Metformin Amplifies Chemotherapy-Induced AMPK Activation and Antitumoral Growth Free

This article is featured in Highlights of This Issue, p. 3853

Metformin is an oral hypoglycemiant agent used as first-line therapy for type 2 diabetes, which is now prescribed to almost 120 million people in the world. There are a large number of epidemiologic studies indicating that diabetics have an increased risk of cancer and cancer mortality (1, 2). Increasing evidence also supports a decreased risk of cancer mortality associated with metformin use in patients with type 2 diabetes (3–6). Furthermore, metformin has been shown to inhibit the growth of cancer cells in vitro and in vivo (7–12) and, while there are still no randomized control trials of metformin as a therapy for cancer, there is intriguing evidence that metformin may enhance chemotherapy for established tumors (13, 14).

Metformin has been found to activate AMP-activated protein kinase (AMPK) signaling (15), and this has become an important focus of interest in carcinogenesis, because AMPK has been implicated in the regulation of mTOR activity, which is frequently activated in cancer (16–20). AMPK is the downstream component of the tumor suppressor, LKB1, which acts as a sensor of cellular energy charge, being activated by increasing AMP, coupled with falling ATP (21). The AMP/LKB1-dependent activation of AMPK results from pathologic stresses such as heat shock, hypoxia, glucose deprivation, and metformin administration (15, 21). AMPK is also activated through Ca⁺²/calmodulin (CaM)-dependent protein kinase kinase (CaMKK), which in contrast to that mediated by AMP/LKB1, is mediated by calcium increases and functions independently of AMP (22, 23). Once activated, AMPK phosphorylates acetyl-CoA carboxylase (ACC) and switches on energy-producing pathways at the expense of energy-depleting processes (24).

Another direct consequence of AMPK activation is the inhibition of the mTOR kinase signaling pathway. mTOR catalytic activity is halted by AMPK activation of the TSC1–TSC2 complex, which inactivates the Rheb GTPase (25, 26). In addition, mTOR activity is positively regulated by growth factors and nutrients (amino acids). PI3K/Akt signaling regulates mTOR through phosphorylation/inactivation of mTOR's negative regulator, TSC2 (17, 27). mTOR activation results in the phosphorylation of the serine/threonine kinase p70S6K and the translational repressor eukaryotic initiation factor (eIF) 4E binding protein (4E-BP1), which have an essential role in regulating cell growth and proliferation by controlling mRNA translation and ribosome biogenesis (28).

To achieve normal cell growth and proliferation, it is critical for cells to have robust antigrowth signaling systems. AMPK has a major role as an antigrowth signal, because it is activated by p53, a sensor of DNA damage stress (29). Recently, the genotoxic stress effect was further evaluated and it has been suggested that the inhibition of mTOR activity occurs through the p53-dependent upregulation of sestrins (SESN1 and SESN2) and consequent activation of AMPK (30). These observations indicate that metformin acts synergistically with chemotherapeutic drugs that increase genotoxic stress through a convergent signaling of metformin-mediated LKB-1/AMPK activation and chemotherapeutic drug activation of SESNs, culminating in an increased AMPK activation and mTOR inhibition. Thus, this study was designed to investigate whether metformin potentiates paclitaxel antitumor effects, a well-known chemotherapeutic drug frequently used in breast and lung cancer patients (31, 32), and to observe whether these drugs share common intracellular signal transduction pathways and to determine whether these signaling systems modulate each other's actions in different cancer cell lineages and in xenografted tumor cells in mice.

All the reagents were from Sigma-Aldrich unless otherwise specified. Paclitaxel was obtained from Laboratório Químico Farmacêutico Bergamo Ltda. Anti–phospho-mTOR, anti-mTOR, anti–phospho-p70S6K, anti-p70S6K, anti–phospho-4E-BP1, anti–4E-BP1, anti–phospho-AMPKα, anti-AMPKα, anti–β-actin, anti–acetyl-lys379-p53, anti–phospho-p53, anti–phospho-ACC, anti–caspase 3, anti–cleaved caspase 3, anti-p27, and anti–phospho-Rb antibodies for immunoblotting were from Cell Signaling Technology; anti-p53 and anti-SESN2 antibodies for immunoblotting were from Santa Cruz Biotechnology; and anti-SESN1 and anti-SESN3 antibodies for immunoblotting were from Abcam.

The human breast cancer cell line MCF-7 (LKB1-positive) and human lung cancer cell line A549 (LKB1-negative) were obtained from ATCC. MCF-7 and A549 cells were cultured in Dulbecco's Modified Eagle's Medium containing 10% FBS with the addition of antibiotics or fungicides. Both cell lines were maintained at 37°C in a humid atmosphere and 5% CO₂.

A total of 3 × 10⁵ cells were seeded in a tissue culture plate in complete growth medium and incubated overnight. On the day of transfection, 200 pmol of siRNA was diluted into OPTI-MEM (Life Technologies) and mixed with 10 μL of Lipofectamine 2000 (Life Technologies) according to supplier's protocol. The transfection medium was then replaced by complete medium and after 24 hours cells were treated with metformin (10 mmol/L) and paclitaxel (1 μmol/L) and incubated for an additional 6 hours. siRNA for AMPK was 5′-AAUUACUUCUGGUGCAGCAUAGCGG-3′ forward and 5′-CCGCUAUGCUGCACCAGAAGUAAUU-3′ reverse; for SESN1, 5′-GAACCUUCUCAGAGUGCUUGAACUG-3′ forward and 5′-CAGUUCAAGCACUCUGAGAAGGUUC-3′ reverse; and for SESN2, 5′- GGAUAGCGAGUAGCCAUGGUCUUCC-3′ forward and 5′-GGAAGACCAGGGCUACUCGCUAUCC-3′ reverse.

Cells were seeded at a density of 2 × 10⁴ cells/well in 24-well plates containing 1 mL of complete medium in triplicate. Cells were allowed to attach overnight before treating with the indicated dose of metformin and paclitaxel for 24 hours. Subsequently, viable cells were counted by using trypan blue staining or they were treated with 0.3 mg/mL of MTT for 4 hours and MTT-formazan conversion was analyzed by spectrophotometry at 570 nm after culture medium was removed and ethanol was added.

Cells were trypsinized, washed in PBS, centrifuged, and pellets were fixed in 200 μL of 70% ethanol and stored at −20°C until use. Cells were centrifuged and pellets resuspended in 200 μL of PBS, and 10 μg/mL of RNAse A was incubated for 1 hour at 37°C. Subsequently, cells were resuspended in propidium iodide solution (0.1% sodium citrate, 0.1% Triton X-100, and 50 μg/mL propidium iodide). Cell-cycle analysis was carried out by flow cytometry (FACScalibur). Data were analyzed by ModFit LT software.

Measurement of oxygen consumption by MCF-7 and A549 after treatment for 24 hours with metformin (10 mmol/L), paclitaxel (1 μ/mol/L), or the combination of both drugs was carried out by a Oxygraph equipped with a Clark-type electrode (Hansatech Instruments Limited) in a closed chamber equipped with magnetic stirrer and temperature control at 37°C. Approximately 2.5 × 10⁶ of viable MCF-7 cells/mL, and 4 × 10⁶ of viable A549 cells/mL, permeable with 10 μmol/L of digitonin, were added in 2 mL of reaction medium containing 125 mmol/L sucrose, 65 mmol/L KCl, 10 mmol/L HEPES, 2.0 mmol/L K2HPO4, 1.0 mmol/L MgCl2 (pH 7.2), 50 μmol/L EGTA, and complex I substrates (2.0 mmol/L malate, 1.0 mmol/L α-ketoglutarate, 1.0 mmol/L pyruvate, and 1.0 mmol/L glutamate). Analyses of oxidative phosphorylation and respiratory activity of mitochondria were made by sequential additions of 100 μmol/L ADP, 2 μg/mL chloramphenicol acetyltransferase, 100 nmol/L carbonylcyanide p-trifluoromethoxyphenylhydrazone, 5 mmol/L succinate, 0.5 μmol/L antimycin, and 200 μmol/L N, N, N′, N′-tetramethyl-p-phenylenediamine/ascorbate. The data were reproduced and calculated by the device's specific software.

Four-week-old male severe combined immunodeficient (SCID) mice were provided by the State University of Campinas—Central Breeding Center. Animals were inoculated in the dorsal region, subcutaneously, with 1 × 10⁶ A549 cells. The mice had ad libitum access to food and water. Once tumors became palpable, tumor volume (V) was calculated daily by measuring the length (L) and width (W) of the tumor with calipers by using the formula: V = [W × L × (W + L)/2)] × 0.52. Each group contained 15 animals.

Treatments began when tumors reached 50 to 100 mm³. Metformin was given daily by gavage at 500 mg/kg body weight. Paclitaxel was given once a week by intraperitoneal (i.p.) injection of 10 mg/kg body weight. All experiments were approved by the Ethics Committee of the State University of Campinas.

Mice were anesthetized with sodium amobarbital (15 mg/kg body weight, i.p.). Tumors were removed, minced coarsely and homogenized in extraction buffer [1% Triton-X 100, 100 mmol/L Tris (pH 7.4), containing 100 mmol/L sodium pyrophosphate, 100 mmol/L sodium fluoride, 10 mmol/L EDTA, 10 mmol/L sodium vanadate, 2 mmol/L PMSF and 0.1 mg of aprotinin/mL]. The extracts were centrifuged at 11,000 rpm and 4°C and the supernatants of these tissues were used.

Whole tissue extracts and cell pellets were homogenized in extraction buffer, treated with Laemmli sample buffer containing 100 mmol/L DTT and heated in a boiling water bath. For total extracts, similar-sized aliquots (50 μg protein) were subjected to SDS-PAGE. Proteins were resolved on 8% to 15% SDS gels and blotted onto nitrocellulose membranes (Bio-Rad). Band intensities were quantified by optical densitometry of developed autoradiographs by using Scion Image software (Scion Corporation).

To detect Ki-67 and cleaved caspase 3, microwave postfixation was carried out by a domestic oven which was delivered to slides immersed in 0.01 mol/L citrate buffer, pH 6.0, in two 7-minute doses separated by a 2-minute break. Sections were then incubated at 4°C overnight with primary monoclonal mouse anti-human Ki-67 clone MIB-1 from Dako (diluted 1:100) and anti-cleaved caspase 3 from Cell Signaling Technology. The slides were then incubated with avidin–biotin complex LSAB+ Kit from Dako Cytomation for 30 minutes followed by the addition of diaminobenzidine tetrahydrochloride as a substrate-chromogen solution. After hematoxylin counterstaining and dehydration, the slides were mounted in Entellan from Merck.

Terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) staining was done by a commercial apoptosis detection kit (Roche), according to the recommendations of the manufacturer. Analysis and documentation of results were carried out by a Leica FW 4500 B microscope.

Data are presented as mean ± SEM of at least 3 independent experiments. All groups were studied in parallel and differences between groups were analyzed by ANOVA, as appropriate, and Bonferroni post hoc tests for multiple unpairwise comparisons of means. The level of significance adopted was P < 0.05.

To examine the effect of metformin on cancer cell growth, MCF-7 breast cancer and A549 lung cancer cell lines were treated with various concentrations of metformin (1–50 mmol/L) for different periods of time (0–8 hours). Metformin treatment resulted in the activation of AMPK, with increased phosphorylation of AMPKα at Thr-172 in a time- and dose-dependent manner. Activation of AMPK is associated with decreased activation of mTOR and p70S6K, a critical translational pathway for protein synthesis (10). Metformin treatment resulted in attenuated activation of mTOR, as shown by the decreased phosphorylation of mTOR, p70S6K, and 4E-BP1, in a time- and dose-dependent manner in treated cancer cells, compared with untreated cells (Fig. 1A–D).

We also observed the effect of 2-deoxy-d-glucose (2-DG), another AMPK activator, in both cell lines at various concentrations, and for different periods of time. As observed for the metformin treatment, 2-DG led to activation of AMPK and inactivation of the mTOR signaling pathway (Supplementary Fig. S1A–D).

A549 lung cancer cell line is negative for LKB1 and recent reports have shown that LKB1 deficiency in hepatocytes impairs metformin action (33). However, in this study we show that AMPK activation in A549 cells after metformin treatment is independent of LKB1 and is dose-responsive, starting from 1 to 100 μmol/L with ACC phosphorylation and mTOR and p70S6K inactivation (Supplementary Fig. S2). The same pattern of AMPK and ACC activation and mTOR and p70S6K inactivation is seen on MCF-7 cell line, which is LKB1 normal (Supplementary Fig. S2).

To investigate the mechanisms underlying the antiproliferative effects of paclitaxel, we characterized the effects of paclitaxel on AMPK and the mTOR pathway. As recently reported, genotoxic stress increases the amount of SESNs, and this effect leads to AMPK activation (30). Our results show that paclitaxel treatment increased the acetylation of p53 at Lys-379, the phosphorylation of p53 at Ser 15, which are known markers of genotoxic stress (34, 35), and the amount of SESN1, SESN2, and SESN3 in a time- and dose-dependent manner in both cell lines. Paclitaxel treatment also resulted in increased phosphorylation of AMPKα at Thr-172, in a time- and dose-dependent manner. The increased activation of AMPK led to inactivation of mTOR as evidenced by diminished phosphorylation of mTOR, p70S6K, and 4E-BP1 also in a time- and dose-dependent manner (Fig. 1E–H).

We next sought to determine the effects of the combined treatment of AMPK activators with paclitaxel. In MCF-7 cells, as shown in Figure 2A, paclitaxel treatment led to a higher increase in acetyl-Lys 379 p53 than metformin or 2-DG treatment, as well as an increase in the amount of SESN2 (Fig. 2B). This increase in SESN2 in paclitaxel-treated cells was followed by an increase in the phosphorylation of Thr-172 of AMPK (Fig. 2C) and inhibition of mTOR (Fig. 2D), p70S6K (Fig. 2E), and 4E-BP1 (Fig. 2F), when compared with metformin or 2-DG treatments alone. Even though metformin or 2-DG treatment do not increase SESN2 (Fig. 2B), we observe an increase in AMPK phosphorylation (Fig. 2C) and inhibition of mTOR (Fig. 2D), p70S6K (Fig. 2E), and 4E-BP1 (Fig. 2F), when compared with vehicle-treated cells. In A549 cells, we also observed an increase in acetyl-Lys379 p-53 and in the amount of SESN2 only in the paclitaxel-treated cells (Fig. 2G and H), when compared with control and metformin and 2-DG alone, and this was correlated with an increase in the Thr-172 phosphorylation of AMPK and decrease in the activation of mTOR, p70S6K, and 4E-BP1 when compared with control. Once again, metformin and 2-DG treatments did not increase SESN2 (Fig. 2H) but they were able to increase AMPK phosphorylation (Fig. 2I) and inhibit mTOR (Fig. 2J), p70S6K (Fig. 2K), and 4E-BP1 (Fig. 2L), when compared with vehicle-treated cells.

To examine the effects of metformin and paclitaxel on cancer cell growth, we treated MCF-7 and A549 cell lines with metformin or paclitaxel alone or in combination and cell viability was determined. As shown in Figure 3A and B, both metformin and paclitaxel inhibited cell viability, as related to vehicle-treated cells. The metformin treatment was statistically significant at 10 mmol/L for the 48- and 72-hour treatment (Fig. 3A) in MCF-7 cells and that a 1 mmol/L dose of metformin was capable of reducing A549 cells proliferation by approximately 20% to 30% for both the 48- and 72-hour treatments. The paclitaxel treatment was effective only at the dose of 10 nmol/L for both 48- and 72-hour treatments (Fig. 3B). Figure 3C and D shows that in the combined treatment metformin potentiates paclitaxel action on MCF-7 (Fig. 3C) and A549 (Fig. 3D) cells, as a dose of 1 nmol/L of paclitaxel is statistically different from the metformin and vehicle-treated cells for both the 48- and 72-hour treatments. At 10 nmol/L of paclitaxel, metformin does not further potentiate paclitaxel treatment on MCF-7 cells. However, in A549 cells, we observed that at the 48-hour treatment metformin (10 mmol/L) and 10 nmol/L paclitaxel is more effective on reducing cell growth than paclitaxel-alone treatment (Fig. 3D). We then analyzed cell viability by trypan blue staining of both cell lines (Fig. 3E), which showed that both metformin and paclitaxel inhibited cell viability, as related to vehicle-treated cells, and the combined treatment was more effective than either treatment alone.

To evaluate the mechanism of growth inhibition by metformin, 2-DG, and paclitaxel, the cell-cycle profile was analyzed by flow cytometry after treatment with metformin, 2-DG, or paclitaxel alone, or the combination of the drugs. Vehicle treatment presented the majority of cells in the G₁ phase of the cell cycle (MCF-7, 68.8%; A549, 71.2%), a small part in the G₂–M phase (MCF-7, 12.4%; A549 8.6%) and the rest of the cells were found to be in the S-phase (MCF-7 18.8%, A549, 20.2%). 2-DG treatment resulted in a slight increase in cells in G₁ phase arrest (MCF-7, 72.1%; A549, 77.4%), with a decrease in S-phase (MCF-7, 15.3%; A549, 15.4%) and no significant alteration in G₂–M phase (MCF-7, 12.6%; A549, 7.2%). Metformin treatment resulted in an increase in the number of cells in the G₁ phase (MCF-7, 80.6%; A549, 81.2%) with almost similar number of cells in the G₂–M phase of MCF-7 cells (MCF-7, 9.5%; A549, 8.7%) and a reduction in the number of cells in the S-phase in both cell lines (MCF-7, 9.9%; A549, 10.2%). Metformin combined with 2-DG (MET + 2-DG) treatment resulted in a decrease in the number of cell in G₁ phase (MCF-7, 53.5%; A549, 52.1%), and increase in cells in the G₂–M phase arrest (MCF-7, 33%; A549, 35.3%) and a decrease in cells in S-phase (MCF-7, 13.4%; A549, 12.6%) as related to vehicle-treated cells. Paclitaxel treatment, as expected, caused an increase in the number of cells in the G₂–M phase (MCF-7, 21.5%; A549, 18.8%) with a reduction in the number of cells in the G₁ phase (MCF-7, 64%; A549, 64.7%) and in the S-Phase (MCF-7, 14.5%; A549, 16.5%) in both cell lines (Fig. 3F).

The combined treatment of 2-DG and paclitaxel, of metformin and paclitaxel and the triple therapy resulted in a synergistic effect of G₂–M cell-cycle arrest. Cells treated with 2-DG and paclitaxel showed a reduction in G₁ phase (MCF-7, 57.1%; A549, 60.4%), there was no significant alteration in the number of cells in S-phase (MCF-7, 15.1%; A549, 16.9%); however, there was an increase in G₂–M phase (MCF-7, 27.8%; A549, 22.7%) compared with either treatment alone. Cells treated with metformin and paclitaxel showed a reduction in G₁ phase arrest, compared with either treatment alone (MCF-7, 54.6%; A549, 45.5%). This treatment resulted in no significant alteration in the number of cells in the S-phase compared with either treatment alone (MCF-7, 13.5%; A549, 12.8%). Additionally, when we analyzed the G₂–M phase, we observed a significant increase in the number of cells in this phase in the metformin combined with paclitaxel treatment, as compared with either treatment alone (MCF-7, 31.9%; A549, 41.8%). Finally, the triple therapy resulted in a decrease in the number of cells in G₁ phase as compared with MET + 2-DG treatment or paclitaxel treatment alone (MCF-7, 40.9%; A549, 39.7%), a slight decrease in the number of cells in S-phase (MCF-7, 10.9%; A549, 12.1%) and a significant increase in the number of cells in G₂–M phase (MCF-7, 48.1%; A549, 48.2%). Thus, Figure 3F shows an increase in cell-cycle arrest in the G₂–M phase, during the combined treatment of MET+2-DG, 2-DG and paclitaxel, metformin and paclitaxel, and the triple therapy and a decrease in the G₁ phase, indicating that cells submitted to these combined treatments were not further undergoing division.

We then examined the protein levels of caspase 3 and cleaved caspase 3, of cyclin D1, of p27, and of phosho-Rb in the cells. After 24 hours, caspase 3 was slightly decreased in the metformin and paclitaxel treatments and notably reduced in both cell lines treated with the combination of metformin and paclitaxel (Fig. 3G). Cleaved caspase 3 was slightly increased with metformin or paclitaxel treatments, and strongly increased in metformin and paclitaxel combined treatment after 24 hours in both cell lines (Fig. 3G). Cyclin D1 levels were only reduced in metformin treatment in both cell lines (Fig. 3G). p27 expression was increased in metformin treatment whereas phosphorylation of Rb was reduced in metformin treatment (Fig. 3G).

To determine whether metformin, paclitaxel, or the combined treatment affect cancer cell metabolism we analyzed their effects on mitochondrial complex I oxygen consumption in MCF-7 and A549 cell lines. Metformin decreased complex I oxygen consumption by 58% in MCF-7 and by 92% in A549, whereas paclitaxel had a modest effect in both cell lines (Fig. 3H). The combined treatment showed no significant difference from the metformin treatment alone (Fig. 3H).

Figure 4A and B shows that p53 is activated with paclitaxel treatment and that further stimulation with metformin + paclitaxel, 2-DG + paclitaxel (double therapy), or a combination of metformin + 2-DG + paclitaxel (triple therapy) does not increase its activation. The same pattern is seen in SESN2 expression. On the contrary, AMPK shows a further increase in activation after double or triple therapies compared with paclitaxel only and vehicle treatments. Additionally, mTOR and its direct substrates p70S6K and 4EBP-1 are inhibited with paclitaxel treatment and further inhibition is observed with the double or triple treatments.

To further evidence the role of AMPK and SESNs in the combination treatment, we treated MCF-7 and A549 cells with siRNA to AMPK and to SESN1 and SESN2 and analyzed SESN1 and SESN2 expressions and AMPK, mTOR, p70S6K, and 4E-BP1 phosphorylation. Figure 4C and D show that treatment with SESN1 and SESN2 siRNA dampers SESN1 and SESN2 expressions, respectively, and reduces AMPK phosphorylation, with an increase in mTOR, p70S6K, and 4E-BP1 phosphorylation. Treatment with AMPK siRNA does not reduce SESN1 or 2 expressions but abolishes AMPK phosphorylation and expression, which increases mTOR, p70S6K, and 4E-BP1 phosphorylation.

Thus, these results clearly show an essential role for SESN and AMPK in the activation of AMPK and inhibition of mTOR, respectively, after the combined treatment of metformin and paclitaxel.

Xenografted SCID mice were treated with control vehicle, metformin, paclitaxel, or metformin and paclitaxel. Treatments began when the tumors presented an average size of 50 mm³ and tumor growth rate was measured daily after the beginning of the treatment. Figure 5A shows that metformin and paclitaxel is clearly more effective in reducing tumor growth, as compared with either paclitaxel alone, metformin alone, or the control. For the entire experiment, the animals treated with combination of metformin and paclitaxel presented almost no tumor growth, with the final tumor volume of 71 mm³ being very close to the tumor volume at the beginning of the treatment, as compared with the final volumes of the control (377 mm³), metformin (203 mm³), and paclitaxel (137 mm³) as shown in Figure 5B. We observed that 2-DG and paclitaxel combination yielded similar results to metformin and paclitaxel combination. However, we did not observe an additive effect with the triple therapy when compared with the double therapy (Supplementary Fig. S3A and B).

The reduced tumor growth following metformin and paclitaxel treatment is due to the reduced proliferation of tumor cells, as shown by Ki-67 staining and quantification (Fig. 5C and D) and increased apoptosis, as quantified by TUNEL staining (Fig. 5E and F) and cleaved caspase 3 staining (Fig. 5G and H). In the control group, Ki-67–positive cells were 25.5% (±2.8) of the total, in metformin these cells were 13.4% (±1.1), whereas in paclitaxel these cells were 10.8% (±1.2) and metformin and paclitaxel presented 7.2% (±0.4) Ki-67–positive cells (Fig. 5C and D). The results of the TUNEL staining experiments show that the control group presented a 9.5% (±1.1) apoptosis, whereas metformin apoptosis was 16.7% (±4.5), paclitaxel apoptosis was 31.8% (±1.8), and metformin and paclitaxel apoptosis was 35.9% (±4.8) (Fig. 5E and F). Similarly, in the cleaved caspase 3 staining experiments, the control group presented a 6.5% (±1.0) positive cells, whereas metformin presented 22.3% (±0.8) positive cells, paclitaxel presented 27.4% (±2.1) positive cells, and metformin and paclitaxel combined presented 39.9% (±3.6) cells positive for cleaved caspase 3 (Fig. 5G and H). These data indicate a reduced proliferation and increased apoptosis in the combined treatment and are consistent in showing that there is a significant advantage in the use of combination treatment with metformin and paclitaxel, as compared with treatment with either agent alone.

As treatment with metformin and paclitaxel inhibited tumor growth, we sought to determine the AMPK/mTOR pathway activation status in the tumor tissue of animals treated with metformin, paclitaxel, and the combination of metformin and paclitaxel. Figure 6A shows that both treatments with paclitaxel resulted in a higher acetylation of Lysine 379 of p-53 and increased quantity of SESN2, as compared with control or metformin (Fig. 6B). The phosphorylation of AMPK at Thr172 was also higher in the metformin and paclitaxel treatment, when compared with paclitaxel alone, metformin alone, or control (Fig. 6C). Both treatments with paclitaxel and metformin also activated AMPK, as compared with the control. Similarly, phosphorylation of mTOR (Fig. 6D), and its direct substrates p70S6K (Fig. 6E) and 4E-BP1 (Fig. 6F), were reduced following metformin and paclitaxel treatments.

In this study, we show that the combination of metformin and paclitaxel has a major antitumor effect in vivo and induces massive cell-cycle arrest in vitro. These effects are correlated with a potent activation of AMPK. Our results show that metformin, which induces a moderate decrease in ATP levels (36), is able to produce molecular activation of AMPK and inactivation of mTOR signaling in breast and lung cancer cells, whereas paclitaxel, through activation of p53 and SESNs, yields similar effects to metformin. Combined treatment with metformin and paclitaxel leads to a quantitative increase in AMPK activation and a drastic reduction of molecular signaling through the mTOR pathway. Likewise, the combination of paclitaxel with 2-DG, which like metformin, leads to intracellular ATP depletion (37, 38), severely inhibited the mTOR signaling pathway.

It was initially shown that metformin was capable of reducing proliferation in different types of cancer including, prostate, colon, and breast cancer cell lines. Subsequently, in vivo experiments with metformin resulted in tumor growth inhibition of up to 55% (12, 39). In accordance with these data, we herein show that metformin treatment resulted in a reduction of A549 and MCF-7 cell viability and a decreased tumor volume (approximately 50%) of A549 tumor when xenografted in SCID mice. These effects were paralleled by a decrease in the central regulator of cell growth and survival, mTOR signaling pathway, as measured by p70S6K and 4EBP-1 phosphorylation.

The mechanisms by which cells protect their genetic material during genotoxic stress include the alert of checkpoint proteins and arrest of cell growth and proliferation (40, 41). The major cellular stress-sensing molecule is p53, which halts cell growth and proliferation by increasing SESNs, thus leading to activation of AMPK and inhibition of mTOR (29, 30). Here we show that paclitaxel induces p53 activation in the cancer cells and activates AMPK. AMPK activation resulted in decreased mTOR pathway activity; this effect may be related to the reduction of cell metabolism that is observed during prolonged mitosis, induced by paclitaxel.

Intracellular interactions between different signaling systems may function as mechanisms for enhancing or counter-regulating signaling pathways. In the case of metformin, the cross-talk with paclitaxel-induced signaling pathways resulted in direct interactions between these drug-induced signaling systems at the level of AMPK. Simultaneous treatment with both drugs led to increased phosphorylation of AMPK and a drastic reduction of mTOR signaling pathway. Furthermore, there was no increase of the effects of the combination of metformin and paclitaxel compared with only metformin on tumor cell metabolism. These results suggest that the positive cross-talk between metformin and paclitaxel-induced signaling was due to additive effects on AMPK activation. Further inhibition of mTOR pathway with the triple therapy does not change the antineoplastic effect of metformin and paclitaxel combination.

The mTOR pathway is a crucial pathway, downstream of several growth factor receptors including epidermal growth factor, platelet-derived growth factor, KIT, and insulin like growth factor I receptor, which coordinate tumor growth (42, 43). The deregulated mTOR pathway is very frequent in human cancer. These alterations include mutational activation of the p110α subunit of phosphoinositide 3-kinase (PI3K), loss of PTEN function, overexpression of PI3K, Akt, eIF4E, and p70S6K, as well as inactivation of tuberous sclerosis 1 or 2 (42, 43). It was also established that the mTOR pathway can be inactivated by AMPK (44), which acts through a PI3K-independent mechanism.

The susceptibility of cancer cells to PI3K inhibitors is highly determined by the presence of mutations in components of the PI3K/Akt/mTOR pathway (45). In contrast, our results showed that a drastic reduction of the mTOR pathway, elicited by the combination of metformin and paclitaxel, yields decreased cell viability in both MCF-7, which has a mutational activation of the PI3K catalytic subunit, and A549 cells, which does not harbor genetic alterations in the PI3K/Akt/mTOR pathway. As the mTOR pathway is essential to cell metabolism and growth, and delayed mitosis induced by paclitaxel is associated with a reduction in gene transcription, our data suggest that the cancer cells may be “pathway addicted,” independently of harboring a mutation in the PI3K/Akt/mTOR pathway during paclitaxel-induced cell-cycle arrest. It is interesting to note that the susceptibility of cancer cells to metformin and paclitaxel combination occurred in a LKB1 independent manner. These data are in accordance with Sanli and colleagues (46) that recently showed that metformin can activate AMPK, probably through action of a metabolite derived from complex I inhibition. Thus, our data suggest that metformin antineoplastic effects are effective even when LKB1 is suppressed.

Toxicity elicited by paclitaxel has been linked to irreversible tubulin polymerization, a cell-cycle block at the metaphase–anaphase transition, and cell death (47, 48). On the contrary, in addition to the metabolic activity of AMPK, there is growing evidence that AMPK has a crucial role in the establishment of cell division, and it has been suggested that AMPK may be essential in the coordination between the sensing of energy resources and genome division (49, 50). Our results show that the combination of metformin and paclitaxel has an additive effect on cell viability and, in accordance with a previous study that combined 2 activators of AMPK, metformin and 2-DG, we observed a compelling accumulation of cells in G₂–M (36). As gene transcription is silenced during mitosis and paclitaxel plus metformin induced a more prolonged division, our results suggest that this event leads to a greater decrease in cell viability.

In conclusion, we observed a convergence of paclitaxel- and metformin- induced signaling at the level of AMPK. This mechanism illustrates how different drugs may cooperate to augment antigrowth signals, suggesting that target activation of AMPK by metformin may be a compelling ally in cancer treatment.

No potential conflicts of interest were disclosed.

We thank Dr. Nicola Conran for English language editing. We also thank Luiz Janeri, Jósimo Pinheiro, and Gerson Ferraz for technical assistance.

This study was supported by grants from Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) and Conselho Nacional de desenvolvimento científico e tecnológico (CNPq).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.