Oncogene (2006) 25, 7434–7439
& 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc
SHORT COMMUNICATION
TRAIL promotes metastasis of human pancreatic ductal adenocarcinoma A Trauzold1, D Siegmund2, B Schniewind1, B Sipos3, J Egberts1, D Zorenkov1, D Emme1, C Ro¨der1, H Kalthoff1,4 and H Wajant2,4 1
Section of Molecular Oncology, Clinic for General Surgery and Thoracic Surgery, University Hospital Schleswig-Holstein, Campus Kiel, Kiel, Germany; 2Department of Molecular Internal Medicine, Medical Clinic and Polyclinic II, University of Wu¨rzburg, Wu¨rzburg, Germany and 3Institute for General Pathology, University Hospital Schleswig-Holstein, Campus Kiel, Kiel, Germany
Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) has attracted considerable attention for its potential use in tumor therapy, as some recombinant variants of this ligand induce apoptosis in tumor cells without harming most normal cells. Here, we show that TRAIL strongly induces the expression of the proinflammatory cytokines interleukin-8 and monocyte chemoattractant protein 1 and enhances the invasion of apoptosis-resistant pancreatic ductal adenocarcinoma cells in vitro by upregulation of the urokinase-type plasminogen activator expression. Most importantly, we also demonstrate for the first time that TRAIL treatment results in strongly increased distant metastasis of pancreatic tumors in vivo. We orthotopically transplanted human pancreatic ductal adenocarcinoma cells to the pancreata of severe combined immunodeficiency mice and observed a dramatic increase in metastatic spread including a sixfold increase in the volume and fourfold increase in the number of liver metastases upon TRAIL treatment. Our results point to the necessity to carefully evaluate in vivo side effects of TRAIL and to select therapy conditions that not only enhance apoptosis induction but in addition prevent proinvasive and proinflammatory non-apoptotic TRAIL signaling. Oncogene (2006) 25, 7434–7439. doi:10.1038/sj.onc.1209719; published online 5 June 2006 Keywords: pancreatic TRAIL
cancer;
metastasis;
invasion;
Tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL), also named Apo-2 ligand, is a typical member of the TNF family that interacts with five members of the TNF receptor superfamily including the death receptors TRAILR1 and TRAILR2 (Almasan and Ashkenazi, 2003). Dependent on aggregation status, membrane association and zinc content, different Correspondence: Dr H Kalthoff, Section of Molecular Oncology, Clinic for General Surgery and Thoracic Surgery, Campus Kiel, Arnold-Heller-Strasse 7, Kiel 24105, Germany. E-mail:
[email protected] 4 These authors contributed equally to this work. Received 17 March 2006; revised 24 April 2006; accepted 27 April 2006; published online 5 June 2006
variants of TRAIL vary in their capability to induce apoptosis (Wajant et al., 2005). For example, membrane TRAIL and aggregated variants of soluble trimeric TRAIL induce apoptosis in primary hepatocytes, whereas non-aggregated trimeric TRAIL does not display any liver toxicity. The latter also does not show significant toxic effects on most other primary cells or after injection in monkeys (Ashkenazi et al., 1999; Walczak et al., 1999; Lawrence et al., 2001; Armeanu et al., 2003). As such ‘hepatocyte-safe’ variants of TRAIL nevertheless induce cell death in a variety of tumor cell lines; TRAIL and TRAIL death receptorspecific agonistic antibodies have attracted considerable attention for their potential use in cancer therapy. Noteworthy, we show that TRAIL and the TRAIL death receptors do not only stimulate apoptosis but can also engage non-apoptotic signaling pathways leading to activation of the protein kinase C, nuclear factor kappaB (NF-kB) and mitogen-activated protein kinases (MAPK) (Trauzold et al., 2001; Siegmund et al., 2002). As these pathways stimulate transcription of genes encoding antiapoptotic, angiogenic, mitogenic and cell migration-stimulating factors, the possibility arises that dependent on the clinical situation, TRAIL treatment may exert protumoral unfavorable effects in patients. Recently, we have shown that TRAIL strongly activates both cell death and NF-kB in Colo357 pancreatic ductal adenocarcinoma (PDAC) cells. (Trauzold et al., 2001). As these effects of TRAIL are inhibitory to each other, selective blockade of each of these responses results in enhancement of the other pathway. In the present work, we analysed putative protumoral effects of TRAIL in these cells. To unmask non-apoptotic TRAIL signaling from superimposed apoptotic mechanisms, we forced Colo357 cells to express Bcl-xL by stable transfection with a retroviral Bcl-xL expression vector. The resulting Colo357/Bcl-xL cells showed similar pathophysiological properties as parental Colo357 cells, except for their profound resistance against apoptosis induction by death receptors and chemotherapeutics in vitro and in vivo (Hinz et al., 2000; Schniewind et al., 2004). In contrast to the observed limited protection of Jurkat T cells from TRAIL-mediated apoptosis by Bcl-2 (Rudner et al., 2005), the overexpression of Bcl-xL did protect PDAC cells even when TRAIL doses up to 1000 ng/ml were
TRAIL promotes human PDAC metastasis A Trauzold et al
7435
used (Hinz et al., 2000). In fact, Bcl-xL protected Colo357 cells from TRAIL-induced apoptosis, even after sensitization with cycloheximide (Figure 1). The observed protection was not caused by changes in the cell surface expression of TRAILR1 and TRAILR2 between both cell lines tested (data not shown). Interestingly, in both, parental Colo357 and Colo357/ Bcl-xL transfectants, TRAIL strongly activated NF-kB and MAPK signaling (data not shown), leading to the mRNA induction of proinflammatory genes including interleukin (IL) 8, monocyte chemoattractant protein 1 (MCP1), urokinase-type plasminogen activator (uPA) as well as IkBa as a bona fide NF-kB target gene (Figure 2a and b). Consequently, we observed a TRAIL-induced secretion of IL8, MCP1 and uPA in both cell lines (Figure 2c). As expected, these responses were stronger in TRAIL-resistant Bcl-xL-overexpressing cells. In accordance with the induction of uPA, we also observed an enhanced invasive potential of TRAIL-treated Colo357/Bcl-xL cells in a cell-based in vitro invasion assay detecting the ability of tumor cells to invade and digest a monolayer of fibroblasts (Figure 3). This TRAIL-induced invasion could be abrogated by neutralizing antibodies against uPA. To determine the possible tumor-promoting effects of the non-apoptotic TRAIL response in vivo, we inoculated severe combined immunodeficiency (SCID) mice orthotopically with Colo357/Bcl-xL cells and treated tumor-bearing mice with recombinant TRAIL or saline 10 days later (according to the scheme in Figure 4a). In accordance with our in vitro data, proinflammatory genes (IL8, MCP1) were induced at the mRNA level in orthotopic Colo357/Bcl-xL tumors 6 h after intraperitoneal
Figure 1 Apoptosis induction by TRAIL in Colo357/wild-type and Colo357/Bcl-xL cells. Parental wild-type Colo357 and Colo357/Bcl-xL cells (established as described by Hinz et al. (2000)) were seeded in 96-well plates (1 104 cells/well). Next day, cells were challenged in triplicates with the indicated concentrations of Killer-TRAIL (Axxora/Alexis, Gru¨nberg, Germany) in the presence or absence of cycloheximide (CHX, 2.5 mM). After additional 16 h, cell viability was determined by crystal violet staining (20% in methanol) for 20 min. After several washes with water, plates were air-dried and the plate-bound dye was dissolved in methanol to measure the absorbance at 595 nm.
injection of TRAIL (Figure 4b). In this study, we used a commercially available aggregated TRAIL (KillerTRAIL) that displayed a higher apoptotic activity compared with soluble trimeric TRAIL on some cell lines (data not shown). Nevertheless, it was well tolerated by all mice treated. Subsequently, to investigate the invasive properties in vivo, tumor cells were orthotopically inoculated in two independent experiments in a total of 34 mice, and 17 of these mice were challenged with TRAIL at day 10, 13 and 16. As control, the remaining 17 animals received saline according to the scheme in Figure 4a. Primary tumors and macroscopic solid metastases were analysed for number, weight and volume. Moreover, ascites formation, peritoneal carcinomatosis and liver micrometastasis were evaluated. The data were statistically analysed by SPSS 11.0 as outlined in the legend of Figure 4. The complete experimental data of all individual animals are shown in tabular form as online Supplementary Information. Upon TRAIL treatment, we observed a profound increase in malignancy of Colo357/Bcl-xL cells. Only a limited therapeutic effect was observed with respect to the mean values of size and volume of the primary tumors (Table 1), but a striking increase in the volume (5.8-fold) and a significant increase in the number (4.1-fold) of liver metastases were found (Table 1 and Figure 4c). Interestingly, hepatic micrometastatic lesions as shown in Figure 4d were also preferentially observed in TRAIL-treated animals. Similarly, the number of spleen metastases increased by a factor of 3. In addition, malignant ascites developed in 70.6% of the TRAIL mice compared to 37.5% in the control group, and macroscopic peritoneal carcinomatosis was exclusively observed in the TRAIL-treated group (Table 1). Our data demonstrate inflammation- and invasionpromoting effects of TRAIL in vitro and in vivo. In our studies, we used Colo357-overexpressing Bcl-xL in order to inhibit apoptosis and to clearly uncover the nonapoptotic TRAIL effects. Similarly, parental Colo357 cells also showed an induction of non-apoptotic signaling after TRAIL treatment, but the specific biological effects of this pathway are difficult to be differentiated as apoptosis induction superimposed non-apoptotic effects in these cells. In vivo, TRAIL-sensitive tumor cells may readily become resistant by challenging interactions with TRAIL-expressing tumor-infiltrating immune cells or the tumor stroma. Accordingly, it has been recently shown that coculture of pancreatic tumor cells with fibroblasts leads to the activation of the antiapoptotic NF-kB pathway and to apoptosis resistance (Mu¨erko¨ster et al., 2004). Indeed, high activity of NF-kB and strong expression of Bcl-xL both are clinical hallmarks of pancreatic tumors with poor prognosis (Miyamoto et al., 1999; Evans et al., 2001). The observed Bcl-xL overexpression in clinical tumor specimens may result from high NF-kB activity as Bcl-xL is a bona fide NF-kB target gene (Chen et al., 2000). TRAIL and other ligands of the TNF family can induce NF-kB activity and thus activate the proinflammatory response. Oncogene
TRAIL promotes human PDAC metastasis A Trauzold et al
7436
Figure 2 Expression of proinflammatory factors in Colo357/wild-type and Colo357/Bcl-xL cells after TRAIL treatment. Cells were stimulated with Killer-TRAIL (100 ng/ml) for 6 h. Total RNA was extracted with peqGOLD RNAPure (PeqLab Biotechnologie GmbH, Erlangen, Germany) according to the manufacturer’s protocol and analysed by ribonulcease protection assay (RPA) using a customer Multi-Probe template set (BD PharMingen, Heidelberg, Germany) (a). Probe synthesis, hybridization and RNase treatment were performed with the RiboQuant Multi-Probe RNase Protection Assay System (BD PharMingen) according to the manufacturer’s recommendations. Protected transcripts were separated by denaturing polyacrylamide gel electrophoresis on 5% acrylamide gels and analysed on a PhosphorImager with the ImageQuant software (Molecular Dynamics, Sunnyvale, CA, USA) (a). Alternatively, cells were stimulated with Killer-TRAIL for 4 h and total RNA was subjected to uPA-specific real-time reverse transcriptase–polymerase chain reaction as described previously (Trauzold et al., 2005) (b). For the determination of protein secretion, Colo357/Bcl-xL cells were treated in triplicates for 16 h with TRAIL (100 ng/ml) and supernatants were analysed using uPA- (American Diagnostica, Pfungstadt, Germany), IL8- or MCP1-Immunoassays (R&D Systems, Wiesbaden, Germany) according to the provided protocols. Measured concentrations were normalized to the cell numbers determined in parallel and are displayed as fold induction compared to untreated cells (c).
In vivo, such response may become highly relevant, especially in cells that gain apoptosis resistance. Consequently, protumoral effects of TRAIL are likely to contribute to tumor progression. Very recently, promotion of migration and invasion in vitro were reported for apoptosis-resistant cholangiocarcinoma cells and were explained as a consequence of TRAIL-induced activation of NF-kB (Ishimura et al., 2006). Moreover, a proinflammatory and proliferationinducing action of TRAIL, mainly attributable to NF-kB activation, has been demonstrated with different TRAIL preparations or TRAILR1- and TRAILR2specific agonistic antibodies (Trauzold et al., 2001, Siegmund et al., 2002; Ehrhardt et al., 2003; Baader et al., 2005). It appears therefore unlikely, albeit Oncogene
desirable, that the beneficial apoptosis-inducing effects of TRAIL can be separated from the unfavorable non-apoptotic properties for the use in tumor therapy either by choosing a special preparation of TRAIL or by direct targeting of TRAILR1, which, from a clinical point of view, has been demonstrated in lymphoid leukemia cells to be most relevant (MacFarlane et al., 2005). In fact, we observed in our SCID mouse orthotopic xenotransplantation model that upon TRAIL administration apoptosis-resistant tumor cells metastasize significantly stronger to distant organs like the liver and spleen. This multistep process clearly implicates more tumor cell activities than can be tested by the current in vitro tests for migration and invasion. Noteworthy,
TRAIL promotes human PDAC metastasis A Trauzold et al
7437
Figure 3 In vitro invasion of Colo357/Bcl-xL cells after treatment with TRAIL and with and without anti-uPA-neutralizing antibodies. The invasive potential of tumor cells was determined using a model for cell invasion based on coculture with fibroblast (Trauzold et al., 2005). Briefly, KiF-5 fibroblasts were grown in 24-well plate (1.5 105 cells/well), permeabilized with 500 ml dimethyl sulfoxide, washed with phosphate-buffered saline and overlaid with tumor cells. After 24 h, cells were treated with TRAIL (R&D Systems; 100 ng/ml), with or without neutralizing anti-uPA antibody (2 mg/ml; American Diagnostica), or culture medium. After 48 h, cells were stained with 0.2% trypan blue (Invitrogen, Karlsruhe, Germany) and photographed. As trypan blue stained the permeabilized cells, the fibroblasts could be distinguished from living carcinoma cells. The observed areas of unstained cells represent regions where fibroblasts were displaced or digested by invasive tumor cells.
the strong increase in distant metastasis in our in vivo model reveals the far-reaching impact of TRAILmediated non-apoptotic signaling effects. Apoptosis induction by ‘hepatocyte-safe’ TRAIL as well as by the more active aggregated TRAIL variants can be significantly enhanced by a variety of established tumor treatments. Consequently, antitumoral therapy with TRAIL is mostly discussed in combination with sensitizing chemotherapeutic drugs, such as doxorubicin, 5-fluorouracil, etoposide or inhibitors of the proteasome, of histone deacetylases (HDACs) or of the NF-kB pathway (Altucci et al., 2001; Wajant et al., 2002; Shankar and Srivastava, 2004; von Haefen et al., 2004; Insinga et al., 2005; Nebbioso et al., 2005). The molecular mechanisms underlying the apoptosis-enhancing effects of these compounds are quite different. Whereas some of these reagents including most conventional chemotherapeutic drugs, retinoids and HDAC inhibitors act predominantly by upregulation of the expression of TRAIL and/or its death receptors, other reagents primarily act by inhibition of antiapoptotic pathways, for example, the NF-kB signaling and apoptosis-inhibitory proteins (Altucci et al., 2001; Wajant et al., 2002; Shankar and Srivastava, 2004; Insinga et al., 2005; Nebbioso et al., 2005). Accordingly, the tentative effects of sensitizing drugs on nonapoptotic TRAIL signaling in apoptosis-resistant cells can also be expected to be very variable. For example, it is conceivable that reagents which act by upregulation of TRAIL or its death receptors also enhance nonapoptotic TRAIL signaling in resistant cells, whereas drugs that block NF-kB activity selectively enhance apoptosis induction by TRAIL without allowing non-apoptotic signaling to become relevant in resistant cells. The use of TRAIL in therapy of cancer, including those malignancies with a particular poor prognosis like PDAC, therefore requires selection of such drugs for TRAIL combination therapies that not only sensitize for apoptosis induction, but in addition block non-apoptotic TRAIL signaling pathways.
Table 1 Summarized computed characteristics of Colo357/BclxL primary tumors and metastases in TRAIL-treated (n ¼ 17) and NaCl-treated (n ¼ 16) SCID/beige mice P-value
Treatment TRAIL Mean Primary tumor volumea (mm3) Primary tumor weight (mg) Number of liver metastases Liver metastases volumea (mm3) Number of spleen metastases Spleen metastases volumea (mm3) Number of mice with malignant ascites (%) Number of mice with peritoneal carcinomatosis (%)
326.7 313.6 2.9 8.1 1.8 5.2
NaCl
Median (min–max) 207.0 295.0 3.0 3.1 1.0 4.0
(118–785) (110–551) (0–9) (0–39) (0–6) (0–23)
12 (70.6) 4 (23.5)
Mean 355.5 391.8 0.7 1.4 0.6 2.9
Median (min–max) 305.0 406.5 0.5 0.5 0.0 0.0
(113–824) (128–691) (0–3) (0–5) (0–3) (0–28)
6 (37.5) 0 (0)
0.396b 0.140b 0.009b 0.042b 0.019b 0.027b 0.056c 0.038c
Abbreviations: SCID, severe combined immunodeficiency; TRAIL, tumor necrosis factor-related apoptosis-inducing ligand. aThe tumor volume was calculated according to the formula of a rotational ellipsoid (V ¼ length height width 0.5236). bWilcoxon–Mann–Whitney U-test; cw2 test. Oncogene
TRAIL promotes human PDAC metastasis A Trauzold et al
7438
Figure 4 TRAIL enhances metastasis of orthotopic PDAC. Orthotopic inoculation of PDAC cells in SCID/beige mice was performed as described before (Tepel et al., 2005). The animals were kept for 10 days until solid orthotopic tumors had developed (inclusion criterion for further analysis). (a) Time course of the animal experiment. (b) RPA analysis of Colo357/Bcl-xL primary tumors. Mice carrying primary tumors were challenged with Killer-TRAIL (15 mg; intraperitoneal, lanes 3–5) or were treated with saline (lanes 1–2). After 6 h, mice were killed and tumors were analysed by RPA for expression of the indicated genes. (c) Box plots of the numbers of liver metastases in TRAIL- (n ¼ 17) and saline-treated (n ¼ 16) animals; thick lines represent medians and asterisks display means. One mouse of the control group was excluded from the study because it did not develop a primary tumor (inclusion criterion). P-value: owing to skewed data distribution (tested by Shapiro–Wilk test) different groups were analysed non-parametrically by Mann–Whitney U-test. As two series of animal experiments were performed, the global level of significance was reduced from 0.05 to 0.025 according to Bonferroni’s inequality correction for multiple tests. Data were analysed using SPSS 11.0 (SPSS Inc., Chicago, IL, USA). (d) Micrometastasis of Colo357/Bcl-xL tumor in the liver (arrow) of a TRAIL-treated animal. Upper inset: macrometastasis in the liver; Met – metastasis, Li – liver parenchyma. Lower inset: macroscopic view of the liver of a TRAIL-treated animal displaying seven macrometastases. Bar mark: 100 mm.
Acknowledgements We thank Beate Bestmann, MA, Reference Center for Quality of Life, University of Kiel, Germany, for helpful advice in
statistical data evaluation. This study was supported by the Deutsche Forschungsgemeinschaft (Grant SFB 415-A3 to HK and Grant SFB 487-B7 to HW) and Deutsche Krebshilfe (Grant 10-1751-Wa 3 to HW).
References Almasan A, Ashkenazi A. (2003). Cytokine Growth Factor Rev 14: 337–348. Altucci L, Rossin A, Raffelsberger W, Reitmair A, Chomienne C, Gronemeyer H. (2001). Nat Med 7: 680–686. Armeanu S, Lauer UM, Smirnow I, Schenk M, Weiss TS, Gregor M et al. (2003). Cancer Res 63: 2369–2372. Ashkenazi A, Pai RC, Fong S, Leung S, Lawrence DA, Marsters SA et al. (1999). J Clin Invest 104: 155–162. Baader E, Toloczko A, Fuchs U, Schmid I, Beltinger C, Ehrhardt H et al. (2005). Cancer Res 65: 7888–7895. Chen C, Edelstein LC, Gelinas C. (2000). Mol Cell Biol 20: 2687–2695. Ehrhardt H, Fulda S, Schmid I, Hiscott J, Debatin KM, Jeremias I. (2003). Oncogene 22: 3842–3852. Evans JD, Cornford PA, Dodson A, Greenhalf W, Foster CS, Neoptolemos JP. (2001). Pancreatology 1: 254–262. Oncogene
von Haefen C, Gillissen B, Hemmati PG, Wendt J, Guner D, Mrozek A et al. (2004). Oncogene 23: 8320–8332. Hinz S, Trauzold A, Boenicke L, Sandberg C, Beckmann S, Bayer E et al. (2000). Oncogene 19: 5477–5486. Insinga A, Monestiroli S, Ronzoni S, Gelmetti V, Marchesi F, Viale A et al. (2005). Nat Med 11: 71–76. Ishimura N, Isomoto H, Bronk SF, Gores GJ. (2006). Am J Physiol Gastrointest Liver Physiol 290: 129–136. Lawrence D, Shahrokh Z, Marsters S, Achilles K, Shih D, Mounho B et al. (2001). Nat Med 7: 383–385. MacFarlane M, Kohlhaas SL, Sutcliffe MJ, Dyer MJS, Cohen GM. (2005). Cancer Res 65: 11265–11270. Miyamoto Y, Hosotani R, Wada M, Lee JU, Koshiba T, Fujimoto K et al. (1999). Oncology 56: 73–82. Mu¨erko¨ster S, Wegehenkel K, Arlt A, Witt M, Sipos B, Kruse ML et al. (2004). Cancer Res 64: 1331–1337.
TRAIL promotes human PDAC metastasis A Trauzold et al
7439 Nebbioso A, Clarke N, Voltz E, Germain E, Ambrosino C, Bontempo P et al. (2005). Nat Med 11: 77–84. Rudner J, Jendrossek V, Lauber K, Daniel PT, Wesselborg S, Belka C. (2005). Oncogene 24: 130–140. Schniewind B, Christgen M, Kurdow R, Haye S, Kremer B, Kalthoff H et al. (2004). Int J Cancer 109: 182–188. Shankar S, Srivastava RK. (2004). Drug Resist Update 7: 139–156. Siegmund D, Wicovsky A, Schmitz I, Schulze-Osthoff K, Kreuz S, Leverkus M et al. (2002). Mol Cell Biol 25: 6363–6379.
Tepel J, Dagvadorj O, Kapischke M, Sipos B, Leins A, Kremer B et al. (2005). Int J Colorectal Dis 16: 1–8. Trauzold A, Ro¨der C, Sipos B, Karsten K, Arlt A, Jiang P et al. (2005). FASEB J 19: 620–622. Trauzold A, Wermann H, Arlt A, Schu¨tze S, Scha¨fer H, Oestern S et al. (2001). Oncogene 20: 4258–4269. Wajant H, Gerspach J, Pfizenmaier K. (2005). Cytokine Growth Factor Rev 16: 55–76. Wajant H, Pfizenmaier K, Scheurich P. (2002). Apoptosis 7: 449–459. Walczak H, Miller RE, Ariail K, Gliniak B, Griffith TS, Kubin M et al. (1999). Nat Med 5: 157–163.
Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).
Oncogene
