Metformin, independent of AMPK, inhibits mTORC1 in a rag GTPase-dependent manner

Abstract

Dysfunctional mTORC1 signaling is associated with a number of human pathologies owing to its central role in controlling cell growth, proliferation, and metabolism. Regulation of mTORC1 is achieved by the integration of multiple inputs, including those of mitogens, nutrients, and energy. It is thought that agents that increase the cellular AMP/ATP ratio, such as the antidiabetic biguanides metformin and phenformin, inhibit mTORC1 through AMPK activation of TSC1/2-dependent or -independent mechanisms. Unexpectedly, we found that biguanides inhibit mTORC1 signaling, not only in the absence of TSC1/2 but also in the absence of AMPK. Consistent with these observations, in two distinct preclinical models of cancer and diabetes, metformin acts to suppress mTORC1 signaling in an AMPK-independent manner. We found that the ability of biguanides to inhibit mTORC1 activation and signaling is, instead, dependent on the Rag GTPases.

Figure 1. Oligomycin-induced dS6K dephosphorylation does not require dTSC1/2

(A) Alignment of Drosophila TSC2 (accession number NP_524177) with the mouse TSC2, isoform 1 (accession number NP_035777). Percentage of sequence identity and similarity (parentheses) of different regions of TSC2 are indicated. The major AMPK phosphorylation sites in TSC2 reported to be affected by energy depletion are in bold and underlined. The mouse 1271 and 1366 phosphorylation sites correspond to the respective rat 1227 and 1345 sites described by Inoki et al. GAP denotes the GTPase-activating domain (B) Analysis of dS6K activity following treatment of Kc167 cells with 10μM oligomycin for 30min or 100nM insulin for 30min. In samples containing 100nM insulin plus 10μM oligomycin, the inhibitor was added for the indicated time prior to the 30min insulin treatment. 20nM rapamycin was added 15min prior to 30min insulin treatment. Western blot analysis of dS6K T398 phosphorylation and dS6K are as indicated (C) Analysis of dS6K activity following dTsc1, dTsc2 dsRNA treatment in the absence or presence of either 100nM insulin or 10μM oligomycin for 30min. Western blot analysis of dS6K T398 phosphorylation, dS6K activity, dS6K and dTSC1 are as indicated. The data shown are representative of two independent experiments.

Figure 2. TSC2 is not required for downregulation of S6K1 T389 by energy-depleting agents

(A) TSC2^+/+ or TSC2^-/- MEFs were treated with either 10μM oligomycin or (B) 100mM 2DG for the times indicated. (C) TSC2^+/+ or TSC2^-/- MEFs were treated with the indicated concentrations of 2DG for 30min. (D) Western blot analysis of the indicated proteins in TSC2^+/+ and TSC2^-/- MEFs. (E and F) TSC2^+/+ MEFs were treated with either 10μM oligomycin or 25mM 2DG at the indicated times. * denotes the addition of a fresh aliquot of either ATP-depleting agent following a 60min incubation with the initial inhibitor. The data shown are representative of at least three independent experiments.

Figure 3. Metformin and phenformin inhibit mTORC1 signaling independent of TSC2

TSC2^+/+ or TSC2^-/- MEFs were treated with either (A) 10mM metformin for 24h or (B) 6mM phenformin for 1h. (C) TSC2^-/- MEFs seeded at increasing density (1.25 ×10⁴, 5 ×10⁴, or 20 ×10⁴ cells/well) were treated as in A. (D) TSC2^-/- MEFs seeded at 20 ×10⁴ cells/well and treated with 10mM metformin for 24h in either the initial or fresh medium. (E) TSC2^+/-ang1 cells were allowed to grow in soft agar for 3 weeks, with or without 20nM rapamycin or 10mM metformin. Colonies were stained with MTT and counted as described in Experimental Procedures. (F) TSC2^+/-ang1 cells were treated with either 20nM rapamycin for 15min or 10mM metformin for 24h. Western blots were performed as described in Experimental Procedures and probed with the indicated antibodies. The data shown are representative of at least two independent experiments.

Figure 4. AMPK is sufficient to inhibit mTORC1 signaling but it is not required for mTORC1 inhibition by energy depletion

(A) WT indicates HEK293 cells in which α, β, and γ subunits of AMPK were co-expressed for 40h. R70Q and R172Q indicate that the WT γ subunit was substituted with the respective γ mutants. After 24h of ectopic expression, cells were deprived of serum overnight and stimulated with 200nM insulin for 30min. C2C12 myoblasts were treated with either (B) 2mM AICAR or (C) 25mM 2DG for the indicated times. (D) Primary AMPK α1 and α2 double knock-out (AMPK dKO) and AMPK WT MEFS were treated with 2DG for 30min at the indicated concentrations. (E) Immortalized AMPK dKO and AMPK WT cells were starved of glucose for 1h. Western blots were performed as described in Experimental Procedures and using the indicated antibodies. The data shown are representative of at least two independent experiments

Figure 5. Metformin and phenformin, but not AICAR, inhibit mTORC1 independent of AMPK

Western blot analysis of extracts from AMPK dKO or AMPK WT MEFs treated either with (A) 10mM metformin for 24h, (B and D) 6mM phenformin for 1h, or (C) 2mM AICAR for 2h. (E) Serum-starved AMPK dKO or AMPK WT MEFs treated for 1h with phenformin or AICAR were stimulated with insulin for 15 min. Raptor was immunoprecipitated, and in vitro mTORC1 activity was assayed as described previously (Gulati et al., 2008; Sancak et al., 2007). In parallel, western blot analyses were performed on cell lysates. The data shown are representative of multiple independent experiments.

Figure 6. Absence of AMPK does not affect metformin-mediated glucose uptake

(A) L6 myotubes were transfected with either non-silencing dsRNA (NS) or with two different dsRNAs directed against either the α1 or the α2 catalytic subunit of AMPK and incubated for 72 hrs. Cells were then treated with either 5mM metformin for the last 24 hrs or 2mM AICAR for the last 2 hrs. Parallel plates were either extracted for western blotting with the indicated antibodies or (B) further processed for glucose-uptake measurements, as described in Experimental Procedures. Results are representative of three independent experiments conducted in triplicate.

Figure 7. Rag GTPases mediate the inhibition of mTORC1 by Phenformin

(A) HEK293T cells were either starved for serum and AAs for 50 min (-AA) and re-stimulated with AAs for 10min (-AA→+AA), or cells in full media treated with 5mM phenformin for 1hr. Cells were processed as described in Experimental Procedures and immunofluorescence assays were performed to detect mTOR (green) (B) HEK293T cells with PRK5 empty vector, used as a transfection control, or expressing epitope-tagged HA-GST-RagB GTP or WT HA- GST-RagB were deprived of AAs for 1hr, or treated with 5mM phenformin for 1hr. Western blots were performed as described in Experimental Procedures and proteins were detected with the indicated antibodies. These studies are representative of at least two independent experiments (C) Model for metformin/phenformin action on mTORC1.