ALCAM/CD166 Is a TGF-B–Responsive Marker and Functional Regulator of Prostate Cancer Metastasis to Bone

Published OnlineFirst January 2, 2014; DOI: 10.1158/0008-5472.CAN-13-1296

Cancer Molecular and Cellular Pathobiology Research

ALCAM/CD166 Is a TGF-b–Responsive Marker and Functional Regulator of Prostate Cancer Metastasis to Bone

Amanda G. Hansen1,2, Shanna A. Arnold1, Ming Jiang3, Trenis D. Palmer1, Tatiana Ketova1, Alyssa Merkel5,6, Michael Pickup2, Susan Samaras1, Yu Shyr4, Harold L. Moses2, Simon W. Hayward2,3, Julie A. Sterling2,5,6, and Andries Zijlstra1,2

Abstract The dissemination of prostate cancer to bone is a common, incurable aspect of advanced disease. Prevention and treatment of this terminal phase of prostate cancer requires improved molecular understanding of the process as well as markers indicative of molecular progression. Through biochemical analyses and loss-of-function in vivo studies, we demonstrate that the cell adhesion molecule, activated leukocyte cell adhesion molecule (ALCAM), is actively shed from metastatic prostate cancer cells by the sheddase ADAM17 in response to TGF-b. Not only is this posttranslational modiﬁcation of ALCAM a marker of prostate cancer progression, the molecule is also required for effective metastasis to bone. Biochemical analysis of prostate cancer cell lines reveals that ALCAM expression and shedding is elevated in response to TGF-b signaling. Both in vitro and in vivo shedding is mediated by ADAM17. Longitudinal analysis of circulating ALCAM in tumor-bearing mice revealed that shedding of tumor, but not host-derived ALCAM is elevated during growth of the cancer. Gene-speciﬁc knockdown of ALCAM in bone-metastatic PC3 cells greatly diminished both skeletal dissemination and tumor growth in bone. The reduced growth of ALCAM knockdown cells corresponded to an increase in apoptosis (caspase-3) and decreased proliferation (Ki67). Together, these data demonstrate that the ALCAM is both a functional regulator as well as marker of prostate cancer progression. Cancer Res; 74(5); 1404–15. 2014 AACR.

Introduction surveillance, which avoids the morbidity associated with sur- Morbidity and mortality among patients with prostate gery. However, without a better understanding of the mechan- cancer are frequently a result of metastatic dissemination to isms that drive progression and biomarkers to predict the risk bone. To date, there is no curative treatment of skeletal of progression, many patients and physicians are not comfort- metastasis and survival from the time of diagnosis is merely able pursuing this option, and instead opt for more invasive fi 3 to 5 years (1). Moreover, skeletal events are associated with "de nitive" interventions. Thus, clinical intervention would fi high morbidity in the form of bone loss, fractures, and pain (2). greatly bene t from further understanding of the molecular Even for patients with organ-confined disease, the risk of mechanisms that drive skeletal metastasis as well as molecular disease progression drives the rigor of clinical intervention. indicators that identify patients at risk of disease progression. Patients have been shown to do well on a regimen of active Because cell motility is an important contributor to metastasis, the activation state of migratory mechanisms could be suitable for both therapeutic intervention as well as a biomarker of metastatic behavior. Authors' Affiliations: Departments of 1Pathology, Microbiology and 2 3 4 Tumor cell metastasis to distant sites, including bone, is a Immunology, Cancer Biology, and Urologic Surgery, Division of Cancer fi Biostatistics, Vanderbilt University; 5Department of Medicine, Division of multistep process. Cancer cells must rst detach from the Clinical Pharmacology, Vanderbilt Center for Bone Biology; and 6Depart- primary tumor site and migrate locally to invade blood vessels. ment of Veterans Affairs, Tennessee Valley Healthcare System, Nashville, Thereafter, tumor cells intravasate into the bloodstream and Tennessee are attracted to preferred sites of metastasis through site- Note: Supplementary data for this article are available at Cancer Research specific cellular and microenvironmental interactions (3, 4). Online (http://cancerres.aacrjournals.org/). Activated leukocyte cell adhesion molecule (ALCAM) is a cell Current address for M. Jiang: Laboratory of Nuclear Receptors and Cancer adhesion molecule that engages in homotypic and heterotypic Research, Center for Medical Research, Nantong University School of Medicine, 19 Qixiu Road, Nantong City, Jiangsu Province 226001, China. cell adhesion in a calcium-independent manner (5). It has been implicated in a number of adhesive and migratory Corresponding Author: Andries Zijlstra, Vanderbilt University Medical Center, C-2102c MCN, 1161 21st Avenue S., Nashville, TN 37232-2561. behaviors including axonal guidance, leukocyte homing, and Phone: 615-322-3295; Fax: 615-936-7040; E-mail: cancer metastasis. ALCAM can be proteolytically cleaved at [email protected] the cell surface by ADAM17 causing the ectodomain to be doi: 10.1158/0008-5472.CAN-13-1296 shed. This shedding can be induced by ionomycin, phorbol 2014 American Association for Cancer Research. 12-myristate 13-acetate (PMA), and EGFs (6). Proteolytic

1404 Cancer Res; 74(5) March 1, 2014

ALCAM Is a Marker and Regulator of Skeletal Metastasis

cleavage/shedding of cell surface proteins is a common reg- acid (BCA) assay (Bio-Rad). Conditioned media samples were ulatory mechanism that can alter the function and localiza- concentrated with Microcon centrifugal filters (Millipore) tion of transmembrane proteins such as cell adhesion mole- following the manufacturer's protocol, which were eluted cules, growth factors, and growth factor receptors. This directly in 5 sample buffer for Western blot analysis. Protein regulation can weaken cell adhesion and destabilize adherens loading for conditioned medium samples for Western blot junctions as in the case of E-cadherin, through the loss of analysis was adjusted according to the total protein in cell homotypic cell adhesion molecular interactions (7). The sol- lysates. Conditioned media and total protein was subjected uble shed ectodomain can act as a soluble antagonist, thereby to SDS-PAGE and electrophoretic transfer to polyvinylidene competing for membrane-bound receptors or cell adhesion difluoride membranes (Immobilon P, Millipore, Inc.). Immu- molecules. Although these processes are important in devel- nodetection was done by conventional chemiluminescence. opment, they have also been associated with tumorigenesis. Functionally, expression of the truncated, trans-membrane Quantitative PCR fragment of ALCAM in BLM melanoma cells results in The mRNA samples were prepared from tumor cells lysed in increased lung metastasis in vivo, whereas overexpression of TRI Reagent (Ambion) and purified using phenol extraction, a soluble extracellular ligand-binding fragment diminished followed by real-time PCR (RT-PCR). The following quantita- metastases (for review see refs. 8 and 9). Previous studies from tive PCR (qPCR) primers were used ALCAM, TCAAGGTGTT- our laboratory and others have investigated the clinical rel- CAAGCAACCA (forward) and CTGAAATGCAGTCACCCAAC evance of ALCAM expression and shedding in a variety of (reverse); ADAM17, ATGTTTCACGTTTGCAGTCTCCA (for- human malignancies, including colorectal, breast, ovarian, ward) and CATGTATCTGTAGAAGCGATGATCTG (reverse); thyroid, and prostate cancer (10–16). However, its molecular and glyceraldehyde-3-phosphate dehydrogenase (GAPDH), contribution to disease progression remains unclear. ATCTTCTTTTGCGTCGCCAG (forward) and TTCCCCATGG- Elevated ALCAM expression in aggressive prostate cancer TGTCTGAGC (reverse). (17) together with its putative role in cell adhesion/migration (8, 18) suggested that ALCAM is a molecular participant in shRNA knockdown prostate cancer metastasis to bone. On the basis of our own To establish cell lines in which ALCAM expression is stably work in colorectal cancer (10) and concomitant work in breast knocked down, cells were transduced with ALCAM-specific (19) as well as ovarian cancer (11), we hypothesized that Mission short hairpin RNA (shRNA; Sigma) lentivirus. Follow- ectodomain shedding of ALCAM is likely to correspond with ing transduction, cells were selected in 10 mg/mL of puromycin. prostate cancer progression. Furthermore, considering the Transduced cells were flow sorted for ALCAM expression and dominant role of TGF-b in driving skeletal metastasis of ALCAM knockdown cells were cultured and maintained in 5 prostate cancer (20, 21), we postulated that this cytokine could mg/mL of puromycin. influence ALCAM expression and/or shedding. The present study investigates ALCAM expression and shedding during Migration assay prostate cancer progression, evaluates the influence of TGF-b Two-dimensional gap closure assays (formerly known as stimulation, and determines the contribution of ALCAM to scratch assays) were conducted by using magnetically attach- skeletal metastasis in a series of orthotopic, and experimental able stencils attached to culture plates (22). In short, 250,000 metastasis models. cells were seeded in 6-well plates and allowed to recover overnight to form a confluent monolayer. Stencils were Materials and Methods removed with tweezers, after which cells were rinsed with Reagents, cell culture PBS to remove detached cells. Culture medium was re-added Full-length purified recombinant porcine TGF-b was and closure of the gap was measured at 8 and 16 hours. Gap obtained from R&D Systems. Antibodies against ALCAM were closure was quantified using TScratch (NIH, Bethesda, MD). obtained from R&D Systems (Clone 105902). The following tumor cell lines were obtained from American Type Culture Histologic analysis of mouse tissue Collection (ATCC) and maintained according to the ATCC's Tumor-bearing tissue and bones were fixed in 10% formalin. recommendations: DU145, LNCaP, and PC3 (prostate metas- Bone specimens were decalcified in 20% EDTA pH 7.4 for 3 to 4 tasis). PC3-luciferase (PC3-luc) cells (Simon Hayward, Vander- days at room temperature. Decalcified bone and tissue were bilt University, Nashville, TN) were cultured in RPMI/10% FBS. dehydrated and embedded in paraffin. Tumor burden was confirmed in 5-mm serial sections stained with hematoxylin SDS-PAGE and immunoblotting and eosin (H&E). Osteoclasts were visualized using a standard Cells (2.5 105) were plated in 6-well dishes. After 24 hours, tartrate-resistant acid phosphatase (TRAP) protocol. All cells were serum-starved in Opti-MEM for 16 hours and then immunohistochemistry and immunofluorescence on tumor treated in the presence or absence of indicated growth factors sections involved antigen retrieval by using a standard pH 6.0 or inhibitors for 48 hours. After that, the cells were lysed in TNE citrate buffer followed by blocking via incubation with 20% lysis buffer [20 mmol/L Tris–Cl (pH 7.4), 0.5 mmol/L EDTA, 1% Aquablock (East Coast Bio). Immunofluorescence data were Triton X-100, 150 mmol/L NaCl, protease inhibitor cocktail obtained using primary antibodies for ALCAM (1:1,000; Leica (Sigma), 1 mmol/L phenylmethylsulfonyl fluoride]. Total pro- Biosystems; Clone, MOG/07), Ki67 (1:500; Fisher; Clone SP6), tein in the lysates was quantified by using the bicinchoninic cleaved caspase-3 (1:200; Cell Signaling Technology; D175), and

www.aacrjournals.org Cancer Res; 74(5) March 1, 2014 1405

Hansen et al.

collagen I (1:1,000; Sigma; C2206) by incubation overnight at either the presence of EDTA as an anticoagulant or a serum 4C. Corresponding Alexa Fluor secondary antibodies were separator tube (Fisher Scientific). Plasma and serum samples used (1:1,000; Invitrogen). Fluorescent imaging was completed were stored at 80C until analyzed. on an Olympus BX61WI upright fluorescent microscope by using Volocity Imaging Software. Micro computed tomography analysis For gross analysis of trabecular bone volume, formalin-fixed ELISA of mouse serum and plasma tibiae were scanned at an isotropic voxel size of 12 mm by using Blood was obtained via the saphenous vein; samples were a micro computed tomography 40 (mCT40; SCANCO Medical). collected in either the presence of EDTA as an anticoagulant or The tissue volume was derived from generating a contour a serum separator tube (Fisher Scientific; cat#1491559 and around the metaphyseal trabecular bone that excluded the 1491553, respectively) and were centrifuged at 1,500 rpm and cortices. The area of measurement began at least 0.2 mm below 4C to remove cells. Plasma and serum samples were stored at the growth plate and was extended by 0.12 mm. The bone 80C until analyzed. Samples were analyzed for soluble volume included all bone tissue that had a material density mouse and human ALCAM using the R&D Systems DuoSet greater than 438.7 mgHA/cm3. following the manufacturer's instructions. Briefly, ELISA plates coated with capture antibody were incubated overnight with Radiographic analysis 100 mL of sera diluted at 1:50. Capture ALCAM was detected Beginning 1 week after tumor cell inoculation, tumor-bear- with biotinylated antibody and peroxidase-conjugated avidin ing animals were subjected to radiographic imaging. Radio- followed by colorimetric detection at 450 nm. graphic images (Faxitron X-ray Corp.) were obtained by using an energy of 35 kV and an exposure time of 8 seconds. Mouse models of prostate cancer and in vivo Osteolytic lesions were quantified bilaterally in the tibia, fibula, quantitation of tumor growth femora, humeri, and pelvis at the endpoint using X-ray images. All experimental protocols were approved by the Vanderbilt Lesion area and lesion numbers were evaluated by using image University Institutional Animal Care and Use Committee. analysis software (Metamorph; Molecular Devices, Inc.). Data Orthotopic prostate xenografts were performed according to presented are the average of lesion area and lesion numbers per Li and colleagues (23). Briefly, 5 104 PC3-luc cells were mouse in each group. suspended in 30 mL of neutralized type I collagen and allowed to polymerize for 16 hours at 37C before implantation into the Statistical analyses prostate of 10-week-old C.B-17/IcrHsd-Prkdc severe combined Expression analysis was performed on datasets GDS1439 immunodeficient mice (SCID) male mice (Harlan Laborato- and GSE10645 available through the Gene Expression Omnibus ries). Tumor growth was monitored weekly by bioluminescent (refs. 26 and 27, respectively). Expression data for selected detection of luciferase-expressing cells. For the xenograft genes from GDS1439 was clustered in software Cluster 3.0 and model, subconfluent PC3-luc cells were trypsinized, washed visualized with software TreeView. For survival analysis, the n ¼ twice in PBS to remove serum, and then resuspended in Hank's patient population of GSE10645 ( 596) was dichotomized Balanced Salt Solution (GIBCO) at a concentration of 1 107 across upper and lower quartile of ALCAM expression. Statis- cells/mL. Of note, 100 mL containing 1 106 PC3 cells in a 50/50 tics were completed by using either R, SPSS, or GraphPad mix of PBS and growth factor–reduced Matrigel (BD Bio- Prism. For all standard bar and box plots, the results were sciences) were injected subcutaneously into the right flank of reported as mean and SEM, unless stated otherwise in the 7-week-old nude male mice (Harlan Laboratories; athymic legend. Comparisons were performed using unpaired two- t – Foxn1 nu/nu). Tumor growth was monitored weekly by caliper sided Student test, nonparametric Mann Whitney test, or R2 P measurements, and tumor volume was calculated on the basis one-way ANOVA. and values were reported from linear of the following formula: (length length width)/6. PC3- regression analysis of mouse data. All statistical tests were fi P < P < P < luciferase shControl (Vector) or PC3-luciferase shALCAM considered signi cant when 0.05: , 0.05; , 0.01; and P < (KD2 or KD3) tumor cells (1 105)ina10mL volume of , 0.001. sterile PBS were injected into the tibia of anesthetized 6-week- old nude male mice (Harlan Laboratories). Skeletal metastasis Results was performed as previously described by Park and colleagues ALCAM gene expression is elevated in advanced prostate (24) and as visualized in ref. 25. Briefly, 1 105 PC-3-luc cells cancer and correlates with poor patient outcome were injected into the left heart ventricle of male nude mice Changes in ALCAM expression have been linked to patient (Harlan Laboratories). Skeletal metastases were monitored by outcome for several malignancies. In prostate cancer, the bioluminescent detection of luciferase-expressing cells and correlation of ALCAM expression with patient outcome is formation of bone lesion by X-ray. Whole-animal luminescent sometimes conflicting. Minner and colleagues (15) conclude imaging was performed with the IVIS system (Caliper Life that reduced ALCAM expression correlates to poor patient Sciences). Luciferin (150 mg/kg in sterile PBS; Biosynth Inter- outcome, whereas the opposite was suggested by Kristiansen national) was delivered via intraperitoneal injection 10 min- and colleagues (17). We evaluated several publicly available utes before imaging. Living Image software (Caliper Life microarray datasets to determine the relationship between Sciences) was used to quantify the luminescence intensity. ALCAM mRNA levels, patient diagnosis, and outcome (Fig. 1). Blood was obtained via the saphenous vein and collected in ALCAM expression seems to be elevated in an experimental

1406 Cancer Res; 74(5) March 1, 2014 Cancer Research

ALCAM Is a Marker and Regulator of Skeletal Metastasis

Figure 1. ALCAM is overexpressed in metastatic prostate cancer and correlates with patient survival. A, ALCAM expression levels were analyzed in publicly available dataset (GDS1439, n ¼ 19) of prostate cancer. Heatmap (A) and corresponding relative expression (B) of ADAM17 (i), ALCAM (ii), N-cadherin (iii), p120 (iv), and E-cadherin (v) indicated by arrows. C, correlation of ALCAM expression to overall survival in a publicly available dataset (GSE10645) composed of 596 patients with prostate cancer. Kaplan–Meier survival curves represent the upper and lower quartile ALCAM expression. D, representative images from immunohistochemical staining of ALCAM membranous and cytoplasmic expression in benign to metastatic prostate cancer. Images obtained through The Human Protein Atlas. model of epithelial–mesenchymal transition performed by the TGF-b induces ALCAM expression and shedding Weinberg's laboratory [(28) GSE9691; Supplementary Fig. S1]. Because ALCAM is associated with disease progression, we set Indeed, a comparison of benign, localized, and metastatic out to determine its contribution to the skeletal metastasis of prostate cancer revealed that the level of ALCAM mRNA prostate cancer. Moreover, because bone metastasis is driven in increased in metastatic disease (Fig. 1A; GDS1439) and coin- large part by TGF-b (2, 20, 21), we investigated the ability of this cided with molecular evidence of a promigratory phenotype cytokine to promote ALCAM shedding in vitro. ALCAM is based on the decreased expression of E-cadherin and p120 with proteolytically shed from PC3 cells (Supplementary Fig. S2). concurrently elevated expression of N-cadherin (Fig. 1B). These Absence of the cytoplasmic tail in the conditioned medium observations were supported by survival analysis for a cohort conﬁrms that the ectodomain is shed (Supplementary Fig. S2A). of 596 patients with prostate cancer (GSE10645), which Moreover, ALCAM is absent from PC3-derived exosomes, ensur- revealed that the high levels of ALCAM mRNA corresponded ing that the ectodomain is shed and not released with cell- with poor patient outcome (Fig. 1C). Immunohistologic stain- derived microparticles (Supplementary Fig. S2B). Selected cyto- ing of prostate cancer tissue microarrays available through the kines thought to be involved in metastasis to bone, including Human Protein Atlas (29) revealed that ALCAM staining is TGF-b, were tested for their abilityto alter ALCAM shedding(Fig. clearly evident in both normal, low-, and medium-grade dis- 2A). ALCAM shed into the conditioned media is detected by ease but is frequently absent from the tumor cell surface in ELISA and normalized to the amount of cellular ALCAM high-grade disease (proteinatlas.org; Fig. 1D). detected in the lysate. Of the eight agents tested, only TGF-b

www.aacrjournals.org Cancer Res; 74(5) March 1, 2014 1407

Hansen et al.

4 A *** D

2 ALCAM

P-Smad2 1

shedding (ng/mL) GAPDH Normalized ALCAM 0 – DMSO –––+– b1 b DGF EGF TGF- ++––– BMP2BMP6VEGFFGF2P SDF-1 b Serum TGF- TGF- Inhib. –++– –

LNCaP PC3 B E ** 4 *** 2.5 * 2.0 3 1.5 2 mRNA 1.0 (ALCAM expression) 1 0.5 ALCAM fold change ALCAM fold change 0 0.0 TGF-b +–+– Vector Ctrl – + – T204D DA – – +

C LNCaP PC3 F

sALCAM Conditioned media (shed ALCAM) ALCAM

ALCAM Cell lysate (intact ALCAM) GAPDH

GAPDH Vector Ctrl – + – T204D DA – – + TGF-b +– +–

Figure 2. Expression and shedding of ALCAM is increased in response to TGF-b. A, ELISA analysis of ALCAM shedding in concentrated conditioned media of PC3 cells treated with indicated exogenous cytokines. TGF-b, P < 0.001. B, expression of ALCAM by RT-PCR fold change relative to GAPDH in PC3 or LNCaP cells treated with or without 10 ng/mL of TGF-b for 48 hours. C, Western blot analysis of shed ALCAM in the conditioned media and intact ALCAM in the cell lysate in LNCaP and PC3 cells. D, Western blot detection of ALCAM and phospho-smad2 expression in PC3 cells treated with 10 ng/mL TGF-b for 48 hours in the presence or absence of 10 mmol/L SB431542, a TGF-b receptor tyrosine kinase inhibitor. E and F, RT-PCR (E) and Western blot analysis (F) of ALCAM expression in LNCaP cells transfected with vector control or T204D, dominant active TGF-b type I receptor; , P < 0.05; , P < 0.01.

was able to promote ALCAM shedding. To further explore the whereas TGF-b–induced ALCAM mRNA and protein expression response to exogenous stimulation with TGF-b, ALCAM expres- could be restored in LNCaP cells when the cells were transfected sion in PC3 cells was compared with ALCAM expression in with dominant-active TGF-b receptor type I (Fig. 2E and F). LNCaP cells that are unable to respond to the cytokine because they lack TGF-b receptor type I (Fig. 2B and C; ref. 30). qRT-PCR ALCAM shedding in vivo correlates with tumor analysis for ALCAM demonstrates that TGF-b was also able to progression induce ALCAM gene transcription in PC3 but not LNCaP cells Published clinical studies have demonstrated that circulat- (Fig. 2B). The absence of any signiﬁcant increase in lysate ing levels of ALCAM are frequently elevated in patients with ALCAM further supports cytokine-induced ALCAM shedding cancer (19, 31, 32). These studies suggest that ALCAM is shed (Fig. 2C). Conversely, LNCaP did not respond to TGF-b even by the tumor. Indeed, experimental models of ovarian cancers though these cells express abundant ALCAM (Fig. 2C). TGF- indicate elevated shedding of ALCAM speciﬁcally from the b–induced expression in PC3 cells could be abrogated with the tumor (11). To determine whether tumor-derived ALCAM is small-molecule inhibitor SB431542 (Fig. 2D, 10 mm; Sigma), the source of elevated circulating ALCAM in prostate cancer,

1408 Cancer Res; 74(5) March 1, 2014 Cancer Research

ALCAM Is a Marker and Regulator of Skeletal Metastasis

Figure 3. ALCAM shedding correlates with tumor burden. –1 0 2 4 5 6 7 8 Circulating ALCAM levels were monitored longitudinally in mice bearing subcutaneously or orthotopically implanted PC3 cells. A, schematic representation of in vivo strategy for subcutaneous and orthotopic tumor models. IVIS, In Vivo Imaging System; PerkinElmer. B and C, circulating BC levels of soluble host and tumor- derived ALCAM detectable in mice bearing subcutaneous injected PC3 cells (B) or PC3 cells orthotopically implanted into the prostate (C). Levels of ALCAM are shown as a function of time (top) or tumor burden at the time of experiment completion (bottom). Each point reﬂects mean of duplicate measurements SD. Each line and corresponding R2 represents a best ﬁt linear regression analysis.

we used species-specific antibodies to monitor circulating derived ALCAM, changes in host-derived ALCAM did not cor- levels of both host (mouse) ALCAM and tumor (human) respond to tumor burden (Fig. 3B; host-derived R2 ¼ 0.03671, n ALCAM longitudinally during orthotopic and subcutaneous ¼ 4 animals; Fig. 3C; host-derived R2 ¼ 0.01358, n ¼ 4 animals). growth of PC3 cells (Fig. 3). To determine that tumor-derived A significantly greater amount of circulating ALCAM was ALCAM could act as a stable biomarker of cancer in vivo,we observed in mice bearing PC3 versus mice bearing LNCaP determined the half-life of human ALCAM in the circulation of tumors (Supplementary Fig. S4B), supporting the relationship its mouse host (Supplementary Fig. S3). Circulating ALCAM between malignancy and ALCAM shedding seen in vitro (Fig. exhibits a 17-hour half-life that is sufficient for monitoring its 2C). Moreover, shedding increased for tumor cells selected release from an endogenous tumor burden. from skeletal metastases (Supplementary Fig. S4F). Shedding is Circulating levels of ALCAM were subsequently monitored on likely to be universally present in solid tumors as it has been a weekly basis (Fig. 3A) in SCID mice bearing subcutaneous (Fig. reported for colon cancer (10), ovarian cancer (11), and is easily 3B, n ¼ 5) or orthotopic xenografts of PC3 (Fig. 3C, n ¼ 8). detected in models of breast and prostate cancer (Supplemen- Animals were bled on a predetermined schedule via saphenous tary Fig. S4C–S4E). vein puncture. Circulating ALCAM levels were detected by ELISA and a comparison with pregrafting baseline levels allowed ALCAM shedding is mediated by ADAM17 in vitro and in for the detection of any increase in host (mouse) ALCAM and the vivo appearance of tumor (human) ALCAM in response to an Because ALCAM is a proteolytic target of ADAM17, we increasing tumor burden. Tumor-derived ALCAM levels showed hypothesized that ADAM17 was responsible for TGF-b–induced significant weekly increases in the serum of tumor-bearing mice cleavage of ALCAM. This hypothesis was supported by pub- (Fig. 3B and C, and Supplementary Fig. S4A; P < 0.0001). lished work demonstrating that TGF-b can increase ADAM17 Regression analysis showed a direct linear relationship between activity by phosphorylation of the protease (33, 34). Indeed, circulating levels of tumor-derived ALCAM and tumor burden knockdown of ADAM17 using siRNA transfection resulted in a for subcutaneous xenografts (Fig. 3B; tumor-derived R2 ¼ 0.707; loss of TGF-b–induced ALCAM shedding (Fig. 4A). Similar P < 0.0001; n ¼ 4) and orthotopic xenografts (Fig. 3C; tumor- results were obtained by using an ADAM17-specific inhibitor derived R2 ¼ 0.7066; P < 0.0001; n ¼ 4). In contrast to tumor- (Fig. 4B, Compound 32, BMS; ref. 35). These studies were

www.aacrjournals.org Cancer Res; 74(5) March 1, 2014 1409

Hansen et al.

ABC

Figure 4. Ectodomain shedding of ALCAM is mediated by ADAM17 in vitro and in vivo. A, Western blot analysis of shed ALCAM in the conditioned media and ADAM17 in the total cell lysate of PC3 transiently transfected with either scrambled siRNA or siRNA targeting ADAM17. B, Western blot analysis of shed ALCAM in PC3 cells treated with an ADAM17-speciﬁc inhibitor (Compound 32, BMS), broad-spectrum matrix metalloproteinase (MMP) inhibitor (GM6001), or diluent control. C, serum levels of tumor-derived ALCAM in 10-week-old SCID mice bearing PC3-luc tumors and treated with DMSO diluent or Compound 32 for 3 days, P ¼ 0.0005.

extended to orthotopic models to confirm that ADAM17 was migration, possibly due to a loss of ALCAM–ALCAM homotypic also the primary protease responsible for ALCAM shedding in interaction on adjacent cells. vivo (Fig. 4C). In vivo dosing and efficacy for Compound 32 was Given the critical importance of cancer cell migration in confirmed by using serum TNF-a (Supplementary Fig. S5, n ¼ malignant tumor expansion and metastasis, we subsequently 6), which demonstrated that 50% inhibition of ADAM17 could evaluated the contribution of ALCAM to primary tumor be achieved for the duration of the ALCAM serum half-life growth and bone metastasis. Primary tumor growth was (Supplementary Fig. S3; 17 hours) without signs of distress or accomplished by using an orthotopic model based on implan- toxicity. Mice were treated twice daily for 3 days with 20 mg/kg tation of tumor cells into the anterior prostate of SCID mice of the ADAM17 inhibitor [Compound 32 or vehicle dimethyl (Fig. 5C, n ¼ 8 for PC3-luc-Control and PC3-luc-ALCAMKD1). sulfoxide (DMSO) control]. Presurgery, weekly, pre-, and post- Skeletal metastasis was accomplished by intracardiac injection treatment saphenous vein bleeds were collected. We found that of tumor cells in nude mice (Fig. 5D, n ¼ 31 for PC3-luc-Control inhibition of ADAM17 resulted in a significant decrease in and n ¼ 25 for PC3-luc-ALCAMKD1). Bioluminescent imaging serum levels of shed ALCAM, approximating the 50% inhibition was used to monitor tumor burden for both models at weekly we achieved with our dosing studies (Supplementary Fig. S5). intervals. Interestingly, reduced ALCAM expression did not Taken together, these data suggest that ALCAM cell surface limit tumor growth within the prostate (Fig. 5C). Whole body shedding is mediated by ADAM17 and promoted by TGF-b. and ex vivo bioluminescent imaging of the orthotopic model upon experiment completion confirmed that both the PC3-luc- Knockdown of ALCAM in PC3 cells inhibits TGF-b– Control and PC3-luc-ALCAMKD1 exhibited similar tumor bur- induced migration and in vivo dissemination to bone den based on luciferase activity (Fig. 5C), and comparable Given that TGF-b is a central driver of tumor cell motility and tumor size based on weight (Fig. 5D). Local invasion and metastasis, we sought to determine the effects of exogenous mesenteric dissemination is common in this model (36) and TGF-b on prostate cancer cells in vitro. To test whether ALCAM was not altered by reduced ALCAM expression. is functionally involved in tumor cell migration and metastasis, In contrast to the orthotopic model, a reduction in ALCAM we knocked down expression in PC3-Luc cells using viral resulted in a significant decrease in skeletal metastasis (Fig. 5E). delivery of shRNA. Three separate stable ALCAM knockdowns Both incidence and metastatic burden were reduced. Approx- were produced (ALCAM KD 1, 2, and 3). Transduced cells were imately, 75% of mice injected with PC3-luc-Control tumor cells selected with puromycin and subsequently subjected to flow- developed bone metastasis, whereas only 17% of the mice sorting to isolate the highest knockdown population (Fig. 5A). injected with PC3-luc-ALCAMKD1 cells developed bone lesions. We pretreated PC3-luc-ALCAMKD1 cells and PC3-luc-ALCAM In addition, in mice that did develop skeletal metastases formed shControl cells with10 mg/mL TGF-b1for16hoursinserum-free by PC3-luc-ALCAMKD1, the number of lesions per mouse was conditions, followed by the analysis of migration by magnetically greatly reduced (0.2 events vs. 1.4 events; Fig. 5F). The orthotopic attachable stencils (Fig. 5B; ref. 22). The analysis revealed a loss and intracardiac experiments were repeated with PC3-luc- TGF-b–induced migration in PC3-luc-ALCAMKD1 cells (Fig. 5B). ALCAMKD3 and similar results were obtained. Similar observations were made in ALCAMKD2 and ALCAMKD3 PC3 cells as well as in MDA-MB-231 and A549 cells that Knockdown of ALCAM in PC3 cells diminishes tumor represent breast and lung cancer, respectively (Supplementary growth in bone Fig. S6). Interestingly, the reduction in ALCAM expression led to To determine the biologic importance of tumor-derived a slight but statistically insignificant increase in spontaneous ALCAM in prostate tumor growth in the bone, PC3-luc-Control

1410 Cancer Res; 74(5) March 1, 2014 Cancer Research

ALCAM Is a Marker and Regulator of Skeletal Metastasis

Prostate xenografts Cell line PC3 shControl PC3 shALCAM 2.0 n.s. A KDCtrl C (–) D 1.5 ALCAM 1.0 GAPDH 0.5 Tumor weight (g) weight Tumor 0.0 shControl shALCAM

Intracardiac injection 100 B *** E F 80 PC3 shControl PC3 shALCAM 2.0 60 *** 40 1.5

% Migrated 20 1.0 0 shControl --++ 0.5 ++-- shALCAM 0.0

TGF-β +-+- # Bone lesions/animal shControl shALCAM

Figure 5. Tumor-derived ALCAM mediates TGF-b–induced migration and skeletal metastasis but not primary tumor growth in the prostate. A, Western blot analysis of lysates from parental PC3-luc and PC3-luc shRNA ALCAM (KD1) knockdown (KD) cells. B, quantitative analysis of tumor cell migration in parental PC3 cells and PC3 cells with shRNA-mediated knockdown of ALCAM treated with or without 10 ng/mL TGF-b. C, whole-animal luciferase imaging of mice bearing orthotopic PC3-luc parental tumors or PC3-luc ALCAM knockdown tumors 6 weeks after surgery. D, primary tumor weights of orthotopic PC3-luc parental tumors (n ¼ 8) and PC3-luc ALCAM knockdown tumors (n ¼ 8). E, representative whole-animal luciferase imaging and matching X-rays 8 weeks after intracardiac injection of PC3-luc parental (shControl) or PC3-luc ALCAM knockdown (shALCAM, KD1). F, average number of bone lesions in mice from PC3-luc (1.42 0.26; n ¼ 31) or PC3-luc shRNA ALCAM (KD1) knockdown cells (0.16 0.07; n ¼ 25). , P < 0.0001; n.s., not signiﬁcant.

(n ¼ 8), PC3-luc-ALCAMKD2 (n ¼ 8), and PC3-luc-ALCAMKD3 tumors, the bone tumors created by both ALCAM KD cells (n ¼ 8) were injected into the tibia of nude mice. Following exhibited elevated levels of cleaved caspase-3, suggesting that intratibial injection, luminescent and X-ray imaging was used these cells are experiencing a reduced ability to survive (Fig. to monitor the tumor burden over time (Fig. 6A). Quantitation 7G). In addition, bone tumors from PC3-luc-ALCAMKD3 had of the bioluminescent signal showed a markedly lower growth significantly lower Ki67 staining, indicating that reduced rate for the ALCAM KD cells (Fig. 6B). At completion of the ALCAM expression diminished the ability of PC3 cells to experiment, 100% of control mice exhibited lesion compared proliferate (Fig. 7H). These observations indicate that ALCAM with an 80% incidence in limbs bearing PC3-luc-ALCAMKD3 contributes to both proliferation and survival. tumor cells (Fig. 6C). To confirm that ALCAM expression was not regained by ALCAMKD in vivo, immunofluorescent staining was performed on intratibial tumors (Fig. 6D). Discussion The data presented here reveal for the first time that ALCAM ALCAM expression contributes to tumor cell survival and plays a important role in prostate cancer establishment in the proliferation in the bone microenvironment bone microenvironment. ALCAM mRNA is elevated in malig- The reduced tumor burden after intratibial injection sug- nant disease, yet by immunohistochemistry it is frequently gested a diminished capacity to grow in the bone. Detailed absent from the tumor cell surface in advanced disease (Fig. imaging of the osteolytic lesions by mCT (Fig. 7A) further 1D; refs. 15–17). We have demonstrated recently that in colo- confirmed decreased lesion area and increased bone volume rectal cancer ectodomain shedding is responsible for the appar- in the bones containing ALCAM KD cells (Fig. 7B and C). ent lossof ALCAM detection(10). This shedding is likelytobe the Histologic visualization of bone tumors generated by control cause for conflicting results reported across several malignan- and ALCAM KD cells resulted in reduced bone tumor size upon cies (37). Indeed, the protease responsible for shedding of ALCAM knockdown (Fig. 7D), however, both control and ALCAM ectodomain (ADAM17) is elevated in advanced prostate ALCAM KD tumors were osteolytic as evidenced by positive cancer (Fig. 1B; ref. 38). These data imply that, while ALCAM staining for the osteoclast marker, tartrate-resistant alkaline gene transcription is elevated in prostate cancer, ectodomain phosphatase (Supplementary Fig. S7). shedding depletes the cell surface of intact protein in advanced To assess changes in survival and proliferation, bone tumors disease. Regression analysis showed a direct linear relationship were stained for cleaved caspase-3 (Fig. 7E, apoptosis) and Ki67 between circulating levels of (human) tumor-derived ALCAM (Fig. 7F, proliferation), respectively. Compared with the control and tumor burden in animals with subcutaneous xenografts

www.aacrjournals.org Cancer Res; 74(5) March 1, 2014 1411

Hansen et al.

A Luminescence X-ray B D Collagen ALCAM ALCAM/collagen/DAPI D7 D28 8.0×1007 Vector 6.0×1007 KD2 KD3 4.0×1007 **

*** Vector Vector 2.0×1007 Total flux (p/s)

0 3020100 Days after injection

C Incidence Lesion area pixels) units, (arbitrary

1.5 20,000 KD2 KD2 Lesion areaLesion 15,000 1.0 10,000 ** ** 0.5 5,000 KD3 KD3 Incidence fraction 0.0 0 Vector KD2 KD3

Figure 6. Tumor-derived ALCAM impacts metastatic growth but not incidence after intratibial injection. A, representative whole-animal luciferase and X-ray imaging of mice postintratibial injection of PC3-luc parental tumors (vector, n ¼ 8), and PC3-luc ALCAM knockdown tumor cells (KD2, n ¼ 8 and KD3, n ¼ 8). B, bioluminescent curve of intratibial tumor development in mice bearing PC3-luc vector, PC3-luc KD2, and KD3 tumors (two-way ANOVA with Bonferroni posttest). C, tumor incidence and average lesion area in the tibias of mice bearing PC3-luc vector, PC3-luc KD2, and KD3 tumors. Data represent the mean SEM (n ¼ 8/group); , P < 0.01; , P < 0.0001. D, collagen I and ALCAM immunoﬂuorescence of tumor cells within the tibias of mice bearing PC3-luc vector or PC3-luc KD2 or KD3 tumors.

(Fig. 3B) and orthotopic xenografts (Fig. 3C). In contrast, the incidence indicates that ALCAM contributes to metastatic host-derived ALCAM did not correspond to tumor burden (Fig. dissemination as well as growth in the metastatic sites. It 3B and C), demonstrating that elevations in circulating ALCAM remains to be determined whether these two biologic con- are tumor-specific. Consistent with this observation, host- tributions are controlled by distinct molecular mechanisms. derived ALCAM does not increase, but rather decreases slightly, Previous work has shown that both soluble ALCAM-Fc and in immunocompetent mice challenged with lipopolysaccharide blocking ALCAM antibodies are able to decrease in vitro (LPS, a model of acute inflammation) or full-thickness skin transendothelial migration of THP1 monocytes (39). Although punch (a model for wound-healing; Supplementary Fig. S8A we could not test the role of ALCAM in extravasation directly, it and S8B, respectively). These data suggest that tumor-derived is possible that the requirement for transendothelial migration ALCAM is a marker specific of tumor burden and that host extends to metastatic tumor cells. The reduced metastatic ALCAM is not significantly shed in response to the tumor incidence after intracardiac injections indeed supports that burden. hypothesis (Fig. 5). The lesions that did arise from circulating Suppression of ALCAM expression using gene-specific PC3-luc-ALCAMKD cells were smaller than those generated shRNAs prevented TGF-b–induced migration in vitro (Fig. from PC3-luc-Control cells, suggesting an additional contri- 5B) and inhibited metastasis as well as tumor growth in bone bution from ALCAM to metastatic growth. Indeed, metastatic in vivo (Figs. 5–7). These observations demonstrate that lesions generated after intratibial injection of PC3-luc- ALCAM is not only a marker of cancer progression but also ALCAMKD cells did not significantly impact tumor incidence a significant regulator of tumor cell migration and metastasis (Fig. 6C) but did dramatically reduce expansion of the meta- to bone. Within the metastatic cascade, there are many static tumor burden (Fig. 7F). sequential steps that can contribute to the overall success of Intratibial tumors generated by PC3-luc-ALCAMKD did not any single metastatic lesion. Evaluation of the primary tumor reexpress ALCAM (Fig. 5D), allowing us to further investigate within the prostate did not reveal any deficiency in growth or any molecular disparities between metastatic lesions that local invasion. Conversely, experimental metastasis by intra- expressed ALCAM and those that did not. The abundant cardiac injection resulted in nearly 10-fold reduction of skeletal presence of osteoclast activity by TRAP staining (Supplemen- metastasis (Fig. 5). This reduced metastatic incidence suggests tary Fig. S7) in metastatic lesions of PC3-luc-ALCAMKD sug- that ALCAM is required for dissemination to the bone. Further gests that there is no deficiency in osteolysis. Nevertheless, examination of tumor growth after intratibial injection (Fig. 6) these lesions remain significantly smaller than those created by revealed a significant inhibition in tumor growth without PC3-luc-Control, suggesting that there was not a lag in tumor affecting the tumor incidence. The reduced metastatic inci- growth but rather a persistent reduced ability to proliferate or, dence from circulating tumor cells (intracardiac injection) conversely, a decreased ability to survive. A review of the together with the reduced growth of the metastatic burden literature reveals one study that suggested that the loss of in the bone (intratibial injection) without further impact on ALCAM may be associated with a proapoptotic behavior in

1412 Cancer Res; 74(5) March 1, 2014 Cancer Research

ALCAM Is a Marker and Regulator of Skeletal Metastasis

A E Cleaved caspase-3 Vector KD1 KD2 Vector KD1 KD2

B C F Ki67 40,000 *** 1.5 * ) ** 3 ** Vector KD1 KD2 30,000 1.0

20,000

0.5 Lesion area 10,000 m CT bone volume (mm (arbitrary units in pixels) 0 0.0 Vector KD1 KD2 Vector KD1 KD2

D Vector KD1 KD2 G H

0.015 *** 0.3 *

*** 0.010 0.2

0.005 0.1 total cell number # apoptotic cells/ total cell number

0.000 # proliferating tumor cells/ 0.0 Vector KD1 KD2 Vector KD1 KD2

Figure 7. Tumor-derived ALCAM impacts tumor survival and proliferation in the bone microenvironment of intratibial bone tumor model. A, representative three-dimensional reconstitutions of mCT images from mice injected with PC3-luc-vector, PC3-luc-KD1, and PC3-luc-KD2 tumor cells. B, boxplots of average BV/TV (bone volume/total volume) by group for the PC3-luc-vector, PC3-luc-KD2, and PC3-luc-KD3 tumor-bearing mice. Data represent quartiles with dots indicating outliers (1.5 upper or lower quartile; n ¼ 16 tibias/group); , P < 0.05; , P < 0.01 (one-way ANOVA, P ¼ 0.0047; post hoc Mann–Whitney). C, a boxplot representing the lesion area calculated from endpoint X-ray images from the same experiment shows dramatic decrease in the lesion area in the KD2 and KD3 tumor lesions compared with vector control (one-way ANOVA, P ¼ 0.0003; post hoc Mann—Whitney, , P < 0.01; , P < 0.0001; n ¼ 8/group). Lesion areas were measured using arbitrary pixel unit. D, representative H&E stains of osteolytic bone lesions in the hind leg of mice. Outlines indicate the osteolytic tumor lesion within the bone. E, representative immunofluorescent staining of cleaved caspase-3–positive cells (red) in tibias of PC3-luc-vector, PC3-luc-KD2, and PC3-luc-KD3 tumor-bearing mice. F, representative immunofluorescent staining of Ki67 proliferating cells (red) in tibias of PC3-luc-vector, PC3-luc-KD2, and PC3-luc-KD3 tumor-bearing mice. Data are mean SEM (n ¼ 8/group); , P < 0.05; , P < 0.01; , P < 0.005 (one- way ANOVA; Mann–Whitney test). G, apoptosis in the tumor-bone microenvironment as a function of total cell number was assessed by staining for cleaved caspase-3 in PC3-luc-vector, PC3-luc-KD2, and PC3-luc-KD3 tumor-bearing tibias of mice 4 weeks after injection. H, proliferation in the tumor-bone microenvironment as a function of total cell number was assessed by staining for Ki67 in PC3-luc-vector, PC3-luc-KD2, and PC3-luc-KD3 tumor- bearing tibias of mice 4 weeks after injection. breast cancer cells (40). Although we did not observe reduced expression is important in the osteotropism and survival of proliferation in vitro (Supplementary Fig. S9), in vivo metastatic tumor cells in the bone (Figs. 5 and 7, respectively), further lesions created by PC3-luc-ALCAMKD did exhibit reduced work is necessary to fully elucidate the mechanisms involved. proliferation and increased apoptosis. Intriguingly, PC3-luc- ALCAMKD2, which retains more ALCAM expression than PC3- luc-ALCAMKD3 (Fig. 6A), did not exhibit reduced proliferation, Conclusion suggesting that the threshold of ALCAM expression that Here, we have presented evidence that ALCAM not only influences cell survival is different from the threshold influ- serves as a longitudinal marker of tumor burden, but also encing cell proliferation. contributes functionally to skeletal metastasis. Clinically, this Together, these observations suggest a dual adhesive and may be of particular importance for prostate cancer in which signaling function of ALCAM, whereby the intracellular sig- metastasis to bone is a frequent aspect of end-stage disease. naling function of ALCAM functions to serve a protective role Monitoring ALCAM status may provide an indicator of met- in apoptosis, and the extracellular adhesive function is astatic potential and the ability to monitor metastatic burden. required for extravasation and subsequent colonization at a The relevance of ALCAM in the bone microenvironment is secondary site. Although we have convincing data that ALCAM coupled with our data showing that ALCAM expression and

www.aacrjournals.org Cancer Res; 74(5) March 1, 2014 1413

Hansen et al.

shedding is driven by TGF-b, a known contributor of the Administrative, technical, or material support (i.e., reporting or orga- nizing data, constructing databases): A.G. Hansen, T. Ketova, S.W. Hayward, vicious cycle in bone metastasis (2, 41, 42). Further elucidation A. Zijlstra of the mechanism by which ALCAM supports metastatic Study supervision: A. Zijlstra dissemination to bone can provide novel means of promoting Acknowledgments cytotoxic therapy in patients with skeletal metastases. The authors thank Drs. Carlos Arteaga (Vanderbilt University) and Guido Swart (Radboud University, Nijmegen, the Netherlands) for scientific insight and direction. We thank Dr. Preston Campbell (Vanderbilt University) for technical Disclosure of Potential Conflicts of Interest assistance. No potential conflicts of interest were disclosed. Grant Support This work was primarily supported by U.S. Public Health Services grants Authors' Contributions CA120711 and CA143081 to A. Zijlstra. A.G. Hansen was supported in part by T32 Conception and design: A.G. Hansen, H.L. Moses, S.W. Hayward, J.A. Sterling, HL007751 and P50 CA098131. T.D. Palmer was supported by CA136228 from the A. Zijlstra NIH. J.A. Sterling was supported by a VA Career Development Award. This work Development of methodology: A.G. Hansen, M. Jiang was supported in part by S10 RR027631 and P01 CA040035 awarded to the Acquisition of data (provided animals, acquired and managed patients, Vanderbilt Center for Bone Biology (VCBB). provided facilities, etc.): A.G. Hansen, S.A. Arnold, M. Jiang, T.D. Palmer, The costs of publication of this article were defrayed in part by the payment of T. Ketova, A. Merkel, M. Pickup, S. Samaras, J.A. Sterling, A. Zijlstra page charges. This article must therefore be hereby marked advertisement in Analysis and interpretation of data (e.g., statistical analysis, biostatistics, accordance with 18 U.S.C. Section 1734 solely to indicate this fact. computational analysis): A.G. Hansen, S.A. Arnold, Y. Shyr, A. Zijlstra Writing, review, and/or revision of the manuscript: A.G. Hansen, T.D. Received May 6, 2013; revised December 15, 2013; accepted December 18, 2013; Palmer, S. Samaras, H.L. Moses, S.W. Hayward, J.A. Sterling, A. Zijlstra published OnlineFirst January 2, 2014.

References 1. Pound CR, Partin AW, Eisenberger MA, Chan DW, Pearson JD, Walsh ciated with advanced tumor stage and early prostate-specific antigen PC. Natural history of progression after PSA elevation following radical relapse in prostate cancer. Hum Pathol 2011;42:1946–52. prostatectomy. JAMA 1999;281:1591–7. 16. Kristiansen G, Pilarsky C, Wissmann C, Stephan C, Weissbach L, Loy 2. Mundy GR. Metastasis to bone: causes, consequences and thera- V, et al. ALCAM/CD166 is up-regulated in low-grade prostate cancer peutic opportunities. Nat Rev Cancer 2002;2:584–93. and progressively lost in high-grade lesions. Prostate 2003;54:34–43. 3. Jin J-K, Dayyani F, Gallick GE. Steps in prostate cancer progression 17. Kristiansen G, Pilarsky C, Wissmann C, Kaiser S, Bruemmendorf T, that lead to bone metastasis. Int J Cancer 2011;128:2545–61. Roepcke S, et al. Expression profiling of microdissected matched 4. Palmer TD, Ashby WJ, Lewis JD, Zijlstra A. Targeting tumor cell motility prostate cancer samples reveals CD166/MEMD and CD24 as new to prevent metastasis. Adv Drug Deliv Rev 2011;63:568–81. prognostic markers for patient survival. J Pathol 2005;205:359–76. 5. van Kempen LC, Nelissen JM, Degen WG, Torensma R, Weidle UH, 18. Lunter PC, van Kilsdonk JW, van Beek H, Cornelissen IM, Bergers M, Bloemers HP, et al. Molecular basis for the homophilic activated Willems PH, et al. Activated leukocyte cell adhesion molecule (ALCAM/ leukocyte cell adhesion molecule (ALCAM)-ALCAM interaction. J Biol CD166/MEMD), a novel actor in invasive growth, controls matrix Chem 2001;276:25783–90. metalloproteinase activity. Cancer Res 2005;65:8801–8. 6. Rosso O, Piazza T, Bongarzone I, Rossello A, Mezzanzanica D, 19. Witzel I, Schroder€ C, Muller€ V, Zander H, Tachezy M, Ihnen M, et al. Canevari S, et al. The ALCAM shedding by the metalloprotease Detection of activated leukocyte cell adhesion molecule in the serum of ADAM17/TACE is involved in motility of ovarian carcinoma cells. Mol breast cancer patients and implications for prognosis. Oncology 2012; Cancer Res 2007;5:1246–53. 82:305–12. 7. Pokutta S, Weis WI. Structure and mechanism of cadherins and 20. Hu Z, Gupta J, Zhang Z, Gerseny H, Berg A, Chen Y-J, et al. Systemic catenins in cell-cell contacts. Annu Rev Cell Dev Biol 2007;23:237–61. delivery of oncolytic adenoviruses targeting transforming growth fac- 8. van Kempen LC, Meier F, Egelad M, Kersten-Niessen MJF, Garbe C, tor-b inhibits established bone metastasis in a prostate cancer mouse Weidle UH, et al. Truncation of activated cell adhesion molecule: a model. Hum Gene Ther 2012;23:871–82. gateway to melanoma metastasis. J Invest Dermatol 2004;122: 21. Wilding G, Zugmeier G, Knabbe C, Flanders K, Gelmann E. Differential 1293–301. effects of transforming growth factor beta on human prostate cancer 9. Hansen A, Swart GW, Zijlstra A. ALCAM. UCSD Nat Mol Pages 2011 cells in vitro. Mol Cell Endocrinol 1989;62:79–87. Feb 1. doi:10.1038/mp.a004126.01. 22. Ashby WJ, Wikswo JP, Zijlstra A. Magnetically attachable stencils and 10. Hansen AG, Freeman TJ, Arnold SA, Starchenko A, Jones-Paris CR, the non-destructive analysis of the contribution made by the under- Gilger MA, et al. Elevated ALCAM shedding in colorectal cancer lying matrix to cell migration. Biomaterials 2012;33:8189–203. correlates with poor patient outcome. Cancer Res 2013;73: 23. Li H, Jiang M, Honorio S, Patrawala L, Jeter CR, Calhoun-Davis T, et al. 2955–64. Methodologies in assaying prostate cancer stem cells. Methods Mol 11. Carbotti G, Orengo AM, Mezzanzanica D, Bagnoli M, Brizzolara A, Biol 2009;568:85–138. Emionite L, et al. Activated leukocyte cell adhesion molecule soluble 24. Park SI, Kim S-J, McCauley LK, Gallick GE. Pre-clinical mouse models form: a potential biomarker of epithelial ovarian cancer is increased in of human prostate cancer and their utility in drug discovery. Curr type II tumors. Int J Cancer 2012;132:2597–605. Protoc Pharmacol 2010;Chapter 14:Unit14.15. 12. King JA, Ofori-Acquah SF, Stevens T, Al-Mehdi A-B, Fodstad O, Jiang 25. Campbell JP, merkel AR, Masood-Campbell SK, Elefteriou F, Sterling WG. Activated leukocyte cell adhesion molecule in breast cancer: JA. Models of bone metastasis. J Vis Exp 2012;e4260. prognostic indicator. Breast Cancer Res 2004;6:R478–87. 26. Varambally S, Yu J, Laxman B, Rhodes DR, Mehra R, Tomlins SA, et al. 13. Mezzanzanica D, Fabbi M, Bagnoli M, Staurengo S, Losa M, Balladore Integrative genomic and proteomic analysis of prostate cancer reveals E, et al. Subcellular localization of activated leukocyte cell adhesion signatures of metastatic progression. Cancer Cell 2005;8:393–406. molecule is a molecular predictor of survival in ovarian carcinoma 27. Nakagawa T, Kollmeyer TM, Morlan BW, Anderson SK, Bergstralh EJ, patients. Clin Cancer Res 2008;14:1726–33. Davis BJ, et al. A tissue biomarker panel predicting systemic progres- 14. Micciche F, Da Riva L, Fabbi M, Pilotti S, Mondellini P, Ferrini S, et al. sion after PSA recurrence post-definitive prostate cancer therapy. Activated leukocyte cell adhesion molecule expression and shedding PLoS ONE 2008;3:e2318. in thyroid tumors. PLoS ONE 2011;6:e17141. 28. Onder TT, Gupta PB, Mani SA, Yang J, Lander ES, Weinberg RA. Loss 15. Minner S, Kraetzig F, Tachezy M, Kilic E, Graefen M, Wilczak W, et al. of E-cadherin promotes metastasis via multiple downstream transcrip- Low activated leukocyte cell adhesion molecule expression is asso- tional pathways. Cancer Res 2008;68:3645–54.

1414 Cancer Res; 74(5) March 1, 2014 Cancer Research

ALCAM Is a Marker and Regulator of Skeletal Metastasis

29. Uhlen M, Oksvold P, Fagerberg L, Lundberg E, Jonasson K, Forsberg zo[d]imidazol-1-yl)methyl)benzamide P10 substituents. Bioorg Med M, et al. Towards a knowledge-based Human Protein Atlas. Nat Chem Lett 2008;18:1577–82. Biotechnol 2010;28:1248–50. 36. Wang Y, Xue H, Cutz J-C, Bayani J, Mawji NR, Chen WG, et al. An 30. Kim IY, Ahn HJ, Zelner DJ, Shaw JW, Sensibar JA, Kim JH, et al. orthotopic metastatic prostate cancer model in SCID mice via grafting Genetic change in transforming growth factor beta (TGF-beta) receptor of a transplantable human prostate tumor line. Lab Invest 2005;85: type I gene correlates with insensitivity to TGF-beta 1 in human 1392–404. prostate cancer cells. Cancer Res 1996;56:44–8. 37. Ofori-Acquah SF, King JA. Activated leukocyte cell adhesion molecule: 31. Tachezy M, Zander H, Gebauer F, Marx A, Kaiﬁ JT, Izbicki JR, et al. a new paradox in cancer. Transl Res 2008;151:122–8. Activated leukocyte cell adhesion molecule (CD166)-Its prognostic 38. Karan D, Lin FC, Bryan M, Ringel J, Moniaux N, Lin M-F, et al. power for colorectal cancer patients. J Surg Res 2012;177:e15–20. Expression of ADAMs (a disintegrin and metalloproteases) and 32. Tachezy M, Effenberger K, Zander H, Minner S, Gebauer F, Vashist YK, TIMP-3 (tissue inhibitor of metalloproteinase-3) in human prostatic et al. ALCAM (CD166) expression and serum levels are markers adenocarcinomas. Int J Oncol 2003;23:1365–71. for poor survival of esophageal cancer patients. Int J Cancer 39. Masedunskas A, King JA, Tan F, Cochran R, Stevens T, Sviridov D, 2012;131:396–405. et al. Activated leukocyte cell adhesion molecule is a component of the 33. Wang SE, Xiang B, Guix M, Olivares MG, Parker J, Chung CH, et al. endothelial junction involved in transendothelial monocyte migration. Transforming growth factor beta engages TACE and ErbB3 to activate FEBS Lett 2006;580:2637–45. phosphatidylinositol-3 kinase/Akt in ErbB2-overexpressing breast 40. Jezierska A, Matysiak W, Motyl T. ALCAM/CD166 protects breast cancer and desensitizes cells to trastuzumab. Mol Cell Biol 2008;28: cancer cells against apoptosis and autophagy. Med Sci Monit 2006; 5605–20. 12:BR263–73. 34. Soond SM, Everson B, Riches DWH, Murphy G. ERK-mediated phos- 41. Bussard KM, Gay CV, Mastro AM. The bone microenvironment in phorylation of Thr735 in TNFalpha-converting enzyme and its potential metastasis; what is special about bone? Cancer Metastasis Rev role in TACE protein trafﬁcking. J Cell Sci 2005;118:2371–80. 2008;27:41–55. 35. Ott GR, Asakawa N, Lu Z, Anand R, Liu R-Q, Covington MB, et al. 42. Buijs JT, Que I, Lowik€ CW, Papapoulos SE, van der Pluijm G. Inhibition Potent, exceptionally selective, orally bioavailable inhibitors of TNF- of bone resorption and growth of breast cancer in the bone microen- alpha converting enzyme (TACE): novel 2-substituted-1H-ben- vironment. Bone 2009;44:380–6.

www.aacrjournals.org Cancer Res; 74(5) March 1, 2014 1415

ALCAM/CD166 Is a TGF- −β Responsive Marker and Functional Regulator of Prostate Cancer Metastasis to Bone

Amanda G. Hansen, Shanna A. Arnold, Ming Jiang, et al.

Cancer Res 2014;74:1404-1415. Published OnlineFirst January 2, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-13-1296

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2014/01/06/0008-5472.CAN-13-1296.DC1

Cited articles This article cites 39 articles, 9 of which you can access for free at: http://cancerres.aacrjournals.org/content/74/5/1404.full#ref-list-1

Citing articles This article has been cited by 4 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/74/5/1404.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/74/5/1404. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.