Co-targeting IGF-1R and c-kit: synergistic inhibition of proliferation and induction of apoptosis in H 209 small cell lung cancer cells

Most small cell lung cancers (SCLC) coexpress the c-kit protein tyrosine receptor kinase and its ligand stem cell factor, resulting in an autocrine loop. As SCLC growth is also driven by insulin-like growth factor-1 receptor (IGF-1R) signalling, tyrphostins AG 1024 and 1296 (inhibitors of IGF-1R and c-kit activity, respectively) were used to co-target these receptors in H 209 SCLC cells. Combination treatment caused synergy in proliferation inhibition and in apoptosis induction, and also enhanced reduction in phosphorylation of Erk1/Erk2, suggesting that co-targeting IGF-1R and c-kit in SCLC may be more effective than single-agent therapies.

The signalling activity of receptor protein tyrosine kinases (PTKs) tightly controls crucial cellular processes such as apoptosis, differentiation and proliferation, and it is therefore not surprising that deregulation of these molecules often leads to neoplastic progression. The involvement of PTKs in the pathology of many cancers has brought attention to these receptors as potential targets for therapeutic intervention (Morin, 2000;Shawver et al, 2002;Dancey and Sausville, 2003). Some neoplastic conditions result from excessive activity of a single PTK (Bcr-Abl in chronic myeloid leukaemia (Melo, 1996), c-kit or PDGFRA in gastrointestinal stromal cell tumours (Hirota et al, 1998;Heinrich et al, 2003)), and are effectively treated using the PTK inhibitor Imatinib (STI 571, Gleevec) (Druker et al, 2001;van Oosterom et al, 2001). However, most cancers involve dysfunction of more than one PTK and crosstalk between their downstream signalling pathways. A therapeutic approach to address this issue involves co-targeting different PTKs (Wu et al, 1995;Ye et al, 1999;Moasser et al, 2001;Moulder et al, 2001;Lu et al, 2001;Chakravarti et al, 2002;Totpal et al, 2002;Camirand et al, 2002). For co-targeting approaches to be useful, careful consideration must be given to the choice of PTKs to be simultaneously blocked (Dancey and Sausville, 2003).
Small cell lung cancer (SCLC) is an aggressive cancer with poor prognosis, which readily metastasises outside the chest, and has a median survival of approximately 9 months (Chute et al, 1999). SCLC has been treated with traditional combination chemotherapy since the 1970s, but less than 3% of patients with extensive-stage SCLC survive more than 3 years (Chute et al, 1999). The molecular pathology of more than 70% of SCLC cases involves autocrine stimulation due to the coexpression of c-kit and its ligand stem cell factor (SCF) (Hibi et al, 1991), and high c-kit expression level is associated with a poor prognosis (Micke et al, 2003). Imatinib (which targets Bcr-Abl, c-kit, and PDGFR (Morin, 2000)), tyrphostin AG 1296 (which inhibits c-kit and PDGFR (Kovalenko et al, 1994)) and indolinones SU 5416 and SU6597 (inhibitors of c-kit and related kinases (Krystal et al, 2001)) reduce SCLC proliferation in vitro (Krystal et al, 1997(Krystal et al, , 2000(Krystal et al, , 2001Wang et al, 2000;Kijima et al, 2002). Phase I and II clinical trials with Imatinib are ongoing for SCLC.
Another growth pathway known to sustain proliferation in SCLC is the IGF-1R signalling cascade (Nakanishi et al, 1988;Macaulay et al, 1990). In many cancers, IGF-1R signalling leads to activation of PI-3 K and MAPK pathways, stimulates proliferation, promotes angiogenesis and metastasis, and inhibits apoptosis (Khandwala et al, 2000;Pollak, 2000;Yu and Rohan, 2000;Wang and Sun, 2002). Preclinical work has demonstrated that IGF-1R could be used as a successful co-target with EGFR (Chakravarti et al, 2002) and HER2/ erbB2 (Lu et al, 2001;Camirand et al, 2002), and in the latter study, synergistic inhibition of proliferation was observed in co-targeted breast cancer cells. The availability of several novel inhibitors of IGF-1R signalling (Ludwig et al, 2003;Garcia-Echeverria et al, 2003;Mitsiades et al, 2003) implies that clinical exploitation of any synergism between co-targeting of c-kit and IGF-1R will be practical.
In this study, we use tyrphostins AG 1024 (IGF-1R inhibitor) and AG 1296 (PDGFR and c-kit inhibitor) to inhibit PTK activity in H 209 SCLC cells which present no PDGFR receptors. We demonstrate that combination treatment causes synergy in inhibition of proliferation and in apoptosis induction in these cells, and we speculate that co-targeting the IGF-1R and c-kit signalling pathways is potentially a more effective approach for this system than single-agent therapies.

MATERIALS AND METHODS
Cell culture and proliferation assays H 209 SCLC cells were obtained from ATCC (Manassas, VA, USA) and cultured at 371C with 5% CO 2 in RPMI 1640 medium with 10% foetal bovine serum (FBS) (InVitrogen, Gaithersburg, MD, USA) except in growth inhibition assays where the FBS supplement was reduced to 1%. Cell proliferation was measured with the Alamar Blue dye reduction method (Biosource International, Camarilo, CA, USA), which is optimal for cells growing in anchorageindependent conditions. Tests were conducted with 10 4 cells well À1 in 200 ml media in 96-well plates (edge wells were not used, but filled with 200 ml PBS) and three replicates per dose combination were used for each experiment. Stock solutions (10 mM) in DMSO of tyrphostins AG 1024 and AG 1296 (Calbiochem, San Diego, CA, USA) were kept at À201C and diluted in 1% medium just before use. DMSO concentration in final culture was kept below 0.2% (v v À1 ). All procedures involving tyrphostins were conducted in low light intensity.

Cell cycle analysis
Approximately 10 6 cells were collected by centrifugation, washed twice in PBS and fixed in ice-cold 70% ethanol for 1 h, then stained for 30 min in the dark at 371C in 1 ml propidium iodide (PI) staining buffer (0.1% Triton X-100, 0.1 mM EDTA pH 7.4 in PBS, to which 50 mg ml À1 PI and 0.05 mg RNAse were added just before use). DNA staining was determined by flow cytometry on a FACSCalibur using CellQuest software (see previous section).

Induction of apoptosis
Approximately 10 6 cells were collected by centrifugation and washed in PBS, then stained with annexin-V-FITC and with PI using the ApoTarget kit (Biosource International, Camarillo, CA, USA). Analysis was conducted by flow cytometry on a FACSCalibur using CellQuest software (see receptor section above).

RESULTS
H 209 cells express no PDGF receptors, but present both c-kit (Kijima et al, 2002) and IGF-1R (Nakanishi et al, 1988;Quinn et al, 1996), as analysed by flow cytometry (Figure 1). Treatment with tyrphostins as single agents revealed an IC 50 for inhibition of cell proliferation of 3.7570.8 mM for AG 1024, and 21.573.3 mM for AG 1296. These two inhibitors were used in a matrix of combinations, and an isobologram plot (Berenbaum, 1981) for IC 50 demonstrated synergistic interactions in growth inhibition (Figure 2A). Combination index (CI) values for several combination treatments were calculated from the classic isobologram equation and proved significantly smaller than 1, indicating synergy. Combination treatments showed no synergy of action on the status of any phase of the cell cycle, as observed by flow cytometry after propidium iodide staining of ethanol-fixed cells (results not shown). In contrast, the synergy in growth reduction was observed to parallel a synergy in apoptosis induction (IC 75 CI ¼ 0.62 at AG 1024 2.5 mM and AG 1296 5.0 mM, Figure 2B), pointing to a cytotoxic effect in these cancer cells associated with the co-targeting of IGF-1R and c-kit. At concentrations where apoptosis is induced synergistically by combinations of the two inhibitors, co-treatment did not enhance inhibition of Akt phosphorylation over levels observed after treatment with AG 1024 alone (results not shown). However, co-treatment reduced the phosphorylation level of p44/42 (Erk1/ Erk2) by 60% with respect to control ( Figure 2C).

DISCUSSION
In this study, tyrphostins AG 1024 (IGF-1R inhibitor) and AG 1296 (PDGFR and c-kit inhibitor) were used in combination for treatment of H 209 SCLC cells. Since specificity of tyrphostin action is not strict, (for example, AG 1296 inhibits c-kit, PDGFR a and PDGFR b with similar Kds (Kovalenko et al, 1994)), we demonstrated the absence of PDGF receptors in the target cells, and confirmed the presence of IGF-1R and c-kit. The data presented here indicate that in H 209 SCLC cells, blockade of IGF-1R signalling synergistically enhances the antiproliferative effects of an anti-c-kit strategy, and that while both anti-IGF-1R and anti-c-kit strategies as single agents can induce apoptosis in this system, combination treatment results in synergy of apoptotic induction. Future work will determine whether these observations can be extended to more SCLC cell lines and have relevance to in vivo behaviour of SCLC.
Cellular pathways by which this observed synergy is achieved may involve downstream effectors differentially influenced by the inhibition of IGF-1R and c-kit. Akt/PKB, an important component of the P1-3K-mediated signal route, is involved in several human cancers, and has recently been shown to be important in the progression of lung cancer (Krystal et al, 2002;Chun et al, 2003). Isobologram plot at IC50 and combination index (CI) of the effects of AG 1094 and AG 1296 on H 209 SCLC cell growth. Tests were conducted in triplicates with 10 4 cells treated with tyrphostins singly or in combination for a total of 72 h in RPMI with 1% FBS, and proliferation of the anchorageindependent cells was assayed by Alamar blue dye reduction (shown is an experiment representative of five). Proliferation CI values were calculated using the classic isobologram equation (Berenbaum, 1981) and are indicated on a graph. CI values o1 indicate synergy. Single-agent dose -response curves on H 209 cell growth for individual tyrphostins are shown on the right. (B) Apoptosis in H 209 cells treated with tyrphostins AG 1024 (0, 0.25, 2.5 mM) and AG 1296 (0, 2.5, 5.0 mM) independently or in combination for a total of 72 h in 3 ml RPMI 1% FBS in six-well plates. Cells were collected by centrifugation, washed in PBS, and stained with annexin V-FITC and PI, and analysed by flow cytometry. Annexin-V-positive cells (M1 peak) represent apoptotic populations. Number above M1 peak is the increase in the percentage of apoptotic cells over the basal control level (Cntrl ¼ 34.5%), (shown is one of two experiments producing similar results). (C) Modification of phosphorylation levels of Erk1/Erk2 in H 209 cells treated for 72 h in media containing 1% FBS with 1.0 mM AG 1024, and/or 5.0 mM AG 1296. Western blot analysis was conducted on 50-mg protein aliquots with anti Erk1/Erk2 (p44/42) and anti-phospho Erk1/Erk2 (Thr202/Tyr204) primary antibodies and HRPconjugated anti-rabbit IgG, and revealed with the ECL chemoluminescence reaction. Densitometric analysis of the bands is shown below. C: control cells treated with vehicle (shown is one of two experiments producing similar results). SCLC therapy by IGF-1R and c-kit co-targeting Combination treatments did not reduce phosphorylation of Akt below levels obtained by AG 1024 inhibition, however, reduction in the phosphorylation of Erk1/Erk2 was enhanced by co-targeting, suggesting that combination of the IGF-1R and c-kit inhibitors (at lower concentrations than needed for single-agent activity) acts through this kinase to promote the greater-than-additive apoptotic effects observed.
It is tempting to speculate that the antiapoptotic effects of IGF-1 hinder cancer therapies that target other PTKs, and that the antineoplastic effects observed when blocking a PTK may be significantly underestimated if examined only under conditions where IGF-IR is fully functional. Previous work indicates that the in vivo targeting of IGF-1R not only has antineoplastic activity (Arteaga et al, 1989;Andrews et al, 2001) but is well-tolerated . These observations, taken together with our results, suggest that combining anti-IGF-1R and anti-kit targeting may be therapeutically useful in SCLC treatment. Targeted singleagent PTK therapies are effective in cancers where activation by mutation of one particular receptor is responsible for neoplastic progression (Druker et al, 2001;van Oosterom et al, 2001;Heinrich et al, 2002); however, in the more common situations where overexpression or deregulation of multiple PTKs are involved, combination approaches may prove to be of value and the IGF-1 receptor appears as a promising co-target.