N-(4-Hydroxyphenyl)retinamide and nitric oxide pro-drugs exhibit apoptoticand anti-invasive effects against bone metastatic breast cancer cells

Ann-Marie Simeone�, Stefano Colella1, Ralf Krahe1,Marcella M.Johnson2, Edna Mora3 and Ana M.Tari

Department of Experimental Therapeutics, 1Department of CancerGenetics and 2Department of Biostatistics, The University of TexasM. D. Anderson Cancer Center, Houston, TX 77030, USA and 3PuertoRico Cancer Center and Surgery Division, The University of PuertoRico, Medical Sciences Campus, San Juan, PR 00936, USA

�To whom correspondence should be addressed.Email: amsimeon@mdanderson.org

Breast cancer most frequently metastasizes to bone causingdecreased quality of life and morbidity. Since currenttreatments are palliative, strategies to prevent bone meta-stases in breast cancer patients are required. There is sub-stantial evidence indicating that high levels of nitric oxide(NO) suppress tumor growth and metastasis in vivo. Wehypothesize that agents that produce high concentrationsof NO could prevent the spread of breast cancer to bone.We previously demonstrated that the synthetic retinoidN-(4-hydroxyphenyl)retinamide (4-HPR) produces highlevels of NO via the induction of NO synthases. NO pro-drugs are designed to produce large amounts of NO with-out inducing NO synthases but upon metabolism by theirintracellular targets. The objective of this study was todetermine the effectiveness of 4-HPR and an NO pro-drug, diethylamineNONOate/AM (NONO-AM), in inhibit-ing the growth and invasiveness of bone metastatic breastcancer cells. Parental MDA-MB-231 breast cancer cellswere resistant to 4-HPR-induced apoptosis at clinicallyrelevant doses, whereas 4-HPR-induced apoptosis in adose-dependent manner in MDA-MB-231/F10 bone meta-static breast cancer cells. Unlike 4-HPR, NONO-AMinduced apoptosis in a dose-dependent manner in bothparental MDA-MB-231 cells and F10 cells. The bonemetastatic F10 cells were more sensitive to the anti-invasiveeffects of 4-HPR and NONO-AM than were MDA-MB-231cells. Although suppression of matrix metalloprotease-9activity may be one mechanism by which 4-HPR decreasesthe invasion of F10 cells, it does not appear to be the anti-invasion mechanism of NONO-AM. These in vitro resultssuggest that 4-HPR and NO pro-drugs may be effectivechemopreventive agents against bone metastatic breastcancer.

Introduction

As many as 85% of patients with advanced breast cancerdevelop bone metastases resulting in complications such as

severe bone pain, fractures, hypercalcemia and spinal cord ornerve compression (1). The morbidity from bone metastasesin patients with breast cancer remains a significant healthproblem, accounting for 470% of breast cancer deaths (2,3).Despite the prevalence of this problem, there are currently nosatisfactory therapies available and most are administered withpalliative intent. Chemotherapy and hormonal therapy are thepreferred systemic treatments for bone metastatic breast can-cer, but they are typically associated with only temporarycontrol of symptomatic disease. Surgery, radiation, bisphos-phonates and analgesics are also used to reduce bone pain andto prevent or repair fractures. Therefore, there is a need for thedevelopment of chemopreventive agents that specificallyprevent the spread of breast cancer to bone.The free radical nitric oxide (NO) plays an important role in

regulating tumor growth and metastasis; however, it remainscontroversial whether NO positively or negatively impactsthese processes. The cellular origin and the concentration ofNO in the tumor microenvironment are crucial in determiningwhether NO exhibits tumor promoting or tumor suppressingeffects. NO is synthesized by three nitric oxide synthase (NOS)isoforms. NOSI and NOSIII are expressed constitutively andproduce trace amounts of NO (low pM for seconds to minutes).In contrast, NOSII is the inducible isoform and can generatelarge amounts of NO (mM for hours to days). Patient samplesand animal models indicate that NOSII expression in tumorcells reduces, while NOSII expression in host cells enhances,breast cancer metastasis (4–7). Low concentrations of NO mayprotect some cell types from apoptosis induced by DNA-damaging agents (8) and may increase the invasiveness andmetastatic potential of murine tumors (9,10). In contrast, highconcentrations of NO are cytotoxic (11–16) and prevent mur-ine tumor metastasis (17). Recently, there have been a numberof studies demonstrating that NOSII-mediated NO suppressestumorigenesis and metastasis in vivo (4,18,19). Therefore,molecules that can induce NOSII-mediated NO production intumor cells could potentially be used as chemopreventiveagents against bone metastatic breast cancer. One suchmolecule is the synthetic retinoid N-(4-hydroxyphenyl)retina-mide (4-HPR). We previously reported that 4-HPR at clinic-ally relevant doses induces apoptosis in breast cancer cells byinducing NOSII-mediated NO production (15). 4-HPR hasalso been shown to inhibit the invasion of several cancer celltypes, including ovarian cancer (20), prostate cancer (21–24)and Kaposi’s sarcoma (25).Unlike 4-HPR, NO donors are designed to spontaneously

generate large amounts of NO in aqueous media at physiolo-gical pH and temperature. However, this indiscriminate releaseof NO can have deleterious effects on the cardiovascular sys-tem and induce severe hypotension. To maximize NO expos-ure to tumor cells while minimizing undesirable side effectselsewhere in the body, NO pro-drugs have been designed togenerate large amounts of NO upon metabolism by intracellu-lar enzymes. The NO pro-drug diethylamine NONOate/AM

Abbreviations: DMEM/F12, Dulbecco’s modified Eagle medium; FBS,fetal bovine serum; MMP, matrix metalloprotease; 4-HPR, N-(4-hydroxyphenyl)retinamide; NO, nitric oxide; NOS, nitric oxide synthase;NONO-AM, diethylamine NONOate/AM.

Carcinogenesis vol.27 no.3 # Oxford University Press 2005; all rights reserved. 568

Carcinogenesis vol.27 no.3 pp.568–577, 2006doi:10.1093/carcin/bgi233Advance Access publication September 30, 2005

by guest on March 24, 2014

http://carcin.oxfordjournals.org/D

ownloaded from

(NONO-AM), which releases NO upon activation by intracel-lular esterases, has been shown to induce apoptosis inleukemia cells (26).Given their ability to produce large amounts of NO, 4-HPR

and NONO-AM may hold promise for the prevention of bonemetastases from breast cancer. The objective of this study wasto determine the effects of 4-HPR and NONO-AM on thein vitro growth and invasiveness of bone metastatic breastcancer cells.

Materials and methods

Reagents

4-HPR was purchased from Sigma Chemical (St Louis, MO). NONO-AM,GM6001, GM6001-negative and matrix metalloprotease (MMP)-2/MMP-9 inhibitor II {(2R)-[(4-biphenylsulfonyl)amino]-N-hydroxy-3-phenylpropionamide} were purchased from EMD Biosciences (La Jolla,CA). Matrigel was purchased from BD Biosciences (Bedford, MA). Hema-3was purchased from Fisher Scientific (Middleton, VA). Casein and gelatinzymogram gels, non-reducing sample buffer, 10� zymogram renaturing buf-fer, 10� zymogram development buffer, Coomassie blue-R250 and Coomassieblue destaining solution were purchased from Bio-Rad Laboratories (Hercules,CA). Stock solutions of 4-HPR (10 mM), NONO-AM (10 mM), GM6001(1 mM), GM6001-negative (1 mM) and MMP-2/MMP-9 inhibitor II (1 mM)were prepared in dimethyl sulfoxide and stored at �20�C. All reagents werediluted in the culture medium to the indicated final concentration.

Cell lines and culture conditions

The human MDA-MB-231 cell line was obtained from American Type CellCulture (Manassas, VA). The MDA-MB-231 cell line is an estrogen independ-ent highly metastatic human breast cancer cell line (27). MDA-MB-231 cellsmetastasize predominantly to bone, but may spread to the brain, ovary andadrenal glands after inoculation into the left ventricle of the heart in femalenude mice (28). Since there are no commercially available bone metastaticbreast cancer cell lines, we used a selective bone metastatic derivative ofMDA-MB-231 cells, MDA-MB-231/F10. The MDA-MB-231/F10 cell linewas established by inoculating MDA-MB-231 cells into the left ventricle ofthe heart of 6-week-old female nude mice. The tumor cells that metastasized tobone were isolated, grown in tissue culture and then reinoculated into the leftventricle of female nude mice. The cells that again metastasized to bone wereisolated, expanded and reinoculated into the left ventricle of the heart. Thisprocess was repeated until no micrometastases were detected histologically orby X-ray in tissues other than bone within 4–5 weeks of injection. The MDA-MB-231/F10 cell line was established and kindly provided by Dr Yoneda (TheUniversity of Texas Health Science Center, San Antonio, Texas). F10 cellsexclusively metastasize to bone after intracardiac inoculation in female nudemice (28). MDA-MB-231 and F10 breast cancer cells were cultured inDulbecco’s modified Eagles medium (DMEM/F12) supplemented with 5%heat-inactivated fetal bovine serum (FBS) at 37�C under 5% CO2 in a humidi-fied incubator.

Cell growth and NO assay

MDA-MB-231 and F10 breast cancer cells were plated at 2.5 � 104 cells/wellin 6-well plates in 2 ml of DMEM/F12 supplemented with 5% FBS. After 24 h,cells were treated with 4-HPR (1, 2.5 mM) or NONO-AM (0–100 mM). After5 days of incubation, cell growth was determined by total live cell counts usingtrypan blue exclusion. Supernatants were collected from untreated and treatedcells and stored at �80�C until NO determination. Total NO was determinedby quantifying nitrite, the stable end product of NO oxidation, spectrophoto-metrically using a Colorimetric non-enzymatic Nitric Oxide Assay Kit (OxfordBiomedical Research, Oxford, MI) as described previously (15). Nitrite valueswere normalized for total cell counts and expressed as mM per one millioncells. The values were reported as the mean (� SD) of experiments performedin triplicate.

Apoptosis assays

MDA-MB-231 and F10 cells were plated at 2.5 � 104 cells/well in 6-wellplates in 2 ml of DMEM/F12 supplemented with 5% FBS. After 24 h, cellswere treated with 4-HPR (1, 2.5 mM) or NONO-AM (40 mM). After 4 days ofincubation, apoptosis was determined by using flow cytometry and immuno-histochemistry. Cells were harvested and prepared for flow cytometry asdescribed previously (15). For immunohistochemistry, cells were collectedand suspended at 1 � 106 cells/ml in phosphate-buffered saline (PBS). Cyto-spins for each treatment were prepared using 100 ml of the appropriate cellsuspension. Slides were quick-fixed in �20�C acetone and stored at �20�C

until immunostaining. The FragELTM DNA Fragmentation Detection Kit(EMD Biosciences, La Jolla, CA) was used to detect apoptotic nuclei accord-ing to the manufacturer’s instructions. This kit allows the recognition of DNAbreaks in apoptotic nuclei. In this assay, terminal deoxynucleotidyl transferasebinds to exposed 30OH ends of DNA fragments and catalyzes the addition ofbiotin-labeled and unlabeled deoxynucleotides. Biotinylated nucleotides aredetected using a streptavidin–horseradish peroxidase conjugate. Diaminoben-zidine reacts with the labeled sample to generate an insoluble brown substrateat the site of DNA fragmentation. Samples were counterstained with methylgreen to aid in distinguishing between normal and apoptotic cells.

Matrigel invasion assay

The effect of 4-HPR and NONO-AM on the invasiveness of breast cancer cellswas determined in vitro by determining the number of MDA-MB-231 and F10cells that invaded through transwell inserts coated with the artificial basementmembrane Matrigel. This assay was also used to determine the importance ofMMP activity on the invasiveness of MDA-MB-231 and F10 cells. Briefly,6-well plate transwell inserts with 8 mm pore-size polycarbonate filters (FisherScientific, Middleton, VA) were coated with Matrigel in cold serum-freeDMEM/F12 at a final concentration of 0.7 mg/ml and placed at room temper-ature for 40 min. MDA-MB-231 and F10 cells were trypsinized, resuspendedin serum-supplemented medium and counted. Cells were then washed threetimes with serum-free medium. Cells (3� 105 in 500 ml) were added into eachtranswell insert and incubated for 72 h in the absence and presence of 4-HPR(1, 2.5 mM) or NONO-AM (20, 40 mM). Cells were also incubated in thepresence of the broad-spectrum MMP inhibitor GM6001 (10 mM) or anidentical amount of its specific product of negative control, GM6001-negative,or a specific MMP-2/MMP-9 inhibitor (10 mM) for 72 h. The lower chamberswere filled with 2 ml of DMEM/F12 medium supplemented with 5% FBS.After incubation, non-invading cells on the upper surface of the filter wereremoved with cotton swabs. Cells that had passed through the pores onto thelower side of the filter were fixed, stained with Hema-3 and photographed.The migrated cells were counted in five fields for each filter under alight microscope at 40� magnification. The invasive ability of the cells wasexpressed as the mean number of cells that had invaded to the lower side of thefilter. The experiments were performed in triplicate and repeated twice.

Zymographic analysis of MMP activity

MDA-MB-231 and F10 cells (3 � 105) were plated in T-25 flasks and incub-ated for 24 h in the absence or presence of 4-HPR (1, 2.5 mM) or NONO-AM(20, 40 mM). MDA-MB-231 and F10 cells were also incubated for 24 h withthe MMP-2/MMP-9 inhibitor (10 mM) to verify the downregulation of MMP-9activity. Cells were changed to serum-free medium the next day for 24 h. Themedium was recovered and centrifuged for 5 min. The conditioned mediumwas collected and concentrated by using spin columns with 10-kDa-cutofffilters (Millipore, Bedford, MA). Concentrated conditioned medium (20 ml)was mixed (1:1) with non-reducing sample buffer, incubated at 37�C for15 min, and then applied to gelatin and casein zymogram gels. After electro-phoresis, the gels were incubated for 3 h in zymogram renaturing buffer atroom temperature with gentle agitation, followed by an overnight incubation inzymogram development buffer at 37�C. The gels were stained with Coomassieblue-R250 for 3 h and then destained with Coomassie blue destaining solution.Gelatinase and caseinase activities were visible as clear bands against the darkblue background, indicating proteolysis of the substrate protein.

Statistical analyses

For statistical analysis of the invasion experiments, the Shapiro–Wilk test wasfirst performed to assess the normality assumption of the data. Given that thedata was normally distributed, two-sample t-tests were performed for each cellline to compare the number of invading cells for the untreated group with thenumber of invading cells for each of the 4-HPR doses. In addition, the numberof invading cells was compared between the two 4-HPR doses. These analyseswere also performed for the NONO-AM, GM6001 and MMP-2/MMP-9 inhib-itor treatments used in the invasion experiments. The significance level foreach individual comparison was adjusted by the Bonferroni method to accountfor multiple testing within each cell line to achieve an overall significance levelof 0.05. All analyses were performed using SAS�.

Results

4-HPR’s induction of apoptosis in F10 bone metastatic breastcancer cells was correlated with NO production

To assess the effect of 4-HPR on the growth of breast cancercells, MDA-MB-231 and F10 cells were treated with 4-HPR.The two experimental doses 1 and 2.5 mM were chosenbecause they are clinically achievable (29). MDA-MB-231

Effects of 4-HPR and NONO-AM on bone metastatic breast cancer cells

569

by guest on March 24, 2014

http://carcin.oxfordjournals.org/D

ownloaded from

cells were resistant to 4-HPR-induced growth inhibition(Table I). In contrast, 4-HPR at 1 and 2.5 mM inhibited thegrowth of F10 cells by 53 and 69%, respectively.NO production increased 3.4- and 5.5-fold in F10 cells

treated with 1 and 2.5 mM 4-HPR, respectively (Table I).In contrast, NO production was not induced in MDA-MB-231 cells treated with 4-HPR. Thus, the growth inhibitoryeffects of 4-HPR were directly correlated with increases inNO production in F10 cells.Flow cytometric analysis of propidium iodide-stained cells

was used to determine whether apoptosis was related tothe growth inhibition found in 4-HPR-treated bonemetastatic breast cancer cells (Figure 1A). 4-HPR increasedthe percentage of apoptotic F10 cells in a dose-dependent

manner (Figure 1A). The number of apoptotic F10 cells wasincreased from 1.1 to 26.7 and 47.3% by the 1 and 2.5 mMdoses of 4-HPR, respectively (Figure 1A). The apoptoticeffects of 4-HPR were confirmed by using DNA fragmentation(Figure 1B). These data indicate that F10 bone metastaticbreast cancer cells were sensitive to pharmacologically achiev-able doses of 4-HPR.

NONO-AM induced apoptosis in MDA-MB-231 and F10breast cancer cells

To assess the effect of NONO-AM on the growth of breastcancer cells, MDA-MB-231 and F10 cells were treated withincreasing concentrations of NONO-AM (0–100 mM). In con-trast to 4-HPR, NONO-AM inhibited the growth of bothMDA-MB-231 and F10 cells in a dose-dependent manner(Figure 2A). The IC50 of NONO-AM was between 20 and40 mM in both cell lines. Growth inhibition in NONO-AM–treated cells was correlated with dose-dependent increases inNO concentration (Figure 2B). To determine whether apop-tosis was related to the growth inhibition found in thebreast cancer cell lines treated with NONO-AM, MDA-MB-231 and F10 cells were treated with NONO-AM (40 mM)and examined by flow cytometry (Figure 2C). NONO-AMincreased the number of apoptotic cells from 2.2 to 35% forMDA-MB-231 cells and from 1.7 to 46.8% for F10 cells(Figure 2C). The apoptotic effects of NONO-AM were con-firmed by using DNA fragmentation (Figure 2D). These dataindicate that both the parental and the F10 bone metastaticbreast cancer cells were sensitive to the apoptotic effects ofNONO-AM.

Table I. 4-HPR selectively induced growth inhibition and NO productionin bone metastatic breast cancer cells

Cell line 4-HPR(mM)

Cell count(�104)a

Nitrite(mM/106 cellsb)

MDA-MB-231 0 53.1 � 1.1 21.7 � 2.41 55.4 � 1.4 18.5 � 1.22.5 52.0 � 1.2 20.7 � 1.4

F10 0 51.8 � 1.3 20.7 � 0.51 24.6 � 0.8 70.2 � 3.12.5 15.9 � 0.7 114.3 � 6.5

aTotal live cell counts were determined by trypan blue exclusion.Valuesshown are the means (� SD) of experiments performed in triplicate.bTotal NO production was determined by measuring its stable end productnitrite. Nitrite values were normalized to cell number. Values shown arethe means (� SD) of experiments performed in triplicate.

peak

A

B

Fig. 1. 4-HPR selectively induced apoptosis in bone metastatic breast cancer cells. F10 cells (2.5 � 104 cells/well) were plated in 6-well plates in DMEM/F12medium supplemented with 5% FBS. The next day, cells were treated with 4-HPR (1, 2.5 mM). After 4 days of incubation, cells were harvested and (A) flowcytometry and (B) immunohistochemical staining of DNA fragmentation were performed to determine the presence of apoptotic cells. Asterisk indicates thepercentage of apoptotic cells (i.e. cells in the sub-G1 peak).

A.M.Simeone et al.

570

by guest on March 24, 2014

http://carcin.oxfordjournals.org/D

ownloaded from

F10 bone metastatic breast cancer cells were more sensitiveto the anti-invasive effects of 4-HPR and NONO-AM

We examined the effect of 4-HPR and NONO-AM on theinvasive ability of MDA-MB-231 and F10 cells by using theMatrigel invasion assay. To avoid confounding any anti-invasive effects of 4-HPR and NONO-AM with their pro-apoptotic effects, these experiments were performed underconditions (initial plating density and treatment duration) inwhich510% growth inhibition was induced (data not shown).Untreated MDA-MB-231 and F10 cells displayed a high invas-ive capacity on Matrigel (Figure 3A). 4-HPR significantly(P 5 0.05) reduced the number of invasive cells in both celllines; however, F10 cells were more sensitive to the anti-invasive effects of 4-HPR (Figure 3A). Significant inhibition(P 50.05) of invasion was observed in F10 cells (74%) and asmaller but significant (P 50.05) decrease in invasion (28%)was observed inMDA-MB-231 cells after treatment with 1mM4-HPR (Figure 3B). The 2.5 mM dose of 4-HPR had a similaranti-invasive effect in MDA-MB-231 (74%) and F10 cells

(77%; Figure 3B). The two doses of 4-HPR did not havesignificantly different anti-invasive effects in F10 cells.In contrast to 4-HPR, NONO-AM inhibited the invasion of

F10 cells but not MDA-MB-231 cells (Figure 4A). At 20 and40 mM, NONO-AM significantly (P 5 0.05) reduced thenumber of invading F10 cells by 60 and 75%, respectively,compared with untreated cells (Figure 4B). The two doses ofNONO-AM did not have significantly different anti-invasiveeffects in F10 cells (Figure 4B).

Effects of 4-HPR and NONO-AM on MMP activities

MMPs, which are essential to the invasive process, areinvolved in the degradation of the extracellular matrix andbasement membranes. To confirm that MMP activity wasimportant for the invasiveness of MDA-MB-231 and F10cells, the broad-spectrum MMP inhibitor GM6001 (10 mM)was used to block MMP activity in a Matrigel invasion assay.GM6001 at the dose used was not cytotoxic to either cell line(data not shown). Inhibition of MMP activity significantly

Fig. 2. NONO-AM induced apoptosis in metastatic breast cancer cells. (A) MDA-MB-231 and F10 cells (2.5 � 104 cells/well) were plated in 6-well plates inDMEM/F12 medium supplemented with 5% FBS. The next day, cells were treated with NONO-AM (0–100 mM). After 5 days of incubation, cell growth wasdetermined by total live cell counts using trypan blue exclusion. The values shown are the mean (� SD) of experiments performed in triplicate. (B) MDA-MB-231 and F10 cells (2.5 � 104 cells/well) were plated in 6-well plates in DMEM/F12 medium supplemented with 5% FBS. The next day, cells were treatedwith NONO-AM (0–100 mM). After 5 days of incubation, NO production was determined by measuring its stable end product nitrite using a ColorimetricNitric Oxide Assay Kit. Nitrite values were normalized to cell number. The values shown are the mean (� SD) of experiments performed in triplicate.MDA-MB-231 and F10 cells (2.5 � 104 cells/well) were plated in 6-well plates in DMEM/F12 medium supplemented with 5% FBS. Cells were treated withNONO-AM (40 mM) 24 h later. After 4 days of incubation, cells were harvested and (C) flow cytometry and (D) immunohistochemical staining of DNAfragmentation were performed to detect the presence of apoptotic cells. Asterisk indicates the percentage of apoptotic cells (i.e. cells in the sub-G1 peak).

Effects of 4-HPR and NONO-AM on bone metastatic breast cancer cells

571

by guest on March 24, 2014

http://carcin.oxfordjournals.org/D

ownloaded from

(P 5 0.05) decreased the invasiveness of MDA-MB-231 andF10 cells (Figure 5A). The number of invaded cells wasdecreased 40 and 54% in MDA-MB-231 and F10 cells,respectively, by treatment with GM6001 (Figure 5B). In con-trast, GM6001-negative, the specific product of negative con-trol for GM6001, did not influence the invasive ability of eithercell line (Figure 5A and B).To determine whether 4-HPR and NONO-AM decreased the

invasive potential of breast cancer cells by blocking MMPactivities, casein and gelatin zymography was performed onuntreated and treated MDA-MB-231 and F10 cells. MMP-9and MMP-2 activities were detected by gelatin zymography inthe supernatants of untreated MDA-MB-231 and F10 cells(Figure 5C). The activity of MMP-9 was greater than that of

MMP-2 in both cell lines and was greater in F10 cells than inMDA-MB-231 cells. MMP-9 activity decreased in adose-dependent manner in MDA-MB-231 cells treated with4-HPR, whereas only the 2.5 mM dose of 4-HPR decreasedMMP-9 activity in F10 cells. NONO-AM had no effect onMMP-9 activity in either cell line. MMP-2 activity was notaffected by treatment with 4-HPR or NONO-AM in eithercell line. Casein zymography revealed two protease bandswith molecular weights of �40–56 kDa in untreated F10cells (Figure 5D). These bands may correspond to the stro-melysins MMP-3 and MMP-10 or may represent an MMP-3doublet (30). Caseinase activities were not altered inMDA-MB-231 and F10 cells by treatment with 4-HPR orNONO-AM.

B

Fig. 3. Bone metastatic breast cancer cells were more sensitive to the anti-invasive effects of 4-HPR than the parental metastatic breast cancer cells.(A) MDA-MB-231 and F10 cells (3 � 105 in 500 ml) were added into transwell inserts (8 mm pore-size) coated with Matrigel. Cells were incubated for72 h in the absence and presence of 4-HPR (1, 2.5 mM). Cells that passed through the pores onto the lower side of the filter were fixed, stained with Hema-3and photographed. (B) The migrated cells were counted in five fields for each filter under a light microscope at 40� magnification. The invasive ability of thecells was expressed as the mean number of cells that invaded to the lower side of the filter. Values shown are the mean (� SD) of experiments performed intriplicate. �, P 50.05 when compared with untreated cells.

A.M.Simeone et al.

572

by guest on March 24, 2014

http://carcin.oxfordjournals.org/D

ownloaded from

Since 4-HPR decreased MMP-9 activity in MDA-MB-231and F10 cells, the importance of MMP-9 in the invasiveness ofthese cell lines was determined. A specific MMP-9 inhibitor(10 mM) was used to block MMP-9 activity of the cell lines ina Matrigel invasion assay. The MMP-9 inhibitor at the doseused was not cytotoxic to either cell line (data not shown). Theinhibitor suppressed MMP-9 activity as revealed by gelatinzymography (data not shown). Inhibition of MMP-9 activitysignificantly (P 50.05) decreased the invasiveness of MDA-MB-231 and F10 cells (Figure 5E). The number of invadedcells was decreased 39 and 41% in MDA-MB-231 and F10cells, respectively, by treatment with the MMP-9 inhibitor

(Figure 5F). Therefore, suppression of MMP-9 activity maybe one mechanism by which 4-HPR decreased the invasivebehavior of the breast cancer cell lines.

Discussion

We found that the F10 bone metastatic breast cancer cells weresensitive to the apoptotic and anti-invasive effects of 4-HPRand NONO-AM. Both drugs produce high levels of NO but bydifferent mechanisms. 4-HPR induces NO production byincreasing NOSII expression in breast cancer cells (15),

Fig. 4. NONO-AM selectively inhibited the invasion of bone metastatic breast cancer cells. (A) MDA-MB-231 and F10 cells (3 � 105 in 500 ml) were addedinto transwell inserts (8 mm pore-size) coated with Matrigel. Cells were incubated for 72 h in the absence and presence of NONO-AM (20, 40 mM). Cells thatpassed through the pores onto the lower side of the filter were fixed, stained with Hema-3 and photographed. (B) The migrated cells were counted in fivefields for each filter under a light microscope at 40� magnification. The invasive ability of the cells is expressed as the mean number of cells that invaded tothe lower side of the filter. Values shown are the mean (� SD) of experiments performed in triplicate. �, P 50.05 when compared with untreated cells.

Effects of 4-HPR and NONO-AM on bone metastatic breast cancer cells

573

by guest on March 24, 2014

http://carcin.oxfordjournals.org/D

ownloaded from

B

C

Fig. 5. Gelatinase and caseinase activities in MDA-MB-231 and F10 cells treated with 4-HPR or NONO-AM. (A) MDA-MB-231 and F10 cells (3 � 105 in500 ml) were added into transwell inserts coated with Matrigel. Cells were incubated for 72 h in the absence and presence of broad-spectrum MMP inhibitorGM6001 (10 mM) or an identical amount of GM6001-negative for 72 h. GM6001-negative is the specific product of negative control for GM6001. Cells thatpassed through the pores onto the lower side of the filter were fixed, stained with Hema-3 and photographed. (B) The migrated MDA-MB-231 and F10 cellswere counted in five fields for each filter under a light microscope at 40� magnification. The invasive ability of the cells was expressed as the mean numberof cells that invaded to the lower side of the filter. Values shown are the means (� SD) of experiments performed in triplicate. �P5 0.05 when compared tountreated cells. (C) The effects of 4-HPR and NONO-AM on gelatinase activitiy were evaluated by zymography. MDA-MB-231 and F10 cells (3 � 105 cells)were plated in T-25 flasks. The next day, cells were treated for 24 h with 4-HPR (1, 2.5 mM) or NONO-AM (20, 40 mM). Cells were then changed to serum-free medium for 24 h. Concentrated conditioned medium (20 ml) was mixed (1:1) with non-reducing sample buffer and applied to gelatin substrate gels. Thegels were incubated in renaturing buffer followed by an overnight incubation in development buffer at 37�C. The gel was stained with Coomassie blue-R250and then destained. Gelatinase activities were visible as clear bands against the dark blue background, indicating proteolysis of the substrate protein. Lane 1,untreated; lane 2, 1 mM 4-HPR; lane 3, 2.5 mM 4-HPR; lane 4, 20 mM NONO-AM; and lane 5, 40 mM NONO-AM. (D) The effect of 4-HPR and NONO-AMon caseinase activity was evaluated by zymography. Casein zymography was performed in an identical manner to gelatin zymography except thatconcentrated conditioned medium was applied to casein substrate gels. Lanes were labeled as in (C). (E) Matrigel invasion assay of MDA-MB-231 and F10cells treated with a specific MMP-9 inhibitor (10 mM) for 72 h. (F) The number of invaded MDA-MB-231 and F10 cells were counted in five fields for eachfilter under a light microscope at 40� magnification. The invasive ability of the cells was expressed as the mean number of cells that invaded to the lowerside of the filter. Values shown are the mean (� SD) of experiments performed in triplicate. �, P 50.05 when compared with untreated cells.

A.M.Simeone et al.

574

by guest on March 24, 2014

http://carcin.oxfordjournals.org/D

ownloaded from

whereas NONO-AM releases NO when activated by intracel-lular esterases (26). We found that 4-HPR selectively inducedapoptosis in F10 cells but not in MDA-MB-231 cells. On theother hand, NONO-AM was effective in inducing apoptosisin both cell lines. These data indicate that both the parentaland the bone metastatic cells are sensitive to NO-mediatedapoptosis. However, the parental cells may possess intrinsicresistance mechanisms that could prevent 4-HPR from

inducing NO, probably by preventing 4-HPR from inducingNOSII expression, as is the case in HER2/neu-overexpressingbreast cancer cells (31).The invasiveness of F10 cells was suppressed by 4-HPR and

NONO-AM, suggesting that the bone metastatic breast cancercells are highly sensitive to NO-mediated anti-invasive effects.In contrast, the invasiveness of the parental MDA-MB-231cells appears to be insensitive to NO, since NONO-AM,

E

Fig. 5. (continued).

Effects of 4-HPR and NONO-AM on bone metastatic breast cancer cells

575

by guest on March 24, 2014

http://carcin.oxfordjournals.org/D

ownloaded from

which could generate large amounts of NO in MDA-MB-231cells, did not decrease the invasive activity of these cells.Surprisingly, 4-HPR decreased the invasive activity of bothMDA-MB-231 and F10 cells, albeit more effectively in F10cells. In unpublished observations, cDNA microarray analysisof 4-HPR-modulated genes in MDA-MB-231 and F10 cellsdid not reveal any gene expression changes 42-fold on Affy-metrix U133A and U133B GeneChips. In MDA-MB-231 cells,4-HPR-modulated the expression of 19 genes by 1.5-fold onthe Affymetrix U133A GeneChip and 15 genes on the Affy-metrix U133B GeneChip. Some of these genes includedthrombospondin-1, interleukin-8 and the C-X-C motif ligands1, 2 and 3. In F10 cells, 4-HPR did not induce any geneexpression changes 41.5-fold in either GeneChip. These datasuggest that the induction of apoptotic and anti-invasiveeffects by 4-HPR in breast cancer cells does not occur at thetranscriptional level.MDA-MB-231 cells have been shown to secrete the gelat-

inase MMP-9 and the caseinase MMP-3 (32–34). In the pre-sent study, suppression of MMP-9 activity, but not MMP-2 orcaseinase activities (MMP-3/MMP-10), was observed inMDA-MB-231 and F10 cells treated with 4-HPR. MMP-9activity was important in the invasiveness of MDA-MB-231and F10 cells since suppression of MMP-9 activity decreasedthe number of invaded cells. In MDA-MB-231 cells, 4-HPRdecreased MMP-9 activity in a dose-dependent manner, whichcorresponded with the dose-dependent decrease in invasion.In F10 cells, invasion was inhibited to a similar degree bytreatment with either 1 or 2.5 mM 4-HPR; however, decreasedMMP-9 activity was observed only at the 2.5 mM dose. Sup-pressing the activity of MMP-9 may be one mechanism bywhich 4-HPR decreases the invasive behavior of MDA-MB-231 and F10 cells. However, it is apparent from these resultsthat 4-HPR uses more than one mechanism to suppress theinvasiveness of metastatic breast cancer cells. MDA-MB-231cells have been shown to secrete the interstitial collagenasesMMP-1 and MMP-13 (32). It is possible that 4-HPR maydecrease the activity of these MMPs to suppress invasion.In addition, 4-HPR inhibited the invasion of transformedBALB/c 3T3 cells and prostate cancer cells by suppressingchemotactic motility (35) and by altering cellular adhesion tothe extracellular matrix (21,35), so the drug may also use thesemechanisms to inhibit the invasive capacity of metastaticbreast cancer cells. Interestingly, NONO-AM did not haveany effect on gelatinase or caseinase activities in F10 cells,indicating that the drug may have affected invasion by alteringthe activity of MMPs that we did not evaluate or by alteringcellular adhesion.Bone is the most frequent site of metastatic disease in

patients with advanced breast cancer, and the prognosis inthese patients is generally poor. According to National CancerInstitute, only 10–20% of women with metastatic breast cancersurvive the disease and achieve permanent remission. Currenttreatments, including chemotherapy, hormonal therapy andradiotherapy, are generally administered with palliative intent;therefore, chemoprevention to prevent the formation of bonemetastases could have a profound effect on the incidence ofbone metastases and associated mortality rate. Bisphosphon-ates are an effective palliative treatment for bone metastaticbreast cancer, but randomized clinical trials investigating theadjuvant use of bisphosphonates to prevent bone metastaseshave yielded conflicting results. Diel et al. (36) and Powleset al. (37) reported a reduction in the occurrence of bone

metastases from breast cancer and an increase in overall sur-vival in patients with primary breast cancer who receivedadjuvant clodronate for 2 years. However, a decrease (38) orlack of effect (37) of clodronate was evident at the 5-yearfollow-up. In contrast, Saarto et al. (39) found that 3 years ofadjuvant clodronate had no effect on the occurrence of bonemetastases and a negative effect on the occurrence of non-bonemetastases as well as survival in breast cancer patients. There-fore, chemopreventive strategies that specifically prevent theoccurrence of bone metastases from breast cancer are urgentlyneeded. The results of the present in vitro studies indicate thepotential of 4-HPR and NO pro-drugs as chemopreventiveagents against bone metastases from breast cancer and warranttheir further evaluation in in vivo chemoprevention studies.

Acknowledgements

The authors thank Drs Douglas Boyd and Lingegowda Mangala for theiradvice with the invasion assay and Dr Arturo Chavez-Reyes for his help withthe DNA fragmentation assay and photographs. The authors thank KarenRamirez and The University of Texas M. D. Anderson Cancer Center Depart-ment of Immunology Flow Cytometry Core Lab for the technical assistance.This research was supported in part by the IDO1-014 U54 RFA CA 096300(A.M.T. and E.M.) and the institutions NCI Core Grant (CA16672).

Conflict of Interest Statement: None declared.

References

1.Coleman,R.E. and Rubens,R.D. (1987) The clinical course of bonemetastases from breast cancer. Br. J. Cancer, 55, 61–66.

2.Aaron,A.D. (1994) The management of cancer metastatic to bone. JAMA,272, 1206–1209.

3.Weiss,L. (1992) Comments on hematogenous metastatic patterns inhumans as revealed by autopsy. Clin. Exp. Metastasis, 10, 191–199.

4.Gauthier,N., Lohm,S., Touzery,C., Chantome,A., Perette,B., Reveneau,S.,Brunotte,F., Juillerat-Jeanneret,L. and Jeannin,J.-F. (2004) Tumour-derived and host-derived nitric oxide differentially regulate breastcarcinoma metastasis to the lungs. Carcinogenesis, 25, 1559–1565.

5.Thomsen,L.L., Miles,D.W., Happerfield,L., Bobrow,L.G., Knowles,R.G.and Moncada,S. (1995) Nitric oxide synthase activity in human breastcancer. Br. J. Cancer, 72, 41–44.

6.Tschugguel,W., Schneeberger,C., Unfried,G., Czerwenka,K.,Weninger,W., Mildner,M., Gruber,D.M., Sator,M.O., Waldhor,T. andHuber,J.C. (1999) Expression of inducible nitric oxide synthase in humanbreast cancer depends on tumor grade. Breast Cancer Res. Treat., 56,145–151.

7.Vakkala,M., Kahlos,K., Lakari,E., Paakko,P., Kinnula,V. and Soini,Y.Inducible nitric oxide synthase expression, apoptosis, and angiogenesis inin situ and invasive breast carcinomas. Clin. Cancer Res., 6, 2408–2416.

8.Xu,W., Liu,L., Smith,G.C. and Charles,G. (2000) Nitric oxide upregulatesexpression of DNA-PKcs to protect cells from DNA-damaging anti-tumour agents. Nat. Cell Biol., 2, 339–345.

9.Orucevic,A., Bechberger,J., Green,A.M., Shapiro,R.A., Billiar,T.R. andLala,P.K. (1999) Nitric-oxide production by murine mammary adenocar-cinoma cells promotes tumor-cell invasiveness. Int. J. Cancer, 81,889–896.

10. Jadeski,L.C., Hum,K.O., Chakraborty,C. and Lala,P.K. (2000) Nitric oxidepromotes mammary tumor growth and metastasis by stimulating tumourcell migration, invasiveness and angiogenesis. Int. J. Cancer, 86, 30–39.

11.Bani,D., Mansini,E., Bello,M.G., Bigazzi,M. and Sacchi,T.B. (1995)Relaxin activates the L-arginine-nitric oxide pathway in human breastcancer cells. Cancer Res., 55, 5272–5275.

12.Adami,A., Crivellente,F., De Prati,A.C., Cabalieri,E., Cuzzolin,L.,Tommasi,M., Suzuki,H. and Benoni,G. (1998) Biotransformation andcytotoxic properties of NO-donors on MCF7 and U251 cell lines. Life Sci.,63, 2097–2105.

13.Binder,C., Schulz,M., Hiddemann,W. and Oellerich,M. (1999) Inductionof inducible nitric oxide synthase is an essential part of tumor necrosisfactor-(-induced apoptosis in MCF-7 and other epithelial tumor cells. Lab.Invest., 79, 1703–1712.

A.M.Simeone et al.

576

by guest on March 24, 2014

http://carcin.oxfordjournals.org/D

ownloaded from

14.Umansky,V., Ushmorov,A., Ratter,F., Chlichlia,K., Bucur,M.,Lichtenauer,A. and Rocha,M. (2000) Nitric oxide-mediated apoptosis inhuman breast cancer cells requires changes in mitochondrial functions andis independent of CD95(APO-1/Fas). Int. J. Oncol., 16, 109–117.

15.Simeone,A., Ekmekcioglu,S., Broemeling,L.D., Grimm,E.A. andTari,A.M. (2002) A novel mechanism by which N-(4-hydroxyphenyl)retinamide inhibits breast cancer cell growth: the produc-tion of nitric oxide. Mol. Cancer Ther., 1, 1009–1017.

16.Hussain,S.P., Hofseth,L.J. and Harris,C.C. (2003) Radical causes ofcancer. Nat. Rev. Cancer, 3, 276–285.

17.Dong,Z., Staroselsky,A.H., Qi,X., Xie,K. and Fidler,I.J. (1994) Inversecorrelation between expression of inducible nitric oxide synthase activityand production metastasis in K-1735 murine melanoma cells. Cancer Res.,54, 789–793.

18.Dhar,A., Brindley,J.M., Stark,C., Citro,M.L., Keefer,L.K. andColburn,N.H. (2003) Nitric oxide does not mediate but inhibitstransformation and tumor phenotype. Mol. Cancer Ther., 2, 1285–1293.

19.Le,X., Wei,D., Huang,S., Lancaster,J.R. and Xie,K. (2005) Nitric oxidesynthase II suppresses the growth and metastasis of human cancerregardless of its upregulation of protumor factors. Proc. Natl Acad. Sci.USA, 102, 8758–8763.

20.Um,S.-J., Lee,S.-Y., Kim,E.-J., Han,H.-S., Koh,Y.-M., Hong,K.-J.,Sin,H.-S. and Park,J.-S. (2001) Antiproliferative mechanism ofretinoid derivatives in ovarian cancer cells. Cancer Lett., 174,127–134.

21.Kim,J.H., Tanabe,T., Chodak,G.W. and Rukstalis,D.B. (1995) In vitroanti-invasive effects of N-(4-hydroxyphenyl)-retinamide on humanprostatic adenocarcinoma. Anticancer Res., 15, 1429–1434.

22.Webber,M.M., Bello-DeOcampo,D., Quader,S., Deocampo,N.D.,Metcalfe,W.S. and Sharp,R.M. (1999) Modulation of the malignantphenotype of human prostate cancer cells by N-(4-hydroxyphenyl)retinamide (4-HPR). Clin. Exp. Metastasis, 17,255–263.

23.Quader,S.T.A., Bello-DeOcampo,D., Williams,D.E., Kleinman,H.K. andWebber,M.M. (2001) Evaluation of the chemopreventive potential ofretinoids using a novel in vitro human prostate carcinogenesis model.Mutat. Res., 496, 153–161.

24.Sharp,R.M., Bello-DeOcampo,D., Quader,S.T.A. and Webber,M.M.(2001) N-(4-hydroxyphenyl)retinamide (4-HPR) decreases neoplasticproperties of human prostate cells: an agent for prevention. Mutat. Res.,496, 163–170.

25.Ferrari,N., Morini,M., Pfeffer,U., Minghelli,S., Noonan,D.M. andAlbini,A. (2003) Inhibition of Kaposi’s sarcoma in vivo by fenretinide.Clin. Cancer Res., 9, 6020–6029.

26.Saavedra,J.E., Shami,P.J., Wang,L.Y., Davies,K.M., Booth,M.N.,Citro,M.L. and Keefer,L.K. (2000) Esterase-sensitive nitric oxide donorsof the diazeniumdiolate family. In vitro antileukemic activity. J. Med.Chem., 43, 261–269.

27.Brunner,N., Thompson,E.W., Spang-Thomsem,M., Rygaard,J., Dano,K.and Zwiebel,J.A. (1992) lacZ transduced human breast cancer xenografts

as an in vivo model for the study of invasion and metastasis. Eur. J.Cancer, 28A, 1989–1995.

28.Yoneda,T., Williams,P.J., Hiraga,T., Niewolna,M. and Nishimura,R.(2001) A bone-seeking clone exhibits different biological properties fromthe MDA-MB-231 parental human breast cancer cells and a brain-seekingclone in vivo and in vitro. J. Bone Min. Res., 16, 1486–1495.

29.Formelli,F., Clerici,M., Campa,T., Di Mauro,M., Magni,A., Mascotti,G.,Moglia,D., De Palo,G., Costa,A. and Veronesi,U. (1993) Five-yearadministration of fenretinide: pharmacokinetics and effects on plasmaretinol concentrations. J. Clin. Oncol., 11, 2036–2042.

30.Hagemann,T., Robinson,S.C., Schulz,M., Trumper,L., Balkwill,F.R. andBinder,C. (2004) Enhanced invasiveness of breast cancer cell linesupon co-cultivation with macrophages is due to TNF-alphadependent upregulation of matrix metalloproteases. Carcinogenesis, 25,1543–1549.

31.Simeone,A., Broemeling,L.D., Rosenblum,J. and Tari,A.M. (2003) HER2/neu reduces the apoptotic effects of N-(4-hydroxyphenyl)retinamide (4-HPR) in breast cancer cells by decreasing nitric oxide production.Oncogene, 22, 6739–6747.

32.Balduyck,M., Zerimach,F., Gouyer,V. et al. (2000) Specific expression ofmatrix metalloproteases 1, 3, 9 and 13 associated with invasiveness ofbreast cancer cells in vitro. Clin. Exp. Metastasis, 18, 171–178.

33.Morini,M., Mottolese,M., Ferrari,N., Ghiorzo,F., Buglioni,S.,Mortarini,R., Noonan,D.M., Natali,P.G. and Albini,A. (2000) The alpha 3beta 1 integrin is associated with mammary carcinoma cell meta-stasis, invasion, and gelatinase B (MMP-9) activity. Int. J. Cancer, 87,336–342.

34.Weber,M.H., Lee,J. and Orr,F.W. (2002) The effect of Neovastat (AE-941) on an experimental metastatic bone tumor model. Int. J. Oncol., 20,299–303.

35.Vaccari,M., Silingardi,P., Argnani,A., Horn,W., Giungi,M.,Mascolo,M.G., Grilli,S. and Colacci,A. (2000) In vitro effects offenretinide on cell-matrix interactions. Anticancer Res., 20, 3059–3066.

36.Diel,I.J., Solomayer,E.F., Costa,S.D., Gollan,C., Goerner,R.,Wallwiener,D., Kaufmann,M. and Bastert,G. (1998) Reduction in newmetastases in breast cancer with adjuvant clodronate treatment. N. Engl. J.Med., 339, 357–363.

37.Powles,T., Paterson,S., Kanis,J.A. et al. (2002) Randomized, placebo-controlled trial of clodronate in patients with primary operable breastcancer. J. Clin. Oncol., 20, 3219–3224.

38.Diel,I.J., Solomyer,E.F. and Bastert,G. (2000) Bisphosphonates and theprevention of metstases: first evidences from preclinical and clinicalstudies. Cancer, 88, 3080–3088.

39.Saarto,T., Blomqvist,C., Virkkunen,P. and Elomaa,I. (2001) Adjuvantclodronate treatment does not reduce the frequency of skeletal metastasesin node-positive breast cancer patients: 5-year results of a randomizedcontrolled trial. J. Clin. Oncol., 19, 10–17.

Received March 28, 2005; revised September 5, 2005;accepted September 24, 2005

Effects of 4-HPR and NONO-AM on bone metastatic breast cancer cells

577

by guest on March 24, 2014

http://carcin.oxfordjournals.org/D

ownloaded from