AMP-activated protein kinase promotes epithelial-mesenchymal

1 downloads 0 Views 4MB Size Report
Jun 27, 2018 - migration of cancer cells, we undertook a scratch-assay with BT474, A549, .... In cycloheximide chase experiments, inhibition of AMPK led to a ...
© 2018. Published by The Company of Biologists Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution and reproduction in any medium provided that the original work is properly attributed.

AMP-activated protein kinase promotes epithelial-mesenchymal transition in cancer cells by Twist1 upregulation

Authors: Meera Saxena1,2†, Sai A Balaji1,†, Neha Deshpande1,†, Santhalakshmi Ranganathan1, Sravanth Kumar Hindupur 1,3 and Annapoorni Rangarajan1,* 1From

the Department of Molecular Reproduction, Development and Genetics,

Indian Institute of Science, Bangalore-560012, Karnataka, India 2

Current affiliation: Department of Biomedicine, University of Basel, Mattenstrasse 28, 4058

Basel, Switzerland 3

Current affiliation: Biozentrum, University of Basel, Klingelbergstrasse 50/70, 4056 Basel,

Switzerland

*To whom correspondence should be addressed. Tel: 91-80-22933263; Fax: 91-8023600999; Email: [email protected]

authors contributed equally to this work

Journal of Cell Science • Accepted manuscript

†These

JCS Advance Online Article. Posted on 27 June 2018

Summary The developmental programme of epithelial-mesenchymal transition (EMT), involving loss of epithelial and acquisition of mesenchymal properties, plays an important role in the invasionmetastasis cascade of cancer cells. Here we show that activation of AMP-activated Protein Kinase (AMPK) using A769662 leads to a concomitant induction of EMT in multiple cancer cell types, as observed by enhanced expression of mesenchymal markers, decrease of epithelial markers, and increase in migration and invasion. In contrast, inhibition or depletion of AMPK led to a reversal of EMT. Importantly, AMPK activity was found to be necessary for the induction of EMT by physiological cues such as hypoxia and TGFβ treatment. Further, AMPK activation increased the expression and nuclear localization of Twist1, an EMT transcription factor. Depletion of Twist1 impaired AMPK-induced EMT phenotypes, suggesting that AMPK might mediate its effects on EMT, at least in part, through Twist1 upregulation. Inhibition or depletion of AMPK also attenuated metastasis. Our data, thus, underscores a central role for AMPK in the induction of EMT and in metastasis, suggesting

Journal of Cell Science • Accepted manuscript

that strategies targeting AMPK might provide novel approaches to curb cancer spread.

Introduction Metastasis, the spread of tumor cells from primary site to a distant organ, is accountable for more than 90% of cancer-related mortality (Chaffer and Weinberg, 2011). This process involves local invasion of cells from a primary tumor into the surrounding tissue, intravasation into the nearby microvasculature, migration through the blood and lymphatic vessels, extravasation at a distant organ site, and their proliferation to form a secondary macroscopic tumor (Chaffer and Weinberg, 2011). It is believed that the evolutionarily conserved developmental programme of epithelial-mesenchymal transition (EMT) bestows cancer cells with such properties that enable them to initiate the invasion-metastasis cascade (Yang et al., 2006). Indeed, EMT is considered to be an important contributor in invasion and metastasis of several types of cancer (Thiery et al., 2009).

During EMT, epithelial cells shed their intercellular contacts, cell-extracellular matrix interactions and apico-basal polarity, and attain more mesenchymal properties like increased expression of mesenchymal markers, motility and invasive capabilities (Polyak and Weinberg, 2009). EMT can be stimulated by a plethora of cues such as hypoxia, signaling through TGFβ, receptor-tyrosine kinases, Notch and Wnt (Polyak and Weinberg, 2009). In response to these cues, several transcription factors like Twist1, Snail, Slug, FOXC2, Zeb1 and Zeb2 (Polyak and Weinberg, 2009) are upregulated which serve as important downstream mediators of EMT. Of these, Twist1 is considered as a master regulator of EMT (Yang et al., 2004). Several mechanisms of regulation of EMT have been elucidated in recent years (Polyak and Weinberg, 2009). However, a discreet hierarchy or a central orchestrator linking

AMP-activated protein kinase (AMPK) is an evolutionarily conserved stress-sensing kinase that gets activated under a variety of stresses, including hypoxia, ischemia and nutrient deprivation, leading to a change in the ATP:AMP ratio (Steinberg and Kemp, 2009). It is a Serine/Threonine protein kinase that functions as a heterotrimeric protein consisting of a catalytic α subunit (α1 or α2), a scaffolding β subunit (β1 or β2) and a nucleotide binding γ subunit (γ1, γ2 or γ3) (Hardie et al., 2012). Besides its pertinent role in the regulation of energy homeostasis, AMPK has been implicated in several physiological processes including cell growth (Xiang et al., 2004), cell division (Jones et al., 2005), maintenance of epithelial cell polarity (Zhang et al., 2006), and autophagy (Liang et al., 2007). Apart from these, AMPK activity has also been associated with cell migration in normal physiology (Nakano et

Journal of Cell Science • Accepted manuscript

these various mechanisms in the EMT interactome is yet to be elaborated.

al., 2010). For instance, AMPK activation was shown to be important for migration of human umbilical vein endothelial cells (HUVECs) (Nagata et al., 2003) and transendothelial lymphocyte migration (Martinelli et al., 2009). In another study, inhibition of AMPK was shown to suppress TGFβ-mediated apoptosis and EMT of normal murine hepatocytes (Wang et al., 2010). Few recent reports have also associated AMPK signaling with cancer cell migration and invasion (Chen et al., 2011; Chiu et al., 2009; Kim et al., 2011). Another recent study showed that indirect activation of AMPK by mitochondrial dysfunction induced EMT in lung cancer cells (He et al., 2016). In contrast, some recent studies have reported an inhibitory role for AMPK in EMT of breast, prostate, and lung cancer cells, mainly using AICAR and metformin as pharmacological activators of AMPK (Banerjee et al., 2016; Chou et al., 2014; Cufi et al., 2010; Han et al., 2015; Lin et al., 2015; Qu et al., 2014). Thus, there are conflicting reports about the role of AMPK signaling in EMT and cancer metastasis, suggesting that AMPK activation might have cell type and context-specific effects. This might, in turn, be dependent on the upstream activators of AMPK used, some of which may additionally have AMPK-independent effects. This incongruity needs to be resolved before AMPK-targeted therapeutics is widely used for cancer treatment.

In light of these conflicting reports, we investigated the role of AMPK in regulating EMT in multiple cancer types. In this study, we show that AMPK activation with a direct, allosteric activator A769662 suffices to induce a functional EMT in multiple cancer cells, whereas, its inhibition or depletion leads to reversal of EMT. Importantly, we show that AMPK is required for EMT-induction by upstream stimuli relevant to cancer progression, such as

upregulation of Twist1, a major regulator of EMT, and increasing its nuclear localization.

Results To explore a possible role for AMPK in EMT, we initiated our study by investigating the effects of AMPK activation and inhibition on the expression of markers associated with EMT in a variety of cancer cell lines including breast cancer cell lines MCF7, T47D, BT474, MDAMB-231, melanoma cell line MDA-MB-435S and lung adenocarcinoma cell line A549. We first determined the endogenous expression of epithelial marker E-cadherin (E-cad) and mesenchymal markers vimentin (Vim) and N-cadherin (N-cad) in these cell lines (Fig. S1A). The cancer cell lines BT474, MCF7 and T47D exhibiting an E-cad+ and Vim- phenotype

Journal of Cell Science • Accepted manuscript

hypoxia and TGFβ. We further show that AMPK mediates its effects, at least in part, through

represent epithelial cell types. A549 exhibited E-cadlow and Vimlow phenotype, reminiscent of a partially EMT’d cell line, while MDA-MB-231 and MDA-MB-435S cells exhibited E-cadand Vimhigh phenotype, representing mesenchymal cell types. The pattern of expression of these markers matched with that reported in the literature (Lombaerts et al., 2006).

To begin our study, we investigated if AMPK activation would promote the induction of EMT in epithelial type cancer cell lines. Most previous studies have used AICAR or metformin for activating AMPK; however, these have additional AMPK-independent effects as well (Foretz et al., 2010; Kalender et al., 2010). To overcome this problem, we used A769662 (Cool et al., 2006), which directly activates the heterotrimeric form of AMPK both allosterically and by inhibiting its dephosphorylation (Sanders et al., 2007). We confirmed treatment of A769662 by performing immunoblotting analyses to detect the levels of phosphorylated Acetyl CoA Carboxylase (ACC), a bonafide substrate of AMPK (Witters et al., 1991). An increase in the phosphorylation of ACC is indicative of AMPK activation (Fig. S1B). Treatment of T47D and MCF7 cells with A769662 led to an increase in the transcript levels of mesenchymal markers N-cad and Vim and EMT transcription factors Snai1, Slug and Zeb1 (Fig. 1A). Moreover, since decrease in the protein levels of epithelial markers and disruption of adherens junctions are hallmarks of EMT (Huang et al., 2012), we also investigated the effects of AMPK activation on the protein levels of epithelial markers E-cad and Zona Occludens-1 (ZO-1). Immunocytochemical analysis revealed a significant reduction in E-cadherin levels as upon AMPK activation for 48 hours in these cell lines (Fig. 1B). Reduced expression of E-cad and ZO-1 was also detected using immunoblotting approach

and E). However, in spite of the changes at transcript levels (Fig. 1A and S1C), AMPK activation for 48 hours did not lead to the detection of mesenchymal markers Vim and N-cad at protein levels in these epithelial cell types (data not shown). Thus, this set of data revealed that AMPK activation in the epithelial cancer cells suffices to bring about a partial EMT involving the down-modulation of epithelial markers.

Next, we investigated the effects of AMPK activation and inhibition on cancer cells that have already undergone EMT to various extents, such as A549, MDA-MB-231 and MDA-MB435S cells. Activation of AMPK in the partially EMT’d A549 cells, bearing an E-cadlow and Vimlow phenotype, led to an increase in the mRNA levels of mesenchymal markers and EMT transcription factors (Fig. 1C). Immunofluorescence analysis also revealed a reduction in E-

Journal of Cell Science • Accepted manuscript

(Fig. S1B). Similar results were obtained with AMPK activation in BT474 cells (Fig. S1C, D

cad and increase in Vim protein expression (Fig. 1D). Reduction in the expression of epithelial markers E-cad and ZO-1, and increase in the expression of mesenchymal markers Vim and N-cad, was also detected using immunoblotting (Fig. S1F). Thus, AMPK activation further promoted the EMT phenotype of A549 cells. In contrast, inhibition of AMPK using 10 μM Compound C, a potent, reversible, and ATP-competitive inhibitor of AMPK (Zhou et al., 2001), led to a reduction of Vim and N-cad expression at both mRNA and protein levels, and an increase in E-cad and ZO-1 protein levels (Fig. 1C, D and S1F). In MDA-MB-231 cells bearing an E-cad- and Vimhigh phenotype, AMPK activation led to a further increase in the mRNA expression of mesenchymal markers Vim and N-cad and EMT transcription factors Snai1, Slug and Zeb1 while AMPK inhibition led to a reduction in their transcript levels (Fig. 1E). At protein levels, we observed a significant increase in Vim expression upon AMPK activation and a moderate decrease upon AMPK inhibition (Fig. 1F and S1G). As reported previously (Sommers et al., 1991), we failed to detect the expression of E-cad in MDA-MB-231 cells (Fig. S1A), and AMPK inhibition also did not lead to its reexpression in these cells (data not shown). However, AMPK activation led to a decrease in the expression of ZO-1, while its inhibition led to its increase in MDA-MB-231 cells (Fig. S1G). Similar results were obtained with another mesenchymal cell line, MDA-MB-435S (Fig. S1H).

Since pharmacological inhibition of AMPK (with Compound C) might have off-target effects (Liu et al., 2014), we additionally undertook RNAi-mediated depletion of AMPK. To do so,

inducible promoters, respectively, each showing ~50% and ~70% knockdown of AMPK (Fig. S1I and S1J). Depletion of AMPKα2 in the epithelial BT474 cells led to a further increase in the expression of epithelial markers E-cad and ZO-1 (Fig. S1I), while we failed to see any expression of Vim (data not shown). In the partially EMT’d A549 cells, AMPK knockdown led to an increase in the expression of E-cad and ZO-1 while reducing the levels of Vim (Fig. S1J; clone#1). Immunofluorescence analysis further revealed an increase in E-cad expression and a decrease in Vim expression upon AMPK depletion in these cells (Fig. S1K). Similar results were obtained in A549 cells where AMPK was stably knocked down using another shRNA oligo sequence (S1L; clone#4). Similarly, mesenchymal MDA-MB-231 cells that stably express AMPKα2 shRNA (Hindupur et al., 2014) and showed ~80% knockdown (S1M) revealed a reduction in the levels of Vim and N-cad (Fig. S1M and S1N). These results

Journal of Cell Science • Accepted manuscript

we used BT474 and A549 cells stably expressing AMPKα2 shRNA under constitutive and

were further validated using an additional, inducible shRNA targeting AMPKα2 (Fig. S1O; clone#4) in MDA-MB-231 cells. Similar results were obtained upon AMPK knockdown in yet another mesenchymal cell type MDA-MB-435S (Fig. S1P). Thus, these data revealed that in cancer cells that have undergone EMT to varying extent, AMPK activation further promoted, while its inactivation or depletion reversed the morphological changes associated with EMT, suggesting that AMPK might play a key role in regulating the process of EMT.

Attainment of motility and invasive capabilities are hallmark features of a functional EMT (Polyak and Weinberg, 2009). To assess whether AMPK can modulate the motility and migration of cancer cells, we undertook a scratch-assay with BT474, A549, MDA-MB-231 and MDA-MB-435S cells treated with either AMPK activator A769662 or its inhibitor Compound C. In keeping with a failure to induce mesenchymal marker expression in epithelial cell lines, AMPK activation failed to alter migration of BT474 cells (Fig. 2A & S2A). In the partially EMT’D A549 cells, A769662 treatment led to a marginal increase in migration whereas Compound C treatment resulted in a significant decrease in the motility (Fig. 2B & S2B). Interestingly, mesenchymal MDA-MB-231 and MDA-MB-435S cells showed a significant increase in migration when treated with A769662 and a significant inhibition of motility with Compound C treatment (Fig. 2C, S2C, 2D and S2D). To further confirm these data, we additionally undertook yet another cell migration assay with MDAMB-435S cells involving quantitative real-time impedance measurement using an ECIS (electric cell-substrate impedance sensing)-based technique (Hong et al., 2011a). In

marked increase in migration of MDA-MB-435S cells (as revealed by an increase in impedance), Compound C treatment led to a decrease in migration (as revealed by a decrease in impedance) compared to the control cells (Fig. S2E). Further, to confirm the results with Compound C, we employed RNAi-mediated knockdown of AMPKα2. Consistent with pharmacological inhibition of AMPK, A549, MDA-MB-231 and MDA-MB-435S cells stably expressing AMPKα2 shRNA or transfected with siRNA targeting AMPKα2 also showed decreased migration compared to scrambled shRNA control cells (Fig. 2E, S2F, 2F, S2G, 2G and S2H). Knockdown of AMPK failed to change the migration in BT474 cells (data not shown). The effect of AMPK inhibition/knockdown on the migratory properties of mesenchymal cells (Fig 2F, S2G, 2G and S2H) was more dramatic compared to the partially EMT’d A549 cells (Fig. 2E and S2F).

Journal of Cell Science • Accepted manuscript

corroboration with the scratch assay results, we found that while A769662 treatment led to a

To determine the effects of AMPK activation and inhibition on the invasive potential of cancer cells, we performed in vitro Boyden chamber invasion assay. Since mesenchymal cells showed maximal effect on AMPK modulation, we undertook invasion assays with these cell types. Compared to vehicle treated control cells, we found that while A769662 treatment caused a significant increase in the invasiveness of MDA-MB-231 and MDA-MB-435S cells, Compound C treatment decreased their invasive potential (Fig. 2H and 2I). Further, knockdown of AMPKα2 in MDA-MB-231 and MDA-MB-435S cells also led to a marked decrease in their invasive potential compared to control cells (Fig. 2J, S2I and 2K). These results collectively demonstrated that while AMPK activation increased the migratory and invasive capability of cancer cells, thereby bringing about a functional EMT, its inhibition or knockdown reversed EMT, suggesting that AMPK might play a critical role in the morphogenetic processes involved with EMT.

We next investigated the role of AMPK in the context of physiological stimuli such as hypoxia that is a known trigger of EMT (Thiery et al., 2009; Yang et al., 2008). Intriguingly, hypoxia is also a well-known trigger for AMPK activation (Mungai et al., 2011). Based on our data above, we hypothesized that hypoxia-induced EMT might be mediated by AMPK activation. To test this, we exposed epithelial BT474 cells stably expressing either scrambled or AMPKα2 shRNA to the hypoxia-mimetic CoCl2. Increase in the levels of Hif1α on treatment with CoCl2 revealed the generation of hypoxic environment (Fig. 3A). Consistent

AMPK activity, as revealed by elevated pACC levels (Fig. 3A). Compared with untreated control cells, a decrease in the expression of E-cad and ZO-1 on CoCl2 treatment revealed the induction of EMT in scrambled shRNA expressing BT474 cells. Interestingly, hypoxiainduced EMT was inhibited upon AMPK knockdown in these cells (Fig. 3A). Immunoblot analysis revealed similar results for partial EMT’d A549 cells stably expressing AMPK2 shRNA and incubated under hypoxic conditions (3% oxygen) in a trigas incubator (Fig. 3B; clone #4). Similarly, in mesenchymal MDA-MB-435S cells, an increase in the expression of Vim and N-cad in 3% oxygen condition revealed the induction of EMT under hypoxia, which was inhibited in the presence of AMPK inhibitor Compound C (Fig. 3C). Similar results were obtained upon treatment with CoCl2 in MDA-MB-231 cells stably expressing AMPKα2

Journal of Cell Science • Accepted manuscript

with literature (Mungai et al., 2011), treatment with CoCl2 also triggered an increase in

shRNA (Fig S3A) or in the presence of Compound C (Fig S3B). These data reveal that AMPK might be required for hypoxia-induced EMT. Besides hypoxia, another potent inducer of EMT is TGFβ (Polyak and Weinberg, 2009; Yang and Weinberg, 2008). It has previously been shown that treatment of murine hepatocytes and AML12 cells with TGFβ leads to activation of AMPK signaling (Wang et al., 2010). We thus speculated that AMPK activity might also be required for TGFβ-induced EMT of cancer cells. To explore this, MDA-MB-435S cells were treated with TGFβ for 48 hours. An increase in the protein levels of pSmad2 (Fig. 3D) confirmed the activation of TGFβ signaling in these cells. In keeping with the data of Wang et al., we also saw an increase in AMPK activity upon TGFβ activation, as revealed by an increase in the levels of pACC (Fig. 3D). Further, increased levels of Vim and N-cad revealed the induction of EMT (Fig. 3D). Intriguingly, TGFβ induced-EMT was significantly inhibited in the presence of AMPK inhibitor Compound C (Fig. 3D). Similar results were obtained upon TGFβ treatment in the presence of AMPK inhibitor, Compound C in A549 (Fig S3C) and MDA-MB-231 cells (Fig. S3D). To further corroborate the data obtained with pharmacological inhibition of AMPK, we addressed the role of AMPK in TGFβ-induced EMT by subjecting MDA-MB-231 cells stably expressing either scrambled or AMPKα2 shRNA to TGFβ treatment for 48 hours. An increase in the levels of Vim and N-cad revealed the induction of EMT marker expression. Depletion of AMPK prevented TGFβ-mediated induction of EMT markers at protein levels (Fig. S3E). Together, these data reveal that AMPK might be required for TGFβ-mediated induction of

Ras signaling has also been shown to promote EMT (Mukherjee, 2010), as well as increase AMPK activity (Rios et al., 2013). To test if AMPK is also important for Ras-induced EMT, we inhibited AMPK in HMLER breast cancer cells (Elenbaas et al., 2001) using Compound C. A reduction in the levels of pAMPK and pACC revealed the effects of Compound C (Fig S3F). We observed that inhibition of AMPK led to a decrease in the expression of mesenchymal markers like Vim and N-cad while increasing the expression of epithelial marker E-cad in HMLER cells (Fig. S3F). Thus, collectively these results pointed towards a critical role for AMPK in the induction of EMT by various upstream stimuli.

Journal of Cell Science • Accepted manuscript

EMT.

Since our study highlighted the role of AMPK in EMT induction, and EMT is considered to be a pre-requisite for metastasis (Polyak and Weinberg, 2009), we next investigated the requirement of AMPK function for metastasis in vivo. To do so, we subjected MDA-MB-231 scrambled and AMPKα2 shRNA expressing stable cells to experimental tail vein metastasis assay. While several microscopic metastatic colonies were detected in the lungs of mice injected with scrambled shRNA cells, no metastatic growth was detected in mice injected with AMPKα2 knockdown cells (Fig. 4A, S4A). Similar results were obtained with MDAMB-435S cells stably expressing scrambled or AMPKα2 shRNA (Fig. 4B, S4B). An MTT assay revealed no significant change in cell proliferation when AMPK is depleted (Fig. S4C), together suggesting that AMPK functions might be required for the metastasis of these cancer cells.

We next investigated the mechanisms by which AMPK might regulate EMT. Induction of EMT is associated with the upregulation of several transcription factors like Twist1, Snail, Slug and Zeb1 (Yang and Weinberg, 2008). Of these, Twist1 has been considered as a major regulator of EMT, invasion and metastasis (Wang et al., 2016; Yang et al., 2004). Therefore, we analyzed the expression of Twist1 under conditions of AMPK activation. We noticed an AMPK-mediated increase in Twist1 transcripts in a variety of cell lines tested including MCF7, BT474, A549, MDA-MB-231 and MDA-MB-435S cells (Fig. 5A). Notably, the EMT’d cells showed a greater increase in Twist1 transcript levels upon AMPK activation compared to the non-invasive cells (Fig. 5A). A western blot analysis revealed elevated

immunocytochemical analysis further confirmed elevated Twist1 levels (Fig. 5C).

To validate if Twist1 regulation is through AMPK activation, we next investigated the effects of AMPK inhibition on Twist1 expression in the EMT’d cancer cells lines A549, MDA-MB231 and MDA-MB-435S. Indeed, in these cell lines, inhibition of AMPK with Compound C led to a significant decrease in the levels of Twist1 transcripts (Fig. 5D). Consistent with this, MDA-MB-231 and MDA-MB-435S cells stably expressing AMPKα2 shRNA also showed less expression of Twist1 compared to scrambled cells (Fig. 5E). Importantly, inhibition of AMPK prevented the induction of Twist1 expression by CoCl2 or TGFβ treatment (Fig. 5F and G). Taken together, these data suggested that AMPK might mediate its effects on EMT through upregulation of Twist1.

Journal of Cell Science • Accepted manuscript

Twist1 protein levels in the presence of A769662 (Fig. 5B and S5A). An

To gauge the role of Twist1 downstream of AMPK in the process of EMT, we investigated the effects of Twist1 depletion in AMPK-mediated EMT response of cancer cells. For this, we generated MDA-MB-231 cells stably expressing Twist1 shRNA (Mani et al., 2008). These cells showed ~90% reduction of Twist1 both at transcript level (Fig. 5H) and protein level (Fig. S5B and S5C) compared to control shGFP cells. Consistent with the literature (Mani et al., 2008), Twist1 knockdown led to a reduction in the expression of EMT markers like Vim, N-cad, Fibronectin 1 (FN1), Snail and Slug (Fig. 5I). As shown previously, AMPK activation with A769662 led to increased EMT marker expression in the control shGFP cells; however, AMPK activation failed to do so in the Twist1 knockdown cells (Fig. 5I). Similar results were obtained in MDA-MB-435S cells expressing Twist1 siRNA oligos (Fig. S5D and S5E) indicating that AMPK might mediate its effects on EMT, at least in part, through Twist1.

Further, to enquire if AMPK requires Twist1 for mediating functional changes associated with EMT, we investigated the effects of AMPK activation in the migration and invasion of Twist1 knockdown cells. As seen before, AMPK activation led to an increase in the migration and invasive potential of control shGFP cells (Fig. 5J, 5K and S5F); however, AMPK activation in shTwist1 cells failed to do so (Fig. 5J, 5K and S5F).

Since Twist1 is a transcription factor and its nuclear localization is critical for its function, we investigated nuclear localization of Twist1 under AMPK activated conditions. As seen in the

noted increased nuclear localization of Twist1 in AMPK activated condition. Since AMPK activation led to a significant increase in Twist1 in MDA-MB-231 cells (Fig. 5A and S5A), we performed subcellular fractionation experiments with the same. Cell fractionation experiments also showed elevated nuclear Twist1 upon AMPK activation (Fig. 6A). Further, immunofluorescence imaging of MDA-MB-231 and MCF7 cells transiently transfected with a GFP-tagged Twist1 construct also showed increased nuclear localization of Twist1 in A769662 treated condition (Fig. 6B), suggesting possible AMPK-mediated post-translational modifications of Twist1 leading to its nuclear localization. Since phosphorylation by Mitogen-activated protein kinases (MAPK), extracellular signal-regulated kinases (ERK), Jun N-terminal kinases (JNKs) has been shown to increase Serine phosphorylation and stability of Twist1 (Hong et al., 2011b), we investigated the effect of AMPK activation on Twist1

Journal of Cell Science • Accepted manuscript

immunofluorescence images (Fig. 5C), in addition to increased levels of Twist1, we also

stability. In cycloheximide chase experiments, inhibition of AMPK led to a decrease in Twist1 stability (Fig. 6C). Moreover, we observed increased Twist1 stability in the presence of AMPK activator (data not shown). Taken together, our data revealed that AMPK activation leads to an increase in the expression, stability, and nuclear localization of Twist1 protein, which could ultimately promote EMT.

In conclusion, our study uncovers a central role for AMPK in the positive regulation of EMT in cancer cells, mediated at least in part, through Twist1 upregulation.

Discussion

Epithelial-mesenchymal transition (EMT) is suggested to be an initial and important step in the invasion-metastasis cascade (Thiery et al., 2009). Therefore, an increased understanding of the molecular players involved in this process is likely to lead to a better understanding of cancer metastasis, as well as identify newer targets for cancer treatment. Though several signaling pathways and physiological conditions have been identified as activators of EMT, and a plethora of transcription factors have been identified as downstream effectors of EMT, it has remained unclear if these various inducers of EMT function independently, work in conjunction, or converge on a common downstream mediator. Our study identifies a central role for AMPK in mediating an EMT downstream of several physiological cues. Even though the LKB1-AMPK axis was initially associated with tumor-suppressive roles (Huang et al., 2008), recent studies have demonstrated pro-tumorigenic functions for AMPK including

cell migration (Hindupur et al., 2014; Jeon and Hay, 2015; Kim et al., 2011; Ng et al., 2012; Wang et al., 2010). In our study, we provide evidence for a direct role of AMPK in the induction of EMT in cancer cells, as well as in the maintenance of the EMT phenotype. We further show that activation of AMPK suffices to bring about a functional EMT characterized by increased migration and invasion potential. Moreover, we show that AMPK activation is mandatory for EMT induction mediated by upstream physiological cues like hypoxia and TGFβ. Additionally, we show that AMPK mediates its effects on EMT via upregulation of Twist1 gene expression and its increased nuclear localization.

Journal of Cell Science • Accepted manuscript

inhibition of apoptosis, promoting anchorage-independent growth, and more recently cancer

AMPK activation with A769662 affected the expression of epithelial markers E-cad and ZO-1 to different extents in different cell lines promoting an EMT. In the epithelial breast cancer cells MCF7, T47D and BT474, AMPK activation decreased the expression of epithelial markers E-cad and ZO-1 and caused dissolution of adherens junctions from the cell membrane (the first hallmark of an active EMT), thus pushing the cells towards initiating an EMT. Further, even though AMPK activation upregulated the transcript levels of mesenchymal markers, (Vim and N-cad) in these cells, their protein levels are not changed (data not shown), and consistent with this, their migratory properties also did not change. Hence, these cells seem to only “partially” undergo an EMT. In metastable A549 cells, which already seem to be in a partial EMT state as evident by E-cadlow and Vimlow phenotype, AMPK activation significantly decreased the protein expression of E-cad and ZO-1 and upregulated mesenchymal markers (Fig. S6). Thus, metastable A549 cells, which seem to be “primed” for EMT, are hence pushed further into EMT in response to AMPK activation. Mesenchymal MDA-MB-231 cells do not express E-cad while expressing ZO-1 at a low basal level. AMPK activation was able to further downregulate ZO-1 expression in these mesenchymal cells, as well as significantly increased the expression of mesenchymal markers (Fig. S6) and their migratory and invasive potential. Hence, it appears that while AMPK activation is able to downregulate the expression of epithelial markers across all states of EMT in different cell types, additional cues might be required to also increase the protein expression (Fournier, 2010) of mesenchymal markers in epithelial cells (MCF7, T47D and BT474) to bring about a complete, functional EMT. Whereas, primed partial EMT’d cells

the requisite machinery to bring about a complete, functional EMT in response to AMPK activation.

On the other hand, AMPK inhibition or depletion increased the epithelial characteristics of the cells while decreasing mesenchymal properties. In the epithelial BT474 cells, AMPK knockdown led to a further increase in epithelial markers (Fig. S6). In the partially EMT’d A549 cells, AMPK inhibition or knockdown led to an increase in epithelial markers E-cad and ZO-1 as well as reduction in mesenchymal markers Vim and N-cad. AMPK inhibition or depletion in the mesenchymal MDA-MB-231 and MDA-MB-435S cell lines decreased the expression of mesenchymal markers Vim and N-cad (Fig. S6); however, it failed to bring back E-cad expression (data now shown). It is known that the E-cad promoter is

Journal of Cell Science • Accepted manuscript

(A549) and mesenchymal cells (MDA-MB-231 and MDA-MB-435S) seem to already possess

hypermethylated in MDA-MB-435S cells completely silencing E-cad expression, whereas it is partially methylated in MDA-MB-231 cells leading to a very low expression of E-cad that cannot be detected by western blotting (Lombaerts et al., 2006). It is possible that AMPK signaling may not be able to alleviate such hypermethylation-induced gene silencing. These data suggested that AMPK is required to maintain the EMT phenotype, while its inhibition leads to reversion of EMT, also known as mesenchymal-to-epithelial transition (MET) (Nieto et al., 2016).

Our data revealed that activation of AMPK using A769662 promotes EMT. In contrast, metformin, also an AMPK activator, has been shown to inhibit EMT phenotype of breast cancer cells (Qu et al., 2014). Another study showed that activation of AMPK by OSU-53, a new allosteric activator of AMPK, reversed EMT (Chou et al., 2014). Metformin was shown to inhibit EMT by AMPK-mediated abrogation of Erk activity leading to downregulation of Snail and Slug in lung and breast cancer cells (Banerjee et al., 2016). AMPK signaling has been shown to inhibit EMT in other cancer types like melanomas, lung and prostate cancer (Chou et al., 2014; Kim et al., 2012; Lin et al., 2015). These discrepancies could be due to the extent of AMPK activation by different pharmacological agents, or due to their non-specific activities. Metformin, a biguanide used as a first-line anti-diabetic drug, is an indirect modulator of AMPK. It inhibits complex I of the mitochondrial respiratory chain causing an increase in intracellular AMP levels, thereby activating AMPK (Owen et al., 2000). However, AMPK-independent roles of metformin have been reported (Foretz et al., 2010; Kalender et al., 2010) and its use as a bonafide AMPK activator needs further investigation. A769662 and

not cause a change in the ATP:ADP ratio or alter mitochondrial function to exert their action. However, AICAR monophosphate (ZMP), the active form of AICAR within cells, is a natural intermediate of purine nucleotide synthesis pathway and is metabolized by AICAR transformylase (Kim et al., 2016). Therefore, the turnover of the active form of AICAR within cells is dependent on their proliferation rate. Thus, these modulators of AMPK might activate it to different extents. Our results with A769662 showed that AMPK is a positive regulator of EMT in multiple cancer cell lines. We have additionally corroborated our results with genetic approaches using either siRNA or inducible shRNA mediated knockdown.

Journal of Cell Science • Accepted manuscript

AICAR, both analogs of AMP, bind directly to the auto-inhibitory domain of AMPK and do

Most importantly, our data revealed that TGFβ, hypoxia, and oncogenic Ras-mediated EMT are significantly inhibited by depletion or inhibition of AMPK. This suggested that activation of AMPK may be a key central event in the EMT programme. Recent studies have identified critical functions of AMPK under patho/physiological stresses associated with tumor progression including nutrient deprivation, hypoxia, and matrix-detachment (Hindupur et al., 2014; Steinberg and Kemp, 2009). Our identification of a stress-response kinase as a necessary component of the EMT programme additionally supports the notion that the phenomenon of EMT might be a stress response that aids cancer cell dissemination and metastasis. Furthermore, given that the earlier studies have linked the attainment of an EMT phenotype with the acquisition of stemness properties (Mani et al., 2008), it is plausible that AMPK activity may determine the differentiation status of epithelial cancers.

EMT is largely effected by a set of transcription factors whose expression and function are under tight control. Twist1, considered as a master regulator of EMT and metastasis (Yang et al., 2004), is expressed in a variety of cancers (Ansieau et al., 2010). Several transcriptional regulators of Twist1 have been identified, including, HIF1α, interleukin-6 (IL-6), NFκB, and Smad-dependent TGFβ pathway (Cheng et al., 2008; Pham et al., 2007; Tan et al., 2012). Our study showed that Twist1 gene expression is transcriptionally regulated by AMPK. Indeed, AMPK has been shown to regulate gene expression by phosphorylating various transcription factors (Canto and Auwerx, 2010), or via its effects on chromatin structure by modulating histone deacetylase activity or histones (Bungard et al., 2010; McGee et al., 2008). However,

elucidated.

Our study additionally revealed that AMPK activation led to increased stability and nuclear localization of Twist1. Such a multi-step regulation of Twist1 by AMPK suggests that the cellular events downstream of the AMPK-Twist1 axis may play a critical role in orchestrating the EMT programme. Twist1 is also known to be phosphorylated by kinases like Akt and MAPK which affect its anti-apoptotic function and stability (Hong et al., 2011; Vichalkovski et al., 2010). Although Twist1 possesses an AMPK consensus motif (data not shown), AMPK-mediated phosphorylation of Twist1 is yet to be specified. Interestingly, we also observed a significant increase or decrease in levels of other EMT transcription factors like Snail, Slug and Zeb1 on AMPK activation or inhibition, respectively. Further, like Twist1,

Journal of Cell Science • Accepted manuscript

the precise mechanisms of Twist1 transcriptional regulation by AMPK remains to be

Snail and Slug also possess an AMPK consensus motif in their protein sequence (data not shown) suggesting that AMPK might additionally regulate these EMT-TFs by phosphorylating them. Hence, while in this study we have investigated the AMPK-Twist1 axis, we speculate that AMPK might also be able to regulate EMT mediated by other transcription factors.

The eventual outgrowth of metastatic colonies at a secondary organ site is the culmination of a series of steps referred to as the invasion-metastasis cascade. EMT is proposed to be one of the early steps that helps initiate this cascade enabling the epithelial cells at the primary tumor site to undergo morphogenetic changes and assume a mesenchymal morphology that enables the cancer cells to migrate and invade following detachment from the extracellular matrix. In fact, previous study from our lab and that of others has shown that AMPK is activated on detachment from matrix (Hindupur et al., 2014; Jeon and Hay, 2012) supporting our finding that AMPK activation leads to EMT. After traversing through blood/lymphatic vessels, the EMT’d cancer cells extravasate at a secondary site, revert their morphology to an epithelial phenotype,

re-attach

to

the

matrix

followed

by

a

clonal

outgrowth

called

micro/macrometastasis. This process is referred to as MET (Nieto et al., 2016). Indeed, we have previously shown that re-attachment to matrix down-modulates AMPK activity (Sundararaman et al., 2016), suggesting that at later stages of metastasis, AMPK might be inactivated, causing MET and thus helping the formation of a new tumor growth at a secondary site. This is consistent with our observation that AMPK inhibition leads to a reduction in EMT markers. Thus, matrix-adhesion state could regulate AMPK activity which

Taken together, our study elicits an important role for AMPK in tumor progression by regulating EMT, migration and invasion of breast cancer cells. The requirement of AMPK downstream of several cues that are known activators of EMT in cancer cells suggests a possible central role for AMPK in orchestrating the EMT programme. Targeting AMPK might thus serve as a novel therapeutic strategy to control cancer metastasis. Since AMPK plays a major role in maintaining energy homeostasis under metabolically stressed conditions, its inhibition for a therapeutic window with minimal systemic effects might be exploited for cancer treatment in the future with the development of specific AMPK inhibitors for clinical use.

Journal of Cell Science • Accepted manuscript

in turn could regulate EMT and MET reversibly, thus enabling metastasis (Fig. 6D).

Materials and Methods

Cell culture and reagents MDA-MB-231, MCF7, T47D (breast epithelial cancer cell lines), MDA-MB-435S (Melanoma), A549 (lung cancer) cells were cultured in DMEM (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% FBS, penicillin (1 KU/ml) and streptomycin (0.1 mg/ml). BT474 (breast cancer) cells were cultured in RPMI (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% FBS, penicillin (1 KU/ml) and streptomycin (0.1 mg/ml). HMLER cells were cultured in DME-F12 medium supplemented with hEGF (10 ng/ml), hydrocortisone (0.5 μg/ml) and insulin (10 μg/ml). Pharmacological chemicals used in this study include A769662 (100 µM; University of Dundee, UK), 6-[4-(2-piperidin-1-ylethoxy-phenyl)]-3-pyridin-4-yl-pyrrazolo [1,5-a]pyrimidine Compound C) (10 µM; Calbiochem, Gibbstown, NJ, USA) and cycloheximide (100 μg/ml; Sigma-Aldrich). Absolute dimethyl sulfoxide (DMSO) (Calbiochem) was used as a vehicle for A769662 and Compound C. Recombinant human TGF-β (R&D Systems, MN, USA) was used at a concentration of 5 ng/ml. For hypoxia-based experiments, cells were either cultured in 3% oxygen conditions using a Trigas incubator (Thermo Fisher, Waltham, MA, USA) or treated with 150 μM cobalt chloride.

Antibodies and immunoblotting Whole cell lysates for Western blotting were prepared with lysis buffer containing 1% NP40 detergent, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM sodium fluoride, 1 mM sodium

(Roche, Mannheim, Germany). Protein concentration was estimated with Bradford reagent and equal amount of protein (25 μg) was resolved by SDS-PAGE using Bio-Rad apparatus, transferred to PVDF membrane (Millipore, Billerica, MA, USA) and probed with appropriate antibodies. For all Western blots involving phospho-protein analysis, the levels of respective total proteins was probed and shown to be equal. α-Tubulin (Calbiochem) served as a loading control. HRP-coupled secondary antibodies were obtained from Jacksons laboratories, and immunoblots were visualized using PICO reagent (Pierce, Waltham, MA, USA). Primary antibodies used were phospho AMPK (Thr172), total AMPKα2, phospho ACC, total ACC (recognizes both isoforms 1 and 2), phospho Smad2, total Smad2 and Vimentin (Cell Signaling Technology, Beverly, MA, USA); N-Cadherin and E-cadherin (Epitomics, Burlingame, CA, USA); Twist1 (Sigma aldrich); HIF1α (Upstate, Billerica, MA, USA).

Journal of Cell Science • Accepted manuscript

orthovanadate, 10 mM sodium pyrophosphate (Sigma-Aldrich) and protease inhibitors

Multiple panels for a given experiment were developed from the same blot by re-probing. Alternately, same lysates were loaded as technical replicates; for each technical run, αTubulin has been probed as loading control. All primary antibodies were used at 1:1000 dilutions in 5% BSA, whereas HRP-conjugated secondary antibodies were used at 1:3000 dilution.

Quantitative densitometric analysis of the EMT-related immunoblots was undertaken using ImageJ software. Protein expression was normalized to α-Tubulin; the normalized expression of control was set to 1, and an increase or decrease in expression under various conditions was calculated relative to the control.

Nuclear-cytoplasmic fractionation For fractionation experiments MDA-MB-435S cells were grown in 90 mm dishes until they reached 80% confluency. The cells were then treated with A769662 for 2 hours and lysed in mild lysis buffer (0.5% NP-40, 10 mM sodium pyrophosphate, 1 mM sodium orthovanadate, protease inhibitor made in PBS). The supernatant obtained after spinning the lysate served as the cytoplasmic fraction. The pellet obtained was gently washed twice with 1X PBS, subjected to a freeze-thaw cycle at -80°C and then resuspended in western lysis buffer cocktail. The supernatant obtained after spinning the lysate served as the nuclear fraction. The cytoplasmic and the nuclear fractions were electrophoresed on an SDS-PAGE and subsequently subjected to immunoblotting.

All plasmid transfections were done using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) as per manufacturer’s instructions.

RNAi experiments AMPKα2 knockdown stable cells were generated by transfecting BT474, MDA-MB-435S and MDA-MB-231 (Hindupur et al., 2014) cells with a pool of four shRNA constructs (HuSH-29 shRNA vectors; Origene, Rockville, MD, USA) targeting AMPKα2. Stable cells were generated using puromycin selection followed by sorting for RFP expression (encoded by the vector) and were expanded and frozen for future use. Knockdown was confirmed by immunoblotting. Simultaneously, all the cell types were transfected with scrambled shRNA

Journal of Cell Science • Accepted manuscript

Transfection

and stable cells were generated as above. MDA-MB-231 and BT474 cells with a stable knockdown of AMPK2 were validated and described earlier (Hindupur et al., 2014). Inducible AMPKα2 knockdown cells were generated by transfecting pTRIPZ vector expressing shRNA against AMPKα2 under Tet-on promoter (Seq #1: V2THS_57638 (Clone#1); Seq #4: V2THS_375319 (Clone#4); Dharmacon, Pittsburgh, United States) in A549 and MDA-MB-231 cells. Stables were selected using puromycin, followed by sorting for RFP (encoded by the vector) after induction with doxycycline (5 µg/ml) for 48 hours. Knockdown was confirmed using immunoblotting. For further experiments, uninduced stable cells were used as control.

For siRNA transfections, cells were seeded in 60 mm dishes at 50% confluence and were transfected with non-targeting control siRNA (Sigma-Aldrich), Twist1 siRNA (SigmaAldrich) or AMPKα2 siRNA (Cell Signaling Technology) using oligofectamine (Invitrogen) as per manufacturer’s instructions.

RNA extraction and RT-PCR Total RNA was isolated using TRI reagent (Sigma-Aldrich). Reverse transcription of mRNA was carried out using Gene-Amp RNA PCR cDNA synthesis kit (Applied Biosystems, Carlsbad, CA, USA). Specific primers (Supplementary Table S1) were designed using Primer 3.0 software. HPRT or β2m were used as normalizing control.

Migration assay was done using imagelock 96 well plate in IncuCyte Zoom (Essen Bioscience) where 2x104 of control and knockdown of MDA-MB-231, BT474 and A549 cells were seeded into each well and treated with 10 µg/ml of Mitomycin C (Calbiochem) for 2 hours. A scratch wound was made using Wound makerTM (Essen Bioscience). Cells were then washed with 1X PBS and incubated with culture media without serum and imaged every 1 hour. Extent of wound healing was determined as wound confluence percentage using IncuCyte software (Essen Bioscience). Each individual experiment had 3-5 technical replicates. Independently, equal number of MDA-MB-435S cells were seeded in 60 mm dishes. siRNA transfection was done for 48 hours. Cells were treated with 10 μg/ml of Mitomycin C

Journal of Cell Science • Accepted manuscript

Wound healing scratch assay

(Calbiochem) for 2 hours to arrest proliferation following which two wounds were made using a P-200 pipette tip. Photomicrographs were taken at 0 and 48 hours of wound generation. The distance migrated was quantified using ProgRes capture software and plotted as a difference of distance migrated between 0 and 48 hours.

ECIS wound-healing assay Real time quantitative wound-healing assays were performed using the Electric Cell-substrate Impedance Sensing (ECIS Model 1600R; Applied BioPhysics, Troy, NY, USA) and the 8W10E ECIS arrays. MDA-MB-435S cells were treated with A769662 or Compound C for 48 hours followed by treatment with Mitomycin C for 2 hours to arrest proliferation, trypsinized and a total of 2x105 live cells seeded in 3 wells of the 8W10E ECIS array. One well with only media (no cells) served as a negative control to give the baseline changes in impedance. Cells were allowed to attach for 24 hours after which a wound was made using elevated AC voltage pulse with a frequency of 40 kHz, and 30 seconds duration. Dead cells were washed away by replacing the media. Migratory response over the next 24 hours was measured in real-time by recording the recovery of electrical impedance.

Matrigel invasion assay MDA-MB-231 and MDA-MB-435S cells treated with A769662 or Compound C for 48 hours were trypsinized and a total of 50,000 live cells were seeded in the BD BioCoatTMMatrigelTM invasion chambers. DMSO served as the vehicle control. After 24 hours, the non-invaded cells were removed, the invaded cells were fixed with 4% paraformaldehyde and stained with

number of invaded cells was counted from these images using ImageJ software and plotted as a graph. MDA-MB-231 (2x104 cells) and MDA-MB-435S (5x104) cells stably expressing control shRNA or AMPKα2 shRNA were similarly subjected to matrigel invasion assay.

Metastasis assay A total of 1×106 cells (control or AMPKα2 knockdown cells) in 50 µl DMEM were injected into the lateral vein of nude mice. The mice were euthanized after nine weeks of injection. The lungs were dissected out, fixed in 4% paraformaldehyde overnight and embedded in paraffin. Hematoxylin and eosin staining was undertaken on multiple lung sections and scored by a pathologist for the detection of metastatic foci. Animal experiments were performed in compliance with CPCSEA guidelines.

Journal of Cell Science • Accepted manuscript

crystal violet blue. Photomicrographs of the invasion chambers were taken at 10X. The

Immunofluorescence Various cancer cell lines or stable cells generated were treated with either A769662 (100µM), Compound C (10µM) or vehicle control DMSO for 24 or 48 hours as mentioned and fixed with 4% paraformaldehyde at room temperature for 2 hours. The cells were permeabilized with PBS containing Triton-X 100, blocked with 1% BSA and probed with antibodies (1:200 dilution) at 4oC overnight. The cells were probed with secondary antibody tagged with FITC or Cy3 (1:200 dilution) for 2 hours at room temperature. The cells were then washed, mounted and visualized under epifluorescence microscope (Olympus IX71). Further image processing and quantification was performed using ImageJ software. For relative fluorescence intensity (RFI) quantification, fluorescence intensity was first normalized to number of nuclei; the normalized intensity of control was set to 1, and the change in intensity under various conditions was plotted relative to control.

Proliferation Assay MDA-MB-231 (5x103), BT474 (8x103) and A549 (5x103) cells expressing scrambled shRNA or AMPKα2 shRNA were seeded into a 96 well plate to measure their proliferation rates. Cells were seeded in triplicates for each time point (day 0- day 4). At each time point, 20 µl of MTT reagent (5 mg/ml) was added to each well of the plate and incubated for 4 hours. The incubation media was then removed and 100 µl of DMSO was added to each well to dissolve the formazan crystals and absorbance was measured using plate reader at 570 nM.

Determination of protein stability For Twist1 protein stability experiments following AMPK inhibition, MCF7 cells were seeded in eight 60 mm dishes at 30-40% confluency, and transfected with 1 μg of pcDNA3Flag Twist1 each from a common DNA-transfection reagent mix. After 24 hours of seeding, four dishes were treated with compound C for 6 hours following which 100 μg/ml cycloheximide (CHX) was added to all the dishes. Compound C treated and untreated dishes were harvested at 0, 2, 4 and 6 hours of cycloheximide treatment for immunoblotting.

Journal of Cell Science • Accepted manuscript

Proliferation rate graph was plotted using Graph-Pad prism software.

Statistics All statistical analysis was performed using ratio t-test one-way ANOVA and two-way ANOVA with GraphPad Prism 5.0 software. All data were presented as mean ± S.E.M of three independent experiments unless otherwise stated. P-values < 0.05 were considered to be statistically significant.

Acknowledgements We thank Prof. Robert Weinberg for HMLER breast epithelial cell lines. We thank Prof. Sendurai A. Mani for Twist-shRNA construct. We acknowledge G S Sravan, D Mohan Pillai and S Pandey for helping in western blot experiments. MS was a recipient of the CSIR Senior Research Fellowship; ASB was a recipient of Prime Minister’s fellowship. This work was supported by the Wellcome Trust/DBT India Alliance Fellowship [grant number 500112-Z09-Z] awarded to AR. We also acknowledge funding from IISc, DBT-IISc partnership programme, and support from Dept. of Science and Technology, Govt. of India, to the Department of MRDG.

Author contributions: M.S., S.H.K. and A.R. conceived and designed experiments. M.S., S.H.K., S.A.B., G.S.S., S.P., N.D., S.R. and D.M.P. performed the experiments. M.S., N.D., S.R. and S.H.K

Journal of Cell Science • Accepted manuscript

performed the data analysis. M.S., S.H.K., N.D., S.R. and A.R. wrote the manuscript.

Ansieau, S., A.P. Morel, G. Hinkal, J. Bastid, and A. Puisieux. 2010. TWISTing an embryonic transcription factor into an oncoprotein. Oncogene. 29:3173-3184. Banerjee, P., H. Surendran, D.R. Chowdhury, K. Prabhakar, and R. Pal. 2016. Metformin mediated reversal of epithelial to mesenchymal transition is triggered by epigenetic changes in E-cadherin promoter. Journal of molecular medicine (Berlin, Germany). 94:1397-1409. Bungard, D., B.J. Fuerth, P.Y. Zeng, B. Faubert, N.L. Maas, B. Viollet, D. Carling, C.B. Thompson, R.G. Jones, and S.L. Berger. 2010. Signaling kinase AMPK activates stress-promoted transcription via histone H2B phosphorylation. Science. 329:12011205. Canto, C., and J. Auwerx. 2010. AMP-activated protein kinase and its downstream transcriptional pathways. Cellular and molecular life sciences : CMLS. 67:3407-3423. Chaffer, C.L., and R.A. Weinberg. 2011. A perspective on cancer cell metastasis. Science. 331:1559-1564. Chen, M.B., X.Y. Wu, J.H. Gu, Q.T. Guo, W.X. Shen, and P.H. Lu. 2011. Activation of AMP-activated protein kinase contributes to doxorubicin-induced cell death and apoptosis in cultured myocardial H9c2 cells. Cell Biochem Biophys. 60:311-322. Cheng, G.Z., W.Z. Zhang, M. Sun, Q. Wang, D. Coppola, M. Mansour, L.M. Xu, C. Costanzo, J.Q. Cheng, and L.H. Wang. 2008. Twist is transcriptionally induced by activation of STAT3 and mediates STAT3 oncogenic function. J Biol Chem. 283:14665-14673. Chiu, Y.C., D.C. Shieh, K.M. Tong, C.P. Chen, K.C. Huang, P.C. Chen, Y.C. Fong, H.C. Hsu, and C.H. Tang. 2009. Involvement of AdipoR receptor in adiponectin-induced motility and alpha2beta1 integrin upregulation in human chondrosarcoma cells. Carcinogenesis. 30:1651-1659. Chou, C.C., K.H. Lee, I.L. Lai, D. Wang, X. Mo, S.K. Kulp, C.L. Shapiro, and C.S. Chen. 2014. AMPK reverses the mesenchymal phenotype of cancer cells by targeting the Akt-MDM2-Foxo3a signaling axis. Cancer Res. 74:4783-4795. Cool, B., B. Zinker, W. Chiou, L. Kifle, N. Cao, M. Perham, R. Dickinson, A. Adler, G. Gagne, R. Iyengar, G. Zhao, K. Marsh, P. Kym, P. Jung, H.S. Camp, and E. Frevert. 2006. Identification and characterization of a small molecule AMPK activator that treats key components of type 2 diabetes and the metabolic syndrome. Cell metabolism. 3:403-416. Cufi, S., A. Vazquez-Martin, C. Oliveras-Ferraros, B. Martin-Castillo, J. Joven, and J.A. Menendez. 2010. Metformin against TGFbeta-induced epithelial-to-mesenchymal transition (EMT): from cancer stem cells to aging-associated fibrosis. Cell cycle. 9:4461-4468. Elenbaas, B., L. Spirio, F. Koerner, M.D. Fleming, D.B. Zimonjic, J.L. Donaher, N.C. Popescu, W.C. Hahn, and R.A. Weinberg. 2001. Human breast cancer cells generated by oncogenic transformation of primary mammary epithelial cells. Genes Dev. 15:5065. Foretz, M., S. Hébrard, J. Leclerc, E. Zarrinpashneh, M. Soty, G. Mithieux, K. Sakamoto, F. Andreelli, and B. Viollet. 2010. Metformin inhibits hepatic gluconeogenesis in mice independently of the LKB1/AMPK pathway via a decrease in hepatic energy state. The Journal of Clinical Investigation. 120:2355-2369. Han, B., H. Cui, L. Kang, X. Zhang, Z. Jin, L. Lu, and Z. Fan. 2015. Metformin inhibits thyroid cancer cell growth, migration, and EMT through the mTOR pathway. Tumor Biology. 36:6295.

Journal of Cell Science • Accepted manuscript

References:

Journal of Cell Science • Accepted manuscript

Hardie, D.G., F.A. Ross, and S.A. Hawley. 2012. AMPK: a nutrient and energy sensor that maintains energy homeostasis. Nat Rev Mol Cell Biol. 13:251-262. He, K., X. Guo, Y. Liu, J. Li, Y. Hu, D. Wang, and J. Song. 2016. TUFM downregulation induces epithelial-mesenchymal transition and invasion in lung cancer cells via a mechanism involving AMPK-GSK3beta signaling. Cellular and molecular life sciences : CMLS. 73:2105-2121. Hindupur, S.K., S.A. Balaji, M. Saxena, S. Pandey, G.S. Sravan, N. Heda, M.V. Kumar, G. Mukherjee, D. Dey, and A. Rangarajan. 2014. Identification of a novel AMPK-PEA15 axis in the anoikis-resistant growth of mammary cells. Breast Cancer Res. 16:420. Hong, J., K. Kandasamy, M. Marimuthu, C.S. Choi, and S. Kim. 2011a. Electrical cellsubstrate impedance sensing as a non-invasive tool for cancer cell study. Analyst. 136:237-245. Hong, J., J. Zhou, J. Fu, T. He, J. Qin, L. Wang, L. Liao, and J. Xu. 2011b. Phosphorylation of serine 68 of Twist1 by MAPKs stabilizes Twist1 protein and promotes breast cancer cell invasiveness. Cancer Res. 71:3980-3990. Huang, R.Y.-J., P. Guilford, and J.P. Thiery. 2012. Early events in cell adhesion and polarity during epithelial-mesenchymal transition. Journal of Cell Science. 125:4417-4422. Huang, X., S. Wullschleger, N. Shpiro, V.A. McGuire, K. Sakamoto, Y.L. Woods, W. McBurnie, S. Fleming, and D.R. Alessi. 2008. Important role of the LKB1-AMPK pathway in suppressing tumorigenesis in PTEN-deficient mice. Biochem J. 412:211221. Jeon, S.M., and N. Hay. 2012. The dark face of AMPK as an essential tumor promoter. Cellular logistics. 2:197-202. Jeon, S.M., and N. Hay. 2015. The double-edged sword of AMPK signaling in cancer and its therapeutic implications. Archives of pharmacal research. 38:346-357. Jones, R.G., D.R. Plas, S. Kubek, M. Buzzai, J. Mu, Y. Xu, M.J. Birnbaum, and C.B. Thompson. 2005. AMP-activated protein kinase induces a p53-dependent metabolic checkpoint. Mol Cell. 18:283-293. Kalender, A., A. Selvaraj, S.Y. Kim, P. Gulati, S. Brule, B. Viollet, B.E. Kemp, N. Bardeesy, P. Dennis, J.J. Schlager, A. Marette, S.C. Kozma, and G. Thomas. 2010. Metformin, independent of AMPK, inhibits mTORC1 in a rag GTPase-dependent manner. Cell metabolism. 11:390-401. Kim, E.K., J.M. Park, S. Lim, J.W. Choi, H.S. Kim, H. Seok, J.K. Seo, K. Oh, D.S. Lee, K.T. Kim, S.H. Ryu, and P.G. Suh. 2011. Activation of AMP-activated protein kinase is essential for lysophosphatidic acid-induced cell migration in ovarian cancer cells. J Biol Chem. 286:24036-24045. Kim, H.S., M.J. Kim, E.J. Kim, Y. Yang, M.S. Lee, and J.S. Lim. 2012. Berberine-induced AMPK activation inhibits the metastatic potential of melanoma cells via reduction of ERK activity and COX-2 protein expression. Biochemical pharmacology. 83:385-394. Kim, J., G. Yang, Y. Kim, J. Kim, and J. Ha. 2016. AMPK activators: mechanisms of action and physiological activities. Experimental & molecular medicine. 48:e224. Liang, J., S.H. Shao, Z.X. Xu, B. Hennessy, Z. Ding, M. Larrea, S. Kondo, D.J. Dumont, J.U. Gutterman, C.L. Walker, J.M. Slingerland, and G.B. Mills. 2007. The energy sensing LKB1-AMPK pathway regulates p27(kip1) phosphorylation mediating the decision to enter autophagy or apoptosis. Nat Cell Biol. 9:218-224. Lin, H., N. Li, H. He, Y. Ying, S. Sunkara, L. Luo, N. Lv, D. Huang, and Z. Luo. 2015. AMPK Inhibits the Stimulatory Effects of TGF-beta on Smad2/3 Activity, Cell Migration, and Epithelial-to-Mesenchymal Transition. Molecular pharmacology. 88:1062-1071.

Journal of Cell Science • Accepted manuscript

Liu, X., R.R. Chhipa, I. Nakano, and B. Dasgupta. 2014. The AMPK inhibitor compound C is a potent AMPK-independent antiglioma agent. Mol Cancer Ther. 13:596-605. Lombaerts, M., T. van Wezel, K. Philippo, J.W. Dierssen, R.M. Zimmerman, J. Oosting, R. van Eijk, P.H. Eilers, B. van de Water, C.J. Cornelisse, and A.M. Cleton-Jansen. 2006. E-cadherin transcriptional downregulation by promoter methylation but not mutation is related to epithelial-to-mesenchymal transition in breast cancer cell lines. British journal of cancer. 94:661-671. Mani, S.A., W. Guo, M.J. Liao, E.N. Eaton, A. Ayyanan, A.Y. Zhou, M. Brooks, F. Reinhard, C.C. Zhang, M. Shipitsin, L.L. Campbell, K. Polyak, C. Brisken, J. Yang, and R.A. Weinberg. 2008. The epithelial-mesenchymal transition generates cells with properties of stem cells. Cell. 133:704-715. Martinelli, R., M. Gegg, R. Longbottom, P. Adamson, P. Turowski, and J. Greenwood. 2009. ICAM-1-mediated endothelial nitric oxide synthase activation via calcium and AMPactivated protein kinase is required for transendothelial lymphocyte migration. Mol Biol Cell. 20:995-1005. McGee, S.L., B.J. van Denderen, K.F. Howlett, J. Mollica, J.D. Schertzer, B.E. Kemp, and M. Hargreaves. 2008. AMP-activated protein kinase regulates GLUT4 transcription by phosphorylating histone deacetylase 5. Diabetes. 57:860-867. Mukherjee, S. 2010. The emperor of all maladies: A biography of cancer. New York, NY: Scribner. Mungai, P.T., G.B. Waypa, A. Jairaman, M. Prakriya, D. Dokic, M.K. Ball, and P.T. Schumacker. 2011. Hypoxia triggers AMPK activation through reactive oxygen species-mediated activation of calcium release-activated calcium channels. Molecular and cellular biology. 31:3531-3545. Nagata, D., M. Mogi, and K. Walsh. 2003. AMP-activated protein kinase (AMPK) signaling in endothelial cells is essential for angiogenesis in response to hypoxic stress. J Biol Chem. 278:31000-31006. Nakano, A., H. Kato, T. Watanabe, K.D. Min, S. Yamazaki, Y. Asano, O. Seguchi, S. Higo, Y. Shintani, H. Asanuma, M. Asakura, T. Minamino, K. Kaibuchi, N. Mochizuki, M. Kitakaze, and S. Takashima. 2010. AMPK controls the speed of microtubule polymerization and directional cell migration through CLIP-170 phosphorylation. Nat Cell Biol. 12:583-590. Ng, T.L., G. Leprivier, M.D. Robertson, C. Chow, M.J. Martin, K.R. Laderoute, E. Davicioni, T.J. Triche, and P.H. Sorensen. 2012. The AMPK stress response pathway mediates anoikis resistance through inhibition of mTOR and suppression of protein synthesis. Cell death and differentiation. 19:501-510. Nieto, M.A., R.Y. Huang, R.A. Jackson, and J.P. Thiery. 2016. Emt: 2016. Cell. 166:21-45. Owen, M.R., E. Doran, and A.P. Halestrap. 2000. Evidence that metformin exerts its antidiabetic effects through inhibition of complex 1 of the mitochondrial respiratory chain. Biochemical Journal. 348:607-614. Pham, C.G., C. Bubici, F. Zazzeroni, J.R. Knabb, S. Papa, C. Kuntzen, and G. Franzoso. 2007. Upregulation of Twist-1 by NF-kappaB blocks cytotoxicity induced by chemotherapeutic drugs. Mol Cell Biol. 27:3920-3935. Polyak, K., and R.A. Weinberg. 2009. Transitions between epithelial and mesenchymal states: acquisition of malignant and stem cell traits. Nat Rev Cancer. 9:265-273. Qu, C., W. Zhang, G. Zheng, Z. Zhang, J. Yin, and Z. He. 2014. Metformin reverses multidrug resistance and epithelial-mesenchymal transition (EMT) via activating AMP-activated protein kinase (AMPK) in human breast cancer cells. Molecular and cellular biochemistry. 386:63-71.

Journal of Cell Science • Accepted manuscript

Rios, M., M. Foretz, B. Viollet, A. Prieto, M. Fraga, J.A. Costoya, and R. Senaris. 2013. AMPK Activation by Oncogenesis Is Required to Maintain Cancer Cell Proliferation in Astrocytic Tumors. Cancer Res. 73:2628-2638. Sanders, M.J., Z.S. Ali, B.D. Hegarty, R. Heath, M.A. Snowden, and D. Carling. 2007. Defining the mechanism of activation of AMP-activated protein kinase by the small molecule A-769662, a member of the thienopyridone family. J Biol Chem. 282:3253932548. Sommers, C.L., E.W. Thompson, J.A. Torri, R. Kemler, E.P. Gelmann, and S.W. Byers. 1991. Cell adhesion molecule uvomorulin expression in human breast cancer cell lines: relationship to morphology and invasive capacities. Cell growth & differentiation : the molecular biology journal of the American Association for Cancer Research. 2:365-372. Steinberg, G.R., and B.E. Kemp. 2009. AMPK in Health and Disease. Physiol Rev. 89:10251078. Sundararaman, A., U. Amirtham, and A. Rangarajan. 2016. Calcium-Oxidant Signaling Network Regulates AMP-activated Protein Kinase (AMPK) Activation upon Matrix Deprivation. J Biol Chem. 291:14410-14429. Tan, E.J., S. Thuault, L. Caja, T. Carletti, C.H. Heldin, and A. Moustakas. 2012. Regulation of transcription factor Twist expression by the DNA architectural protein high mobility group A2 during epithelial-to-mesenchymal transition. J Biol Chem. 287:7134-7145. Thiery, J.P., H. Acloque, R.Y. Huang, and M.A. Nieto. 2009. Epithelial-mesenchymal transitions in development and disease. Cell. 139:871-890. Wang, X., X. Pan, and J. Song. 2010. AMP-activated protein kinase is required for induction of apoptosis and epithelial-to-mesenchymal transition. Cell Signal. 22:1790-1797. Wang, Y., J. Liu, X. Ying, P.C. Lin, and B.P. Zhou. 2016. Twist-mediated Epithelialmesenchymal Transition Promotes Breast Tumor Cell Invasion via Inhibition of Hippo Pathway. Sci Rep. 6:24606. Witters, L.A., A.C. Nordlund, and L. Marshall. 1991. Regulation of intracellular acetyl-CoA carboxylase by ATP depletors mimics the action of the 5'-AMP-activated protein kinase. Biochem Biophys Res Commun. 181:1486-1492. Xiang, X., A.K. Saha, R. Wen, N.B. Ruderman, and Z. Luo. 2004. AMP-activated protein kinase activators can inhibit the growth of prostate cancer cells by multiple mechanisms. Biochem Biophys Res Commun. 321:161-167. Yang, J., S.A. Mani, J.L. Donaher, S. Ramaswamy, R.A. Itzykson, C. Come, P. Savagner, I. Gitelman, A. Richardson, and R.A. Weinberg. 2004. Twist, a master regulator of morphogenesis, plays an essential role in tumor metastasis. Cell. 117:927-939. Yang, J., S.A. Mani, and R.A. Weinberg. 2006. Exploring a new twist on tumor metastasis. Cancer Res. 66:4549-4552. Yang, J., and R.A. Weinberg. 2008. Epithelial-mesenchymal transition: at the crossroads of development and tumor metastasis. Developmental cell. 14:818-829. Yang, M.H., M.Z. Wu, S.H. Chiou, P.M. Chen, S.Y. Chang, C.J. Liu, S.C. Teng, and K.J. Wu. 2008. Direct regulation of TWIST by HIF-1alpha promotes metastasis. Nat Cell Biol. 10:295-305. Zhang, H., R.R. Singh, A.H. Talukder, and R. Kumar. 2006. Metastatic tumor antigen 3 is a direct corepressor of the Wnt4 pathway. Genes Dev. 20:2943-2948. Zhou, G., R. Myers, Y. Li, Y. Chen, X. Shen, J. Fenyk-Melody, M. Wu, J. Ventre, T. Doebber, N. Fujii, N. Musi, M.F. Hirshman, L.J. Goodyear, and D.E. Moller. 2001. Role of AMP-activated protein kinase in mechanism of metformin action. J Clin Invest. 108:1167-1174.

Journal of Cell Science • Accepted manuscript

Figures

Figure 1: AMPK activity modulates EMT marker expression

Journal of Cell Science • Accepted manuscript

(A-F) Cancer cell lines T47D, MCF7 (A and B), A549 (C and D) and MDA-MB-231 (E and F) were cultured in the presence of 100 μM AMPK activator (A769662), 10 μM AMPK inhibitor (Compound C) or DMSO (vehicle control) as specified for 48 hours. Thereafter, cells were harvested and subjected to qRT-PCR analysis for the specified transcripts. Graphs in A, C and E represent fold change in gene expression normalized to β2m; error bars represent SEM, n=3. Parallel dishes were fixed using 4% PFA to undertake immunocytochemistry analysis for the specified proteins (B, D and F). Photomicrographs show representative fluorescent images taken using Olympus 1X71 microscope at 20X magnification and processed using ImageJ software; Hoechst 33342 was used for nuclear staining. Graphs in B, D and F represent relative fluorescence intensity measurements normalized to number of nuclei. Error bars represent SEM, n=3.

(A) Scratch assay was performed with BT474 cells after treatment with 100µM AMPK activator (A769662) or DMSO (vehicle control). Graph represents time kinetics of wound confluence percentage, calculated by Incucyte ZOOM software. Error bars represent SEM, n=2.

Journal of Cell Science • Accepted manuscript

Figure 2: AMPK induces a functional EMT

(B) Scratch assay was performed with A549 cells after treatment with 100µM AMPK activator (A769662), 10 µM AMPK inhibitor (Compound C) or DMSO (vehicle control). Graph represents time kinetics of wound confluence percentage, calculated by Incucyte ZOOM software. Error bars represent SEM, n=2. (C and D) Scratch assay was performed with MDA-MB-231 (C) and MDA-MB-435S (D) cells after treatment with 100µM AMPK activator (A769662), 10 µM AMPK inhibitor (Compound C) or DMSO (vehicle control). Graphs represent time kinetics of wound confluence percentage, calculated by Incucyte ZOOM software. Error bars represent SEM, n=3 and 2 respectively. (E) Scratch assay was performed with A549 cells stably expressing inducible shRNA against AMPKα2 (clone #4) cultured with and without doxycycline. Graph represents time kinetics of wound confluence percentage, calculated by Incucyte ZOOM software. Error bars represent SEM, n=2. (F) Scratch assay was performed with MDA-MB-231 cells stably expressing scrambled shRNA or AMPKα2 shRNA. Graph represents time kinetics of wound confluence percentage, calculated by Incucyte ZOOM software. Error bars represent SEM, n=2. (G) Scratch assay was performed with MDA-MB-435S cells transfected with control siRNA or AMPKα2 siRNA. Graphs represent the distance migrated between the cell edges by 48 hours, calculated by ImageJ software. Error bars represent SEM, n=3. (H-I) Boyden chamber invasion assay was performed for 24 hours with MDA-MB-231 (H) and MDA-MB-435S cells (I) treated with 100 µM AMPK activator (A769662), 10 µM AMPK inhibitor (Compound C) or DMSO (vehicle control). Graphs represent number of cells invaded per field (9-10 fields were counted per experiment). Error bars represent SEM, n=3.

For all scratch/ migration assay experiments above, cells were pretreated with 10 g/ml Mitomycin C for 2 hours before the scratch/ wound was made. For all migration assay experiments performed using Incucyte ZOOM, 10X magnification (wide mode) was used.

Journal of Cell Science • Accepted manuscript

(J-K) Boyden chamber invasion assay was performed for 24 hours with MDA-MB-231 cells stably expressing scrambled shRNA or AMPKα2 shRNA; n=2 (J) or MDA-MB-435S cells transfected with control siRNA or AMPKα2 siRNA; n=3 (K). Graphs represent number of cells invaded per field (9-10 fields were counted per experiment). Error bars represent SEM.

Journal of Cell Science • Accepted manuscript

Figure 3: AMPK is necessary for EMT induction by different upstream stimuli (A) BT474 cells stably expressing scrambled shRNA or AMPKα2 shRNA were cultured in the presence of 150 µM CoCl2 for 48 hours. Thereafter, cells were harvested and immunoblot analysis was undertaken. Graph represents densitometric quantification of the specified proteins normalized to α-Tubulin; error bars represent SEM; n=2. (B) A549 cells stably expressing inducible shRNA against AMPKα2 (clone #4) were cultured with and without doxycycline in ambient oxygen (20% O2) or in hypoxic (3% O2) conditions in a trigas incubator for 48 hours. Thereafter, cells were harvested and subjected to immunoblot analysis. Graph represents densitometric quantification of the specified proteins normalized to α-Tubulin; error bars represent SEM; n=2. (C) MDA-MB-435S cells were cultured in ambient oxygen (20% O2) or in hypoxic (3% O2) conditions in a trigas incubator in the presence of 10 µM AMPK inhibitor (Compound C) for 24 hours. Thereafter, cells were harvested and subjected to immunoblot analysis. Graph represents densitometric quantification of the specified proteins normalized to α-Tubulin; error bars represent SEM; n=4.

Journal of Cell Science • Accepted manuscript

(D) MDA-MB-435S cells were cultured with 5 ng/ml TGFβ in the presence or absence of 10 µM AMPK inhibitor (Compound C) for 24 hours. Thereafter, immunoblot analysis was undertaken for the specified proteins. Graph represents densitometric quantification of the specified proteins normalized to α-Tubulin; error bars represent SEM; n=4.

Figure 4: AMPK knockdown impairs metastasis of breast cancer cells

Journal of Cell Science • Accepted manuscript

(A and B) MDA-MB-231 and MDA-MB-435S cells stably expressing scrambled shRNA or AMPKα2 shRNA were injected through tail vein of nude mice and scored for lung metastasis. Photomicrographs show histologic sectioning of representative lung metastases originating from MDA-MB-231 (A) and MDA-MB-435S (B) cells (Magnification, x20; H&E staining). Graphs represent incidence of lung metastases found in mice in experiment A and B; n=5 mice for each group.

Journal of Cell Science • Accepted manuscript

(A) Various cancer cell lines (MCF7, BT474, A549, MDA-MB-231 and MDA-MB-435S) were cultured in the presence of 100 µM AMPK activator (A769662) or DMSO (vehicle control) for 48 hours. Thereafter, cells were harvested for RNA isolation and subjected to qPCR analysis for Twist1 gene expression. Graph represents fold changes (normalized to housekeeping gene β2M) of Twist1 gene expression in the specified cell lines; error bars represent SEM, n=3. (B) MDA-MB-435S cells were cultured in the presence of 100 μM AMPK activator (A769662) or DMSO (vehicle control) for 48 hours. Thereafter, cells were harvested and subjected to immunoblot analysis for Twist1 and α-Tubulin. Graph represents densitometric quantification of the Twist1 protein normalized to α-Tubulin; error bars represent SEM, n=3. (C) MDA-MB-231 cells were cultured in the presence of 100 µM AMPK activator (A769662) or DMSO (vehicle control) for 24 hours. Thereafter, the dishes were subjected to immunocytochemistry analysis for Twist1 protein. Photomicrographs show representative fluorescent images taken at 20X magnification using Olympus 1X71 microscope; Hoechst 33342 was used for nuclear staining, and was pseudo colored green for representation. Graph represents cytoplasmic and nuclear Fluorescence intensity (FI) measurements per cell, n=30 cells quantified in each of the three biological replicates. (D) qPCR analysis for Twist1 gene expression in A549, MDA-MB-231 and MDA-MB-435S cells cultured in the presence of 10 μM AMPK inhibitor (Compound C) or DMSO (vehicle control) for 48 hours. Graph represents fold change (normalized to β2M); error bar represents SEM, n=3. (E) qPCR analysis for Twist1 gene expression in MDA-MB-231 and MDA-MB-435S cells stably expressing scrambled shRNA or AMPKα2 shRNA. Graph represents fold change (normalized to β2M); error bar represents SEM, n=3. (F and G) MDA-MB-231 and BT474 cells were cultured in the presence of 10 µM AMPK inhibitor (Compound C) with 150 µM CoCl2 (F) or TGFβ (G) for 48 hours. Thereafter, cells were harvested and quantified for Twist1 mRNA levels by qPCR. Graph represents fold change (normalized to β2M); error bar represents SEM, n=3. (H) Quantification of Twist1 mRNA levels by qPCR in MDA-MB-231 cells stably expressing GFP shRNA or Twist1 shRNA. Graph represents fold change of Twist1 gene expression (normalized to β2M); error bars represent SEM, n=3. (I) MDA-MB-231 cells stably expressing GFP shRNA or Twist1 shRNA were cultured in the presence of 100 µM AMPK activator (A769662) or DMSO (vehicle control) for 48 hours. Thereafter, the cells were harvested and qPCR analysis was undertaken for the specified transcripts. Graph represents relative gene expression of the specified genes (normalized to β2M expression); error bars represent SEM, n=4. (J-K) MDA-MB-231 cells stably expressing GFP shRNA or Twist1 shRNA were cultured in the presence of 100 µM AMPK activator (A769662) or DMSO (vehicle control). (J) Graph represents the distance migrated between the cell edges in a scratch assay in 48 hours, calculated by Image J software; error bar represents SEM, n=3. (K) The specified cells treated similarly were also subjected to Boyden chamber invasion assay. Graph represents total number cells invaded per field; error bar represents SEM, n=4.

Journal of Cell Science • Accepted manuscript

Figure 5: AMPK regulates Twist1 expression

Journal of Cell Science • Accepted manuscript

Figure 6: AMPK mediates EMT via increased Twist1 stability and its nuclear translocation (A) MDA-MB-231 cells were cultured in the presence of 100 μM AMPK activator (A769662) or DMSO (vehicle control) for 48 hours. Thereafter, cells were harvested, cytoplasmic and nuclear proteins were fractionated and subjected to immunoblot analysis for Twist1, pAMPKα, Lamin and α-Tubulin, n=3. Graph represents densitometric quantification of nuclear and cytoplasmic Twist1 levels; error bar represents SEM, n=4. (B) MCF7 and MDA-MB-231 cells were transiently transfected with Twist1 tagged to GFP and subsequently cultured in the presence of 100 µM AMPK activator (A769662) or DMSO (vehicle control) for 24 hours. Thereafter, the cells were fixed and localization of GFP fluorescence was analyzed. Photomicrographs show representative fluorescent images taken at 20X magnification using Olympus 1X71 microscope; Hoechst 33342 was used for nuclear staining, and was pseudo colored red for representation. Graph represents nuclear and cytoplasmic localization of Twist1-GFP, n=30 cells quantified in each of the three biological replicates. (C) MDA-MB-435S cells were cultured in the presence of 10 μM AMPK inhibitor (Compound C) or DMSO (vehicle control) for 6 hours followed by inhibition of protein translation by cycloheximide treatment for 0, 2, 4 and 6 hours. Cells were harvested and subjected to immunoblot analysis for Twist1 and α-Tubulin. Graph represents densitometric quantification of Twist1 levels normalized to α-Tubulin.

Journal of Cell Science • Accepted manuscript

(D) Model: Stresses faced by tumor cells in the primary tumor environment, such as, hypoxia, nutrient deprivation, and matrix-detachment, activate AMPK, which in turn induces EMT, allowing tumor cells to initiate the invasion-metastasis cascade. Re-attachment at distant sites causes inhibition of AMPK signaling, enabling reversal of EMT and establishment of secondary tumors.

Journal of Cell Science • Supplementary information

J. Cell Sci. 131: doi:10.1242/jcs.208314: Supplementary information

Journal of Cell Science • Supplementary information

J. Cell Sci. 131: doi:10.1242/jcs.208314: Supplementary information

J. Cell Sci. 131: doi:10.1242/jcs.208314: Supplementary information

Figure S1: (S1A) Cancer cell lines MDA-MB-435S, MDA-MB-231, BT474, MCF7, T47D and A549 were subjected to immunoblot analysis for the detection of endogenous levels of the specified proteins, n=3. (S1B) T47D and MCF7 cells were cultured in the presence of 100 µM AMPK activator (A769662) or DMSO (vehicle control) for 48 hours, followed by immunoblot analysis for the specified proteins, n=4. Band intensities are represented below the relevant panels. (S1 C, D and E) BT474 cells were cultured in the presence of 100 µM AMPK activator (A769662) or DMSO (vehicle control) for 48 hours. Thereafter, qPCR analysis (C) was undertaken for the specified transcripts, n=3. Parallel dishes were subjected to immunoblot analysis for the specified proteins, n=3 (D), and immunocytochemical analysis for E-cadherin protein (E). Photomicrographs show representative fluorescent images taken at 20X magnification using Olympus 1X71 microscope; Hoechst 33342 was used for nuclear staining. Graph represents relative fluorescence intensity (RFI) measurements normalized to number of nuclei, n=3. (S1 F and G) A549 (F) and MDA-MB-231 (G) cells were cultured in the presence of 100 μM AMPK activator (A769662), 10 μM AMPK inhibitor (Compound C) or DMSO (vehicle control) for 48 hours, followed by immunoblot analysis for the specified proteins; n=4. Same DMSO treated samples were used as a control for A769662 or Compound C treated samples, but developed with different exposures to capture the differences. Band intensities are represented below the relevant panels. (S1H) MDA-MB-435S cells were cultured in the presence of 100 μM AMPK activator (A769662), 10 μM AMPK inhibitor (Compound C) or DMSO (vehicle control) for 48 hours followed by immunoblot analysis for the specified proteins; n=4.

(S1J and K) A549 cells stably expressing inducible shRNA against AMPKα2 (clone#1) cultured with and without doxycycline for 48 hours, were harvested and subjected to immunoblot analysis for the specified proteins, n=2 (J). Parallel dishes were subjeted to immunocytochemical analsysis for E-cadherin and vimentin protein, n=2 (K). Photomicrographs show representative fluorescent images taken at 20X magnification using Olympus 1X71 microscope; Hoechst 33342 was used for nuclear staining. (S1L) A549 cells stably expressing inducible shRNA against AMPKα2 (clone#4) cultured with and without doxycycline for 48 hours, were harvested and subjected to immunoblot analysis for the specified proteins, n=2 (S1M and N) MDA-MB-231 cells stably expressing scrambled shRNA or AMPKα2 shRNA were harvested and subjected to immunoblot analysis for the specified proteins, n=3 (K). Parallel dishes were subjected to immunocytochemistry analysis for Vimentin, n=3 (L). (S1O) MDA-MB-231 cells stably expressing inducible shRNA against AMPKα2 (clone #4) cultured with and without doxycycline for 48 hours, were harvested and subjected to immunoblot analysis for the specified proteins, n=2. (S1P) MDA-MB-435S cells transfected with control siRNA or AMPKα2 siRNA for 48 hours were harvested and subjected to immunoblot analysis for the specified proteins; n=4.

Journal of Cell Science • Supplementary information

(S1I) BT474 cells stably expressing scrambled shRNA or AMPKα2 shRNA were harvested and subjected to immunoblot analysis for the specified proteins, n=2.

Journal of Cell Science • Supplementary information

J. Cell Sci. 131: doi:10.1242/jcs.208314: Supplementary information

J. Cell Sci. 131: doi:10.1242/jcs.208314: Supplementary information

Figure S2: (S2A) Representative phase contrast microscopic images of the scratch assay performed with BT474 cells treated with 100 µM AMPK activator (A769662) or DMSO (vehicle control) taken between 0 and 24 hours. (S2B) Representative phase contrast microscopic images of the scratch assay performed with A549 cells treated with 100 µM AMPK activator (A769662), 10 µM AMPK inhibitor (Compound C) or DMSO (vehicle control) taken between 0 and 24 hours. (S2 C and D) Representative phase contrast microscopic images of the scratch assay performed with MDA-MB231 (C) and MDA-MB-435S (D) cells treated with 100 µM AMPK activator (A769662), 10 µM AMPK inhibitor (Compound C) or DMSO (vehicle control) taken between 0 and 24 hours. (S2E) ECIS-based migration assay was performed with MDA-MB-435S cells treated with 100 μM AMPK activator (A769662), 10 μM AMPK inhibitor (Compound C) or DMSO (vehicle control) for 24 hours. Graph shows a representative experiment of migration assessed over 24 hours from the time of wound generation; n=4. (S2F) Representative phase contrast microscopic images of the scratch assay performed with A549 cells stably expressing inducible shRNA against AMPKα2 (clone #4) cultured with and without doxycyclin taken between 0 and 24 hours. (S2G) Representative phase contrast microscopic images of the scratch assay performed with MDA-MB-231 cells stably expressing scrambled shRNA or AMPKα2 shRNA taken between 0 and 24 hours.

(S2I) Representative bright field images of the invasion assay performed with MDA-MB-231 cells stably expressing scrambled shRNA or AMPKα2 shRNA (in Fig. 2J).

Journal of Cell Science • Supplementary information

(S2H) Representative phase contrast microscopic images of the scratch assay performed with MDA-MB-435S cells transfected with control siRNA or AMPKα2 siRNA taken between 0 and 48 hours.

Journal of Cell Science • Supplementary information

J. Cell Sci. 131: doi:10.1242/jcs.208314: Supplementary information

J. Cell Sci. 131: doi:10.1242/jcs.208314: Supplementary information

Figure S3: (S3A) MDA-MB-231 cells stably expressing scrambled shRNA or AMPKα2 shRNA were cultured in the presence of 150 µM CoCl2 for 48 hours. Thereafter, cells were harvested and immunoblot analysis was undertaken. Graph represents densitometric quantification of the specified proteins normalized to α-Tubulin; error bars represent SEM, n=4. (S3B) MDA-MB 231 cells were cultured with 150 µM CoCl2 in the presence or absence of 10 µM AMPK inhibitor (Compound C) for 48 hours. Thereafter, cells were harvested for RNA isolation and subjected to qRT-PCR for specified transcripts. Graph represents fold change in gene expression normalized to β2m; error bars represent SEM, n=3. (S3C) A549 cells were cultured with 5 ng/ml TGFβ in the presence or absence of 10 µM AMPK inhibitor (Compound C) for 24 hours. Thereafter, cells were harvested and immunoblot analysis was undertaken for EMT marker proteins. Band intensities are represented below the relevant panels; n=3. (S3D) MDA-MB-231 cells were cultured with 5 ng/ml TGFβ in the presence or absence of 10 µM AMPK inhibitor (Compound C) for 24 hours. Thereafter, cells were harvested for RNA isolation and subjected to qRTPCR for specified transcripts. Graph represents fold change in gene expression normalized to β2m; error bars represent SEM, n=3. (S3E) MDA-MB-231 cells stably expressing scrambled shRNA or AMPKα2 shRNA were cultured in the presence of 5 ng/ml of TGFβ for 24 hours. Thereafter, the cells were harvested and subjected to immunoblot analysis. Graph represents densitometric quantification of the specified proteins normalized to α-Tubulin; error bars represent SEM, n=3.

Journal of Cell Science • Supplementary information

(S3F) HMLER cells (HMLE cells overexpressing oncogenic Ras) were cultured in the presence of 10 μM AMPK inhibitor (Compound C) or DMSO (vehicle control) for 24 hours. Thereafter, cells were harvested and subjected to immunoblot analysis for the specified proteins; n=3.

Figure S4: (S4 A and B) Graphs represent the number of macroscopic lung nodules counted on the surface of lungs of mice injected with MDA-MB-231 (A) or MDA-MB-435S (B) cells stably expressing scrambled shRNA or AMPKα2 shRNA in Fig. 4. Each data point represents one mouse. (S4C) Graph represents proliferation assay using MTT performed for MDA-MB-231 stably expressing scrambled shRNA or AMPKα2 shRNA; error bars represent SEM; n=3.

Journal of Cell Science • Supplementary information

J. Cell Sci. 131: doi:10.1242/jcs.208314: Supplementary information

Journal of Cell Science • Supplementary information

J. Cell Sci. 131: doi:10.1242/jcs.208314: Supplementary information

J. Cell Sci. 131: doi:10.1242/jcs.208314: Supplementary information

Figure S5: (S5A) MDA-MB-231 and BT474 cells were cultured in the presence of 100 μM AMPK activator (A769662) or DMSO (vehicle control) for 48 hours. Thereafter, cells were harvested and subjected to immunoblot analysis for Twist1 and α-Tubulin. Graph represents densitometric quantification of the Twist1 protein normalized to α-Tubulin; error bars represent SEM, n=4. (S5B) MDA-MB-231 cells stably expressing GFP shRNA or Twist1 shRNA were subjected to immunoblot analysis for Twist1 and α-Tubulin, n=3. (S5C) MDA-MB-231 cells stably expressing GFP shRNA or Twist1 shRNA were subjected to immunocytochemistry analysis for Twist1 protein. Photomicrograph shows representative fluorescent image taken at 20X magnification; Hoechst 33342 was used for nuclear staining, n=3. (S5D) MDA-MB-435S cells were transfected with control siRNA or Twist1 siRNA. After 48 hours of transfection, cells were harvested and subjected to semi-quantitative PCR analysis for the specified transcripts, n=4. (S5E) MDA-MB-435S cells transfected with control siRNA or Twist1 siRNA were cultured in the presence of 100 µM AMPK activator (A769662) or DMSO (vehicle control) for 48 hours. Thereafter, the cells were harvested and subjected to immunoblot analysis for the specified proteins; n=4.

Journal of Cell Science • Supplementary information

(S5F) Photomicrographs show representative phase contrast images taken at 0 and 48 hours of scratch assay performed with MDA-MB-231 cells stably expressing GFP shRNA or Twist1 shRNA cultured in the presence of 100 μM AMPK activator (A769662) or DMSO (vehicle control) for 48 hours. Magnification, x20; n=3.

J. Cell Sci. 131: doi:10.1242/jcs.208314: Supplementary information

Figure S6:

Journal of Cell Science • Supplementary information

(S6) A schematic representation of EMT marker expression changes at protein level upon AMPK modulation across various cancer cell lines. The basal expression level is depicted based on Fig. S1A, and alteration in protein expression is modeled based on western blot quantifications in Fig. S1B, S1D, S1F, S1G and S1H. ‘N.D’ signifies ‘not detected’ and ‘nd’ signifies ‘not done’.

J. Cell Sci. 131: doi:10.1242/jcs.208314: Supplementary information

Table S1: Primer sequences Forward primer sequence 5’-GAGAACTTTGCCGTTGAAGC-3’ 5’-GACAATGCCCCTCAAGTGTT-3’ 5’-ACCCCACATCCTTCTCACTG-3’ 5’-CTTTTTCTTGCCCTCACTGC-3’ 5’-TGCACTGAGTGTGGAAAAGC-3’ 5’-GAATCTTACGGGAAGATCAACCC-3’ 5’-GGACAAGTGACCACAGGA-3’ 5’-GTCCGCAGTCTTACGAGGAG-3’ 5’-CCTGAATTGCTATGTGTCT-3’ 5’- TGCTCGAGATGTGATGAAGG -3’

Reverse primer sequence 5’-TCCAGCAGCTTCCTGTAGGT-3’ 5’-CCATTAAGCCGAGTGATGGT-3’ 5’-TACAAAAACCCACGCAGACA-3’ 5’-GCTTCGGAGTGAAGAAATGC-3’ 5’-TGGTGATGCTGAAAGAGACG-3’ 5’-CGCAGAGTCGGCTAAGGTG-3’ 5’-GGAGAAAATCAAGTCGTG-3’ 5’-CCAGCTTGAGGGTCTGAATC-3’ 5’-TGATGCTGCTTACATGTCT-3’ 5’- TCCCCTGTTGACTGGTCATT -3’

Journal of Cell Science • Supplementary information

Gene Vimentin N-cadherin Snai1 Slug Zeb1 Smad7 Hif1α Twist1 β2m HPRT