Hedgehog Pathway Inhibition in Chondrosarcoma Using the Smoothened Inhibitor IPI-926 Directly Inhibits Sarcoma Cell Growth

Published OnlineFirst March 14, 2014; DOI: 10.1158/1535-7163.MCT-13-0731

Molecular Cancer Cancer Biology and Signal Transduction Therapeutics

Veronica T. Campbell1, Puviindran Nadesan3, S. Amanda Ali3, Chang Ye Yale Wang3, Heather Whetstone3, Raymond Poon3, Qingxia Wei3, John Keilty1, Jennifer Proctor1, Lauren W. Wang2, Suneel S. Apte2, Karen McGovern1, Benjamin A. Alman3,5, and Jay S. Wunder4,5

Abstract Hedgehog (Hh) pathway inhibition in cancer has been evaluated in both the ligand-independent and ligand- dependent settings, where Hh signaling occurs either directly within the cancer cells or within the nonmalignant cells of the tumor microenvironment. Chondrosarcoma is a malignant tumor of cartilage in which there is ligand- dependent activation of Hh signaling. IPI-926 is a potent, orally delivered small molecule that inhibits Hh pathway signaling by binding to Smoothened (SMO). Here, the impact of Hh pathway inhibition on primary chondrosarcoma xenografts was assessed. Mice bearing primary human chondrosarcoma xenografts were treated with IPI- 926. The expression levels of known Hh pathway genes, in both the tumor and stroma, and endpoint tumor volumes were measured. Gene expression profiling of tumors from IPI-926–treated mice was conducted to identify potential novel Hh target genes. Hh target genes were studied to determine their contribution to the chondrosarcoma neoplastic phenotype. IPI-926 administration results in downmodulation of the Hh pathway in primary chondrosarcoma xenografts, as demonstrated by evaluation of the Hh target genes GLI1 and PTCH1,aswellas inhibition of tumor growth. Chondrosarcomas exhibited autocrine and paracrine Hh signaling, and both were affected by IPI-926. Decreased tumor growth is accompanied by histopathologic changes, including calcification and loss of tumor cells. Gene profiling studies identified genes differentially expressed in chondrosarcomas following IPI-926 treatment, one of which, ADAMTSL1, regulates chondrosarcoma cell proliferation. These studies provide further insight into the role of the Hh pathway in chondrosarcoma and provide a scientific rationale for targeting the Hh pathway in chondrosarcoma. Mol Cancer Ther; 13(5); 1259–69. 2014 AACR.

Introduction transmembrane receptor Patched (PTCH) is localized in The Hedgehog (Hh) signal transduction pathway the primary cilia and inhibits the activity of the G-pro- plays a critical role in cell differentiation and patterning tein–coupled receptor-like protein Smoothened (SMO), during development, but is inactive in many adult cells which is sequestered within cytosolic vesicles (2). Inhi- (1). There are three Hh ligands—Sonic Hedgehog (SHH), bition of the key signaling protein SMO ensures that GLI Indian Hedgehog (IHH), and Desert Hedgehog (DHH)— transcription factors are held in an inactive form via a that activate the pathway. In the absence of Hh ligand, the complex that contains Suppressor of Fused (SUFU). When Hh ligands bind PTCH, repression of SMO is relieved, resulting in signaling that activates the GLI Authors' Affiliations: 1Infinity Pharmaceuticals, Inc., Cambridge, Massachu- transcription factors, leading to increased expression of setts; 2Department of Biomedical Engineering, Cleveland Clinic, Cleveland, Ohio; 3Program in Developmental and Stem Cell Biology, Hospital for Sick many genes involved in growth and development. A role Children; 4University Musculoskeletal Oncology Unit and Lunenfeld Research for Hh pathway activation has been demonstrated in a 5 Institute, Mount Sinai Hospital; and Division of Orthopaedic Surgery, Depart- broad range of cancers. Ligand-independent signaling ment of Surgery, University of Toronto, Toronto, Ontario Canada occurs in some tumors due to genetic mutations in key Note: Supplementary data for this article are available at Molecular Cancer pathway members such as PTCH (3–5); however, in Therapeutics Online (http://mct.aacrjournals.org/). many tumors, ligand-dependent Hh signaling occurs Current address for B.A. Alman: Department of Orthopaedic Surgery, Duke either directly within the tumor cells, or involves the University, DUMC 2888, 200 Trent Drive, Orange Zone 5th Floor, Durham, cells within the tumor microenvironment. NC 27710. Emerging evidence points to a critical role for Hh Corresponding Author: Jay S. Wunder, University Musculoskeletal Oncology Unit, Mount Sinai Hospital, 476-600 University Avenue, Toronto, signaling in the pathogenesis of chondrosarcoma, a malig- Ontario M5G 1X5, Canada. Phone: 416-586-5995; Fax: 416-586-8611; nant tumor of cartilaginous origin (6). A potential role for E-mail: [email protected] Hh in chondrosarcoma was first implied by studies that doi: 10.1158/1535-7163.MCT-13-0731 evaluated Hh signaling in the biology of chondrocytes 2014 American Association for Cancer Research. within the growth plate. IHH regulates the differentiation

www.aacrjournals.org 1259

Campbell et al.

of chondrocytes during endochondral bone development. Armand Frappier (IAF) (Laval, Quebec, Canada). The During normal bone development, expression of IHH by study was conducted in accordance with current standard prehypertrophic chondrocytes leads to proliferation of operating procedures. Six-week-old CD-1 mice were ran- growth plate chondrocytes. IHH subsequently upregu- domized to receive IPI-926 by oral gavage at 30 mg/kg for lates PTHrP, which inhibits chondrocyte differentiation 14 consecutive days, and the vehicle control mice were and downregulates IHH, thus completing a tightly con- treated with 0.5% methylcellulose, 5% Tween 80 in SWFI, trolled negative feedback loop. Thus, IHH ligand func- USP. On day 14 of dosing, the mice were sacrificed and tions as a central regulator of endochondral ossification tibiofemoral (knee) joints were collected into 10% forma- and regulates the proliferation and differentiation of lin and prepared for histologic evaluation. chondrocytes via a feedback loop essential for mainte- nance of the growth plate (7–9). The importance of Hh Primary chondrosarcoma xenograft mouse model signaling in developing bone was further substantiated in and drug treatment recent studies where young mice treated with a Smo With institutional review board approval, 6 human inhibitor developed severe bone defects, including pre- chondrosarcoma tumor samples were obtained fresh from mature differentiation of chondrocytes, thinning of cor- the operating room at Mount Sinai Hospital (Toronto, tical bone, and fusion of the growth plate (10). Canada). After surgical removal, tumor samples were In chondrosarcoma, signaling through the IHH/PTHrP divided into explants of 5 5 5 mm each, and implanted axis is dysregulated. The ability of PTHrP to inhibit IHH into the subcutaneous tissues on the back of either non- expression is lost, leading to constitutive Hh signaling obese diabetic/severe combined immunodeficient (NOD/ (11). High expression levels of the Hh-regulated genes SCID) or interleukin-2 receptor gamma chain (gamma)- PTCH1 and GLI1 provide evidence for constitutive Hh null NOD/SCID (NSG) mice. For each condition tested in pathway signaling in chondrosarcoma (12). Chondrosar- vivo, 10 mice with tumor explants from each human coma tumor cell proliferation is increased by Hh ligand chondrosarcoma were utilized. In all chondrosarcoma in and decreased by inhibitors of the Hh pathway in cell vivo studies, IPI-926 was administered orally by gavage at culture. Studies using tumor explants found that exposure 40 mg/kg, five times a week on a Monday–Friday sched- to Hh inhibitors downregulated expression of GLI1 and ule. The vehicle control mice were treated with 5% cyclo- PTCH1, supporting the existence of ligand-dependent dextrin in PBS. Treatment was initiated on the day of tumor autocrine Hh signaling in chondrosarcoma. implant, one month after implantation, or after one pas- There are few therapeutic options for patients with sage through the NOD/SCID mice and transplantation chondrosarcoma. These tumors are largely resistant to into NSG mice. Chemotherapies were injected intrave- chemotherapy and radiotherapy, therefore radical surgery nously. Doxorubicin was administered at 5 mg/kg three constitutes the mainstay of therapy in patients. Although times a week on a Monday, Wednesday, and Friday chemotherapy may be beneficial for patients with the rare schedule. Cisplatin was administered at 8 mg/kg once a tumor variant known as mesenchymal chondrosarcoma, week. After sacrifice and removal of the xenograft, tumor there is little evidence of efficacy against dedifferentiated volume was calculated using three-dimensional measure- chondrosarcoma or the most common subtype of conven- ments as previously described (12). For the MIA PaCa-2 tional chondrosarcoma (13–15). In patients with unresect- pancreatic carcinoma cell line in vivo study, tumor cells able disease, there is no treatment option, and the outcome were obtained from American Type Culture Collection is ultimately fatal. The relatively low incidence of chon- (ATCC) and cultured in RPMI-1640 media supplemented drosarcoma has made the study of new therapeutic agents with 10% FBS and 1% penicillin/streptomycin. Cells were difficult. Given the role of the Hh pathway in the biology of prepared in sterile PBS and implanted subcutaneously at a normal chondrocytes, and the emerging evidence for the concentration of 107 cells per Ncr Nude mouse. IPI-926 was importance of the Hh pathway in chondrosarcoma, a administered orally at 40 mg/kg/d for a total of 6 weeks. clinically relevant animal model of chondrosarcoma was Twenty-four hours after the last dose, tumors were placed developed to test the effect of IPI-926, a potent, orally in 10% neutral-buffered formalin for histology or snap delivered small molecule that targets the Hh pathway by frozen in liquid nitrogen for RNA isolation. Sections were specifically inhibiting SMO (5, 16). Studies were conducted prepared by standard methods and stained with hema- in primary chondrosarcoma xenografts to assess the activ- toxylin and eosin for histologic evaluation. ity of IPI-926 on tumor establishment and growth, and to further understand the mechanism of Hh inhibition. Cell line and other primary cancer xenograft mouse models and drug treatment The following cell lines were obtained from ATCC: Materials and Methods Panc1, BxPC3, MIA PaCa-2, L3.6PL, Colo205, SW480, and Two-week oral gavage study in CD-1 mice for bone HT29, and cultured as recommended by the supplier. The plate analysis purchased cell lines were not authenticated by the The in vivo portion of this study was performed by the authors. The cell lines were not cultured beyond 20 pas- Charles River Laboratories Preclinical Services Montreal, sages for any experimental work. The cells were collected Inc. at the Experimental Biology Center of the Institute in sterile PBS and implanted subcutaneously into Ncr

1260 Mol Cancer Ther; 13(5) May 2014 Molecular Cancer Therapeutics

The Hh Pathway and IPI-926 in Chondrosarcoma

Nude mice at 2 106 or 1 107 cells total per mouse. The Technologies). Images were captured at 40 or 63 using primary SCLC (LX22) xenograft cells were a gift from Neil the TissueGnostics imaging system. The presence of cilia Watkins (Monash Institute of Medical Research, Monash in the xenograft sections and TMA spots was assessed in 5 Medical Center, Clayton, Victoria, Australia) received in separate 40 fields of view (FOV). The tumors were 2006 (17). The LX22 primary tumor was not further designated to be positive if cilia were detected in any of authenticated by the authors. The primary colon and the 5 FOV and negative if no cilia were identified. þ ER breast cancer models were established by purchasing fresh surgical specimen material from Asterand, dividing Gene expression microarray analysis the explants each into 5 5 5 mm fragments and Gene expression profiling was carried out on primary implanting them subcutaneously into NOD/SCID mice. chondrosarcoma tumors from mice treated once a day All primary xenograft models were propagated by orally with control or IPI-926 at 40 mg/kg on a Monday– mouse-to-mouse passaging until enough mice were Friday schedule (n ¼ 4/group) for 4 weeks. The tumor implanted for a treatment study (n ¼ 7–10 per group). RNA was used to generate cDNA which was examined on IPI-926 was dosed at 40 mg/kg/d for a total of 21 to 41 Agilent Whole Human Genome (4 44 K) chips. Differ- days. The control mice were treated with 5% cyclodextrin entially expressed genes were determined through the use in PBS. The primary chondrosarcoma xenograft studies of a two-sided moderated t test. Multiple test correction were established and treated as previously described. was applied using the Benjamini–Hochberg method to Tumors were collected and snap frozen for mRNA assess- generate adjusted P values. A cut-off value of 0.05 was ment, following IPI-926 treatment. used to select significantly differentially expressed genes. All calculations were carried out using R version 2.14.1 RNA isolation and quantitative real-time PCR (http://www.r-project.org) limma package (version RNA was isolated from tumor tissue using the RNAqu- 3.10.0). The data have been deposited in NCBI’s Gene eous-4PCR kit (Ambion). Fifty ng of RNA was added Expression Omnibus (18) and are accessible through GEO to each 25 mL reaction in One-step master mix and run Series Accession Number GSE44581. on the 7300RT cycler (Applied Biosystems). RNA was isolated from at least 3 independent experiments, and Chromatin immunoprecipitation analysis was analyzed in separate PCR reactions, with a minimum For chromatin immunoprecipitation (ChIP) analysis, cells of triplicates for each treatment condition and primer were fixed in formaldehyde to cross-link and preserve pro- set. All primer and probe sets were purchased from Applied tein/DNA interactions. The protein/DNA complexes were Biosystems: Human GLI1 (Hs00171790_m1), human IHH sheared into small fragments using sonication to produce (Hs01081801_m1), human SHH (Hs00179843_m1) human DNA fragments of approximately 1 kb. A specific antibody PTCH1 (Hs00970980_m1), human SMO (Hs00170665_m1), against GLI1 or GLI2 and negative control IgG were used human ADAMTSL1 (Hs00364599_ml), human ADAMTSL3 to directly immunoprecipitate the DNA-binding protein. (Hs00324954_ml), human GAPDH (4310884E), murine Gli1 Following immunoprecipitation, cross-linking was reversed (Mm00494645_m1), and murine Gapdh (4352339E). All by adding NaCl, the proteins were removed by treatment genes tested were normalized to glyceraldehyde-3-phos- with Proteinase K, and the DNA was purified. The DNA phate dehydrogenase (GAPDH) and reported as relative was then analyzed by sequencing using the Solexa/Illumina DD expression, or the formula, 2 CT, where CT refers to Genome Analyzer II, and the resultant sequences were com- threshold cycle, was applied to calculate the fold change pared with data from the human genome. of gene expression compared with vehicle control. Explant culture proliferation assay Immunofluorescence staining Human chondrosarcoma tissue organ cultures were Formalin-fixed, paraffin-embedded human chondro- established from primary tumor explants (12). Five sarcoma tissue microarrays (TMA) were purchased from explant tumor tissue cultures, each derived from a differ- Folio Biosciences for cilia staining. Before staining, all ent patient, were treated for 24 hours with 1 mmol/L Hh slides were placed in a 60C oven and baked for 1 hour. antagonist IPI-926, 0.5 mg/mL recombinant Hh ligand Sections were deparaffinized in xylene (3 5 minutes) SHH-N, or 100 mL of ADAMTSL1 conditioned medium, and rehydrated through a graded series of ethanols for 3 as reported elsewhere (19–21). A dose–response experi- minutes each. Heat-induced epitope retrieval was carried ment was conducted with 0, 0.5, 1, 2, 4, and 8 mg/mL out in the DIVA Decloaker solution using the Decloaking purified ADAMTSL1 protein. A proliferation assay was Chamber from Biocare Medical. Sections were blocked for performed 24 hours following treatment (totaling 48 1 hour. Anti-acetylated tubulin (Sigma) at 1:2,000 and anti- hours treatment) with ADAMTSL1 according to the man- CROCC (Sigma) at 1:75 were incubated overnight at 4C. ufacturer’s protocol [Cell Proliferation ELISA, BrdU (col- Sections were washed with Tris-buffered saline Tween-20 orimetric), Roche]. (TBST, 3 5 minutes), and anti-mouse Alexa Fluor 488 and anti-rabbit Alexa Fluor 555 were added at 1:1,000. Statistical analysis Sections were rinsed in TBST (3 5 minutes) and mounted A Student t test or ANOVA with Student t test was run using ProLong Gold anti-fade reagent with DAPI (Life using the JMP software to determine the statistical

www.aacrjournals.org Mol Cancer Ther; 13(5) May 2014 1261

Campbell et al.

Figure 1. Hh pathway members and primary cilia are expressed in human chondrosarcoma tumor cells. RNA was extracted from primary chondrosarcoma xenograft tissues and human Hh pathway expression was measured by qRT- PCR. A, both SHH and IHH were analyzed; however, only IHH mRNA levels were measurable. B, GLI1, PTCH1,andSMO mRNA expression were also detected. C and D, immunofluorescence staining to detect cilia was performed, and shown are representative images from a primary chondrosarcoma xenograft (C) and a human chondrosarcoma TMA spot (D). The basal body of the cilium is detected by rootletin (red) and the microtubules of the cilium are detected by acetylated tubulin (green). Cilia were scored as present if rootletin (red) and acetylated tubulin (green) immunofluorescence were in close proximity. Magnification, 63.

significance of each experiment. The level of significance antibody against acetylated tubulin and the basal bodies at was set at 0.05 for all experiments. The in vivo tumor data the base of the cilium with an antibody against rootletin. are expressed as mean SEM. Immunofluorescence staining demonstrated that primary cilia were present in chondrosarcoma tumor cells in threeof the xenografts available for testing (Fig. 1C). To investigate Results the prevalence of primary cilia in a larger set of primary Primary chondrosarcoma tumor cells contain chondrosarcoma tumor samples, commercially obtained essential Hh pathway signaling components conventional chondrosarcoma TMAs were stained. Micro- The Hh pathway is active in normal chondrocyte devel- scopic analysis of chondrosarcomas revealed that a major- opment and promotes active bone plate growth. Studies in ity (21/28; 75%) exhibited cilia expression (Fig. 1D). juvenile mice demonstrated that Hh pathway blockade leads to premature growth plate closure in the long bones IPI-926 treatment inhibits Hh pathway signaling in (10). Indeed, Hh signaling in normal mouse chondrocytes chondrosarcoma xenografts is inhibited by IPI-926 treatment resulting in a decrease in To determine whether IPI-926 could inhibit Hh signal- proliferating chondrocytes and premature growth plate ing in chondrosarcoma tumor cells, human GLI1, PTCH1, closure (Supplementary Fig. S1). and SMO mRNA levels were assessed in xenograft tumor To investigate whether there is evidence of active Hh tissue following treatment with IPI-926 at 40 mg/kg for a pathway signaling in primary chondrosarcoma xeno- minimum period of 3 weeks. Twenty-four hours follow- grafts, human Hh pathway gene expression was assessed ing the last dose of vehicle or IPI-926 treatment, tumors using quantitative real-time PCR (qRT-PCR). Bulk tumor were harvested and mRNA levels of human Hh genes specimens from surgically resected human conventional were assessed. Treatment with IPI-926 resulted in a 5- to chondrosarcomas were implanted subcutaneously into 12.5-fold downregulation of human GLI1 (P < 0.01) and a NOD/SCID immunocompromised mice. Following tumor 2- to 3-fold downregulation of PTCH1 (P ¼ 0.05) expres- engraftment and growth for at least 6 weeks, tumors were sion, compared with vehicle controls (Fig. 2A and B). A excised and RNA was extracted. Similar to normal chon- lack of corresponding decrease in expression of SMO was drocytes within the growth plate, the chondrosarcoma anticipated, but SMO expression actually increased fol- tumor cells expressed high levels of human IHH mRNA lowing treatment with IPI-926 (Fig. 2B). This increase in compared with SHH mRNA (Fig. 1A). In addition, high SMO expression occurred in response to the effect of IPI- levels of human PTCH1, SMO,andGLI1 mRNA were also 926, which inhibits Hh signaling by targeting SMO at the detected (Fig. 1B). protein level. These data suggest that inhibition of the Hh Chondrosarcoma xenografts were also evaluated for the pathway occurs in the cancer cells themselves. presence of primary cilia, as these provide an important Hh pathway inhibition in reactive stroma was also cellular scaffold for Hh-induced pathway activation (2). evaluated by analyzing murine Gli1 mRNA using spe- Cilia were visualized by staining the microtubules with an cies-specific primers. Murine Gli1 expression levels in

1262 Mol Cancer Ther; 13(5) May 2014 Molecular Cancer Therapeutics

The Hh Pathway and IPI-926 in Chondrosarcoma

chondrosarcoma xenografts were found to be substantial- ly lower than the human levels, but were still signiﬁcantly downregulated in the IPI-926–treated group (Fig. 2C). In addition, the level of murine Gli1 expression was lower in the chondrosarcoma xenografts than that observed in epithelial-derived tumor xenografts, such as the MIA PaCa-2 pancreatic carcinoma xenograft. Although treatment with IPI-926 resulted in downregulation of murine Gli1 expression in both chondrosarcoma and pancreatic carcinoma xenografts (2.4-fold, P < 0.04, and 42-fold, P < 0.0001, respectively), only in the chondrosarcoma cells was human GLI1 expression similarly diminished (P < 0.04). These results are consistent with the histology of chondrosarcoma tumors, which primarily consist of tumor cells embedded in a relatively acellular matrix with few surrounding stromal cells present (Supplementary Fig. S2). The results of human and murine GLI1 expression analysis in a variety of human tumor xenografts treated with IPI-926 are summarized in Supplementary Table S1. Collectively, these data indicate that the majority of the chondrosarcoma tumor cells tested are directly respon- sive to Hh pathway inhibition by IPI-926, whereas epithelial-derived cancer cells, like pancreatic tumors, are not. In addition, these data are consistent with the pro- posal that there is ligand-dependent activation of Hh signaling in chondrosarcoma, which occurs through both autocrine and paracrine mechanisms as signaling can be inhibited by IPI-926 treatment.

Hh inhibition in primary chondrosarcoma xenografts leads to decreased tumor volume, decreased tumor cellularity, and tumoral calcification Experiments were designed to determine whether IPI- 926 has an effect on chondrosarcoma xenograft establishment and tumor growth. Two study designs in mice were tested: in one study design, treatment was initiated on the day of implantation, and in the second study design, treatment was initiated one month after implantation to allow time for tumor establishment. IPI-926 was administered for 6 weeks. Tumors were excised, and the end- Figure 2. IPI-926 treatment leads to Hh pathway inhibition in point tumor volume was recorded. There were no chondrosarcoma cancer cells. Mice bearing primary human chondrosarcoma tumors or pancreatic carcinoma tumors derived from observed effects on tumor establishment as all chondro- the MIA PaCa-2 cell line were treated orally with 40 mg/kg of IPI-926 or sarcomas were able to grow as xenografts despite treat- vehicle control. Twenty-four hours after the last dose, tumors were ment with IPI-926 starting on the day of implantation. In excised and RNA was extracted for qRT-PCR analysis. A, IPI-926 both study designs, IPI-926 administration resulted in treatment led to a significant reduction of human GLI1 mRNA expression in both primary chondrosarcoma models tested (P < 0.01). B, PTCH1 significantly decreased tumor volumes compared with mRNA levels were also significantly downregulated with IPI-926 controls. Overall, there was a mean reduction in tumor treatment (P ¼ 0.05), while SMO mRNA levels increased. C, human and volume of 44% with IPI-926 treatment (n ¼ 5, P < 0.02; Fig. murine mRNA expression comparison following IPI-926 treatment. 3A). At the end of each study, tumors were also collected Gli1 fi Murine levels were detected using species-speci c primers. IPI-926 and processed for histologic evaluation. Hematoxylin and was able to downregulate murine Gli1 mRNA expression in chondrosarcoma, however not to the same extent as human GLI1 due to eosin–stained samples revealed decreased tumor cellu- the fact that much higher levels of human GLI1 were expressed in larity, and interestingly, increased areas of calcification in chondrosarcoma xenografts. Note that murine Gli1 expression is the tumors from drug-treated mice (Fig. 3B). These obser- higher in the pancreatic xenograft compared with the primary vations suggest that IPI-926 may drive chondrosarcoma chondrosarcoma xenograft (P < 0.04 for human GLI1 in chondrosarcoma, P < 0.04 for murine Gli1 in chondrosarcoma and P < 0.001 for murine cells to differentiate, leading to calcification, a process Gli1 in pancreatic xenograft). There was no modulation of human GLI1 in observed in normal growth plate chondrocytes. the IPI-926–treated MIA PaCa-2 tumors. , significant differences in The response of chondrosarcoma xenografts to IPI-926 expression. in comparison with the chemotherapeutic agents,

www.aacrjournals.org Mol Cancer Ther; 13(5) May 2014 1263

Campbell et al.

Figure 3. IPI-926 treatment of primary chondrosarcoma xenografts at the time of implantation, or after tumor establishment, leads to reduced tumor size, decreased cellularity and increased areas of calcification. A, treatment with IPI-926 led to a 44% mean tumor growth inhibition compared with the control group (n ¼ 5, P < 0.02). B, hematoxylin and eosin (H&E) stain of the IPI-926–treated tumors shows overall less tumor cellularity and increased areas of calcification compared with the vehicle control–treated tumors. C, IPI-926 treatment was compared with other chemotherapies, and only IPI-926 and cisplatin administration resulted in a statistically significant reduction in endpoint tumor weights (P ¼ 0009 and P ¼ 0.01, respectively). D, qRT-PCR analysis showed that only IPI-926 treatment led to a significant (P ¼ 0.0001) downmodulation of human GLI1 mRNA expression. , significant differences in expression.

doxorubicin and cisplatin, was also investigated. Alt- genes in the chondrosarcoma tumor cells. To identify such hough it is generally accepted that chondrosarcoma is targets, human gene profiling was carried out by expres- resistant to conventional sarcoma chemotherapy, it was sion microarray analysis of primary chondrosarcoma important to determine whether there was any chemo- tissue treated with IPI-926 or control. Using the analysis therapeutic response of the primary chondrosarcoma described in the Materials and Methods resulted in an tumors relative to the changes already described above initial list of 28,587 differentially expressed genes that was by Hh inhibition with IPI-926. Mice in which chondro- filtered down to 9,240 genes. A cut-off value of 0.05 was sarcoma xenografts were established were treated with used to select the significantly differentially expressed doxorubicin, cisplatin, or IPI-926 for a duration of 3 weeks. genes, and the top 10 upregulated or downregulated by Twenty-four hours after the last dose of each therapy, IPI-926 treatment are listed in Table 1. tumors were collected and weighed and then processed One of the 10 most downregulated genes under Hh for mRNA expression analysis. Compared with the con- blockade was ADAMTSL1. We chose to study this gene trol group, only IPI-926 and cisplatin-treated xenografts further because its Caenorhabditis elegans homolog is showed significant tumor growth inhibition of 46% (P ¼ known to be involved in apoptosis and as such it could 0.009) and 54% (P ¼ 0.01), respectively (Fig. 3C). qRT-PCR play an important role in neoplasia; however, it is not analysis confirmed that only IPI-926 significantly inhib- known to be a Hh target gene (22, 23). This gene encodes a ited human GLI1 expression (P ¼ 0.0001) in the chondro- secreted protein whose function is unclear, but shares sarcoma cells (Fig. 3D). similarities with members of the disintegrin-like and metalloproteinase domain with thrombospondin type 1 Identification of ADAMTSL1 as a Hh-regulated gene motif (ADAMTS) family (24). In chondrosarcoma xeno- that alters chondrosarcoma proliferation grafts from mice treated with IPI-926, downregulation of Hh signaling might contribute to the neoplastic pheno- expression of ADAMTSL1, but not its closest homolog type by regulating the expression of downstream target ADAMTSL3, was verified by qRT-PCR (P ¼ 0.002;

1264 Mol Cancer Ther; 13(5) May 2014 Molecular Cancer Therapeutics

The Hh Pathway and IPI-926 in Chondrosarcoma

Table 1. The top 20 genes modulated with IPI-926 treatment, rank ordered by fold change with treatment

Top 10 upregulated genes Gene symbol Log fold change Description ITIH5a 7.02 Inter-alpha-trypsin inhibitor heavy chain family, member 5 SLC2A4a 2.63 Solute carrier family 2 (facilitated glucose transporter), member 4 ANGPTL7a 2.13 Angiopoietin-like 7 HAS1 2.11 Hyaluronan synthase 1 FGF18 1.98 Fibroblast growth factor 18 KLRC1 1.92 Killer cell lectin-like receptor subfamily C, member 1 UBD 1.88 Ubiquitin D C1QTNF9B 1.58 C1q and tumor necrosis factor–related protein 9B SPRR2D 1.32 Small proline-rich protein 2D ADCY4 1.39 Adenylate cyclase 4

Top 10 downregulated genes Gene symbol Log fold change Description

GLI1a 3.27 GLI family zinc ﬁnger 1 KCNIP1a 3.25 Kv channel interacting protein 1 PLCXD3a 2.91 Phosphatidylinositol-speciﬁc phospholipase C, X domain containing 3 SLC13A5a 2.86 Solute carrier family 13 (sodium-dependent citrate transporter), member 5 HHIPa 2.64 Hedgehog interacting protein GDF10a 2.43 Growth differentiation factor 10 C7a 2.36 Complement component 7 FAM150Ba 2.30 Family with sequence similarity 150, member B HHIPL2a 2.03 HHIP-like 2 ADAMTSL1a 1.81 ADAMTS-like 1

aAdjusted P value <0.05.

Fig. 4A). Chondrosarcoma xenografts were grown as ADAMTSL1, but not ADAMTSL3 (data not shown), led tissue explants in culture (12) and subjected to either Hh to a 2-fold increase in cell proliferation as measured by blockade with IPI-926 or Hh stimulation with recombi- BrdUrd incorporation (P ¼ 0.04; Fig. 5A). When treated nant SHh-N. Hh blockade resulted in significant down- with increasing concentrations of purified ADAMTSL1 regulation of GLI1 and ADAMTSL1 expression, while protein, cell proliferation increased in a dose-dependent pathway stimulation led to significant upregulation of manner (P < 0.005; Fig. 5B). Thus, in chondrosarcoma, Hh- both genes (P < 0.01; Fig. 4A). regulated genes include those whose protein product The MULAN website (http://mulan.dcode.org/) was modulates the neoplastic phenotype. This supports the used to examine the promoter region of human concept that Hh signaling within chondrosarcoma cells ADAMTSL1. Multiple GLI consensus binding sites were themselves alters cellular behavior in a manner that drives identified in its promoter region (Fig. 4B), and ChIP was the neoplastic phenotype. used to verify GLI binding to the ADAMTSL1 promoter. ChIP with primer pairs adjacent to the GLI1 and GLI2 consensus binding site were identified (Fig. 4B) and an Discussion antibody to GLI1 or GLI2 detected both GLI1 and GLI2 In the current study, we demonstrated that a targeted bound to this site (Fig. 4C). Fragments immunoprecipi- and specific Hh inhibitor, IPI-926, had significant antitu- tated using antibodies against GLI1, GLI2, and IgG were mor effects against human chondrosarcoma. IPI-926 is a sequenced and the GLI1 and GLI2 fractions aligned novel potent small-molecule inhibitor of the Hh pathway uniquely to the intergenic region of the ADAMTSL1 gene. that directly targets SMO and has previously shown This verified binding of both GLI1 and GLI2 to the con- activity in xenograft models of basal cell carcinoma and sensus sequence site in the ADAMTSL1 promoter. To medulloblastoma, two tumors that, like chondrosarcoma, determine whether this protein has functional significance are also dependent on constitutive Hh signaling (5, 25). in chondrosarcoma, primary tumors were grown in cell We found IPI-926 to be a very effective antitumor agent in culture and treated with ADAMTSL1 or the related protein primary chondrosarcoma xenografts, regardless of ADAMTSL3-conditioned medium. Treatment with whether treatment was initiated at the time of tumor

www.aacrjournals.org Mol Cancer Ther; 13(5) May 2014 1265

Campbell et al.

Figure 4. Positive correlation between Hh activity and human ADAMTSL1 expression. A, qRT-PCR analysis of chondrosarcoma xenografts from mice treated with IPI-926 confirms the downregulation of human ADAMTSL1 expression (P ¼ 0.002), but not ADAMTSL3. Treatment of chondrosarcoma explants in culture with the Hh antagonist IPI-926 significantly reduced GLI1 and ADAMTSL1 expression (P < 0.01), whereas the Hh agonist SHh significantly increased GLI1 and ADAMTSL1 expression (P < 0.01). B, diagrammatic representation of the human ADAMTSL1 gene promoter showing putative Gli consensus binding sites from MULAN analysis. The vertical black line indicates a GLI1 and GLI2 consensus binding site in the intergenic DNA. C, ChIP was performed on a primary chondrosarcoma treated with SHh ligand. A PCR and gel electrophoresis is shown for primers at the GLI1 and GLI2 consensus binding site from fragments immunoprecipitated using a GLI1, GLI2 or IgG antibody. Amplification was seen with the GLI1 and GLI2 antibody, but not control IgG (input DNA is a positive control). Sequence reads of fragments from GLI1 and GLI2 immunoprecipitated fractions aligned uniquely to the intergenic region of the ADAMTSL1 gene.

implantation or 6 weeks following engraftment. Histo- For most tumors, it remains controversial whether Hh logic analysis identified changes in tumor morphology, signaling occurs in the tumor cells themselves, stromal including decreased cellularity and increased calcifica- cells, or both. Chondrosarcoma is an example of a tumor tion, following IPI-926 treatment. Both of these features with ligand-dependent Hh signaling as it expresses high are indicative of induced cartilage cell differentiation as is levels of Ihh, and only rarely harbors mutations in path- normally seen in growth plate chondrocytes in response to way-specific genes (11, 26–28). We utilized species-spe- Hh blockade. Further evidence for an on-target effect of cific RT-PCR primers to address the controversial issue of IPI-926 was decreased proliferation of chondrocytes and whether Hh functions through an autocrine or paracrine premature growth plate closure in the mice following mechanism in chondrosarcoma. This tumor is unique in treatment, consistent with the expected normal physio- that it exhibits evidence of both autocrine and paracrine logic role of Hh blockade (10). Hh signaling. IPI-926 led to significant reductions in

Figure 5. ADAMTSL1 increases proliferation of chondrosarcoma cells. Conditioned media (A) or purified ADAMTSL1 protein (rADAMTSL1; B) was added to the media of chondrosarcoma explants and proliferation rate was assessed by a BrdUrd uptake assay. Treatment with ADAMTSL1 leads to a 2-fold increase in cell proliferation (P ¼ 0.04). B, dose-dependent increase in proliferation was observed when increasing doses of purified ADAMTSL1 were added to three explants derived from different donors (P < 0.005). The average increase in proliferation of the three donors is graphed. , significant differences in expression.

1266 Mol Cancer Ther; 13(5) May 2014 Molecular Cancer Therapeutics

The Hh Pathway and IPI-926 in Chondrosarcoma

expression of PTCH1 and GLI1 in both human tumor cells which codes for a secreted protein of unknown function and mouse stroma in chondrosarcoma xenografts, indi- that is a member of the ADAMTS super family (24). cating that active Hh pathway signaling was present and ADAMTSL1 lacks the proteolytic activity typical of most functional in each cellular compartment. However, there ADAMTS family members but has the characteristic was a much greater reduction in expression of the human thrombospondin type-1 repeats (TSR). ADAMTSL1 is a Hh target gene GLI1 in chondrosarcoma tumor cells com- mammalian ortholog of MADD-4, a recently discovered pared with murine Gli1 in the stroma, likely due to the protein which functions through the netrin receptor UNC- much higher levels of GLI1 expression in the human 40/DCC in C. elegans (23). As a potential mediator of tumor cells at baseline and the relative acellularity of netrins, ADAMTSL1 may play an important role in cell stromal tissues seen histologically in chondrosarcoma. survival and neoplasia (22). Targeting ADAMTSL1 and a We found murine Gli1 expression to be dramatically putative receptor that lies downstream of Hh pathway higher in the MIA PaCa-2 pancreatic cancer xenografts signaling may represent a promising therapeutic strategy compared with the chondrosarcoma tissue, and this was to pursue for patients with chondrosarcoma. predominantly due to the higher prevalence of mouse A potential means of improving the poor results asso- stromal cells. This is also consistent with the extensive ciated with systemic therapy for patients with chondro- desmoplasia typical of pancreatic cancer that is known to sarcoma would be to consider targeting Hh or ADAMTSL1 be Hh-driven and constitutes an important feature of this together with conventional chemotherapy, as the combi- particular tumor microenvironment (29, 30). Interesting- nation could act synergistically. Compared with the sig- ly, IPI-926 treatment of mice bearing MIA PaCa-2 pan- nificant antitumor effects of IPI-926 in the chondrosarcoma creatic xenografts led to diminished Hh signaling only in xenograft model, cisplatin also led to a significant reduc- stromal cells but not in the pancreatic cancer cells them- tion in tumor growth independent of Hh inhibition, as selves, indicating that paracrine Hh signaling is likely there was not a concomitant decrease in Gli1 expression. involved in pancreatic cancer. Other solid tumor xeno- Given that chondrosarcoma is considered to be chemo- grafts that we examined from colon, breast, and lung therapy resistant, these data suggest that further investi- cancers demonstrated similar findings, in that Hh-respon- gation may be warranted into the potential efficacy of sive signaling was limited to the mouse stroma in the cisplatin in chondrosarcoma and whether combination tumor microenvironment (Supplementary Table S1). This with IPI-926 or other novel therapeutics could provide is consistent with recent data showing that in most solid additional benefits. For example, inhibition of Hh signal- epithelial-derived malignancies, tumor cell-derived Hh ing in combination with chemotherapy has been shown to ligands do not stimulate the tumor cells themselves, but reverse chemoresistance in tumor models of pancreatic instead activate the Hh pathway in the surrounding cancer, lung cancer, and lymphoma (38, 44, 45). In tumors stromal cells in a paracrine manner (31–35). such as chondrosarcoma, which rely on ligand-dependent There are few examples of autocrine Hh signaling in Hh activation and exhibit high levels of GLI1 and PTCH1 tumor cells beyond the demonstration of chondrosarcoma expression, downregulation of GLI1 and PTCH1 could in the current study. Recent evidence for constitutive Hh serve as biomarkers for responseto Hh pathway inhibition. activation was observed in models of small and non–small In addition, as most chondrosarcomas have an identifiable cell lung cancer, melanoma, and chronic myelogenous primary cilium, the presence of cilia may serve as a bio- and lymphocytic leukemia, in which the pathway is marker for response to Hh blockade, as recently demon- activated in the tumor cells (36–40). However, the pres- strated in a transgenic mouse model of cartilage neoplasia ence of transcripts for Hh target genes does not necessarily (46). As a potential treatment for patients, IPI-926, a potent represent active signaling, without evidence that the and selective SMO antagonist, may have advantages as expression can be modulated by pathway inhibition targeted therapy for Hh-dependent tumors compared (35, 41). In addition, results from studies using high doses with other pathway inhibitors. For example, triparanol of cyclopamine may be confounded by the off-target has significant toxicity in humans in the form of inducing effects of cyclopamine at concentrations in the micromolar alopecia and cataracts (12), whereas cyclopamine is pla- range. In some tumors where paracrine and autocrine Hh gued by a lack of selectivity at exposures necessary for Hh signaling coexists, the balance between the two may pathway inhibition (35). regulate cancer progression (42, 43). Chondrosarcoma is a malignant tumor of cartilage in The antitumor activity of IPI-926 in chondrosarcoma which there is ligand-dependent activation of Hh signal- xenografts could be due to multiple factors and pathways ing both within the tumor cells and the surrounding known to be regulated by Hh pathway blockade. To stromal compartment. The results of these studies provide identify potential downstream targets of the Hh pathway the rationale for investigating Hh pathway blockade as a in chondrosarcoma, differentially expressed genes were novel therapeutic option for patients with chondrosar- identified in tumor xenografts using microarray analysis. coma. Although the early results of a single-arm trial with Not surprisingly, GLI1 and two Hh associated proteins the Hh blocking drug vismodegib found that 9 of 30 (HHIP and HHIPL2) were among the most downregulated chondrosarcoma patients were progression-free at 6 genes. Another gene which was highly downregulated as months follow-up, longer follow-up of this single-arm a result of Hh blockade with IPI-926 was ADAMTSL1, trial and as a randomized trial with IPI-926 in patients

www.aacrjournals.org Mol Cancer Ther; 13(5) May 2014 1267

Campbell et al.

with advanced chondrosarcoma were both stopped as Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): V. Campbell, P. Nadesan, S.A. Ali, C.Y.Y. Wang, they did not meet their clinical objectives (47, 48). Ongoing H. Whetstone, R. Poon, Q. Wei, J.L. Proctor, B.A. Alman, J. Wunder translational research on tumor samples from these two Analysis and interpretation of data (e.g., statistical analysis, biostatis- tics, computational analysis): V. Campbell, P. Nadesan, S.A. Ali, C.Y.Y. studies may identify patient subsets that have derived Wang, R. Poon, J. Keilty, B.A. Alman, J. Wunder benefit from Hh blockade. Despite this setback, new Writing, review, and/or revision of the manuscript: V. Campbell, S.A. Ali, research findings in chondrosarcoma may lead to the J. Keilty, J.L. Proctor, S. Apte, K. McGovern, B.A. Alman, J. Wunder Administrative, technical, or material support (i.e., reporting or orga- development of other opportunities for novel treatments. nizing data, constructing databases): C.Y.Y. Wang, J. Keilty, L.W. Wang, For example, isocitrate dehydrogenase (IDH) mutations J. Wunder were recently identified in approximately 50% of chon- Study supervision: J. Wunder drosarcomas and drugs are currently under development to inhibit mutant IDH enzymatic activity (49). In addition, Acknowledgments genomic sequencing and functional evaluation of signal- The authors thank the Canadian Institutes of Health Research for supporting this research project. ing pathways have both led to the identification of specific genes that may be targetable as part of novel therapeutic strategies (27, 50). Grant Support This research project was funded by the Canadian Institutes of Health fl Research through operating grant #MOP-37913 (awarded to J.S. Wunder Disclosure of Potential Con icts of Interest and B.A. Alman) and by Infinity Pharmaceuticals, Inc. No potential conflicts of interest were disclosed. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked Authors' Contributions advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate Conception and design: V. Campbell, C.Y.Y. Wang, Q. Wei, K. McGovern, this fact. B.A. Alman, J. Wunder Development of methodology: V. Campbell, P. Nadesan, C.Y.Y. Wang, Received September 18, 2013; revised February 7, 2014; accepted March R. Poon, Q. Wei, B.A. Alman, J. Wunder 2, 2014; published OnlineFirst March 14, 2014.

References 1. Jiang J, Hui C-c. Hedgehog signaling in development and cancer. Dev 13. Dantonello TM, Int-Veen C, Leuschner I, Schuck A, Furtwaengler R, Cell 2008;15:801–12. Claviez A, et al. Mesenchymal chondrosarcoma of soft tissues and 2. Corbit KC, Aanstad P, Singla V, Norman AR, Stainier DY, Reiter JF. bone in children, adolescents, and young adults: experiences of the Vertebrate Smoothened functions at the primary cilium. Nature CWS and COSS study groups. Cancer 2008;112:2424–31. 2005;437:1018–21. 14. GrimerRJ,GoshegerG,TaminiauA,BiauD,MatejovskyZ,Kol- 3. Bale AE, Yu KP. The hedgehog pathway and basal cell carcinomas. lender Y, et al. Dedifferentiated chondrosarcoma: prognostic fac- Hum Mol Genet 2001;10:757–62. tors and outcome from a European group. Eur J Cancer 2007;43: 4. Hahn H, Wicking C, Zaphiropoulous PG, Gailani MR, Shanley S, 2060–5. Chidambaram A, et al. Mutations of the human homolog of Drosophila 15. Italiano A, Mir O, Ciofﬁ A, Palmerini E, Piperno-Neumann S, Perrin C, patched in the nevoid basal cell carcinoma syndrome. Cell 1996;85: et al. Advanced chondrosarcomas: role of chemotherapy and survival. 841–51. Ann Oncol 2013;24:2916–22. 5. Lee MJ, Hatton BA, Villavicencio EH, Khanna PC, Friedman SD, Ditzler 16. Tremblay MR, Lescarbeau A, Grogan MJ, Tan E, Lin G, Austad BC, S, et al. Hedgehog pathway inhibitor saridegib (IPI-926) increases et al. Discovery of a potent and orally active hedgehog pathway lifespan in a mouse medulloblastoma model. Proc Natl Acad Sci antagonist (IPI-926). J Med Chem 2009;52:4400–18. U S A 2012;109:7859–64. 17. Daniel VC, Marchionni L, Hierman JS, Rhodes JT, Devereux WL, Rudin 6. Gelderblom H, Hogendoorn PC, Dijkstra SD, van Rijswijk CS, Krol AD, CM, et al. A primary xenograft model of small-cell lung cancer reveals Taminiau AH, et al. The clinical approach towards chondrosarcoma. irreversible changes in gene expression imposed by culture in vitro. Oncologist 2008;13:320–9. Cancer Res 2009;69:3364–73. 7. Farquharson C, Jefferies D, Seawright E, Houston B. Regulation of 18. Edgar R, Domrachev M, Lash AE. Gene Expression Omnibus: NCBI chondrocyte terminal differentiation in the postembryonic growth gene expression and hybridization array data repository. Nucleic Acids plate: the role of the PTHrP-indian hedgehog axis. Endocrinology Res 2002;30:207–10. 2001;142:4131–40. 19. Ho L, Stojanovski A, Whetstone H, Wei QX, Mau E, Wunder JS, et al. 8. Long F, Zhang XM, Karp S, Yang Y, McMahon AP. Genetic manipu- Gli2 and p53 cooperate to regulate IGFBP-3- mediated chondrocyte lation of hedgehog signaling in the endochondral skeleton reveals a apoptosis in the progression from benign to malignant cartilage direct role in the regulation of chondrocyte proliferation. Development tumors. Cancer Cell 2009;16:126–36. 2001;128:5099–108. 20. Lin TL, Means RE, Wang QH, Brown P, Peacock C, Merchant AA, et al. 9. Maeda Y, Nakamura E, Nguyen MT, Suva LJ, Swain FL, Razzaque MS, Self-renewal of acute lymphocytic leukemia cells is limited by the et al. Indian Hedgehog produced by postnatal chondrocytes is essen- Hedgehog pathway inhibitors cyclopamine and IPI-926. PLoS ONE tial for maintaining a growth plate and trabecular bone. Proc Natl Acad 2010;5:e15262. Sci U S A 2007;104:6382–7. 21. Wang LW, Leonhard-Melief C, Haltiwanger RS, Apte SS. Post-trans- 10. Kimura H, Ng JM, Curran T. Transient inhibition of the Hedgehog lational modiﬁcation of thrombospondin type-1 repeats in ADAMTS- pathway in young mice causes permanent defects in bone structure. like 1/punctin-1 by C-mannosylation of tryptophan. J Biol Chem Cancer Cell 2008;13:249–60. 2009;284:30004–15. 11. Hopyan S, Gokgoz N, Poon R, Gensure RC, Yu C, Cole WG, et al. A 22. Llambi F, Causeret F, Bloch-Gallego E, Mehlen P. Netrin-1 acts as a mutant PTH/PTHrP type I receptor in enchondromatosis. Nat Genet survival factor via its receptors UNC5H and DCC. EMBO J 2001;20: 2002;30:306–10. 2715–22. 12. Tiet TD, Hopyan S, Nadesan P, Gokgoz N, Poon R, Lin AC, et al. 23. Seetharaman A, Selman G, Puckrin R, Barbier L, Wong E, D'Souza SA, Constitutive hedgehog signaling in chondrosarcoma up-regulates et al. MADD-4 is a secreted cue required for midline-oriented guidance tumor cell proliferation. Am J Pathol 2006;168:321–30. in Caenorhabditis elegans. Dev Cell 2011;21:669–80.

1268 Mol Cancer Ther; 13(5) May 2014 Molecular Cancer Therapeutics

The Hh Pathway and IPI-926 in Chondrosarcoma

24. Hirohata S, Wang LW, Miyagi M, Yan L, Seldin MF, Keene DR, et al. 38. Park KS, Martelotto LG, Peifer M, Sos ML, Karnezis AN, Mahjoub MR, Punctin, a novel ADAMTS-like molecule, ADAMTSL-1, in extracellular et al. A crucial requirement for Hedgehog signaling in small cell lung matrix. J Biol Chem 2002;277:12182–9. cancer. Nat Med 2011;17:1504–8. 25. Weiss GJ. Hunting Hedgehog in the Field of Skin Cancer, New 39. Singh S, Wang Z, Liang Fei D, Black KE, Goetz JA, Tokhunts R, et al. Concepts Session, Advanced in Skin Cancers: New Treatment, Hedgehog-producing cancer cells respond to and require autocrine New Virus, New Mouse Models. Presented at the 101st Annual Hedgehog activity. Cancer Res 2011;71:4454–63. Meeting of the American Association for Cancer Research; 2010 40. Stecca B, Mas C, Clement V, Zbinden M, Correa R, Piguet V, et al. Apr 17–21; Washington, DC. Philadelphia (PA): 2010. Melanomas require HEDGEHOG-GLI signaling regulated by interac- 26. Couvineau A, Wouters V, Bertrand G, Rouyer C, Gerard B, Boon LM, tions between GLI1 and the RAS-MEK/AKT pathways. Proc Natl Acad et al. PTHR1 mutations associated with Ollier disease result in receptor Sci U S A 2007;104:5895–900. loss of function. Hum Mol Genet 2008;17:2766–75. 41. Zhang J, Lipinski R, Shaw A, Gipp J, Bushman W. Lack of demon- 27. Tarpey PS, Behjati S, Cooke SL, Van Loo P, Wedge DC, Pillay N, et al. strable autocrine hedgehog signaling in human prostate cancer cell Frequent mutation of the major cartilage collagen gene COL2A1 in lines. J Urol 2007;177:1179–85. chondrosarcoma. Nat Genet 2013;45:923–6. 42. Lauth M, Bergstrom A, Shimokawa T, Tostar U, Jin Q, Fendrich V, et al. 28. Yan T, Angelini M, Alman BA, Andrulis IL, Wunder JS. Patched-one or DYRK1B-dependent autocrine-to-paracrine shift of Hedgehog signal- smoothened gene mutations are infrequent in chondrosarcoma. Clin ing by mutant RAS. Nat Struct Mol Biol 2010;17:718–25. Orthop Relat Res 2008;466:2184–9. 43. Tzelepi V, Karlou M, Wen S, Hoang A, Logothetis C, Troncoso P, et al. 29. Neureiter D, Zopf S, Dimmler A, Stintzing S, Hahn EG, Kirchner T, et al. Expression of hedgehog pathway components in prostate carcinoma Different capabilities of morphological pattern formation and its asso- microenvironment: shifting the balance towards autocrine signalling. ciation with the expression of differentiation markers in a xenograft Histopathology 2011;58:1037–47. model of human pancreatic cancer cell lines. Pancreatology 2005; 44. Olive KP, Jacobetz MA, Davidson CJ, Gopinathan A, McIntyre D, 5:387–97. Honess D, et al. Inhibition of Hedgehog signaling enhances delivery 30. Wei HJ, Yin T, Zhu Z, Shi PF, Tian Y, Wang CY. Expression of CD44, of chemotherapy in a mouse model of pancreatic cancer. Science CD24 and ESA in pancreatic adenocarcinoma cell lines varies with local 2009;324:1457–61. microenvironment. Hepatobiliary Pancreat Dis Int 2011;10:428–34. 45. Singh RR, Kunkalla K, Qu C, Schlette E, Neelapu SS, Samaniego F, 31. Fan L. Hedgehog signaling promotes prostate xenograft tumor growth. et al. ABCG2 is a direct transcriptional target of hedgehog signaling Endocrinology 2004;145:3961–70. and involved in stroma-induced drug tolerance in diffuse large B-cell 32. Nolan-Stevaux O, Lau J, Truitt ML, Chu GC, Hebrok M, Fernandez- lymphoma. Oncogene 2011;30:4874–86. Zapico ME, et al. GLI1 is regulated through Smoothened-independent 46. Ho L, Ali SA, Al-Jazrawe M, Kandel R, Wunder JS, Alman BA. Primary mechanisms in neoplastic pancreatic ducts and mediates PDAC cell cilia attenuate hedgehog signalling in neoplastic chondrocytes. Onco- survival and transformation. Genes Dev 2009;23:24–36. gene 2012;32:5388–96. 33. O'Toole SA, Machalek DA, Shearer RF, Millar EK, Nair R, Schofield P, 47. Infinity Stops Phase 2 Trials of Saridegib in Chondrosarcoma et al. Hedgehog overexpression is associated with stromal interactions and Myelofibrosis [press release]. Cambridge, MA: Business Wire; and predicts for poor outcome in breast cancer. Cancer Res 2011; 18 Jun 2012. Available from: http://www.businesswire.com/news/ 71:4002–14. home/20120618005411/en/Infinity-Stops-Phase-2-Trials-Saridegib- 34. Tian H, Callahan CA, DuPree KJ, Darbonne WC, Ahn CP, Scales SJ, Chondrosarcoma#.Uy-VwlJOW70. et al. Hedgehog signaling is restricted to the stromal compartment 48. Italiano A, Le Cesne A, Bellera C, Piperno-Neumann S, Duffaud F, during pancreatic carcinogenesis. Proc Natl Acad Sci U S A 2009; Penel N, et al. GDC-0449 in patients with advanced chondrosar- 106:4254–9. comas: a French Sarcoma Group/US and French National Cancer 35. Yauch RL, Gould SE, Scales SJ, Tang T, Tian H, Ahn CP, et al. A Institute Single-Arm Phase II Collaborative Study. Ann Oncol paracrine requirement for hedgehog signalling in cancer. Nature 2008; 2013;24:2922–6. 455:406–10. 49. Amary MF, Bacsi K, Maggiani F, Damato S, Halai D, Berisha F, et al. 36. Decker S, Zirlik K, Djebatchie L, Hartmann D, Ihorst G, Schmitt-Graeff IDH1 and IDH2 mutations are frequent events in central chondrosar- A, et al. Trisomy 12 and elevated GLI1 and PTCH1 transcript levels are coma and central and periosteal chondromas but not in other mes- biomarkers for Hedgehog-inhibitor responsiveness in CLL. Blood enchymal tumours. J Pathol 2011;224:334–43. 2012;119:997–1007. 50. Zhang YX, van Oosterwijk JG, Sicinska E, Moss S, Remillard SP, van 37. Dierks C, Beigi R, Guo GR, Zirlik K, Stegert MR, Manley P, et al. Wezel T, et al. Functional profiling of receptor tyrosine kinases Expansion of Bcr-Abl-positive leukemic stem cells is dependent on and downstream signaling in human chondrosarcomas identifies path- Hedgehog pathway activation. Cancer Cell 2008;14:238–49. ways for rational targeted therapy. Clin Cancer Res 2013;19:3796–807.

www.aacrjournals.org Mol Cancer Ther; 13(5) May 2014 1269

Hedgehog Pathway Inhibition in Chondrosarcoma Using the Smoothened Inhibitor IPI-926 Directly Inhibits Sarcoma Cell Growth

Veronica T. Campbell, Puviindran Nadesan, S. Amanda Ali, et al.

Mol Cancer Ther 2014;13:1259-1269. Published OnlineFirst March 14, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-13-0731

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2014/03/14/1535-7163.MCT-13-0731.DC1

Cited articles This article cites 48 articles, 16 of which you can access for free at: http://mct.aacrjournals.org/content/13/5/1259.full#ref-list-1

Citing articles This article has been cited by 3 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/13/5/1259.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://mct.aacrjournals.org/content/13/5/1259. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.