Too sweet for autophagy: hexokinase inhibition of mTORC1 activates autophagy

Autophagy, a catabolic process that sequesters and degrades cellular contents, is a vital contributor to cellular energy homeostasis. Autophagy flux increases in response to energy depletion, such as that accompanying starvation, and promotes survival by eliminating damaged organelles and recovering metabolites, including amino acids. Induction of autophagy in response to depletion of ATP, glucose and/or amino acids is directed by the activities of the energy sensing kinase AMPK and the nutrient sensing kinase mTORC1. AMPK is activated in response to a decline in cellular ATP levels, while mTORC1 activity is modulated by amino acid levels through a pathway that depends on Rag GTPases, the Regulator complex and the v-ATPase (Inoki et al., 2012; Bar-Peled et al., 2012). While it is well accepted that glucose withdrawal inhibits mTORC1 activity and in most circumstances promotes autophagy, the relationship between glucose metabolism and autophagy remains poorly characterized. Roberts et al. (2014) elucidate a previously unrecognized direct link between glucose metabolism and autophagy.

The study is based on the unexpected observation that the glycolysis inhibitor 2-deoxyglucose (2-DG) prevents autophagy induction in response to glucose starvation in cardiac mycocytes. 2-DG is a glucose analog in which the 2-hydroxyl group is replaced by hydrogen. 2-DG is phosphorylated by hexokinase, but not further metabolized by glucose-6-phosphate isomerase, thus shutting down glycolysis. It had been anticipated, therefore, that 2-DG treatment would deplete ATP levels, resulting in activation of AMPK and subsequent induction of autophagy via phosphorylation of ULK1 (Kim et al., 2011). In resolving the basis for paradoxical inhibition of autophagy in the setting of reduced glycolysis, the authors illuminated an AMPK-independent role for the enzyme Hexokinase-II in directly regulating mTORC1.

Glucose deprivation typically results in autophagy induction to maintain energy homeostasis. Here the authors show that RNAi-mediated depletion of Hexokinase-II (HK-II) mitigates autophagy induction in response to glucose deprivation. Consistent with this observation, overexpression of HK-II potentiates the autophagy response and promotes cell survival in these nutrient-limited conditions. Notably, these effects are not mediated by AMPK, since activity of this enzyme is not impacted by manipulation of HK-II levels. Instead, the authors demonstrate that HK-II levels directly regulate the activity of mTORC1. The authors show that in glucose-limiting conditions HK-II binds mTORC1 and inhibits its activity, promoting autophagy induction. This direct inhibition of mTORC1 does not depend on the catalytic activity of HK-II. Enigmatically, however, the authors also found that inhibiting HK-II catalytic activity under glucose-replete conditions promotes autophagy. How does HK-II promote autophagy in glucose-limiting conditions, yet facilitate the suppression autophagy in glucose-replete conditions? The authors show that HK-II binding to mTORC1 is inhibited by glucose-6-phosphate (G6P), the substrate HK-II. Thus, inhibition of mTORC1 by HK-II is regulated by the availability of its substrate.

The findings of Roberts et al. (2014) are intriguing for two reasons. First, they elucidate a mechanism where by glucose deprivation directly induces autophagy. This mechanism could be vital for survival of cells and organisms in starvation conditions, for example with acute postnatal interruption of placental nutrients (Efeyan et al., 2012) or in poorly vascularized regions of tumors. Second, and perhaps more importantly, the findings of Robert et al. imply that a mismatch between glucose uptake and entry into the glycolytic pathway, with resultant accumulation of G6P, may release a metabolic brake on mTOR activity, permitting excess glucose to fuel cellular growth (see Figure 1).

One may speculate on the potential importance of the HKII-dependent regulation of mTORC1 activity in certain diseases, such as diabetic cardiomyopathy or glycolytic enzyme deficiencies. It is established that the mTORC1 pathway promotes cell growth by activating protein translation and oxidative phosphorylation while simultaneously suppressing autophagy (Dibble et al., 2013; Morita et al., 2013). Disturbance of HKII-dependent regulation of mTORC1 could upset the normal balance of anabolism and catabolism and drive pathophysiology. For example, in the heart mTORC1 is a key determinant of cardiac hypertrophy that results from chronic increase in hemodynamic load. Hemodynamic stress increases the rate of myocardial glucose uptake beyond the heart’s oxidative capacity, increasing G6P levels and load-induced mTORC1 activation (Sen et al., 2013). Moreover, the regression of hypertrophy that follows mechanical unloading relies on autophagy (Hariharan et al., 2013) and is associated with a decrease in G6P levels and mTORC1 activity (Sen et al., 2013). The findings of Roberts et al. (2014) provide the missing link between G6P, mTORC1 activity and autophagy that may be relevant for the development of diabetic cardiomyopathy, which develops in patients with diabetes and hypertension. G6P isomerase (GPI) deficiency is an autosomal recessive disorder characterized by a non-spherocytichemolytic anemia, neuromuscular dysfunction and mental retardation of variable severity. The mechanism by which GPI deficiency causes symptoms is unknown, although GPI deficient cells have 10-fold increase in G6P levels and associated mTOR activation (Haller et al., 2011). In light of the findings in Roberts et al., it will be interesting to determine whether G6P accumulates in the relevant cell types of patients with GPI deficiency and causes a disturbance in the normal regulation of mTORC1 by HK-II.

G6P sits at the entry point of three key metabolic pathways: glycolysis for immediate ATP production, pentose phosphate pathway for the generation of reducing equivalents, and glycogen synthesis for energy storage (see Figure 1). Therefore, it is intriguing to speculate that flux through the autophagy pathway may be influenced by altering cellular levels of G6P resulting from perturbations in any of these pathways and contribute to the pathogenesis of other diseases associated with dysregulation of autophagy and mTORC1 signaling – including cancer, metabolic diseases, neurological diseases, infection and autoimmunity.