Cell death assays
Detection of DNA fragmentation (SubG1 cells) was performed in propidium iodide stained cells as previously described [9]. In vitro determination of cytoplasmic histone-associated-DNAfragments (mono- and oligonucleosomes) was performed using Cell Death Detection ELISAPLUS (Roche Applied Science, Mississauga, ON, Canada). Both adherent and floating cells were collected and the ELISA was then carried out as per the instructions of the manufacturer.Fluorescent detection of lysosomes and autophagy
To detect and quantify AVOs, cells were stained with acridine orange (final concentration of 1 mg/ml in PBS, 15 min). Cells were then analyzed with a fluorescent microscope (Olympus, Cell software) or trypsinized and analyzed by flow cytometry (FACScan, Becton Dickinson; CellQuest software). To visualize lysosomes, cells were fixed with paraformaldehyde (4%; 20 min), permeabilized with ice cold methanol, blocked with 10% FBS and 1% BSA in PBS and subsequently incubated overnight with primary anti-LAMP-1 or anti-LAMP-2 antibody. After incubation with the AlexaFluor488-labeled secondary antibody, cells were post-fixed with paraformaldehyde, counterstained with DAPI, mounted with Prolong Gold Antifade reagent and analyzed using an Olympus Fluoview FV1000 confocal microscope (Melville, NY). To analyze autophagic flux, MET4 cells were transfected with mRFP-GFP-LC3 (also known as tandem fluorescence-LC3, tfLC3) expressing plasmid and Fugene HD (2/6 ratio) as per the manufacturer’s instructions (Roche, Germany). Cells were pooled, seeded in chamber slides and left untreated (to avoid transfectioninduced autophagy). Autophagic flux was determined by evaluating the punctuated pattern of GFP and mRFP (punctae/cell were counted). Analysis of fluorescence was done on an Olympus (Aartselaar, Belgium) cell imaging station using Cell M software.

Caspase 3 activation and poly(adenosine diphosphate-ribose)polymerase (Parp) cleavage, two hallmarks of apoptotic cell death, were observed in all SCC cell lines treated with LUT (50 mM or higher), indicating that the decrease in viability upon LUT treatment was due to apoptosis (Figure 1C). Cleavage and activation of executioner caspase 3 can be the result of two converging pathways activating specific initiator caspases via the formation of a molecular platform. We used a luminescent assay to measure the activity of caspase 8 and caspase 9, the initiator caspase of the extrinsic and the intrinsic apoptotic pathway respectively. Although the increase in caspase activity after LUT treatment differed between the used SCC cell lines and was LUT dose-dependent, we detected a significant 2- to 3-fold induction of both caspase 8 and caspase 9 (Figure 1D), suggesting the involvement of intrinsic as well as extrinsic apoptotic signaling. Furthermore, zVAD-fmk, a broad spectrum caspase inhibitor, blocked LUT-induced cell death as detected with trypan blue exclusion assay, indicating cell death was caspase dependent (Figure 1E). Hence, LUT causes a caspase dependent form of cell death specifically in malignant, SCC-derived keratinocytes.

LUT mediated apoptosis involves interference with AKT signaling
Since our previous study revealed that enhanced AKT activation parallels progressive tumor stage and increased resistance of MET1 and MET4 cells to therapeutic drugs [24], we investigated the impact of LUT treatment on the phosphorylation and thus activation of AKT (Ser473) using western blot analysis (Figure 2A). As soon as 30 minutes after LUT addition, AKT became progressively dephosphorylated in both MET1 and MET4 cells. Interestingly, the exhaustive dephosphorylation of AKT after 6 h LUT treatment in MET1 cells corresponded well with the appearance of apoptotic markers, which suggests the importance of AKT as an inhibitor of apoptosis. To rule out cell specific effect, we treated A431 and A253 cells with increasing amounts of LUT and detected also in these SCC cell lines a progressive down regulation of AKT signaling (data not shown). To confirm that LUT decreased AKT-mediated signaling, we monitored phosphorylation of the AKT substrate mTOR, and its downstream target p70S6 kinase. LUT clearly inhibited AKT signaling as indicated by the reduced phosphorylation of mTOR and p70S6 kinase in both MET1 and MET4 cells as depicted in Figure 2B. To further investigate whether MET1 and MET4 cells were dependent on AKT signaling for survival, we used a combination of an isozyme-specific AKT1/2 inhibitor (AI) to block basal AKT phosphorylation and LUT. This treatment resulted in additional cell death in MET1 and MET4 cells compared to LUT alone (Figure 2C�D). In agreement with the fact that MET4 cells are more resistant, a higher LUT dose (50 mM) had to be used to reveal the additive effect on protein level (Figure 2D). In addition, we transiently transfected MET1 and MET4 cells with a construct expressing AKT in a membrane-targeted form (myrAKT-HA). The capability of LUT to induce apoptosis active AKT was reduced in both the MET1 and MET4-cells overexpressing constitutively active AKT (Figure 2E). Altogether these data showed that LUT suppressed AKTsignaling in both the primary and metastatic SCC cells. However, MET1 primary tumor cells were more sensitive to LUT-induced apoptosis (Figure 1 & 2), suggesting that induction of survival mechanisms in advanced MET4 cells overcome LUT-induced cell death.

Statistical analysis
The data were expressed as means 6 S.D. Statistical analysis was performed by using Student’s t-test (two-tailed). The criterion for statistical significance was taken as (p,0.05) unless stated otherwise.Results Luteolin induces apoptotic cell death in SCC cells, but sensitivity decreases with SCC tumor progression
Recent studies indicate that LUT has potential anticancer effects in different tumor cell types, however these effects have not been investigated in human cutaneous SCC cells [27]. To address the therapeutic effects of LUT in SCC, we first investigated the cell death inducing effects of LUT on cell lines derived from cutaneous SCC: MET1, MET4 cells and A253 cells. Treatment of those cell lines with LUT (10?00 mM) decreased metabolic activity in all cell lines, but was more pronounced in the primary tumor cells (MET1) compared to the metastatic MET4 in A253 cells (Figure 1A). In a concentration range above 50 mM LUT was clearly cytotoxic in all tested SCC cell lines because the percentage propidium iodide (PI) positive cells increased significantly (Figure 1B). Interestingly, normal human keratinocytes (NHKs) were completely resistant to the doses of LUT killing the malignant keratinocytes (Figure 1B), in agreement with our previous results on the differential protective effect of LUT against UVB-induced apoptosis in normal and not in malignant keratinocytes [9].

Figure 1. LUT induces apoptotic cell death specifically in malignant keratinocytes. A. Detection of metabolic activity by MTT assay, 24 hours after treatment with LUT (10?00 mM) or DMSO (equal amount as highest LUT concentration) in different SCC cell lines. (Representive experiment out of 3, n = 16) B. Flow cytometric analysis of propidium iodide (PI) stained cells, 24 hours after treatment with LUT (50 mM). PI-positive cells = dead cells. (Representive experiment is shown, n = 3) C.