[Arg6, D-Trp7,9, NmePhe8]-substance P (6Ð11) (antagonist G) induces AP-1 transcription and sensitizes cells to chemotherapy

Arg6, D-Trp7,9, NmePhe8]-substance P (6–11) (antagonist G) inhibits small cell lung cancer (SCLC) growth and is entering Phase II clinical investigation for the treatment of SCLC. As well as acting as a neuropeptide receptor antagonist, antagonist G stimulates c-jun-Nterminal kinase (JNK) activity and apoptosis in SCLC cells. We extend these findings and show that the stimulation of JNK and apoptosis by antagonist G is dependent upon the generation of reactive oxygen species (ROS) being inhibited either by anoxia or the presence of N-acetyl cysteine (n-AC). Antagonist G is not intrinsically a free radical oxygen donor but stimulates free radical generation specifically within SCLC cells (6.2-fold) and increases the activity of the redox-sensitive transcription factor AP-1 by 61%. In keeping with this, antagonist G reduces cellular glutathione (GSH) levels (38% reduction) and stimulates ceramide production and lipid peroxidation (112% increase). At plasma concentrations achieved clinically in the phase I studies, antagonist G augments, more than additively, growth inhibition induced by etoposide. Our results suggest that antagonist G may be particularly effective as an additional treatment with standard chemotherapy in SCLC. These novel findings will be important for the clinical application of this new and exciting compound and for the future drug development of new agents to treat this aggressive cancer. © 2000 Cancer Research Campaign

inhibiting neuropeptide stimulated mitogenic signalling (Mitchell et al, 1995;Seckl et al, 1995;. However, recently it has been shown that in addition to inhibiting neuropeptide responses, antagonist G, and its analogue antagonist D, also have important agonist activity, activating c-jun N terminal kinase (JNK) in a neuropeptide-independent manner (Jarpe et al, 1998a;MacKinnon et al, 1999). These studies provide the molecular basis to understanding the mechanisms by which antagonist G inhibits both basal and neuropeptide-stimulated SCLC growth (Sethi et al, 1992;Langdon et al, 1994).
Apoptosis is an active cell-death programme which is initiated via specific signal transduction cascades (Kerr et al, 1972;Kyprianou et al, 1992). Factors affecting the balance between SCLC cell proliferation and apoptosis will have a profound effect on tumour growth. The stress-activated protein kinases or c-jun N-terminal kinases (JNKs) have been shown to be activated in response to apoptotic stimuli in several cell types (Verheij et al, 1996;Le-Niculescu, 1999). Previous studies have shown that expression of a dominant negative JNK1 mutant in SCLC cells inhibits UV-induced apoptosis, suggesting an essential role of JNK1 in the induction of apoptosis in SCLC cells (Butterfield et al, 1997). Our previous work has shown that antagonist G activates JNK1 in SCLC cells (MacKinnon et al, 1999).
The role of reactive oxygen species (ROS) in pathways leading to apoptosis is particularly pertinent in cancers where the oxygen tension at the centre of tumours may be particularly low (Bush et al, 1978). Although the exact role of ROS in the induction of apoptosis is unclear, it is becoming well established that ROS may be involved in the relay of early apoptotic signals by acting as signalling intermediates (Lo and Cruz, 1995). Recent evidence [Arg 6 ,9 , N me Phe 8 ]-substance P (6Ð11) (antagonist G) induces AP-1 transcription and sensitizes cells to chemotherapy suggests that different forms of chemotherapy primarily exert their cytotoxic effects by inducing apoptosis (Friesen et al, 1999). The generation of intracellular ROS and/or the depletion of intracellular antioxidants may sensitize tumours to the pro-apoptotic effects of chemotherapeutic agents (Carmichael et al, 1988).
The aim of this study was to define the molecular mechanisms underlying these observations and their relevance to the use of antagonist G in clinical practice. Our results show that antagonist G induces ROS generation and stimulates lipid peroxidation while reducing GSH levels. This is cell type-specific and plays an important part in the apoptotic effect of antagonist G. In addition we show that antagonist G sensitizes SCLC cells to chemotherapeutic agents. These findings have important implications in the future clinical use and development of antagonist G and in the design of novel therapeutic agents to treat this aggressive cancer.

Measurement of apoptosis
SCLC cells (10 6 cells ml -1 ) were incubated in quiescent medium in the presence or absence of antagonist G or other test agents for 24 h. CHO-K1 cells were incubated with antagonist G in DMEM containing 0.1% fetal bovine serum for 24 h.

Morphology
Cells were cytocentrifuged onto glass slides, fixed with methanol, stained with May-Grünwald-Giemsa stain and examined using an Olympus BH-2 microscope, at a magnification of × 400. Apoptotic SCLC cells displayed typical morphological features, including cell shrinkage and chromatin condensation as previously described (Tallet et al, 1996).
Annexin V binding H69 cells and trypsinized CHO-K1 cells were washed in PBS and incubated with annexin V-FITC (1:250) for 10 min at 4°C then fixed in 3% paraformaldehyde. The percentage of annexin V positive cells were determined by flow cytometry using an EPICS Profile II flow cytometer (Coulter Electronics, Luton, UK).
DNA laddering DNA ladders from antagonist G-treated cells were prepared as described (Tallet et al, 1996).

Measurement of intracellular reactive oxygen species (ROS)
Intracellular ROS were measured by dihydrorhodamine fluorescence. H69 SCLC cells and trypsinized CHO-K1 cells were washed and incubated (10 6 cells ml -1 ) with 1 µM dihydrorhodamine (DHR) for 5 min at 37°C in phosphate buffered saline (PBS) prior to the addition of antagonist G or H 2 O 2 for a further 10 min at 37°C. Fluorescence was measured by flow cytometry using an EPICS Profile II flow cytometer (Coulter Electronics, Luton, UK).

Measurement of c-jun-N terminal kinase (JNK)
Quiesced H69 cells were washed twice in PBS (pH 7.4), and incubated for various times with test agents as indicated in the figure legends. Following immunoprecipitation of JNK1 from whole cell lysates, kinase activity was carried out as described (Coso et al, 1995;MacKinnon et al, 1999) 1 µg GST-c-jun(79) substrate. Phosphorylated c-jun was identified from autoradiographs of Coomasie Blue stained-SDS-PAGE gels, and quantified by optical densitometry.

Measurement of lipid peroxidation and cellular glutathione levels
SCLC cells (10 7 cells ml -1 ) in serum-free media were incubated with antagonist G for 24 h. Lipid peroxides were assayed as thiobarbituric acid reactive substances (TBARS) as previously described (Ohkawa et al, 1979). Following precipitation of proteins with 0.1% (v/v) Triton X-100 and 0.6% (w/v) sulphosalicylic acid, total cellular glutathione was measured as described (Tietze, 1969).

Ceramide quantitation
SCLC cells (10 7 cells ml -1 ) in SITA medium were treated with test agents as indicated and the lipids extracted (Bligh et al, 1959) and dried under vacuum. Ceramide was quantitated using the diacylglycerol (DAG) kinase assay as previously described (Jayadev et al, 1994).

Chemosensitivity
SCLC cells, 3-5 days post-passage, were washed and resuspended in SITA medium at a density of 5 × 10 4 cells ml -1 in the presence of 0-100 µg ml -1 etoposide with or without antagonist G. At each time-point cell number was determined using a Coulter counter (Coulter Electronics).
Antagonist G-induced apoptosis was blocked by co-incubation with the ROS scavenger N-acetyl cysteine (10 mM n-AC) and by incubation under anoxic conditions ( Figure 1B). This extends our previous work and suggests that similar to JNK activation (MacKinnon et al, 1999), antagonist G-induced apoptosis occurs via an oxidant-dependent mechanism in SCLC cells.

Antagonist G-induced generation of intracellular ROS in SCLC cells is cell type-specific
The finding that the apoptotic effect of antagonist G was abrogated by n-AC and anoxia suggests that antagonist G requires the generation of ROS to induce apoptosis in SCLC cells. This was formally examined by measuring intracellular ROS generation. Cells were preloaded with the cell-permeable oxidant-sensitive fluorescent dye dihydrorhodamine (DHR) and ROS generation followed as an increase in fluorescence measured by flow cytometry. H 2 O 2 (200 µM) increased DHR fluorescence, causing a 5.1fold increase in H69 cells (from 1.8 ± 0.09 to 9.2 ± 3.2 mean fluorescence units; n = 4, Figure 2B). Similar results were shown in the H510 cell line (data not shown). At concentrations which induce apoptosis, antagonist G produced a dose-dependent increase in DHR fluorescence ( Figure 2A) causing a 6.2-fold increase with 100 µM antagonist G in SCLC cells. The finding that this effect was evident within 15 min exposure to antagonist G and well before morphological changes were apparent suggests that this is a primary event involved in a pro-apoptotic pathway in SCLC cells. This was emphasized by the finding that although the antagonist G-induced increase in DHR fluorescence and apoptosis was blocked by n-acetyl cysteine ( Figure 2B and 2D), only the apoptotic effect was inhibited by the caspase inhibitor ZVAD-fmk ( Figure 1C). ZVAD-fmk had no effect on antagonist-G induced DHR fluorescence (data not shown). Antagonist G was not acting as an oxygen donor itself as it had no effect on cytochrome c oxidation in vitro (data not shown). However, antagonist G at concentrations which increased ROS generation in the SCLC cell lines had no effect on DHR fluorescence in CHO-K1 cells ( Figure  2A). In addition, antagonist G did not induce apoptosis in CHO-K1 cells ( Figure 2C). These experiments indicate that the oxidantdependent apoptotic effect of antagonist G is not non-specific and generally cytotoxic, but is cell type-specific.

Antagonist G depletes cellular glutathione
In keeping with a ROS-dependent mechanism, antagonist G (30 µM) increased peroxyl radical generation by 112% in H510 cells ( Figure 3C) and depleted the levels of GSH by 38% in H510 cells. This reduction in cellular GSH levels was similar to that observed with a maximal concentration of L-buthionine-(SR)sulfoximine (BSO, 0.1 mM), an inhibitor of the GSH synthesizing enzyme glutamylcysteine synthetase ( Figure 3A). Ceramide is an important lipid second messenger which has been implicated in the apoptotic response to a wide variety of cellular stresses including oxidant stress (Haimovitz-Friedman et al, 1994; Verheij et al, 1996). The JNK cascade is one of the downstream signalling pathways activated by ceramide. We show that antagonist G (50 µM) increases ceramide levels in SCLC cells after 1 h and 24 h treatment ( Figure 3B). Figure 4A shows that antagonist G (25 µM) and H 2 O 2 (200 µM) stimulate JNK activity in H69 cells. The transcription factor AP-1 has been associated with apoptosis (Staal et al, 1990;Devary et al, 1991;Jarpe et al, 1998b) and its transcriptional activity is increased by JNK (Staal et al, 1990;Devary et al, 1991). Figure 4B shows that antagonist G (10 µM) increased AP-1 activation by 61 ± 14% (n = 4) compared to an 84 ± 18% increase induced by H 2 O 2 (200 µM, n = 4, Figure 4B). The specificity was confirmed by the inhibition of binding by unlabelled AP-1 but not NFκβ ( Figure 4B). Thus, antagonist G stimulates JNK and AP-1 activation via a redox-sensitive mechanism in SCLC cells.

Antagonist G sensitizes SCLC cells to chemotherapy
Previous work has shown that modulation of the redox balance by altering the levels of GSH changes the response of cells to cytotoxic drugs (Russo et al, 1984). Moreover, the levels of GSH have been shown to increase in SCLC and NSCLC cells as they become resistant to chemotherapy (Ikegaki et al, 1994). The ability of antagonist G to increase ROS and deplete GSH suggests that antagonist G may sensitize SCLC cells to chemotherapeutic agents. We therefore examined whether antagonist G could potentiate growth inhibition induced by the chemotherapeutic agent etoposide. Etoposide (0.3 µg ml -1 ), or antagonist G (5 µM) had no effect on H69 SCLC cell growth individually, however, when added in combination, H69 SCLC cell proliferation was inhibited by 24% ( Figure 5A). 50 µM antagonist G did inhibit growth by 59% and this was potentiated to 79% inhibition in the presence of 0.3 µg ml -1 etoposide. The effect of antagonist G on etoposideinduced growth inhibition was more than additive and was synergistic showing an augmentation of the maximum response to etoposide rather than a change in affinity ( Figure 2B). These data show that antagonist G can sensitize SCLC cells to etoposideinduced growth inhibition, and suggest an additional role for antagonist G to be co-administered with standard chemotherapy in the treatment of SCLC.

DISCUSSION
We demonstrate that altering the redox status of the cell toward a more reduced environment modulates the pro-apoptotic effect of antagonist G. This was demonstrated from observations that the induction of apoptosis by antagonist G is inhibited under anoxic conditions and that antagonist G directly increases the generation of ROS within the cell. Antagonist G is not an oxygen donor itself as it has no effect on isolated cytochrome c activity in vitro. We suggest that the ability of antagonist G to stimulate intracellular ROS is not a general cytotoxic effect as it is unable to increase ROS generation and induce apoptosis in CHO-K1 cells and in isolated human neutrophils (unpublished observations). Therefore, it appears that antagonist G increases cellular ROS specifically in SCLC cells. This, coupled with the finding that anoxia and Nacetyl cysteine block the ability of antagonist G to induce apoptosis, suggests that this is an important part of its pro-apoptotic mechanism of action, which occurs upstream of caspase activation.
The transcription factor AP-1 is regulated by JNK and its transcriptional activity is increased by ROS in many cell types (Staal et al, 1990;Devary, 1991;Buttke and Sandstrom, 1994;Cossarizza et al, 1995;Lo and Cruz, 1995). Although increased JNK and AP-1 activity is associated with cellular proliferation in many cell types, in neuronal cells such as PC12 and in certain models of stress-induced apoptosis, increased activity of JNK and AP-1 are associated with apoptosis (Verheij et al, 1996; Jarpe  , 1998b;Le-Niculescu et al, 1999). This study shows that antagonist G increases AP-1 activity at concentrations which induce apoptosis and activate JNK in SCLC cells, and gives further evidence for the importance of ROS in the mediation of antagonist G's effects. We show that antagonist G depletes the levels of the major cellular antioxidant glutathione. There is evidence from other systems that such a response can lead to a stimulation of the stress pathway and programmed cell death. For example it has been shown that glutathione directly inhibits sphingomyelinase (Liu and Hannun, 1997), thereby reducing the levels of ceramide, an important inducer of apoptosis in many cells (Haimowitz-Friedman et al, 1994;Pena et al, 1997). In turn, ceramide stimulates the electron transport chain of mitochondria, leaking free  antagonist G. Total cell number was determined at day 9. The data are expressed as cell number ml -1 and represent the mean ± SEM of three independent experiments electrons to generate superoxide anion (Zamzami et al, 1995) which leads to a depletion of GSH and an increase in the generation of ROS (Quillet-May et al, 1997). In Jurkat T cells, the cellular GSH level was suggested to be critical for ROS activation of JNK in a pathway involving ras or other GTPases such as rac1 or cdc 42 as redox sensors (Lander et al, 1996). Therefore, it appears that the level of GSH within the cell plays an important role in modulating stress responses leading to apoptosis and cell death. Small cell lung cancer is characteristically resistant to oxidative damage and lipid peroxidation (Galeotti et al, 1991). Patients with lung cancer have increased levels of antioxidant GSH in the lungs, which renders them more resistant to oxidant-or chemotherapyinduced cell death (Russo et al, 1984;Melloni et al, 1996). Moreover, modulation of GSH levels has been shown to change the response of cells to cytotoxic drugs and ionizing radiation (Carmichael et al, 1988). L-buthionine-SR-sulfoximine (BSO), an inhibitor of γ-glutamylcysteine synthetase, has undergone Phase I clinical development to lower elevated GSH in patients with lung cancer (Gallo et al, 1995), however, there is no compound currently available which employs the dual function of increasing ROS and depleting GSH with the potential to cause persistent activation of JNK and apoptosis. Here we show for the first time an explicitly novel mechanism of action of antagonist G to induce apoptosis through a redox-sensitive pathway. We also show that antagonist G (at concentrations achieved clinically in the Phase I studies (Clive et al, 1999) is able to synergistically sensitize SCLC cells to etoposide at concentrations which were otherwise ineffective. This will have important implications for the future clinical use of antagonist G as tumours progress from chemosensitive to chemoresistant phenotypes. This suggests a role for antagonist G not only as a sole treatment for SCLC but as an additional treatment with chemotherapy. Antagonist G may therefore have a dual beneficial role in SCLC and may also be of benefit in a variety of other tumours by sensitizing them to subsequent chemotherapy.