPs (37), and Writing, review, and/or revision of the manuscript: X.-Q. Li, Y.-M. Feng MSP-a (38) in cancer cells, providing additional evidence sup- Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): P.-Z. Kong, Y.-M. Feng porting the role of ITGBL1 in cancer cell–triggered osteolytic Study supervision: Y.-M. Feng lesions in bone metastatic events. b The role of the TGF signaling pathway in bone metastasis has Grant Support b been clearly outlined. TGF s promote the formation of osteomi- This work was supported by the National Natural Science Foundation of metic breast cancer by inducing the expression of bone matrix China (nos. 30872518, 81201647, 81272357, and 81472680), the Major proteins (17) and also contribute to the "vicious cycle" and Program of Applied Basic Research Projects of Tianjin (nos. 09JCZDJC19800 osteoclastic resorption (22). In this study, we found that ITGBL1 and 13JCZDJC30100), and the Specialized Research Fund for the Doctoral increased the expression and secretion of TGFb1 and TGFb3in Program of Higher Education of the Ministry of Education of China (no. breast cancer cells. Blockade of the TGFb pathway by SB431542 20121202120011). The costs of publication of this article were defrayed in part by the payment of weakened the ITGBL1-induced bone metastatic potential of breast page charges. This article must therefore be hereby marked advertisement in cancer cells, suggesting a modulatory role for the TGFb pathway in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ITGBL1-driven bone-specific breast cancer metastasis. However, the detailed mechanism of TGFb upregulation and TGFb pathway Received January 23, 2015; revised June 4, 2015; accepted June 4, 2015; activation by ITGBL1 need to be further explored in future studies. published OnlineFirst June 9, 2015.

References 1. Sethi N, Dai X, Winter CG, Kang Y. Tumor-derived JAGGED1 promotes 5. Feng Y, Li X, Sun B, Wang Y, Zhang L, Pan X, et al. Evidence for a osteolytic bone metastasis of breast cancer by engaging notch signaling in transcriptional signature of breast cancer. Breast Cancer Res Treat bone cells. Cancer Cell 2011;19:192–205. 2010;122:65–75. 2. Fradet A, Sorel H, Bouazza L, Goehrig D, Depalle B, Bellahcene A, et al. Dual 6. Garcia T, Jackson A, Bachelier R, Clement-Lacroix P, Baron R, Clezardin P, function of ERRalpha in breast cancer and bone metastasis formation: et al. A convenient clinically relevant model of human breast cancer bone implication of VEGF and osteoprotegerin. Cancer Res 2011;71:5728–38. metastasis. Clin Exp Metastasis 2008;25:33–42. 3. Berg RW, Leung E, Gough S, Morris C, Yao WP, Wang SX, et al. Cloning and 7. Franceschi RT, Xiao G. Regulation of the osteoblast-speciﬁc transcription characterization of a novel beta integrin-related cDNA coding for the factor, Runx2: responsiveness to multiple signal transduction pathways. protein TIED ("ten beta integrin EGF-like repeat domains") that maps to J Cell Biochem 2003;88:446–54. chromosome band 13q33: a divergent stand-alone integrin stalk structure. 8. Martin JW, Zielenska M, Stein GS, van Wijnen AJ, Squire JA. The role of Genomics 1999;56:169–78. RUNX2 in osteosarcoma oncogenesis. Sarcoma 2011;2011:282745. 4. Takagi J, Beglova N, Yalamanchili P, Blacklow SC, Springer TA. Deﬁnition 9. Barnes GL, Hebert KE, Kamal M, Javed A, Einhorn TA, Lian JB, et al. Fidelity of EGF-like, closely interacting modules that bear activation epitopes in of Runx2 activity in breast cancer cells is required for the generation of integrin beta subunits. Proc Natl Acad Sci U S A 2001;98:11175–80. metastases-associated osteolytic disease. Cancer Res 2004;64:4506–13.

3312 Cancer Res; 75(16) August 15, 2015 Cancer Research

ITGBL1 in Breast Cancer Bone Metastasis

10. Pratap J, Wixted JJ, Gaur T, Zaidi SK, Dobson J, Gokul KD, et al. Runx2 27. Zhao Y, Bachelier R, Treilleux I, Pujuguet P, Peyruchaud O, Baron R, et al. transcriptional activation of Indian Hedgehog and a downstream Tumor alphavbeta3 integrin is a therapeutic target for breast cancer bone bone metastatic pathway in breast cancer cells. Cancer Res 2008;68: metastases. Cancer Res 2007;67:5821–30. 7795–802. 28. Sloan EK, Pouliot N, Stanley KL, Chia J, Moseley JM, Hards DK, et al. 11. Bauerle T, Adwan H, Kiessling F, Hilbig H, Armbruster FP, Berger MR. Tumor-specific expression of alphavbeta3 integrin promotes spontaneous Characterization of a rat model with site-specific bone metastasis metastasis of breast cancer to bone. Breast Cancer Res 2006;8:R20. induced by MDA-MB-231 breast cancer cells and its application to the 29. Korah R, Boots M, Wieder R. Integrin alpha5beta1 promotes survival of effects of an antibody against bone sialoprotein. Int J Cancer 2005; growth-arrested breast cancer cells: an in vitro paradigm for breast cancer 115:177–86. dormancy in bone marrow. Cancer Res 2004;64:4514–22. 12. Pratap J, Lian JB, Javed A, Barnes GL, van Wijnen AJ, Stein JL, et al. 30. Nakamura H, Hiraga T, Ninomiya T, Hosoya A, Fujisaki N, Yoneda T, et al. Regulatory roles of Runx2 in metastatic tumor and cancer cell interactions Involvement of cell-cell and cell-matrix interactions in bone destruction with bone. Cancer Metastasis Rev 2006;25:589–600. induced by metastatic MDA-MB-231 human breast cancer cells in nude 13. Ports MO, Nagle RB, Pond GD, Cress AE. Extracellular engagement of mice. J Bone Miner Metab 2008;26:642–7. alpha6 integrin inhibited urokinase-type plasminogen activator-mediated 31. Lu X, Mu E, Wei Y, Riethdorf S, Yang Q, Yuan M, et al. VCAM-1 promotes cleavage and delayed human prostate bone metastasis. Cancer Res 2009; osteolytic expansion of indolent bone micrometastasis of breast cancer by 69:5007–14. engaging alpha4beta1-positive osteoclast progenitors. Cancer Cell 14. Yuan J, Ossendorf C, Szatkowski JP, Bronk JT, Maran A, Yaszemski M, et al. 2011;20:701–14. Osteoblastic and osteolytic human osteosarcomas can be studied with 32. Aguirre-Ghiso JA. Models, mechanisms and clinical evidence for cancer a new xenograft mouse model producing spontaneous metastases. dormancy. Nat Rev Cancer 2007;7:834–46. Cancer Invest 2009;27:435–42. 33. Wittrant Y, Theoleyre S, Chipoy C, Padrines M, Blanchard F, Heymann D, 15. Ringner M, Fredlund E, Hakkinen J, Borg A, Staaf J. GOBO: gene expression- et al. RANKL/RANK/OPG: new therapeutic targets in bone tumours and based outcome for breast cancer online. PLoS ONE 2011;6:e17911. associated osteolysis. Biochim Biophys Acta 2004;1704:49–57. 16. Bellahcene A, Bachelier R, Detry C, Lidereau R, Clezardin P, Castronovo V. 34. Atanga E, Dolder S, Dauwalder T, Wetterwald A, Hofstetter W. TNFalpha Transcriptome analysis reveals an osteoblast-like phenotype for human inhibits the development of osteoclasts through osteoblast-derived GM- osteotropic breast cancer cells. Breast Cancer Res Treat 2007;101:135–48. CSF. Bone 2011;49:1090–100. 17. Rucci N, Teti A. Osteomimicry: how tumor cells try to deceive the bone. 35. Boucharaba A, Guillet B, Menaa F, Hneino M, van Wijnen AJ, Clezardin P, Front Biosci 2010;2:907–15. et al. Bioactive lipids lysophosphatidic acid and sphingosine 1-phosphate 18. Wang Y, Klijn JG, Zhang Y, Sieuwerts AM, Look MP, Yang F, et al. Gene- mediate breast cancer cell biological functions through distinct mechan- expression profiles to predict distant metastasis of lymph-node-negative isms. Oncol Res 2009;18:173–84. primary breast cancer. Lancet 2005;365:671–9. 36. Bajayo A, Bar A, Denes A, Bachar M, Kram V, Attar-Namdar M, et al. Skeletal 19. Minn AJ, Gupta GP, Siegel PM, Bos PD, Shu W, Giri DD, et al. Genes that parasympathetic innervation communicates central IL-1 signals regulating mediate breast cancer metastasis to lung. Nature 2005;436:518–24. bone mass accrual. Proc Natl Acad Sci U S A 2012;109:15455–60. 20. Zhang XH, Wang Q, Gerald W, Hudis CA, Norton L, Smid M, et al. Latent 37. Morrison NA, Day CJ, Nicholson GC. Dominant negative MCP-1 blocks bone metastasis in breast cancer tied to Src-dependent survival signals. human osteoclast differentiation. J Cell Biochem 2014;115:303–12. Cancer Cell 2009;16:67–78. 38. Kretschmann KL, Eyob H, Buys SS, Welm AL. The macrophage stimulating 21. Kang Y, Siegel PM, Shu W, Drobnjak M, Kakonen SM, Cordon-Cardo C, protein/Ron pathway as a potential therapeutic target to impede multiple et al. A multigenic program mediating breast cancer metastasis to bone. mechanisms involved in breast cancer progression. Curr Drug Targets Cancer Cell 2003;3:537–49. 2010;11:1157–68. 22. Weilbaecher KN, Guise TA, McCauley LK. Cancer to bone: a fatal attraction. 39. Shore P. A role for Runx2 in normal mammary gland and breast cancer Nat Rev Cancer 2011;11:411–25. bone metastasis. J Cell Biochem 2005;96:484–9. 23. Gassmann P, Enns A, Haier J. Role of tumor cell adhesion and migration in 40. Clezardin P, Teti A. Bone metastasis: pathogenesis and therapeutic impli- organ-specific metastasis formation. Onkologie 2004;27:577–82. cations. Clin Exp Metastasis 2007;24:599–608. 24. Hiraga T, Ito S, Nakamura H. Cancer stem-like cell marker CD44 promotes 41. Leong DT, Lim J, Goh X, Pratap J, Pereira BP, Kwok HS, et al. Cancer-related bone metastases by enhancing tumorigenicity, cell motility, and hyalur- ectopic expression of the bone-related transcription factor RUNX2 in non- onan production. Cancer Res 2013;73:4112–22. osseous metastatic tumor cells is linked to cell proliferation and motility. 25. Hirbe AC, Morgan EA, Weilbaecher KN. The CXCR4/SDF-1 chemokine Breast Cancer Res 2010;12:R89. axis: a potential therapeutic target for bone metastases? Curr Pharm Des 42. Javed A, Barnes GL, Pratap J, Antkowiak T, Gerstenfeld LC, van Wijnen AJ, 2010;16:1284–90. et al. Impaired intranuclear trafficking of Runx2 (AML3/CBFA1) transcrip- 26. Olechnowicz SW, Edwards CM. Contributions of the host microenviron- tion factors in breast cancer cells inhibits osteolysis in vivo. Proc Natl Acad ment to cancer-induced bone disease. Cancer Res 2014;74:1625–31. Sci U S A 2005;102:1454–9.

www.aacrjournals.org Cancer Res; 75(16) August 15, 2015 3313

ITGBL1 Is a Runx2 Transcriptional Target and Promotes Breast Cancer Bone Metastasis by Activating the TGF β Signaling Pathway

Xiao-Qing Li, Xin Du, Dong-Mei Li, et al.

Cancer Res 2015;75:3302-3313. Published OnlineFirst June 9, 2015.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-15-0240

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2015/06/09/0008-5472.CAN-15-0240.DC1

Cited articles This article cites 42 articles, 10 of which you can access for free at: http://cancerres.aacrjournals.org/content/75/16/3302.full#ref-list-1

Citing articles This article has been cited by 5 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/75/16/3302.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 Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/75/16/3302. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.