IJC International Journal of Cancer
Protease-activated receptor-1 drives pancreatic cancer progression and chemoresistance Karla C.S. Queiroz1, Kun Shi1, JanWillem Duitman1, Hella L. Aberson1, Johanna W. Wilmink2, Carel J.M. van Noesel3, Dick J. Richel2 and C. Arnold Spek1 1
Center for Experimental and Molecular Medicine, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands Department of Medical Oncology, Academic Medical Center, Amsterdam, The Netherlands 3 Department of Pathology, Academic Medical Center, Amsterdam, The Netherlands 2
Cancer Cell Biology
Protease activated receptor (PAR)-1 expression in tumor cells is associated with disease progression and overall survival in a variety of cancers of epithelial origin; however, the importance of PAR-1 in the tumor microenvironment remains unexplored. Utilizing an orthotopic pancreatic cancer model in which tumor cells are PAR-1 positive whereas stromal cells are PAR-1 negative, we show that PAR-1 expression in the microenvironment drives progression and induces chemoresistance of pancreatic cancer. PAR-1 enhances monocyte recruitment into the tumor microenvironment by regulating monocyte migration and fibroblast dependent chemokine production thereby inducing chemoresistance. Overall, our data identify a novel role of PAR-1 in the pancreatic tumor microenvironment and suggest that PAR-1 may be an attractive target to reduce drug resistance in pancreatic cancer.
Pancreatic cancer is a devastating disease with the worst survival outcome of any human cancer. The incidence, which is around 10 per 100,000 individuals, is rising in developed countries,1 with around 44,000 new cases and 37,000 deaths in the United States in 2011.2 The 5-year survival rate is less than 5%, and overall mortality approaches 99%.3 Based on a small survival benefit of gemcitabine as compared to 5fluorouracil in a Phase III study performed more than 10 years ago,4 gemcitabine is first-line standard of care for inoperable pancreatic cancer but the observed benefits are small5 and seem restricted to patients with a good performance status.6 Protease activated receptor (PAR)-1 is a receptor that belongs to the family of G protein-coupled receptors.7 Four different PARs have been identified of which PAR-1 is the prototype of this family of receptors. Activation of PARs is Key words: protease activated receptors, microenvironment, macrophage, chemoresistance Additional Supporting Information may be found in the online version of this article. Grant sponsor: Top Institute Pharma; Grant number: T1-215; Grant sponsor: Dutch Cancer Society; Grant number: AMC 20094324 DOI: 10.1002/ijc.28726 History: Received 7 Aug 2013; Accepted 2 Jan 2014; Online 16 Jan 2014 Correspondence to: Karla C.S. Queiroz, Center for Experimental and Molecular Medicine, Academic Medical Center, Room L01-145, University of Amsterdam, Meibergdreef 9, NL-1105 AZ Amsterdam, The Netherlands, Tel.: 131-20-5664903, E-mail: [email protected]
C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
unique as it requires proteolytic cleavage rather than ligand binding. Removal of the N-terminal extracellular region releases a new tethered ligand that interacts with the body of the receptor to induce transmembrane signaling thereby triggering a range of signaling pathways affecting multiple pathophysiological responses.8 PAR-1 expression levels are increased in a variety of cancers including breast,9 lung,10,11 ovarian12 and prostate cancer.13 More interestingly, PAR-1 levels are correlated with disease progression and overall survival in breast,14 lung,11 gallbladder,15 melanoma16 and gastric cancer.17 In line with a potential important role of tumor cell PAR-1, pharmacological targeting of PAR-1 in preclinical animal models shows promising effects. Indeed, P1pal-7 (pepducin targeting PAR-1) treatment significantly inhibited growth and metastasis to the lungs18 of breast cancer xenografts. Moreover, P1pal-7 provided inhibition of lung tumor growth in nude mice,10 whereas PAR-1 siRNA decreased tumor growth and metastasis to the lung in a murine melanoma model.19 PAR-1 expression on tumor cells thus seems to drive tumor progression and metastasis. A hallmark of pancreatic cancer is the strong desmoplastic reaction causing dense fibrosis around the tumor cells (reviewed in Ref. 20). Activation of (myo)fibroblasts in the stroma, or tumor microenvironment, leads to the secretion of collagen and other extracellular matrix components. Interestingly, the stroma is not simply a mechanical barrier but actually constitutes a dynamic compartment involved in tumor progression, metastasis and drug resistance.21 Indeed, growth factor production by stromal cells activates oncogenic signaling pathways in proliferating tumor cells leading to worse prognosis and resistance to treatment.
2295
Queiroz et al.
What’s new? Few people survive a diagnosis of pancreatic cancer. In the search for new avenues to attack the disease, researchers have noticed that the receptor PAR-1 promotes tumor growth and proliferation. It is also expressed in the tumor microenvironment, where its role is not yet fully explored. In this study, the authors showed that PAR-1 outside the tumor contributes to drug resistance by bringing monocytes onto the scene. Their results suggest that targeting PAR-1 could boost the effectiveness of chemotherapy against pancreatic cancer.
Material and Methods Patients
Tumor sections of pancreatic cancer patients obtained during surgery were used to prepare a tissue microarray according to the guidelines of the Medical Ethical Committee of the Academic Medical Center of Amsterdam. Biopsies were obtained from pancreatic cancer patients after informed consent. Animals
PAR-1 deficient mice generated as described previously28 and C57BL/6 mice (purchased from Charles River) were maintained at the animal facility of the Academic Medical Center of Amsterdam with free access to food and water. All animal experiments were approved by the facility’s Institutional Animal Care and Use Committee. Orthotopic pancreatic cancer model
Confluent cultures of PANC02 cells (>90% viable; kindly provided by Dr. Schmitz, Universit€atsklinikum Bonn) were incubated with trypsin, pelleted, washed twice in phosphate buffered saline (PBS) and resuspended in 0.9% sterile saline (Sigma, St Louis, MO). Tumor cells (4 3 105) were injected directly into the tail of the pancreas of 8- to 10-week-old mice. Mice were evaluated for changes in body weight and signs of discomfort or morbidity, and they were euthanized C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
4 to 5 weeks after tumor cell injection. The pancreas, including the tumor, was removed, measured and weighed. Metastasis in distant organs was assessed macroscopically by visual examination. If indicated, mice were treated with gemcitabine (120 mg/kg in PBS; i.p.) or saline (control) starting from 7 days after tumor cell injection and this treatment was repeated twice weekly. Both saline and gemcitabine experiments were performed simultaneously. (Immuno)histochemistry
Endogenous peroxidase activity was quenched with 0.3% hydrogen peroxide for 15 min at room temperature, and antigen retrieval was performed for 10 min at 100 C in 10 mM sodium citrate buffer, pH 7.4. Next, slides were blocked for 10 minu with Ultra V block (Thermo Scientific, Runcorn, UK) for PAR-1, F4/80, CD31 and LYVE-1 or for 30 min with 5% (for collagen I) or 10% (for aSMA) normal goat serum. Primary antibodies against PAR-1 (1:200; SC-13503, Santa Cruz, San Diego, CA), F4/80 (1:500, clone CI:A3-1, AbD Serotec, Kidlington, UK), collagen I (1:400, GTX41286, Bioconnect, Huissen, The Netherlands), cleaved caspase 3 (1:100, 9661, Cell Signaling Technology, Boston, MA), Ki67 (1:500, clone Sp6; Neomarkers, Fremont, CA) or aSMA (1:800, Clone 1A4, DAKO, Heverlee, Belgium) were added for overnight incubation at 4 C. For CD31 (1:1,000; sc-1506R, Santa Cruz Biotechnology) or LYVE-1 (1:500; #4509 American Diagnostica; Stamford, CT) primary antibodies were incubated for 90 min at room temperature. Slides were subsequently incubated with appropriate HRP-conjugated secondary antibodies and DAB staining was used to visualize peroxidase activity. Slides were photographed with a microscope equipped with a digital camera (Leica CTR500, Leica Microsystems, Wetzlar, Germany). The number of aSMA positive cells (excluding blood vessels) were counted in 10 different fields at 4003 magnification. CD31 and LYVE-1 positive vessels were counted in 10 different fields at 4003 and 2003 magnification, respectively. Cleaved caspase-3 positive cells were counted in 10 different fields at 4003 magnification. Ki67 positive cells were counted with the ImageJ cell counter tool, positive cells were counted on 2003 magnification pictures in at least five different fields of 0.257 cm2. F4/ 80 staining was analyzed with ImageJ and expressed as percentage of the surface area. The average of 10 pictures at 2003 magnification was used for analysis. Masson trichrome and collagen I stained sections were as graded on a 0 to 3 scale in which 0 indicates a section with no staining, whereas
Cancer Cell Biology
In addition to an important role in tumor cell biology, PAR-1 is emerging as a key player in fibroproliferative disease.22 Indeed, activated PAR-1 induces proliferation, migration and extracellular matrix production of fibroblasts and PAR-1 is upregulated in fibrotic regions in different pathologies.23–25 Importantly, genetic ablation of PAR-1 affords protection from bleomycin-induced lung fibrosis, and limits carbon tetrachloride-induced liver fibrosis26 and bile duct ligation induced cirrhosis.27 PAR-1 expression in the microenvironment of pancreatic cancers may thus drive the desmoplastic reaction thereby promoting tumor growth and chemoresistance. We challenged this hypothesis by evaluating tumor growth, metastasis and drug resistance in an orthotopic pancreatic cancer model in which tumor cells are PAR-1 positive whereas stromal cells are PAR-1 negative. We show that PAR-1 expression in the microenvironment indeed drives tumor growth, metastasis and chemoresistance but the tumor promoting effects of PAR-1 seem independent of the desmoplastic reaction.
2296
1 indicates weak staining, 2 indicates median staining, and 3 indicates intense staining. Cell culture studies
PANC02, RAW264.7 and C3H10T1/2 cells were cultured in RPMI, IMDM and DMEM, respectively, containing 10% FCS, 2 mM L-glutamine and antibiotics. For conditioned medium preparation, cells were grown overnight under normal growth conditions. After 24 hr, the cells were washed, serum starved for 4 hr, and incubated for 24 hr in medium with or without thrombin (1 U/ml) in the presence or absence of P1pal-12 (2.5 lM). For cell viability assays, 1 3 103 PANC02 cells were seeded in 96-well plates and, after overnight attachment, treated for 48 hr with gemcitabine in control or conditioned medium. RAW264.7 cell viability or proliferation was assessed using the same conditions in the absence or presence of thrombin (0–1 U/ml) for 24 hr.
Cancer Cell Biology
Cell viability and proliferation assays
Cell viability was measured using MTT reduction assays29 and by crystal violet staining.30 Cell proliferation was determined with a Cell Proliferation ELISA kit from Roche (BrdU; chemiluminescent) following the manufacturer recommendations. Cell viability and proliferation assays were measured with microplate reader (Synergy HT, BioTek). Transwell migration assays
In detail, 2 3 104 Raw 264.7 cells, serum starved for 4 hr, were transferred to 8 mM pore size Cell Culture Inserts (Greiner) coated with 0.1% collagen. The inserts were incubated at 37 C for 16 hr in 24-well plates containing medium with or without MCP-1 (50 ng/ml) or C3H10T1/2 conditioned medium as chemoattractants. Next, cells on the upper side of the transwell membrane were removed with a cotton swab and the inserts were fixed and stained in a crystal violet solution. The membranes were subsequently mounted on a glass slide, and migrated cells were counted by light microscopy. Cells were counted in five different fields using a 2003 magnification. Cytokines and chemokines measurement
MCP-1 was measured using ELISA (R&D Systems, Minneapolis, MN) according to manufacturer’s instructions with a detection limit of 31.25 pg/ml. Data analysis
Statistical analyses were performed using one-way analysis of variance, followed by the Mann-Whitney U test. Statistical significance was set at p < 0.05.
Results PAR-1 is abundantly expressed in the microenvironment of both primary pancreatic tumors and metastatic sites
To assess the potential importance of PAR-1 in the microenvironment of pancreatic cancer, we determined PAR-1 expression in 13 human pancreatic tumors. As shown in Figure 1,
PAR-1 drives pancreatic cancer progression
the tumor microenvironment is rather desmoplastic as evident from Vimentin, Collagen I and aSMA stainings. Ten of these tumors were PAR-1 positive, and all these tumors showed PAR-1 expression in the microenvironment. Indeed, PAR-1 coincides with the expression pattern of the stromal markers (Fig. 1). Tumor cells were PAR-1 positive in five tumors only (Fig. 1a). Interestingly, PAR-1 is also highly expressed in metastatic lesions in the liver of pancreatic cancer patients (Supporting Information Fig. S1). In all 13 liver biopsies analyzed, PAR-1 is expressed by cells in the tumor microenvironment (fibroblast-like cells, inflammatory cells and endothelial cells) and in 6 of them PAR-1 is also expressed by tumor cells. Genetic ablation of PAR-1 from the tumor microenvironment limits tumor growth and metastasis
Due to the omnipresence of PAR-1 in the microenvironment of human pancreatic tumors, we addressed the possibility that PAR-1 in the microenvironment may drive tumor growth and/or metastasis. To this end, we used a murine orthotopic pancreatic tumor model in which PAR-1 expressing pancreatic tumor cells are injected directly into the pancreas of wildtype or PAR-1 deficient mice (Fig. 2a). Four weeks after tumor cell injection, wildtype mice had large tumors with an average weight of 1.96 6 0.36 g and a volume of 1.95 6 0.43 cm3 (Figs. 2b–2d). Interestingly, tumors in PAR-1 deficient animals were significantly smaller with an average weight of 0.95 6 0.14 g and a volume of 0.94 6 0.15 cm3 (Figs. 2b–2d). Histological examination of hematoxylin and eosin (H&E)-stained sections showed that tumors from wild-type animals appear largely similar to tumors from PAR-1 deficient mice. Interestingly however, cellularity was reduced in the core of PAR-1 deficient tumors (Fig. 2e). In line with this observation, tumor cell proliferation as assessed by Ki67 stainings showed a decreased number of positive cells in the tumor core of PAR-1 deficient mice (Fig. 2f) but not at the periphery of the tumor. Cleaved caspase-3 stainings, showing apoptotic cells, did not show differences between the different tumors (Supporting Information Fig. S2). The majority of deaths following pancreatic cancer are due to metastatic spread and consequently we assessed the importance of PAR-1 in the microenvironment for metastasis. Macrometastasis were observed in the spleen of all seven wild-type mice, whereas six wild-type mice showed duodenal macrometastasis (Fig. 2g). In contrast, none of the PAR-1 deficient animals had metastases in their spleen and only one out of eight showed metastasis to the duodenum. Histological examination of the metastases indicates that the tumor cells are actively invading into the different organs (Fig. 2h). Thus, PAR-1 expression in the tumor microenvironment promotes pancreatic cancer growth and metastasis. PAR-1 expression in the tumor microenvironment enhances angiogenesis
PAR-1 plays a critical role in blood vessel formation during embryonic development31 and tumor angiogenesis.32 As C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
2297
Cancer Cell Biology
Queiroz et al.
Figure 1. PAR-1 expression in primary pancreatic tumors. (a) Paraffin sections obtained from patients with resectable pancreatic cancer were stained for alpha-SMA, Vimentin and Collagen I, and (b) PAR-1 (brown). Tumor cells are indicated by an asterisk whereas stromal cells are indicated by solid arrowheads, vessels are indicated by (v) and inflammatory infiltrate by (i). Depicted is a 1003 magnification of primary tumor tissue. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
angiogenesis is a hallmark of cancer and a prerequisite for tumor growth,33 we evaluated whether the reduced tumor growth in PAR-1 deficient mice did correlate with angiogenic changes. To this end, CD31 expression in pancreatic tumor sections of wild-type and PAR-1 deficient mice was compared. Indeed the number of CD31 positive vessels in wildtype tumors (42.3 6 1.5) was around 1.5-fold higher than that in PAR-1 deficient tumors (27.5 6 0.8; Fig. 3). Interestingly, LYVE-1 expression, a marker for lymphatic vessels, did not differ between wild-type and PAR-1 deficient tumors (Fig. 3). Taken together, these data suggest that PAR-1C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
driven angiogenesis explains the decrease in tumor growth and metastasis in PAR-1 deficient mice. Genetic ablation of PAR-1 from the tumor microenvironment improves response to gemcitabine
Gemcitabine chemotherapy is used as standard of care for pancreatic cancer patients, although the effect on overall survival is small.5 To assess whether PAR-1 expression in the tumor microenvironment affects the intrinsic drug resistance of pancreatic cancer cells, tumor bearing wild-type and PAR1 deficient mice were treated with gemcitabine (Fig. 4a).
PAR-1 drives pancreatic cancer progression
Cancer Cell Biology
2298
Figure 2. PAR-1 in the tumor microenvironment potentiates pancreatic tumor growth and metastasis. (a) Schematic representation of the experimental set-up. (b) Pancreatic tumors derived from wildtype (top) or PAR-1 deficient mice (bottom). (c) Weight and (d) size of the tumors depicted in (b) (mean 6 SEM; n 5 7–8 per group) *p < 0.05. (e) H&E stained sections obtained from tumor bearing wild-type (top) or PAR-1 deficient mice (bottom). (f) Paraffin sections obtained from tumor bearing wildtype or PAR-1 deficient mice stained for Ki67. Quantification of the slides (as indicted in the Material and Methods section) is shown on the right. Indicated is the mean 6 SEM (n 5 7–8 per group). **p < 0.01. (g) H&E stained sections of metastases obtained from tumor bearing wild-type mice. The black line differentiates invading tumor cells from normal cells present in the individual organs. (h) Number of wildtype and PAR-1 deficient mice with macroscopic metastasis. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
Interestingly, gemcitabine treatment reduced tumor growth by around twofold (tumor weight and volume of 0.58 6 0.18 g and 0.58 6 0.21 cm3, respectively) in wildtype animals. Genetic ablation of PAR-1 from the microenvironment potentiated the effect of gemcitabine and treatment almost completely blocked cancer growth in PAR-1 deficient mice
(Figs. 4b and 4c). Indeed, only very small tumors were observed in two PAR-1 deficient mice (weight of 0.03 6 0.02 g and volume of 0.02 6 0.01 cm3), whereas no tumors were observed in the other six PAR-1 deficient mice. Histological examination of the tumors shows highly proliferative tumors in wild-type animals (Fig. 4d). In contrast, on H&E-stained C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
2299
Figure 3. PAR-1 in the tumor microenvironment induces angiogenesis but not lymphangiogenesis. Paraffin sections obtained from tumor bearing wildtype or PAR-1 deficient mice stained for CD31 (angiogenesis) and LYVE-1 (lymphangiogenesis). Quantification of the slides (as indicated in the Material and Methods section) is shown on the right. Indicated is the mean 6 SEM (n 5 7–8 per group). ***p < 0.001. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
Figure 4. PAR-1 in the tumor microenvironment induces drug resistance of pancreatic tumors. (a) Schematic representation of the experimental set-up. (b) Pancreatic tumors or pancreas derived from wild-type (left) or PAR-1 deficient (right) mice after gemcitabine treatment. (c) Weight and size of the tumors depicted in (b) (mean 6 SEM; n 5 8 per group). *p < 0.05; **p < 0.01. (d) H&E stained sections obtained from gemcitabine treated tumor bearing mice. The top left panel shows a proliferative tumor obtained from a wild-type mouse. The top right panel shows one of the small tumors (circled in black) observed in PAR-1 deficient mice. The bottom left panel shows some residual necrotic tumor cells within the pancreas of PAR-1 deficient mice and the bottom right panel shows a completely healthy and tumor cell free pancreas observed in three PAR-1 deficient mice. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
pancreatic tissue slides from PAR-1 deficient animals no tumor cells could be detected in four out of eight mice, whereas a small necrotic tumor was observed in one animal. C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
Our data thus show that PAR-1 expression in the microenvironment of pancreatic tumors limits gemcitabine-induced cell death.
Cancer Cell Biology
Queiroz et al.
2300
Cancer Cell Biology
PAR-1 does not modify the fibrotic response during pancreatic cancer development
Drug resistance of pancreatic cancer is suggested to rely on the characteristic desmoplastic reaction seen in pancreatic tumors34 and cancer-associated stromal fibroblasts do promote pancreatic tumor progression.35 As PAR-1 is emerging as a key factor in fibroproliferative disease,22 PAR-1 driven fibrosis may explain the low efficacy of gemcitabine in wildtype mice. The number of matrix producing alpha-smooth muscle actin (aSMA) positive myofibroblasts was however not different between tumors of wildtype and PAR-1 deficient mice (Supporting Information Fig. S3A). In line with these findings, histological examination of Masson trichromestained pancreatic cancer sections showed that collagen deposition is comparable in tumors of wildtype and PAR-1 deficient mice (Supporting Information Fig. S3B), whereas collagen 1 immunostainings also showed similar amounts of collagen in tumors of wildtype and PAR-1 deficient mice (Supporting Information Fig. S3C). Although collagen deposition is not modified by PAR-1 expression in the microenvironment, differential growth factor expression could explain the difference in gemcitabine efficacy in wildtype versus PAR-1 deficient mice. Consequently, we determined whether fibroblast conditioned medium would protect PANC02 cells from gemcitabine-induced cell death. Indeed, conditioned medium from C3H10T1/2 fibroblasts reduced gemcitabine-induced PANC02 cell death (Supporting Information Fig. S3D). Importantly however, this effect is PAR-1 independent as fibroblast conditioned medium prepared in the presence of a PAR-1 agonist (thrombin) was as effective in preventing gemcitabine-induced cell death as conditioned medium prepared in the absence of PAR-1 agonist (Supporting Information Fig. S3D). Overall, PAR-1 expression in the microenvironment does not modify the desmoplastic reaction and gemcitabine sensitivity in mice that lack PAR-1 in their tumor microenvironment seems not dependent on a different fibrotic response. PAR-1 regulates macrophage recruitment and consequently drives chemoresistance in pancreatic cancer
In addition to fibroblasts and extracellular matrix, the pancreatic cancer microenvironment contains a strong inflammatory component represented by myeloid cells. Tumor-associated macrophages represent a predominant population of inflammatory cells in solid tumors, and these tumor-associated macrophages are important drivers of pancreatic cancer progression and drug resistance.36 Consequently, we assessed macrophage recruitment into tumors of wild-type and PAR-1 deficient mice. Immunohistochemical examination of F4/80stained pancreatic cancer sections showed increased macrophage numbers in tumors of wild-type mice (Fig. 5a). Indeed, macrophage numbers are around threefold lower in mice that lack PAR-1 in their microenvironment (Fig. 5b). Macrophages can be classified as either M1 or M2 polarized and M2-like tumor-associated macrophages are consid-
PAR-1 drives pancreatic cancer progression
ered to be pro-tumorigenic.37 To assess whether macrophage polarization is different in wildtype versus PAR-1 deficient tumors, we performed iNOS (M1 marker) and Arg1 (M2 marker) stainings on the pancreatic cancer slides. As shown in Figure 5c, the Arg1/iNOS ratio is >95% in both wildtype and PAR-1 deficient tumors suggesting PAR-1 does not modify macrophage polarization in pancreatic cancer. Macrophages may still proliferate after differentiation38 and PAR-1-driven macrophage proliferation may thus explain the difference in macrophage numbers in tumors of wildtype and PAR-1 deficient animals. Stimulation of RAW264.7 macrophages by thrombin did however not induce proliferation (Fig. 5c). Alternatively, PAR-1 dependent monocyte/macrophage recruitment into the tumor may explain the observed reduction in macrophage numbers in PAR-1 deficient mice. The PAR-1 agonist thrombin does not act as chemoattractant for RAW264.7 cells (data not shown) but actually induces chemokinesis. Indeed, stimulation of RAW264.7 cells with thrombin potentiated migration towards MCP-1 by around fourfold (Fig. 6a). Importantly, thrombin-driven migration was inhibited by pre-treatment with the specific PAR-1 inhibitor P1pal-12 (Fig. 6a). Chemoattractant production in the microenvironment of the tumor is a prerequisite for macrophage recruitment. Interestingly, fibroblasts secrete MCP-1 and PAR-1 stimulation induced MCP-1 secretion by fibroblasts (Fig. 6b). Interestingly, RAW264.7 cells also secrete MCP-1 which is increased by PAR-1 stimulation and pre-treatment with P1pal-12 inhibit this increase in a dose-dependent manner (Fig. 6b). In line, fibroblast conditioned medium is a chemoattractant for RAW264.7 cells and fibroblast conditioned medium prepared in the presence of thrombin further enhanced migration of RAW264.7 cells (Fig. 6c). Finally, we assessed whether the decreased number of macrophages in tumors derived from PAR-1 deficient animals could be involved in the increased gemcitabine sensitivity. Interestingly, conditioned medium from RAW264.7 cells reduced gemcitabine-induced PANC02 cell death (Fig. 6d). Macrophage induced gemcitabine resistance itself is however PAR-1 independent as RAW264.7 conditioned medium prepared in the presence of thrombin was as effective in preventing gemcitabine-induced PANC02 cell death as conditioned medium prepared in the absence of thrombin (Fig. 6d). Overall, PAR-1 expression in monocyte/macrophage seems to augment macrophage recruitment by the tumor, and the macrophages may subsequently diminish gemcitabineinduced cell death in a PAR-1 independent manner (Fig. 6e).
Discussion Our data suggest a model in which PAR-1 acts on endothelial cells to induce angiogenesis thereby enhancing tumor growth and metastasis, whereas PAR-1 acts on monocytes and tumor associated fibroblasts to induce chemoresistance. C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
2301
Figure 5. PAR-1 regulates macrophage recruitment. (a) Sections obtained from tumor bearing wild-type or PAR-1 deficient mice stained for the macrophage marker F4/80. (b) Quantification of the slides (mean 6 SEM; n 5 7–8 per group). (c) Ratio of the number of Arg1 (M2 marker) and iNOS (M1 marker) positive cells in tumor sections from wildtype or PAR-1 deficient mice. (d) Proliferation of RAW264.7 cells in the presence of thrombin as evaluated by BrdU incorporation and MTT reduction assays (mean 6 SEM of an experiment performed two times in octoplo). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
PAR-1 expression on tumor cells is associated with disease progression and inhibition of tumor cell PAR-1 limits tumor growth and metastasis in experimental animal models.10 Interestingly however, PAR-1 is highly expressed in the microenvironment of pancreatic tumors (Fig. 1). Indeed, in 77% of human pancreatic tumors, PAR-1 is expressed in the microenvironment, whereas PAR-1 is expressed in tumor cells in 38% of these tumors only. Moreover, all liver metastasis examined showed PAR-1 expression in the microenvironment of the tumor. Importantly, PAR-1 expression by cells composing the microenvironment is not an epiphenomenon of tumor growth but actually contributes to tumor progression and metastasis (Fig. 2). Tumors in mice that lack PAR-1 in the microenvironment are reduced in size by around twofold, whereas the number of metastases is reduced around 20-fold. In line with literature showing that PAR-1 drives angiogenesis during embryonic development31 and tumor angiogenesis,19,32 tumors in wild-type mice are better vascularized as compared to tumors in PAR-1 deficient animals (Fig. 3). Considering the critical role of vascularization, our data suggest that (stromal) PAR-1-dependent angiogenesis drives pancreatic tumor growth and metastasis. Although drug responses are related to (reduced) angiogenesis in pancreatic C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
cancer,39 the importance of angiogenesis in pancreatic cancer is under debate40 mainly because of the fact that antiangiogenic treatment with Sunitinib did not reduce tumor burden in pancreatic ductal adenocarcinoma.41 Pancreatic cancer is a devastating disease, which remains refractory to treatment and has consequently the worst survival outcome of human cancer. The poor prognosis and limited treatment effects are largely due to an intrinsic drug resistance of pancreatic tumors. Recent data suggest that the microenvironment of pancreatic tumors plays a crucial role in determining intrinsic chemoresistance.20,40,42 In line, we show that ablation of PAR-1 from the tumor microenvironment sensitizes pancreatic tumors for gemcitabine treatment. Indeed, gemcitabine treatment is highly effective in PAR-1 deficient animals and only in two out of eight mice a tumor was observed macroscopically. However, the volume and weight of these tumors was reduced around 30- and 20-fold compared to tumors in wildtype mice treated with gemcitabine, respectively (Fig. 4). In two of the other PAR-1 deficient mice, some residual tumor cells could be observed microscopically, whereas no tumor cells were observed in the pancreas of the remaining four mice. It is well-accepted that a strong desmoplastic reaction, which is characteristic for pancreatic tumors, may in part explain drug resistance.
Cancer Cell Biology
Queiroz et al.
PAR-1 drives pancreatic cancer progression
Cancer Cell Biology
2302
Figure 6. PAR-1, macrophage recruitment and chemoresistance in pancreatic cancer. (a) Thrombin-induced migration of RAW264.7 cells toward MCP-1 in the absence or presence of P1pal-12 (mean 6 SEM of an experiment performed three times). Representative pictures of cells migrated through the transwell are shown on the right. (b) MCP-1 levels secreted by C3H10T1/2 and RAW264.7 cells after PAR-1 stimulation by thrombin or control stimulation. Shown is the mean 6 SEM of an experiment performed three times. (c) Migration of RAW264.7 cells towards fibroblast conditioned medium (C3H10T1/2 CM, red bars) prepared in the absence or presence of thrombin. RAW264.7 cell migration towards plain medium (control medium, grey bars) without or with thrombin 1 U/ml was used as experimental control. Shown is the mean 6 SEM of an experiment performed three times. Representative pictures of cells migrated through the transwell are shown on the right. (d) RAW264.7 conditioned medium induces PANC02 cells resistance to gemcitabine (compare solid black squares with open white squares) independent from PAR-1 (compare solid black dots with open white dots). Shown is the mean 6 SEM of an experiment performed three times in octoplo. (e) Suggested role of PAR-1 in macrophage dependent drug resistance (1.), whereas it further potentiates monocyte/ macrophage recruitment by enhancing MCP-1 production by fibroblasts and macrophages (2.). Finally, the recruited macrophages induce drug resistance independent of PAR-1 (3.). *p < 0.05, **p < 0.01, ***p < 0.001. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
Indeed, mesenchymal cells of the stroma secrete molecules like cytokines, chemokines and growth factors35 involved in tumor progression, metastasis and drug resistance.21 Moreover, fibrosis in and around the tumor may limit effective drug delivery.34,43 Targeting the tumor stroma by IPI-926, a drug that depletes tumor-associated stromal tissue by inhibiting Hedgehog signaling, increased the intratumoral concentration of gemcitabine and increased overall survival in experimental pancreatic cancer.34 However, a clinical trial combining IPI-926 with gemcitabine was stopped due to lack of efficiency44 questioning the importance of the stroma for efficient drug delivery. Interestingly, PAR-1 emerges as a key driver of fibroproliferative disease22,23,27 and differential PAR-1 expression was shown in stromal fibroblasts from normal, benign and malignant human tissues.45 Consequently we hypothesized that PAR-1 driven fibrosis would induce gemcitabine resistance. Indeed, tumors in both wild-type and PAR-1 deficient mice are desmoplastic with evident collagen deposition but no quantitative differences could be observed between tumors from wild-type or PAR-1 deficient mice (Supporting Information Fig. S3). Moreover, pancreatic cancer cells stimulated with fibroblast conditioned medium were less sensitive for gemcitabine (which is in line with stromal derived growth factor being responsible for drug resistance in pancreatic cancer) but this effect was independent from PAR-1. Thus, PAR-1 does not seem to drive drug resistance of pancreatic tumors due to the induction of fibrosis. Of the different stromal cell types common to solid tumors, tumor-associated macrophages are significant for fostering tumor progression.46 Experimental studies indicate that these macrophages contribute to drug-resistance36,47,48 and increased numbers of tumor-associated macrophages are correlated with therapy failure and poor prognosis.36,49,50 Interestingly, the number of macrophages in tumors derived from PAR-1 deficient mice was reduced compared to tumors from wildtype mice (Fig. 5). The difference in macrophage numbers seems not to depend on macrophage proliferation but may be due to PAR-1 driven monocyte migration. Indeed, monocytes are recruited from the circulation after which they differentiate within the tumor microenvironment into tumor-associated macrophages.37 In line with an important role of PAR-1 in monocyte/macrophage recruitment, PAR-1 stimulation of monocytes enhanced their migration towards MCP-1 (Fig. 6). Interestingly, fibroblasts secrete MCP-1 in a PAR-1
dependent manner thereby initiating monocyte recruitment. Once monocytes have migrated into the tumor microenvironment, they enhance subsequent monocyte migration by secreting MCP-1 in part in a PAR-1-dependent manner. Subsequently, monocytes/macrophages induce drug resistance as evident from experiments in which monocyte conditioned medium reduced gemcitabine-induced pancreatic cancer cell death (Fig. 6d). Overall our data suggest a model (Fig. 6e) in which PAR-1 dependent recruitment of tumorassociated macrophages is dependent on an enhanced migratory capacity of monocytes in combination with increased chemokine production by cells in the tumor microenvironment (both fibroblasts and macrophages). The recruited macrophages subsequently seem to induce drug resistance in a PAR-1 independent manner. In addition to PAR-1, PAR-2 also emerges as a key factor in cancer progression. Interestingly, pancreatic cancer cells express high levels of PAR-251 and activation of PAR-2 induces proliferation52,53 and migration54 of pancreatic cancer cells. Ongoing experiments in our laboratory should elucidate whether these pro-tumorigenic effects of PAR-2 in vitro do translate into reduced pancreatic cancer development in PAR-2 deficient mice. Overall, PAR-1 expression in the microenvironment of pancreatic tumors drives tumor progression and limits drug responses. We suggest that PAR-1 induces gemcitabine resistance of experimental pancreatic tumors by enhancing the intrinsic capacity of monocytes to migrate and by providing a proper tumor microenvironment for monocyte migration. These data point to a novel role of PAR-1 in the tumor microenvironment and indicate that targeting PAR-1 may be particular effective in combination with routine chemotherapy.
Acknowledgements We thank Joost Daalhuisen, Marieke ten Brink and Alex R Musler for expert technical assistance. K.C.S.Q. acquisition of data; analysis and interpretation of data and drafting of the manuscript. K.S. acquisition of data. J.D. acquisition of data. H.L.A. technical support. J.W.W. critical revision of the manuscript and material support. C.J.M.N. acquisition of data and material support D.J.R. critical revision of the manuscript and material support C.A.S. study concept; analysis and interpretation of data; drafting of the manuscript; obtained funding and study supervision.
References 1. 2. 3.
4.
Jemal A, Bray F, Center MM, et al. Global cancer statistics. CA Cancer J Clin 2011;61:69–90. Society AC. Cancer facts & figures. Atlanta: American Cancer Society, 2011. Ghaneh P, Costello E, Neoptolemos JP. Biology and management of pancreatic cancer. Gut 2007; 56:1134–52. Burris HA3, Moore MJ, Andersen J, et al. Improvements in survival and clinical benefit
5.
with gemcitabine as first-line therapy for patients with advanced pancreas cancer: a randomized trial. J Clin Oncol 1997;15: 2403–13. Cunningham D, Chau I, Stocken DD, et al. Phase III randomized comparison of gemcitabine versus gemcitabine plus capecitabine in patients with advanced pancreatic cancer. J Clin Oncol 2009;27: 5513–18.
C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
6.
7.
8.
Heinemann V, Haas M, Boeck S. Systemic treatment of advanced pancreatic cancer. Cancer Treat Rev 2012;38:843–53. Vu TK, Hung DT, Wheaton VI, et al. Molecular cloning of a functional thrombin receptor reveals a novel proteolytic mechanism of receptor activation. Cell 1991;64:1057–68. Ramachandran R, Hollenberg MD. Proteinases and signalling: pathophysiological and therapeutic
Cancer Cell Biology
2303
Queiroz et al.
2304
9.
10.
11.
12.
13.
Cancer Cell Biology
14.
15.
16.
17.
18.
19.
20. 21.
22.
23.
implications via PARs and more. Br J Pharmacol 2008;153(Suppl 1):S263–S282. Boire A, Covic L, Agarwal A, et al. PAR1 is a matrix metalloprotease-1 receptor that promotes invasion and tumorigenesis of breast cancer cells. Cell 2005;120:303–13. Cisowski J, O’Callaghan K, Kuliopulos A, et al. Targeting protease-activated receptor-1 with cellpenetrating pepducins in lung cancer. Am J Pathol 2011;179:513–23. Ghio P, Cappia S, Selvaggi G, et al. Prognostic role of protease-activated receptors 1 and 4 in resected stage IB non-small-cell lung cancer. Clin Lung Cancer 2006;7:395–400. Grisaru-Granovsky S, Salah Z, Maoz M, et al. Differential expression of protease activated receptor 1 (Par1) and pY397FAK in benign and malignant human ovarian tissue samples. Int J Cancer 2005;113:372–8. Black PC, Mize GJ, Karlin P, et al. Overexpression of protease-activated receptors-1,-2, and-4 (PAR-1, -2, and -4) in prostate cancer. Prostate 2007;67:743–56. Eroglu A, Karabiyik A, Akar N. The association of protease activated receptor 1 gene -506 I/D polymorphism with disease-free survival in breast cancer patients. Ann Surg Oncol 2012;19:1365–9. Du X, Wang S, Lu J, et al. Correlation between MMP1-PAR1 axis and clinical outcome of primary gallbladder carcinoma. Jpn J Clin Oncol 2011;41:1086–93. Massi D, Naldini A, Ardinghi C, et al. Expression of protease-activated receptors 1 and 2 in melanocytic nevi and malignant melanoma. Hum Pathol 2005;36:676–85. Fujimoto D, Hirono Y, Goi T, et al. Prognostic value of protease-activated receptor-1 (PAR-1) and matrix metalloproteinase-1 (MMP-1) in gastric cancer. Anticancer Res 2008;28:847–54. Yang E, Boire A, Agarwal A, et al. Blockade of PAR1 signaling with cell-penetrating pepducins inhibits Akt survival pathways in breast cancer cells and suppresses tumor survival and metastasis. Cancer Res 2009;69:6223–31. Villares GJ, Zigler M, Wang H, et al. Targeting melanoma growth and metastasis with systemic delivery of liposome-incorporated protease-activated receptor-1 small interfering RNA. Cancer Res 2008;68:9078–86. Hidalgo M. Pancreatic cancer. N Engl J Med 2010;362:1605–17. Mahadevan D, Hoff Von DD. Tumor-stroma interactions in pancreatic ductal adenocarcinoma. Mol Cancer Ther 2007;6:1186–97. Chambers RC. Procoagulant signalling mechanisms in lung inflammation and fibrosis: novel opportunities for pharmacological intervention? Br J Pharmacol 2008;153(Suppl 1):S367–S378. Howell DC, Johns RH, Lasky JA, et al. Absence of proteinase-activated receptor-1 signaling affords protection from bleomycin-induced lung inflammation and fibrosis. Am J Pathol 2005;166: 1353–65.
PAR-1 drives pancreatic cancer progression
24. Mercer PF, Johns RH, Scotton CJ, et al. Pulmonary epithelium is a prominent source of proteinase-activated receptor-1-inducible CCL2 in pulmonary fibrosis. Am J Respir Crit Care Med 2009;179:414–25. 25. Snead AN, Insel PA. Defining the cellular repertoire of GPCRs identifies a profibrotic role for the most highly expressed receptor, proteaseactivated receptor 1, in cardiac fibroblasts. FASEB J 2012;26:4540–7. 26. Rullier A, Gillibert-Duplantier J, Costet P, et al. Protease-activated receptor 1 knockout reduces experimentally induced liver fibrosis. Am J Physiol Gastrointest Liver Physiol 2008;294:G226–G235. 27. Fiorucci S, Antonelli E, Distrutti E, et al. PAR1 antagonism protects against experimental liver fibrosis. Role of proteinase receptors in stellate cell activation. Hepatology 2004;39:365–75. 28. Trejo J, Connolly AJ, Coughlin SR. The cloned thrombin receptor is necessary and sufficient for activation of mitogen-activated protein kinase and mitogenesis in mouse lung fibroblasts. Loss of responses in fibroblasts from receptor knockout mice. J Biol Chem 1996;271:21536–41. 29. Borensztajn K, Stiekema J, Nijmeijer S, et al. Factor Xa stimulates proinflammatory and profibrotic responses in fibroblasts via proteaseactivated receptor-2 activation. Am J Pathol 2008; 172:309–20. 30. Kim YM, Talanian RV, Billiar TR. Nitric oxide inhibits apoptosis by preventing increases in caspase-3-like activity via two distinct mechanisms. J Biol Chem 1997;272:31138–48. 31. Griffin CT, Srinivasan Y, Zheng YW, et al. A role for thrombin receptor signaling in endothelial cells during embryonic development. Science 2001;293:1666–70. 32. Yin YJ, Salah Z, Maoz M, et al. Oncogenic transformation induces tumor angiogenesis: a role for PAR1 activation. FASEB J 2003;17:163–74. 33. Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases. Nature 2000;407:249–57. 34. Olive KP, Jacobetz MA, Davidson CJ, et al. Inhibition of Hedgehog signaling enhances delivery of chemotherapy in a mouse model of pancreatic cancer. Science 2009;324:1457–61. 35. Hwang RF, Moore T, Arumugam T, et al. Cancerassociated stromal fibroblasts promote pancreatic tumor progression. Cancer Res 2008;68:918–26. 36. Tang X, Mo C, Wang Y, et al. Anti-tumour strategies aiming to target tumour-associated macrophages. Immunology 2012; 138:93-104. 37. Sica A, Bronte V. Altered macrophage differentiation and immune dysfunction in tumor development. J Clin Invest 2007;117:1155–66. 38. Jenkins SJ, Ruckerl D, Cook PC, et al. Local macrophage proliferation, rather than recruitment from the blood, is a signature of TH2 inflammation. Science 2011;332:1284–8. 39. Khorana AA, Ahrendt SA, Ryan CK, et al. Tissue factor expression, angiogenesis, and thrombosis in pancreatic cancer. Clin Cancer Res 2007;13: 2870–5.
40. Feig C, Gopinathan A, Neesse A, et al. The pancreas cancer microenvironment. Clin Cancer Res 2012;18:4266–76. 41. Olson P, Chu GC, Perry SR, et al. Imaging guided trials of the angiogenesis inhibitor sunitinib in mouse models predict efficacy in pancreatic neuroendocrine but not ductal carcinoma. Proc Natl Acad Sci USA 2011;108:E1275–E1284. 42. Hidalgo M, Maitra A. The hedgehog pathway and pancreatic cancer. N Engl J Med 2009;361: 2094–6. 43. Frese KK, Neesse A, Cook N, et al. nab-Paclitaxel potentiates gemcitabine activity by reducing cytidine deaminase levels in a mouse model of pancreatic cancer. Cancer Discov 2012;2:260–9. 44. Queiroz KC, Spek CA, Peppelenbosch MP. Targeting Hedgehog signaling and understanding refractory response to treatment with Hedgehog pathway inhibitors. Drug Resist Updat 2012;15: 211–22. 45. D’Andrea MR, Derian CK, Santulli RJ, et al. Differential expression of protease-activated receptors-1 and -2 in stromal fibroblasts of normal, benign, and malignant human tissues. Am J Pathol 2001;158:2031–41. 46. Ruffell B, Affara NI, Coussens LM. Differential macrophage programming in the tumor microenvironment. Trends Immunol 2012;33:119–26. 47. Zhang P, Gruber A, Kasuda S, et al. Suppression of arterial thrombosis without affecting hemostatic parameters with a cell-penetrating PAR1 pepducin. Circulation 2012;126:83–91. 48. Zhang W, Zhu XD, Sun HC, et al. Depletion of tumor-associated macrophages enhances the effect of sorafenib in metastatic liver cancer models by antimetastatic and antiangiogenic effects. Clin Cancer Res 2010;16:3420–30. 49. Hanada T, Nakagawa M, Emoto A, et al. Prognostic value of tumor-associated macrophage count in human bladder cancer. Int J Urol 2000; 7:263–9. 50. Kurahara H, Shinchi H, Mataki Y, et al. Significance of M2-polarized tumor-associated macrophage in pancreatic cancer. J Surg Res 2011;167: e211–e219. 51. Ikeda O, Egami H, Ishiko T, et al. Expression of proteinase-activated receptor-2 in human pancreatic cancer: a possible relation to cancer invasion and induction of fibrosis. Int J Oncol 2003;22: 295–300. 52. Shimamoto R, Sawada T. A role for proteaseactivated receptor-2 in pancreatic cancer cell proliferation. Int J Oncol 2004; 24:1401-6. 53. Darmoul D, Gratio V, Devaud H, et al. Proteaseactivated receptor 2 in colon cancer trypsininduced mapk phosphorylation and cell proliferation are mediated by epidermal. J Biol Chem 2004;279:20927-34. 54. Shi K, Queiroz KCS, Stap J, Richel DJ, Spek CA. Protease-activated receptor-2 induces migration of pancreatic cancer cells in an extracellular ATPdependent manner. J Thromb Haemost 2013;11: 1892–902.
C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
Protease-activated receptor-1 drives pancreatic cancer progression and chemoresistance Karla C.S. Queiroz1, Kun Shi1, JanWillem Duitman1, Hella L. Aberson1, Johanna W. Wilmink2, Carel J.M. van Noesel3, Dick J. Richel2 and C. Arnold Spek1 1
Center for Experimental and Molecular Medicine, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands Department of Medical Oncology, Academic Medical Center, Amsterdam, The Netherlands 3 Department of Pathology, Academic Medical Center, Amsterdam, The Netherlands 2
Cancer Cell Biology
Protease activated receptor (PAR)-1 expression in tumor cells is associated with disease progression and overall survival in a variety of cancers of epithelial origin; however, the importance of PAR-1 in the tumor microenvironment remains unexplored. Utilizing an orthotopic pancreatic cancer model in which tumor cells are PAR-1 positive whereas stromal cells are PAR-1 negative, we show that PAR-1 expression in the microenvironment drives progression and induces chemoresistance of pancreatic cancer. PAR-1 enhances monocyte recruitment into the tumor microenvironment by regulating monocyte migration and fibroblast dependent chemokine production thereby inducing chemoresistance. Overall, our data identify a novel role of PAR-1 in the pancreatic tumor microenvironment and suggest that PAR-1 may be an attractive target to reduce drug resistance in pancreatic cancer.
Pancreatic cancer is a devastating disease with the worst survival outcome of any human cancer. The incidence, which is around 10 per 100,000 individuals, is rising in developed countries,1 with around 44,000 new cases and 37,000 deaths in the United States in 2011.2 The 5-year survival rate is less than 5%, and overall mortality approaches 99%.3 Based on a small survival benefit of gemcitabine as compared to 5fluorouracil in a Phase III study performed more than 10 years ago,4 gemcitabine is first-line standard of care for inoperable pancreatic cancer but the observed benefits are small5 and seem restricted to patients with a good performance status.6 Protease activated receptor (PAR)-1 is a receptor that belongs to the family of G protein-coupled receptors.7 Four different PARs have been identified of which PAR-1 is the prototype of this family of receptors. Activation of PARs is Key words: protease activated receptors, microenvironment, macrophage, chemoresistance Additional Supporting Information may be found in the online version of this article. Grant sponsor: Top Institute Pharma; Grant number: T1-215; Grant sponsor: Dutch Cancer Society; Grant number: AMC 20094324 DOI: 10.1002/ijc.28726 History: Received 7 Aug 2013; Accepted 2 Jan 2014; Online 16 Jan 2014 Correspondence to: Karla C.S. Queiroz, Center for Experimental and Molecular Medicine, Academic Medical Center, Room L01-145, University of Amsterdam, Meibergdreef 9, NL-1105 AZ Amsterdam, The Netherlands, Tel.: 131-20-5664903, E-mail: [email protected]
C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
unique as it requires proteolytic cleavage rather than ligand binding. Removal of the N-terminal extracellular region releases a new tethered ligand that interacts with the body of the receptor to induce transmembrane signaling thereby triggering a range of signaling pathways affecting multiple pathophysiological responses.8 PAR-1 expression levels are increased in a variety of cancers including breast,9 lung,10,11 ovarian12 and prostate cancer.13 More interestingly, PAR-1 levels are correlated with disease progression and overall survival in breast,14 lung,11 gallbladder,15 melanoma16 and gastric cancer.17 In line with a potential important role of tumor cell PAR-1, pharmacological targeting of PAR-1 in preclinical animal models shows promising effects. Indeed, P1pal-7 (pepducin targeting PAR-1) treatment significantly inhibited growth and metastasis to the lungs18 of breast cancer xenografts. Moreover, P1pal-7 provided inhibition of lung tumor growth in nude mice,10 whereas PAR-1 siRNA decreased tumor growth and metastasis to the lung in a murine melanoma model.19 PAR-1 expression on tumor cells thus seems to drive tumor progression and metastasis. A hallmark of pancreatic cancer is the strong desmoplastic reaction causing dense fibrosis around the tumor cells (reviewed in Ref. 20). Activation of (myo)fibroblasts in the stroma, or tumor microenvironment, leads to the secretion of collagen and other extracellular matrix components. Interestingly, the stroma is not simply a mechanical barrier but actually constitutes a dynamic compartment involved in tumor progression, metastasis and drug resistance.21 Indeed, growth factor production by stromal cells activates oncogenic signaling pathways in proliferating tumor cells leading to worse prognosis and resistance to treatment.
2295
Queiroz et al.
What’s new? Few people survive a diagnosis of pancreatic cancer. In the search for new avenues to attack the disease, researchers have noticed that the receptor PAR-1 promotes tumor growth and proliferation. It is also expressed in the tumor microenvironment, where its role is not yet fully explored. In this study, the authors showed that PAR-1 outside the tumor contributes to drug resistance by bringing monocytes onto the scene. Their results suggest that targeting PAR-1 could boost the effectiveness of chemotherapy against pancreatic cancer.
Material and Methods Patients
Tumor sections of pancreatic cancer patients obtained during surgery were used to prepare a tissue microarray according to the guidelines of the Medical Ethical Committee of the Academic Medical Center of Amsterdam. Biopsies were obtained from pancreatic cancer patients after informed consent. Animals
PAR-1 deficient mice generated as described previously28 and C57BL/6 mice (purchased from Charles River) were maintained at the animal facility of the Academic Medical Center of Amsterdam with free access to food and water. All animal experiments were approved by the facility’s Institutional Animal Care and Use Committee. Orthotopic pancreatic cancer model
Confluent cultures of PANC02 cells (>90% viable; kindly provided by Dr. Schmitz, Universit€atsklinikum Bonn) were incubated with trypsin, pelleted, washed twice in phosphate buffered saline (PBS) and resuspended in 0.9% sterile saline (Sigma, St Louis, MO). Tumor cells (4 3 105) were injected directly into the tail of the pancreas of 8- to 10-week-old mice. Mice were evaluated for changes in body weight and signs of discomfort or morbidity, and they were euthanized C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
4 to 5 weeks after tumor cell injection. The pancreas, including the tumor, was removed, measured and weighed. Metastasis in distant organs was assessed macroscopically by visual examination. If indicated, mice were treated with gemcitabine (120 mg/kg in PBS; i.p.) or saline (control) starting from 7 days after tumor cell injection and this treatment was repeated twice weekly. Both saline and gemcitabine experiments were performed simultaneously. (Immuno)histochemistry
Endogenous peroxidase activity was quenched with 0.3% hydrogen peroxide for 15 min at room temperature, and antigen retrieval was performed for 10 min at 100 C in 10 mM sodium citrate buffer, pH 7.4. Next, slides were blocked for 10 minu with Ultra V block (Thermo Scientific, Runcorn, UK) for PAR-1, F4/80, CD31 and LYVE-1 or for 30 min with 5% (for collagen I) or 10% (for aSMA) normal goat serum. Primary antibodies against PAR-1 (1:200; SC-13503, Santa Cruz, San Diego, CA), F4/80 (1:500, clone CI:A3-1, AbD Serotec, Kidlington, UK), collagen I (1:400, GTX41286, Bioconnect, Huissen, The Netherlands), cleaved caspase 3 (1:100, 9661, Cell Signaling Technology, Boston, MA), Ki67 (1:500, clone Sp6; Neomarkers, Fremont, CA) or aSMA (1:800, Clone 1A4, DAKO, Heverlee, Belgium) were added for overnight incubation at 4 C. For CD31 (1:1,000; sc-1506R, Santa Cruz Biotechnology) or LYVE-1 (1:500; #4509 American Diagnostica; Stamford, CT) primary antibodies were incubated for 90 min at room temperature. Slides were subsequently incubated with appropriate HRP-conjugated secondary antibodies and DAB staining was used to visualize peroxidase activity. Slides were photographed with a microscope equipped with a digital camera (Leica CTR500, Leica Microsystems, Wetzlar, Germany). The number of aSMA positive cells (excluding blood vessels) were counted in 10 different fields at 4003 magnification. CD31 and LYVE-1 positive vessels were counted in 10 different fields at 4003 and 2003 magnification, respectively. Cleaved caspase-3 positive cells were counted in 10 different fields at 4003 magnification. Ki67 positive cells were counted with the ImageJ cell counter tool, positive cells were counted on 2003 magnification pictures in at least five different fields of 0.257 cm2. F4/ 80 staining was analyzed with ImageJ and expressed as percentage of the surface area. The average of 10 pictures at 2003 magnification was used for analysis. Masson trichrome and collagen I stained sections were as graded on a 0 to 3 scale in which 0 indicates a section with no staining, whereas
Cancer Cell Biology
In addition to an important role in tumor cell biology, PAR-1 is emerging as a key player in fibroproliferative disease.22 Indeed, activated PAR-1 induces proliferation, migration and extracellular matrix production of fibroblasts and PAR-1 is upregulated in fibrotic regions in different pathologies.23–25 Importantly, genetic ablation of PAR-1 affords protection from bleomycin-induced lung fibrosis, and limits carbon tetrachloride-induced liver fibrosis26 and bile duct ligation induced cirrhosis.27 PAR-1 expression in the microenvironment of pancreatic cancers may thus drive the desmoplastic reaction thereby promoting tumor growth and chemoresistance. We challenged this hypothesis by evaluating tumor growth, metastasis and drug resistance in an orthotopic pancreatic cancer model in which tumor cells are PAR-1 positive whereas stromal cells are PAR-1 negative. We show that PAR-1 expression in the microenvironment indeed drives tumor growth, metastasis and chemoresistance but the tumor promoting effects of PAR-1 seem independent of the desmoplastic reaction.
2296
1 indicates weak staining, 2 indicates median staining, and 3 indicates intense staining. Cell culture studies
PANC02, RAW264.7 and C3H10T1/2 cells were cultured in RPMI, IMDM and DMEM, respectively, containing 10% FCS, 2 mM L-glutamine and antibiotics. For conditioned medium preparation, cells were grown overnight under normal growth conditions. After 24 hr, the cells were washed, serum starved for 4 hr, and incubated for 24 hr in medium with or without thrombin (1 U/ml) in the presence or absence of P1pal-12 (2.5 lM). For cell viability assays, 1 3 103 PANC02 cells were seeded in 96-well plates and, after overnight attachment, treated for 48 hr with gemcitabine in control or conditioned medium. RAW264.7 cell viability or proliferation was assessed using the same conditions in the absence or presence of thrombin (0–1 U/ml) for 24 hr.
Cancer Cell Biology
Cell viability and proliferation assays
Cell viability was measured using MTT reduction assays29 and by crystal violet staining.30 Cell proliferation was determined with a Cell Proliferation ELISA kit from Roche (BrdU; chemiluminescent) following the manufacturer recommendations. Cell viability and proliferation assays were measured with microplate reader (Synergy HT, BioTek). Transwell migration assays
In detail, 2 3 104 Raw 264.7 cells, serum starved for 4 hr, were transferred to 8 mM pore size Cell Culture Inserts (Greiner) coated with 0.1% collagen. The inserts were incubated at 37 C for 16 hr in 24-well plates containing medium with or without MCP-1 (50 ng/ml) or C3H10T1/2 conditioned medium as chemoattractants. Next, cells on the upper side of the transwell membrane were removed with a cotton swab and the inserts were fixed and stained in a crystal violet solution. The membranes were subsequently mounted on a glass slide, and migrated cells were counted by light microscopy. Cells were counted in five different fields using a 2003 magnification. Cytokines and chemokines measurement
MCP-1 was measured using ELISA (R&D Systems, Minneapolis, MN) according to manufacturer’s instructions with a detection limit of 31.25 pg/ml. Data analysis
Statistical analyses were performed using one-way analysis of variance, followed by the Mann-Whitney U test. Statistical significance was set at p < 0.05.
Results PAR-1 is abundantly expressed in the microenvironment of both primary pancreatic tumors and metastatic sites
To assess the potential importance of PAR-1 in the microenvironment of pancreatic cancer, we determined PAR-1 expression in 13 human pancreatic tumors. As shown in Figure 1,
PAR-1 drives pancreatic cancer progression
the tumor microenvironment is rather desmoplastic as evident from Vimentin, Collagen I and aSMA stainings. Ten of these tumors were PAR-1 positive, and all these tumors showed PAR-1 expression in the microenvironment. Indeed, PAR-1 coincides with the expression pattern of the stromal markers (Fig. 1). Tumor cells were PAR-1 positive in five tumors only (Fig. 1a). Interestingly, PAR-1 is also highly expressed in metastatic lesions in the liver of pancreatic cancer patients (Supporting Information Fig. S1). In all 13 liver biopsies analyzed, PAR-1 is expressed by cells in the tumor microenvironment (fibroblast-like cells, inflammatory cells and endothelial cells) and in 6 of them PAR-1 is also expressed by tumor cells. Genetic ablation of PAR-1 from the tumor microenvironment limits tumor growth and metastasis
Due to the omnipresence of PAR-1 in the microenvironment of human pancreatic tumors, we addressed the possibility that PAR-1 in the microenvironment may drive tumor growth and/or metastasis. To this end, we used a murine orthotopic pancreatic tumor model in which PAR-1 expressing pancreatic tumor cells are injected directly into the pancreas of wildtype or PAR-1 deficient mice (Fig. 2a). Four weeks after tumor cell injection, wildtype mice had large tumors with an average weight of 1.96 6 0.36 g and a volume of 1.95 6 0.43 cm3 (Figs. 2b–2d). Interestingly, tumors in PAR-1 deficient animals were significantly smaller with an average weight of 0.95 6 0.14 g and a volume of 0.94 6 0.15 cm3 (Figs. 2b–2d). Histological examination of hematoxylin and eosin (H&E)-stained sections showed that tumors from wild-type animals appear largely similar to tumors from PAR-1 deficient mice. Interestingly however, cellularity was reduced in the core of PAR-1 deficient tumors (Fig. 2e). In line with this observation, tumor cell proliferation as assessed by Ki67 stainings showed a decreased number of positive cells in the tumor core of PAR-1 deficient mice (Fig. 2f) but not at the periphery of the tumor. Cleaved caspase-3 stainings, showing apoptotic cells, did not show differences between the different tumors (Supporting Information Fig. S2). The majority of deaths following pancreatic cancer are due to metastatic spread and consequently we assessed the importance of PAR-1 in the microenvironment for metastasis. Macrometastasis were observed in the spleen of all seven wild-type mice, whereas six wild-type mice showed duodenal macrometastasis (Fig. 2g). In contrast, none of the PAR-1 deficient animals had metastases in their spleen and only one out of eight showed metastasis to the duodenum. Histological examination of the metastases indicates that the tumor cells are actively invading into the different organs (Fig. 2h). Thus, PAR-1 expression in the tumor microenvironment promotes pancreatic cancer growth and metastasis. PAR-1 expression in the tumor microenvironment enhances angiogenesis
PAR-1 plays a critical role in blood vessel formation during embryonic development31 and tumor angiogenesis.32 As C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
2297
Cancer Cell Biology
Queiroz et al.
Figure 1. PAR-1 expression in primary pancreatic tumors. (a) Paraffin sections obtained from patients with resectable pancreatic cancer were stained for alpha-SMA, Vimentin and Collagen I, and (b) PAR-1 (brown). Tumor cells are indicated by an asterisk whereas stromal cells are indicated by solid arrowheads, vessels are indicated by (v) and inflammatory infiltrate by (i). Depicted is a 1003 magnification of primary tumor tissue. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
angiogenesis is a hallmark of cancer and a prerequisite for tumor growth,33 we evaluated whether the reduced tumor growth in PAR-1 deficient mice did correlate with angiogenic changes. To this end, CD31 expression in pancreatic tumor sections of wild-type and PAR-1 deficient mice was compared. Indeed the number of CD31 positive vessels in wildtype tumors (42.3 6 1.5) was around 1.5-fold higher than that in PAR-1 deficient tumors (27.5 6 0.8; Fig. 3). Interestingly, LYVE-1 expression, a marker for lymphatic vessels, did not differ between wild-type and PAR-1 deficient tumors (Fig. 3). Taken together, these data suggest that PAR-1C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
driven angiogenesis explains the decrease in tumor growth and metastasis in PAR-1 deficient mice. Genetic ablation of PAR-1 from the tumor microenvironment improves response to gemcitabine
Gemcitabine chemotherapy is used as standard of care for pancreatic cancer patients, although the effect on overall survival is small.5 To assess whether PAR-1 expression in the tumor microenvironment affects the intrinsic drug resistance of pancreatic cancer cells, tumor bearing wild-type and PAR1 deficient mice were treated with gemcitabine (Fig. 4a).
PAR-1 drives pancreatic cancer progression
Cancer Cell Biology
2298
Figure 2. PAR-1 in the tumor microenvironment potentiates pancreatic tumor growth and metastasis. (a) Schematic representation of the experimental set-up. (b) Pancreatic tumors derived from wildtype (top) or PAR-1 deficient mice (bottom). (c) Weight and (d) size of the tumors depicted in (b) (mean 6 SEM; n 5 7–8 per group) *p < 0.05. (e) H&E stained sections obtained from tumor bearing wild-type (top) or PAR-1 deficient mice (bottom). (f) Paraffin sections obtained from tumor bearing wildtype or PAR-1 deficient mice stained for Ki67. Quantification of the slides (as indicted in the Material and Methods section) is shown on the right. Indicated is the mean 6 SEM (n 5 7–8 per group). **p < 0.01. (g) H&E stained sections of metastases obtained from tumor bearing wild-type mice. The black line differentiates invading tumor cells from normal cells present in the individual organs. (h) Number of wildtype and PAR-1 deficient mice with macroscopic metastasis. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
Interestingly, gemcitabine treatment reduced tumor growth by around twofold (tumor weight and volume of 0.58 6 0.18 g and 0.58 6 0.21 cm3, respectively) in wildtype animals. Genetic ablation of PAR-1 from the microenvironment potentiated the effect of gemcitabine and treatment almost completely blocked cancer growth in PAR-1 deficient mice
(Figs. 4b and 4c). Indeed, only very small tumors were observed in two PAR-1 deficient mice (weight of 0.03 6 0.02 g and volume of 0.02 6 0.01 cm3), whereas no tumors were observed in the other six PAR-1 deficient mice. Histological examination of the tumors shows highly proliferative tumors in wild-type animals (Fig. 4d). In contrast, on H&E-stained C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
2299
Figure 3. PAR-1 in the tumor microenvironment induces angiogenesis but not lymphangiogenesis. Paraffin sections obtained from tumor bearing wildtype or PAR-1 deficient mice stained for CD31 (angiogenesis) and LYVE-1 (lymphangiogenesis). Quantification of the slides (as indicated in the Material and Methods section) is shown on the right. Indicated is the mean 6 SEM (n 5 7–8 per group). ***p < 0.001. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
Figure 4. PAR-1 in the tumor microenvironment induces drug resistance of pancreatic tumors. (a) Schematic representation of the experimental set-up. (b) Pancreatic tumors or pancreas derived from wild-type (left) or PAR-1 deficient (right) mice after gemcitabine treatment. (c) Weight and size of the tumors depicted in (b) (mean 6 SEM; n 5 8 per group). *p < 0.05; **p < 0.01. (d) H&E stained sections obtained from gemcitabine treated tumor bearing mice. The top left panel shows a proliferative tumor obtained from a wild-type mouse. The top right panel shows one of the small tumors (circled in black) observed in PAR-1 deficient mice. The bottom left panel shows some residual necrotic tumor cells within the pancreas of PAR-1 deficient mice and the bottom right panel shows a completely healthy and tumor cell free pancreas observed in three PAR-1 deficient mice. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
pancreatic tissue slides from PAR-1 deficient animals no tumor cells could be detected in four out of eight mice, whereas a small necrotic tumor was observed in one animal. C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
Our data thus show that PAR-1 expression in the microenvironment of pancreatic tumors limits gemcitabine-induced cell death.
Cancer Cell Biology
Queiroz et al.
2300
Cancer Cell Biology
PAR-1 does not modify the fibrotic response during pancreatic cancer development
Drug resistance of pancreatic cancer is suggested to rely on the characteristic desmoplastic reaction seen in pancreatic tumors34 and cancer-associated stromal fibroblasts do promote pancreatic tumor progression.35 As PAR-1 is emerging as a key factor in fibroproliferative disease,22 PAR-1 driven fibrosis may explain the low efficacy of gemcitabine in wildtype mice. The number of matrix producing alpha-smooth muscle actin (aSMA) positive myofibroblasts was however not different between tumors of wildtype and PAR-1 deficient mice (Supporting Information Fig. S3A). In line with these findings, histological examination of Masson trichromestained pancreatic cancer sections showed that collagen deposition is comparable in tumors of wildtype and PAR-1 deficient mice (Supporting Information Fig. S3B), whereas collagen 1 immunostainings also showed similar amounts of collagen in tumors of wildtype and PAR-1 deficient mice (Supporting Information Fig. S3C). Although collagen deposition is not modified by PAR-1 expression in the microenvironment, differential growth factor expression could explain the difference in gemcitabine efficacy in wildtype versus PAR-1 deficient mice. Consequently, we determined whether fibroblast conditioned medium would protect PANC02 cells from gemcitabine-induced cell death. Indeed, conditioned medium from C3H10T1/2 fibroblasts reduced gemcitabine-induced PANC02 cell death (Supporting Information Fig. S3D). Importantly however, this effect is PAR-1 independent as fibroblast conditioned medium prepared in the presence of a PAR-1 agonist (thrombin) was as effective in preventing gemcitabine-induced cell death as conditioned medium prepared in the absence of PAR-1 agonist (Supporting Information Fig. S3D). Overall, PAR-1 expression in the microenvironment does not modify the desmoplastic reaction and gemcitabine sensitivity in mice that lack PAR-1 in their tumor microenvironment seems not dependent on a different fibrotic response. PAR-1 regulates macrophage recruitment and consequently drives chemoresistance in pancreatic cancer
In addition to fibroblasts and extracellular matrix, the pancreatic cancer microenvironment contains a strong inflammatory component represented by myeloid cells. Tumor-associated macrophages represent a predominant population of inflammatory cells in solid tumors, and these tumor-associated macrophages are important drivers of pancreatic cancer progression and drug resistance.36 Consequently, we assessed macrophage recruitment into tumors of wild-type and PAR-1 deficient mice. Immunohistochemical examination of F4/80stained pancreatic cancer sections showed increased macrophage numbers in tumors of wild-type mice (Fig. 5a). Indeed, macrophage numbers are around threefold lower in mice that lack PAR-1 in their microenvironment (Fig. 5b). Macrophages can be classified as either M1 or M2 polarized and M2-like tumor-associated macrophages are consid-
PAR-1 drives pancreatic cancer progression
ered to be pro-tumorigenic.37 To assess whether macrophage polarization is different in wildtype versus PAR-1 deficient tumors, we performed iNOS (M1 marker) and Arg1 (M2 marker) stainings on the pancreatic cancer slides. As shown in Figure 5c, the Arg1/iNOS ratio is >95% in both wildtype and PAR-1 deficient tumors suggesting PAR-1 does not modify macrophage polarization in pancreatic cancer. Macrophages may still proliferate after differentiation38 and PAR-1-driven macrophage proliferation may thus explain the difference in macrophage numbers in tumors of wildtype and PAR-1 deficient animals. Stimulation of RAW264.7 macrophages by thrombin did however not induce proliferation (Fig. 5c). Alternatively, PAR-1 dependent monocyte/macrophage recruitment into the tumor may explain the observed reduction in macrophage numbers in PAR-1 deficient mice. The PAR-1 agonist thrombin does not act as chemoattractant for RAW264.7 cells (data not shown) but actually induces chemokinesis. Indeed, stimulation of RAW264.7 cells with thrombin potentiated migration towards MCP-1 by around fourfold (Fig. 6a). Importantly, thrombin-driven migration was inhibited by pre-treatment with the specific PAR-1 inhibitor P1pal-12 (Fig. 6a). Chemoattractant production in the microenvironment of the tumor is a prerequisite for macrophage recruitment. Interestingly, fibroblasts secrete MCP-1 and PAR-1 stimulation induced MCP-1 secretion by fibroblasts (Fig. 6b). Interestingly, RAW264.7 cells also secrete MCP-1 which is increased by PAR-1 stimulation and pre-treatment with P1pal-12 inhibit this increase in a dose-dependent manner (Fig. 6b). In line, fibroblast conditioned medium is a chemoattractant for RAW264.7 cells and fibroblast conditioned medium prepared in the presence of thrombin further enhanced migration of RAW264.7 cells (Fig. 6c). Finally, we assessed whether the decreased number of macrophages in tumors derived from PAR-1 deficient animals could be involved in the increased gemcitabine sensitivity. Interestingly, conditioned medium from RAW264.7 cells reduced gemcitabine-induced PANC02 cell death (Fig. 6d). Macrophage induced gemcitabine resistance itself is however PAR-1 independent as RAW264.7 conditioned medium prepared in the presence of thrombin was as effective in preventing gemcitabine-induced PANC02 cell death as conditioned medium prepared in the absence of thrombin (Fig. 6d). Overall, PAR-1 expression in monocyte/macrophage seems to augment macrophage recruitment by the tumor, and the macrophages may subsequently diminish gemcitabineinduced cell death in a PAR-1 independent manner (Fig. 6e).
Discussion Our data suggest a model in which PAR-1 acts on endothelial cells to induce angiogenesis thereby enhancing tumor growth and metastasis, whereas PAR-1 acts on monocytes and tumor associated fibroblasts to induce chemoresistance. C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
2301
Figure 5. PAR-1 regulates macrophage recruitment. (a) Sections obtained from tumor bearing wild-type or PAR-1 deficient mice stained for the macrophage marker F4/80. (b) Quantification of the slides (mean 6 SEM; n 5 7–8 per group). (c) Ratio of the number of Arg1 (M2 marker) and iNOS (M1 marker) positive cells in tumor sections from wildtype or PAR-1 deficient mice. (d) Proliferation of RAW264.7 cells in the presence of thrombin as evaluated by BrdU incorporation and MTT reduction assays (mean 6 SEM of an experiment performed two times in octoplo). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
PAR-1 expression on tumor cells is associated with disease progression and inhibition of tumor cell PAR-1 limits tumor growth and metastasis in experimental animal models.10 Interestingly however, PAR-1 is highly expressed in the microenvironment of pancreatic tumors (Fig. 1). Indeed, in 77% of human pancreatic tumors, PAR-1 is expressed in the microenvironment, whereas PAR-1 is expressed in tumor cells in 38% of these tumors only. Moreover, all liver metastasis examined showed PAR-1 expression in the microenvironment of the tumor. Importantly, PAR-1 expression by cells composing the microenvironment is not an epiphenomenon of tumor growth but actually contributes to tumor progression and metastasis (Fig. 2). Tumors in mice that lack PAR-1 in the microenvironment are reduced in size by around twofold, whereas the number of metastases is reduced around 20-fold. In line with literature showing that PAR-1 drives angiogenesis during embryonic development31 and tumor angiogenesis,19,32 tumors in wild-type mice are better vascularized as compared to tumors in PAR-1 deficient animals (Fig. 3). Considering the critical role of vascularization, our data suggest that (stromal) PAR-1-dependent angiogenesis drives pancreatic tumor growth and metastasis. Although drug responses are related to (reduced) angiogenesis in pancreatic C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
cancer,39 the importance of angiogenesis in pancreatic cancer is under debate40 mainly because of the fact that antiangiogenic treatment with Sunitinib did not reduce tumor burden in pancreatic ductal adenocarcinoma.41 Pancreatic cancer is a devastating disease, which remains refractory to treatment and has consequently the worst survival outcome of human cancer. The poor prognosis and limited treatment effects are largely due to an intrinsic drug resistance of pancreatic tumors. Recent data suggest that the microenvironment of pancreatic tumors plays a crucial role in determining intrinsic chemoresistance.20,40,42 In line, we show that ablation of PAR-1 from the tumor microenvironment sensitizes pancreatic tumors for gemcitabine treatment. Indeed, gemcitabine treatment is highly effective in PAR-1 deficient animals and only in two out of eight mice a tumor was observed macroscopically. However, the volume and weight of these tumors was reduced around 30- and 20-fold compared to tumors in wildtype mice treated with gemcitabine, respectively (Fig. 4). In two of the other PAR-1 deficient mice, some residual tumor cells could be observed microscopically, whereas no tumor cells were observed in the pancreas of the remaining four mice. It is well-accepted that a strong desmoplastic reaction, which is characteristic for pancreatic tumors, may in part explain drug resistance.
Cancer Cell Biology
Queiroz et al.
PAR-1 drives pancreatic cancer progression
Cancer Cell Biology
2302
Figure 6. PAR-1, macrophage recruitment and chemoresistance in pancreatic cancer. (a) Thrombin-induced migration of RAW264.7 cells toward MCP-1 in the absence or presence of P1pal-12 (mean 6 SEM of an experiment performed three times). Representative pictures of cells migrated through the transwell are shown on the right. (b) MCP-1 levels secreted by C3H10T1/2 and RAW264.7 cells after PAR-1 stimulation by thrombin or control stimulation. Shown is the mean 6 SEM of an experiment performed three times. (c) Migration of RAW264.7 cells towards fibroblast conditioned medium (C3H10T1/2 CM, red bars) prepared in the absence or presence of thrombin. RAW264.7 cell migration towards plain medium (control medium, grey bars) without or with thrombin 1 U/ml was used as experimental control. Shown is the mean 6 SEM of an experiment performed three times. Representative pictures of cells migrated through the transwell are shown on the right. (d) RAW264.7 conditioned medium induces PANC02 cells resistance to gemcitabine (compare solid black squares with open white squares) independent from PAR-1 (compare solid black dots with open white dots). Shown is the mean 6 SEM of an experiment performed three times in octoplo. (e) Suggested role of PAR-1 in macrophage dependent drug resistance (1.), whereas it further potentiates monocyte/ macrophage recruitment by enhancing MCP-1 production by fibroblasts and macrophages (2.). Finally, the recruited macrophages induce drug resistance independent of PAR-1 (3.). *p < 0.05, **p < 0.01, ***p < 0.001. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
Indeed, mesenchymal cells of the stroma secrete molecules like cytokines, chemokines and growth factors35 involved in tumor progression, metastasis and drug resistance.21 Moreover, fibrosis in and around the tumor may limit effective drug delivery.34,43 Targeting the tumor stroma by IPI-926, a drug that depletes tumor-associated stromal tissue by inhibiting Hedgehog signaling, increased the intratumoral concentration of gemcitabine and increased overall survival in experimental pancreatic cancer.34 However, a clinical trial combining IPI-926 with gemcitabine was stopped due to lack of efficiency44 questioning the importance of the stroma for efficient drug delivery. Interestingly, PAR-1 emerges as a key driver of fibroproliferative disease22,23,27 and differential PAR-1 expression was shown in stromal fibroblasts from normal, benign and malignant human tissues.45 Consequently we hypothesized that PAR-1 driven fibrosis would induce gemcitabine resistance. Indeed, tumors in both wild-type and PAR-1 deficient mice are desmoplastic with evident collagen deposition but no quantitative differences could be observed between tumors from wild-type or PAR-1 deficient mice (Supporting Information Fig. S3). Moreover, pancreatic cancer cells stimulated with fibroblast conditioned medium were less sensitive for gemcitabine (which is in line with stromal derived growth factor being responsible for drug resistance in pancreatic cancer) but this effect was independent from PAR-1. Thus, PAR-1 does not seem to drive drug resistance of pancreatic tumors due to the induction of fibrosis. Of the different stromal cell types common to solid tumors, tumor-associated macrophages are significant for fostering tumor progression.46 Experimental studies indicate that these macrophages contribute to drug-resistance36,47,48 and increased numbers of tumor-associated macrophages are correlated with therapy failure and poor prognosis.36,49,50 Interestingly, the number of macrophages in tumors derived from PAR-1 deficient mice was reduced compared to tumors from wildtype mice (Fig. 5). The difference in macrophage numbers seems not to depend on macrophage proliferation but may be due to PAR-1 driven monocyte migration. Indeed, monocytes are recruited from the circulation after which they differentiate within the tumor microenvironment into tumor-associated macrophages.37 In line with an important role of PAR-1 in monocyte/macrophage recruitment, PAR-1 stimulation of monocytes enhanced their migration towards MCP-1 (Fig. 6). Interestingly, fibroblasts secrete MCP-1 in a PAR-1
dependent manner thereby initiating monocyte recruitment. Once monocytes have migrated into the tumor microenvironment, they enhance subsequent monocyte migration by secreting MCP-1 in part in a PAR-1-dependent manner. Subsequently, monocytes/macrophages induce drug resistance as evident from experiments in which monocyte conditioned medium reduced gemcitabine-induced pancreatic cancer cell death (Fig. 6d). Overall our data suggest a model (Fig. 6e) in which PAR-1 dependent recruitment of tumorassociated macrophages is dependent on an enhanced migratory capacity of monocytes in combination with increased chemokine production by cells in the tumor microenvironment (both fibroblasts and macrophages). The recruited macrophages subsequently seem to induce drug resistance in a PAR-1 independent manner. In addition to PAR-1, PAR-2 also emerges as a key factor in cancer progression. Interestingly, pancreatic cancer cells express high levels of PAR-251 and activation of PAR-2 induces proliferation52,53 and migration54 of pancreatic cancer cells. Ongoing experiments in our laboratory should elucidate whether these pro-tumorigenic effects of PAR-2 in vitro do translate into reduced pancreatic cancer development in PAR-2 deficient mice. Overall, PAR-1 expression in the microenvironment of pancreatic tumors drives tumor progression and limits drug responses. We suggest that PAR-1 induces gemcitabine resistance of experimental pancreatic tumors by enhancing the intrinsic capacity of monocytes to migrate and by providing a proper tumor microenvironment for monocyte migration. These data point to a novel role of PAR-1 in the tumor microenvironment and indicate that targeting PAR-1 may be particular effective in combination with routine chemotherapy.
Acknowledgements We thank Joost Daalhuisen, Marieke ten Brink and Alex R Musler for expert technical assistance. K.C.S.Q. acquisition of data; analysis and interpretation of data and drafting of the manuscript. K.S. acquisition of data. J.D. acquisition of data. H.L.A. technical support. J.W.W. critical revision of the manuscript and material support. C.J.M.N. acquisition of data and material support D.J.R. critical revision of the manuscript and material support C.A.S. study concept; analysis and interpretation of data; drafting of the manuscript; obtained funding and study supervision.
References 1. 2. 3.
4.
Jemal A, Bray F, Center MM, et al. Global cancer statistics. CA Cancer J Clin 2011;61:69–90. Society AC. Cancer facts & figures. Atlanta: American Cancer Society, 2011. Ghaneh P, Costello E, Neoptolemos JP. Biology and management of pancreatic cancer. Gut 2007; 56:1134–52. Burris HA3, Moore MJ, Andersen J, et al. Improvements in survival and clinical benefit
5.
with gemcitabine as first-line therapy for patients with advanced pancreas cancer: a randomized trial. J Clin Oncol 1997;15: 2403–13. Cunningham D, Chau I, Stocken DD, et al. Phase III randomized comparison of gemcitabine versus gemcitabine plus capecitabine in patients with advanced pancreatic cancer. J Clin Oncol 2009;27: 5513–18.
C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
6.
7.
8.
Heinemann V, Haas M, Boeck S. Systemic treatment of advanced pancreatic cancer. Cancer Treat Rev 2012;38:843–53. Vu TK, Hung DT, Wheaton VI, et al. Molecular cloning of a functional thrombin receptor reveals a novel proteolytic mechanism of receptor activation. Cell 1991;64:1057–68. Ramachandran R, Hollenberg MD. Proteinases and signalling: pathophysiological and therapeutic
Cancer Cell Biology
2303
Queiroz et al.
2304
9.
10.
11.
12.
13.
Cancer Cell Biology
14.
15.
16.
17.
18.
19.
20. 21.
22.
23.
implications via PARs and more. Br J Pharmacol 2008;153(Suppl 1):S263–S282. Boire A, Covic L, Agarwal A, et al. PAR1 is a matrix metalloprotease-1 receptor that promotes invasion and tumorigenesis of breast cancer cells. Cell 2005;120:303–13. Cisowski J, O’Callaghan K, Kuliopulos A, et al. Targeting protease-activated receptor-1 with cellpenetrating pepducins in lung cancer. Am J Pathol 2011;179:513–23. Ghio P, Cappia S, Selvaggi G, et al. Prognostic role of protease-activated receptors 1 and 4 in resected stage IB non-small-cell lung cancer. Clin Lung Cancer 2006;7:395–400. Grisaru-Granovsky S, Salah Z, Maoz M, et al. Differential expression of protease activated receptor 1 (Par1) and pY397FAK in benign and malignant human ovarian tissue samples. Int J Cancer 2005;113:372–8. Black PC, Mize GJ, Karlin P, et al. Overexpression of protease-activated receptors-1,-2, and-4 (PAR-1, -2, and -4) in prostate cancer. Prostate 2007;67:743–56. Eroglu A, Karabiyik A, Akar N. The association of protease activated receptor 1 gene -506 I/D polymorphism with disease-free survival in breast cancer patients. Ann Surg Oncol 2012;19:1365–9. Du X, Wang S, Lu J, et al. Correlation between MMP1-PAR1 axis and clinical outcome of primary gallbladder carcinoma. Jpn J Clin Oncol 2011;41:1086–93. Massi D, Naldini A, Ardinghi C, et al. Expression of protease-activated receptors 1 and 2 in melanocytic nevi and malignant melanoma. Hum Pathol 2005;36:676–85. Fujimoto D, Hirono Y, Goi T, et al. Prognostic value of protease-activated receptor-1 (PAR-1) and matrix metalloproteinase-1 (MMP-1) in gastric cancer. Anticancer Res 2008;28:847–54. Yang E, Boire A, Agarwal A, et al. Blockade of PAR1 signaling with cell-penetrating pepducins inhibits Akt survival pathways in breast cancer cells and suppresses tumor survival and metastasis. Cancer Res 2009;69:6223–31. Villares GJ, Zigler M, Wang H, et al. Targeting melanoma growth and metastasis with systemic delivery of liposome-incorporated protease-activated receptor-1 small interfering RNA. Cancer Res 2008;68:9078–86. Hidalgo M. Pancreatic cancer. N Engl J Med 2010;362:1605–17. Mahadevan D, Hoff Von DD. Tumor-stroma interactions in pancreatic ductal adenocarcinoma. Mol Cancer Ther 2007;6:1186–97. Chambers RC. Procoagulant signalling mechanisms in lung inflammation and fibrosis: novel opportunities for pharmacological intervention? Br J Pharmacol 2008;153(Suppl 1):S367–S378. Howell DC, Johns RH, Lasky JA, et al. Absence of proteinase-activated receptor-1 signaling affords protection from bleomycin-induced lung inflammation and fibrosis. Am J Pathol 2005;166: 1353–65.
PAR-1 drives pancreatic cancer progression
24. Mercer PF, Johns RH, Scotton CJ, et al. Pulmonary epithelium is a prominent source of proteinase-activated receptor-1-inducible CCL2 in pulmonary fibrosis. Am J Respir Crit Care Med 2009;179:414–25. 25. Snead AN, Insel PA. Defining the cellular repertoire of GPCRs identifies a profibrotic role for the most highly expressed receptor, proteaseactivated receptor 1, in cardiac fibroblasts. FASEB J 2012;26:4540–7. 26. Rullier A, Gillibert-Duplantier J, Costet P, et al. Protease-activated receptor 1 knockout reduces experimentally induced liver fibrosis. Am J Physiol Gastrointest Liver Physiol 2008;294:G226–G235. 27. Fiorucci S, Antonelli E, Distrutti E, et al. PAR1 antagonism protects against experimental liver fibrosis. Role of proteinase receptors in stellate cell activation. Hepatology 2004;39:365–75. 28. Trejo J, Connolly AJ, Coughlin SR. The cloned thrombin receptor is necessary and sufficient for activation of mitogen-activated protein kinase and mitogenesis in mouse lung fibroblasts. Loss of responses in fibroblasts from receptor knockout mice. J Biol Chem 1996;271:21536–41. 29. Borensztajn K, Stiekema J, Nijmeijer S, et al. Factor Xa stimulates proinflammatory and profibrotic responses in fibroblasts via proteaseactivated receptor-2 activation. Am J Pathol 2008; 172:309–20. 30. Kim YM, Talanian RV, Billiar TR. Nitric oxide inhibits apoptosis by preventing increases in caspase-3-like activity via two distinct mechanisms. J Biol Chem 1997;272:31138–48. 31. Griffin CT, Srinivasan Y, Zheng YW, et al. A role for thrombin receptor signaling in endothelial cells during embryonic development. Science 2001;293:1666–70. 32. Yin YJ, Salah Z, Maoz M, et al. Oncogenic transformation induces tumor angiogenesis: a role for PAR1 activation. FASEB J 2003;17:163–74. 33. Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases. Nature 2000;407:249–57. 34. Olive KP, Jacobetz MA, Davidson CJ, et al. Inhibition of Hedgehog signaling enhances delivery of chemotherapy in a mouse model of pancreatic cancer. Science 2009;324:1457–61. 35. Hwang RF, Moore T, Arumugam T, et al. Cancerassociated stromal fibroblasts promote pancreatic tumor progression. Cancer Res 2008;68:918–26. 36. Tang X, Mo C, Wang Y, et al. Anti-tumour strategies aiming to target tumour-associated macrophages. Immunology 2012; 138:93-104. 37. Sica A, Bronte V. Altered macrophage differentiation and immune dysfunction in tumor development. J Clin Invest 2007;117:1155–66. 38. Jenkins SJ, Ruckerl D, Cook PC, et al. Local macrophage proliferation, rather than recruitment from the blood, is a signature of TH2 inflammation. Science 2011;332:1284–8. 39. Khorana AA, Ahrendt SA, Ryan CK, et al. Tissue factor expression, angiogenesis, and thrombosis in pancreatic cancer. Clin Cancer Res 2007;13: 2870–5.
40. Feig C, Gopinathan A, Neesse A, et al. The pancreas cancer microenvironment. Clin Cancer Res 2012;18:4266–76. 41. Olson P, Chu GC, Perry SR, et al. Imaging guided trials of the angiogenesis inhibitor sunitinib in mouse models predict efficacy in pancreatic neuroendocrine but not ductal carcinoma. Proc Natl Acad Sci USA 2011;108:E1275–E1284. 42. Hidalgo M, Maitra A. The hedgehog pathway and pancreatic cancer. N Engl J Med 2009;361: 2094–6. 43. Frese KK, Neesse A, Cook N, et al. nab-Paclitaxel potentiates gemcitabine activity by reducing cytidine deaminase levels in a mouse model of pancreatic cancer. Cancer Discov 2012;2:260–9. 44. Queiroz KC, Spek CA, Peppelenbosch MP. Targeting Hedgehog signaling and understanding refractory response to treatment with Hedgehog pathway inhibitors. Drug Resist Updat 2012;15: 211–22. 45. D’Andrea MR, Derian CK, Santulli RJ, et al. Differential expression of protease-activated receptors-1 and -2 in stromal fibroblasts of normal, benign, and malignant human tissues. Am J Pathol 2001;158:2031–41. 46. Ruffell B, Affara NI, Coussens LM. Differential macrophage programming in the tumor microenvironment. Trends Immunol 2012;33:119–26. 47. Zhang P, Gruber A, Kasuda S, et al. Suppression of arterial thrombosis without affecting hemostatic parameters with a cell-penetrating PAR1 pepducin. Circulation 2012;126:83–91. 48. Zhang W, Zhu XD, Sun HC, et al. Depletion of tumor-associated macrophages enhances the effect of sorafenib in metastatic liver cancer models by antimetastatic and antiangiogenic effects. Clin Cancer Res 2010;16:3420–30. 49. Hanada T, Nakagawa M, Emoto A, et al. Prognostic value of tumor-associated macrophage count in human bladder cancer. Int J Urol 2000; 7:263–9. 50. Kurahara H, Shinchi H, Mataki Y, et al. Significance of M2-polarized tumor-associated macrophage in pancreatic cancer. J Surg Res 2011;167: e211–e219. 51. Ikeda O, Egami H, Ishiko T, et al. Expression of proteinase-activated receptor-2 in human pancreatic cancer: a possible relation to cancer invasion and induction of fibrosis. Int J Oncol 2003;22: 295–300. 52. Shimamoto R, Sawada T. A role for proteaseactivated receptor-2 in pancreatic cancer cell proliferation. Int J Oncol 2004; 24:1401-6. 53. Darmoul D, Gratio V, Devaud H, et al. Proteaseactivated receptor 2 in colon cancer trypsininduced mapk phosphorylation and cell proliferation are mediated by epidermal. J Biol Chem 2004;279:20927-34. 54. Shi K, Queiroz KCS, Stap J, Richel DJ, Spek CA. Protease-activated receptor-2 induces migration of pancreatic cancer cells in an extracellular ATPdependent manner. J Thromb Haemost 2013;11: 1892–902.
C 2014 UICC Int. J. Cancer: 135, 2294–2304 (2014) V
