Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Resistance of cyclooxygenase-2 expressing pancreatic ductal adenocarcinoma cells against γδ T cell cytotoxicity

Resistance of cyclooxygenase-2 expressing pancreatic ductal adenocarcinoma cells against γδ T... ORIGINAL RESEARCH OncoImmunology 4:3, e988460; March 2015; © 2015 Taylor & Francis Group, LLC Resistance of cyclooxygenase-2 expressing pancreatic ductal adenocarcinoma cells against gd T cell cytotoxicity 1 1 2 2 3 1 Daniel Gonnermann , Hans-Heinrich Oberg , Christian Kellner , Matthias Peipp , Susanne Sebens , Dieter Kabelitz , 1, and Daniela Wesch 1 2 Institute of Immunology; Christian-Albrechts-University; Kiel, Germany; Division of Stem Cell Transplantation and Immunotherapy; Christian-Albrechts-University; Kiel, Germany; Institute for Experimental Medicine; Christian-Albrechts-University; Kiel Keywords: Cox-2, cytotoxicity, gammadelta T lymphocytes, human, pancreatic ductal adenocarcinoma, PGE2 Abbreviations: BrHPP, bromohydrin-pyrophosphate; Cox, cyclooxygenase; n-BP, nitrogen-containing bisphosphonates; PAg, phosphorylated antigen; PDAC, pancreatic ductal adenocarcinoma; PG, prostaglandins; RTCA, Real Time Cell Analyzer; TCR, T cell receptor. The prostaglandin (PG) synthetase cyclooxygenase 2 (Cox-2) promotes tumorigenesis, tumor progression, and metastasis in a variety of human cancer entities including pancreatic ductal adenocarcinoma (PDAC). In this study, we demonstrate that in PDAC cells such as Colo357 cells, enhanced Cox-2 expression and increased release of the Cox-2 metabolite prostaglandin E2 (PGE2) promotes resistance against gd T cell-mediated lysis. Co-culture with activated gd T cells induced an upregulation of Cox-2 expression in Colo357 cells, and thereby an enhanced PGE2 release, in response to tumor necrosis factor a .TNFa/ secretion from gd T cells. The PGE2-mediated inhibition of gd T cell cytotoxicity against Cox-2-expressing PDAC cells can be partially overcome by Cox-2 inhibitors. Our results show that differences between PDAC cells in regards to sensitivity to gd T-cell cytotoxicity can be due to distinct levels of Cox-2 expression associated with varying amounts of PGE2 release. While gd T cell cytotoxicity against PDAC cells expressing low levels of Cox-2 can be effectively enhanced by tribody [(Her2) £Vg9] with specificity for Vg9 T cell receptor and HER-2/neu on PDAC cells, a combination of tribody [(Her2) £Vg9] and Cox-2 inhibitor is necessary to induce complete lysis of Cox-2 high expressing Colo357. In conclusion, our results suggest that the application of tribody [(Her2) £Vg9] that enhances gd T-cell cytotoxicity and Cox-2 inhibitors that overcome PGE2-mediated resistance of PDAC cells to the cytotoxic activity of gd T cells might offer a promising combined immunotherapy for pancreatic cancer. Introduction cells, the function of the numerically small population of gd T lymphocytes is also inhibited by PGE2 secreting and Cox-2 Arachidonic acid is released from plasma membrane phospho- expressing mesenchymal stem cells involved in the development 9,10 lipids and is further converted by cyclooxygenase (Cox) enzymes of tumor-promoting cancer stroma. PGE2 binds to prosta- Cox-1 and Cox-2 to prostaglandins (PG), prostacyclins and glandin E2 and E4 receptors expressed on activated gd T cells, thromboxanes. Cox-1 is constitutively expressed in most mam- and thereby inhibits T cell receptor (TCR)-activated cytotoxic malian tissues, whereas prostaglandin endoperoxidase synthase activity of gd T cells by cyclic adenosine monophosphate enzyme 2 (PTGS2, better known as Cox-2) is induced by a vari- (cAMP)-mediated protein kinase A type I-dependent signaling. 1,2 ety of pro-inflammatory stimuli. Cox-2 is overexpressed in However, such an enhanced PGE2 synthesis could be potently diverse tumor entities including ductal pancreatic adenocarcino- decreased by the Cox-2 inhibitor celecoxib in a pancreatic cancer mas (PDAC), which is mostly associated with a more aggressive xenograft mouse model. Although selective Cox-2 inhibitors 3-5 tumor stage and worse prognosis. Cox-2 overexpression leads such as celecoxib and DuP697 showed anti-proliferative activity to an enhanced production of prostaglandin E2 (PGE2), which toward different tumor entities and against several PDAC cell is involved in tumor progression as well as in tumor evasion of lines in vitro as well as in vivo, celecoxib as a single agent failed to immunosurveillance. The release of PGE2 by cancer cells indu- exert anti-proliferative or antitumor effects toward pancreatic ces and recruits regulatory T cells to the tumor site and suppresses adenocarcinomas resected from patients and implanted in nude immune responses of CD8 ab T cells and natural killer (NK) mice in a xenograft model, despite a decrease in PGE2 synthe- 2,7,8 11-13 cells as well as maturation of dendritic cells. Similar to ab T sis. Nevertheless, selective Cox-2 inhibitors have been *Correspondence to: Daniela Wesch; Email: [email protected] Submitted: 10/16/2014; Revised: 11/08/2014; Accepted: 11/12/2014 http://dx.doi.org/10.4161/2162402X.2014.988460 www.tandfonline.com OncoImmunology e988460-1 included in clinical trials (Phase II) as combinatorial therapies, monolayers can be dynamically and quantitatively monitored on such as to increase the efficiency of the topoisomerase I inhibitor the sensor plate in real time for up to several hours or days with- irinotecan or chemotherapy for the treatment of patients with out the incorporation of labels. After 20 h of adherence, PDAC 12,14,15 advanced or metastatic pancreatic adenocarcinoma. cells were co-cultured with newly established gd T cell lines from PDAC is a highly malignant gastrointestinal tumor character- healthy donors and PDAC patients together with the previously ized by thepresenceofadense desmoplasticstromacomposedof titrated optimal concentrations of the indicated stimuli. By using extracellular matrix and diverse non-neoplastic inflammatory cells the RTCA assay, we were able to distinguish between 3 different mostly displaying an immunosuppressive phenotype. The overall sensitivities of PDAC cells against gd T cell-mediated cytotoxic- 5-year survival rate is less than 5%. The poor prognosis of PDAC ity as follows: (i) lysis of Panc89 cells as well as of Pt45P1 cells by is due to the absence of specific symptoms and a rapid progression gd T cell lines was drastically enhanced in the presence of PAg of this extremely aggressive disease, essentially limiting therapeutic BrHPP or tribody [(Her2) £Vg9)] (Fig. 1A, data not shown) options. Moreover, PDAC is characterized by a profound resis- and the cell index decreased nearly to the level of maximal lysis 16,17 tancetowardcurrent chemo- andtargetedtherapies. We induced by the positive control with 1% Triton X-100, (ii) the recently reported that the adoptive transfer of gd T cells in combi- gd T cell-mediated lysis of PancTu-I cells was slightly enhanced nation with application of a novel tribody [(Her2) £Vg9] by BrHPP, but potentiated much more strongly by reducedgrowthofPDACcells graftedintoseverecombined [(Her2) £Vg9)] (Fig. 1B), (iii) the lysis of Colo357 cells by gd immunodeficiency disease (SCID)-Beige immunodeficient mice. T cells was not increased by BrHPP and only slightly by We designed a unique bispecific antibody [(Her2) £Vg9] in the [(Her2) £Vg9)] (Fig. 1C). Also, [(Her2) £Vg9)] as well as 2 2 2 so-called tribody format that has one binding site for the Vg9T- BrHPP had no effect on PDAC cells in the absence of gd T cell cell receptor (TCR) element on gd T cells and 2 binding sites for lines (data not shown). Additionally, a similarly designed control HER-2/neu (ERBB2) expressed by PDAC cells. Our intention construct [(Her2) £CD89] that does not bind to gd T cells had was to selectively target gd T cells to HER-2/neu-expressing no effect, suggesting that the tribody [(Her2) £Vg9)] triggered tumor cells and to enhance their cytotoxicity more efficiently than target lysis by T cell receptor (TCR) engagement ( and data selective activators of gd T cells, such as phosphorylated antigens not shown). In the course of these studies, we analyzed the cyto- (PAg). Despite promising results in these studies, we identified toxicity against these 4 different PDAC cells with various gd T PDAC cells that were nearly resistant to gd T-cell cytotoxicity. cell lines established from different healthy donors (n D 10) and In this study, we examined whether an intrinsic tumor resis- PDAC patients (n D 15). To demonstrate that the different sen- tance can explain the inability of [(Her2) £Vg9] activated-gd T sitivity of the PDAC cells to gd T cell cytotoxicity was not due to cell lines to completely lyse several PDAC cells lines. We investi- differences between individual gd T cell lines established from gated if overexpression of Cox-2 and enhanced PGE2 release by different donors, gd T cell lines from the same 4 representative PDAC cells was responsible for the resistance of several PDAC donors (2 healthy and 2 PDAC patient) were used for the lysis of cells against gd T-cell cytotoxicity. the various PDAC cells as indicated. Differential Cox-2 expression and PGE2 release by PDAC cells Results Next, we examined whether PDAC cells differ in their Cox-2 expression, thereby releasing different amounts of PGE2 which Different sensitivity of PDAC cells against gd T could be one explanation for the different sensitivity toward gd cell-mediated lysis T cell lysis. The expression of Cox-2, the inducible enzyme in the PGE2 producing pathway, was analyzed in all PDAC cells by Our recent studies revealed that PDAC cells differ in their sensitivity against gd T cell-mediated cytotoxicity. Weak gd T flow cytometry as well as by Western Blot. Regarding the mean cell-mediated lysis of PDAC cells could be enhanced in several fluorescence intensity determined by cytometric fluorimetric but not all PDAC cells by the addition of a novel bispecific anti- analysis, we observed a very low expression of intracellular Cox-2 body [Her2£CD3] or tribody [(Her2) £Vg9]. The aim of in PancTu-I as well as in Panc89 cells compared to Colo357 cells, our current study was to analyze the underlying mechanisms of which highly expressed intracellular Cox-2 (Fig. 2A and B). the differential sensitivity of PDAC cells against gd T cell-medi- These results were verified with an additional Cox-2 antibody by ated lysis. We analyzed 4 PDAC cell lines of distinct origins and Western Blot analysis confirming the highest Cox-2 expression grades of differentiation, which we previously used in our recent level in Colo357 cells (Fig. 2C). The parallel analysis of Cox-1, 18,19 publications. PancTu-I and Pt45P1 cells were initially estab- the constitutively active enzyme of the PGE2 producing pathway, lished from primary tumors (stage G2-G3), whereas Panc89 and revealed a similar expression in all PDAC cell lines (Fig. 2D). Colo357 cells were generated out of lymph node metastases Next, we investigated whether the differences in intracellular (stage G1-G2). For our analysis, we used the previously Cox-2 expression were reflected by differential release of PGE2 described Real Time Cell Analyzer (RTCA) system which uses from PDAC cells. In line with the low intracellular Cox-2 expres- impedance technology. An electronic sensor measures the sion, PancTu-I and Panc89 cells secreted very low amounts of impedance, which reflects changes in cellular parameters of PGE2 (< 0.5 ng/mL), whereas Cox-2 high expressing Colo357 adherent PDAC cells, but not of suspended cells such as gd T cells released relatively high amounts of PGE2 (> 6 ng/mL) cells. Therefore, gd T cell-mediated lysis of PDAC cell (Fig. 2E). e988460-2 OncoImmunology Volume 4 Issue 3 3 Figure 1. Differential susceptibility of PDAC cells against gd T cell-mediated lysis. Five £ 10 (A) Panc89, (B) PancTu-I and (C) Colo357 pancreatic ductal adenocarcinoma (PDAC) cells were cultured in complete medium for 24 h on an E-plate before they were treated with medium (green line) or with gd T cell lines together with medium (dark blue line), 300 nM phosphorylated antigen (PAg) BrHPP (orange line), 1 mg/mL [(Her2) £Vg9)] (light blue line) of 2 representative healthy donor (HD1 and 2) out of 10 and 2 representative PDAC patients (PC1 and 2) out of 15 at an effector to target (E:T) cell ratio of 12.5:1 in a Real Time Cell Analyzer (RTCA) electrical impedance assay over the indicated time. As a positive control for maximal lysis, PDAC cells were treated with Triton X-100. The cell index was determined every 5 min before addition of different substances and gd T cell lines and, thereafter, every minute for > 6 h. The addition of substances or gd T cells is indicated by an arrow. The cell index was normalized to 1 shortly before the time of addition of substances or gd T cell lines as presented by the vertical black line. The average of triplicate determination with the standard deviation (SD) of one rep- resentative experiment is shown. Taken together, the 3 different PDAC cell lines differed sig- different PDAC cell lines directly after seeding of the PDAC cells nificantly from each other regarding Cox-2 expression and PGE2 for 20 h (in comparison to DMSO as a control) and before release. BrHPP-stimulated gd T cell lines were added and cell lysis was determined by RTCA assay (data not shown). Additionally, we tested the effect of PGE2 on PDAC cells alone showing that PGE2 inhibited gd T cell-mediated lysis of PDAC cells PGE2 has no effect on the impedance, and thus on the cell death abolished by Cox-2 inhibitors of PDAC cells, in the absence of gd T cells (data not shown). The enhanced release of PGE2 by Colo357 cells being in line However, the presence of 0.1–1 mg/mL PGE2 strongly reduced with the reduced sensitivity toward gd T cell lysis (Fig. 1) the lysis of Panc89 as well as of PancTu-I cells by BrHPP-stimu- prompted us to examine the effect of PGE2 on the gd T cell- lated gd T cell lines compared to the stimulation in the absence mediated lysis of PDAC cells. Different concentrations (0.1– of PGE2 and in the presence of medium or DMSO control 10 mg/mL) of PGE2 (dissolved in DMSO) were added to the 3 www.tandfonline.com OncoImmunology e988460-3 we observed no difference between Cox-1/2 and Cox- 2 inhibitors on the enhanced lysis suggesting that only Cox-2 expression influences the cytotoxic activity of gd T cells toward PDAC cells. To analyze whether the Cox-inhibitors effect PGE2 production by Colo357, the tumor cells were cultured alone or co-cultured with BrHPP-stimulated gd T cell lines in the presence of the indicated Cox-inhibitor and PGE2 release was deter- mined in the supernatants. Interestingly, PGE2 release by Colo357 alone was not Figure 2. Increased Cox-2 expression and PGE2 release of Colo357. (A, B, and D) Cytofluorimetric analysis to deter- reduced after the addition mine the intracellular expression levels of (A/B) Cox-2 or (C) Cox-1 in the indicated pancreatic ductal adenocarci- of the indicated concentra- noma (PDAC) cells. Protein levels were detected in (A) one representative experiment (bold line; the thin line tions of Cox-inhibitors. indicates the isotype control) and (B/D) in 5 additional experiments (results display the mean § SEM) with an anti- However, the co-culture Cox-1-FITC and anti-Cox-2-PE mAb mixture (AS66/AS67, BD Biosciences) by flow cytometry. (C) Whole protein (20 mg) from the indicated PDAC cells were used for Western Blot, and Cox-2 protein was analyzed with anti-Cox-2 with BrHPP-stimulated gd mAb (33/Cox-2, BD Biosciences). b-actin was used as a loading control. (E) PDAC cells at 1.5 £ 10 cells/well were T cell lines enhanced the incubated in 24-well plates for 24 h. PGE2 release was determined in the supernatant after overnight culture. PGE2 release dramatically Bars § SD present the mean value of 4 independent experiments. which could be significantly reduced in the presence of (Fig. 3A and B and data not shown). In contrast to the lysis of Cox-2 inhibitors (Fig. 4). In line with the enhanced PGE2 PancTu-I and Panc89, the weak lysis of Colo357 cells was virtu- release, Cox-2 expression in Colo357 cells was drastically ally unaltered by additional treatment with PGE2 (Fig. 3C). enhanced after co-culturing them with BrHPP-stimulated gd T These results fit well with the indicated high release of PGE2 by cell lines (data not shown), which underlines the results by others Colo357 (Fig. 2E) and demonstrate that an additional applica- that Cox-2 expression correlates with PGE2 production. tion of PGE2 to the already high level PGE2-producing These data demonstrate that an enhanced PGE2 release Colo357 only slightly affected the nearly complete resistance induced by an intensified expression of Cox-2 in Colo357 cells against gd T cell-mediated lysis (Fig. 3C). co-cultured with gd T cell lines decreased gd T-cell cytotoxicity Since Cox-2 expression goes along with the release of PGE2, that could be abolished by Cox-2 inhibitors. we tested whether Cox inhibitors were able to increase gd T cell- mediated lysis of PDAC cells. First, we studied the effect of dif- TNFa released by gd T cells enhanced Cox-2 expression in ferent concentrations (5, 10, 50 and 200 mM) of the Cox-1/2 Colo357 PDAC cells inhibitor Indomethacin and the selective Cox-2 inhibitor In light of our observations that Colo357 PDAC cells co-cul- DuP697 on PDAC cells alone to exclude toxic effects on the tured with gd T cells displayed enhanced production of PGE2 tumor cells potentially induced by these inhibitors. None of the and intracellular Cox-2 expression taken together with reports by tested concentrations of the inhibitors (dissolved in DMSO) others that Cox-2 expression can be stimulated by various anti- 1,6,22 revealed an influence on PDAC cell growth except 200 mM gens and cytokines, we next set out to investigate the role of DuP697, which caused a growth inhibition on all PDAC cells, cytokines on the regulation of Cox-2 expression in Colo357 cells. and 50 mM DuP697, which reduced Panc89 cell growth (data We focused on tumor necrosis factor a .TNFa/ as well as inter- not shown). feron g .IFNg/ because gd T cell lines produce both cytokines 23,24 Accordingly, pretreatment of PDAC cells with 50 mM Indo- after activation with PAg, such as BrHPP. We found that methacin or of PancTu-I with 50 mM and Panc89 with 10 mM TNFa, but not IFNg, induced a dose-dependent and significant DuP697 revealed no effect on the lysis of these cells by BrHPP- upregulation of intracellular Cox-2 expression in Colo357, but stimulated gd T cell lines (data not shown), whereas the lysis of not in PancTu-I or Panc89 cells (Fig. 5A, data not shown). Colo357 was strongly enhanced by gd T cell lines in the presence When IFNg was added to Colo357 cells, Cox-2 expression of Indomethacin or 50 mM DuP697 (Fig. 3D and E). Moreover, decreased significantly. However, concomitant stimulation of e988460-4 OncoImmunology Volume 4 Issue 3 3 Figure 3. Effect of PGE2 and Cox-2 inhibitors on gd T cell cytotoxicity. Five £ 10 (A) Panc89, (B) PancTu-I and (C-E) Colo357 pancreatic ductal adenocar- cinoma (PDAC) cells were seeded for approximately 24 to 44 h on an E-plate of a Real Time Cell Analyzer (RTCA) electrical impedance assay. After reach- ing a cell index of one, tumor cells were treated with medium (green line) or gd T cell lines at an effector to target (E:T) cell ratio of 25:1 added together with 300 nM phosphorylated antigen (PAg) BrHPP in the presence of medium (dark blue line), DMSO (as solvent control, light blue line) or (A–C)1 mg/ mL prostaglandin E2 (PGE2; pink line) as well as (D-E) 50 mM Indomethacin (violet line) and 50 mM DuP697 (red line). PDAC cells treated with Triton X- 100, which was used as control for maximum cell lysis, are illustrated as a black line. The cell index was analyzed every 5 min during the whole time course and in one min steps for 6 h after addition of the different stimuli, which assigned the time point of cell index normalization (black vertical line). The average of 3 replicates with SD is presented in (A–C) of one representative out of 6 independent experiments or in (D and E) of 2 representative out of 8 experiments. gd T cell lines of the same healthy donor were used in (A, C and E) and from an additional healthy donor in (B and D). Similar reactivity patterns were obtained in 6 to 8 independent experiments with other gd T cell lines of healthy donors as well as of PDAC patients. Colo357 cells with TNFa applied together with IFNg led to a Cox-2 inhibitor DuP697 together with [(Her2) £Vg9] similar increase of Cox-2 expression as TNFa stimulation alone overcome the resistance toward gd T cell-mediated lysis of (Fig. 5A). In accordance with these results, application of the Colo357 TNFa-blocker Infliximab during co-culture of Colo357 cells To investigate whether the addition of the Cox-2 inhibitor and TNFa producing gd T cell lines clearly prevented the DuP697 co-administered together with the tribody increase of Cox-2 expression, an effect that was not observed after [(Her2) £Vg9] could overcome the resistance of Colo357 cells treatment with control Abs (Fig. 5B). toward gd T cell-cytotoxicity, we activated several gd T cell lines The inhibition by Infliximab demonstrates that TNFa from different healthy donors with BrHPP in the absence or pres- released by activated gd T cell lines accounts for the strong ence of DuP697, [(Her2) £Vg9], or with the combination of induction of Cox-2 expression in Colo357 cells. both. As expected, gd T cell lines only weakly lysed the tumor www.tandfonline.com OncoImmunology e988460-5 Figure 5. Enhanced Cox-2 expression in Colo357 induced by TNFa is inhibited by TNFa blocker Infliximab. Cytofluorimetric analysis of the effects of cytokine levels on cyclooxygenase 2 (Cox-2) expression. (A) Figure 4. Upregulation of PGE2 release by Colo357 is inhibited by Cox- Colo357 at 1.5 £ 10 cells/well were seeded in medium or together with inhibitors. The effects of cyclooxygenase (Cox) inhibitors on prostaglan- the indicated concentrations of tumor necrosis factor a .TNFa/, inter- din E2 (PGE2) release by pancreatic cancer cells was tested by culturing feron g (IFNg) or the combination of both in 24-well plates. After 24 h of 1.5 £ 10 Colo357 cells with either medium, 50 mM Indomethacin or culture, mean fluorescence intensity (MFI) § standard error of mean 50 mM DuP697 for 20 h. After pre-treatment, 300 nM phosphorylated (SEM) of Cox-2 expression was determined by intracellular staining of antigen (PAg) BrHPP with or without gd T cell lines at an effector to tar- Colo357 with fluorochrome-conjugated anti-Cox-2 monoclonal antibody get (E:T) cell ratio of 5:1 were added. PGE2 release in the supernatant (mAb) and measured by flow cytometry. After subtraction of isotype con- was determined after 24 h by ELISA. Representative results of experi- trols, statistical analysis of differences in MFI were determined by t-test ments with 4 donors § SD are shown. Statistical analysis was performed and P values were calculated in relation to the medium control in 3 inde- pendent experiments. Levels of significance are presented as *P < 0.05; by Student’s t-test and significance presented as *P < 0.05. **P < 0.01. (B) Colo357 were cultured overnight before the addition of 10 mg/mL Infliximab or 10 mg/mL IgG1 as a control followed by cells after activation with BrHPP. The additional treatment with medium-cultured or phosphorylated antigen (PAg; 300 nM BrHPP) cul- DuP697 or [(Her2) £Vg9] strongly enhanced the cytotoxic 2 tured gd T cell lines from 4 different donors at an effector to target (E:T) cell ratio of 5:1. MFI § SEM of Cox-2 expression of 6 independent experi- activity of gd T cells toward Colo357 cells (Fig. 6). Similar ments are presented. Significances are shown as *P < 0.05. results were obtained with gd T cell lines from PDAC patients (data not shown). In the absence of BrHPP, we observed no enhancing effect of DuP697, whereas [(Her2) £Vg9] with or without BrHPP similarly increased the cytotoxic effects gd T Discussion cells toward Colo357 cells, as we previously showed. Interest- ingly, the combination of DuP697 and [Her2) £Vg9] most 2 Our study indicates that the inhibition of the PGE2 pathway prominently enhanced the gd T cell-mediated lysis of the natu- with Cox-2 inhibitor DuP697 together with [(Her2) £Vg9], an rally resistant Colo357 cells. Similar results were obtained by enhancer of gd T cell cytotoxicity, abolished the resistance of the using gd T cell lines derived from PDAC patients. We conclude PDAC cell line Colo357 against gd T cell-mediated lysis. that the killing of Cox-2 high PDAC cells by gd T cell lines is gd T lymphocytes have raised substantial interest for immu- more efficient in the presence of DuP697 together with notherapy based on their capacity to kill (radio- and chemother- [(Her2) £Vg9] than with [(Her2) £Vg9] alone. 2 2 apy resistant) PDAC cells in an HLA-independent manner. We e988460-6 OncoImmunology Volume 4 Issue 3 Figure 6. [(Her2) £Vg9)] together with Cox-2 inhibitors overcome the resistance of Colo357 toward gd T cell-mediated lysis. After culturing Colo357 overnight, cells were left untreated (green line) or were co-cultured with phosphorylated antigen (PAg; 300 nM BrHPP) stimulated gd T cell lines at an effector to target (E:T) cell ratio of 25:1 in the presence of 50 IU/mL IL-2 with medium (dark blue line), 1 mg/mL [(Her2) £Vg9)] (light blue line), 50 mM DuP697 (red line) or the combination of [(Her2) £Vg9)] and DuP697 (pink line). The cell index (as measured by electrical impedance) was analyzed in 5 min steps over 24 h and was normalized at the time of addition of substances and gd T cell lines. Thereafter, cell index was measured in 1 min steps for 6 h. Five different individual experiments with Colo357 are shown. The arrow indicates addition of substances and/or gd T cells. previously reported that gd T cells in PDAC tissues are predomi- [(Her2) £Vg9] selectively enhanced the cytotoxicity of gd T nantly extensively distributed in the ductal epithelium and the cells in vitro as well as in vivo upon transfer into immunocompro- stroma close to the ductal epithelium, which demonstrates mobi- mised mice. Although our prior study revealed promising lization and infiltration to the tumor site. However, PDAC is results, heterogeneity between the various PDAC cell lines was characterized by the presence of dense desmoplastic stroma com- observed. Several of the PDAC cell lines, such as Colo357 cells, posed of extracellular matrix and diverse (immunosuppressive) were neither completely lysed by gd T cells alone nor in combi- 18,25 cells dampening cytotoxic activity of gd T cells. An efficient nation with enhancers of their cytotoxicity suggesting that the strategy to overcome immunosuppression by the stromal compo- malignant cells themselves actively promote resistance. Therefore, sition on gd T-cell activity could be the usage of enhancers of we analyzed the direct cross-talk between PDAC cells and gd T gd T-cell cytotoxicity. Recently, we demonstrated that cells in the absence of other cell types in vitro. We observed that www.tandfonline.com OncoImmunology e988460-7 the intracellular Cox-2 expression directly correlated with an of the TCR signaling kinase LCK, the Z chain TCR associated 35,36 increased PGE2 release by Colo357 cells and their resistance kinase ZAP-70 and AKT in T cells. Although EP2 and EP4 against gd T cell-mediated lysis. Similar to the report of Marti- receptors share the same signaling pathway, a distinguishing fea- net and colleagues, who demonstrated an inhibitory effect of ture of the EP4 receptor is its activation of the PI3K signaling PGE2 on gd T-cell cytotoxicity, we observed that the addition pathway resulting in subsequent nuclear factor kB (NF-kB) acti- of PGE2 to PDAC cells releasing scarce native PGE2, such as vation and thus TNFa release. Martinet and co-workers Panc89 and PancTu-I cells, robustly inhibited gd T cell-medi- reported that PGE2 inhibits TCR-activated gd T cell-cytotoxic- ated lysis. In contrast, the addition of exogenous PGE2 to ity by a cAMP-mediated protein kinase A type I-dependent sig- PDAC cells that already released high concentrations of native naling. In addition, EP2- and EP4-specific agonists reduced PGE2 only slightly modulated gd T-cell cytotoxicity, suggesting intracellular IFNg production in activated gd T cells compara- that the endogenous release of PGE2 by Colo357 cells is suffi- ble to the addition of exogenous PGE2. An enhanced release of cient to potently inhibit gd T-cell cytotoxicity. This hypothesis IFNg often leads to reduced intracellular stores of IFNg.To was confirmed by the experiment with the inhibitors Indometh- this end, Martinet and colleagues analyzed whether IFNg modu- acin and DuP697 that partially abrogated the resistance of lates the Cox-2 expression or PGE2 secretion of mesenchymal Colo357 against gd T-cell mediated lysis. Furthermore, since stem cells, which are involved in the tumor stroma development the Cox-1/2 inhibitor Indomethacin and the selective Cox-2 and inhibit gd T cell-mediated cytotoxicity by an enhanced inhibitor DuP697 had similar effects on gd T cell-mediated PGE2 release. In their experiments, they observed that 100 ng/ cytotoxicity against PDAC cells we can rule out that Cox-1 is mL IFNg as well as 100 ng/mL TNFa only slightly enhanced involved in this context, although Cox-1 reportedly also contrib- intracellular Cox-2 expression in mesenchymal stem cells, utes to tumorgenesis. Cox-1 is constitutively expressed in whereas the combined addition of both cytokines significantly many tissues, whereas inducible Cox-2 has been reported to be increased Cox-2 expression suggesting that both cytokines overexpressed in »70% of human pancreatic cancer fulfilling an together produced by activated gd T cells are necessary to induce intrinsic role in tumor initiation, development and progression Cox-2 expression in mesenchymal stem cells. Additionally, we through the activation of the phosphatidylinositol 3 kinase observed that IFNg alone rather diminished intracellular Cox-2 1,28,29 (PI3K)/AKT pathway. While anti-Cox-2 therapy has been expression in Colo357 tumor cells, whereas TNFa significantly consistently shown to inhibit PGE2 synthesis in xenografted increased Cox-2 expression. Further, the combination of both pancreatic tumors and in pancreatic cancer patients, conflicting cytokines enhanced Cox-2 expression in Colo357 cells to levels findings have been reported in regards to the antitumor, anti- comparable to that of TNFa stimulation alone. TNFa pro- proliferative and anti-angiogenic effect of Cox-2 inhibitors in duced by macrophages has been reported to enhance Cox-2 xenografted pancreatic carcinomas, a result that may be due to expression in tissues normally Cox-2 negative. In our study, the usage of PDAC cell lines versus primary PDAC cells in these the importance of TNFa in enhancing Cox-2 expression in 11,13 two contradictory reports. In our experiments, we did not Colo357 cells was supported by the effect of the anti-TNFa focus on the direct toxic effect of Cox-2 inhibitors on PDAC blockers Infliximab and Adalimumab (Fig. 5B and data not cells but on the effect on gd T cell-mediated cytotoxicity. Based shown) that abrogated the enhanced intracellular Cox-2 expres- on previous titration results, we used Cox-2 inhibitors at con- sion in Colo357 cells mediated by exogenous TNFa. In line centrations that were not toxic for PDAC cells but potently with the concept that Infliximab specifically reduced the TNFa- inhibited their PGE2 release. This allowed us to examine the mediated Cox-2 up-regulation but not the constitutive expres- effect of PGE2 on gd T cell-mediated lysis of Colo357 in more sion of Cox-2, the addition of Infliximab in the cytotoxic assay detail. We observed that the co-culture of activated gd T cell only marginally affected the gd T cell-cytotoxicity against lines with Colo357 cells enhanced the PGE2 release, which Colo357 cells (data not shown). Besides the potential upregula- could be significantly inhibited by the Cox-inhibitors Indometh- tion of Cox-2 expression in pancreatic tumors by cytokines, acin and DuP697. As shown by others, PGE2 binds to specific other factors such as genetic alterations during the pancreatic G-protein-coupled receptors termed prostaglandin E receptor 2 carcinogenesis or mutations in p53 and KRAS could play a role. (PTGER2, better known as EP2) and prostaglandin E receptor The oncogene KRAS (also known as Ki-ras), for example, has 4 (PTGER4, better known as EP4), both of which are expressed been previously described to stimulate Cox-2 expression. on activated gd T cells. Activation of EP2 as well as the EP4 EP receptor antagonists have been tested for their ability to 40,41 receptor induces adenylate cyclase and thereby the concomitant suppress breast cancer metastasis. Although the targeted 6,30,31 increase in the secondary messenger cAMP. Subsequently, modulation of T-cell function by blocking specific EP receptor cAMP mediates the dissociation of the regulatory and catalytic signaling seems to be a promising approach, there are several lim- subunits of protein kinase A, which subsequently initiates the itations in the usage of EP receptor antagonist for therapy. EP corresponding transactivation of the transcription factor cAMP receptor antagonists display many compensatory and opposing responsive element binding (CREB). CREB proteins are essen- roles, such as their ability to ameliorate not only Th1 responses tial regulators for T-cell function and cytokine production (e.g., but also Th17 responses. Moreover, these agents typically 33,34 of IFNg). Moreover, PGE2-induced cAMP was described inhibit only one or 2 specific receptors, which might be not as to also activate the raft-associated enzyme c-src tyrosine kinase efficient as Cox-inhibitors, which inhibit all downstream (CSK), which negatively regulates the phosphorylating activities prostaglandins. e988460-8 OncoImmunology Volume 4 Issue 3 Several Cox-2 inhibitors such as Celecoxib (Celebrex ) have Prof. Dr. Kalthoff, Section of Molecular Oncology, Kiel. The been included in several clinical protocols, e.g., in a clinical Phase genotype of PDAC-cell lines was recently confirmed by short tan- II study combined with chemotherapy (paclitaxel, carboplatin) dem repeats analysis. 0.05% trypsin/ 0.02% EDTA was used to and radiotherapy for patients with inoperable stage IIIA/B non- remove adherent PDAC cells from flasks. small-cell lung cancer or with chemotherapy (gemcitabine) com- bined with topoisomerase I inhibitor irinotecan or not for Flow cytometry patients with advanced or metastatic pancreatic adenocarci- For the analysis of the purity of the gd T cell lines, cells were 14,15,43,44 noma. However, the application of Celecoxib over a stained with the following phycoerythrin (PE), allophycocyanin prolonged period of time seems to be critical since it may cause (APC), fluorescein isothiocyanate (FITC) or Per-CP fluoro- cardiovascular side effects. Moreover, the effect of Celecoxib chrome-conjugated monoclonal antibodies (mAb): anti-CD3 12,45 alone induced different effects on pancreatic cancer cells. In (clone SK7; PE: #345765 or APC: # 344812, BD Biosciences), our experiments, we used different Cox inhibitors, which did not anti-TCRgd .clone 11F2; FITC: #347903 or APC: custom induce cell death or cell cycle arrest in PDAC cells in the used research conjugate, BD Biosciences), anti-TCRab .clone IP26, concentrations but inhibited the PGE2 release. Our data demon- #306706, Biolegend), anti-TCRVd2 (clone Immu389, FITC: strate that a combined immunotherapy with Cox-2 inhibitors #PNIM1464, Beckman Coulter or clone B6, PerCP: #331410), together with the newly designed tribody [(Her2) £Vg9] gives anti-TCRVd1 (clone TS8.2, #TCR2730, Thermo Fisher Scien- us a tool to efficiently induce the gd T cell-mediated lysis of tific) or corresponding isotype controls (BD Biosciences or Colo357 PDAC cells. Biolegend). 5 6 Taken together, our results indicate that the usage of the tri- For intracellular staining, 10 -10 tumor cells were washed, body [(Her2) £Vg9], which re-directs gd T cells to target cancer permeabilized and fixed with Cytofix/Cytoperm kit (#554714, cells and enhances their cytotoxic activity toward PDAC cells, BD Biosciences) and stained with 20 mL anti-Cox-1-FITC/anti- administered together with Cox-2 inhibitors that reduce PGE2 Cox-2-PE mAb (clone AS70/AS67, #334090, BD Biosciences) production might be a promising approach to target gd T cell- or the appropriate isotype controls. After washing, all samples resistant Cox-2 expressing PDAC cells. were analyzed on a FACS Calibur flow cytometer (BD Bioscien- ces) using Cell Quest Pro software. To analyze the modulation of intracellular Cox-2 expression, 1.5 £ 10 / 500 mL of PDAC cells Materials and Methods were seeded in complete medium in 24-well plates together with a final concentration of 10 or 100 ng/mL recombinant TNFa or Isolation of PBMC and establishment of gd T cell lines IFNg (#210-TA/CF or 285-IF, R&D Systems) or the combina- Short-term gd T cell lines were established from peripheral tion of both for 24 h before intracellular Cox-2 expression was blood mononuclear cells (PBMC) of healthy donors or patients determined by flow cytometry. For blocking experiments during with PDAC. PBMC were isolated from leukocyte concentrates co-culture of Colo357 cells with gd T cell lines (effector/target or from heparinized blood by Ficoll-Hypaque (#6115, Bio- ratio of 5:1), the chimeric mAb Infliximab (Remicade , Janssen chrom) density gradient centrifugation. Blood from healthy Biologics B.V.) as well as the human mAb Adalimumab donors was provided by the Department of Transfusion Medi- (Humira , Abbott GmbH & Co KG), both of which bind cine or the Institute of Immunology, whereas heparinized blood TNFa, were used to prevent the activation of the TNFa recep- from PDAC patients was obtained from the Department of Gen- tor. As an irrelevant control human IgG (#1–001-A, R&D Sys- eral Surgery of the municipal hospital. Informed consent was tems) was added. The optimal concentrations of the mAb for obtained from all donors, and the research was approved by the inhibition were previously titrated using final concentrations of relevant institutional review boards (code number: D 405/10, 2, 5, 10 and 50 mg/mL. D404/14). PBMC were cultured in RPMI 1640 supplemented with Western blot analysis 2 mM L-glutamine, 25 mM HEPES, antibiotics, 10% FCS For detection of Cox-2 in PDAC cells by Western Blot, (complete medium), and stimulated with 5 mM Nitrogen-con- PDAC cells were cultured for 24 h, trypsinized and lysed with taining-bisphosphonate (n-BP) zoledronic acid (Novartis) to TNE lysis buffer (50 mM Tris, 1% (v/v) NP-40, 150 mM selectively activate Vg9Vd2 gd T lymphocytes within the NaCl, 2 mM EDTA and protease inhibitors sodium fluoride PBMC. Since initially Vg9Vd2 gd T cells produced only low [10 mM] and Na VO [1 mM]). Protein concentration was ana- 3 4 amounts of IL-2, 50 U/mL rIL-2 (Novartis) were added every 2 lyzed by Coomassie Protein Assay Reagent (#1856209, Thermo d over a culture period of 14 to 21 d After the culture period, Fisher Scientific). Twenty mg of protein was separated by 10% most short-term activated gd T cell lines (termed gd T cell lines) SDS-PAGE, transferred to nitrocellulose membrane comprise > 97% Vg9Vd2 gd T cells as analyzed by flow (#10600001, GE Healthcare) and detected with 1 mg/mL cytometry. unconjugated anti-Cox-2 mAb 33/Cox-2 (#G10203, BD Biosciences) followed by peroxidase (POD)-conjugated rabbit PDAC cell lines anti-mouse (0.24 mg/mL, #315–005–020, Dianova) and chemi- Colo357, Pt45P1, PancTu-I and Panc89 cells were cultured luminescence detection reagent (ECL Plus, #RPN2106, GE in complete medium. All PDAC cells were kindly provided by Healthcare). As control for protein load, Ab-linked POD was www.tandfonline.com OncoImmunology e988460-9 inactivated by 15% H O and reprobed with anti-b-Actin mAb plates for 20 h, and were stimulated with either medium or the 2 2 (AC-15, 0.29 mg/mL, #A5441, Sigma-Aldrich). previously titrated and indicated concentrations of Cox-inhibi- tors Indomethacin or DuP697 in the absence or presence of Real time cell analyzer BrHPP-stimulated gd T cell lines (at an E/T ratio of 5:1) Five £ 10 adherent PDAC cells were seeded in 96-well E- together with 12.5 U/mL rIL-2 for the last 4 h. After a total plates covered at the bottom with gold-electrodes (ACEA), which incubation time of 24 h, supernatants were collected and PGE2 measured the impedance changes of attached PDAC cells. The was measured by Prostaglandin E Parameter Assay Kit in dupli- impedance reflects cellular parameters such as cell number, size cates following the procedures outlined by the manufacturer and shape, which was monitored with the Real Time Cell Ana- (#SKGE004B, R&D Systems). lyzer (RTCA) single-plate (SP) system (ACEA) every 15 min for up to 24 h. The RTCA-SP system displays impedance in arbi- Statistical analysis trary cell index units, which correlates with cell number. After Since no violation of normal distribution assumption was the PDAC cells reached linear growth phase, PGE2 (#2296/10, found (Shapiro-Wilk test), all statistical comparisons were done Tocris, R&D Systems), Indomethacin (#1708/100, Cox-1/2 parametrically by using t-tests. All statistical tests were 2-sided inhibitor), DuP697 (#1430/10, Cox-2 inhibitor) (both from and the level of significance was set at 5%, not corrected for mul- Tocris, R&D Systems), PAg Bromohydrin-pyrophosphate tiple testing. (BrHPP, selective Vg9Vd2 gd T cell stimulator, Innate Pharma), tribody [(Her2) £Vg9] [consisting of 2 single chain variable fragments derived from trastuzumab (Herceptin) V-regions genetically fused to a Vg9 Fab fragment via flexible linkers ]or Disclosure of Potential Conflicts of Interest 1:500 DMSO (#1.02952.1000, Merck) were pre-incubated over No potential conflicts of interest were disclosed. night or added directly in the previously titrated optimal concen- trations as indicated. Thereafter, gd T cell lines as effector cells together with 12.5 IU/mL IL-2 were added to the RTCA at the indicated effector/target (E/T) ratio. Cell index was used to quan- Acknowledgments titatively monitor cytotoxicity induced by added compounds This work forms part of the master thesis of DG. We grate- and/or gd T cells, which themselves did not interact with the fully acknowledge Dr. Ilka Vogel and Elfi Jerg for organizing and electrodes due to their non-adherent phenotype. Whenever effec- providing blood from PDAC patients. We also thank Dr. Holger tor cells induced lysis of the tumor cells, the loss of impedance of Kalthoff and Dr. Christian R€oder for providing PDAC cell lines. tumor cells was recorded. The cells were monitored every minute BrHPP was kindly provided by Innate Pharma (Marseille, for 6 h and, thereafter, every 5 min for up to an additional 20 h. France). 1% Triton X-100 (X100, Sigma) was used as a positive control for induction of maximal lysis. Cell index was analyzed using the RTCA software 1.2 (www.aceabio.com). Funding ELISA This work was supported by the Medical Faculty of Kiel Uni- To quantify PGE2 released by PDAC cells, 1.5 £ 10 PDAC versity (DW), and the DFG Pancreatic Cancer Consortium Kiel cells in 500 mL of complete medium were seeded into 24-well (DK, DW; WE 3559/2–1; SS; SE 1831/4–1). References 5. Koshiba T, Hosotani R, Miyamoto Y, Wada M, Lee 9. Martinet L, Fleury-Cappellesso S, Gadelorge M, JU, Fujimoto K, Tsuji S, Nakajima S, Doi R, Imamura Dietrich G, Bourin P, Fournie JJ, Poupot R. A reg- 1. Wang MT, Honn KV, Nie D. Cyclooxygenases, pros- M. Immunohistochemical analysis of cyclooxygenase-2 ulatory cross-talk between Vg9Vd2 T lymphocytes tanoids, and tumor progression. Cancer Metastasis Rev expression in pancreatic tumors. Int J Pancreatol 1999; and mesenchymal stem cells. Eur J Immunol 2009; 2007; 26:525-34; PMID:17763971; http://dx.doi.org/ 26:69-76; PMID:10597402; http://dx.doi.org/ 39:752-62; PMID:19197941; http://dx.doi.org/ 10.1007/s10555-007-9096-5 10.1007/BF02781733 10.1002/eji.200838812 2. Harris SG, Padilla J, Koumas L, Ray D, Phipps RP. 6. Sreeramkumar V, Fresno M, Cuesta N. Prostaglandin 10. Kabashima-Niibe A, Higuchi H, Takaishi H, Masugi Prostaglandins as modulators of immunity. Trends E2 and T cells: friends or foes? Immunol Cell Biol Y, Matsuzaki Y, Mabuchi Y, Funakoshi S, Adachi M, Immunol 2002; 23:144-50; PMID:11864843; http:// 2012; 90:579-86; PMID:21946663; http://dx.doi.org/ Hamamoto Y, Kawachi S, et al. Mesenchymal stem dx.doi.org/10.1016/S1471-4906(01)02154-8 10.1038/icb.2011.75 cells regulate epithelial-mesenchymal transition and 3. Merati K, Said SM, Andea A, Sarkar F, Ben-Josef E, 7. Ahmadi M, Emery DC, Morgan DJ. Prevention of tumor progression of pancreatic cancer cells. Cancer Mohammad R, Philip P, Shields AF, Vaitkevicius V, both direct and cross-priming of antitumor CD8C T- Sci 2013; 104:157-64; PMID:23121112; http://dx. Grignon DJ, Adsay NV. Expression of inflammatory cell responses following overproduction of prostaglan- doi.org/10.1111/cas.12059 modulator COX-2 in pancreatic ductal adenocarci- din E2 by tumor cells in vivo. Cancer Res 2008; 11. Jimeno A, Amador ML, Kulesza P, Wang X, Rubio- noma and its relationship to pathologic and clinical 68:7520-9; PMID:18794140; http://dx.doi.org/ Viqueira B, Zhang X, Chan A, Wheelhouse J, Kuramo- parameters. Am J Clin Oncol 2001; 24:447-52; 10.1158/0008-5472.CAN-08-1060 chi H, Tanaka K, et al. Assessment of celecoxib phar- PMID:11586094; http://dx.doi.org/10.1097/ 8. Baratelli F, Lin Y, Zhu L, Yang SC, Heuze-Vourc’h N, macodynamics in pancreatic cancer. Mol Cancer Ther 00000421-200110000-00007 Zeng G, Reckamp K, Dohadwala M, Sharma S, Dubi- 2006; 5:3240-7; PMID:17172427; http://dx.doi.org/ 4. Molina MA, Sitja-Arnau M, Lemoine MG, Frazier nett SM. Prostaglandin E2 induces FOXP3 gene 10.1158/1535-7163.MCT-06-0565 ML, Sinicrope FA. Increased cyclooxygenase-2 expres- expression and T regulatory cell function in human 12. Vosooghi M, Amini M. The discovery and develop- sion in human pancreatic carcinomas and cell lines: CD4C T cells. J Immunol 2005; 175:1483-90; ment of cyclooxygenase-2 inhibitors as potential anti- growth inhibition by nonsteroidal anti-inflammatory PMID:16034085; http://dx.doi.org/10.4049/ cancer therapies. Expert Opin Drug Discov 2014; drugs. Cancer Res 1999; 59:4356-62; jimmunol.175.3.1483 PMID:10485483 e988460-10 OncoImmunology Volume 4 Issue 3 9:255-67; PMID:24483845; http://dx.doi.org/ 24. Wesch D, Beetz S, Oberg HH, Marget M, Krengel K, melanoma-infiltrating T lymphocytes. Cancer Immu- 10.1517/17460441.2014.883377 Kabelitz D. Direct costimulatory effect of TLR3 ligand nol Immunother 2011; 60:111-22; PMID:20960188; 13. Xu XF, Xie CG, Wang XP, Liu J, Yu YC, Hu HL, Guo poly(I:C) on human gd T lymphocytes. J Immunol http://dx.doi.org/10.1007/s00262-010-0924-z CY. Selective inhibition of cyclooxygenase-2 suppresses 2006; 176:1348-54; PMID:16424161; http://dx.doi. 36. Vang T, Abrahamsen H, Myklebust S, Horejsi V, the growth of pancreatic cancer cells in vitro and in org/10.4049/jimmunol.176.3.1348 Tasken K. Combined spatial and enzymatic regulation vivo. Tohoku J Exp Med 2008; 215:149-57; 25. Helm O, Mennrich R, Petrick D, Goebel L, Freitag- of Csk by cAMP and protein kinase a inhibits T cell PMID:18577844; http://dx.doi.org/10.1620/tjem. Wolf S, Roder C, Kalthoff H, R€ocken C, Sipos B, receptor signaling. J Biol Chem 2003; 278:17597-600; 215.149 Kabelitz D, et al. Comparative characterization of PMID:12665526; http://dx.doi.org/10.1074/jbc. 14. Dragovich T, Burris H, III, Loehrer P, Von Hoff DD, stroma cells and ductal epithelium in chronic pancreati- C300077200 Chow S, Stratton S, Green S, Obregon Y, Alvarez I, tis and pancreatic ductal adenocarcinoma. PLoS One 37. Nishigaki N, Negishi M, Ichikawa A. Two Gs-coupled Gordon M. Gemcitabine plus celecoxib in patients 2014; 9:e94357; PMID:24797069; http://dx.doi.org/ prostaglandin E receptor subtypes, EP2 and EP4, differ with advanced or metastatic pancreatic adenocarci- 10.1371/journal.pone.0094357 in desensitization and sensitivity to the metabolic inac- noma: results of a phase II trial. Am J Clin Oncol 26. Martinet L, Jean C, Dietrich G, Fournie JJ, Poupot R. tivation of the agonist. Mol Pharmacol 1996; 50:1031- 2008; 31:157-62; PMID:18391600; http://dx.doi.org/ PGE2 inhibits natural killer and gd T cell cytotoxicity 7; PMID:8863851 10.1097/COC.0b013e31815878c9 triggered by NKR and TCR through a cAMP-mediated 38. Patrignani P, Tacconelli S, Sciulli MG, Capone ML. 15. Lipton A, Campbell-Baird C, Witters L, Harvey H, Ali PKA type I-dependent signaling. Biochem Pharmacol New insights into COX-2 biology and inhibition. S. Phase II trial of gemcitabine, irinotecan, and cele- 2010; 80:838-45; PMID:20470757; http://dx.doi.org/ Brain Res Brain Res Rev 2005; 48:352-9; coxib in patients with advanced pancreatic cancer. J 10.1016/j.bcp.2010.05.002 PMID:15850674; http://dx.doi.org/10.1016/j. Clin Gastroenterol 2010; 44:286-8; PMID:20216081; 27. Chulada PC, Thompson MB, Mahler JF, Doyle CM, brainresrev.2004.12.024 http://dx.doi.org/10.1097/MCG.0b013e3181cda097 Gaul BW, Lee C, Tiano HF, Morham SG, Smithies O, 39. Yamaguchi K, Nakano K, Nagai E, Chijiiwa K, 16. Lunardi S, Muschel RJ, Brunner TB. The stromal com- Langenbach R. Genetic disruption of Ptgs-1, as well as Kinoshita M, Ohta M, Tanaka M. Ki-ras mutations in partments in pancreatic cancer: are there any therapeu- Ptgs-2, reduces intestinal tumorigenesis in Min mice. codon 12 and p53 mutations (biomarkers) and cytol- tic targets? Cancer Lett 2014; 343:147-55; Cancer Res 2000; 60:4705-8; PMID:10987272 ogy in bile in patients with hepatobiliary-pancreatic car- PMID:24141189; http://dx.doi.org/10.1016/j.canlet. 28. Crofford LJ. COX-1 and COX-2 tissue expression: cinoma. Hepatogastroenterology 2005; 52:713-8; 2013.09.039 implications and predictions. J Rheumatol Suppl 1997; PMID:15966189 17. Siegel R, Ma J, Zou Z, Jemal A. Cancer statistics, 2014. 49:15-9; PMID:9249646 40. Kundu N, Ma X, Kochel T, Goloubeva O, Staats P, CA Cancer J Clin 2014; 64:9-29; PMID:24399786; 29. Hill R, Li Y, Tran LM, Dry S, Calvopina JH, Garcia A, Thompson K, Martin S, Reader J, Take Y, Collin P, http://dx.doi.org/10.3322/caac.21208 Kim C, Wang Y, Donahue TR, Herschman HR, et al. et al. Prostaglandin E receptor EP4 is a therapeutic tar- 18. Oberg HH, Peipp M, Kellner C, Sebens S, Krause S, Cell intrinsic role of COX-2 in pancreatic cancer devel- get in breast cancer cells with stem-like properties. Petrick D, Adam-Klages S, R€ocken C, Becker T, Vogel opment. Mol Cancer Ther 2012; 11:2127-37; Breast Cancer Res Treat 2014; 143:19-31; I, et al. Novel bispecific antibodies increase gd T-cell PMID:22784710; http://dx.doi.org/10.1158/1535- PMID:24281828; http://dx.doi.org/10.1007/s10549- cytotoxicity against pancreatic cancer cells. Cancer Res 7163.MCT-12-0342 013-2779-4 2014; 74:1349-60; PMID:24448235; http://dx.doi. 30. Alfranca A, Iniguez MA, Fresno M, Redondo JM. Pros- 41. Ma X, Holt D, Kundu N, Reader J, Goloubeva O, org/10.1158/0008-5472.CAN-13-0675 tanoid signal transduction and gene expression in the Take Y, Fulton AM. A prostaglandin E (PGE) receptor 19. Shojaei H, Oberg HH, Juricke M, Marischen L, Kunz endothelium: role in cardiovascular diseases. Cardio- EP4 antagonist protects natural killer cells from PGE- M, Mundhenke C, Gieseler F, Kabelitz D, Wesch D. vasc Res 2006; 70:446-56; PMID:16458868; http:// mediated immunosuppression and inhibits breast can- Toll-like receptors 3 and 7 agonists enhance tumor cell dx.doi.org/10.1016/j.cardiores.2005.12.020 cer metastasis. Oncoimmunology 2013; 2:e22647; lysis by human gd T cells. Cancer Res 2009; 69:8710- 31. Narumiya S, Sugimoto Y, Ushikubi F. Prostanoid PMID:23482441; http://dx.doi.org/10.4161/onci. 7; PMID:19887600; http://dx.doi.org/10.1158/0008- receptors: structures, properties, and functions. Physiol 22647 5472.CAN-09-1602 Rev 1999; 79:1193-226; PMID:10508233 42. Chen Q, Muramoto K, Masaaki N, Ding Y, Yang H, 20. Sipos B, Moser S, Kalthoff H, Torok V, Lohr M, Klop- 32. Pino MS, Nawrocki ST, Cognetti F, Abruzzese JL, Mackey M, Li W, Inoue Y, Ackermann K, Shirota H, pel G. A comprehensive characterization of pancreatic Xiong HQ, McConkey DJ. Prostaglandin E2 drives et al. A novel antagonist of the prostaglandin E(2) EP ductal carcinoma cell lines: towards the establishment cyclooxygenase-2 expression via cyclic AMP response (4) receptor inhibits Th1 differentiation and Th17 of an in vitro research platform. Virchows Arch 2003; element activation in human pancreatic cancer cells. expansion and is orally active in arthritis models. Br J 442:444-52; PMID:12692724 Cancer Biol Ther 2005; 4:1263-9; PMID:16319525; Pharmacol 2010; 160:292-310; PMID:20423341; 21. Murakami M, Naraba H, Tanioka T, Semmyo N, http://dx.doi.org/10.4161/cbt.4.11.2138 http://dx.doi.org/10.1111/j.1476-5381.2010.00647.x Nakatani Y, Kojima F, Ikeda T, Fueki M, Ueno A, Oh 33. Ponti C, Gibellini D, Boin F, Melloni E, Manzoli FA, 43. Sobolewski C, Cerella C, Dicato M, Ghibelli L, Die- S, Kudo I. Regulation of prostaglandin E2 biosynthesis Cocco L, Zauli G, Vitale M. Role of CREB transcrip- derich M. The role of cyclooxygenase-2 in cell prolifer- by inducible membrane-associated prostaglandin E2 tion factor in c-fos activation in natural killer cells. Eur ation and cell death in human malignancies. Int J Cell synthase that acts in concert with cyclooxygenase-2. J J Immunol 2002; 32:3358-65; PMID:12432566; Biol 2010; 2010:215158; PMID:20339581; http://dx. Biol Chem 2000; 275:32783-92; PMID:10869354; http://dx.doi.org/10.1002/1521-4141(200212) doi.org/10.1155/2010/215158 http://dx.doi.org/10.1074/jbc.M003505200 32:12%3c3358::AID-IMMU3358%3e3.0.CO;2-Q 44. Lipton A, Harvey H, Witters L, Kerr S, Legore K, 22. Gupta NK, Dixit VK. Bioavailability enhancement of 34. Samten B, Howard ST, Weis SE, Wu S, Shams H, Campbell C. Gemcitabine/Irinotecan/celecoxib in pan- curcumin by complexation with phosphatidyl choline. Townsend JC, Safi H, Barnes PF. Cyclic AMP response creatic cancer. Oncology (Williston Park) 2004; 18:43- J Pharm Sci 2011; 100:1987-95; PMID:21374628; element-binding protein positively regulates produc- 5; PMID:15685834 http://dx.doi.org/10.1002/jps.22393 tion of IFN-g by T cells in response to a microbial 45. Xu XF, Xie CG, Wang XP, Liu J, Yu YC, Hu HL, Guo 23. Pietschmann K, Beetz S, Welte S, Martens I, Gruen J, pathogen. J Immunol 2005; 174:6357-63; CY. Selective inhibition of cyclooxygenase-2 suppresses Oberg HH, Wesch D, Kabelitz D. Toll-like receptor PMID:15879136; http://dx.doi.org/10.4049/ the growth of pancreatic cancer cells in vitro and in expression and function in subsets of human gd T lym- jimmunol.174.10.6357 vivo. Tohoku J Exp Med 2008; 215:149-57; phocytes. Scand J Immunol 2009; 70:245-55; 35. Su Y, Jackson EK, Gorelik E. Receptor desensitization PMID:18577844; http://dx.doi.org/10.1620/tjem. PMID:19703014; http://dx.doi.org/10.1111/j.1365- and blockade of the suppressive effects of prostaglandin 215.149 3083.2009.02290.x E(2) and adenosine on the cytotoxic activity of human www.tandfonline.com OncoImmunology e988460-11 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png OncoImmunology Taylor & Francis

Resistance of cyclooxygenase-2 expressing pancreatic ductal adenocarcinoma cells against γδ T cell cytotoxicity

Download PDF
11 pages

Loading next page...
 
/lp/taylor-francis/resistance-of-cyclooxygenase-2-expressing-pancreatic-ductal-oPwfvWMvR1

References (46)

Publisher
Taylor & Francis
Copyright
© 2015 Taylor & Francis Group, LLC
ISSN
2162-402X
DOI
10.4161/2162402X.2014.988460
Publisher site
See Article on Publisher Site

Abstract

ORIGINAL RESEARCH OncoImmunology 4:3, e988460; March 2015; © 2015 Taylor & Francis Group, LLC Resistance of cyclooxygenase-2 expressing pancreatic ductal adenocarcinoma cells against gd T cell cytotoxicity 1 1 2 2 3 1 Daniel Gonnermann , Hans-Heinrich Oberg , Christian Kellner , Matthias Peipp , Susanne Sebens , Dieter Kabelitz , 1, and Daniela Wesch 1 2 Institute of Immunology; Christian-Albrechts-University; Kiel, Germany; Division of Stem Cell Transplantation and Immunotherapy; Christian-Albrechts-University; Kiel, Germany; Institute for Experimental Medicine; Christian-Albrechts-University; Kiel Keywords: Cox-2, cytotoxicity, gammadelta T lymphocytes, human, pancreatic ductal adenocarcinoma, PGE2 Abbreviations: BrHPP, bromohydrin-pyrophosphate; Cox, cyclooxygenase; n-BP, nitrogen-containing bisphosphonates; PAg, phosphorylated antigen; PDAC, pancreatic ductal adenocarcinoma; PG, prostaglandins; RTCA, Real Time Cell Analyzer; TCR, T cell receptor. The prostaglandin (PG) synthetase cyclooxygenase 2 (Cox-2) promotes tumorigenesis, tumor progression, and metastasis in a variety of human cancer entities including pancreatic ductal adenocarcinoma (PDAC). In this study, we demonstrate that in PDAC cells such as Colo357 cells, enhanced Cox-2 expression and increased release of the Cox-2 metabolite prostaglandin E2 (PGE2) promotes resistance against gd T cell-mediated lysis. Co-culture with activated gd T cells induced an upregulation of Cox-2 expression in Colo357 cells, and thereby an enhanced PGE2 release, in response to tumor necrosis factor a .TNFa/ secretion from gd T cells. The PGE2-mediated inhibition of gd T cell cytotoxicity against Cox-2-expressing PDAC cells can be partially overcome by Cox-2 inhibitors. Our results show that differences between PDAC cells in regards to sensitivity to gd T-cell cytotoxicity can be due to distinct levels of Cox-2 expression associated with varying amounts of PGE2 release. While gd T cell cytotoxicity against PDAC cells expressing low levels of Cox-2 can be effectively enhanced by tribody [(Her2) £Vg9] with specificity for Vg9 T cell receptor and HER-2/neu on PDAC cells, a combination of tribody [(Her2) £Vg9] and Cox-2 inhibitor is necessary to induce complete lysis of Cox-2 high expressing Colo357. In conclusion, our results suggest that the application of tribody [(Her2) £Vg9] that enhances gd T-cell cytotoxicity and Cox-2 inhibitors that overcome PGE2-mediated resistance of PDAC cells to the cytotoxic activity of gd T cells might offer a promising combined immunotherapy for pancreatic cancer. Introduction cells, the function of the numerically small population of gd T lymphocytes is also inhibited by PGE2 secreting and Cox-2 Arachidonic acid is released from plasma membrane phospho- expressing mesenchymal stem cells involved in the development 9,10 lipids and is further converted by cyclooxygenase (Cox) enzymes of tumor-promoting cancer stroma. PGE2 binds to prosta- Cox-1 and Cox-2 to prostaglandins (PG), prostacyclins and glandin E2 and E4 receptors expressed on activated gd T cells, thromboxanes. Cox-1 is constitutively expressed in most mam- and thereby inhibits T cell receptor (TCR)-activated cytotoxic malian tissues, whereas prostaglandin endoperoxidase synthase activity of gd T cells by cyclic adenosine monophosphate enzyme 2 (PTGS2, better known as Cox-2) is induced by a vari- (cAMP)-mediated protein kinase A type I-dependent signaling. 1,2 ety of pro-inflammatory stimuli. Cox-2 is overexpressed in However, such an enhanced PGE2 synthesis could be potently diverse tumor entities including ductal pancreatic adenocarcino- decreased by the Cox-2 inhibitor celecoxib in a pancreatic cancer mas (PDAC), which is mostly associated with a more aggressive xenograft mouse model. Although selective Cox-2 inhibitors 3-5 tumor stage and worse prognosis. Cox-2 overexpression leads such as celecoxib and DuP697 showed anti-proliferative activity to an enhanced production of prostaglandin E2 (PGE2), which toward different tumor entities and against several PDAC cell is involved in tumor progression as well as in tumor evasion of lines in vitro as well as in vivo, celecoxib as a single agent failed to immunosurveillance. The release of PGE2 by cancer cells indu- exert anti-proliferative or antitumor effects toward pancreatic ces and recruits regulatory T cells to the tumor site and suppresses adenocarcinomas resected from patients and implanted in nude immune responses of CD8 ab T cells and natural killer (NK) mice in a xenograft model, despite a decrease in PGE2 synthe- 2,7,8 11-13 cells as well as maturation of dendritic cells. Similar to ab T sis. Nevertheless, selective Cox-2 inhibitors have been *Correspondence to: Daniela Wesch; Email: [email protected] Submitted: 10/16/2014; Revised: 11/08/2014; Accepted: 11/12/2014 http://dx.doi.org/10.4161/2162402X.2014.988460 www.tandfonline.com OncoImmunology e988460-1 included in clinical trials (Phase II) as combinatorial therapies, monolayers can be dynamically and quantitatively monitored on such as to increase the efficiency of the topoisomerase I inhibitor the sensor plate in real time for up to several hours or days with- irinotecan or chemotherapy for the treatment of patients with out the incorporation of labels. After 20 h of adherence, PDAC 12,14,15 advanced or metastatic pancreatic adenocarcinoma. cells were co-cultured with newly established gd T cell lines from PDAC is a highly malignant gastrointestinal tumor character- healthy donors and PDAC patients together with the previously ized by thepresenceofadense desmoplasticstromacomposedof titrated optimal concentrations of the indicated stimuli. By using extracellular matrix and diverse non-neoplastic inflammatory cells the RTCA assay, we were able to distinguish between 3 different mostly displaying an immunosuppressive phenotype. The overall sensitivities of PDAC cells against gd T cell-mediated cytotoxic- 5-year survival rate is less than 5%. The poor prognosis of PDAC ity as follows: (i) lysis of Panc89 cells as well as of Pt45P1 cells by is due to the absence of specific symptoms and a rapid progression gd T cell lines was drastically enhanced in the presence of PAg of this extremely aggressive disease, essentially limiting therapeutic BrHPP or tribody [(Her2) £Vg9)] (Fig. 1A, data not shown) options. Moreover, PDAC is characterized by a profound resis- and the cell index decreased nearly to the level of maximal lysis 16,17 tancetowardcurrent chemo- andtargetedtherapies. We induced by the positive control with 1% Triton X-100, (ii) the recently reported that the adoptive transfer of gd T cells in combi- gd T cell-mediated lysis of PancTu-I cells was slightly enhanced nation with application of a novel tribody [(Her2) £Vg9] by BrHPP, but potentiated much more strongly by reducedgrowthofPDACcells graftedintoseverecombined [(Her2) £Vg9)] (Fig. 1B), (iii) the lysis of Colo357 cells by gd immunodeficiency disease (SCID)-Beige immunodeficient mice. T cells was not increased by BrHPP and only slightly by We designed a unique bispecific antibody [(Her2) £Vg9] in the [(Her2) £Vg9)] (Fig. 1C). Also, [(Her2) £Vg9)] as well as 2 2 2 so-called tribody format that has one binding site for the Vg9T- BrHPP had no effect on PDAC cells in the absence of gd T cell cell receptor (TCR) element on gd T cells and 2 binding sites for lines (data not shown). Additionally, a similarly designed control HER-2/neu (ERBB2) expressed by PDAC cells. Our intention construct [(Her2) £CD89] that does not bind to gd T cells had was to selectively target gd T cells to HER-2/neu-expressing no effect, suggesting that the tribody [(Her2) £Vg9)] triggered tumor cells and to enhance their cytotoxicity more efficiently than target lysis by T cell receptor (TCR) engagement ( and data selective activators of gd T cells, such as phosphorylated antigens not shown). In the course of these studies, we analyzed the cyto- (PAg). Despite promising results in these studies, we identified toxicity against these 4 different PDAC cells with various gd T PDAC cells that were nearly resistant to gd T-cell cytotoxicity. cell lines established from different healthy donors (n D 10) and In this study, we examined whether an intrinsic tumor resis- PDAC patients (n D 15). To demonstrate that the different sen- tance can explain the inability of [(Her2) £Vg9] activated-gd T sitivity of the PDAC cells to gd T cell cytotoxicity was not due to cell lines to completely lyse several PDAC cells lines. We investi- differences between individual gd T cell lines established from gated if overexpression of Cox-2 and enhanced PGE2 release by different donors, gd T cell lines from the same 4 representative PDAC cells was responsible for the resistance of several PDAC donors (2 healthy and 2 PDAC patient) were used for the lysis of cells against gd T-cell cytotoxicity. the various PDAC cells as indicated. Differential Cox-2 expression and PGE2 release by PDAC cells Results Next, we examined whether PDAC cells differ in their Cox-2 expression, thereby releasing different amounts of PGE2 which Different sensitivity of PDAC cells against gd T could be one explanation for the different sensitivity toward gd cell-mediated lysis T cell lysis. The expression of Cox-2, the inducible enzyme in the PGE2 producing pathway, was analyzed in all PDAC cells by Our recent studies revealed that PDAC cells differ in their sensitivity against gd T cell-mediated cytotoxicity. Weak gd T flow cytometry as well as by Western Blot. Regarding the mean cell-mediated lysis of PDAC cells could be enhanced in several fluorescence intensity determined by cytometric fluorimetric but not all PDAC cells by the addition of a novel bispecific anti- analysis, we observed a very low expression of intracellular Cox-2 body [Her2£CD3] or tribody [(Her2) £Vg9]. The aim of in PancTu-I as well as in Panc89 cells compared to Colo357 cells, our current study was to analyze the underlying mechanisms of which highly expressed intracellular Cox-2 (Fig. 2A and B). the differential sensitivity of PDAC cells against gd T cell-medi- These results were verified with an additional Cox-2 antibody by ated lysis. We analyzed 4 PDAC cell lines of distinct origins and Western Blot analysis confirming the highest Cox-2 expression grades of differentiation, which we previously used in our recent level in Colo357 cells (Fig. 2C). The parallel analysis of Cox-1, 18,19 publications. PancTu-I and Pt45P1 cells were initially estab- the constitutively active enzyme of the PGE2 producing pathway, lished from primary tumors (stage G2-G3), whereas Panc89 and revealed a similar expression in all PDAC cell lines (Fig. 2D). Colo357 cells were generated out of lymph node metastases Next, we investigated whether the differences in intracellular (stage G1-G2). For our analysis, we used the previously Cox-2 expression were reflected by differential release of PGE2 described Real Time Cell Analyzer (RTCA) system which uses from PDAC cells. In line with the low intracellular Cox-2 expres- impedance technology. An electronic sensor measures the sion, PancTu-I and Panc89 cells secreted very low amounts of impedance, which reflects changes in cellular parameters of PGE2 (< 0.5 ng/mL), whereas Cox-2 high expressing Colo357 adherent PDAC cells, but not of suspended cells such as gd T cells released relatively high amounts of PGE2 (> 6 ng/mL) cells. Therefore, gd T cell-mediated lysis of PDAC cell (Fig. 2E). e988460-2 OncoImmunology Volume 4 Issue 3 3 Figure 1. Differential susceptibility of PDAC cells against gd T cell-mediated lysis. Five £ 10 (A) Panc89, (B) PancTu-I and (C) Colo357 pancreatic ductal adenocarcinoma (PDAC) cells were cultured in complete medium for 24 h on an E-plate before they were treated with medium (green line) or with gd T cell lines together with medium (dark blue line), 300 nM phosphorylated antigen (PAg) BrHPP (orange line), 1 mg/mL [(Her2) £Vg9)] (light blue line) of 2 representative healthy donor (HD1 and 2) out of 10 and 2 representative PDAC patients (PC1 and 2) out of 15 at an effector to target (E:T) cell ratio of 12.5:1 in a Real Time Cell Analyzer (RTCA) electrical impedance assay over the indicated time. As a positive control for maximal lysis, PDAC cells were treated with Triton X-100. The cell index was determined every 5 min before addition of different substances and gd T cell lines and, thereafter, every minute for > 6 h. The addition of substances or gd T cells is indicated by an arrow. The cell index was normalized to 1 shortly before the time of addition of substances or gd T cell lines as presented by the vertical black line. The average of triplicate determination with the standard deviation (SD) of one rep- resentative experiment is shown. Taken together, the 3 different PDAC cell lines differed sig- different PDAC cell lines directly after seeding of the PDAC cells nificantly from each other regarding Cox-2 expression and PGE2 for 20 h (in comparison to DMSO as a control) and before release. BrHPP-stimulated gd T cell lines were added and cell lysis was determined by RTCA assay (data not shown). Additionally, we tested the effect of PGE2 on PDAC cells alone showing that PGE2 inhibited gd T cell-mediated lysis of PDAC cells PGE2 has no effect on the impedance, and thus on the cell death abolished by Cox-2 inhibitors of PDAC cells, in the absence of gd T cells (data not shown). The enhanced release of PGE2 by Colo357 cells being in line However, the presence of 0.1–1 mg/mL PGE2 strongly reduced with the reduced sensitivity toward gd T cell lysis (Fig. 1) the lysis of Panc89 as well as of PancTu-I cells by BrHPP-stimu- prompted us to examine the effect of PGE2 on the gd T cell- lated gd T cell lines compared to the stimulation in the absence mediated lysis of PDAC cells. Different concentrations (0.1– of PGE2 and in the presence of medium or DMSO control 10 mg/mL) of PGE2 (dissolved in DMSO) were added to the 3 www.tandfonline.com OncoImmunology e988460-3 we observed no difference between Cox-1/2 and Cox- 2 inhibitors on the enhanced lysis suggesting that only Cox-2 expression influences the cytotoxic activity of gd T cells toward PDAC cells. To analyze whether the Cox-inhibitors effect PGE2 production by Colo357, the tumor cells were cultured alone or co-cultured with BrHPP-stimulated gd T cell lines in the presence of the indicated Cox-inhibitor and PGE2 release was deter- mined in the supernatants. Interestingly, PGE2 release by Colo357 alone was not Figure 2. Increased Cox-2 expression and PGE2 release of Colo357. (A, B, and D) Cytofluorimetric analysis to deter- reduced after the addition mine the intracellular expression levels of (A/B) Cox-2 or (C) Cox-1 in the indicated pancreatic ductal adenocarci- of the indicated concentra- noma (PDAC) cells. Protein levels were detected in (A) one representative experiment (bold line; the thin line tions of Cox-inhibitors. indicates the isotype control) and (B/D) in 5 additional experiments (results display the mean § SEM) with an anti- However, the co-culture Cox-1-FITC and anti-Cox-2-PE mAb mixture (AS66/AS67, BD Biosciences) by flow cytometry. (C) Whole protein (20 mg) from the indicated PDAC cells were used for Western Blot, and Cox-2 protein was analyzed with anti-Cox-2 with BrHPP-stimulated gd mAb (33/Cox-2, BD Biosciences). b-actin was used as a loading control. (E) PDAC cells at 1.5 £ 10 cells/well were T cell lines enhanced the incubated in 24-well plates for 24 h. PGE2 release was determined in the supernatant after overnight culture. PGE2 release dramatically Bars § SD present the mean value of 4 independent experiments. which could be significantly reduced in the presence of (Fig. 3A and B and data not shown). In contrast to the lysis of Cox-2 inhibitors (Fig. 4). In line with the enhanced PGE2 PancTu-I and Panc89, the weak lysis of Colo357 cells was virtu- release, Cox-2 expression in Colo357 cells was drastically ally unaltered by additional treatment with PGE2 (Fig. 3C). enhanced after co-culturing them with BrHPP-stimulated gd T These results fit well with the indicated high release of PGE2 by cell lines (data not shown), which underlines the results by others Colo357 (Fig. 2E) and demonstrate that an additional applica- that Cox-2 expression correlates with PGE2 production. tion of PGE2 to the already high level PGE2-producing These data demonstrate that an enhanced PGE2 release Colo357 only slightly affected the nearly complete resistance induced by an intensified expression of Cox-2 in Colo357 cells against gd T cell-mediated lysis (Fig. 3C). co-cultured with gd T cell lines decreased gd T-cell cytotoxicity Since Cox-2 expression goes along with the release of PGE2, that could be abolished by Cox-2 inhibitors. we tested whether Cox inhibitors were able to increase gd T cell- mediated lysis of PDAC cells. First, we studied the effect of dif- TNFa released by gd T cells enhanced Cox-2 expression in ferent concentrations (5, 10, 50 and 200 mM) of the Cox-1/2 Colo357 PDAC cells inhibitor Indomethacin and the selective Cox-2 inhibitor In light of our observations that Colo357 PDAC cells co-cul- DuP697 on PDAC cells alone to exclude toxic effects on the tured with gd T cells displayed enhanced production of PGE2 tumor cells potentially induced by these inhibitors. None of the and intracellular Cox-2 expression taken together with reports by tested concentrations of the inhibitors (dissolved in DMSO) others that Cox-2 expression can be stimulated by various anti- 1,6,22 revealed an influence on PDAC cell growth except 200 mM gens and cytokines, we next set out to investigate the role of DuP697, which caused a growth inhibition on all PDAC cells, cytokines on the regulation of Cox-2 expression in Colo357 cells. and 50 mM DuP697, which reduced Panc89 cell growth (data We focused on tumor necrosis factor a .TNFa/ as well as inter- not shown). feron g .IFNg/ because gd T cell lines produce both cytokines 23,24 Accordingly, pretreatment of PDAC cells with 50 mM Indo- after activation with PAg, such as BrHPP. We found that methacin or of PancTu-I with 50 mM and Panc89 with 10 mM TNFa, but not IFNg, induced a dose-dependent and significant DuP697 revealed no effect on the lysis of these cells by BrHPP- upregulation of intracellular Cox-2 expression in Colo357, but stimulated gd T cell lines (data not shown), whereas the lysis of not in PancTu-I or Panc89 cells (Fig. 5A, data not shown). Colo357 was strongly enhanced by gd T cell lines in the presence When IFNg was added to Colo357 cells, Cox-2 expression of Indomethacin or 50 mM DuP697 (Fig. 3D and E). Moreover, decreased significantly. However, concomitant stimulation of e988460-4 OncoImmunology Volume 4 Issue 3 3 Figure 3. Effect of PGE2 and Cox-2 inhibitors on gd T cell cytotoxicity. Five £ 10 (A) Panc89, (B) PancTu-I and (C-E) Colo357 pancreatic ductal adenocar- cinoma (PDAC) cells were seeded for approximately 24 to 44 h on an E-plate of a Real Time Cell Analyzer (RTCA) electrical impedance assay. After reach- ing a cell index of one, tumor cells were treated with medium (green line) or gd T cell lines at an effector to target (E:T) cell ratio of 25:1 added together with 300 nM phosphorylated antigen (PAg) BrHPP in the presence of medium (dark blue line), DMSO (as solvent control, light blue line) or (A–C)1 mg/ mL prostaglandin E2 (PGE2; pink line) as well as (D-E) 50 mM Indomethacin (violet line) and 50 mM DuP697 (red line). PDAC cells treated with Triton X- 100, which was used as control for maximum cell lysis, are illustrated as a black line. The cell index was analyzed every 5 min during the whole time course and in one min steps for 6 h after addition of the different stimuli, which assigned the time point of cell index normalization (black vertical line). The average of 3 replicates with SD is presented in (A–C) of one representative out of 6 independent experiments or in (D and E) of 2 representative out of 8 experiments. gd T cell lines of the same healthy donor were used in (A, C and E) and from an additional healthy donor in (B and D). Similar reactivity patterns were obtained in 6 to 8 independent experiments with other gd T cell lines of healthy donors as well as of PDAC patients. Colo357 cells with TNFa applied together with IFNg led to a Cox-2 inhibitor DuP697 together with [(Her2) £Vg9] similar increase of Cox-2 expression as TNFa stimulation alone overcome the resistance toward gd T cell-mediated lysis of (Fig. 5A). In accordance with these results, application of the Colo357 TNFa-blocker Infliximab during co-culture of Colo357 cells To investigate whether the addition of the Cox-2 inhibitor and TNFa producing gd T cell lines clearly prevented the DuP697 co-administered together with the tribody increase of Cox-2 expression, an effect that was not observed after [(Her2) £Vg9] could overcome the resistance of Colo357 cells treatment with control Abs (Fig. 5B). toward gd T cell-cytotoxicity, we activated several gd T cell lines The inhibition by Infliximab demonstrates that TNFa from different healthy donors with BrHPP in the absence or pres- released by activated gd T cell lines accounts for the strong ence of DuP697, [(Her2) £Vg9], or with the combination of induction of Cox-2 expression in Colo357 cells. both. As expected, gd T cell lines only weakly lysed the tumor www.tandfonline.com OncoImmunology e988460-5 Figure 5. Enhanced Cox-2 expression in Colo357 induced by TNFa is inhibited by TNFa blocker Infliximab. Cytofluorimetric analysis of the effects of cytokine levels on cyclooxygenase 2 (Cox-2) expression. (A) Figure 4. Upregulation of PGE2 release by Colo357 is inhibited by Cox- Colo357 at 1.5 £ 10 cells/well were seeded in medium or together with inhibitors. The effects of cyclooxygenase (Cox) inhibitors on prostaglan- the indicated concentrations of tumor necrosis factor a .TNFa/, inter- din E2 (PGE2) release by pancreatic cancer cells was tested by culturing feron g (IFNg) or the combination of both in 24-well plates. After 24 h of 1.5 £ 10 Colo357 cells with either medium, 50 mM Indomethacin or culture, mean fluorescence intensity (MFI) § standard error of mean 50 mM DuP697 for 20 h. After pre-treatment, 300 nM phosphorylated (SEM) of Cox-2 expression was determined by intracellular staining of antigen (PAg) BrHPP with or without gd T cell lines at an effector to tar- Colo357 with fluorochrome-conjugated anti-Cox-2 monoclonal antibody get (E:T) cell ratio of 5:1 were added. PGE2 release in the supernatant (mAb) and measured by flow cytometry. After subtraction of isotype con- was determined after 24 h by ELISA. Representative results of experi- trols, statistical analysis of differences in MFI were determined by t-test ments with 4 donors § SD are shown. Statistical analysis was performed and P values were calculated in relation to the medium control in 3 inde- pendent experiments. Levels of significance are presented as *P < 0.05; by Student’s t-test and significance presented as *P < 0.05. **P < 0.01. (B) Colo357 were cultured overnight before the addition of 10 mg/mL Infliximab or 10 mg/mL IgG1 as a control followed by cells after activation with BrHPP. The additional treatment with medium-cultured or phosphorylated antigen (PAg; 300 nM BrHPP) cul- DuP697 or [(Her2) £Vg9] strongly enhanced the cytotoxic 2 tured gd T cell lines from 4 different donors at an effector to target (E:T) cell ratio of 5:1. MFI § SEM of Cox-2 expression of 6 independent experi- activity of gd T cells toward Colo357 cells (Fig. 6). Similar ments are presented. Significances are shown as *P < 0.05. results were obtained with gd T cell lines from PDAC patients (data not shown). In the absence of BrHPP, we observed no enhancing effect of DuP697, whereas [(Her2) £Vg9] with or without BrHPP similarly increased the cytotoxic effects gd T Discussion cells toward Colo357 cells, as we previously showed. Interest- ingly, the combination of DuP697 and [Her2) £Vg9] most 2 Our study indicates that the inhibition of the PGE2 pathway prominently enhanced the gd T cell-mediated lysis of the natu- with Cox-2 inhibitor DuP697 together with [(Her2) £Vg9], an rally resistant Colo357 cells. Similar results were obtained by enhancer of gd T cell cytotoxicity, abolished the resistance of the using gd T cell lines derived from PDAC patients. We conclude PDAC cell line Colo357 against gd T cell-mediated lysis. that the killing of Cox-2 high PDAC cells by gd T cell lines is gd T lymphocytes have raised substantial interest for immu- more efficient in the presence of DuP697 together with notherapy based on their capacity to kill (radio- and chemother- [(Her2) £Vg9] than with [(Her2) £Vg9] alone. 2 2 apy resistant) PDAC cells in an HLA-independent manner. We e988460-6 OncoImmunology Volume 4 Issue 3 Figure 6. [(Her2) £Vg9)] together with Cox-2 inhibitors overcome the resistance of Colo357 toward gd T cell-mediated lysis. After culturing Colo357 overnight, cells were left untreated (green line) or were co-cultured with phosphorylated antigen (PAg; 300 nM BrHPP) stimulated gd T cell lines at an effector to target (E:T) cell ratio of 25:1 in the presence of 50 IU/mL IL-2 with medium (dark blue line), 1 mg/mL [(Her2) £Vg9)] (light blue line), 50 mM DuP697 (red line) or the combination of [(Her2) £Vg9)] and DuP697 (pink line). The cell index (as measured by electrical impedance) was analyzed in 5 min steps over 24 h and was normalized at the time of addition of substances and gd T cell lines. Thereafter, cell index was measured in 1 min steps for 6 h. Five different individual experiments with Colo357 are shown. The arrow indicates addition of substances and/or gd T cells. previously reported that gd T cells in PDAC tissues are predomi- [(Her2) £Vg9] selectively enhanced the cytotoxicity of gd T nantly extensively distributed in the ductal epithelium and the cells in vitro as well as in vivo upon transfer into immunocompro- stroma close to the ductal epithelium, which demonstrates mobi- mised mice. Although our prior study revealed promising lization and infiltration to the tumor site. However, PDAC is results, heterogeneity between the various PDAC cell lines was characterized by the presence of dense desmoplastic stroma com- observed. Several of the PDAC cell lines, such as Colo357 cells, posed of extracellular matrix and diverse (immunosuppressive) were neither completely lysed by gd T cells alone nor in combi- 18,25 cells dampening cytotoxic activity of gd T cells. An efficient nation with enhancers of their cytotoxicity suggesting that the strategy to overcome immunosuppression by the stromal compo- malignant cells themselves actively promote resistance. Therefore, sition on gd T-cell activity could be the usage of enhancers of we analyzed the direct cross-talk between PDAC cells and gd T gd T-cell cytotoxicity. Recently, we demonstrated that cells in the absence of other cell types in vitro. We observed that www.tandfonline.com OncoImmunology e988460-7 the intracellular Cox-2 expression directly correlated with an of the TCR signaling kinase LCK, the Z chain TCR associated 35,36 increased PGE2 release by Colo357 cells and their resistance kinase ZAP-70 and AKT in T cells. Although EP2 and EP4 against gd T cell-mediated lysis. Similar to the report of Marti- receptors share the same signaling pathway, a distinguishing fea- net and colleagues, who demonstrated an inhibitory effect of ture of the EP4 receptor is its activation of the PI3K signaling PGE2 on gd T-cell cytotoxicity, we observed that the addition pathway resulting in subsequent nuclear factor kB (NF-kB) acti- of PGE2 to PDAC cells releasing scarce native PGE2, such as vation and thus TNFa release. Martinet and co-workers Panc89 and PancTu-I cells, robustly inhibited gd T cell-medi- reported that PGE2 inhibits TCR-activated gd T cell-cytotoxic- ated lysis. In contrast, the addition of exogenous PGE2 to ity by a cAMP-mediated protein kinase A type I-dependent sig- PDAC cells that already released high concentrations of native naling. In addition, EP2- and EP4-specific agonists reduced PGE2 only slightly modulated gd T-cell cytotoxicity, suggesting intracellular IFNg production in activated gd T cells compara- that the endogenous release of PGE2 by Colo357 cells is suffi- ble to the addition of exogenous PGE2. An enhanced release of cient to potently inhibit gd T-cell cytotoxicity. This hypothesis IFNg often leads to reduced intracellular stores of IFNg.To was confirmed by the experiment with the inhibitors Indometh- this end, Martinet and colleagues analyzed whether IFNg modu- acin and DuP697 that partially abrogated the resistance of lates the Cox-2 expression or PGE2 secretion of mesenchymal Colo357 against gd T-cell mediated lysis. Furthermore, since stem cells, which are involved in the tumor stroma development the Cox-1/2 inhibitor Indomethacin and the selective Cox-2 and inhibit gd T cell-mediated cytotoxicity by an enhanced inhibitor DuP697 had similar effects on gd T cell-mediated PGE2 release. In their experiments, they observed that 100 ng/ cytotoxicity against PDAC cells we can rule out that Cox-1 is mL IFNg as well as 100 ng/mL TNFa only slightly enhanced involved in this context, although Cox-1 reportedly also contrib- intracellular Cox-2 expression in mesenchymal stem cells, utes to tumorgenesis. Cox-1 is constitutively expressed in whereas the combined addition of both cytokines significantly many tissues, whereas inducible Cox-2 has been reported to be increased Cox-2 expression suggesting that both cytokines overexpressed in »70% of human pancreatic cancer fulfilling an together produced by activated gd T cells are necessary to induce intrinsic role in tumor initiation, development and progression Cox-2 expression in mesenchymal stem cells. Additionally, we through the activation of the phosphatidylinositol 3 kinase observed that IFNg alone rather diminished intracellular Cox-2 1,28,29 (PI3K)/AKT pathway. While anti-Cox-2 therapy has been expression in Colo357 tumor cells, whereas TNFa significantly consistently shown to inhibit PGE2 synthesis in xenografted increased Cox-2 expression. Further, the combination of both pancreatic tumors and in pancreatic cancer patients, conflicting cytokines enhanced Cox-2 expression in Colo357 cells to levels findings have been reported in regards to the antitumor, anti- comparable to that of TNFa stimulation alone. TNFa pro- proliferative and anti-angiogenic effect of Cox-2 inhibitors in duced by macrophages has been reported to enhance Cox-2 xenografted pancreatic carcinomas, a result that may be due to expression in tissues normally Cox-2 negative. In our study, the usage of PDAC cell lines versus primary PDAC cells in these the importance of TNFa in enhancing Cox-2 expression in 11,13 two contradictory reports. In our experiments, we did not Colo357 cells was supported by the effect of the anti-TNFa focus on the direct toxic effect of Cox-2 inhibitors on PDAC blockers Infliximab and Adalimumab (Fig. 5B and data not cells but on the effect on gd T cell-mediated cytotoxicity. Based shown) that abrogated the enhanced intracellular Cox-2 expres- on previous titration results, we used Cox-2 inhibitors at con- sion in Colo357 cells mediated by exogenous TNFa. In line centrations that were not toxic for PDAC cells but potently with the concept that Infliximab specifically reduced the TNFa- inhibited their PGE2 release. This allowed us to examine the mediated Cox-2 up-regulation but not the constitutive expres- effect of PGE2 on gd T cell-mediated lysis of Colo357 in more sion of Cox-2, the addition of Infliximab in the cytotoxic assay detail. We observed that the co-culture of activated gd T cell only marginally affected the gd T cell-cytotoxicity against lines with Colo357 cells enhanced the PGE2 release, which Colo357 cells (data not shown). Besides the potential upregula- could be significantly inhibited by the Cox-inhibitors Indometh- tion of Cox-2 expression in pancreatic tumors by cytokines, acin and DuP697. As shown by others, PGE2 binds to specific other factors such as genetic alterations during the pancreatic G-protein-coupled receptors termed prostaglandin E receptor 2 carcinogenesis or mutations in p53 and KRAS could play a role. (PTGER2, better known as EP2) and prostaglandin E receptor The oncogene KRAS (also known as Ki-ras), for example, has 4 (PTGER4, better known as EP4), both of which are expressed been previously described to stimulate Cox-2 expression. on activated gd T cells. Activation of EP2 as well as the EP4 EP receptor antagonists have been tested for their ability to 40,41 receptor induces adenylate cyclase and thereby the concomitant suppress breast cancer metastasis. Although the targeted 6,30,31 increase in the secondary messenger cAMP. Subsequently, modulation of T-cell function by blocking specific EP receptor cAMP mediates the dissociation of the regulatory and catalytic signaling seems to be a promising approach, there are several lim- subunits of protein kinase A, which subsequently initiates the itations in the usage of EP receptor antagonist for therapy. EP corresponding transactivation of the transcription factor cAMP receptor antagonists display many compensatory and opposing responsive element binding (CREB). CREB proteins are essen- roles, such as their ability to ameliorate not only Th1 responses tial regulators for T-cell function and cytokine production (e.g., but also Th17 responses. Moreover, these agents typically 33,34 of IFNg). Moreover, PGE2-induced cAMP was described inhibit only one or 2 specific receptors, which might be not as to also activate the raft-associated enzyme c-src tyrosine kinase efficient as Cox-inhibitors, which inhibit all downstream (CSK), which negatively regulates the phosphorylating activities prostaglandins. e988460-8 OncoImmunology Volume 4 Issue 3 Several Cox-2 inhibitors such as Celecoxib (Celebrex ) have Prof. Dr. Kalthoff, Section of Molecular Oncology, Kiel. The been included in several clinical protocols, e.g., in a clinical Phase genotype of PDAC-cell lines was recently confirmed by short tan- II study combined with chemotherapy (paclitaxel, carboplatin) dem repeats analysis. 0.05% trypsin/ 0.02% EDTA was used to and radiotherapy for patients with inoperable stage IIIA/B non- remove adherent PDAC cells from flasks. small-cell lung cancer or with chemotherapy (gemcitabine) com- bined with topoisomerase I inhibitor irinotecan or not for Flow cytometry patients with advanced or metastatic pancreatic adenocarci- For the analysis of the purity of the gd T cell lines, cells were 14,15,43,44 noma. However, the application of Celecoxib over a stained with the following phycoerythrin (PE), allophycocyanin prolonged period of time seems to be critical since it may cause (APC), fluorescein isothiocyanate (FITC) or Per-CP fluoro- cardiovascular side effects. Moreover, the effect of Celecoxib chrome-conjugated monoclonal antibodies (mAb): anti-CD3 12,45 alone induced different effects on pancreatic cancer cells. In (clone SK7; PE: #345765 or APC: # 344812, BD Biosciences), our experiments, we used different Cox inhibitors, which did not anti-TCRgd .clone 11F2; FITC: #347903 or APC: custom induce cell death or cell cycle arrest in PDAC cells in the used research conjugate, BD Biosciences), anti-TCRab .clone IP26, concentrations but inhibited the PGE2 release. Our data demon- #306706, Biolegend), anti-TCRVd2 (clone Immu389, FITC: strate that a combined immunotherapy with Cox-2 inhibitors #PNIM1464, Beckman Coulter or clone B6, PerCP: #331410), together with the newly designed tribody [(Her2) £Vg9] gives anti-TCRVd1 (clone TS8.2, #TCR2730, Thermo Fisher Scien- us a tool to efficiently induce the gd T cell-mediated lysis of tific) or corresponding isotype controls (BD Biosciences or Colo357 PDAC cells. Biolegend). 5 6 Taken together, our results indicate that the usage of the tri- For intracellular staining, 10 -10 tumor cells were washed, body [(Her2) £Vg9], which re-directs gd T cells to target cancer permeabilized and fixed with Cytofix/Cytoperm kit (#554714, cells and enhances their cytotoxic activity toward PDAC cells, BD Biosciences) and stained with 20 mL anti-Cox-1-FITC/anti- administered together with Cox-2 inhibitors that reduce PGE2 Cox-2-PE mAb (clone AS70/AS67, #334090, BD Biosciences) production might be a promising approach to target gd T cell- or the appropriate isotype controls. After washing, all samples resistant Cox-2 expressing PDAC cells. were analyzed on a FACS Calibur flow cytometer (BD Bioscien- ces) using Cell Quest Pro software. To analyze the modulation of intracellular Cox-2 expression, 1.5 £ 10 / 500 mL of PDAC cells Materials and Methods were seeded in complete medium in 24-well plates together with a final concentration of 10 or 100 ng/mL recombinant TNFa or Isolation of PBMC and establishment of gd T cell lines IFNg (#210-TA/CF or 285-IF, R&D Systems) or the combina- Short-term gd T cell lines were established from peripheral tion of both for 24 h before intracellular Cox-2 expression was blood mononuclear cells (PBMC) of healthy donors or patients determined by flow cytometry. For blocking experiments during with PDAC. PBMC were isolated from leukocyte concentrates co-culture of Colo357 cells with gd T cell lines (effector/target or from heparinized blood by Ficoll-Hypaque (#6115, Bio- ratio of 5:1), the chimeric mAb Infliximab (Remicade , Janssen chrom) density gradient centrifugation. Blood from healthy Biologics B.V.) as well as the human mAb Adalimumab donors was provided by the Department of Transfusion Medi- (Humira , Abbott GmbH & Co KG), both of which bind cine or the Institute of Immunology, whereas heparinized blood TNFa, were used to prevent the activation of the TNFa recep- from PDAC patients was obtained from the Department of Gen- tor. As an irrelevant control human IgG (#1–001-A, R&D Sys- eral Surgery of the municipal hospital. Informed consent was tems) was added. The optimal concentrations of the mAb for obtained from all donors, and the research was approved by the inhibition were previously titrated using final concentrations of relevant institutional review boards (code number: D 405/10, 2, 5, 10 and 50 mg/mL. D404/14). PBMC were cultured in RPMI 1640 supplemented with Western blot analysis 2 mM L-glutamine, 25 mM HEPES, antibiotics, 10% FCS For detection of Cox-2 in PDAC cells by Western Blot, (complete medium), and stimulated with 5 mM Nitrogen-con- PDAC cells were cultured for 24 h, trypsinized and lysed with taining-bisphosphonate (n-BP) zoledronic acid (Novartis) to TNE lysis buffer (50 mM Tris, 1% (v/v) NP-40, 150 mM selectively activate Vg9Vd2 gd T lymphocytes within the NaCl, 2 mM EDTA and protease inhibitors sodium fluoride PBMC. Since initially Vg9Vd2 gd T cells produced only low [10 mM] and Na VO [1 mM]). Protein concentration was ana- 3 4 amounts of IL-2, 50 U/mL rIL-2 (Novartis) were added every 2 lyzed by Coomassie Protein Assay Reagent (#1856209, Thermo d over a culture period of 14 to 21 d After the culture period, Fisher Scientific). Twenty mg of protein was separated by 10% most short-term activated gd T cell lines (termed gd T cell lines) SDS-PAGE, transferred to nitrocellulose membrane comprise > 97% Vg9Vd2 gd T cells as analyzed by flow (#10600001, GE Healthcare) and detected with 1 mg/mL cytometry. unconjugated anti-Cox-2 mAb 33/Cox-2 (#G10203, BD Biosciences) followed by peroxidase (POD)-conjugated rabbit PDAC cell lines anti-mouse (0.24 mg/mL, #315–005–020, Dianova) and chemi- Colo357, Pt45P1, PancTu-I and Panc89 cells were cultured luminescence detection reagent (ECL Plus, #RPN2106, GE in complete medium. All PDAC cells were kindly provided by Healthcare). As control for protein load, Ab-linked POD was www.tandfonline.com OncoImmunology e988460-9 inactivated by 15% H O and reprobed with anti-b-Actin mAb plates for 20 h, and were stimulated with either medium or the 2 2 (AC-15, 0.29 mg/mL, #A5441, Sigma-Aldrich). previously titrated and indicated concentrations of Cox-inhibi- tors Indomethacin or DuP697 in the absence or presence of Real time cell analyzer BrHPP-stimulated gd T cell lines (at an E/T ratio of 5:1) Five £ 10 adherent PDAC cells were seeded in 96-well E- together with 12.5 U/mL rIL-2 for the last 4 h. After a total plates covered at the bottom with gold-electrodes (ACEA), which incubation time of 24 h, supernatants were collected and PGE2 measured the impedance changes of attached PDAC cells. The was measured by Prostaglandin E Parameter Assay Kit in dupli- impedance reflects cellular parameters such as cell number, size cates following the procedures outlined by the manufacturer and shape, which was monitored with the Real Time Cell Ana- (#SKGE004B, R&D Systems). lyzer (RTCA) single-plate (SP) system (ACEA) every 15 min for up to 24 h. The RTCA-SP system displays impedance in arbi- Statistical analysis trary cell index units, which correlates with cell number. After Since no violation of normal distribution assumption was the PDAC cells reached linear growth phase, PGE2 (#2296/10, found (Shapiro-Wilk test), all statistical comparisons were done Tocris, R&D Systems), Indomethacin (#1708/100, Cox-1/2 parametrically by using t-tests. All statistical tests were 2-sided inhibitor), DuP697 (#1430/10, Cox-2 inhibitor) (both from and the level of significance was set at 5%, not corrected for mul- Tocris, R&D Systems), PAg Bromohydrin-pyrophosphate tiple testing. (BrHPP, selective Vg9Vd2 gd T cell stimulator, Innate Pharma), tribody [(Her2) £Vg9] [consisting of 2 single chain variable fragments derived from trastuzumab (Herceptin) V-regions genetically fused to a Vg9 Fab fragment via flexible linkers ]or Disclosure of Potential Conflicts of Interest 1:500 DMSO (#1.02952.1000, Merck) were pre-incubated over No potential conflicts of interest were disclosed. night or added directly in the previously titrated optimal concen- trations as indicated. Thereafter, gd T cell lines as effector cells together with 12.5 IU/mL IL-2 were added to the RTCA at the indicated effector/target (E/T) ratio. Cell index was used to quan- Acknowledgments titatively monitor cytotoxicity induced by added compounds This work forms part of the master thesis of DG. We grate- and/or gd T cells, which themselves did not interact with the fully acknowledge Dr. Ilka Vogel and Elfi Jerg for organizing and electrodes due to their non-adherent phenotype. Whenever effec- providing blood from PDAC patients. We also thank Dr. Holger tor cells induced lysis of the tumor cells, the loss of impedance of Kalthoff and Dr. Christian R€oder for providing PDAC cell lines. tumor cells was recorded. The cells were monitored every minute BrHPP was kindly provided by Innate Pharma (Marseille, for 6 h and, thereafter, every 5 min for up to an additional 20 h. France). 1% Triton X-100 (X100, Sigma) was used as a positive control for induction of maximal lysis. Cell index was analyzed using the RTCA software 1.2 (www.aceabio.com). Funding ELISA This work was supported by the Medical Faculty of Kiel Uni- To quantify PGE2 released by PDAC cells, 1.5 £ 10 PDAC versity (DW), and the DFG Pancreatic Cancer Consortium Kiel cells in 500 mL of complete medium were seeded into 24-well (DK, DW; WE 3559/2–1; SS; SE 1831/4–1). References 5. Koshiba T, Hosotani R, Miyamoto Y, Wada M, Lee 9. Martinet L, Fleury-Cappellesso S, Gadelorge M, JU, Fujimoto K, Tsuji S, Nakajima S, Doi R, Imamura Dietrich G, Bourin P, Fournie JJ, Poupot R. A reg- 1. Wang MT, Honn KV, Nie D. Cyclooxygenases, pros- M. Immunohistochemical analysis of cyclooxygenase-2 ulatory cross-talk between Vg9Vd2 T lymphocytes tanoids, and tumor progression. Cancer Metastasis Rev expression in pancreatic tumors. Int J Pancreatol 1999; and mesenchymal stem cells. Eur J Immunol 2009; 2007; 26:525-34; PMID:17763971; http://dx.doi.org/ 26:69-76; PMID:10597402; http://dx.doi.org/ 39:752-62; PMID:19197941; http://dx.doi.org/ 10.1007/s10555-007-9096-5 10.1007/BF02781733 10.1002/eji.200838812 2. Harris SG, Padilla J, Koumas L, Ray D, Phipps RP. 6. Sreeramkumar V, Fresno M, Cuesta N. Prostaglandin 10. Kabashima-Niibe A, Higuchi H, Takaishi H, Masugi Prostaglandins as modulators of immunity. Trends E2 and T cells: friends or foes? Immunol Cell Biol Y, Matsuzaki Y, Mabuchi Y, Funakoshi S, Adachi M, Immunol 2002; 23:144-50; PMID:11864843; http:// 2012; 90:579-86; PMID:21946663; http://dx.doi.org/ Hamamoto Y, Kawachi S, et al. Mesenchymal stem dx.doi.org/10.1016/S1471-4906(01)02154-8 10.1038/icb.2011.75 cells regulate epithelial-mesenchymal transition and 3. Merati K, Said SM, Andea A, Sarkar F, Ben-Josef E, 7. Ahmadi M, Emery DC, Morgan DJ. Prevention of tumor progression of pancreatic cancer cells. Cancer Mohammad R, Philip P, Shields AF, Vaitkevicius V, both direct and cross-priming of antitumor CD8C T- Sci 2013; 104:157-64; PMID:23121112; http://dx. Grignon DJ, Adsay NV. Expression of inflammatory cell responses following overproduction of prostaglan- doi.org/10.1111/cas.12059 modulator COX-2 in pancreatic ductal adenocarci- din E2 by tumor cells in vivo. Cancer Res 2008; 11. Jimeno A, Amador ML, Kulesza P, Wang X, Rubio- noma and its relationship to pathologic and clinical 68:7520-9; PMID:18794140; http://dx.doi.org/ Viqueira B, Zhang X, Chan A, Wheelhouse J, Kuramo- parameters. Am J Clin Oncol 2001; 24:447-52; 10.1158/0008-5472.CAN-08-1060 chi H, Tanaka K, et al. Assessment of celecoxib phar- PMID:11586094; http://dx.doi.org/10.1097/ 8. Baratelli F, Lin Y, Zhu L, Yang SC, Heuze-Vourc’h N, macodynamics in pancreatic cancer. Mol Cancer Ther 00000421-200110000-00007 Zeng G, Reckamp K, Dohadwala M, Sharma S, Dubi- 2006; 5:3240-7; PMID:17172427; http://dx.doi.org/ 4. Molina MA, Sitja-Arnau M, Lemoine MG, Frazier nett SM. Prostaglandin E2 induces FOXP3 gene 10.1158/1535-7163.MCT-06-0565 ML, Sinicrope FA. Increased cyclooxygenase-2 expres- expression and T regulatory cell function in human 12. Vosooghi M, Amini M. The discovery and develop- sion in human pancreatic carcinomas and cell lines: CD4C T cells. J Immunol 2005; 175:1483-90; ment of cyclooxygenase-2 inhibitors as potential anti- growth inhibition by nonsteroidal anti-inflammatory PMID:16034085; http://dx.doi.org/10.4049/ cancer therapies. Expert Opin Drug Discov 2014; drugs. Cancer Res 1999; 59:4356-62; jimmunol.175.3.1483 PMID:10485483 e988460-10 OncoImmunology Volume 4 Issue 3 9:255-67; PMID:24483845; http://dx.doi.org/ 24. Wesch D, Beetz S, Oberg HH, Marget M, Krengel K, melanoma-infiltrating T lymphocytes. Cancer Immu- 10.1517/17460441.2014.883377 Kabelitz D. Direct costimulatory effect of TLR3 ligand nol Immunother 2011; 60:111-22; PMID:20960188; 13. Xu XF, Xie CG, Wang XP, Liu J, Yu YC, Hu HL, Guo poly(I:C) on human gd T lymphocytes. J Immunol http://dx.doi.org/10.1007/s00262-010-0924-z CY. Selective inhibition of cyclooxygenase-2 suppresses 2006; 176:1348-54; PMID:16424161; http://dx.doi. 36. Vang T, Abrahamsen H, Myklebust S, Horejsi V, the growth of pancreatic cancer cells in vitro and in org/10.4049/jimmunol.176.3.1348 Tasken K. Combined spatial and enzymatic regulation vivo. Tohoku J Exp Med 2008; 215:149-57; 25. Helm O, Mennrich R, Petrick D, Goebel L, Freitag- of Csk by cAMP and protein kinase a inhibits T cell PMID:18577844; http://dx.doi.org/10.1620/tjem. Wolf S, Roder C, Kalthoff H, R€ocken C, Sipos B, receptor signaling. J Biol Chem 2003; 278:17597-600; 215.149 Kabelitz D, et al. Comparative characterization of PMID:12665526; http://dx.doi.org/10.1074/jbc. 14. Dragovich T, Burris H, III, Loehrer P, Von Hoff DD, stroma cells and ductal epithelium in chronic pancreati- C300077200 Chow S, Stratton S, Green S, Obregon Y, Alvarez I, tis and pancreatic ductal adenocarcinoma. PLoS One 37. Nishigaki N, Negishi M, Ichikawa A. Two Gs-coupled Gordon M. Gemcitabine plus celecoxib in patients 2014; 9:e94357; PMID:24797069; http://dx.doi.org/ prostaglandin E receptor subtypes, EP2 and EP4, differ with advanced or metastatic pancreatic adenocarci- 10.1371/journal.pone.0094357 in desensitization and sensitivity to the metabolic inac- noma: results of a phase II trial. Am J Clin Oncol 26. Martinet L, Jean C, Dietrich G, Fournie JJ, Poupot R. tivation of the agonist. Mol Pharmacol 1996; 50:1031- 2008; 31:157-62; PMID:18391600; http://dx.doi.org/ PGE2 inhibits natural killer and gd T cell cytotoxicity 7; PMID:8863851 10.1097/COC.0b013e31815878c9 triggered by NKR and TCR through a cAMP-mediated 38. Patrignani P, Tacconelli S, Sciulli MG, Capone ML. 15. Lipton A, Campbell-Baird C, Witters L, Harvey H, Ali PKA type I-dependent signaling. Biochem Pharmacol New insights into COX-2 biology and inhibition. S. Phase II trial of gemcitabine, irinotecan, and cele- 2010; 80:838-45; PMID:20470757; http://dx.doi.org/ Brain Res Brain Res Rev 2005; 48:352-9; coxib in patients with advanced pancreatic cancer. J 10.1016/j.bcp.2010.05.002 PMID:15850674; http://dx.doi.org/10.1016/j. Clin Gastroenterol 2010; 44:286-8; PMID:20216081; 27. Chulada PC, Thompson MB, Mahler JF, Doyle CM, brainresrev.2004.12.024 http://dx.doi.org/10.1097/MCG.0b013e3181cda097 Gaul BW, Lee C, Tiano HF, Morham SG, Smithies O, 39. Yamaguchi K, Nakano K, Nagai E, Chijiiwa K, 16. Lunardi S, Muschel RJ, Brunner TB. The stromal com- Langenbach R. Genetic disruption of Ptgs-1, as well as Kinoshita M, Ohta M, Tanaka M. Ki-ras mutations in partments in pancreatic cancer: are there any therapeu- Ptgs-2, reduces intestinal tumorigenesis in Min mice. codon 12 and p53 mutations (biomarkers) and cytol- tic targets? Cancer Lett 2014; 343:147-55; Cancer Res 2000; 60:4705-8; PMID:10987272 ogy in bile in patients with hepatobiliary-pancreatic car- PMID:24141189; http://dx.doi.org/10.1016/j.canlet. 28. Crofford LJ. COX-1 and COX-2 tissue expression: cinoma. Hepatogastroenterology 2005; 52:713-8; 2013.09.039 implications and predictions. J Rheumatol Suppl 1997; PMID:15966189 17. Siegel R, Ma J, Zou Z, Jemal A. Cancer statistics, 2014. 49:15-9; PMID:9249646 40. Kundu N, Ma X, Kochel T, Goloubeva O, Staats P, CA Cancer J Clin 2014; 64:9-29; PMID:24399786; 29. Hill R, Li Y, Tran LM, Dry S, Calvopina JH, Garcia A, Thompson K, Martin S, Reader J, Take Y, Collin P, http://dx.doi.org/10.3322/caac.21208 Kim C, Wang Y, Donahue TR, Herschman HR, et al. et al. Prostaglandin E receptor EP4 is a therapeutic tar- 18. Oberg HH, Peipp M, Kellner C, Sebens S, Krause S, Cell intrinsic role of COX-2 in pancreatic cancer devel- get in breast cancer cells with stem-like properties. Petrick D, Adam-Klages S, R€ocken C, Becker T, Vogel opment. Mol Cancer Ther 2012; 11:2127-37; Breast Cancer Res Treat 2014; 143:19-31; I, et al. Novel bispecific antibodies increase gd T-cell PMID:22784710; http://dx.doi.org/10.1158/1535- PMID:24281828; http://dx.doi.org/10.1007/s10549- cytotoxicity against pancreatic cancer cells. Cancer Res 7163.MCT-12-0342 013-2779-4 2014; 74:1349-60; PMID:24448235; http://dx.doi. 30. Alfranca A, Iniguez MA, Fresno M, Redondo JM. Pros- 41. Ma X, Holt D, Kundu N, Reader J, Goloubeva O, org/10.1158/0008-5472.CAN-13-0675 tanoid signal transduction and gene expression in the Take Y, Fulton AM. A prostaglandin E (PGE) receptor 19. Shojaei H, Oberg HH, Juricke M, Marischen L, Kunz endothelium: role in cardiovascular diseases. Cardio- EP4 antagonist protects natural killer cells from PGE- M, Mundhenke C, Gieseler F, Kabelitz D, Wesch D. vasc Res 2006; 70:446-56; PMID:16458868; http:// mediated immunosuppression and inhibits breast can- Toll-like receptors 3 and 7 agonists enhance tumor cell dx.doi.org/10.1016/j.cardiores.2005.12.020 cer metastasis. Oncoimmunology 2013; 2:e22647; lysis by human gd T cells. Cancer Res 2009; 69:8710- 31. Narumiya S, Sugimoto Y, Ushikubi F. Prostanoid PMID:23482441; http://dx.doi.org/10.4161/onci. 7; PMID:19887600; http://dx.doi.org/10.1158/0008- receptors: structures, properties, and functions. Physiol 22647 5472.CAN-09-1602 Rev 1999; 79:1193-226; PMID:10508233 42. Chen Q, Muramoto K, Masaaki N, Ding Y, Yang H, 20. Sipos B, Moser S, Kalthoff H, Torok V, Lohr M, Klop- 32. Pino MS, Nawrocki ST, Cognetti F, Abruzzese JL, Mackey M, Li W, Inoue Y, Ackermann K, Shirota H, pel G. A comprehensive characterization of pancreatic Xiong HQ, McConkey DJ. Prostaglandin E2 drives et al. A novel antagonist of the prostaglandin E(2) EP ductal carcinoma cell lines: towards the establishment cyclooxygenase-2 expression via cyclic AMP response (4) receptor inhibits Th1 differentiation and Th17 of an in vitro research platform. Virchows Arch 2003; element activation in human pancreatic cancer cells. expansion and is orally active in arthritis models. Br J 442:444-52; PMID:12692724 Cancer Biol Ther 2005; 4:1263-9; PMID:16319525; Pharmacol 2010; 160:292-310; PMID:20423341; 21. Murakami M, Naraba H, Tanioka T, Semmyo N, http://dx.doi.org/10.4161/cbt.4.11.2138 http://dx.doi.org/10.1111/j.1476-5381.2010.00647.x Nakatani Y, Kojima F, Ikeda T, Fueki M, Ueno A, Oh 33. Ponti C, Gibellini D, Boin F, Melloni E, Manzoli FA, 43. Sobolewski C, Cerella C, Dicato M, Ghibelli L, Die- S, Kudo I. Regulation of prostaglandin E2 biosynthesis Cocco L, Zauli G, Vitale M. Role of CREB transcrip- derich M. The role of cyclooxygenase-2 in cell prolifer- by inducible membrane-associated prostaglandin E2 tion factor in c-fos activation in natural killer cells. Eur ation and cell death in human malignancies. Int J Cell synthase that acts in concert with cyclooxygenase-2. J J Immunol 2002; 32:3358-65; PMID:12432566; Biol 2010; 2010:215158; PMID:20339581; http://dx. Biol Chem 2000; 275:32783-92; PMID:10869354; http://dx.doi.org/10.1002/1521-4141(200212) doi.org/10.1155/2010/215158 http://dx.doi.org/10.1074/jbc.M003505200 32:12%3c3358::AID-IMMU3358%3e3.0.CO;2-Q 44. Lipton A, Harvey H, Witters L, Kerr S, Legore K, 22. Gupta NK, Dixit VK. Bioavailability enhancement of 34. Samten B, Howard ST, Weis SE, Wu S, Shams H, Campbell C. Gemcitabine/Irinotecan/celecoxib in pan- curcumin by complexation with phosphatidyl choline. Townsend JC, Safi H, Barnes PF. Cyclic AMP response creatic cancer. Oncology (Williston Park) 2004; 18:43- J Pharm Sci 2011; 100:1987-95; PMID:21374628; element-binding protein positively regulates produc- 5; PMID:15685834 http://dx.doi.org/10.1002/jps.22393 tion of IFN-g by T cells in response to a microbial 45. Xu XF, Xie CG, Wang XP, Liu J, Yu YC, Hu HL, Guo 23. Pietschmann K, Beetz S, Welte S, Martens I, Gruen J, pathogen. J Immunol 2005; 174:6357-63; CY. Selective inhibition of cyclooxygenase-2 suppresses Oberg HH, Wesch D, Kabelitz D. Toll-like receptor PMID:15879136; http://dx.doi.org/10.4049/ the growth of pancreatic cancer cells in vitro and in expression and function in subsets of human gd T lym- jimmunol.174.10.6357 vivo. Tohoku J Exp Med 2008; 215:149-57; phocytes. Scand J Immunol 2009; 70:245-55; 35. Su Y, Jackson EK, Gorelik E. Receptor desensitization PMID:18577844; http://dx.doi.org/10.1620/tjem. PMID:19703014; http://dx.doi.org/10.1111/j.1365- and blockade of the suppressive effects of prostaglandin 215.149 3083.2009.02290.x E(2) and adenosine on the cytotoxic activity of human www.tandfonline.com OncoImmunology e988460-11

Journal

OncoImmunology – Taylor & Francis

Published: Mar 4, 2015

Keywords: Cox-2; cytotoxicity; gammadelta T lymphocytes; human; pancreatic ductal adenocarcinoma; PGE2; BrHPP, bromohydrin-pyrophosphate; Cox, cyclooxygenase; n-BP, nitrogen-containing bisphosphonates; PAg, phosphorylated antigen; PDAC, pancreatic ductal adenocarcinoma; PG, prostaglandins; RTCA, Real Time Cell Analyzer; TCR, T cell receptor.

There are no references for this article.