Regulation of the TSC pathway by LKB1: evidence of a molecular link between tuberous sclerosis complex and Peutz-Jeghers syndrome

Abstract

Tuberous sclerosis complex (TSC) and Peutz-Jeghers syndrome (PJS) are dominantly inherited benign tumor syndromes that share striking histopathological similarities. Here we show that LKB1, the gene mutated in PJS, acts as a tumor suppressor by activating TSC2, the gene mutated in TSC. Like TSC2, LKB1 inhibits the phosphorylation of the key translational regulators S6K and 4EBP1. Furthermore, we show that LKB1 activates TSC2 through the AMP-dependent protein kinase (AMPK), indicating that LKB1 plays a role in cell growth regulation in response to cellular energy levels. Our results suggest that PJS and other benign tumor syndromes could be caused by dysregulation of the TSC2/mTOR pathway.

Figure 1.

LKB1 inhibits the phosphorylation of S6K and 4EBP1. (A) LKB1 decreases the activation status of coexpressed S6K. HA-S6K and Flag-LKB1 were cotransfected into HEK293 cells, and the phosphorylation of S6K was monitored with a phospho-specific antibody for the S6K activation site, T389. The level of HA-S6K in cell lysates was determined by immunoblotting with an anti-HA antibody, and LKB1 expression was monitored with an anti-Flag antibody. (B) LKB1 decreases the phosphorylation of 4EBP1. Flag-4EBP1 and Flag-LKB1 were cotransfected into HEK293 cells, and Flag antibody was used to immunoprecipitate (IP) 4EBP1 and LKB1. The activation state of 4EBP1 was monitored with phospho-specific antibodies against S65 and T37/41. Levels of 4EBP1 and LKB1 were monitored via an anti-Flag immunoblot (IB). (C) LKB1 kinase activity is required for full inhibition of S6K. HA-S6K was cotransfected into HEK293 cells with either wild-type LKB1 or one of three kinase inactive mutants: D194A, a catalytic mutant; K81R, an ATP loading mutant; and Dbl, a double mutant incorporating both mutations. Kinase inactive mutants of LKB1 were unable to significantly inhibit phosphorylation of S6K on T389. Quantitation was performed on the sample with higher expression for each construct.

Figure 2.

LKB1 functions through AMPK to inhibit S6K. (A) Down-regulation of LKB1 with siRNA causes an increase in endogenous S6K activation. LKB1 siRNAs were transfected into HEK293 cells, and the activation status of endogenous S6K was monitored. Two oligonucleotides were tested: LKB1iSP is a pool of several siRNAs, and LKB1icust is a single, custom-designed oligonucleotide. Both LKB1icust and LKB1iSP were able to cause a decrease in the endogenous LKB1 protein and an increase in S6K activation on T389. Expression levels of LKB1 and S6K were monitored by Western analysis with the indicated antibodies. The high concentration of LKB1 antibody used in this figure (1:5000) detects a nonspecific band (n.s. band) just above the LKB1 protein (as previously described in Bardeesy et al. 2002). (B) An AMPK inhibitor blocks the inhibitory effect of LKB1 on S6K. Flag-LKB1 and HA-S6K were cotransfected into HEK293 cells and treated with 40 μM compound C, an AMPK inhibitor (AMPKi), as indicated. Immunoblots were performed with the indicated antibodies. (C) An AMPK dominant-negative construct blocks the inhibitory effect of LKB1 on S6K. HA-S6K expressed in HEK293 cells was immunoprecipitated by anti-HA antibody, and the activation status of S6K was monitored with phospho-S6K antibody. The levels of S6K in immunoprecipitants were monitored by anti-S6K antibody. The levels of Flag-LKB1, HA-S6K, and HA-AMPK-DN (which is a weak band just above HA-S6K) in lysates are also shown.

Figure 3.

LKB1 inhibits the phosphorylation of S6K in HeLa cells and LKB1^-/- MEFs. (A) LKB1 inhibits S6K phosphorylation in HeLa cells and LKB1^-/- MEFs under basal conditions. HeLa cells and LKB1^-/- MEFs were infected with either HA-LKB1 or empty vector-containing retroviruses. Phosphorylation of endogenous proteins was monitored with the indicated phospho-specific antibodies. LKB1 expression was confirmed by HA-antibody. (B,C) LKB1 enhances ATP depletion-induced S6K dephosphorylation. Vector or HA-LKB1-expressing HeLa cells (B) and LKB1^-/- MEF cells (C) were treated with various concentrations (1–25 mM) of 2-DG for 10 min. Phosphorylation of endogenous proteins was determined by the indicated phospho-specific antibodies. LKB1 expression was monitored with anti-HA antibody. The exposures selected in B and C were chosen to best illustrate the energy depletion effect on S6K, AMPK, and 4EBP1 phosphorylation; therefore, the Western blots of LKB1-expressing cells were exposed longer than the LKB1-null cells to compare similar basal signal levels.

Figure 4.

LKB1 functions through TSC2. (A) A kinase inactive LKB1 construct, LKB1-D194A, causes a downshift of cotransfected TSC2. LKB1 or LKB1-D194A was coexpressed in HEK293 cells with HA-TSC2 and Myc-TSC1. Mobility shift assay of HA-TSC2 was performed by running lysates on 6% SDS-PAGE followed by immunoblotting with anti-HA antibody. (B) LKB1-D194A causes a down-shift of coexpressed wild-type HA-TSC2 but not of HA-TSC 2A, a mutant of TSC2 that eliminates the AMPK phosphorylation residues T1227 and S1345. The experimental design is the same as in A. (C) LKB1-D194A affects T1227 and S1345 phosphorylation on TSC2. HA-TSC2 and Myc-TSC1 were coexpressed with LKB1 or LKB1-D194A in HEK293 cells. In vivo labeling was performed, and immunoprecipitated HA-TSC2 was subjected to two-dimensional phospho-peptide mapping. Phosphorylation of the shaped spots depicted in panel V were enhanced by 2-DG treatment. Phosphorylation of the corresponding spots were decreased by LKB1-D194A (panels I,VI) and eliminated by (2A) T1227A/S1345A mutations (panel IV). Panel VII represents a schema for indicating AMPK-dependent phospho-peptide changes (also depicted in panels IV and V; Inoki et al. 2003). (D) TSC2 is required for LKB1 to fully inhibit S6K. TSC2 was down-regulated in HEK293 cells by RNA interference in the presence or absence of cotransfected LKB1. The activation status of HA-S6K was monitored on T389. The levels of S6K, LKB1, and endogenous TSC2 were monitored via immunoblot with the indicated antibodies.

Figure 5.

Cells null for TSC2 or LKB1 display similar phenotypes for certain biological responses. (A) TSC2 and LKB1 play essential roles in protecting cells from glucose depletion-induced cell death. LEF cells (TSC2^-/-) stably expressing empty vector or TSC2, and MEF cells (LKB1^-/-) stably expressing empty vector or LKB1 were cultured in 25 mM glucose (glucose+) or glucose-free medium (glucose-). Photos were taken after 72 h in culture for LEF cells and 48 h for MEF cells. (B) Rapamycin inhibits glucose depletion-induced caspase-3 activation in LKB^-/- MEF cells. MEF (LKB^-/-) cells stably expressing vector or LKB1 were cultured for 30 h in media containing various concentrations of glucose (1–25 mM) in the presence or absence of rapamycin (20 nM). Immunoblots of caspase-3, cleaved caspase-3, and actin are shown. (C) Rapamycin inhibits VEGF secretion in LKB1^-/- MEF cells. Equal numbers of LKB^-/- MEF cells stably expressing empty vector or LKB1 were cultured for 36 h with or without rapamycin. The concentration of secreted VEGF in the medium was determined by ELISA. Data are expressed as a mean, and one standard deviation is indicated by the error bars (n = 3). The asterisk indicates that the reduction in VEGF secretion induced by rapamycin in LKB1-null cells is significant, as well as the increased expression of VEGF in the LKB1-null cells relative to the reexpressed cells (p < 0.0001). (D) Rapamycin inhibits VEGF expression in LKB1^-/- MEF cells. Equal numbers of MEFs stably expressing empty vector or LKB1 were cultured for 36 h with or without rapamycin. VEGF expression in cell lysates was monitored by VEGF immunoblot, and protein level was monitored by β-tubulin immunoblot. (E) A proposed model for a role of LKB1 in the TSC pathway. Loss of LKB1 is a hallmark of Peutz-Jeghers syndrome. Less LKB1 would cause a decrease in the activation status of AMPK on T172 and of TSC2 on S1227 and S1335, relieving TSC2's inhibitory effect on Rheb. Higher Rheb activity would subsequently drive mTOR activation. Two key targets of mTOR—S6K and 4EBP1—are shown. We propose a model in which dysregulated increases in the activity S6K, 4EBP1, and other mTOR targets could lead to hamartoma formation in PJS in an analogous way to that of TSC. In a similar way, hamartoma-causing mutations in PTEN affect the activation status of AKT, a molecule that also acts on TSC2. Thus, the TSC2/mTOR pathway could represent a common pathway for hamartoma formation. Pointed arrowheads indicate activation, and flat arrowheads indicate inhibition.