Metformin Activates AMPK through the Lysosomal Pathway

Metformin (N, N-dimethylbiguanide) is the most widely used first-line drug for treatment of type 2 diabetes (T2D). This anti-hyperglycemic drug offers clinical superiority over other glucose-lowering drugs, with little induction of hypoglycemia or weight gain and with the effects to reverse fatty liver, improve insulin sensitivity, and ameliorate cardiovascular dysfunctions associated with T2D (extensively reviewed by Foretz et al., 2014Foretz M. Guigas B. Bertrand L. Pollak M. Viollet B. Cell Metab. 2014; 20: 953-966Abstract Full Text Full Text PDF PubMed Scopus (804) Google Scholar). Administration of metformin to nematodes (C. elegans) and mice gave rise to extended lifespan and health span (Barzilai et al., 2016Barzilai N. Crandall J.P. Kritchevsky S.B. Espeland M.A. Cell Metab. 2016; 23: 1060-1065Abstract Full Text Full Text PDF PubMed Scopus (551) Google Scholar). Moreover, cancer incidence was found to be decreased in patients treated with metformin (Foretz et al., 2014Foretz M. Guigas B. Bertrand L. Pollak M. Viollet B. Cell Metab. 2014; 20: 953-966Abstract Full Text Full Text PDF PubMed Scopus (804) Google Scholar). Based on the promising research results, a clinical trial named TAME (Targeting Aging with Metformin; http://www.afar.org/natgeo) has been proposed to test whether metformin can delay the onset of age-related diseases and conditions, including cancer, cardiovascular disease, and Alzheimer’s disease. The AMP-activated protein kinase (AMPK) is a master controller of various metabolic pathways (Hardie, 2014Hardie D.G. Cell Metab. 2014; 20: 939-952Abstract Full Text Full Text PDF PubMed Scopus (404) Google Scholar). The beneficial effects of metformin can be attributed, at least in part, to its cause of AMPK activation (Foretz et al., 2014Foretz M. Guigas B. Bertrand L. Pollak M. Viollet B. Cell Metab. 2014; 20: 953-966Abstract Full Text Full Text PDF PubMed Scopus (804) Google Scholar), which depends on LKB1, the liver kinase B1 (Shaw et al., 2005Shaw R.J. Lamia K.A. Vasquez D. Koo S.H. Bardeesy N. Depinho R.A. Montminy M. Cantley L.C. Science. 2005; 310: 1642-1646Crossref PubMed Scopus (1554) Google Scholar). For example, activated AMPK is indispensable for the attenuation of hepatic steatosis and atherosclerosis after metformin treatment (Li et al., 2011Li Y. Xu S. Mihaylova M.M. Zheng B. Hou X. Jiang B. Park O. Luo Z. Lefai E. Shyy J.Y. et al.Cell Metab. 2011; 13: 376-388Abstract Full Text Full Text PDF PubMed Scopus (1133) Google Scholar). One recent study further revealed that a single mutation on the AMPK-mediated phosphorylation sites of ACC1/2 strongly blocks the metformin-improved insulin action and glucose tolerance in diabetic mice (Fullerton et al., 2013Fullerton M.D. Galic S. Marcinko K. Sikkema S. Pulinilkunnil T. Chen Z.P. O’Neill H.M. Ford R.J. Palanivel R. O’Brien M. et al.Nat. Med. 2013; 19: 1649-1654Crossref PubMed Scopus (560) Google Scholar). Similarly, metformin retards aging in C. elegans in an AMPK-dependent manner (reviewed by Burkewitz et al., 2014Burkewitz K. Zhang Y. Mair W.B. Cell Metab. 2014; 20: 10-25Abstract Full Text Full Text PDF PubMed Scopus (288) Google Scholar). However, the mechanism for metformin to activate AMPK remains unclear and even controversial. It was shown that metformin treatment increases cellular levels of AMP through inhibiting complex I of the electron transport chain, by which ATP synthesis is uncoupled, leading to a drop in cellular ATP concentration and hence an increase of AMP levels (Foretz et al., 2014Foretz M. Guigas B. Bertrand L. Pollak M. Viollet B. Cell Metab. 2014; 20: 953-966Abstract Full Text Full Text PDF PubMed Scopus (804) Google Scholar). However, some studies failed to detect the alteration of AMP levels in cultured cell lines treated with metformin (He and Wondisford, 2015He L. Wondisford F.E. Cell Metab. 2015; 21: 159-162Abstract Full Text Full Text PDF PubMed Scopus (282) Google Scholar). In addition, recent studies showed that treatment of primary hepatocytes with clinically relevant concentrations of metformin, ∼70 μM (detected in portal vein after a therapeutic dose), efficiently activates AMPK without disrupting energy state (He and Wondisford, 2015He L. Wondisford F.E. Cell Metab. 2015; 21: 159-162Abstract Full Text Full Text PDF PubMed Scopus (282) Google Scholar). Similar observations were also obtained in the liver of mice after chronic administration of 50 mg/kg of metformin, within the clinical dose range (He and Wondisford, 2015He L. Wondisford F.E. Cell Metab. 2015; 21: 159-162Abstract Full Text Full Text PDF PubMed Scopus (282) Google Scholar). Furthermore, some studies were unable to detect direct inhibition of metformin on complex I in isolated mitochondria, unless at high concentrations (∼5 mM) (He and Wondisford, 2015He L. Wondisford F.E. Cell Metab. 2015; 21: 159-162Abstract Full Text Full Text PDF PubMed Scopus (282) Google Scholar). It is therefore evident that a clear understanding on how metformin activates AMPK awaits further studies. The scaffold protein AXIN plays an essential role in glucose starvation-induced AMPK activation by co-translocating LKB1 to the surface of lysosome in that AXIN also docks onto the lysosomal v-ATPase-Ragulator complex (Zhang et al., 2014Zhang C.S. Jiang B. Li M. Zhu M. Peng Y. Zhang Y.L. Wu Y.Q. Li T.Y. Liang Y. Lu Z. et al.Cell Metab. 2014; 20: 526-540Abstract Full Text Full Text PDF PubMed Scopus (313) Google Scholar, Zhang et al., 2013Zhang Y.L. Guo H. Zhang C.S. Lin S.Y. Yin Z. Peng Y. Luo H. Shi Y. Lian G. Zhang C. et al.Cell Metab. 2013; 18: 546-555Abstract Full Text Full Text PDF PubMed Scopus (175) Google Scholar). It must be noted that the v-ATPase-Ragulator complex must be primed by glucose starvation or treatment of the v-ATPase inhibitor concanamycin A (conA), likely through conformational change upon glucose starvation, for AXIN to anchor onto (Zhang et al., 2014Zhang C.S. Jiang B. Li M. Zhu M. Peng Y. Zhang Y.L. Wu Y.Q. Li T.Y. Liang Y. Lu Z. et al.Cell Metab. 2014; 20: 526-540Abstract Full Text Full Text PDF PubMed Scopus (313) Google Scholar). Here, we first explored whether AXIN is required for metformin-mediated AMPK activation by examining AMPK activation in the liver of metformin-treated AXIN-liver-specific knockout mice (AXINLKO). Protocols for animal experiments were approved by the Institutional Animal Care and the Animal Committee of Xiamen University. The AXINLKO mice were intraperitoneally injected with 50 mg/kg/day of metformin, within the clinical dose range (He and Wondisford, 2015He L. Wondisford F.E. Cell Metab. 2015; 21: 159-162Abstract Full Text Full Text PDF PubMed Scopus (282) Google Scholar), for 30 days. We observed that, in the liver homogenates of AXINLKO mice, the induction of p-T172-AMPK by metformin administration was virtually abolished but was intact in their wild-type littermates (Figure S1A). We also examined the role of AXIN in AMPK activation in primary hepatocytes and found that depletion of AXIN abolished AMPK activation under treatment of 70 μM metformin (Figure S1B). To test whether failure of AMPK activation by metformin in AXIN-deficient mouse liver was merely due to the absence of the tethering role of AXIN, we asked whether the AXIN-based lysosomal pathway also operates in this process of AMPK activation. Mice with liver-specific knockout of LAMTOR1 (a critical component of Ragulator, LAMTOR1LKO) were treated with metformin in the same way as for AXINLKO mice. It was found that activation of AMPK was fully blocked in these mice (Figure S1C). Consistently, metformin-induced activation of AMPK was abolished in LAMTOR1−/− hepatocytes (Figure S1D). We also examined the role of v-ATPase in metformin-induced AMPK activation in HEK293T cells, a cell line used because of available antibody. We showed that metformin failed to activate AMPK when the v0c subunit of v-ATPase (ATP6v0c) was knocked down (Figure S1E). As MEFs were to be used for performing immunoprecipitation assays and preparation of light organelles and cytosols for in vitro reconstitution experiments, we also examined metformin-induced AMPK activation in various MEFs, showing that AXIN and LAMTOR1 are both needed for metformin-mediated AMPK activation in MEFs (Figures S1F and S1G). It is noteworthy that owing to the reduced expression of OCT1 (the metformin transporter) (Foretz et al., 2014Foretz M. Guigas B. Bertrand L. Pollak M. Viollet B. Cell Metab. 2014; 20: 953-966Abstract Full Text Full Text PDF PubMed Scopus (804) Google Scholar) in MEFs or HEK293T, higher concentrations of metformin are commonly used for activating AMPK in these cells. Taken together, these results indicate that the AXIN-based lysosomal pathway is required for AMPK activation after metformin treatment. To monitor the formation of AMPK-activating complex, we next carried out analysis of the formation of v-ATPase-Ragulator-AXIN/LKB1-AMPK complex in different MEF cells before and after metformin treatment. We found that increased amounts of LAMTOR1 along with the representative V1B2 subunit of v-ATPase and AMPK were co-immunoprecipitated with AXIN in MEFs treated with metformin (Figure S1H). Metformin-promoted complex formation was abolished in MEFs lacking AXIN or LAMTOR1 (Figures S1H and S1I), consistent with AMPK being unable to be activated in these cells. This finding was further supported by immunofluorescent staining assay showing that AXIN failed to colocalize with LAMP2, a lysosome marker, in LAMTOR1−/− MEFs treated with metformin (Figure S1J). Similarly, when ATP6v0c was knocked down, the metformin-enhanced complex formation was severely impaired (Figure S1K). The results above suggest that metformin activates AMPK through promoting the formation of the v-ATPase-Ragulator-AXIN/LKB1-AMPK complex. In addition, RAPTOR and mTOR of the mTORC1 complex were dissociated from v-ATPase/Ragulator after metformin treatment (Figure S1I), indicative of a direct inactivation of mTORC1 by metformin as seen in response to glucose deprivation (Zhang et al., 2014Zhang C.S. Jiang B. Li M. Zhu M. Peng Y. Zhang Y.L. Wu Y.Q. Li T.Y. Liang Y. Lu Z. et al.Cell Metab. 2014; 20: 526-540Abstract Full Text Full Text PDF PubMed Scopus (313) Google Scholar). We next reconstituted the process of metformin-enhanced AXIN binding to v-ATPase-Ragulator by preparation of endosome/lysosome-containing light organelles from AXIN-deficient MEFs and AXIN-containing cytosols from WT MEFs. The purified light organelles were pre-treated for 30 min with 300 μM metformin, within the concentration that can be detected in the liver from mice after administration of a single dose of 50 mg/kg (Foretz et al., 2014Foretz M. Guigas B. Bertrand L. Pollak M. Viollet B. Cell Metab. 2014; 20: 953-966Abstract Full Text Full Text PDF PubMed Scopus (804) Google Scholar), and mixed with AXIN-containing cytosol from untreated MEFs. It was found that AXIN strongly interacted with LAMTOR1 and V1B2 (a representative subunit of v-ATPase) (Figure S1L). As a control, AXIN could hardly interact with the proteins in untreated light organelles. In addition, metformin failed to further promote the interaction between AXIN and LAMTOR1/V1B2 on light organelles isolated from cells pretreated with conA, which is known to inhibit v-ATPase (Figure S1M). These results clearly suggest that v-ATPase acts as the sensor or effector of metformin in AMPK activation. It is of great interest to further define the specific subunit(s) of v-ATPase responsible for metformin binding in the future. Our current study unequivocally demonstrates that metformin activates AMPK through an “active” process, but not a mere consequence of disruption of metabolic processes such as ATP synthesis through the oxidative phosphorylation. Metformin can bring about a state of mimicry of austere nutrient supply, as it can directly act on v-ATPase and promote the translocation of AXIN/LKB1 onto the surface of lysosome to form complex with v-ATPase-Ragulator, ultimately leading to AMPK activation (Figure S1N). Once occupied by AXIN, the v-ATPase-Ragulator complex dissociates Raptor and mTOR thereof, turning off the activity of mTORC1, a master regulator for anabolic pathways (Zhang et al., 2014Zhang C.S. Jiang B. Li M. Zhu M. Peng Y. Zhang Y.L. Wu Y.Q. Li T.Y. Liang Y. Lu Z. et al.Cell Metab. 2014; 20: 526-540Abstract Full Text Full Text PDF PubMed Scopus (313) Google Scholar). Our finding thus demonstrates that metformin not only activates AMPK, but also inactivates mTORC1 through the AXIN/LKB1-v-ATPase-Ragulator pathway, providing valuable molecular insights into how metformin offers a myriad of benefits to users. C.-S.Z., M.L., T.M., S.-Y.L., and S.-C.L. conceived the project and designed the experiments. C.-S.Z., M.L., Y.Z., J.C., J.-W.F., and Y.-Q.W performed cellular experiments, and T.M. performed the animal experiments. C.-S.Z., M.L., T.M., S.-Y.L., and S.-C.L. analyzed the data and wrote the manuscript. We thank all other members of S.-C.L. laboratory for suggestions and technical assistance. This work was supported by grants from the National Key Research and Development Project of China (2016YFA0502001) and National Natural Science Foundation of China (#31430094, #31571214, #31601152, and #J1310027). Download .pdf (1.92 MB) Help with pdf files Document S1. Supplemental Experimental Procedures and Figures S1