The kinase LKB1 mediates glucose homeostasis in liver and therapeutic effects of metformin

Abstract

The Peutz-Jegher syndrome tumor-suppressor gene encodes a protein-threonine kinase, LKB1, which phosphorylates and activates AMPK [adenosine monophosphate (AMP)-activated protein kinase]. The deletion of LKB1 in the liver of adult mice resulted in a nearly complete loss of AMPK activity. Loss of LKB1 function resulted in hyperglycemia with increased gluconeogenic and lipogenic gene expression. In LKB1-deficient livers, TORC2, a transcriptional coactivator of CREB (cAMP response element-binding protein), was dephosphorylated and entered the nucleus, driving the expression of peroxisome proliferator-activated receptor-gamma coactivator 1alpha (PGC-1alpha), which in turn drives gluconeogenesis. Adenoviral small hairpin RNA (shRNA) for TORC2 reduced PGC-1alpha expression and normalized blood glucose levels in mice with deleted liver LKB1, indicating that TORC2 is a critical target of LKB1/AMPK signals in the regulation of gluconeogenesis. Finally, we show that metformin, one of the most widely prescribed type 2 diabetes therapeutics, requires LKB1 in the liver to lower blood glucose levels.

Figures

Fig. 1

Requirement of LKB1 for AMPK activation in liver. (A) Immunoblotting of liver and muscle protein lysates. Two weeks after adenoviral Cre injection, organs were collected from the mice of the indicated genotypes and immediately frozen in liquid nitrogen. Proteins from total-cell extracts of liver or muscle were immunoblotted for LKB1, phospho-Thr172 AMPK (P-AMPK), or total AMPKα. (B) Regulation of mTOR signaling. The mTOR kinase directly phosphorylates two effectors: the translational initiation inhibitor 4E-BP1, and the ribosomal S6 kinase (S6K1). mTOR-activated S6K1 then phosphorylates ribosomal S6. Thus, the level of phosphorylation of 4EBP1, S6K1, and S6 reflect the level of mTOR activation within the cell. Two weeks after adenoviral Cre injection, mice of the indicated genotypes were fasted for 18 hours overnight or fed ad libitium (ad lib) and then killed. Total-cell extracts were made from liver or muscle and immunoblotted with indicated antibodies to examine AMPK activation and mTOR activation in wild-type and LKB1-deficient livers after an 18-hour fast or under ad lib fed conditions, as indicated.

Fig. 2

Glucose homeostasis defects in mice lacking LKB1 in liver. (A) Mice of the indicated genotypes were fasted for 18 hours at indicated times after adenoviral Cre injection, and fasting blood glucose was measured. P < 0.01 at all time points. (B) Glucose-tolerance test (GTT) on mice of indicated genotypes 2 weeks after adenoviral Cre injection. (C) Insulin-tolerance test (ITT) on mice of indicated genotypes 2 weeks after adenoviral Cre injection. No significant difference was observed. Data represents the mean + SEM for six mice of each genotype. Average T0 glucose levels for the wild-type mice = 180 mg/dl; average T0 glucose levels for the L/L mice = 355 mg/dl.

Fig. 3

Gluconeogenic and lipogenic gene expression is elevated in LKB1-deficient livers. (A) The left panel shows qRT-PCR examining the expression levels of mRNA for indicated gluconeogenic genes from the livers of mice of indicated genotypes and conditions 3 weeks after Cre administration. Expression was normalized to hypoxanthine-guanine phosphoribosyl transferase (HPRT) and equilibrated to the lowest-value condition for each gene. The right panel shows immunoblot analysis of PGC-1α or eIF4E (loading control) from fasted mice of indicated genotypes 6 weeks after Cre injection. (B) qRT-PCR for lipogenic target genes on liver samples as in (A). Mice were either fasted or fed ad lib. The induction of lipogenic targets is lower under ad lib fed conditions than for mice re-fed after fasting conditions (49). qRT-PCR data represent the mean + SEM for samples analyzed in triplicate.

Fig. 4

Decreased phosphorylation of the TORC2 transcriptional coactivator and increased gluconeogenesis in the LKB1-deficient livers. (A) Requirement of LKB1 for TORC2 phosphorylation at Ser171 and associated mobility shift on SDS-polyacrylamide gel elecrophoresis. (Left) Wild-type (WT) or catalytically-inactive (KD) LKB1 was cotransfected with FLAG-tagged wild-type or Ser171 Ala mutant TORC2 into −/− HeLa cells. Total cell lysates were immunoblotted with antibodies to LKB1 or FLAG. The correlation between the mobility-shifted TORC2 and phosphorylation at Ser171 has also been previously described (37). (Right) Immunoblot of endogenous TORC2 protein from liver extracts of (ad lib fed) mice of indicated genotypes 2 weeks after adenoviral Cre injection. eIF4E levels were examined as a loading control. (B) Immunohistochemistry of endogenous TORC2 in livers of indicated genotypes analogous to those extracted and immunoblotted in (A). Hoechst dye stains only nuclear DNA; thus, by merging the anti-TORC2 localization with the Hoechst, we observed the presence of TORC2 in the nucleus. In wild-type mice, the Hoescht dye and TORC2 staining are mutually exclusive, whereas in LKB1-deficient livers, there is near complete overlap in the nucleus. (C) TORC2 shRNA in liver reduces TORC2 and PGC1α protein levels. Adenovirus encoding TORC2 shRNA or a control scrambled shRNA (US ctl) was introduced by tail-vein injection into L/L mice that had been tail-vein injected with adenoviral Cre 2 weeks earlier. Five days after adenoviral shRNA injection, mice were fasted for 18 hours. Total-cell lysates were made from the liver and immunoblotted with indicated antibodies. (D) Blood glucose levels in Cre-injected L/L mice reduced by TORC2 shRNA. Adenovirus shRNA was administered as above and fasting glucose levels were monitored. Data represent the mean + SEM for five mice of each group. P < 0.01, t test.

Fig. 5

LKB1 in liver is required for the ability of metformin to lower blood glucose. (A) LKB1 is required for metformin to activate AMPK in liver in mice. 8-week-old +/+ or L/L littermate male mice were injected in the tail vein with adenoviral Cre. Two weeks after Cre administration, mice were injected intraperitoneally with 250 mg metformin per kg body weight (mg/kg) in saline or just saline for 3 consecutive days. On the third day, total-cell extracts were made from collected livers 1 hour after metformin administration. Liver extracts were immunoblotted with antibody to phospho-Thr172 AMPK or total AMPK antibodies. (B) LKB1 in liver is required for the ability of metformin to lower blood glucose. 8-week-old +/+ or L/L littermate male mice were injected in the tail vein with adenoviral Cre and placed on a high-fat diet (55% fat, 24% carbohydrate, Harlan Teklad, Madison, WI) for 6 weeks. 8-week-old ob/ob male mice were obtained from Jackson Laboratory. Mice were fasted for 18 hours, then blood glucose was measured. Starting the next day, mice were injected intraperitoneally with 250 mg/kg metformin in saline or just saline for 3 consecutive days. On the third day, mice were fasted overnight and blood glucose was measured. Data represent the mean + SEM for five mice of each group. *P < 0.001, t test.