Combining EGFR and Mtor Blockade for the Treatment of Epithelioid Sarcoma

Published OnlineFirst August 5, 2011; DOI: 10.1158/1078-0432.CCR-11-0660

Clinical Cancer Cancer Therapy: Preclinical Research

Combining EGFR and mTOR Blockade for the Treatment of Epithelioid Sarcoma

Xianbiao Xie1,5, Markus P.H. Ghadimi1,5, Eric D. Young1,5, Roman Belousov1,5, Quan-sheng Zhu1,5, Juehui Liu1,5, Gonzalo Lopez2,5,6, Chiara Colombo1,5, Tingsheng Peng1,5, David Reynoso3,5, Jason L. Hornick7, Alexander J. Lazar4,5,6, and Dina Lev2,5,6

Abstract Purpose: Molecular deregulations underlying epithelioid sarcoma (ES) progression are poorly under- stood yet critically needed to develop new therapies. Epidermal growth factor receptor (EGFR) is over- expressed in ES; using preclinical models, we examined the ES EGFR role and assessed anti-ES EGFR blockade effects, alone and with mTOR inhibition. Experimental Design: EGFR and mTOR expression/activation was examined via tissue microarray (n ¼ 27 human ES specimens; immunohistochemistry) and in human ES cell lines (Western blot and quantitative reverse transcriptase PCR). Cell proliferation, survival, migration, and invasion effects of EGFR and mTOR activation treated with erlotinib (anti-EGFR small-molecule inhibitor) alone and combined with rapamycin were assessed in cell culture assays. In vivo growth effects of erlotinib alone or with rapamycin were evaluated using severe combined immunodeficient mouse ES xenograft models. Results: EGFR was expressed and activated in ES specimens and cell lines. EGFR activation increased ES cell proliferation, motility, and invasion and induced cyclin D1, matrix metalloproteinase (MMP) 2, and MMP9 expression. EGFR blockade inhibited these processes and caused significant cytostatic ES growth inhibition in vivo. mTOR pathway activation at varying levels was identified in all tissue microarray– evaluable ES tissues; 88% of samples had no or reduced PTEN expression. Similarly, both ES cell lines showed enhanced mTOR activity; VAESBJ cells exhibited constitutive mTOR activation uncoupled from EGFR signaling. Most importantly, combined erlotinib/rapamycin resulted in synergistic anti-ES effects in vitro and induced superior tumor growth inhibition in vivo versus single agent administration. Conclusions: EGFR and mTOR signaling pathways are deregulated in ES. Preclinical ES model–derived insights suggest that combined inhibition of these targets might be beneficial, supporting evaluations in clinical trials. Clin Cancer Res; 17(18); 5901–12. 2011 AACR.

Introduction the cell lineage and cell of origin of ES are unclear. Recent insights into ES molecular underpinnings suggest that First described almost 50 years ago, epithelioid sarcoma loss of INI1 expression, a component of a chromatin (ES) is a malignancy exhibiting both epithelial (keratins remodeling complex regulating transcription machinery, and epithelial membrane antigen) and mesenchymal may be involved in disease inception (2). This distinct (most notably vimentin and CD34) differentiation mar- and uncommon soft tissue sarcoma (STS) histologic kers and remains a biological and clinical enigma (1). Both subtype is further classified into the distal and the more aggressive proximal subtypes, typically develops in young adults (third and fourth decades of life), and exhibits a Authors' Affiliations: Departments of 1Surgical Oncology, 2Cancer Bi- propensity for local recurrence, lymphatic spread, and 3 4 5 ology, Sarcoma Medical Oncology, and Pathology, Adult Sarcoma pulmonary metastasis (3, 4). Although complete surgical Research Center, The University of Texas MD Anderson Cancer Center; 6The University of Texas Graduate School of Biomedical Sciences, Hous- resection of localized tumors results in relatively favor- ton, Texas; and 7Department of Pathology, Brigham and Women's Hos- able 5-year outcome rates (75%–88%), local and systemic pital, Harvard Medical School, Boston, Massachusetts recurrences tend to occur later in the course of disease and Note: Supplementary data for this article are available at Clinical Cancer underlie a dismal 10-year disease-specific survival rate of Research Online (http://clincancerres.aacrjournals.org/). less than 50% (1, 3–5). Unresectable locally advanced Corresponding Author: Dina Lev, Department of Cancer Biology, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, and metastatic ES are usually chemoresistant and conse- Unit 1104, Houston, TX 77030. Phone: 713-792-1637; Fax: 713-563-1185; quentially are generally fatal (3). No major improve- E-mail: [email protected] ments in the therapy of ES have been made since doi: 10.1158/1078-0432.CCR-11-0660 the early 1970s and novel therapeutic approaches are 2011 American Association for Cancer Research. therefore critically needed.

www.aacrjournals.org 5901

Xie et al.

Although not yet extensively investigated in the context Translational Relevance of ES, a recent report has identified enhanced EGFR expression in a cohort of human ES tumor samples (13). This Epithelioid sarcomas (ES), while uncommon, are initial finding has provided the impetus for the studies unfavorable malignancies affecting young adults with presented here, seeking to ascertain EGFR and activated a marked propensity for local recurrence and distant EGFR (pEGFR) expression levels in an independent subset metastasis. Complete surgical resection, the only poten- of human ES specimens. Most importantly, the ES-associ- tially curative therapy, is often not achievable, pointing ated function of EGFR and the effects of an EGFR inhibitor to a critical need for more effective therapeutic strate- on ES growth were evaluated using cellular and xenograft gies. Studies reported here suggest a role for epidermal ES models. Finally, we sought to also identify an EGFR growth factor receptor (EGFR) activation in promoting blockade containing therapeutic combination whose anti- the ES aggressive/metastatic phenotype and show EGFR ES effects were superior compared with single agent EGFR blockade to induce anti-ES effects including abrogated inhibition. cell growth, survival, migration, and invasion. Further- more, mTOR pathway activation was found to com- Materials and Methods monly occur in ES, possibly mediated, at least in part, by reduced or lost PTEN expression. Most importantly, Cell lines and reagents dual targeting of EGFR and mTOR significantly and The human ES cell lines VAESBJ (American Type Culture in vitro in vivo synergistically abrogates ES growth and . Collection) and Epi544 (developed in our laboratory) were Taken together, these data offer new insights into ES maintained and propagated as previously described (4). molecular deregulations and support further evaluation Several additional human cancer cell lines and normal cell of therapeutic combinations targeting EGFR and mTOR primary cultures were used as controls (see Supplementary in the ES clinical setting. Information). Authentication of ES cell lines was conducted utilizing short tandem repeat DNA fingerprinting (STR; Supplementary Data). Recombinant human EGF was The rapidly increasing awareness of malignancy-related purchased from R&D Systems, the EGFR inhibitor erlotinib genetic and epigenetic deregulations has driven the inclu- (Tarceva) from LC Laboratories, and the mTOR inhibitor sion of biologically based treatment approaches as an rapamycin from the University of Texas MD Anderson integral and desired component of anticancer therapeutic Cancer Center (UTMDACC) pharmacy. Commercially strategies (6). Such tumor-tailored, target-orientated available antibodies were used for immunoblot or immu- approaches are aimed specifically at molecules which are nohistochemical detection of: pEGFR (Tyr1173), pERK involved in the initiation and maintenance of a given (Thr202/Tyr404), ERK, pAKT (Ser473), AKT, p4EBP1 cancer and are easily "druggable," thereby offering hope (Thr70), 4EBP1, p70S6K (Thr389), 70S6K, pSRP for achieving the overarching goal of cancer therapy which (Ser235/236), CDK2, CDK4, PARP, and PTEN (all from is to develop drugs that eliminate tumor cells while sparing Cell Signaling); MMP2, MMP9 (R&D Systems); Ki67 normal tissues (7). One molecular target that has attracted (Thermo/Lab Vision); EGFR, cyclin D1, cyclin B1, p27, considerable investigative interest is the epidermal growth and b-actin (Santa Cruz Biotechnology); cyclin A1, cleaved factor receptor (EGFR), an HER receptor family member PARP (Abcam); cleaved caspase 3 (Biocare Medical); and, (HER1/EGFR, HER2/neu, HER3, and HER4; ref. 8). Upon INI1 (BD Transduction Laboratories). ligand binding, EGFR phosphorylation triggers the activation of downstream signaling pathways involved in critical Immunohistochemistry cellular functions such as proliferation, survival, angiogen- A recently established tissue microarray (TMA) contain- esis (9). Functional EGFR dysregulation is frequently ob- ing tissues retrieved from human ES surgical specimens was served in human cancers, with overexpression and used to immunohistochemically evaluate biomarker ex- activation by mutations or autocrine/paracrine growth pression (4). Of note, 8 breast adenocarcinoma specimens factor loops identified in approximately 50% of epithelial were incorporated into this TMA as controls. TMA immu- malignancies, suggesting a pivotal role in tumorigenesis nostaining and xenograft-derived specimen immunohis- and disease progression (10). Similarly, increased EGFR tochemistry were conducted as previously described (14). expression has been described as occurring in several STS For TMA analysis, most biomarkers (see exceptions de- histologic subtypes (11). Consequentially, significant scribed below) were scored by 2 independent observers efforts have been directed toward development and eval- (A.J. Lazar and X. Xie) after excluding samples with insuf- uation of anti-EGFR therapeutic approaches, specifically ficient tumor tissue. Intensity was graded as none (=0), small molecule tyrosine kinase inhibitors and receptor- weak/low (=1), or moderate-high (=2). This system was blocking monoclonal antibodies (12). Multiple preclinical modified for Ki67 and cyclin D1, and the percentage of and even clinical studies have shown therapeutic efficacy of positive tumor nuclei was scored. For INI1, the tumors were EGFR inhibition especially in lung, head and neck, and scored as having nuclear expression intact or absent. As- colorectal carcinoma where such therapies have been in- sessment of PTEN categorized tumors as: 0, no expression; tegrated into the therapeutic armamentarium (12). 1, greatly reduced; and 2, relatively normal or increased

5902 Clin Cancer Res; 17(18) September 15, 2011 Clinical Cancer Research

Targeted Therapy for Epithelioid Sarcoma

expression; intratumoral blood vessel endothelium was used as an internal positive control.

Cellular assays A panel of in vitro cell culture–based assays was used. These included the following: MTS and clonogenicity assays to determine cell growth; propidium iodide (PI) staining and PI/Annexin V staining fluorescence-activated cell-sorting (FACS) analyses to evaluate cell-cycle progression and rate of apoptosis, respectively; and migration and invasion assays to assess these respective cellular pheno- types. Western blot analyses were used to evaluate levels of protein expression and phosphorylation, and quantitative real time PCR (qRT-PCR) was used to determine matrix metalloproteinase (MMP) 2 and MMP9 mRNA expression levels. All these experiments were carried out as we have previously described (15); further information is available as Supplementary Data.

In vivo animal models All animal procedures/care were approved by UTM- DACC Institutional Animal Care and Usage Committee. Animals received humane care as per the Animal Welfare Act and the NIH "Guide for the Care and Use of Laboratory Animals." Animal models were used as previously described (16). Information about the xenograft model and therapeutic schemas are provided in Supplementary Data.

Statistics Cell culture–based assays were repeated at least twice; mean SD was calculated. Cell lines were examined separately. For outcomes that were measured at a single time point, 2-sample Student’s t tests were used to assess the differences. To determine whether the cytotoxic interac- Figure 1. EGFR is highly expressed and activated in human ES. A, tions of erlotinib and rapamycin in ES cells were synergis- representative photographs of ES TMA EGFR and pEGFR (Tyr1173) tic, additive, or antagonistic, drug effects were examined immunostaining (original images were captured at 200 magnification) using the combination index (CI) method of Chou and depicting low and high expressing tumors. B, Western blot analyses showing INI1 loss and pEGFR and EGFR expression in protein extracts of Talalay (17–19). Briefly, the fraction affected (Fa) was human ES cell lines (VAESBJ and Epi544) grown under regular serum calculated from cell viability assays, and CIs were generated containing conditions and compared with expression in normal cell using CalcuSyn software (Biosoft). Differences in xenograft controls (NHDF and HC-SMC) and other cancer cell lines [SW872 size, weight, and lung weight in vivo were assessed using a (liposarcoma), SKLMS1 (leiomyosarcoma), MESA (uterine sarcoma), and 2-tailed Student’s t test. Significance was set at P 0.05. A549 (lung carcinoma)]. C, Western blot analyses depicting enhanced EGFR phosphorylation and downstream signaling in response to EGF. Cells were grown in serum-free media overnight prior to EGF stimulation. Results Interestingly, a high basal level of pAKT was noted in VAESBJ cells even under serum-free conditions. EGF stimulation induced only a minimal EGFR is highly expressed and activated in human ES increase in pAKT expression in VAESBJ cells in contrast to the marked Before evaluating the potential utility of EGFR as an ES pAKT induction observed in Epi544 cells. therapeutic target, we sought to expand previous observations of enhanced EGFR expression in this STS histologic subtype (13). A preconstructed TMA containing human ES on the TMA exhibited EGFR expression. To further deter- specimens retrieved from 27 patients was used for immu- mine whether ES-expressed EGFR is activated we evaluated nohistochemical analysis (Fig. 1A and Supplementary the expression of pEGFR in ES samples: 95% (19, 20) of Table S2). Twenty of the evaluable specimens (77%) EGFR-expressing specimens exhibited positive staining at expressed EGFR (11 moderate to high expression and 9 varying levels whereas no EGFR phosphorylation was low); only 6 did not express EGFR, corroborating the above observed in only 1 sample (Fig. 1A). Taken together, these described previously published observation. Of potential results suggest that the EGFR is both expressed and importance, only 2 of the breast cancer samples included activated in human ES.

www.aacrjournals.org Clin Cancer Res; 17(18) September 15, 2011 5903

Xie et al.

Next, we analyzed the expression of EGFR in 2 human ES hours. As depicted in Figure 2C, EGF enhanced the migra- cell lines available to us. Immunoblotting showed loss of tion and invasion of both ES cell lines tested (P < 0.05). INI1 expression in both of these cell lines as compared with Metalloproteinases play an important role in cellular in- normal human cells (NHDF and HC-SMC) as well as a vasive capacity; consequently, we evaluated the effect of panel of randomly selected cancer cell lines representing EGFR activation on both MMP2 and MMP9 expression. A both sarcomas (SW872, SKLMS1, and MESA) and carcino- significant increase in the mRNA levels of these 2 enzymes mas (A549), a finding which supports their ES origin was noted after 24 hours of EGF stimulation (Fig. 2D). As in (Fig. 1B). Increased EGFR expression was noted in ES cells the case of cyclin D1, MMP2 and MMP9 were found to be as compared with normal controls at levels comparable commonly expressed in human ES specimens upon im- with those observed in several other cancer cell lines munohistochemical staining of TMA (Fig. 2D). In summa- (Fig. 1B). Increased EGFR expression was similarly noted ry, EGFR signaling was found to enhance ES proliferation, in 2 human ES primary cultures (early passages; Supple- cell-cycle progression, migration, and invasion—all func- mentary Fig. S1); both these cell strains exhibit INI1 protein tions critical for tumor growth and progression. loss confirming their ES origin. Under serum-containing growth conditions, pEGFR expression was found in both ES EGFR blockade inhibits ES cell growth in vitro and cell lines although it was more pronounced in VAESBJ as in vivo compared with Epi544. We evaluated whether ES cell– The orally available small molecule EGFR inhibitor expressed EGFR can be further activated with the addition erlotinib was used to evaluate the impact of EGFR blockade of an appropriate ligand. Toward that end, cells were in our ES preclinical models. ES cells were treated with cultured under serum-free conditions overnight and were incremental drug doses for 4 hours; decreases in EGF- then stimulated with EGF; a marked increase in EGFR induced pEGFR were observed even at the lowest dose phosphorylation was observed and a consequent activation (1 mmol/L) tested (Fig. 3A). Furthermore, EGFR blockade of the downstream extracellular signal–regulated kinase resulted in a dose-dependent inhibition of pERK in both (ERK) signaling pathways was found in both cell lines cell lines tested (Fig. 3A). Interestingly, while a marked (Fig. 1C). Interestingly, even under serum-free conditions, effect on the phosphorylation of AKT and its downstream we noted a high basal level of pAKT in VAESBJ cells and mTOR-regulated effectors P70S6 kinase and 4EBP1 was EGF stimulation induced only a minimal additional phos- observed in Epi544 cells, only a limited effect on this phorylation in contrast to the marked pAKT induction signaling pathway was seen in VAESBJ cells (Fig. 3A). observed in Epi544 cells. These initial findings showed Functionally, a dose-dependent decrease in ES cell that EGFR is both highly expressed and functional in growth in response to erlotinib (96 hours) was observed human ES cell lines, thereby providing a rationale for in both cell lines, although more pronouncedly in Epi544 examining the effect of EGFR activation and blockade on (Fig. 3B). Similarly, EGFR blockade inhibited the colony the ES protumorigenic phenotype. formation capacity of ES cells (Fig. 3B). Next, the effects of EGFR inhibition on cell-cycle progression and apoptosis EGFR activation induces the proliferation, cell-cycle were evaluated. Erlotinib treatment (24 hours) resulted in a progression, migration, and invasion of ES cells G1 cell-cycle arrest in both cell lines (Fig. 3B). This effect The functional effects of EGFR activation on ES cells were could, at least in part, be secondary to the observed next examined. A significant (P < 0.05) increase in ES cell decrease in cyclin D1 expression in treated cells growth was observed after EGF stimulation as per MTS (100 (Fig. 3C). Furthermore, PI/Annexin V staining FACS anal- ng/mL/24 h) and clonogenicity (100 ng/mL/14 d) assays ysis showed a significant (P < 0.05) increase in apoptosis (Fig. 2A). EGF induced G1 cell-cycle progression in serum- induced by erlotinib in the ES cells after 96 hours (Fig. 3C). starved ES cells possibly, at least in part, through increase in Concordantly, increased levels of PARP cleavage were the expression of the cell-cycle regulator cyclin D1 (Fig. 2B). noted in both cell lines in response to EGFR blockade Immunohistochemical analysis of the ES TMA also showed (96 hours) further confirming the presence of treatment- increased nuclear cyclin D1 expression in human tumors induced apoptosis. (Fig. 2B and Supplementary Table S2), possibly suggesting To evaluate the effects of erlotinib on ES cell migration that a functional EGFR signaling network is operative in situ and invasion while taking into account the effects of this although additional mechanisms enhancing cyclin D1 inhibitor on ES cell growth and survival, assays were expression might also be at play. Western blot analyses conducted with cells pretreated with erlotinib for 4 hours identified EGF to affect the expression of several additional (only viable cells were used). A significant decrease in G1 cell-cycle regulators; increased cyclin A1, cyclin B1, EGF-stimulated migration and invasion was identified cyclin-dependent kinase (CDK) 2, and CDK4 and de- (P < 0.05); qRT-PCR showed a marked decrease in creased p27 expression were noted, especially in the MMP2 and MMP9 mRNA levels (Fig. 3D). Taken together, Epi544 cell line (Supplementary Fig. S1) these results suggest that EGFR blockade results in decreased Modified Boyden chamber assays were conducted to ES cell proliferation, G1 cell-cycle arrest, enhanced apopto- evaluate effects of EGF on cell migration and invasion. sis, and abrogated migration and invasion capacities. To assure that the effects seen were not secondary to EGF On the basis of the above findings, we next sought to impact on cell proliferation, experiments were limited to 6 evaluate whether the impact of EGFR blockade can also be

5904 Clin Cancer Res; 17(18) September 15, 2011 Clinical Cancer Research

Targeted Therapy for Epithelioid Sarcoma

Figure 2. EGFR activation enhances ES cell growth, cell-cycle progression, migration, and invasion. A, a significant increase in ES cell growth was observed after EGF stimulation as per MTS (top) and clonogenicity assays (bottom). B, EGF stimulation induced G1 cell-cycle progression in serum-starved ES cells (PI staining FACS analysis; top). Western blotting showed increase in cyclin D1 expression in response to EGF (middle). Representative photographs of ES TMA cyclin D1 immunostaining depicting low and high expressing tumors (bottom). C, enhanced ES cell migration and invasion in response to EGF stimulation was identified using modified Boyden chamber assays. D, EGF stimulation enhances MMP2 and MMP9 mRNA expression (qRT-PCR) in ES cells (top). Representative photographs of ES TMA MMP2 and MMP9 immunostaining depicting low and high expressing tumors (bottom). Graphs represent the average of at least 2 repeated experiments SD; *, statistically significant effects (P < 0.05).

www.aacrjournals.org Clin Cancer Res; 17(18) September 15, 2011 5905

Xie et al.

Figure 3. EGFR inhibition exerts anti-ES effects in vitro. A, Western blot analysis identified EGFR phosphorylation inhibition in both cell lines response to erlotinib treatment. Furthermore, EGFR blockade resulted in a dose-dependent inhibition of pERK. Interestingly, whereas a marked effect on AKT phosphorylation and its downstream mTOR-regulated effectors p70S6 kinase (p70S6K) and 4EBP1 was observed in Epi544 cells, only a limited effect on this signaling pathway was seen in VAESBJ cells. B, erlotinib induced a dose-dependent decrease in ES cell growth (MTS; top) which was more pronounced in

Epi544 cells. Similarly, EGFR blockade inhibited the colony formation capacity of ES cells (top middle). Furthermore, erlotinib treatment resultedinaG1 cell-cycle arrest in both cell lines (PI staining FACS analysis; bottom middle) and a dose-dependent decrease in cyclin D1 expression (Western blot; bottom). C, PI/Annexin V staining FACS analyses showed a significant increase in ES cell apoptosis induced by erlotinib (top). Concordantly, increased levelsof PARP cleavage were noted in both cell lines (Western blot; bottom). D, a significant decrease in EGF-stimulated ES cell migration and invasion was observed in response to erlotinib (top). qRT-PCR showed a marked decrease in MMP2 and MMP9 mRNA levels (bottom). Graphs represent the average of at least 2 repeated experiments SD; *, statistically significant effects (P < 0.05).

5906 Clin Cancer Res; 17(18) September 15, 2011 Clinical Cancer Research

Targeted Therapy for Epithelioid Sarcoma

Figure 4. EGFR blockade delays ES growth in vivo. A, VAESBJ xenografts (once average size tumor reached 5 mm in largest dimension) were treated with erlotinib (n ¼ 10) or vehicle control (n ¼ 8). Erlotinib-treated mice exhibited slower growth (left graph; P ¼ 0.002) and a significantly decreased tumor weight at study termination (*, P < 0.05; right graph) as compared with control treated mice. B, immunohistochemical staining confirmed decreased pEGFR expression in erlotinib-treated tumors without significant change in total EGFR expression. Erlotinib-treated tumors exhibited decreased proliferation (measured by Ki67), increased apoptosis [measured via cleaved caspase 3 (CC3)], and decreased cyclin D1, MMP2, and MMP9 expression (original images were captured at 200 magnification). observed in vivo. Using a recently published ES xenograft Taken together, the impact of EGFR blockade on ES animal model resulting from the subcutaneous injection of xenografts mirrors the effects noted in cell culture. However, VAESBJ cells (4), we compared the effect of erlotinib on taking into account the increasing body of clinical evidence, tumor growth in severe combined immunodeficient mice suggesting only minimal effects of compounds inducing with the impact of control vehicle administration. Therapy EGFR blockade as single anticancer agents (20, 21), and our was initiated after tumor establishment (5 mm in greatest data showing tumor growth delay but not tumor abrogation dimension). Mice in both groups were followed for tumor in response to erlotinib suggested that it might be pertinent size and toxicity; treatment was terminated when tumors in to identify EGFR blockade containing therapeutic combina- control group reached an average of 1.5 cm in largest tions that possessed superior anti-ES effects. dimension. Treatment with erlotinib induced significant (P ¼ 0.002) tumor growth delay as compared with control mTOR pathway is commonly deregulated in ES vehicle–treated tumors (Fig. 4A). Average tumor weights We sought to identify additional targets that could be recorded at termination of the study were control group: blocked in combination with EGFR blockade as a treatment 1.19 0.45 g versus erlotinib group: 0.63 0.3 g strategy for ES. Insights from experiments conducted above (P < 0.05; Fig. 4A). have suggested evaluating the potential deregulation of the Next, formalin-fixed, paraffin-embedded tumor sections AKT/mTOR in ES. As depicted in Figure 5A, both ES cell from mice of both study arms were immunohistochemi- lines express pAKT and activated mTOR-regulated down- cally evaluated. Decrease in pEGFR expression in erlotinib- stream targets. However, expression levels are markedly treated tumors confirmed EGFR activation blockade in vivo more pronounced in VAESBJ cells where pAKT expression (Fig. 4B); as expected, no decrease in the total EGFR is shown to be independent of serum conditions (Fig. 5A). expression level was observed. A decrease in the number In addition, 1 of the 2 ES cell strains showed pronounced of Ki67 (a nuclear marker for proliferation)-expressing AKT/mTOR activation (Supplementary Fig. S1). Phosphor- tumor cells and an increase in the number of cleaved ylation of the AKT/mTOR pathway in Epi544 cells is caspase 3 (marker for apoptosis)-positive cells were ob- decreased under serum-free culture conditions, suggesting served in erlotinib-treated samples (Fig. 4B). Furthermore, that its activation might possibly occur via upstream as was found in vitro, erlotinib induced a decrease in cyclin signaling. VAESBJ-expressed pAKT/p4EBP1/pP70S6K levels D1, MMP2, and MMP9 expression in vivo (Fig. 4B). are similar to those observed in sarcoma cells exhibiting

www.aacrjournals.org Clin Cancer Res; 17(18) September 15, 2011 5907

Xie et al.

Figure 5. mTOR pathway is commonly deregulated in ES. A, Western blot analyses showed increased AKT/mTOR activation in ES cells which is more pronounced in VAESBJ cells as compared with Epi544 (top). VAESBJ pAKT expression level is independent of serum conditions (bottom) whereas in Epi544 cells, it is decreased under serum-free culture conditions (bottom). VAESBJ-expressed pAKT/p4EBP1/pP70S6K levels are similar to those observed in sarcoma cells exhibiting loss of PTEN (SW872 and MESA) and while not completely lost, PTEN expression in VAESBJ cells is lower than in Epi544 cells (top). B, representative photographs of ES TMA pSRP and p4EBP1 immunostaining depicting low and high expressing tumors (top). PTEN immunohistochemistry (bottom) showed decreased to absent PTEN expression in the majority of ES samples as compared with levels observed in normal cells and other cancer tissues included on the TMA (PTEN expression level in a breast sample is depicted for comparison). C, rapamycin blocks mTOR signaling in ES cell lines (Western blot; left) and inhibits tumor cell growth (MTS; middle) and colony formation capacity (right).

loss of PTEN (SW872 and MESA; ref. 22) and although not are not just reflective of cell culture artifact, we deter- completely lost, PTEN expression in VAESBJ cells is much mined the expression of p4EBP1 and pSRP (surrogate lower than in Epi544 cells (Fig. 5A). biomarkers for mTOR pathway activation) in our TMA To confirm that mTOR deregulation is a bona fide (Fig. 5B and Supplementary Table S2). p4EBP1 and pSRP common event in human ES and that the results above expression was 100% in both the samples (sample

Figure 6. Combined EGFR and mTOR targeting results in superior anti-ES effects as compared with monotherapy. A, combining erlotinib and rapamycin resulted in inhibition of both EGFR and mTOR signaling (Western blot). B, dual blockade resulted in a markedly superior ES cell growth inhibition as compared with each agent alone (MTS, P < 0.05; top). Isobologram analysis (middle) revealed that the growth inhibitory effect of this therapeutic combination is strongly synergistic in both cell lines, although more pronouncedly in VAESBJ cells, with CI < 0.5 for most combinations tested. Similarly, combination therapy induced

a superior inhibitory effect on colony formation as compared with each agent alone (bottom). C, the most pronounced G1 cell-cycle arrest was found to occur in response to erlotinib and rapamycin combination (PI staining FACS analysis; top) as is also reflected in the most significant decrease in cyclin D1 expression observed in response to this regimen (Western blot; top middle). Similarly, erlotinib and rapamycin combination resulted in a more pronounced apoptotic rate as compared with monotherapy (PI/Annexin V staining FACS analysis; bottom middle) and PARP cleavage (bottom). D, the impact of combined therapy was assessed in vivo using VAESBJ xenografts growing in SCID mice. Both erlotinib and rapamycin as single agents inhibited tumor growth as compared with control (P < 0.05; left and middle top; v, vehicle; R, rapamycin; E, erlotinib; C, combination). Most importantly, combination therapy resulted in significant growth abrogation as compared with erlotinib, rapamycin, or control (P < 0.05). Combination-treated mice exhibited the most significant decrease in tumor weight as compared with all other therapeutic groups (right top). Immunohistochemical analyses showed a decrease in Ki67-positive staining cells and an increase in cleaved caspase 3 (CC3)-positive cells most pronounced in combination-treated tumors (bottom). *, statistically significant effects (P < 0.05).

5908 Clin Cancer Res; 17(18) September 15, 2011 Clinical Cancer Research

Targeted Therapy for Epithelioid Sarcoma

www.aacrjournals.org Clin Cancer Res; 17(18) September 15, 2011 5909

Xie et al.

1: moderate-high ¼ 78%, low ¼ 22%; sample 2: moder- Although each compound individually leads to increased ate-high ¼ 87%, low ¼ 13%). Furthermore, complete loss ES G1 cell-cycle arrest, the most pronounced effect was of PTEN expression in tumor cells was found in 40% of ES observed after combination therapy (24 hours; Fig. 6C). specimens. Interestingly, of those samples expressing Similarly, the most significant decrease in cyclin D1 PTEN (n ¼ 15), 80% exhibited a relative low staining expression was noted in combination-treated cells intensity whereas in only 20% expression was noted that (Fig. 6C). To determine the effect of the therapeutic com- was equivalent to the level observed in intratumoral and bination on cell death PI/Annexin V staining FACS analyses surrounding normal cells (lymphocytes and endothelial were conducted after 96 hours of treatment: no apoptosis cells) used as controls (Fig. 5B and Supplementary was found in response to rapamycin alone whereas similar Table S1). These findings suggest that PTEN deregulation to the findings described above, erlotinib did result in ES is a common molecular aberration in human ES. Similar cell apoptosis. However, a significantly higher apoptotic to ES expression pattern, pSRP and p4EBP1 were com- rate was observed after combination therapy (P < 0.05; monly expressed in the breast cancer samples evaluated: Fig. 6C). This observation was further strengthened by the all 8 specimens were found to express moderate-high marked increase in PARP cleavage also selectively observed levels of pSRP; low p4EBP1 expression was noted in 1 after combination therapy (Fig. 6C). breast cancer sample whereas moderate-high expression Finally, a 4-armed therapeutic study was conducted to was found in 7 samples. None of these samples exhibited determine the comparative impact on tumor growth in vivo, loss of PTEN expression (reduced expression as compared assessing the effect of combination treatment to each drug with normal stroma was noted in 3 cases), suggesting that alone or vehicle control. Both erlotinib and rapamycin as other molecular mechanisms possibly drive AKT/mTOR single agents inhibited VAESBJ xenograft growth as com- activation in these tumors. pared with control (P < 0.05; Fig. 6D). Most importantly, To further confirm that mTOR activation has a function- combination therapy resulted in significant growth abro- al role in ES, we sought to test the effects of the mTOR gation as compared with erlotinib, rapamycin, or control inhibitor rapamycin on tumor cell growth. As shown in alone (P < 0.05). Average tumor weights recorded at Figure 5C, rapamycin blocked mTOR signaling in ES cell termination of the study were control group: 1.21 0.4 lines. A rapamycin dose-dependent decrease (one that g; erlotinib group: 0.53 0.3 g; rapamycin group: 0.38 plateaus after 10 nmol/L) in tumor cell growth and colony 0.09 g; and combination group: 0.15 0.11 g (Fig. 6D). formation capacity was observed in both cell lines. Of note, Immunohistochemical analysis showed a decrease in Ki67- VAESBJ cells exhibited a more pronounced sensitivity to the positive staining cells and an increase in cleaved caspase 3– drug possibly secondary to PTEN loss (Fig. 5C). In sum, our positive cells which was most pronounced in combination data suggest that mTOR pathway activation is a common treatment tumors. These data suggest that EGFR blockade molecular deregulation in human ES contributing to tumor in combination with mTOR inhibition results in significant cell growth, thereby providing the rationale for testing anti-ES effects in vitro and in vivo, a finding of potential therapeutic combinations that target this pathway in con- clinical utility. junction with EGFR blockade. Discussion Combining EGFR and mTOR blockade results in superior anti-ES effects The current study highlights several ES-associated mo- We tested the effects of erlotinib and rapamycin combi- lecular deregulations of potential translational and clinical nations on ES cells. As anticipated and depicted in Figure importance. First, a recent observation (13) identifying 6A, combining these 2 compounds resulted in the inhibi- EGFR overexpression as commonly occurring in ES was tion of both EGFR and mTOR signaling. Of importance, validated and was further expanded to show receptor this therapeutic combination abrogated rapamycin "feed- activation in human ES samples. Moreover, a role for EGFR back" induction of AKT phosphorylation. To determine activation in supporting the malignant phenotype of ES effects on cell growth, ES cells were treated for 96 hours was shown. Most importantly, EGFR blockade was found with increasing doses of erlotinib alone, rapamycin alone, to inhibit the growth of ES cells in vitro and in vivo, inducing or with different dose combinations of these 2 drugs. In arrested cell-cycle progression, enhanced apoptosis, and both cell lines, combined therapy resulted in markedly abrogated tumor cell migration and invasion. While the superior effects as compared with each agent alone limitations of only a small cohort of ES cell lines available (P < 0.05; Fig. 6B). We further determined whether erlo- for testing should be taken into consideration, these tinib and rapamycin interactions were additive or syner- findings are encouraging and possibly support further gistic. Isobologram analysis revealed that the growth investigation of EGFR blockade in the clinical ES context. inhibitory effect of these drugs was strongly synergistic ClinicalTrials.gov (http://clinicaltrials.gov/NCT00148109) in both cell lines although more pronouncedly in VAESBJ reports a currently active phase II clinical trial using an cells, with CI < 0.5 for most combinations tested (Fig. 6B EGFR monoclonal antibody for the treatment of metastatic and Supplementary Table S3). Similarly, combination or locally advanced bone and STSs. It is not certain whether therapy induced a superior inhibitory effect on colony any ES patients were included in this study; however, on formation as compared with each agent alone (Fig. 6B). the basis of observations made here, it might be useful to

5910 Clin Cancer Res; 17(18) September 15, 2011 Clinical Cancer Research

Targeted Therapy for Epithelioid Sarcoma

establish the response rate of this subcohort, perhaps as a possible role this epigenetic modifier plays in the activation separate evaluation. Moreover, it is pertinent that a large of the PI3K/mTOR pathway and in decreased PTEN expres- body of evidence suggests that despite promising responses sion is currently unknown and should be further explored. to monotherapy EGFR manipulation in multiple preclin- Interestingly, response to EGFR inhibitors has been ical epithelial cancer models in which the receptor is over- shown to inversely correlate with PI3K/mTOR signaling expressed (23, 24), the clinical therapeutic benefit is overactivation (33–36). For example, breast cancer cells minimal, even with high levels of EGFR expression within exhibiting enhanced mTOR activation secondary to PTEN the tumor, and is mostly observed in tumors harboring a loss were found to be EGFR blockade resistant; PTEN genetic deregulation of the receptor, that is, amplifications reintroduction into these cells results in increased response and activating mutations (11, 25). Furthermore, even to EGFR inhibition (33). Similarly, loss of PTEN has been tumors harboring such genetic modifications are bound found to predict resistance to EGFR blockade in colorectal to develop resistance to EGFR blockade (26). It is of note cancer, lung cancer, and glioblastoma (34–36); our studies that no EGFR amplifications or gene mutations were expand these observations to now include ES—a tumor previously identified in ES specimens (13). These findings lacking effective systemic treatment. Although erlotinib suggest that enhanced EGFR expression and activation per inhibited the growth of both available ES cell lines, a more se are not sufficient to predict response to EGFR blockade limited response was observed in VAESBJ cells. These cells and that additional molecular deregulations may possibly exhibit uncoupling of EGFR from its downstream PI3K/ decrease the efficacy of EGFR inhibitors as single agents. mTOR signaling, a higher level of constitutive PI3K/mTOR In the past few years, we have learned that rational activation, and consequently survive independent of EGFR. combinations of targeted therapeutics may achieve more Multiple preclinical studies have shown that dual targeting potent antitumor effects and help overcome initial of EGFR and mTOR results in enhanced anticancer effects therapeutic resistance (16); by implication, identifying in several different cancer types (37–39). Such an approach other ES-associated molecular deregulations that can be was shown to be effective in experimental models exhi- targeted in combination with EGFR might be of great biting relative sensitivity to EGFR blockade as a single benefit. agent and also for targeting resistant cells where mTOR The pivotal mTOR pathway, a convergence downstream blockade restores therapeutic sensitivity (37, 39). In light node of a large number of molecular processes including of these findings, several phase I/II trials combining EGFR EGFR signaling, has been shown to play a major role in and mTOR blockade for the treatment of a variety of cancer progression and metastasis (27). mTOR is known to epithelial cancers have recently been initiated (Clinical- act as a master switch for cellular catabolism and anabolism, Trials.gov). In the present study, we found EGFR inhibi- enhancing cancer cell growth and proliferation (27, 28). In tion in combination with mTOR blockade to exert a addition, mTOR can induce cell-cycle progression, enhance synergistic anti-ES effect in vitro and in vivo.Thiscombi- cell survival, and block cell death as dictated by the cellular nation resulted in significant ES cell growth arrest, inhi- context and specific downstream targets (27, 28). Rapamy- bition of cell-cycle progression, enhanced apoptosis, and cin and its derivatives that block the mTORC1 complex have blocked migration and invasion. Together, these data been developed and are currently being evaluated in several support the inclusion of ES patients in future EGFR/ clinical trials for a variety of hematologic and solid malig- mTOR blockade clinical trials. nancies (29, 30). To the best of our knowledge, no previous studies have evaluated mTOR deregulation and function in Disclosure of Potential Conflicts of Interest ES. Our findings identify ubiquitous phosphoinositide-3- No potential conflicts of interest were disclosed. kinase (PI3K)/mTOR pathway activation in human ES specimens (to levels similar to those found in breast cancer Acknowledgments samples) and cell lines possibly mediated, at least in part, by reduced or lost expression of the negative regulator PTEN, We appreciate the expert assistance provided by Mr. Paul Cuevas in the although additional molecular mechanisms are potentially preparation and submission of the manuscript and Ms. Kim Vu for her aid in figure preparation. relevant. Of potential importance, we found that the ES cell line exhibiting PTEN loss (VAESBJ) was most pronouncedly Grant Support sensitive to rapamycin. To that end, it is possible that PTEN loss should be further evaluated as a molecular biomarker The manuscript was supported in part by an NIH/NCI RO1CA138345 (to D. for appropriate patient selection for inclusion in mTOR Lev), an Amschwand Foundation seed grant (to D. Lev), a Chinese scholarship council grant (to X. Xie), a Deutsche Forschungsgemeinschaft training grant (to blockade–based human ES clinical trials. Reduced PTEN M.P.H. Ghadimi), an UTMDACC Physician-Scientist grant (to A.J. Lazar), and an expression in cancer can occur through genomic alterations AIRC fellowship grant (to C. Colombo). MDACC cell line characterization Core or epigenetic modifications such as DNA methylation (31). Facility was further supported by a National Cancer Institute Cancer Center support grant (CA16672). Loss of INI1, the product of the hSNF5/SMARCB1/BAF47 The costs of publication of this article were defrayed in part by the payment gene, occurs in the vast majority of ES (2). INI1 is a of page charges. This article must therefore be hereby marked advertisement in component of the hSWI/SNF chromatin remodeling com- accordance with 18 U.S.C. Section 1734 solely to indicate this fact. plex regulating gene transcription by modulating nucleo- Received March 11, 2011; revised July 21, 2011; accepted July 25, 2011; somal structures in an ATP-dependent manner (32). The published OnlineFirst August 5, 2011.

www.aacrjournals.org Clin Cancer Res; 17(18) September 15, 2011 5911

Xie et al.

References 1. Halling AC, Wollan PC, Pritchard DJ, Vlasak R, Nascimento AG. 22. Zhu QS, Ren W, Korchin B, Lahat G, Dicker A, Lu Y, et al. Soft tissue Epithelioid sarcoma: a clinicopathologic review of 55 cases. Mayo sarcoma cells are highly sensitive to AKT blockade: a role for p53- Clin Proc 1996;71:636–42. independent up-regulation of GADD45 alpha. Cancer Res 2. Hornick JL, Dal Cin P, Fletcher CD. Loss of INI1 expression is 2008;68:2895–903. characteristic of both conventional and proximal-type epithelioid 23. Gridelli C, Bareschino MA, Schettino C, Rossi A, Maione P, Ciardiello sarcoma. Am J Surg Pathol 2009;33:542–50. F. Erlotinib in non-small cell lung cancer treatment: current status and 3. Wolf PS, Flum DR, Tanas MR, Rubin BP, Mann GN. Epithelioid future development. Oncologist 2007;12:840–9. sarcoma: the University of Washington experience. Am J Surg 24. Galizia G, Lieto E, De Vita F, Orditura M, Castellano P, Troiani T, et al. 2008;196:407–12. Cetuximab, a chimeric human mouse anti-epidermal growth factor 4. Sakharpe A, Lahat G, Gulamhusein T, Liu P, Bolshakov S, Nguyen T, receptor monoclonal antibody, in the treatment of human colorectal et al. Epithelioid sarcoma and unclassified sarcoma with epithelioid cancer. Oncogene 2007;26:3654–60. features: clinicopathological variables, molecular markers, and a new 25. Lynch TJ, Bell DW, Sordella R, Gurubhagavatula S, Okimoto RA, experimental model. Oncologist 2011;16:512–22. Brannigan BW, et al. Activating mutations in the epidermal growth 5. Fisher C. Epithelioid sarcoma of Enzinger. Adv Anat Pathol 2006;13: factor receptor underlying responsiveness of non-small-cell lung 114–21. cancer to gefitinib. N Engl J Med 2004;350:2129–39. 6. Morse DL, Gillies RJ. Molecular imaging and targeted therapies. 26. Pao W, Miller VA, Politi KA, Riely GJ, Somwar R, Zakowski MF, et al. Biochem Pharmacol 2010;80:731–8. Acquired resistance of lung adenocarcinomas to gefitinib or erlotinib 7. Blagosklonny MV. Target for cancer therapy: proliferating cells or is associated with a second mutation in the EGFR kinase domain. stem cells. Leukemia 2006;20:385–91. PLoS Med 2005;2:e73. 8. Salomon DS, Brandt R, Ciardiello F, Normanno N. Epidermal growth 27. Bjornsti MA, Houghton PJ. The TOR pathway: a target for cancer factor-related peptides and their receptors in human malignancies. therapy. Nat Rev Cancer 2004;4:335–48. Crit Rev Oncol Hematol 1995;19:183–232. 28. Schmelzle T, Hall MN. TOR, a central controller of cell growth. Cell 9. Citri A, Yarden Y. EGF-ERBB signalling: towards the systems level. 2000;103:253–62. Nat Rev Mol Cell Biol 2006;7:505–16. 29. Chapuis N, Tamburini J, Green AS, Willems L, Bardet V, Park S, et al. 10. Hynes NE, Stern DF. The biology of erbB-2/neu/HER-2 and its role in Perspectives on inhibiting mTOR as a future treatment strategy for cancer. Biochim Biophys Acta 1994;1198:165–84. hematological malignancies. Leukemia 2010;24:1686–99. 11. Dobashi Y, Takei N, Suzuki S, Yoneyama H, Hanawa M, Ooi A. 30. Fouladi M, Laningham F, Wu J, O'Shaughnessy MA, Molina K, Aberration of epidermal growth factor receptor expression in bone Broniscer A, et al. Phase I study of everolimus in pediatric patients and soft-tissue tumors: protein overexpression, gene amplification with refractory solid tumors. J Clin Oncol 2007;25:4806–12. and activation of downstream molecules. Mod Pathol 2004;17: 31. Soria JC, Lee HY, Lee JI, Wang L, Issa JP, Kemp BL, et al. Lack of 1497–505. PTEN expression in non-small cell lung cancer could be related to 12. Ciardiello F, Tortora G. EGFR antagonists in cancer treatment. N Engl promoter methylation. Clin Cancer Res 2002;8:1178–84. J Med 2008;358:1160–74. 32. Carlson M, Laurent BC. The SNF/SWI family of global transcriptional 13. Cascio MJ, O’Donnell RJ, Horvai AE. Epithelioid sarcoma expresses activators. Curr Opin Cell Biol 1994;6:396–402. epidermal growth factor receptor but gene amplification and kinase 33. She QB, Solit D, Basso A, Moasser MM. Resistance to gefitinib in domain mutations are rare. Mod Pathol 2010;23:574–80. PTEN-null HER-overexpressing tumor cells can be overcome through 14. Zou C, Smith KD, Liu J, Lahat G, Myers S, Wang WL, et al. Clinical, restoration of PTEN function or pharmacologic modulation of consti- pathological, and molecular variables predictive of malignant periph- tutive phosphatidylinositol 30-kinase/Akt pathway signaling. Clin Can- eral nerve sheath tumor outcome. Ann Surg 2009;249:1014–22. cer Res 2003;9:4340–6. 15. Jin Z, Lahat G, Korchin B, Nguyen T, Zhu QS, Wang X, et al. Midkine 34. Mellinghoff IK, Wang MY, Vivanco I, Haas-Kogan DA, Zhu S, Dia EQ, enhances soft-tissue sarcoma growth: a possible novel therapeutic et al. Molecular determinants of the response of glioblastomas to target. Clin Cancer Res 2008;14:5033–42. EGFR kinase inhibitors. N Engl J Med 2005;353:2012–24. 16. Ren W, Korchin B, Lahat G, Wei C, Bolshakov S, Nguyen T, et al. 35. Sos ML, Koker M, Weir BA, Heynck S, Rabinovsky R, Zander T, et al. Combined vascular endothelial growth factor receptor/epidermal PTEN loss contributes to erlotinib resistance in EGFR-mutant growth factor receptor blockade with chemotherapy for treatment lung cancer by activation of Akt and EGFR. Cancer Res 2009;69: of local, uterine, and metastatic soft tissue sarcoma. Clin Cancer Res 3256–61. 2008;14:5466–75. 36. Jhawer M, Goel S, Wilson AJ, Montagna C, Ling YH, Byun DS, et al. 17. Chou TC. Preclinical versus clinical drug combination studies. Leuk PIK3CA mutation/PTEN expression status predicts response of colon Lymphoma 2008;49:2059–80. cancer cells to the epidermal growth factor receptor inhibitor cetux- 18. Chou TC, Talalay P. Generalized equations for the analysis of inhibi- imab. Cancer Res 2008;68:1953–61. tions of Michaelis-Menten and higher-order kinetic systems with two 37. Bianco R, Garofalo S, Rosa R, Damiano V, Gelardi T, Daniele G, et al. or more mutually exclusive and nonexclusive inhibitors. Eur J Biochem Inhibition of mTOR pathway by everolimus cooperates with EGFR 1981;115:207–16. inhibitors in human tumours sensitive and resistant to anti-EGFR 19. Reynolds CP, Maurer BJ. Evaluating response to antineoplastic drug drugs. Br J Cancer 2008;98:923–30. combinations in tissue culture models. Methods Mol Med 2005; 38. Chiu CW, Nozawa H, Hanahan D. Survival benefit with proapoptotic 110:173–83. molecular and pathologic responses from dual targeting of mamma- 20. Rich JN, Reardon DA, Peery T, Dowell JM, Quinn JA, Penne KL, et al. lian target of rapamycin and epidermal growth factor receptor in a Phase II trial of gefitinib in recurrent glioblastoma. J Clin Oncol preclinical model of pancreatic neuroendocrine carcinogenesis. J Clin 2004;22:133–42. Oncol 2010;28:4425–33. 21. Moore MJ, Goldstein D, Hamm J, Figer A, Hecht JR, Gallinger S, et al. 39. Wang MY, Lu KV, Zhu S, Dia EQ, Vivanco I, Shackleford GM, et al. Erlotinib plus gemcitabine compared with gemcitabine alone in Mammalian target of rapamycin inhibition promotes response to patients with advanced pancreatic cancer: a phase III trial of the epidermal growth factor receptor kinase inhibitors in PTEN-defi- National Cancer Institute of Canada Clinical Trials Group. J Clin Oncol cient and PTEN-intact glioblastoma cells. Cancer Res 2006;66: 2007;25:1960–6. 7864–9.

5912 Clin Cancer Res; 17(18) September 15, 2011 Clinical Cancer Research

Combining EGFR and mTOR Blockade for the Treatment of Epithelioid Sarcoma

Xianbiao Xie, Markus P.H. Ghadimi, Eric D. Young, et al.

Clin Cancer Res 2011;17:5901-5912. Published OnlineFirst August 5, 2011.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-11-0660

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2011/08/05/1078-0432.CCR-11-0660.DC1

Cited articles This article cites 39 articles, 14 of which you can access for free at: http://clincancerres.aacrjournals.org/content/17/18/5901.full#ref-list-1

Citing articles This article has been cited by 5 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/17/18/5901.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://clincancerres.aacrjournals.org/content/17/18/5901. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.