ORIGINAL ARTICLE

Aromatase Inhibitor Exemestane has Antiproliferative Effectson Human Mesothelioma Cells

Daniela Stoppoloni, BSc,* Luisa Salvatori, BSc,*† Annamaria Biroccio, BSc,‡Carmen D’Angelo, BSc,‡ Paola Muti, MD,§ Alessandra Verdina, BSc,* Ada Sacchi, BSc,*

Bruno Vincenzi, MD,� Alfonso Baldi, MD,¶ and Rossella Galati, BSc*

Purpose: The aim of this study was to investigate the expressionand biological activity of aromatase (CYP19A1) in malignant me-sothelioma (MM).Experimental Design: We found CYP19A1 in five human MM celllines using reverse transcription polymerase chain reaction andWestern immunoblots and in a group of samples from patients withMM by immunohistochemistry. Aromatization activity was determinedin MM cells by enzyme-linked immunosorbent assay as a measure ofestradiol (E2) product, in basal condition and after addition of cytokine,prostaglandin-E2, and epidermal growth factor to MM cells. Treatmentof MM cells with exemestane, a CYP19A1 inhibitor, was assessed bycell proliferation kit, cell cycle analysis, and Western blot for caspase,poly(ADP-ribose)polymerase, Bcl-xL, and v-akt murin thymoma viraloncogene homolog (Akt).Results: Biological activity of CYP19A1, already present in basalcondition, was increased in MPP89 and Ist-Mes1 cells after treat-ment with cytokine, in all MM cells on prostaglandin-E2 treatment,and in MPP89, Ist-Mes2, and Ist-Mes1 after addition of epidermalgrowth factor. Treatment of MM cells with exemestane led tosignificant reduction of tumor cell growth, perturbation of cell cycle,caspase activation, poly(ADP-ribose)polymerase cleavage, anddown-regulation of phosphorylation of Akt and Bcl-xL. In tumortissues, we found a cytoplasmic localization of CYP19A1. Byunivariate analysis, overall survival resulted to be strongly influ-enced by high CYP19A1 expression (p � 0.001).Conclusion: These findings show that CYP19A1 is present in MM andthat cell growth can be down-regulated by exemestane. As Akt pathwayand Bcl-xL are implicated in conferring resistance to conventionalchemotherapy, exemestane could open new treatment strategies to beassociated with standard therapy for patients afflicted with MM.

Key Words: Mesothelioma, Aromatase, Exemestane.

(J Thorac Oncol. 2011;6: 583–591)

Aromatase (CYP19A1) is the cytochrome P450 enzymecomplex that converts C19 androgens to C18 estrogens.

The human CYP19A1 gene, located in the 21.2 region on thelong arm of chromosome 15 (15q21.2), spans a region thatconsists of a 30 kb coding region and a 93 kb regulatoryregion. Its regulatory region contains at least 10 distinctpromoters regulated in a tissue- or signaling pathway-specificmanner. Each promoter is regulated by a distinct set ofregulatory sequences in DNA and transcription factors thatbind to these specific sequences. These partially tissue-spe-cific promoters are used in the gonads, bone, brain, vasculartissue, adipose tissue, skin, fetal liver, and placenta forestrogen biosynthesis necessary for human physiology.1 Es-trogens contribute to differentiation and maturation in normallung2 and also stimulate growth and progression of lungtumors.3,4 These biological effects are mediated by estrogenreceptors (ERs), with ER transcripts and proteins reported inmost non-small cell lung cancers.3,4 As in breast, CYP19A1mediates synthesis of estrogens in lung tissues,4,5 and localproduction of estrogens in women and in men could affectlung tumor progression in ER-expressing malignancies.4,6

Cytokines, such as interleukin (IL)-6 and tumor necrosis factor(TNF)-�, have an important role in regulating estrogen synthesisin peripheral tissues. The activity of the CYP19A1 is increasedby IL-6 and TNF-�.7 Multiple factors, including epidermalgrowth factor (EGF) and prostaglandin-E2 (PGE2), would syn-ergistically up-regulate CYP19A1 expression.8–10

Pleural mesothelioma is not just a limiting protectivelayer for lung but a dynamic cellular structure regulatingserial responses to injury, infection, and disease. Mesothelialcells are biologically active because they can sense andrespond to signals within their microenvironment. The devel-opment of malignant mesothelioma (MM) is associated inmost patients with a history of asbestos exposure.11 Researchhas demonstrated that asbestos exposure generates reactiveoxygen species and activates macrophages and other celltypes to produce these compounds, and cytokines and growthfactors.12 Furthermore, the deposition of insoluble amphibolefibers results in a chronic inflammatory state in exposedindividuals.13 The existence of inflammation has been asso-

*Department for the Development of Therapeutic Programs, Regina ElenaCancer Institute; †CNR, Institute of Molecular Biology and Pathology;‡Experimental Chemotherapy Laboratory, Regina Elena Cancer Insti-tute; §Regina Elena Cancer Institute; �Section of Oncology, CampusBiomedico University, Rome, Italy; and ¶Department of Biochemistryand Biophysics, Section of Pathology, Second University of Naples,Italy.

Disclosure: The authors declare no conflicts of interest.Address for correspondence: Rossella Galati, BSc, Regina Elena Cancer

Institute, Via E. Chianesi 53, Rome 00144, Italy. E-mail: galati@ifo.itCopyright © 2011 by the International Association for the Study of LungCancerISSN: 1556-0864/11/0603-0583

Journal of Thoracic Oncology • Volume 6, Number 3, March 2011 583

CORE Metadata, citation and similar papers at core.ac.uk

Provided by Elsevier - Publisher Connector

ciated with up-regulation of the inducible cyclooxygenase-2,leading to an increase in its product PGE2,14 and is associatedwith an increased risk of cancer.15 Even with widespreadasbestos abatement efforts, the increase in MM incidence islikely to continue in Western Europe and the United Stateswell into the next decade, and projections suggest that theincidence will peak around 2020.16 Other factors, such aschemical carcinogens, ionizing radiation, chronic inflamma-tion, and SV40 viral exposure, may contribute to the devel-opment of MM.17 Pleural MM is 10- to 30-fold more com-mon than their peritoneal counterparts.18,19 Regardless of siteof origin, the prognosis is usually poor with a median survivalof 4 to 12 months for pleural tumors20 and less than 1 year forperitoneal tumors.19 Systemic chemotherapy and radiation dolittle to improve the outcome. The resistance to conventionalchemotherapy and most chemotherapeutic agents have beenattributed to activation of v-akt murin thymoma viral onco-gene homolog (Akt) and overexpression of prosurvival Bcl2family members, such as Bcl-xL and Bcl2.21,22 The phospha-tidylinositol 3-kinase/Akt signaling pathway regulates funda-mental cellular processes linked to tumorigenesis, includingcell cycle progression; cell survival, adhesion, motility, andspreading; angiogenesis; glucose homeostasis; and cell andorgan size control.23,24 The Bcl2 family proteins are pivotalregulators of apoptotic cell death that are considered asattractive targets for drug design.25

Thus, it clearly seems that, to improve the clinicaloutcome in the pharmacological treatment of this refractorytumor, drugs directed against novel tumor-specific cellulartargets and/or characterized by a more specific mechanism ofaction are needed.

In this study, we have investigated the possibility thatthe growth of MM cells may be influenced by CYP19A1.NCI-H2452, MPP89, Ist-Mes1, Ist-Mes2, and MSTO-211Hcell lines were chosen as a model. Our experiments show thatMM cell lines, as well as human MM tissues, express ERsand CYP19A1. The data produced suggest a specific role ofthe CYP19A1 in the growth of MM. Furthermore, the anti-proliferative action of exemestane provides strong evidencethat anti-CYP19A1 drugs inhibit tumor growth and may be anovel strategy to sensitize the MM cells to standard therapy.

MATERIALS AND METHODS

Cell LinesThe human MM cell lines MSTO-211H and NCI-

H2452 were obtained from the American Type Culture Col-lection (ATCC) (Rockville, MD). Cells were cultured asmonolayers in flasks using American Type Culture Collectioncomplete growth medium in a humidified atmosphere con-taining 5% CO2 at 37°C. The human MM cell lines Ist-Mes1,Ist-Mes2, and MPP89 were from Genova Institute CultureCollection. Ist-Mes1 and Ist-Mes2 were replaced by Dul-becco modified eagle medium with piruvate supplementedwith 10% fetal bovine serum, glutamine (2 mM) 1% nones-sential amino acids, and antibiotics (0.02 IU/ml penicillin and0.02 mg/ml streptomycin), whereas the established cell line,MPP89, was maintained in Ham’s F10 with 15% fetal bovineserum and supplemented with glutamine (2 mM) and antibi-

otics (0.02 IU/ml penicillin and 0.02 mg/ml streptomycin ina humidified atmosphere containing 5% CO2 at 37°C).

Protein Extraction and Western Blot AnalysisCell lysates were prepared by treating cells with ice-

cold lysis buffer (20mM Tris pH 8, 1% NP40, 10% glycerol,137 mM NaCl, 10 mM ethylenediaminetetraacetic acid, andinhibitor of protease and phosphatase) for 20 minutes fol-lowed by centrifugation at 4°C for 15 minutes. Proteins (80�g) were separated on 10% sodium dodecyl sulfate-polyac-rylamide gel electrophoresis gels and then transferred onpolyvinylidenedifluoride membrane. Membranes were incu-bated with antibodies of interest. Goat anti-mouse/rabbitimmunoglobulin-G horseradish peroxidase-conjugated sec-ondary antibodies (1:3000) (Bio-Rad Laboratories; Hercules,CA) were used. Antibody reaction was visualized using ECLWestern blotting detection reagents (Amersham-Pharmacia,Uppsala, Sweden). Membranes were stripped by incubationin 1 M Tris-HCl (pH 6.8), 10% sodium dodecyl sulfate, and10 mM dithiothreitol for 30 minutes at 55°C, and reprobedwith different antibodies. The blots were, then, reacted withECL Western blotting detection reagents. Actin was used asa loading control. The experiments were performed in tripli-cate. Proteins were probed with antibodies againstCYP19A1(Santa Cruz Biotechnology, Santa Cruz, CA), anti-ER� (F-10) (Santa Cruz Biotechnology, Santa Cruz, CA),anti-ER�1 (Serotec), anti-Akt, and phosphorylation of Akt(p-Akt) (Cell Signaling Technology, Danvers, MA) anti-Bcl-xL (Cell Signaling Technology) anti-caspase-3 (E-8)(Santa Cruz Biotechnology, Santa Cruz, CA), anti-poly-(ADP-ribose)polymerase (PARP) (Cell Signaling Technol-ogy), and with antiactin antibody (Sigma, St. Louis, MO) tonormalize the sample loading and then detected with horse-radish peroxidase-conjugated secondary antibodies fromSanta Cruz Biotechnology, Santa Cruz, CA. The experimentswere done in triplicate.

RNA Isolation and Reverse TranscriptionPolymerase Chain Reaction Assay of CYP19A1

Total RNA was prepared from cultured MPP89, Ist-Mes2, NCI-H2452, MSTO-211H, and Ist-Mes2 using TRIzolReagent (Invitrogen Life Technologies, Paisley, UK) accord-ing to the manufacturer’s protocols as described.26 Briefly,reverse transcription of RNA, for first-strand complementaryDNA (cDNA) synthesis was performed using 4 �g total RNAand 0.5 �g oligo (dT) 12 to 18 primer (Invitrogen LifeTechnologies, Paisley, UK), 10 mM deoxynucleotidetriphos-phate mix (Invitrogen Life Technologies, Paisley, UK) in afinal volume of 12 �l. The reaction was incubated at 70°C for10 minutes and immediately chilled on ice. Primer extensionwas then performed 10 minutes at temperature room and 42°Cfor 2 minutes after addition of first-strand buffer, 10 mMdithiothreitol, and 40 U RNase OUT Recombinant RibonucleaseInhibitor (Invitrogen Life Technologies, Paisley, UK) in a finalvolume of 19 �l. Then 1 �l (200U) SuperScript II ReverseTranscriptase (Invitrogen Life Technologies, Paisley, UK) wasadded and incubated at 42°C for 50 minutes. The reaction wasinactivated by heating at 70°C for10 minutes. cDNA was storedat �20°C.

Stoppoloni et al. Journal of Thoracic Oncology • Volume 6, Number 3, March 2011

Copyright © 2011 by the International Association for the Study of Lung Cancer584

Quantitative polymerase chain reaction (PCR) was con-ducted in a volume of 25 �l containing 40 ng cDNA (1/100dilution of reverse transcription mixture), 1.25 �l ofCYP19A1 primer, and 12.5 �l TaqMan Universal PCR Mas-ter Mix (Applied Biosystems, Foster City, CA) in the follow-ing sequence: 2 minutes at 50°C and denaturation for 10minutes at 95°C followed by 40 cycles of the amplificationstep at 95° for 15 seconds (denaturation) and then at 60°C for60 seconds (annealing/extension) in 96-well plates with theABI PRISM 7000 sequence Detection System (Applied Bio-systems, Foster City, CA). Quantitative PCR for the endog-enous control glyceraldehyde-3-phosphate dehydrogenase(GAPDH) was carried out under the same conditions, using aGADPH Assay on Demand (Applied Biosystems, FosterCity, CA). A standard curve for CYP19A1 was constructedusing serial dilutions (200-40-8-1.6 ng) of a pool of cDNAsfrom MM cells. Results have been analyzed using the Ap-plied Biosystems analysis software and expression levelscalculated from a linear regression of the standard curve.Results are given as CYP19A1 gene expression versusGAPDH expression to correct for differences in the quantityof cDNA used in the PCR reaction. All quantitative PCRreactions for each sample were performed in triplicate.

Cell Stimulation by CytokineSubconfluent MM cells were stimulated with a mixture

of cytokines (cytomix) as described.27 Briefly, cytomix (14ng/ml TNF-�, Alexis Biochemical, Lausen, Switzerland), 1ng/�l IL-1beta (Alexis Biochemical), 27 ng/ml human inter-feron gamma (PBL Biochemical Laboratories, Piscataway,NJ), and 160 ng lipopolysaccharide (Alexis Biochemical)were added to cell medium for 3 hours. Then, the cells wereremoved by trypsin for mRNA extraction. Controls weretreated in the same manner without cytomix. Experiments ineach MM cell line were repeated three times.

CYP19A1 ActivityAs CYP19A1 catalyzes conversion of androstenedione

and testosterone to estrone and estradiol, respectively, itsactivity was determined by production of estradiol aftertreatment with testosterone. Cells were plated in mediumcomplete and the day after were treated with 10 nM testos-terone (Sigma, St. Louis, MO) alone and in association with10 nM PGE2 or EGF (Sigma, St. Louis, MO) for 24 hours.Then, supernatant fluids were harvested and frozen at �80°C.CYP19A1 activity was measured by a competitive enzymeimmunoassay, Estradiol 17� EIA kit (ALPCO Diagnostics,Windham, NH). This kit was used for quantitative determi-nation of estradiol in culture supernatant. This kit was basedon a competitive enzyme immunoassay for quantitative de-termination of estradiol in culture supernatant.

Clinical Data and Tumor Specimen AcquisitionAll patients were treated at the Second University of

Naples between 1980 and 1996. Clinical data were obtainedby retrospective chart review. Survival was determined fromthe date of initial surgery. Indeed, surgery/biopsy was the firststep of diagnosis in all patients. As a consequence, surgery/biopsy and diagnosis were overlapped. Follow-up was avail-

able for all patients. Two subjects who died of causes otherthan mesothelioma during the follow-up period were ex-cluded from the study. All patients were treated at least withcytoreductive surgery, and 13 patients were then treated withradiotherapy or chemotherapy. Tissues from 29 MM speci-mens (16 epithelioid, six sarcomatoid, and seven biphasicmesothelioma) obtained from open biopsies or pleurectomieswere collected and fixed in 10% formalin before beingembedded in paraffin.

HistologyThe formalin-fixed, paraffin-embedded samples were

sectioned at 5 �m and stained with hematoxylin and eosin.The histological diagnosis was reexamined by a pathologist(A.B.) according to the World Health Organization (Histo-logical typing of tumors, 2004). In addition, the most repre-sentative blocks were selected to be cut into new 5-�m-thicksections for immunohistochemical studies.

ImmunohistochemistryAll 29 cases have been assessed by immunohistochem-

istry for the presence of CYP19A1. Sections from eachspecimen were cut at 5 �m, mounted on glass, and driedovernight at 37°C. All sections were then deparaffinized inxylene, rehydrated through a graded alcohol series, andwashed in phosphate-buffered saline (PBS). PBS was usedfor all subsequent washes and for antibody dilution. Endog-enous peroxidase activity was blocked by 5% hydrogenperoxide. The goat antihuman CYP19A1 (CYP19A1A) anti-body C-16 (Santa Cruz Biotechnology Inc., Santa Cruz, CA)was applied at 4°C for 12 hours at the dilution of 1:100 afterantigen retrieval in citrate buffer in a pressure cooker for 5minutes. The optimal working dilution was defined on thebasis of titration experiments. Then, the sections were immu-nostained with the streptavidin-biotin system (Dako, Carpin-tera, CA), using diaminobenzidine as the final chromogen andhematoxylin as the nuclear counterstain. Negative controlsfor each tissue section were prepared by leaving out theprimary antibody. A suitable positive control was run witheach set of slides. All samples were processed under the sameconditions. Cytoplasmic CYP19A1 staining was determinedbased on staining intensity (0, below the level of detection; 1,weak; 2, moderate; and 3, strong).

In Vitro Cytotoxicity Assays and FlowCytometric Analysis

The in vitro drug sensitivity was assessed by CellProliferation kit (XTT) (Roche Molecular Biochemicals, In-dianapolis, IN), a colorimetric method based on the tetrazo-lium salt (XTT), using the manufacturer’s instructions. Cellswere seeded at 2500 to 20,000 cells/well in 96-well flat-bottomed plate (Corning Inc., Corning, NY) to allow expo-nential growth for the 3 days of the assay to give an absor-bance of 1.0 to 2.2. The optimum number of cells required toreach an absorbance between 1.0 and 2.2 was determined foreach cell line (data not shown). In a typical experiment, cellswere trypsinized, seeded in 96-well plates, and allowed torecover for 24 hours before the addition of exemestane(Sequoia research product, UK). Drug concentrations ranged

Journal of Thoracic Oncology • Volume 6, Number 3, March 2011 Aromatase Inhibitor Exemestane in Mesothelioma

Copyright © 2011 by the International Association for the Study of Lung Cancer 585

from 2.8 � 10�4 to 2.8 � 10�1 mM. All experimental pointswere quantified fivefold. Every single point was comparedwith the respective control with the same amount of DMSO.All experiments were repeated three times. The assay wasdeveloped after 48-hour incubation, and the absorbance wasthen measured.

Cell cycle analysis was performed by flow cytom-etry. Cells were washed in PBS and fixed in 70% ethanolin PBS; 1 � 106 cells were pelleted and resuspended in astaining solution (5.0 �g/ml propidium iodide, 75 KU/mlRNase A in PBS) for 30 minutes at room temperature inthe dark and analyzed by flow cytometry using FACScali-bur (Becton-Dickinson Immunocytometry System-BDIS,San Jose, CA). For each analysis, 20,000 events werecollected. Cell cycle distribution and percentage of apo-ptotic cells were analyzed using Cell Quest (BDIS) andModFit LT (Verity Software House, Topsham, ME).

Statistical AnalysisComparisons of treatment outcome were tested for statis-

tical differences using the Student’s t test for paired data.Statistical significance were assumed at a p value of �0.05.

Fisher’s exact test was used to assess relationshipbetween ordinal data (correlation matrix between immuno-staining parameters). An univariate survival analysis for eachprognostic variable on overall survival was estimated accord-ing to the Kaplan-Meier method. The terminal event wasdeath attributable to cancer. The statistical significance of thedifferences in survival distribution among the prognosticgroups was evaluated by the log-rank test. p values less than0.05 was regarded as statistical significant in two-tailed tests.SPSS software (version 10.00, SPSS, Chicago, IL) was usedfor statistical analysis.

RESULTS

CYP19A1 and ER Expression in MM Cell LinesCYP19A1, ER�, and ER� protein expressions were

assessed by Western blotting in MM cells, using as a controlthe breast cancer cell line MCF-7. In particular, CYP19A1was detected in all MM cell lines, as well as in MCF-7 cells,displaying a molecular size of 54 kDa6 (Figure 1A).

The classic 67 kDa and a variant 46 kDa of ER�(Figure 1B) were expressed in MM cell lines and in MCF-7cells.4 A variant ER� 46 was identified in MCF-7 in which itwas coexpressed with the ER� 66.28 Interestingly, expressionof ER� 46 has been also reported in osteoclasts29 and endo-thelial cells.30,31 No information exists on the function ofER� 46 in epithelia breast cancer cells but is known that ER�66/ER� 46 ratio change with the cell growth status of theMCF-7 cell lines.28 For this study, the MM and MCF-7 cellextract from nonconfluent (20% confluence) cells growing innormal serum were evaluated for ER� protein content byWestern blot analysis. Results showed an excess of the ER�46 compared with ER� 66. The ER� forms at predominantly59 kDa (Figure 1C) were detected in all cell lines. To furtherconfirm these findings, reverse transcription PCR was used todetect CYP19A1 in MM cell lines (Figure 1D).

CYP19A1 Activity in MM Cell LinesThe CYP19A1 activity was measured in the medium of

MM cell lines before and after the addition of 10 nMtestosterone. With the exception of MPP89 and MSTO-211H,difference between treatment and control was not observed,leading us to believe that the levels of testosterone in themedium were sufficient to stimulate the activity ofCYP19A1. Subsequently, on testosterone treatment, wetested the effect of PGE2 or EGF addition, on CYP19A1activity. Figure 2A shows a significant increase of CYP19A1activity in MPP89, Ist-Mes2, and Ist-Mes1 after EGF andPGE2 stimulation. PGE2 induced an increase in CYP19A1activity in NCI-H2452 and MSTO-211H, whereas in thesame cell lines, EGF had no effect. The effect of cytokines on

FIGURE 1. Aromatase and estrogen receptors are expressedin human mesothelioma cells. A, Using gel electrophoresiswith Western blot, CYP19A1 expression is shown in five rep-resentative malignant mesothelioma (MM) cell lines. Breastcancer cell line MCF-7 with known expression of CYP19A1,positive control. MPP89, Ist-Mes2, NCI-H2452, MSTO-211H,and Ist-Mes1 MM cell lines as noted. B, Using electrophore-sis and Western blot, ER� expression is shown in five repre-sentative MM cell lines and MCF-7, positive control withknown high expression of Er�. The classical 67 KDa ER�, aswell as a 46 KDa variant, occurs in MCF-7and MM cell lines.C, Using electrophoresis with Western blot, expression ofER� is shown in MM cell lines and MCF7, positive control.Bottom, actin loading controls. Multiple signals wereachieved by stripping and reprobing the same blot. D, Re-verse transcription polymerase chain reaction (RT-PCR) am-plification for assay of aromatase mRNA from MCF7and MMcells. mRNA CYP indicates aromatase gene expression versusGAPDH. The standard deviation expresses the result of threedifferent quantization. Note that the same representativeMM cancer cell lines were used for RT-PCR and for Westernimmunoblots.

Stoppoloni et al. Journal of Thoracic Oncology • Volume 6, Number 3, March 2011

Copyright © 2011 by the International Association for the Study of Lung Cancer586

the CYP19A1 mRNA level and activity of CYP19A1 weretested in MM cell lines. Cytokines increased CYP19A1mRNA in MPP89 and Ist-Mes2 after 3 hours of treatment. InIst-Mes1, the effect was observed already after 1 hour (Figure2B). In all cell lines, the CYP19A1 activity was measuredafter 3 hours of treatment with cytomix. At this time, a

TABLE 1. Characteristics of the Patients Enrolled in theStudy

Patients Histology

Score Aromatase(0 � Absent; 1 � Low;

and 2 � High)Survival

Time (mo)

1 Epithelial 1 18

2 Epithelial 0 15

3 Epithelial 2 6

4 Epithelial 2 1

5 Epithelial 1 12

6 Epithelial 2 10

7 Epithelial 0 18

8 Epithelial 2 9

9 Epithelial 2 7

10 Epithelial 1 14

11 Epithelial 0 13

12 Epithelial 2 4

13 Epithelial 2 3

14 Epithelial 2 5

15 Epithelial 2 10

16 Epithelial 1 14

17 Sarcomatoid 0 9

18 Sarcomatoid 2 3

19 Sarcomatoid 2 5

20 Sarcomatoid 2 4

21 Sarcomatoid 2 8

22 Sarcomatoid 0 16

23 Biphasic 2 6

24 Biphasic 2 4

25 Biphasic 2 1

26 Biphasic 2 11

27 Biphasic 2 9

28 Biphasic 2 15

29 Biphasic 2 8

level after treatment with cytomix at 1, 2, and 3 hours. Lev-els of CYP19A1 mRNA were compared with that in MM cellsnot treated with cytomix. Values were reported as means �SD of three independent experiments, and p value indicatessignificant difference (p � 0.05) of the treatment with cyto-kines versus control calculated by Student’s t test. C, Effectof cytokines in MM cell lines was measured by production ofestradiol after treatment with cytomix at 3 hours. Aromataseactivity in MM cells was measured by production of estradiolafter treatment with cytomix at 3 hours. See text for details.Values were reported as means � SD of three independentexperiments, and p values indicate significant difference(p � 0.05) of the treatment with cytokines versus controlcalculated by Student’s t test.

FIGURE 2. Aromatase activity occurs in human malignantmesothelioma (MM) cells. A, Aromatase activity in MM cellswas measured by production of estradiol after treatmentwith testosterone. The greatest increment in the productionof estradiol occurred after 48 hours treatment with 10nmol/L testosterone and prostaglandin-E2 (PGE2) in all MMcell lines. Testosterone in combination with epidermalgrowth factor (EGF) induced a statistically significant in-crease in aromatase activity in MPP89, Ist-Mes2 and Ist-Mes1. Estradiol standards were provided by the manufac-turer, and controls included cells grown but not treated withtestosterone. Values were reported as means � SD of threeindependent experiments, and asterisks indicate significantdifference (p � 0.05) of the treatment with EGF or PGE2versus testosterone calculated by Student’s t test. B, Effect ofcytokines in MM cell lines was measured by CYP19A1 mRNA

Journal of Thoracic Oncology • Volume 6, Number 3, March 2011 Aromatase Inhibitor Exemestane in Mesothelioma

Copyright © 2011 by the International Association for the Study of Lung Cancer 587

significant increase in CYP19A1 activity was determined inIst-Mes1 and MPP89 (Figure 2C).

Analysis of CYP19A1 Expression in HumanMesothelioma Samples

Table 1 presents the characteristics of the patientsenrolled in this study and summary results from immunohis-tochemical analysis of the 29 mesothelioma specimens. His-tologically, tumor specimens contained 16 epithelioid, sixsarcomatoid, and seven biphasic mesothelioma. CYP19A1staining was either negative or cytoplasmic. Figure 3A pre-sents some typical immunohistochemical staining forCYP19A1. Normal mesothelium exhibited a weak positivityfor CYP19A1 (data not shown). By rank correlation matrix,correlation has been recorded between immunohistochemicalparameters and mesothelioma histology type. These resultsare summarized in Figure 3B. By univariate analysis, overallsurvival seems to be influenced by CYP19A1 expression. Themedian survival in patients with low to absent CYP19A1expression was longer than in those patients with highCYP19A1 expression (p � 0.001). Interestingly, the histo-logic type did not influence the overall survival in our patients’population. These data are summarized in Figure 3C. Figure 4depicts Kaplan-Meier survival plots for all patients showing a

statistically significant association between high expression ofCYP19A1 and poor outcome (p � 0.001).

In Vitro Cytotoxicity Assays and FlowCytometric Analysis

To assess direct effects on an CYP19A1 inhibitor onmesothelioma cell proliferation, tumor cells were treatedwithout or with exemestane (doses ranging from 0 to 2.8 �10�1 mM) in vitro. Cell growth of MM cells was determinedby the cell proliferation kit in MPP89, Ist-Mes1, and Ist-Mes2. In these cell lines, CYP19A1 activity was up-regulatedby PGE2 and EGF and, therefore, similar to experimentalmodels of breast cancer in which aromatase inhibitors areeffective. Figure 5A shows the effect of dose-dependentexemestane on the survival of MM cells. To analyze whetherinhibition of cell growth by treatment with exemestane inmesothelioma cell lines was accompanied by alterationsin cell cycle distribution, we analyzed the percentage of cellsin the different stages of cell cycle by flow cytometry. Figure5B shows the histograms of DNA content in MPP, Ist-Mes1,and Ist-Mes2 cells treated with 2.8 � 10�1 mM of exemes-tane for 48 hours. Analysis of cell percentages in the differentphases of the cell cycle revealed that exemestane induces anaccumulation of cells in the S-G2/M proliferative compart-ments of the cell cycle with the concomitant decrease in theG0/G1 phase in all the lines used. Moreover, a fraction of cellpopulation resides to the sub-G1 compartment, suggestingthat exemestane also induced cell death. Western blot anal-ysis showed that there was activation of caspase-3 and PARPin MM treated with 2.8 � 10�1 mM of exemestane (Figure5C). Taken together, exemestane induces apoptosis in MMcells. Figure 5C shows a significant decreasing of Bcl-xL inboth cell lines and strongly inhibited constitutive p-AKTin Ist-Mes-1 and MPP89. No basal p-AKT was detected inIst-Mes-2 cell line.26

DISCUSSIONMM is one of the most lethal human tumors. Recent

randomized studies on the treatment of mesothelioma withcombined chemotherapy demonstrated a survival benefit

FIGURE 3. Aromatase expression by immunohistochemistryis present in most archival malignant mesothelioma (MM)tumor specimens. A, Immunohistochemical staining of aro-matase in human pleural mesothelioma. 1 � strong cyto-plasmic expression (original magnification, �200); 2 � neg-ative staining (original magnification, �200). B, Correlationbetween aromatase expression and histotype. C, Survivaland pathological and immunohistochemical parameters inpatients with MM in an univariate analysis.

FIGURE 4. Kaplan-Meier survival plot for aromatase.

Stoppoloni et al. Journal of Thoracic Oncology • Volume 6, Number 3, March 2011

Copyright © 2011 by the International Association for the Study of Lung Cancer588

when a combination of cisplatin and antifolate drugs wasused32,33 and when surgery was supplemented by postopera-tive radiotherapy in patients who underwent incomplete re-section.34 Unfortunately, none of those forms of treatmenthad a significant impact on the progression and outcome ofmesothelioma, and new therapeutic approaches must be in-vestigated for a more successful treatment of this disease.In this scenario, we demonstrated the presence of CYP19A1in MM cell lines and in a group of samples from patients withMM. Furthermore, ER� and ER� were also detected in MMcell lines. This finding was confirmed by evidences support-ing a role of ER� expression in MM.35 Over the last decade,many studies have been carried out to identify potentialCYP19A1 stimulatory factors: IL-6 was the most potentfactor detected that could stimulate CYP19A1 activity.36

Other cytokines and growth factors can stimulate CYP19A1activity.37 Notable, in our experimental setting, the stimula-tion with cytomix induced an increase of CYP19A1 mRNAlevels. It is well known that expression of the CYP19A1 geneis regulated in a tissue-specific manner by the use of alterna-tive promoters.38 PGE2 is thought to be an important regu-lator of CYP19A1 gene expression.39 Interestingly, we foundthat PGE2 increased CYP19A1 activity level in all five MMcell lines. A growing number of reports on the tumor biologyof MMs have identified many cytokines involved in tumorgrowth or the spread of this disease.40–44

The MM cell lines were capable of releasing a consti-tutively high amount of IL-6 (�1100 pg/ml/supernatant ofconfluent cultures).45 This could explain the presence ofCYP19A1 in MM cells. On the other hand, previous studieshave shown that CYP19A1 is present and biologically activein human non-small cell lung cancer cell and that tumorgrowth can be down-regulated by specific inhibition ofCYP19A1.46

To investigate whether these in vitro observations couldhave some clinical relevance, we decided to analyzeCYP19A1 expression in a group of human MM samples. Wefound that CYP19A1 was expressed in the majority of MMsamples as a cytoplasmic protein. Interestingly, the cytoplas-mic expression of CYP19A1 significantly correlated withpoor survival. The World Health Organization classifies MMinto epithelial, sarcomatoid, and biphasic types, each ofwhich can be subdivided further.47 This classification hasimplications for both diagnosis and prognosis. Prognosis ispoor for all MMs, but sarcomatoid MMs have a particularlypoor response rate to treatment.48 Interestingly, a significantassociation between high expression of CYP19A1 and sarco-

the different phases of cell cycle was reported inside the rel-ative histogram. A representative out of three independentexperiments is shown. C, Effect of exemestane (Exe) oncaspase-3, poly(ADP-ribose) polymerase (PARP), Bcl-xL, andphosphorylation of Akt (p-Akt). The cells were cultured in com-plete medium for 24 hours and then treated with exemestane(2.8 � 10�1 mM) for 48 hours. Western blot of total lysatesindicates that the addition of exemestane activates caspasepathways and PARP cleavage and significantly decreases p-Akt andBcl-xL. Ist-Mes2 has no detectable basal p-Akt.

FIGURE 5. Aromatase inhibitor, exemestane, blocks growthof malignant mesothelioma (MM) cell in vitro. A, Exemes-tane reduces growth of MPP89, Ist-Mes1, and Ist-Mes2 celllines in vitro. Both cell lines were treated with exemestanefor 48 hours (72 hours after seeding the cells in flask) atconcentrations ranging from 2.8 � 10�4 to 2.8 � 10�1mM.B, Effect of exemestane on the cell cycle in MM cell lines.Cell cycle analysis after propidium iodide staining was per-formed by flow cytometry in MPP89 (white histograms), Ist-Mes1 (gray histograms), and Ist-Mes2 (black histograms)cells untreated or treated with exemestane at the dose of2.8 � 10�1 mM for 48 hours. The percentages of cells in

Journal of Thoracic Oncology • Volume 6, Number 3, March 2011 Aromatase Inhibitor Exemestane in Mesothelioma

Copyright © 2011 by the International Association for the Study of Lung Cancer 589

matoid MMs was found. These observations strongly suggestthat CYP19A1 plays a role in tumor progression in MM. Itmust be noted that the data described are retrospective andessentially observational in nature, and, therefore, they can-not explain functional mechanisms by which CYP19A1 ac-tually promotes tumor growth. Therefore, additional studiesperformed with larger independent groups of patients areneeded to decipher the role of CYP19A1 expression in MMprogression. Nevertheless, in vitro and in vivo data on patientstrongly suggest that the CYP19A1 can be an interestinglytherapeutic target for MM. To test whether CYP19A1 inhib-itors reduce mesothelioma cell growth, MM cell lines weretreated with exemestane, a type I CYP19A1 inhibitor thatirreversible binds to the CYP19A1 enzyme, causing perma-nent inactivation. In our study, mesothelioma cell prolifera-tion was significantly reduced by exemestane treatment in allthe lines used. Inhibition of cell proliferation well correlatedwith the perturbation of cell cycle induced by exemestane.Specifically, we found that inhibition of CYP19A1 caused ablock of cells in S-G2M phase of cell cycle. We speculatethat exemestane-induced proliferative block was irreversibleas a significative increase of a population with a sub-G1 DNAcontent that are indicative of apoptosis was observed in thetreated compared with untreated cells. Evidence of apoptosisin MM cells on exemestane treatment was observed in West-ern blot. Exemestane induces apoptosis by caspase-dependentpathway. In our experiments, caspase activation and PARPcleavage were detected in MM cell lines. The prosurvivalBclII family members, such as Bcl-xL and Bcl2, are pivotalregulators of apoptotic cell death that are considered asattractive targets for drug design.25,31 Exemestane induces inMM cell lines a down-regulation of Bcl-xL. Recent reportshave indicated that activation of Akt pathway is implicated inconferring resistance to conventional chemotherapy and mul-tiple chemotherapeutic agents (5-fluorouracil, adriamycin,mitomycin C, and cisplatinum) on cancer cells.49,50 Akt ishyperactivated in a wide range of human tumors as a result ofconstitutive activation of growth receptors, mutation of phos-phatidylinositol 3-kinase, and inactivation or loss of phospha-tase and tensin homolog phosphatase.21 Exemestane inducesin MM cell lines a down-regulation of p-Akt. Overall, thesedata lead us to believe that this CYP19A1 inhibitor can beused to sensitize the cells to standard therapy. The combina-tion of chemotherapy with antiestrogen has been alreadyexploited in clinical trials on patients with MM.51 A phase IIItrial comparing maximum debulking surgery and postopera-tive cisplatin, interferon alpha-2b, and tamoxifen immuno-chemotherapy with and without intraoperative photodynamictherapy was described. The results indicate that the aggres-sive multimodal therapy can be delivered for patients withhigher stage MM.52

Finally, it must be underlined that the analysis ofCYP19A1 expression was performed on a small number ofpatients treated heterogeneously; therefore, further studies areurgently required both in vitro and in vivo level to confirmthese observations and eventually to propose exemestane as aconcrete target for MM therapy.

REFERENCES1. Bulun SE, Takayama K, Suzuki T, et al. Organization of the human

aromatase p450 (CYP19A1) gene. Semin Reprod Med 2004;22:5–9.2. Patrone C, Cassel TN, Pettersson K, et al. Regulation of postnatal lung

development and homeostasis by estrogen receptor �. Mol Cell Biol2003;23:8542–8552.

3. Stabile LP, Davis AL, Gubish CT, et al. Human non-small cell lungtumors and cells derived from normal lung express both estrogenreceptor � and � and show biological responses to estrogen. Cancer Res2002;62:2141–2150.

4. Pietras RJ, Marquez DC, Chen H-W, et al. Estrogen and growth factorreceptor interactions in human breast and non-small cell lung cancercells. Steroids 2005;70:372–381.

5. Pezzi V, Mathis JM, Rainey WE, et al. Profiling transcript levels forsteroidogenic enzymes in fetal tissues. J Steroid Biochem Mol Biol2003;87:181–189.

6. Brodie A, Long B, Lu Q. Aromatase expression in the human breast.Breast Cancer Res Treat 1998;49:S85–S91.

7. Purohit A, Newman SP, Reed MJ. The role of cytokines in regulatingestrogen synthesis: implications for the etiology of breast cancer. BreastCancer Res 2002;4:65–69.

8. Watanabe M, Noda M, Nakajin S. Effect of epidermal growth factor andprostaglandin on the expression of aromatase (CYP19A1) in humanadrenocortical carcinoma cell line NCI-H295R cells. J Endocrinol 2006;188:59–68.

9. Richards JA, Brueggemeier RW. Prostaglandin E2 regulates aromataseactivity and expression in human adipose stromal cells via two distinctreceptor subtypes. J Clin Endocrinol Metab 2003;88:2810–2816.

10. Richards JA, Petrel TA, Brueggemeier RW. Signaling pathways regu-lating aromatase and cyclooxygenases in normal and malignant breastcells. J Steroid Biochem Mol Biol 2002;80:203–212.

11. Mossman BT, Kamp DW, Weitzman SA. Mechanisms of carcinogenesisand clinical features of asbestos-associated cancers. Cancer Invest 1996;14:466–480.

12. Kamp DW, Weitzman SA. The molecular basis of asbestos induced lunginjury. Thorax 1999;54:638–652.

13. Mossman BT, Churg A. Mechanisms in the pathogenesis of asbestosisand silicosis. Am J Respir Crit Care Med 1998;157:1666–1680.

14. Vane JR, Mitchell JA, Appleton I, et al. Inducible isoforms of cycloox-ygenase and nitric-oxide synthase in inflammation. Proc Natl Acad SciUSA 1994;91:2046–2050.

15. Ambs S, Hussain SP, Marrogi AJ, et al. Cancer-prone oxyradicaloverload disease. IARC Sci Publ 1999;150:295–302.

16. M. Britton. The epidemiology of mesothelioma. Semin Oncol 2002;29:18–25.

17. Bielefeldt-Ohmann H, Jarnicki AG, Fitzpatrick DR. Molecular pathobi-ology and immunology of malignant mesothelioma. J Pathol 1996;178:369–378.

18. Loggie BW. Malignant peritoneal mesothelioma. Curr Treat OptionsOncol 2001;2:395–399.

19. Sebbag G, Yan H, Shmookler BM, et al. Results of treatment of 33patients with peritoneal mesothelioma. Br J Surg 2000;87:1587–1593.

20. Zellos LS, Sugarbaker DJ. Diffuse malignant mesothelioma of thepleural space and its management. Oncology 2002;16:907–925.

21. Opitz I, Soltermann A, Abaecherli M, et al. PTEN expression is a strongpredictor of survival in mesothelioma patients. Eur J Cardiothorac Surg2008;33:502–506.

22. Varin E, Denoyelle C, Brotin E, et al. Down-regulation of Bcl-xL andMcl-1 is sufficient to induce cell death in mesothelioma cells highlyrefractory to conventional chemotherapy. Carcinogenesis 2010;31:984–993.

23. Hennessy BT, Smith DL, Ram PT, et al. Exploiting the PI3K/AKTpathway for cancer drug discovery. Nat Rev Drug Discov 2005;4:988–1004.

24. Luo HR, Hattori H, Hossain MA, et al. Akt as a mediator of cell death.Proc Natl Acad Sci USA 2003;100:11712–11717.

25. Manion MK, Hockenbery DM. Targeting BCL-2- related proteins incancer therapy. Cancer Biol Ther 2003;2:S105–S114.

26. Stoppoloni D, Canino C, Cardillo I, et al. Synergistic effect of gefitiniband rofecoxib in mesothelioma cells. Mol Cancer 2010;9:27.

27. Stoppoloni D, Cardillo I, Verdina A, et al. Expression of the embryonic

Stoppoloni et al. Journal of Thoracic Oncology • Volume 6, Number 3, March 2011

Copyright © 2011 by the International Association for the Study of Lung Cancer590

lethal abnormal vision-like protein HuR in human mesothelioma: asso-ciation with cyclooxygenase-2 and prognosis. Cancer 2008;113:2761–2769.

28. Flouriot G, Brand H, Denger S, et al. Identification of a new isoform of thehuman estrogen receptor-� (hER-�) that is encoded by distinct transcriptsand that is able to repress hER-� activation function 1. EMBO J 2000;19:4688–4700.

29. Denger S, Reid G, Kos M, et al. ER� gene expression in human primaryosteoblasts: evidence for the expression of two receptor proteins. MolEndocrinol 2001;15:2064–2077.

30. Figtree GA, McDonald D, Watkins H, et al. Truncated estrogen receptor�46-kDa isoform in human endothelial cells: relationship to acuteactivation of nitric oxide synthase. Circulation 2003;107:120–126.

31. Li L, Haynes P, Bender JR. Plasma membrane localization and functionof the estrogen receptor � variant (ER46) in human endothelial cells.Proc Natl Acad Sci USA 2003;100:4807–4812.

32. Vogelzang NJ, Rusthoven JJ, Symanowski J, et al. Phase III study ofpemetrexed in combination with cisplatin versus cisplatin alone inpatients with malignant pleural mesothelioma. J Clin Oncol 2003;21:2636–2644.

33. van Meerbeeck JP, Gaafar R, Manegold C, et al. Randomized phase IIIstudy of cisplatin with or without raltitrexed in patients with malignantpleural mesothelioma: an intergroup study of the European Organisationfor Research and Treatment of Cancer Lung Cancer Group and theNational Cancer Institute of Canada. J Clin Oncol 2005;23:6881–6889.

34. Weder W, Stahel RA, Bernhard J, et al. Multicenter trial of neo-adjuvantchemotherapy followed by extrapleural pneumonectomy in malignantpleural mesothelioma. Ann Oncol 2007;7:1196–1202.

35. Pinton G, Brunelli E, Murer B, et al. Estrogen receptor-beta affects theprognosis of human malignant mesothelioma. Cancer Res 2009;69:4598–4604.

36. Reed MJ, Coldham NG, Patel SR, et al. Interleukin-1 and interleukin-6in breast cyst fluid: their role in regulating aromatase activity in breastcancer cells. J Endocrinol 1992;132:R5–R8.

37. Zhao Y, Nichols JE, Bulnn SE, et al. Aromatase P450 gene expressionin human adipose tissue. J Biol Chem 1995;270:16449–16457.

38. Simpson ER, Mahendroo MS, Means GD, et al. Aromatase cytochromeP450, the enzyme responsible for estrogen biosynthesis. Endocr Rev1994;15:342–355.

39. Zhao Y, Agarwal VR, Mendelson CR, et al. Estrogen biosynthesisproximal to a breast tumour is stimulated by PGE2 via cyclic AMPleading to activation of promoter II of the CYP19A1 (aromatase) gene.Endocrinology 1996;137:5739–5742.

40. Fitzpatrick DR, Peroni DJ, Bielefeldt-Ohmann H. The role of growthfactors and cytokines in the tumorigenesis and immunobiology ofmalignant mesothelioma. Am J Respir Cell Mol Biol 1995;12:455–460.

41. Versnel MA, Claesson-Welsh L, Hammacher A, et al. Human malignantmesothelioma cell lines express PDGF �-receptors whereas culturednormal mesothelial cells express predominantly PDGF �-receptors.Oncogene 1991;6:2005–2011.

42. Lee TC, Zhang Y, Aston C, et al. Normal human mesothelial cells andmesothelioma cell lines express insulin-like growth factor 1 and asso-ciated molecules. Cancer Res 1993;53:2858–2864.

43. Morocz IA, Schmitter D, Lauber B, et al. Autocrine stimulation of ahuman lung mesothelioma cell line is mediated through the transforminggrowth factor a/epidermal growth factor receptor mitogenic pathway.Br J Cancer 1994;70:850–856.

44. Galffy G, Mohammed KA, Dowling PA, et al. Interleukin-8: an auto-crine growth factor for malignant mesothelioma. Cancer Res 1999;59:367–371.

45. Orengo AM, Spoletini L, Procopio A, et al. Establishment of four newmesothelioma cell lines: characterization by ultrastructural and immu-nophenotypic analysis. Eur Respir J 1999;13:527–534.

46. Weinberg OK, Marquez-Garban DC, Fishbein MC, et al. Aromatase inhib-itors in human lung cancer therapy. Cancer Res 2005;65:11287–11291.

47. Travis WD, Colby TV, Corrin B. Histological Typing of Lung andPleural Tumours, 3rd ed. Berlin: Springer, 1999.

48. Neragi-Miandoab S, Richards WG, Sugarbaker DJ. Morbidity, mortal-ity, mean survival, and the impact of histology on survival after pleu-rectomy in 64 patients with malignant pleural mesothelioma. Int J Surg2008;6:293–297.

49. Oki E, Baba H, Tokunaga E, et al. Akt phosphorylation associates withLOH of PTEN and leads to chemoresistance for gastric cancer. Int JCancer 2005;117:376–380.

50. Kai K, D’Costa S, Sills RC, Kim Y. Inhibition of the insulin-like growthfactor 1 receptor pathway enhances the antitumor effect of cisplatin inhuman malignant mesothelioma cell lines. Cancer Lett 2009;278:49–55.

51. Gertler SZ. Combination Chemotherapy and Tamoxifen in TreatingPatients With Solid Tumors Study NCT00002608 by National CancerInstitute (NCI). Available at: www.clinicaltrials.gov.

52. Pass HI, Temeck BK, Kranda K, et al. Phase III randomized trial ofsurgery with or without intraoperative photodynamic therapy and post-operative immunochemotherapy for malignant pleural mesothelioma.Ann Surg Oncol 1997;4:628–633.

Journal of Thoracic Oncology • Volume 6, Number 3, March 2011 Aromatase Inhibitor Exemestane in Mesothelioma

Copyright © 2011 by the International Association for the Study of Lung Cancer 591