Concomitant inhibition of AKT and autophagy is required for efficient cisplatin-induced apoptosis of metastatic skin carcinoma



Cutaneous squamous cell carcinoma (cSCC) is one of the most common cancers in the Caucasian population. Although early stages of skin cancer have a high curability and excellent prognosis, advanced cSCC shows resistance to chemotherapy, including cisplatin. The PI3-K/AKT pathway is known to have a role in both skin cancer development and resistance to therapeutic drugs. In this study, we used isogenic cell lines representing different stages of malignant transformation of the keratinocytes that were derived from dysplastic forehead skin (PM1), primary cutaneous SCC (MET1) and its lymph node metastasis (MET4) of an immunosuppressed patient. We show that skin tumor progression parallels enhanced AKT activation and increased resistance to cisplatin-induced apoptosis. Pharmacological AKT inhibition, or specific AKT1 knock down, sensitizes the apoptosis-resistant MET1 and, to a lesser extent, MET4 cells to cisplatin-mediated cell death. Concomitantly autophagy induction was observed in MET4, as demonstrated by accumulation of the autophagic protein marker LC3-II, by analysis of full autophagosome maturation process using tandem mRFP-GFP fluorescence microscopy and by electron microscopy. Counteracting the autophagic process by 3-methyladenine or specific ATG5 knock down enhanced cytotoxicity of cisplatin combined with AKT inhibitor, thus revealing a key role for autophagy in chemoresistance. Taken together, these results indicate that concomitant inhibition of autophagy is required to increase the therapeutic benefit of AKT inhibition for combination therapy with the standard chemotherapeutic agent cisplatin in advanced skin carcinoma.

Cutaneous squamous cell carcinoma (cSCC) is one of the most common cancers in the Caucasian population and its incidence is still increasing worldwide,1, 2 and is predicted to increase even further.3 Early SCC of the skin has a high curability and relatively low overall metastatic rate of 3–5%.4 However, certain tumor and patient characteristics predispose to the development of nodal disease and distant metastases,5 which portends a poor prognosis, with 5-year survival ranging from 14 to 39%, regardless of the treatment used. Current standard initial treatments for advanced SCC are surgical resection, radiation therapy or both. Chemotherapy (with cisplatin, 5-Fluoro-Uracil, Paclitaxel, or a combination thereof) has been integrated in the standard treatment of advanced cSCC in an attempt to improve survival, but unfortunately with poor response rates and high recurrence rates, necessitating the search for new therapeutic targets.4, 6, 7 Cisplatin is one of the best studied and currently most used (alone or in combination) chemotherapeuticum in the treatment of epithelial tumors, including SCC of the skin. Cisplatin causes DNA damage, interstrand and intrastrand DNA crosslinks,8 which is thought to induce cell death via apoptosis. Resistance to cisplatin has been shown to arise, due to several mechanisms including progressive acquisition of molecular changes downregulating proapoptotic pathways and activating survival pathways, including the PI3-K/AKT signaling cascade, as demonstrated in various non-SCC tumors.9

The PI3-K/AKT signaling cascade involves activation of phosphatidylinositol 3-kinase (PI3-K), which generates phosphatidylinositol 3,4,5-trisphosphate (PIP3) leading to the recruitment of AKT to the plasma membrane where it is phosphorylated. Double phosphorylated (Ser473 and Thr308) activated AKT exerts its effects on a multitude of substrates that ultimately orchestrate the diverse cellular roles of AKT, including cell survival, growth, proliferation, angiogenesis, metabolism and migration.10 The PI3-K/AKT pathway can be deregulated via different mechanisms as demonstrated for various cancers, including constitutive activation of growth factor receptors, PI3-K amplification/mutation, inactivation of PTEN, amplification of AKT, and, as reported recently, mutational activation of AKT itself.11, 12 Constitutive AKT activation is reported in many malignant cancers.13 Recently, it has been suggested that phosphorylated AKT may play an important role in the pathogenesis of malignant tumors of the epidermis.14 The important role of PI3-K/AKT signaling in the process of cutaneous carcinogenesis is further supported by studies in mice.15, 16 In addition to its prominent role in the inhibition of apoptosis, AKT also negatively regulates macroautophagy (hereafter called autophagy) mainly through the regulation of the mTOR signaling.17 Autophagy is a highly conserved catabolic program for the degradation and recycling of cellular constituents such as long-lived proteins and organelles. Depending on the context, autophagy has been described as a prosurvival process or as a process to induce non apoptotic cell death (Type II cell death).18–20 The role of autophagy in cancer development and in the response to therapy is still controversial requiring further investigation.21, 22

In previous work, we have shown the importance of PI3-K/AKT signaling in cell death response in normal keratinocytes upon UVB irradiation.23 In this study, we investigated the involvement of PI3-K/AKT signaling in the cell death/survival response of progressive stages of skin photocarcinogenesis upon exposure to the chemotherapeutic agent cisplatin. To this end, we used different isogenic cSCC cell lines established from different stages of malignant progression of the epidermis of an immunosuppressed patient.24 We found that the increased resistance to cisplatin-induced cell death, displayed by the cells of the more malignant stages of epidermal transformation, is associated with an increased activation status of AKT. AKT knock down by siRNA or inhibition using an isozyme-selective AKT inhibitor, sensitized significantly primary cancer cells, and to lesser extent metastatic cells, to cisplatin-induced apoptosis, which can be specifically attributed to AKT1 inhibition as shown via AKT1 knock down in the metastatic cells. However, AKT inhibition also resulted in the stimulation of autophagy whose inhibition increased dramatically MET4 cells susceptibility to undergo cisplatin-induced cell death, thus revealing a chemoresistant role for autophagy in advanced cancer of the skin.


AI: AKT inhibitor VIII; AVOs: acidic vesicular organelles; CDDP: cis-diamminedichloroplatinum; cSCC: cutaneous squamous cell carcinoma; DiOC6: 3,3′-dihexyloxacarboxyanine iodide; LC3: microtubule-associated protein 1 light chain 3; 3-MA: 3-methyladenine; PARP: poly-ADP ribose polymerase; PI3-K: phosphatidylinositol 3-kinase

Material and Methods

Cells and cell culture

PM1, MET1 and MET4 keratinocyte cell lines were respectively derived from forehead skin that showed moderate to severe dysplasia, from a primary squamous cell carcinoma of the back of the left hand and its metastasis within the left axillary lymph nodes from a single immunosuppressed individual.24 MET4 is clonally derived from MET1.25 These cell lines were grown in Dulbecco's modified Eagle's medium plus HAMS F12 medium containing 10% fetal calf serum, hydrocortisone, mouse EGF and antibiotics in a 37°C incubator at 5% CO2. Cell lines were validated by STR DNA fingerprinting using the AmpFℓSTR Identifiler kit according to manufacturer instructions (Applied Biosystems (Foster City, CA) cat 4322288). The STR profiles were compared to known ATCC fingerprints (, to the Cell Line Integrated Molecular Authentication database (CLIMA) version 0.1.200808 ( (Nucleic Acids Research 37:D925-D932 PMCID: PMC2686526) and to the MD Anderson fingerprint database. The STR profiles for the PM1, MET1 and MET4 cell lines were unique. SCC12B2 cell line is derived from a human facial SCC and was grown in the same medium. Primary human keratinocytes were isolated and pooled from foreskins of 5 different donors (less than 6 years) as described.26 Third to fifth passage cells were used in experiments. Primary human keratinocytes were cultured in serum-free and growth factor-containing medium (Keratinocyte-SFM; Invitrogen, Belgium), which harbors several growth factors (5 μg/ml insulin, 74 ng/ml hydrocortisone, 6.7 ng/ml triiodo-L-thyronine, 50 μg/ml bovine pituitary extract and 5 ng/ml human recombinant epidermal growth factor). The procedure has been approved by the ethical committee of the University of Leuven. Experiments performed adhered to the Declaration of Helsinki Principles.


We obtained cis-diamminedichloroplatinum (CDDP) and acridine orange from Sigma (St. Louis, MO) and dissolved it in PBS and distillated water, respectively. All experiments were performed using 20 μg/ml CDDP. AKT inhibitor VIII (Calbiochem, San Diego, CA) was dissolved in DMSO. 3-Methyladenine (3-MA) was obtained from Sigma and dissolved in the medium used for the cell lines. FuGENE HD transfection reagent was purchased from Roche (Mannheim, Germany).

In vitro nonradioactive AKT kinase assay

PM1, MET1 and MET4 cells were treated with 20 μg/ml cisplatin for 24 hr. Cells were harvested and the AKT kinase assay from Cell Signaling Technology (Beverly, MA) was used according to manufacturer's instructions. Briefly, AKT was immunoprecipitated from cell lysates using immobilized AKT mAb. Then, an in vitro kinase assay is performed using GSK-3 fusion protein as a substrate and phosphorylation of GSK-3 is measured by Western blotting, using phospho-GSK-3α/β (Ser21/9) antibody.

Viability assays

Metabolic activity was assessed using the tetrazolium salt MTT (Sigma, St. Louis, MO). Cells were seeded in 96-well plates. After 24-hr treatment, cells were incubated with MTT (1 mg/ml) in PBS for 1 hr. Cleavage of MTT by dehydrogenase enzymes of metabolically active cells yields a blue formazan product. The formazan product was dissolved in DMSO and the absorbance at 550 nm was measured by spectrophotometry.

In addition, cell viability was assessed using the trypan blue exclusion assay. Cells were seeded in plates and treated as described. Cells were harvested using trypsin-EDTA and resuspended in 1 ml PBS and analyzed using a Vi-cell XR cell viability analyzer (Beckman Coulter, Fullerton, CA).

Detection of DNA fragmentation

Both adherent and floating MET1 cells were collected and washed twice in PBS. Cells were fixed with ice cold ethanol and stored on ice for 30 min. After 2 wash cycles in PBS containing 0.01 % Tween20 (PBS-tween), the cell pellet was dissolved in 1 ml PBS-tween containing 0.5 mg propidium iodide and 0.1 mg RNAse. Samples were stored overnight at 4°C and analyzed with a flow cytometer (FACScan, Becton Dickinson, Franklin Lakes, NJ).

Determination of loss of mitochondrial membrane potential

At the indicated time point after treatment, cells were incubated in 20 or 40 nM 3,3′-dihexyloxacarboxyanine iodide (DiOC6) at 37°C for 30 min, harvested by trypsinization and washed with cold PBS solution. Mitochondrial membrane potential was determined by flow cytometry (FACScan, Becton Dickinson, Franklin Lakes, NJ).

Cell death detection ELISA

Apoptosis in PM1, MET1 and MET4 cells was determined by in vitro determination of cytoplasmic histone-associated-DNA-fragments (mono- and oligonucleosomes) 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.

Sequence analysis

The AKT1 open reading frame was amplified with PCR and AmpliTaq Gold DNA polymerase and the GeneAmp PCR System 2400 (Applied Biosystems, Foster City, CA). The PCR products were purified and sequenced in both directions on the ABI Prism BigDye (Terminator Cycle Sequencing Kit version 1.1) on an ABI PRISM 3100 Genetic Analyser (Applied Biosystems, Foster City, CA).

Western blot analysis

Cells were scraped in medium, spun down and proteins were isolated using lysis buffer [25 mM N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid, 0.3 mM NaCl, 1.5 mM MgCl2, 20 mM β-glycerolphosphate, 2 mM EDTA and 2 mM EGTA (pH 7.5)] containing 1% Triton, 10% glycerol, 1 mM Na3VO4, 0.5 mM dithiothreitol, 10 μg/ml Leupeptin, 10 μg/ml Aprotinin, and 10 μg/ml Antipain and phosphatase inhibitor cocktail (phosSTOP) (Roche, Mannheim, Germany). Extracts were incubated on ice for 20 min and spun down at 20,800g for 20 min. Protein concentration was determined using the BCA Protein Assay Reagent (Pierce Chemical Company, Rockford, IL). Equal amounts of protein from each sample were separated by electrophoresis through SDS-PAGE gels (Invitrogen, Merelbeke, Belgium) and transferred to Hybond-C Super membrane (Amersham Pharmacia Biotech, Piscataway, NJ). Equal loading of proteins was verified using Ponceau-S. Membranes were blocked for 1 hr at room temperature in Tris-buffered saline containing 0.1% Tween-20 and 5% nonfat dry milk. The membrane was incubated overnight at 4°C with the primary antibody diluted in 5% nonfat dry milk or 5% BSA in 1× TBS plus 0.1% Tween 20. We purchased the PARP and anti-caspase-8 antibody from BD Biosciences (San Jose, CA), and LC3 antibody from Nanotools (Teningen, Germany). Antibodies against phospho-AKT (Ser473), total AKT, AKT1, AKT2, PTEN, phospho-p70S6K (Thr389), caspase-9 and caspase-3 were obtained from Cell Signaling Technology (Beverly, MA). We purchased active caspase-3 and p70S6K antibody from Epitomics (Burlingame, CA). The primary antibody against actin (JLA20) was purchased from Developmental Studies Hybridoma Bank at the University of Iowa. The membranes were washed, and incubated for 1 hr at room temperature with the peroxidase-conjugated secondary antibody (Cell Signaling Technology, Beverly, MA). Protein bands were visualized using enhanced chemiluminescence as described by the supplier (Amersham Pharmacia Biotech, Piscataway, NJ).

Flow cytometric quantification of acidic vesicular organelles (AVOs)

To detect and quantify AVOs, cells were washed with PBS, stained with acridine orange at a final concentration of 1 μg/ml in PBS for a period of 15 min. Then, the cells were removed from the plate with trypsin-EDTA, collected, spun down and resuspended in 1 ml PBS and analyzed by flow cytometry (FACScan, Becton Dickinson, Franklin Lakes, NJ) and data were analyzed using CellQuest (BD Biosciences, San Jose, CA) software.

Monitoring autophagic flux using tandem mRFP-GFP-tagged LC3 (tfLC3)

After transfection of MET4 cells in a 100-mm cell culture dish with 4 μg tfLC3 expressing plasmid and Fugene HD (12 μl/P100) as per the manufacturer's instructions (Roche, Mannheim, Germany), medium was changed after 24 hr and cells were left untreated for another 24 hr (to avoid transfection-induced autophagy). Cells were then trypsinized and split in a 96-well plate. After a 24-hr culture and treatment, images were acquired on the IN Cell Analyzer 1000. Autophagic flux was determined by evaluating the punctuated pattern of GFP and mRFP. The tfLC3 construct was a kind gift from Dr. Tamotsu Yoshimori (Osaka University).

Transient transfection of siRNA

MET4 cells in exponential phase of growth were transfected with siRNA (Dharmacon, Lafayette, CO) targeting AKT1, AKT2, ATG5 or scrambled siRNA (-pool), using DharmaFECT (Dharmacon) or Lipofectamine™ RNAiMAX (Invitrogen, Merelbeke, Belgium) according to manufacturer's instructions. Cells were split allowing the same amount of transfection efficiency for the different conditions. The day after, MET4 cells were treated with cisplatin in the presence or absence of AI and 3-MA and cells were harvested 24 hr later.

Transmission electron microscopy

Cells were grown and treated as described. After trypsinization and centrifugation, the cells were immediately fixed in 2.5% glutaraldehyde, 0.1 mol/l phosphate buffer, and pH 7.2 at 4°C overnight. After 1-hr postfixation in 1% osmium tetroxide, 0.1 mol/l phosphate buffer at 4°C, the samples were dehydrated in graded series of alcohol and embedded in epoxy resin. Ultra-thin sections 50–60 nm were cut, stained with uranyl acetate and lead citrate and examined at 50 kV using a Zeiss EM 900 electron microscope (Oberkochen, Germany). Images were recorded digitally using a Jenoptik Progress C14 camera system (Jena, Germany) operated using Image-Pro express software (Media cybernetics, USA).

Mass spectroscopy-based approach evaluating single nucleotide polymorphisms

A mass spectroscopy-based approach evaluating single nucleotide polymorphisms (SNP) was used to detect the AKT1_E17K mutation and “hotspot mutations” in exon 9 and exon 20 of PIK3CA (PIK3CA_E542K, PIK3CA_E545K, PIK3CA_E542Q, PIK3CA_H1047R, PIK3CA_H1047L, PIK3CA_H1047Y). PCR and extension primers for AKT and PIK3CA were designed using Sequenom, Inc. Assay Design. PCR-amplified DNA was cleaned using EXO-SAP (Sequenom, San Diego, CA), and primer was extended by IPLEX chemistry, desalted using Clean Resin (Sequenom, San Diego, CA), and spotted onto Spectrochip matrix chips using a nanodispenser (Samsung, Irvine, CA). Chips were run in duplicate on a Sequenom MassArray MALDI-TOF MassArray system. Sequenom Typer Software and visual inspection were used to interpret mass spectra. Reactions where >15% of the resultant mass ran in the mutant site in both reactions were scored as positive.

Statistical analysis

The data were expressed as means ± S.D. Statistical analysis was performed by using Student's t-test (2-tailed). The criterion for statistical significance was taken as p < 0.05.


Increased resistance to cisplatin-induced apoptosis is associated with progressive stages of keratinocyte malignancy

The effect of cisplatin was studied in the epidermal cell lines PM1, MET1 and MET4, respectively derived from dysplastic forehead skin, primary SCC of the back of the left hand, and its lymph node metastasis from a single immunosuppressed patient. Hence, PM1 as an early stage, MET1 as a late and MET4 as very late stage of cSCC represent a unique model for skin tumor progression. Treatment of these cSCCs with CDDP decreased both metabolic activity and cell viability as demonstrated respectively by MTT assay (Supporting Information Fig. S1a) and trypan blue exclusion assay (Fig. 1a). However, the cytotoxic effect of CDDP was most pronounced in the PM1 cells, and progressively less in MET1 and MET4, indicating that the metastatic cancer cells (MET4) are the most resistant to CDDP-induced cell death. Since CDDP is known to induce apoptotic cell death,9 we investigated whether the decrease in viability in the different cancer cell lines, treated with CDDP, was due to the induction of apoptosis, using independent assays to monitor and quantify apoptosis. Quantification of the amount of cytoplasmic histone-associated DNA fragments (a hallmark for apoptosis) via Cell death detection ELISA, showed a significant decrease in the apoptotic response to CDDP of cSCC cells at more progressive stages of malignancy (Fig. 1b). These data were further complemented in a consistent way by determination of the fraction of cells harboring a sub-G1 DNA amount (data not shown) and the determination of the loss of mitochondrial membrane potential (Supporting Information Fig. S1b) via FACS analysis, and Western blot analysis for PARP cleavage and caspase-3 activation (Fig. 1c). Treatment of these cells with zVAD-fmk, a broad-spectrum caspase inhibitor, blocked the ability of CDDP to induce caspase-3 activation, and PARP cleavage in all cell lines and in parallel inhibited apoptosis, thus indicating that CDDP induced apoptosis required caspase signaling (data not shown). Consistent with the lower level of effector caspase-3, a reduced level of cleaved initiators caspase-8 and -9 was observed in the MET1 and MET4 cells following CDDP treatment (Fig. 1c), thus suggesting that the ability to promptly activate caspase signaling and apoptosis is gradually lost during cSCC progression.

Figure 1.

Effect of cisplatin (CDDP) on viability, apoptosis and AKT activity status. (a) PM1, MET1 and MET4 cells were exposed to 20 μg/ml CDDP, trypsinized after 24, 48 and 72 hr and cell viability was measured using trypan blue exclusion assay. Ctr, control; points, mean of 4 independent experiments; bars, standard deviation. (b) Cells were treated with CDDP for 24 hr and the amount of DNA fragments was determined by the Cell death detection ELISA as described in Material and Methods. Columns, mean of 2 experiments performed in duplicate; bars mean standard deviation. * means p < 0.05. (c) CDDP was supplemented to the different cell lines (PM1, MET1 and MET4). Cells were harvested 24 hr later and protein lysates were separated by SDS-PAGE and analyzed by immunoblotting with antibodies against PARP, cleaved caspase-3, caspase-8 and caspase-9, phospho-AKT (Ser473) and AKT. β-Actin was used as loading control. AKT kinase assay was performed as described in Material and Methods and phosphorylation of GSK-3 was measured by Western blotting, using phospho-GSK-3α/β (Ser21/9) antibody. Densitometric analysis of the Western blots was performed and the amount of pSer473AKT was compared to total AKT protein. Relative amount of pSer473AKT from untreated PM1 cells was set as 1 and relative fold was measured for the other conditions.

AKT activation status correlates to SCC tumor progression and resistance to cisplatin-induced apoptosis

To investigate whether AKT activation may play a role in the resistance to CDDP in the more advanced stages of carcinogenesis, we investigated the Ser473 phosphorylation status by immunoblotting and kinase activity of AKT by monitoring phospho-GSK3α/β as a substrate, in PM1, MET1 and MET4 before and after CDDP treatment (Fig. 1c). The MET1 and MET4 cells showed a higher basal level of phosphorylated AKT, and displayed a higher AKT kinase activity as compared to the PM1 cells. CDDP treatment resulted in a decrease in the phosphorylation status of AKT, and AKT kinase activity relative to untreated cells in all 3 cancer cell lines, but to a greater extent in PM1 than in MET1 and MET4. Loss of total AKT protein in the PM1 cells was observed in parallel with extensive induction of cell death after CDDP treatment (Fig. 1c).

Activation of the PI3-K/AKT signal transduction pathway in human cancer can be ascribed to various mechanisms including stimulation of receptor tyrosine kinases, overexpression of growth factor receptors and abnormalities in PI3-K, PTEN and AKT.11, 12, 27, 28 In our study, PTEN protein levels were not altered as assessed by Western blot (Supporting Information Fig. S2a). In addition, no mutation in the pleckstrin homology domain of AKT could be found as assessed by sequence analysis (Supporting Information Fig. S2b) and confirmed by a mass spectroscopy-based approach using methods designed to detect SNPs, which is more sensitive than conventional Sanger sequencing29, 30 (Supporting Information Fig. S2c). Analysis of PIK3CA “hot spot” mutations sites28 in exon 9 (encodes the PI3-K helical domain) and exon 20 (encodes the PI3-K kinase domain) did not reveal any mutation either (Supporting Information Fig. S2c). This suggests that a downregulation of PTEN protein or mutations in the PI3-K or AKT genes are not likely responsible for hyperactivation of AKT observed during cSCC progression.

Inhibition of AKT sensitizes MET1 and MET4 cells to cisplatin-induced apoptosis

To further examine the involvement of the AKT pathway in the resistance to CDDP, we investigated whether specific inhibition of AKT sensitizes MET1 and MET4 cells to CDDP-induced cell death, using an isozyme-selective AKT1/2 inhibitor (AI).31 The metabolic activity of PM1, MET1 and MET4 cells after AKT inhibition was assessed in the absence or presence of CDDP using MTT assay. AI treatment by itself resulted in a dose-dependent decrease of metabolic activity in PM1 (Fig. 2a), MET1 (Fig. 2b) and MET4 (Fig. 2c) cells. Interestingly, the effect of AKT inhibition was found to be cell type dependent, since cSCC cells harboring high levels of AKT phosphorylation/activation (MET1 and MET4) (Figs. 2b and 2c) were more sensitive to AKT inhibition than the PM1 cells exhibiting a lower phospho-AKT level. Combination of AKT inhibition and CDDP had an additional inhibitory effect on the metabolic activity that was most pronounced in the primary cancer cells (MET1) (Fig. 2b).

Figure 2.

Inhibition of AKT with an isozyme-selective AKT inhibitor reduces metabolic activity of PM1, MET1 and MET4 and sensitizes MET1 and MET4, but not PM1 to CDDP induced-cell death and apoptosis. PM1 (a), MET1 (b) and MET4 (c) cells were treated with different concentrations of AKT inhibitor (AI) 1 hr before the addition of 20 μg/ml CDDP. Metabolic activity was measured by MTT assay. The metabolic activity of the untreated cells was regarded as 100%. Points, mean of 8 data points; bars, standard deviation. * means, p < 0.05 in comparison of metabolic activity in cells with AI alone with that in cells treated with combination of AI and CDDP. (d) The different cell lines were treated 1 hr with 1 μM AI (AI1) or 10 μM AI (AI10) before the addition of CDDP (20 μg/ml). The amount of death cells was assayed 24 hr later by trypan blue exclusion assay. Columns, mean of at least 6 independent experiments; bars denote standard deviation; * means p < 0.05. (e) PM1, MET1 and MET4 cells were treated as in (d) and the amount of DNA fragmentation was determined by the Cell death detection ELISA as described in Material and Methods. Columns, mean of 2 independent experiments performed in duplicate; bars: standard deviation; * means p < 0.05.

Next we analyzed whether the AI could increase the sensitivity of the cells to CDDP-induced cell death. Addition of 1 or 10 μM AI, which efficiently blocked AKT phosphorylation (Fig. 3a), to CDDP-treated PM1 cells did not enhance CDDP-induced cell death (Fig. 2d), nor did it affect other apoptotic parameters such as caspase-3 or PARP cleavage (Fig. 3a) or the amount of cytoplasmic histone-associated DNA fragments (Fig. 2e). In contrast, in CDDP-treated MET1 cells, AKT inhibition resulted in a significant and dose-dependent increase in all tested cell death parameters (Figs. 2d and 2e, Fig. 3a and Supporting Information Fig. S3). The sensitizing effect of AI to CDDP-mediated apoptotic cell death (Figs. 2d and 2e and Fig. 3a) was significantly weaker in MET4 than in MET1 cells.

Figure 3.

Inhibition of AKT and AKT1 knock down increases the sensitivity to CDDP-induced apoptosis. (a) PM1, MET1 and MET4 were pretreated with 1 or 10 μM AI and then treated with CDDP (20 μg/ml). Cell lysates were collected after 24 hr and protein lysates were analyzed by immunoblotting with antibodies against phospho-AKT (Ser473), AKT, PARP and cleaved caspase-3. β-Actin was used to confirm equal loading of proteins. (b) Knock down of AKT1 and AKT2 was performed using siRNA against AKT1 and AKT2 in MET4 cells. MET4 cells were transfected with siRNA and treated as described in Material and Methods. Scrambled siRNA was used as a negative control (-pool). Protein lysates were analyzed by Western blot using antibodies against AKT1, AKT2, PARP, cleaved caspase-3 and actin (loading control). AKT1 and AKT2 levels were measured by densitometric analysis of the Western blots and compared to actin levels. AKT1 and AKT2 levels of scrambled siRNA transfected and untreated MET4 cells were considered as 1.

AI has been shown to inhibit both AKT1 and AKT231 at the used concentrations. To investigate which AKT isoform is responsible for the resistance of MET4 cells against CDDP-induced apoptosis, we reduced specifically the expression levels of AKT1 or AKT2 in the MET4 cells using siRNA targeting these genes (Fig. 3b). Remarkably, compared to the knock down of AKT2, which even appears to reduce apoptotic sensitivity to CDDP (Fig. 3b and Supporting Information Fig. S4a), only the knock down of AKT1 resulted in an evident increase in apoptotic markers, such as processing of caspase-3 and PARP (Fig. 3b), and apoptotic cell death (Supporting Information Fig. S4b) after CDDP treatment. Although knock down of AKT2 also had a partial effect on the level of AKT1, clearly it did not contribute to a higher sensitivity of the cells to CDDP (Fig. 3b and Supporting Information Fig. S4a). In addition, this partial decrease in AKT1 level after AKT2 silencing shows that only a robust AKT1 knock down is required to increase the sensitivity of the MET4 cells to CDDP. These results indicate that the increased sensitivity to CDDP-induced apoptosis by AI is due to the specific inhibition of AKT1 and not AKT2.

Autophagy stimulation by AKT inhibition is involved in chemoresistance against CDDP-mediated cell death

AKT is not only known to be an inhibitor of apoptosis, but also of autophagy.32 Therefore, we investigated whether AKT inhibition also resulted in the induction of autophagy in the different cell lines. AKT inhibition in PM1, MET1 and MET4 cells (Supporting Information Fig. S5) resulted in the induction of autophagy as assessed by detection of endogenous LC3 lipidation33 by Western blot analysis, which correlates well with the number of autophagosomes present in the cell. Induction of autophagy in all 3 cell lines also paralleled the inhibition of the mTOR/p70S6K pathway downstream of AKT, as assessed by the phosphorylation status of p70S6K (Supporting Information Fig. S5). Since autophagy is thought to be an important survival mechanism correlated to the problem of chemoresistance, we further investigated the role of autophagy specifically in the metastatic cell line MET4, with its inherent problem of chemoresistance. Electron microscopy analysis of MET4 cells showed that autophagic vacuoles were not clearly evident in untreated cells (Fig. 4a, detail a′) and only weakly present after CDDP treatment (Fig. 4b, detail b′). Following inhibition of AKT, the number of MET4 cells displaying autophagic vacuoles in their cytoplasm greatly increased (Fig. 4c, detail c′). The cotreatment of CDDP and AI resulted in a population of cells containing many autophagic vacuoles with electron dense material together with ultrastructural features of apoptosis. In addition, the MET4 cells displayed dilated rough endoplasmatic reticulum (RER) cisternae, suggesting CDDP-mediated injury to this organelle (Fig. 4d, detail d′).

Figure 4.

Induction of autophagy upon AKT inhibition in MET4 cells. (ad) Electron micrographs showing the ultrastructure of MET4 cells with and without treatment. (a) Control MET4 cells. (a′) Detail of cytoplasmic organelles; the rough endoplasmic reticulum (RER) is slightly dilated (arrowhead) and autophagic vacuoles are lacking. (b) MET4 cells treated with 20 μg/ml CDDP for 24 hr. A moderate number of cells contain autophagic vacuoles (arrow). The other cells have the same appearance as the control cells. (b′) Detail of the cytoplasm containing several autophagic vacuoles (arrow). (c) MET4 cells treated with 10 μM AI (AI10) for 24 hr. A high number of cells contain numerous autophagic vacuoles (arrow). (c′) Detail of the cytoplasm containing many autophagic vacuoles (arrow). (d) MET4 cells treated with CDDP (20 μg/ml) and AI (10 μM) for 24 hr. A high number of cells contain autophagic vacuoles (arrow). An apoptotic cell (big arrow) is obvious with some remnants of autophagic vacuoles in the cytoplasm (open arrow). (d′) Detail of the cytoplasm containing autophagic vacuoles (arrow) and dilated RER cisternae (arrowhead). Scale bar in (a, b, c, d): 2.50 μm; scale bar in (a′, b′, c′, d′): 1 μm. (eh) MET4 cells were transfected with a plasmid expressing tfLC3 as described in Material and Methods and the cells were untreated (e), treated with 20 μg/ml CDDP (f), 10 μM AI (g) or a combination of both CDDP and AI (h) for 24 hr, fixed and analyzed by the IN Cell Analyzer 1000.

An increase in autophagosome accumulation may result either from the upregulation of autophagosome formation or from the blockage of autophagic degradation.33 To distinguish between these 2 processes, we transiently transfected MET4 cells with an mRFP-GFP-LC3 construct (tf-LC3), a novel marker that allows assessing autophagic flux, that is, the complete processing of autophagosomes after fusing with lysosomes, using fluorescence microscopy. Using this procedure autophagosomes are imaged by the colocalization of mRFP and GFP signals, whereas autophagolysosomes show attenuated GFP signals and retain mRFP signals alone since mRFP fluorescence is relatively stable in acidic conditions and not degraded by lysosomal hydrolases.34, 35 Untreated (Fig. 4e) and CDDP-treated MET4 cells (Fig. 4f) showed only few colocalizing GFP and mRFP puncta. In contrast, following treatment of MET4 cells with AI in the absence (Fig. 4g) or presence of CDDP (Fig. 4h), mRFP puncta clearly accumulated indicating that the fusion of autophagosomes with lysosomes had taken place. These results demonstrate that the autophagic maturation process is completed in cells where AKT is inhibited with or without CDDP treatment.

To further clarify the role of massive autophagy in conditions of AKT inhibition in terms of cell death versus survival, we inhibited pharmacologically autophagy with 3-MA. We detected endogenous LC3 lipidation by Western blot analysis,36 and the quantification of AVOs by flow cytometry. MET4 cells treated with AI in the absence or presence of CDDP, demonstrated a clear accumulation of the faster migrating lipidated form of LC3 (LC3-II) (Fig. 5a), and a substantial increase in cells showing AVO formation (Fig. 5b). Inhibition of autophagy by 3-MA, as demonstrated by a reduction of LC3-II accumulation (Fig. 5a) and a significant inhibition of AVO formation (Fig. 5b), further increased CDDP-mediated apoptosis in the presence of AI (Fig. 5c) or after siRNA-mediated AKT1 knock down (Supporting Information Fig. S4b). Knock down of ATG5 further validated the survival function of autophagy induced upon AKT inhibition combined with CDDP (Fig. 5d). In conformity with these data, we also showed that pharmacological inhibition of autophagy upon combined treatment with CDDP and AI increased apoptosis in the MET1 cells, as evidenced by an increase in active caspase-3 (Supporting Information Fig. S6a) and DNA fragmentation (Supporting Information Fig. S6b). Moreover, autophagy inhibition also reduced the metabolic activity of another cSCC cell line (SCC12B2) (Supporting Information Fig. S7). All together these results suggest that autophagy induced by AI acts as a prosurvival process in skin cancer that provides these cells with an increased ability to cope with CDDP-mediated cellular damage.

Figure 5.

Inhibition of AI-induced autophagy with 3-MA or specific ATG5 knock down results in further increased CDDP-induced cell death in MET4 cells. (ac) MET4 cells were treated with CDDP (20 μg/ml), 10 μM AI (AI10) or 10 mM 3-MA alone or with a combination of CDDP and AI with or without 3-MA. 3-MA and AI10 were added 1 hr before the addition of CDDP. (a) Cells were harvested 24 hr later and proteins were analyzed for the accumulation of LC3-II by western blot. β-Actin was used as a loading control. Relative fold means amount of LC3-II compared to β-actin and Ctr condition was regarded as 1. (b) MET4 cells were trypsinized and stained after 24 hr with acridine orange (1 μg/ml) for 15 min and analyzed for the presence of AVOs by flow cytometric analysis. Results shown are mean of 2 measurements from a representative of 2 independent experiments. Bars mean standard deviation; * means p < 0.05. (c) After 24 hr, the amount of DNA fragments was determined by the Cell death detection ELISA as described in Material and Methods. Results shown are mean of 2 measurements from a representative of 2 independent experiments. Bars mean standard deviation; * means p < 0.05. (d) MET4 cells were transfected with scrambled siRNA (-pool) or siRNA against ATG5 as described in Material and Methods and treated with 20 μg/ml CDDP and 10 μM AI (AI10). The amount of DNA fragments was determined after 24 hr by the Cell death detection ELISA. Results shown are a representative of 2 independent experiments done in duplicate. Bars mean standard deviation.


One of the major goals in cancer research is the identification of new strategies for the treatment of advanced cancer such as metastatic SCC of the skin. Certain tumor and patient characteristics, including chronic immunosuppression predispose to advanced SCC of the skin, with poor prognosis due to a high recurrence rate after surgery and radiotherapy and resistance to chemotherapy, including treatment with the genotoxic agent cisplatin. Identification of new strategies that overcome chemoresistance and induce selective cancer cell death is of utmost importance to reduce mortality rate of advanced cancer. This requires further insight into the molecular processes leading to cell death.

The underlying mechanisms of chemoresistance or the decreased capacity of a chemotherapeutic drug to kill tumor cells is, at least partly, due to the progressive acquisition of multiple genetic changes downregulating proapoptotic pathways and activating survival pathways, such as the AKT/PKB pathway. In this study, we investigated the role of PI3-K/AKT signaling in the cell death/survival response to cisplatin in a unique series of epidermal cell lines representing different stages of malignant transformation.24 Our results reveal an increasing innate resistance to CDDP-induced cell death in the more malignant stages of epidermal transformation corresponding with the increased activation status of AKT. The dysplastic PM1 cells remained highly sensitive to cisplatin-induced cell death and apoptosis, even more sensitive than normal human keratinocytes (Supporting Information Fig. S8), consistent with general clinical experience that topical application of chemotherapeutic agents, such as 5-fluorouracil, selectively induces cell death of the dysplastic keratinocytes (actinic keratoses and SCC in situ), while saving normal surrounding epidermis. Increased activation status of AKT, paralleled by augmented resistance to cisplatin-induced cell death in MET1 and MET4, is consistent with earlier described correlations between AKT activity and advanced disease and/or poor prognosis in some tumor types27, 37 and chemoresistance.38

Levels of phosphorylated AKT and its kinase activity upon CDDP treatment decrease in PM1 cells and to a lesser extent in the MET1 and MET4 cells. This is consistent with other reports showing that certain compounds, including cisplatin, can reduce AKT signaling in cell lines.39–42 Reduced AKT signaling can be the result of a negative interaction between the PI3-K/AKT pathway with other kinase-regulated signals, including p38 MAPK,39 PKC40 or casein kinase I epsilon (CKIε),41 as reported for other apoptotic paradigms. Another possible explanation is that AKT may be inhibited by the cisplatin-induced activation of specific AKT inhibitors such as MST1.42 Moreover, elevated phosphorylation levels of AKT, like in MET1 and MET4, have been recently reported to induce resistance to cleavage at the caspase-3 site within the kinase domain.43 This raises the intriguing possibility that in the process of skin carcinogenesis (e.g., at the MET1 stage) the higher phosphorylation status of AKT, renders this kinase refractory to undergo caspase-mediated downregulation, thereby sustaining its prosurvival function.

Targeting the AKT pathway alone or in combination with other chemotherapeutic drugs, might be a promising strategy to improve therapeutic outcome since many human cancers show an aberrant activation of the PI3-K/AKT pathway resulting from different mechanisms including stimulation of receptor tyrosine kinases, overexpression of growth factor receptors and abnormalities in PI3-K, PTEN and AKT.11, 12, 27, 44 Because of the limited specificity of the widely used PI3-K inhibitor for class I PI3-K, wortmannin, we preferred to target directly AKT by using an AKT inhibitor (AI) that is a cell-permeable, potent and selective inhibitor by binding to the pleckstrin homology domain of both AKT1 and AKT2. AI has been reported to selectively induce apoptosis in tumor cells but not in normal cells.45 In our experimental setup, AI supplementation results in a significant increase in apoptotic cell death in CDDP-treated primary tumor cells (MET1) and, to a lesser extent, in the metastatic MET4 cells. Inhibition of AKT does not increase the amount of cisplatin-induced cell death in the dysplastic PM1 cells, which are already highly sensitive to cisplatin and have a low amount of phosphorylated and active AKT. Both AKT1 and AKT2 have been described to have a role in the resistance to cisplatin.38, 46, 47 We could show that a downregulation of the expression of AKT1 by a specific siRNA approach was able to sensitize MET4 cells to CDDP-induced apoptosis, pointing to a specific role of AKT1 in the resistance of metastatic skin cancer cells to CDDP-induced apoptosis. In contrast, knock down of AKT2 appears to decrease apoptotic sensitivity to CDDP, pointing toward distinct roles of AKT1 and AKT2 in regulating apoptosis, which needs to be further investigated in future studies.

The protein kinase AKT is not only an inhibitor of apoptosis, but also of autophagy, which is a highly conserved catabolic program for the degradation and recycling of cellular constituents such as long-lived proteins and organelles. This process is linked to both health (e.g., cellular stress, development, differentiation) and disease (e.g., infectious diseases, neurodegenerative disorders, cancer, etc).21 Its role in cancer development48 and in the response to cancer treatment49 is still a point of debate requiring further investigation. Although autophagy induction by conditions of metabolic stress, such as starvation, has in most instance a prosurvival function, autophagy may be turned into a mechanism of cell death (autophagic cell death) in response to severe cellular stress, as for example, in response to chemotherapy.50 Thus, the function of autophagy is dependent on the nature of the stress stimulus, on the type of cell and the underlying genetic defects leading to tumor development. In our study, inhibition of AKT does not only result in the induction of apoptosis, but it also promotes autophagy in the metastatic cells as demonstrated by accumulation of the autophagic protein marker LC3-II and a substantial increase in AVOs formation. Electron microscopy data of metastatic cells show massive accumulation of autophagic vacuoles in cells treated with the AKT inhibitor. Accumulation of these autophagic vacuoles in MET4 cells in the presence of AI resulted from autophagic flux and was not the result of a blockade in or a deficiency in the autophagic process, since we could demonstrate that in these conditions the autophagic maturation process was completed. PM1 and MET1 cells also revealed the accumulation of the faster migrating lipidated form of LC3 (LC3-II) in the presence of the AKT inhibitor (Supporting Information Fig. S5), however the accumulation was most pronounced in MET4. Concomitant with a decrease in AKT phosphorylation, we also observed a reduction in phosphorylated p70S6K (Supporting Information Fig. S5), a downstream target of the mTOR signaling pathway, thus suggesting that AI effectively inhibits the AKT-mTOR/p70S6K/S6K pathway, which is known to downregulate autophagy.17 Autophagy upon AKT inhibition could be induced in another cutaneous cancer cell line, SCC12B2, demonstrating that this effect was not restricted to the isogenic PM1, MET1 and MET4 cancer cells.

In recent years, targeting the PI3-K/AKT/mTOR pathway has emerged as an attractive opportunity for cancer therapy and several drugs are in clinical testing. Among agents targeting this axis, mTOR inhibitors are furthest in development. Also several lipid based and peptide based AKT inhibitors are being evaluated mostly preclinically, while Perifosine is the best characterized AKT inhibitor in human testing so far.51 Numerous preclinical reports, including studies in nonsmall cell lung cancer cells,38 melanoma cells52 and oral SCC53 have documented that inhibition of AKT plays an important role in the sensitization to various proapoptotic insults. In our study AKT inhibition also sensitized MET1 and to a lesser extent MET4 to CDDP apoptosis, but not to the level as was observed in the PM1 cells. However, our results indicate that AKT inhibition can concomitantly stimulate autophagy both on its own as well as in combination with cisplatin, especially in metastatic skin cancer cells harboring elevated basal levels of AKT activity. To realize a better patient outcome after cancer treatment, it is important to know whether autophagy induction contributes to survival or death of the tumor cells.50 We provide evidence that 3-MA and ATG5 knock down mediated attenuation of AI-induced autophagy in MET4 cells increases significantly cisplatin-induced apoptosis, which argues for a protective role of the autophagic response in metastatic cancer cells. This prosurvival role for autophagy was also observed in the MET1 cells, derived from the primary cSCC and another cutaneous cancer cell line SCC12B2, indicating a general role of autophagy as a survival strategy in skin cancer cells responding to this chemotherapeutic. This is an important observation that supports the combined inhibition of autophagy and AKT as new molecular strategy to potentiate cisplatin-mediated killing of metastatic skin cancer. This is consistent with the report that AKT inhibition by siRNA or small molecule inhibitors promotes autophagy and sensitizes tumor cells to autophagy inhibitors.32

Taken altogether, our results (summarized schematically in Fig. 6) indicate that sensitization of cancer cells to cisplatin-induced cell death is not only achieved by the induction of apoptosis after AKT inhibition, but can even be increased when the AKT inhibitor is combined with an inhibitor of autophagy. From a therapeutic standpoint, these observations advocate the balanced use of AKT inhibitors to sensitize cancer cells to therapeutic approaches using classical anticancer agents, such as cisplatin, in combination with blockers of autophagy.

Figure 6.

A model for the role of AKT in the chemoresistance of progressive stages of cutaneous SCC to cisplatin. Increased skin tumor progression parallels increased resistance to apoptotic cell death upon CDDP treatment and increased activation status of AKT. Concomitant with the role of AKT in inhibiting apoptosis, but also its negative regulating effect on autophagy, via regulation of mTOR signaling, inhibition of AKT with the pharmacological inhibitor AI increases the apoptotic response of the MET1 and MET4 cells to CDDP, but also induces massive autophagy in MET1 and MET4. Pharmacological inhibition of autophagy using 3-MA further increases apoptotic sensitivity to CDDP in MET1 and MET4.


OT/04/42/BOF (Sofie Claerhout), SB/61432/IWT (Lien Verschooten) and OT/06/49 (Patrizia Agostinis). Funding as an Odyssey Fellow (Sofie Claerhout) was supported by the Odyssey Program and the Theodore N. Law Endowment for Scientific Achievement at The University of Texas M.D. Anderson Cancer Center. This research is supported by the Stichting tegen Kanker 188-2008 to Patrizia Agostinis and by F.W.O grant G.0491.05. This paper presents research results of the IAP6/18, funded by the Interuniversity Attraction Poles Program, initiated by the Belgian State, Science Policy Office. The scientific responsibility rests with its author(s).

We thank Dr. W. Bossuyt for the help with the sequence analysis of AKT and PI3-K and for critical reading of the manuscript and Dr. Tamotsu Yoshimori (Osaka University) for the tfLC3 construct. We also thank the Developmental Studies Hybridoma Bank for the monoclonal actin antibody. STR DNA fingerprinting was done by the Cancer Center Support grant funded Characterized Cell Line core, NCI # CA16672.